The Fibromyalgia Syndrome-Diagnosis Pathophysiology PowerPoint Presentation

Slide 1 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center

Slide 2 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic

Slide 3 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999)

Slide 4 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999)

Slide 5 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36

Slide 6 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 7 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be

Slide 8 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain

Slide 9 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change

Slide 10 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90%

Slide 11 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data

Slide 12 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject

Slide 13 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper

Slide 14 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures

Slide 15 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper

Slide 16 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration.

Slide 17 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 18 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points)

Slide 19 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM

Slide 20 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed

Slide 21 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function

Slide 22 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ

Slide 23 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC

Slide 24 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances

Slide 25 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms

Slide 26 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms

Slide 27 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances

Slide 28 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening.

Slide 29 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase.

Slide 30 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V)

Slide 31 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments.

Slide 32 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092)

Slide 33 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care.

Slide 34 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01

Slide 35 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders

Slide 36 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 37 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD

Slide 38 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms

Slide 39 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA

Slide 40 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR

Slide 41 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk

Slide 42 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts.

Slide 43 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns).

Slide 44 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity

Slide 45 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * *

Slide 46 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic

Slide 47 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls.

Slide 48 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring

Slide 49 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 50 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 51 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 52 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 53 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging

Slide 54 -

Outcome Measures in Fibromyalgia Daniel J. Clauw, MD Professor of Medicine, Division of Rheumatology Director, Chronic Pain and Fatigue Research Program University of Michigan Medical Center Chronic Pain – Defined by Mechanisms Peripheral (nociceptive) Primarily due to inflammation or damage in periphery NSAID, opioid responsive Behavioral factors minor Examples OA Acute pain models (e.g. third molar, post-surgery) RA Cancer pain Central (non-nociceptive) Primarily due to a central disturbance in pain processing Tricyclic responsive Behavioral factors more prominent Examples Fibromyalgia Irritable bowel syndrome Tension and migraine headache Interstitial cystitis / vulvodynia, non-cardiac chest pain / etc. Mixed Neuropathic Effect Sizes of Various Treatments in Fibromyalgia - I (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia - II (Rossy et. Al. Ann Behav Med 1999) Effect Sizes of Various Treatments in Fibromyalgia – III (Arnold et. al. Psychosomatics 2000) Pooled effects sizes of nine trials of tricyclics Sleep .69 Physician global .64 Pain .57 Fatigue .52 Patient global .50 Tenderness .36 Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Visual Analog Scale No Pain Pain as bad as it could be Problems with Current Methods of Pain Measurement VAS not a good measure from reliability, validity standpoint Anchor is something that patient may have never experienced (i.e. worst pain imaginable) Scaling problems - Patients only use a portion of scale, and different portions of the scales, no linearity to scale VAS only captures a single dimension of pain experience Multidimensional scales e.g. McGill Pain Questionnaire Problems with retrospective report of any symptom, and lack of accounting for variability in pain over time Miss other important domains of pain that may be as important to outcomes as intensity of pain Better scales Add verbal anchors so that choosing a point in scale is not an exercise in imagination and fractionation Make scale logarithmic so that a wider range can be used, and so that each interval in scale represents the same magnitude of change Patient Compliance with Paper and Pencil Diaries Stone et. al. (BMJ 2002) performed study of 80 chronic pain pts. over 21 days, asking them to record both paper and electronic entries of pain levels Unbeknownst to subjects, there was microchip imbedded into paper diaries that could tell when diary was opened Pts. recorded 89% compliance with entries within 30 minute window, yet actual compliance was 11% On 32% of days binder was not opened, yet compliance recordings for those days averaged 90% Patient Experience Diary Prompts subjects at any pre-determined interval to answer any number of questions When device is placed in cradle each night modem downloads information to central location InVivo Data Example of data from single fibromyalgia subject Conclusions I – Various Pain Measures Less frequent sampling of pain leads to increased baseline scores (Cypress Phase II data; n=125) 15.0 Baseline score (0-20) Assessments used 50 14 2 1 Recall interval (days) 0 1 7 7/30 RP Daily Weekly Weekly/ Paper Comparison of RP to Weekly Diary Pain Sampling Measures Conclusions II – Various Pain Measures Less frequent sampling of pain leads to increased endpoint scores, BUT not to the same degree as what is seen at baseline Change from baseline (clinical improvement, all patents) 2.0 2.4 2.4 3.7 11.3 Endpoint score RP Daily Weekly Weekly/ Paper Summary Random prompt pain is extremely variable in fibromyalgia patients, much moreso than paper values The clinic (paper) weekly and monthly pain values were higher than random prompt values from the same days and weeks, with an average increase of over 4 units (0-20 scale) at the beginning of the trial, and 2 units at the end. We speculate that elevated initial scores on the paper clinic assessments may relate to anxiety, initial lack of familiarity with the assessment scales, and/or demand characteristics. As an artificially elevated baseline value would affect interpretation of all later results during an interventional trial, this observation merits further explanation and consideration. Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging The Neurobiological / Psychobehavioral Continuum Neurobiological factors Abnormal sensory processing Autonomic dysfunction HPA dysfunction ? Peripheral factors Psychobehavioral factors General “distress” Cognitive factors Psychiatric comorbidities Maladaptive illness behaviors Secondary gain issues Population Primary Care Tertiary Care Definition factors (e.g., tender points) Symptoms Psychological and Behavioral Consequences Decreased activity Poor sleep Increased distress Maladaptive illness behaviors Interaction between Symptoms and Function in FM Potential Functional Status Measures in FM Fibromyalgia Impact Questionnaire (FIQ) Able to: do shopping, do laundry, prepare meals, wash dishes by hand, vacuum a rug, make beds, walk several blocks, visit friends and relatives, do yard work, drive a car Seven VAS measuring how much pain interfered with job, pain, tiredness, restedness upon awakening, stiffness, tense/anxious, depressed Problems with “Floor Effect” for FIQ Results of FIQ and SF-36 PCS at end of milnacipran Phase II trial Lower number = higher function Higher number = higher function Potential Functional Status Measures in FM SF-36 Physical Component Summary (PCS) Score (physical functioning, role limitations due to physical problems, bodily pain, general health and vitality) Health Assessment Questionnaire Fibromyalgia HAQ Effectiveness of Aerobic Exercise and Cognitive Behavioral Therapy in Chronic Multisymptom Illnesses: Results from CSP #470 Sam Donta 3, Daniel J Clauw 1, Charles C Engel 2,, Andre Barkhuizen 4, James S Skinner 5, Peter Peduzzi 6, Peter Guarino 6, David A Williams 1, Thomas Taylor 7, Lew Kazis 8, John R Feussner 9, and the CSP #470 Study Group 1University of Michigan, Ann Arbor, MI;2Walter Reed Army Medical Center, Washington, DC;3VAMC Boston, Boston, MA;4Oregon Health Sciences University, Portland, OR;5Indiana University, Indianapolis, IN;6VA Cooperative Trials Coordinating Center, West Haven, CT;7White River Junction VA, White River, VT;8VAMC Bedford, Bedford, MA;9VA Research and Development, Washington, DC The U.S. Gulf War Experience - I In 1990 and 1991, over 700,000 U.S. troops were deployed to the Persian Gulf Although there were very few combat casualties, within months of returning from the war many soldiers were complaining of illnesses The primary symptoms seen were joint and muscle pain, headaches, fatigue, difficulties with memory, rash, and gastrointestinal disturbances The U.S. Gulf War Experience - II After all of this research, several facts are now clear: The symptoms that Gulf War veterans suffer from represent the same clusters of symptoms that occur in the general population, and go by names such as fibromyalgia, chronic fatigue syndrome, somatoform disorders These symptoms are indeed more common in Gulf War veterans, but in fact have been seen in veterans of every war that the U.S. has ever been involved in No specific exposures (except a single study implicating vaccines) have been shown to lead to this constellation of symptoms Chronic Multi-symptom Illnesses (CMI) Term coined by the CDC in 1999 to describe multiple somatic symptoms in Gulf War veterans (Fukuda et. al. JAMA 1999) This study and subsequent studies in the general population using factor analytic techniques (e.g., Doebbling et. al. Am J Med 2000) identified 3 – 4 symptom factors that cluster in the populations Multifocal pain Fatigue Cognitive difficulties Psychological symptoms This and subsequent studies demonstrated that approximately 10 – 15% of the population suffers from a syndrome characterized by two or more of these symptoms Overlap between Fibromyalgia and Other “Systemic” Syndromes: Chronic Multi-symptom Illnesses FIBROMYALGIA 2 - 4% of population; defined by widespread pain and tenderness EXPOSURE SYNDROMES e.g. Gulf War Illnesses, silicone breast implants, sick building syndrome CHRONIC FATIGUE SYNDROME 1% of population; fatigue and 4/8 “minor criteria” SOMATOFORM DISORDERS 4% of population; multiple unexplained symptoms - no organic findings MULTIPLE CHEMICAL SENSITIVITY - symptoms in multiple organ systems in response to multiple substances Inclusion criteria To be eligible veterans had have been deployed to the Gulf War between August 1990 and August 1991, and to endorse > 2 of the following symptoms: fatigue limiting usual activity pain in > 2 body regions neurocognitive symptoms These symptoms had to begin after August 1990, last for more than six months, and be present at the time of screening. Subjects / Methods 1092 veterans who satisfied the eligibility criteria and gave written informed consent were randomized to one of four treatment arms: 1) CBT alone, 2) exercise alone, 3) CBT + exercise, or 4) usual care. Both CBT and exercise were delivered in groups of three to eight participants. CBT Treatment sessions were 60-90 minutes long and met weekly for 12 weeks. Exercise prescriptions focusing on low impact exercise were individualized for each participant after they performed a submaximal cycle ergometer exercise test at baseline. Veterans in the exercise group were asked to exercise once/wk in the presence of the exercise therapist, and 2 – 3X / wk independently during the 12-week treatment phase. Outcome measures Treatments were given for three months using standard protocols and participants were evaluated at baseline, 3, 6 and 12 months. The primary endpoint was the proportion of participants who improved more than 7 units on the physical component summary scale of the Veterans Short Form 36-item (SF-36) Health Survey at 12 months after randomization. Secondary outcomes were standardized measures of: Pain (McGill Pain Questionnaire) Fatigue (Multidimensional Fatigue Inventory) Cognitive symptoms (Cognitive Failures Questionnaire) Distress (Mental Health Inventory – 5 of the SF-36V) Mental health functioning (Mental component score of the SF-36V) Results – Demographics of Participants 85% male Mean age 40.7 81% presented with all three cardinal symptoms of GWVI at the time of screening The mean duration of symptoms was 6.7 years Based on the Prime MD: 45% percent of veterans had either a major depressive disorder or dysthymia, 35% had an anxiety disorder 43% had posttraumatic stress disorder 24% percent of veterans had a pending disability claim and 42% were receiving disability payments. Physical Component Summary (PCS) of the SF-36 50.1 Healthy Normals (n=2,329) Mean Score 43.7 Type II Diabetes (n=123) 38.3 Congestive Heart Failure (n=69) 28.7 Vets FMS (n=4,195) Hypertension (n=816) Myocardial Infarction (n=50) 45.6 42.7 Kazis, (1999; P.C.); Ware, Kosinski, Keller, 1995 33.7 Vets GWI (n=1092) Response to Treatment There was a modest difference in the proportion of veterans who reported an improvement in physical function at one year among the CBT groups: 11.5% for usual care 11.7% for exercise 18.4% for CBT 18.5% for CBT + exercise More significant improvements in fatigue, cognitive symptoms, distress, and mental health functioning were observed with exercise alone, and with exercise plus CBT compared to usual care. Correlations between changes in outcome measures in CMI Correlation of change in symptoms with change in PCS score (12 months to baseline) Pain .34 General fatigue .40 Physical fatigue .42 Cognitive dysfunction -.35 MCS .01 Conclusions This cohort with CMI had extremely low levels of self-report function, like other cohorts with FM CBT specifically aimed at improving physical function had only a marginally significant impact on self-reported physical function for veterans with GWVI. Exercise, with or without CBT, resulted in improvement in fatigue, cognitive symptoms, distress and mental health functioning. There were no additive or synergistic effects between the two treatments. This and other studies suggest weaker correlations between improvements in symptoms (e.g. pain, fatigue, etc.) and improvement in function in FM than in other rheumatic disorders Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Objective - Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging How do Fibromyalgia Patients Really “Function”? Angela Lyden, M.S.1, Ali Berlin2, Kirsten Ambrose, M.S.1, Willem J. Kop, Ph.D.2, Daniel J. Clauw, M.D.1 1Chronic Pain and Fatigue Research Program University of Michigan, Ann Arbor, MI 2Department of Medical and Clinical Psychology Uniformed Services University of the Health Sciences, Bethesda, MD Relationship between symptoms, self-reported, and objective measures of activity Patients with FM have amongst the lowest self-reported activity levels of any chronic illness This parameter has been very difficult to improve in interventional studies How is self-reported activity related to: Objective measures of activity Specific symptoms Actigraphy Designed for long-term monitoring of gross motor activity Omnidirectional wristwatch-like device Accelerometer monitors the occurrence and degree of motion; sensor integrates information to produce an electrical current of varying magnitude Greater the degree = higher voltage Sensitive device, although not specific Results highly correlated with actual physical activity in most settings, including modest correlation with activity in RA Actiwatch®-Score Mini-Mitter Co., Inc. Bend, OR Actogram I Running Swimming Office work-desk Walking Preparing dinner Couch sitting; reading In bed; reading Sleeping Got up Getting ready Walking Office work-desk Methods / Subjects Thirty patients with FM (mean age=41.5) were compared with 29 control participants (mean age=38.9) not engaging in high-exercise activities. Actigraphs were worn for 5 consecutive days and four consecutive nights. Activity levels were sampled over 5 min epochs. Participants rated symptoms ("pain", "tired", "stressed") on 10-point scales 5 times/day based on actigraph-driven alerts. Results - Activity Average daytime and nighttime activity levels were nearly identical in the patient and the control groups (p=ns). Peak activity was significantly lower in the patient group relative to the control group (p=0.008). 7870  3223 vs. 12178  7862 activity units Variability of peak activity was significantly different between groups Levene’s test on SDs, p=0.001 Peak Activity Average and Diurnal Peak Activity Levels of Fibromyalgia Compared to Controls *p<0.05; Error Bars=SEM * * * * Actogram II FM patient Days of higher activity followed by days of less activity Control Higher peak activity, less sporadic Relationship of Activity to Symptoms Peak and average ratings of pain, fatigue and stress were higher in the patient group relative to the control group, but these symptoms were not related to activity in either patient or control groups. Actigraphy results (average or peak) were not significantly correlated with self-report function (SF-36) in either patients or controls. Conclusions – Function in FM FM patients rate their function as being very low This domain has been the most difficult to improve in clinical trials Dysfunction in FM patients is fundamentally different than dysfunction in other rheumatic diseases . . . there is less of a relationship between improvements in symptoms and improvements in function in FM It is not clear what these self-report measures of function are actually measuring Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Potential Outcome Measures in FM Pain Type of scale Measured how? Functional status Subjective Activity monitoring Patient global improvement Other symptoms Fatigue Sleep Cognitive symptoms Process / surrogate outcome measures Evoked pain Functional imaging Options for Primary Outcome in FM Trials Where should we set the bar? This is a legitimate syndrome with a large unmet therapeutic need (RA or OA 30 years ago), thus improvement in a single defining domain is adequate (Migraine, IBS, Depression precedents) USE 2003 RCT STANDARDS (ITT, Responder) PAIN Primary Outcome Must make sure that we are not making “whole person” worse while making their pain better DEFINE RESPONDER BY BOTH 1) MCID IMPROVEMENT IN PAIN AND 2) IMPROVEMENT IN PT. GLOBAL Use standards being proposed for analgesics (drawn from experiences in nociceptive pain disorders, diseases with effective therapies), that propose that there is improvement in 1) pain, 2) global well being, and 3) “function” MUST HAVE VERY LIBERAL INTERPRETATION OF “FUNCTION” AND LOW EXPECTATIONS FOR IMPROVEMENT (i.e. allow use of > measure because of floor effect of FIQ) AND/OR CONSIDER ACR-20 TYPE OF COMPOSITE (RECOGNIZE PROBLEMS WITH THIS PRECEDENT VIS-A-VIS DRUGS THAT MIGHT WORK EXTREMELY WELL IN A SINGLE SYMPTOM DOMAIN)