Original Article
Arthritis Care & Research DOI 10.1002/acr.22375
EXERCISE THERAPY REDUCES PAIN SENSITIVITY IN PATIENTS WITH KNEE OSTEOARTHRITIS: A RANDOMIZED CONTROLLED TRIAL Marius Henriksen PhD1, Louise Klokker MSc1, Thomas Graven-Nielsen DMSc2, Cecilie Bartholdy PT1, Tanja Schjødt Jørgensen MSc1,2, Elisabeth Bandak MSc1, Bente DanneskioldSamsøe DMSc1, Robin Christensen PhD1, Henning Bliddal DMSc1,2 1 Clinical
Motor Function Laboratory, The Parker Institute, Department of Rheumatology, Copenhagen University Hospitals Bispebjerg and Frederiksberg, Copenhagen, Denmark 2 Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology, Faculty of Medicine, Aalborg University, 9220 Aalborg, Denmark. Original paper for: Arthritis Care & Research Correspondence: Marius Henriksen, PT, PhD Clinical Motor Function Laboratory The Parker Institute, Department of Rheumatology Copenhagen University Hospital at Frederiksberg Ndr. Fasanvej 57 DK-2000 Copenhagen F Denmark Tel: +45 3816 4160 Fax: +45 3816 4159 email: [email protected] Word count: 3,796 Running Head: Exercise, pain sensitivity and knee osteoarthritis
FUNDING This study was supported by a grants from The Danish Council for Independent Research | Medical Sciences (Grant number: 10-093704), Danish Physiotherapists Association, The Lundbeck Foundation, and The Oak Foundation to cover salaries and running costs.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an ‘Accepted Article’, doi: 10.1002/acr.22375 © 2014 American College of Rheumatology Received: Jan 13, 2014; Revised: Apr 28, 2014; Accepted: May 27, 2014
Arthritis Care & Research
ABSTRACT Background: Exercise has beneficial effects on pain in knee osteoarthritis (OA), yet the underlying mechanisms are unclear. The purpose of this study was to investigate the effects of exercise on pressure pain sensitivity in patients with knee OA. Methods: In a randomized controlled trial participants were assigned to 12 weeks of supervised exercise therapy (ET, 36 sessions), or a no attention control group (CG). Pressure pain sensitivity was assessed by cuff pressure algometry on the calf of the most symptomatic leg. The co-primary outcomes were 1) pressure pain thresholds (PPT), and 2) cumulated visual analogue scale pain scores during constant pressure for 6 minutes at 125% of the PPT as a measure of temporal summation (TS) of pressure pain. Secondary outcomes included self-reported pain using the KOOS questionnaire. Analyses were based on the ‘Per-Protocol’ population (participants following the protocol). Results: 60 participants were randomized (31 ET/29 CG), and the Per-Protocol population included 48 participants (25 ET/23 CG). At follow-up mean group differences in the change from baseline were 3.1 kPa (95% CI: 0.2 to 6.0; P=0.038) for the PPT, 2,608 mm*s (95%CI: 458 to 4,758; P=0.019) for TS, and 6.8 points (95% CI 1.2 to 12.4; P=0.018), for KOOS pain, all in favor of ET. Conclusion: Pressure pain sensitivity, temporal summation and self-reported pain are reduced among patients completing a 12-week supervised exercise program compared to a no-attention control group. These results demonstrate beneficial effects of exercise on basic pain mechanisms, and further exploration may provide basis for optimized treatment. Clinicaltrials.gov: NCT01545258
2 John Wiley & Sons, Inc.
Page 2 of 43
Page 3 of 43
Arthritis Care & Research
SIGNIFICANCE AND INNOVATION This randomized study is the first to evaluate the effect of exercise on pain sensitivity in knee osteoarthritis (OA). A 12-week supervised exercise program reduces the pressure pain sensitivity compared to a no attention control group, in adjunct to an effect on self-reported pain. This study suggests that exercise therapy has beneficial effects on basic pain mechanisms, and further exploration of this effect may provide basis for optimized treatment.
3 John Wiley & Sons, Inc.
Arthritis Care & Research
Osteoarthritis (OA) is the most frequent painful musculoskeletal disease (1), with a massive socioeconomic burden (2). Because there is no cure, clinical management of OA is an enormous challenge. The ultimate treatment is surgical joint replacement but conservative treatments are recommended before surgery (3). A highly recommended treatment for OA is exercise (4;5), with well-proven beneficial effects on pain and disability (6). Although exercise is more cost-effective than surgery (5), the effect sizes of exercise are moderate and optimization of the exercise paradigm is needed. While the link between exercise and reduced disability is mediated by e.g. increased muscle strength and endurance (7), the mechanisms linking exercise therapy and pain reduction remain to be explored. It has recently been shown that knee OA patients may have both peripheral and central sensitization of pain mechanisms resulting in spreading hyperalgesia (8). Thus, targeted pain treatment may focus on both peripheral and central mechanisms. Presently it is not known if or to what extent exercise affects these mechanisms in patients with knee OA. Recent clinical trials have shown that strengthening exercises of muscles remote to the knee (isolated hip muscle strengthening) caused reductions in knee OA pain over 8-12 weeks (9;10). One suggested mechanism behind the clinical effects of exercise is biomechanical optimization, however, in a number of studies no effects of exercise on knee biomechanics have been reported (11-15). Another possible mechanism is that exercise leads to widespread adaptations in the central nervous system resulting in reduced pain sensitivity. Joint nociception may cause sensitization of pain mechanisms in the central nervous system (16). In addition, peripheral sensitization includes structural and functional changes in the peripheral nociceptors, e.g. caused by inflammation (16). In systemic low-grade inflammatory diseases (e.g. diabetes) anti-inflammatory effects of regular exercises have been suggested
4 John Wiley & Sons, Inc.
Page 4 of 43
Page 5 of 43
Arthritis Care & Research
(17), which in turn may lead to reduced peripheral sensitization. In healthy young volunteers a single bout of resistance exercise is capable of modifying the sensation of experimentally induced pain (18). However, the long term effects of regular exercise therapy on sensitization in knee OA are unknown. A deeper understanding of the mechanisms involved in the painrelieving effects of therapeutic exercises in patients with knee OA is imperative to optimize the exercise paradigm in OA research and clinical management. The purpose of this study was to investigate the effects of a 12-week therapeutic exercise program on the pain sensitivity in patients with knee OA that adhere to the exercise protocol. It was hypothesized that exercise therapy would reduce the patients’ sensitivity to painful pressure stimuli.
METHODS The logistics of this study and preparation of the manuscript are based on the pre-specified protocol including all steps (19). The protocol is registered and available on www.Clinicaltrials.Gov (NCT01545258); the manuscript conforms to the CONSORT reporting guidelines (20).
Trial Design and Participants The study was a single-center randomized controlled, parallel group trial with blinded outcome assessors conducted in Copenhagen, Denmark. Participants were randomly assigned (1:1) to one of two groups; an exercise therapy group (ET) receiving exercise therapy for 12 weeks, or a control group (CG) receiving no attention for 12 weeks. Participants were recruited from the OA outpatient clinic of Copenhagen University Hospital at Frederiksberg, Copenhagen, Denmark, through advertisements in
5 John Wiley & Sons, Inc.
Arthritis Care & Research
newspapers. Eligible participants were adults aged 40 year or more with a clinical diagnosis of tibiofemoral OA confirmed by radiography assessed by an experienced radiologist, and a body mass index (BMI) between 20 and 35 kg/m2. Exclusion criteria included participation in exercise therapy within the previous 3 months, systemic inflammatory and autoimmune diseases, lower extremity joint replacement, significant cardiovascular, neurological or psychiatric disease, cervical or lumbar nerve root compression syndromes, and widespread or regional pain syndromes, e.g. fibromyalgia. At inclusion, the participants’ most symptomatic knee defined the treatment knee that was the target of all subsequent assessments. A clinical ultrasound examination of the treatment knee was performed after baseline assessments assessing presence of effusion, and excessive joint fluid was aspirated ultrasound guided if required. All participants gave written informed consent prior to inclusion and the study was approved by the regional health research ethics committee (H-2-2011-159).
Intervention The participants assigned to ET were offered facility based functional and individualized exercise therapy supervised by a physiotherapist 3 times weekly for 12 weeks. The exercise was group-based and the participants consecutively joined the group as they were included in the study. The exercise program lasted approximately 1 hour and consisted of a 10 minutes warm up phase (bicycle ergometer at moderate intensity) followed by a circuit training program focusing on strength and coordination exercises of the trunk, hips and knees. The exercises were performed with free weights, elastic rubber bands or body weight as resistance. Progression of resistance or coordination difficulty was made on an individual basis, according to a pre-specified progression protocol. The level of each exercise, including
6 John Wiley & Sons, Inc.
Page 6 of 43
Page 7 of 43
Arthritis Care & Research
external load, number of repetitions or duration, was recorded for each patient at each visit in a personal training diary together with current knee pain before an exercise session on a 0-10 numeric rating scale (NRS). If a participant reported symptomatic exacerbation upon attending an exercise session (defined as current knee pain exceeding 5), a ‘rescue’ training program was applied for that session, including only warm up, trunk and hip exercises, repeated twice. A full description of the exercise program including progression guidelines is provided as supplementary online material. Attendance to the exercise sessions was emphasized to the participants and attendance to minimum 24 sessions (out of 36 possible) was defined as protocol adherence. The participants assigned to the CG received no attention from the study during the 12 weeks but were invited to participate in another study on exercise (http://clinicaltrials.gov/ct2/show/NCT01945749) after completing the present study. The CG participants were encouraged not to engage in therapeutic exercise during the study, and were asked about this at the follow-up visit.
Outcomes The co-primary outcomes were changes from baseline in pressure pain sensitivity, assessed by 1) pressure pain thresholds (PPT), and 2) indices of temporal summation (TS) of pain upon sustained noxious pressure stimulation using cuff pressure algometry (21). Secondary outcomes included changes from baseline in patient reported pain, symptoms, and disease impact using the Knee Osteoarthritis Outcome Score (KOOS) (22).
7 John Wiley & Sons, Inc.
Arthritis Care & Research
Cuff algometry Computerized cuff algometry (21) consists of a double chamber tourniquet cuff (VBM Medizintechnik GmbH, Germany) and a computer-controlled air compressor (DoloCuff, Unique Electronic Aps, Denmark). The cuff was wrapped around the calf at the bulky part of the gastrocnemius muscle. Both chambers of the cuff were automatically inflated simultaneously (compression rate: 1 kPa/s) until the subject reported the first sensation of pain by pressing a push-button. This recorded pressure defined the PPT measured in kPa. The PPT was recorded three times separated by at least 60 seconds. The first PPT was discarded to account for learning effects and the two last PPT values were averaged for further analysis. Based on a previous group of 25 knee OA patients assessed twice separated by one week prior to the study, the intraclass correlation coefficient for the PPT using cuff algometry is 0.72 (95%CI:0.64-0.87) with a minimal detectable change of 2.2 kPa; similar to other PPT assessment methods in knee OA (23). Temporal summation of pressure pain (TS) was assessed on the same site as the PPT. Constant pressure stimulation at 125% of the PPT pressure (assessed same day) was applied for 6 minutes. The participants rated the pain intensity continuously during the constant pressure stimulation on an electronic 0-100 mm VAS with 0 and 100 mm anchored with “no pain” and “maximal pain”, respectively. The VAS signal was sampled by a computer at 10 Hz. To quantify TS of pressure pain, the area under the time-VAS curve was calculated, and expressed in mm×s. There are no available data on reliability or minimal detectable change for the TS measures.
8 John Wiley & Sons, Inc.
Page 8 of 43
Page 9 of 43
Arthritis Care & Research
KOOS questionnaire The KOOS (22) was used to assess patient-reported symptoms. KOOS consists of 5 subscales: Pain; Symptoms; Function in daily living; Function in sports and recreation; and Knee related quality of life, each containing questions that are scored on 0-4 Likert scales and summarized into continuous scores ranging from 0 (worst) to 100 (best). Reliability of KOOS in an OA population has been reported as acceptable (24).
Sample size There are limited data available upon which an adequate sample size could be estimated. Total knee replacement in knee OA has been estimated to yield an increase in PPT of 10 kPa using cuff pressure algometry (25). Based on an expected difference between exercise and control groups in the changes from baseline in PPT of 5 kPa (assuming a standard deviation of 5) , 23 participants per group were needed to detect a statistical significant group difference at P=0.05 with a power of at least 90%. To account for possible dropouts we increased the sample size to 60 participants.
Randomization and Masking For group allocation, a computer-generated list of random numbers was used. Randomization sequence was created using SAS (version 9.2) statistical software, stratified by sex with a 1:1 allocation using random block sizes of 2, 4, and 6. The allocation was concealed from study staff enrolling participants, in sequentially numbered, opaque (aluminum foil inside), and sealed envelopes. To prevent subversion of the allocation sequence, the name and date of birth of the participant was written on the envelope. Corresponding envelopes were opened only after the included participants had completed all baseline assessments. Whereas the
9 John Wiley & Sons, Inc.
Arthritis Care & Research
participants were aware of their group allocation, outcome assessors and data analysts were kept blinded to the allocation.
Statistical analyses As this study aims at exploring the underlying mechanisms of exercise on pain sensitivity, the analyses were based on the per protocol population, defined as those participants that had followed the study protocol (12 or less incidences of nonattendance to exercise sessions in the exercise group; no exercise in the control group) and with complete data sets at baseline and follow-up. The analyses of PPT, TS, and KOOS focused on group mean differences in changes from baseline, calculated as the baseline value subtracted from the follow-up value. General linear models were used with a factor for group and adjusting for the baseline value. The analyses were repeated adjusting for age and gender as potential confounders. As recommended from the EQUATOR network we present both analyses for the primary outcomes (26). As the results were unchanged, we only present the adjusted analyses for the secondary outcomes. For statistically significant outcomes, we report effect sizes (standardized mean differences). All analyses were done using SAS software, and statistical significance was accepted at P
Arthritis Care & Research DOI 10.1002/acr.22375
EXERCISE THERAPY REDUCES PAIN SENSITIVITY IN PATIENTS WITH KNEE OSTEOARTHRITIS: A RANDOMIZED CONTROLLED TRIAL Marius Henriksen PhD1, Louise Klokker MSc1, Thomas Graven-Nielsen DMSc2, Cecilie Bartholdy PT1, Tanja Schjødt Jørgensen MSc1,2, Elisabeth Bandak MSc1, Bente DanneskioldSamsøe DMSc1, Robin Christensen PhD1, Henning Bliddal DMSc1,2 1 Clinical
Motor Function Laboratory, The Parker Institute, Department of Rheumatology, Copenhagen University Hospitals Bispebjerg and Frederiksberg, Copenhagen, Denmark 2 Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology, Faculty of Medicine, Aalborg University, 9220 Aalborg, Denmark. Original paper for: Arthritis Care & Research Correspondence: Marius Henriksen, PT, PhD Clinical Motor Function Laboratory The Parker Institute, Department of Rheumatology Copenhagen University Hospital at Frederiksberg Ndr. Fasanvej 57 DK-2000 Copenhagen F Denmark Tel: +45 3816 4160 Fax: +45 3816 4159 email: [email protected] Word count: 3,796 Running Head: Exercise, pain sensitivity and knee osteoarthritis
FUNDING This study was supported by a grants from The Danish Council for Independent Research | Medical Sciences (Grant number: 10-093704), Danish Physiotherapists Association, The Lundbeck Foundation, and The Oak Foundation to cover salaries and running costs.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an ‘Accepted Article’, doi: 10.1002/acr.22375 © 2014 American College of Rheumatology Received: Jan 13, 2014; Revised: Apr 28, 2014; Accepted: May 27, 2014
Arthritis Care & Research
ABSTRACT Background: Exercise has beneficial effects on pain in knee osteoarthritis (OA), yet the underlying mechanisms are unclear. The purpose of this study was to investigate the effects of exercise on pressure pain sensitivity in patients with knee OA. Methods: In a randomized controlled trial participants were assigned to 12 weeks of supervised exercise therapy (ET, 36 sessions), or a no attention control group (CG). Pressure pain sensitivity was assessed by cuff pressure algometry on the calf of the most symptomatic leg. The co-primary outcomes were 1) pressure pain thresholds (PPT), and 2) cumulated visual analogue scale pain scores during constant pressure for 6 minutes at 125% of the PPT as a measure of temporal summation (TS) of pressure pain. Secondary outcomes included self-reported pain using the KOOS questionnaire. Analyses were based on the ‘Per-Protocol’ population (participants following the protocol). Results: 60 participants were randomized (31 ET/29 CG), and the Per-Protocol population included 48 participants (25 ET/23 CG). At follow-up mean group differences in the change from baseline were 3.1 kPa (95% CI: 0.2 to 6.0; P=0.038) for the PPT, 2,608 mm*s (95%CI: 458 to 4,758; P=0.019) for TS, and 6.8 points (95% CI 1.2 to 12.4; P=0.018), for KOOS pain, all in favor of ET. Conclusion: Pressure pain sensitivity, temporal summation and self-reported pain are reduced among patients completing a 12-week supervised exercise program compared to a no-attention control group. These results demonstrate beneficial effects of exercise on basic pain mechanisms, and further exploration may provide basis for optimized treatment. Clinicaltrials.gov: NCT01545258
2 John Wiley & Sons, Inc.
Page 2 of 43
Page 3 of 43
Arthritis Care & Research
SIGNIFICANCE AND INNOVATION This randomized study is the first to evaluate the effect of exercise on pain sensitivity in knee osteoarthritis (OA). A 12-week supervised exercise program reduces the pressure pain sensitivity compared to a no attention control group, in adjunct to an effect on self-reported pain. This study suggests that exercise therapy has beneficial effects on basic pain mechanisms, and further exploration of this effect may provide basis for optimized treatment.
3 John Wiley & Sons, Inc.
Arthritis Care & Research
Osteoarthritis (OA) is the most frequent painful musculoskeletal disease (1), with a massive socioeconomic burden (2). Because there is no cure, clinical management of OA is an enormous challenge. The ultimate treatment is surgical joint replacement but conservative treatments are recommended before surgery (3). A highly recommended treatment for OA is exercise (4;5), with well-proven beneficial effects on pain and disability (6). Although exercise is more cost-effective than surgery (5), the effect sizes of exercise are moderate and optimization of the exercise paradigm is needed. While the link between exercise and reduced disability is mediated by e.g. increased muscle strength and endurance (7), the mechanisms linking exercise therapy and pain reduction remain to be explored. It has recently been shown that knee OA patients may have both peripheral and central sensitization of pain mechanisms resulting in spreading hyperalgesia (8). Thus, targeted pain treatment may focus on both peripheral and central mechanisms. Presently it is not known if or to what extent exercise affects these mechanisms in patients with knee OA. Recent clinical trials have shown that strengthening exercises of muscles remote to the knee (isolated hip muscle strengthening) caused reductions in knee OA pain over 8-12 weeks (9;10). One suggested mechanism behind the clinical effects of exercise is biomechanical optimization, however, in a number of studies no effects of exercise on knee biomechanics have been reported (11-15). Another possible mechanism is that exercise leads to widespread adaptations in the central nervous system resulting in reduced pain sensitivity. Joint nociception may cause sensitization of pain mechanisms in the central nervous system (16). In addition, peripheral sensitization includes structural and functional changes in the peripheral nociceptors, e.g. caused by inflammation (16). In systemic low-grade inflammatory diseases (e.g. diabetes) anti-inflammatory effects of regular exercises have been suggested
4 John Wiley & Sons, Inc.
Page 4 of 43
Page 5 of 43
Arthritis Care & Research
(17), which in turn may lead to reduced peripheral sensitization. In healthy young volunteers a single bout of resistance exercise is capable of modifying the sensation of experimentally induced pain (18). However, the long term effects of regular exercise therapy on sensitization in knee OA are unknown. A deeper understanding of the mechanisms involved in the painrelieving effects of therapeutic exercises in patients with knee OA is imperative to optimize the exercise paradigm in OA research and clinical management. The purpose of this study was to investigate the effects of a 12-week therapeutic exercise program on the pain sensitivity in patients with knee OA that adhere to the exercise protocol. It was hypothesized that exercise therapy would reduce the patients’ sensitivity to painful pressure stimuli.
METHODS The logistics of this study and preparation of the manuscript are based on the pre-specified protocol including all steps (19). The protocol is registered and available on www.Clinicaltrials.Gov (NCT01545258); the manuscript conforms to the CONSORT reporting guidelines (20).
Trial Design and Participants The study was a single-center randomized controlled, parallel group trial with blinded outcome assessors conducted in Copenhagen, Denmark. Participants were randomly assigned (1:1) to one of two groups; an exercise therapy group (ET) receiving exercise therapy for 12 weeks, or a control group (CG) receiving no attention for 12 weeks. Participants were recruited from the OA outpatient clinic of Copenhagen University Hospital at Frederiksberg, Copenhagen, Denmark, through advertisements in
5 John Wiley & Sons, Inc.
Arthritis Care & Research
newspapers. Eligible participants were adults aged 40 year or more with a clinical diagnosis of tibiofemoral OA confirmed by radiography assessed by an experienced radiologist, and a body mass index (BMI) between 20 and 35 kg/m2. Exclusion criteria included participation in exercise therapy within the previous 3 months, systemic inflammatory and autoimmune diseases, lower extremity joint replacement, significant cardiovascular, neurological or psychiatric disease, cervical or lumbar nerve root compression syndromes, and widespread or regional pain syndromes, e.g. fibromyalgia. At inclusion, the participants’ most symptomatic knee defined the treatment knee that was the target of all subsequent assessments. A clinical ultrasound examination of the treatment knee was performed after baseline assessments assessing presence of effusion, and excessive joint fluid was aspirated ultrasound guided if required. All participants gave written informed consent prior to inclusion and the study was approved by the regional health research ethics committee (H-2-2011-159).
Intervention The participants assigned to ET were offered facility based functional and individualized exercise therapy supervised by a physiotherapist 3 times weekly for 12 weeks. The exercise was group-based and the participants consecutively joined the group as they were included in the study. The exercise program lasted approximately 1 hour and consisted of a 10 minutes warm up phase (bicycle ergometer at moderate intensity) followed by a circuit training program focusing on strength and coordination exercises of the trunk, hips and knees. The exercises were performed with free weights, elastic rubber bands or body weight as resistance. Progression of resistance or coordination difficulty was made on an individual basis, according to a pre-specified progression protocol. The level of each exercise, including
6 John Wiley & Sons, Inc.
Page 6 of 43
Page 7 of 43
Arthritis Care & Research
external load, number of repetitions or duration, was recorded for each patient at each visit in a personal training diary together with current knee pain before an exercise session on a 0-10 numeric rating scale (NRS). If a participant reported symptomatic exacerbation upon attending an exercise session (defined as current knee pain exceeding 5), a ‘rescue’ training program was applied for that session, including only warm up, trunk and hip exercises, repeated twice. A full description of the exercise program including progression guidelines is provided as supplementary online material. Attendance to the exercise sessions was emphasized to the participants and attendance to minimum 24 sessions (out of 36 possible) was defined as protocol adherence. The participants assigned to the CG received no attention from the study during the 12 weeks but were invited to participate in another study on exercise (http://clinicaltrials.gov/ct2/show/NCT01945749) after completing the present study. The CG participants were encouraged not to engage in therapeutic exercise during the study, and were asked about this at the follow-up visit.
Outcomes The co-primary outcomes were changes from baseline in pressure pain sensitivity, assessed by 1) pressure pain thresholds (PPT), and 2) indices of temporal summation (TS) of pain upon sustained noxious pressure stimulation using cuff pressure algometry (21). Secondary outcomes included changes from baseline in patient reported pain, symptoms, and disease impact using the Knee Osteoarthritis Outcome Score (KOOS) (22).
7 John Wiley & Sons, Inc.
Arthritis Care & Research
Cuff algometry Computerized cuff algometry (21) consists of a double chamber tourniquet cuff (VBM Medizintechnik GmbH, Germany) and a computer-controlled air compressor (DoloCuff, Unique Electronic Aps, Denmark). The cuff was wrapped around the calf at the bulky part of the gastrocnemius muscle. Both chambers of the cuff were automatically inflated simultaneously (compression rate: 1 kPa/s) until the subject reported the first sensation of pain by pressing a push-button. This recorded pressure defined the PPT measured in kPa. The PPT was recorded three times separated by at least 60 seconds. The first PPT was discarded to account for learning effects and the two last PPT values were averaged for further analysis. Based on a previous group of 25 knee OA patients assessed twice separated by one week prior to the study, the intraclass correlation coefficient for the PPT using cuff algometry is 0.72 (95%CI:0.64-0.87) with a minimal detectable change of 2.2 kPa; similar to other PPT assessment methods in knee OA (23). Temporal summation of pressure pain (TS) was assessed on the same site as the PPT. Constant pressure stimulation at 125% of the PPT pressure (assessed same day) was applied for 6 minutes. The participants rated the pain intensity continuously during the constant pressure stimulation on an electronic 0-100 mm VAS with 0 and 100 mm anchored with “no pain” and “maximal pain”, respectively. The VAS signal was sampled by a computer at 10 Hz. To quantify TS of pressure pain, the area under the time-VAS curve was calculated, and expressed in mm×s. There are no available data on reliability or minimal detectable change for the TS measures.
8 John Wiley & Sons, Inc.
Page 8 of 43
Page 9 of 43
Arthritis Care & Research
KOOS questionnaire The KOOS (22) was used to assess patient-reported symptoms. KOOS consists of 5 subscales: Pain; Symptoms; Function in daily living; Function in sports and recreation; and Knee related quality of life, each containing questions that are scored on 0-4 Likert scales and summarized into continuous scores ranging from 0 (worst) to 100 (best). Reliability of KOOS in an OA population has been reported as acceptable (24).
Sample size There are limited data available upon which an adequate sample size could be estimated. Total knee replacement in knee OA has been estimated to yield an increase in PPT of 10 kPa using cuff pressure algometry (25). Based on an expected difference between exercise and control groups in the changes from baseline in PPT of 5 kPa (assuming a standard deviation of 5) , 23 participants per group were needed to detect a statistical significant group difference at P=0.05 with a power of at least 90%. To account for possible dropouts we increased the sample size to 60 participants.
Randomization and Masking For group allocation, a computer-generated list of random numbers was used. Randomization sequence was created using SAS (version 9.2) statistical software, stratified by sex with a 1:1 allocation using random block sizes of 2, 4, and 6. The allocation was concealed from study staff enrolling participants, in sequentially numbered, opaque (aluminum foil inside), and sealed envelopes. To prevent subversion of the allocation sequence, the name and date of birth of the participant was written on the envelope. Corresponding envelopes were opened only after the included participants had completed all baseline assessments. Whereas the
9 John Wiley & Sons, Inc.
Arthritis Care & Research
participants were aware of their group allocation, outcome assessors and data analysts were kept blinded to the allocation.
Statistical analyses As this study aims at exploring the underlying mechanisms of exercise on pain sensitivity, the analyses were based on the per protocol population, defined as those participants that had followed the study protocol (12 or less incidences of nonattendance to exercise sessions in the exercise group; no exercise in the control group) and with complete data sets at baseline and follow-up. The analyses of PPT, TS, and KOOS focused on group mean differences in changes from baseline, calculated as the baseline value subtracted from the follow-up value. General linear models were used with a factor for group and adjusting for the baseline value. The analyses were repeated adjusting for age and gender as potential confounders. As recommended from the EQUATOR network we present both analyses for the primary outcomes (26). As the results were unchanged, we only present the adjusted analyses for the secondary outcomes. For statistically significant outcomes, we report effect sizes (standardized mean differences). All analyses were done using SAS software, and statistical significance was accepted at P
