The effect of land-based exercise on pain and function outcomes in hip osteoarthritis: a systematic review and meta-analysis

Abstract

s / Osteoarthritis and Cartilage 22 (2014) S7–S56 S21 number of steps per day. Worsening was defined for the period 60-84 months as any increase in Osteoarthritis Research International (OARSI) JSN grade (in two knees) and any increase in Whole-Organ Magnetic Resonance Imaging (MRI) Score (WORMS) for cartilage morphology (in one knee). For cartilage morphology, separate analyses were conducted for medial and lateral compartments. Odds ratios (OR) and 95% confidence intervals (CI) for JSN and cartilage morphology were calculated using logistic regression models. Generalized estimating equations accounted for two knees per individual for JSN and for 5 regions per knee for cartilage morphology. Covariates included age, sex, body mass index (BMI), alignment, knee pain at 60 months, knee injury, and 60month KL grades. For cartilage morphology, 60-month cartilage morphology status was additionally included as a covariate. Results: The 1179 study participants (1019 left and 989 right knees) (59% females) were on average 67.0 ( 7.6) years, with a mean BMI of 29.8 ( 5.3) kg/m2. Mean steps per day were 9569 ( 3657). At 60 months, 51.6%, 20.5%, and 27.9% of knees had KL grade 0, 1, and 2, respectively. No significant associations were found between activity level at 60months and worsening of JSN or lateral cartilage morphology between 60-84 months. For medial cartilage morphology, those in the lowest tertile ( 10580 steps per day) showed higher odds of worsening of medial cartilage morphology compared to those with moderate activity level (OR 1.05 95% CI 1.00, 1.10). Conclusion: Those in the lowest and highest tertiles of physical activity showed a weak but significant association with worsening of medial cartilage morphology compared with those in the middle tertile. No significant associations were revealed for physical activity level and worsening of JSN or lateral cartilage morphology in individuals with mild osteoarthritis or at risk for knee osteoarthritis. 28 THE EFFECT OF LAND-BASED EXERCISE ON PAIN AND FUNCTION OUTCOMES IN HIP OSTEOARTHRITIS: A SYSTEMATIC REVIEW AND META-ANALYSIS R. Leigh y, K. Mills z, R. Ferber y. yUniv. of Calgary, Calgary, AB, Canada; zMacquarie Univ., Sydney, Australia Purpose: The purpose of the present systematic review and metaanalysis was to 1) determine the efficacy of exercise on pain and function outcomes (self-report and performance-based) in individuals with hip osteoarthritis (OA) and 2) to determine if the point estimate of effect of exercise changes when exercise is semi-tailored and/or when recruitment is limited to hip OA subjects only (and not knee OA). Methods: Electronic databases were searched for randomized controlled trials (RCTs) that investigated the effect of land-based exercise on pain, self-reported and performance-based physical function outcomes in individuals with hip OA. RCTs that included subjects with a radiographic or clinical diagnosis of unilateral or bilateral hip OA were eligible for inclusion in the present review. Only RCTs that compared the effect of a land-based exercise program (e.g. strengthening, stretching, aerobic, balance training) with a control group were eligible for inclusion.Group means, standard deviations, and sample sizes were extracted from each study. From these data, mean differences, standardized mean differences (SMD 1⁄4 mean difference divided by the pooled standard deviation) and their 95% confidence intervals were calculated. When 2 studies measured the same outcome at time points within 6-weeks of each other, meta-analyses of their SMDs were performed in Cochrane Review Manager (V5.2) using an inverse variance random effects model. Confidence intervals that included zero were interpreted as having no effect.An a priori sub-group analysis of treatment effect was performed in studies that recruited hip OA patients only (and not knee OA). A post-hoc sub-group analysis examined whether semi-tailored/hip-specific exercise programs (as seen in the more recent RCTs) resulted in different treatment effects as compared with programs that previously prescribed generic lower extremity exercises. This analysis was performed given the effort made in recent RCTs to tailor hip OA treatments to subject assessment findings. Results: Following a comprehensive search strategy, 8 studies were included in the review. All studies scored 7/10 on the PEDro quality index tool. No evidence of publication bias was found for pain outcomes. Pooled data revealed that exercise interventions had no effect on pain outcomes (SMD 1⁄4 -0.15 (-0.60 to 0.29) post-intervention or at follow-up (SMD 1⁄4 -0.15 (-0.55 to 0.25)) when compared with controls. There was also no effect of exercise on self-reported function post intervention (SMD 1⁄4 -0.10 (-0.41 to 0.22), at 6-months (SMD 1⁄4 -0.13 (-0.43 to 0.18) or 10-12 months follow-up (SMD 1⁄4 -0.15 (0.55 to 0.26)). Pooled data revealed exercise had no effect on performance-based physical function outcomes post-intervention or at follow-up. Results of the sub-analyses revealed isolated recruitment to individuals with only hip OA did not alter the results of the initial analysis. Exercise continued to have no effect on pain outcomes post-intervention (SMD 1⁄4 -0.04 (-0.55 to 0.48)) or at follow-up (SMD 1⁄4 -0.20 (-0.84 to 0.44)) and there was no effect on functional outcomes. However, when only those studies that prescribed a semi-tailored exercise program were considered, data pooling of revealed that exercise resulted in moderate improvements in pain (SMD 1⁄4 -0.78 (-1.21 to -0.35)) immediately postintervention. Conclusions: Low risk of bias studies provided no evidence of an effect of land-based exercise on pain and function (both self-reported and performance-based) post-intervention. This lack of effect was also observed when studies limited recruitment to hip OA subjects only. However, exercise that is based on patient assessment findings (semitailored) does appear to impart a moderate beneficial effect on pain post-intervention and should be considered an important component of exercise based interventions in hip OA populations. 29 PSYCHOSOCIAL FACTORS AND PAIN EXACERBATION IN KNEE OSTEOARTHRITIS: A WEB BASED CASE-CROSSOVER STUDY D.J. Hunter y, K. Bennell z, J. Makovey y, B. Metcalf z, J. Chen y, L. March y, F. Keefe x, A. Williams k, Y. Zhang{. yUniv. of Sydney, St Leonards, Australia; zUniv. of Melbourne, Melbourne, Australia; xDuke Univ., Durham, NC, USA; kUniv. of New South Wales, Kensington, Australia; {Boston Univ., Boston, MA, USA Purpose: While pain from knee osteoarthritis (OA) has long been considered a persistent condition, the symptoms experienced by OA patients are neither constant nor stable. If the risk factors for these pain exacerbations could be identified and avoided, many such episodes could be prevented. A range of psychological and social environmental processes can potentially affect a patient’s response to pain. Using aweb based case-crossover approach, we evaluated whether psychosocial factors (specifically pain coping and mood) are risk factors for pain exacerbations in people with knee OA. Methods: Participants with a diagnosis of symptomatic knee OA (i.e., Kellgren and Lawrence> 2 or radiographic patellofemoral OA and knee pain on most days on the past month) were recruited and followed for 3 months at 10 day intervals (control periods). Subjects were instructed to log on to the study website if they experienced a knee pain exacerbation during the follow-up period (hazard periods). Via the internet we collected data on triggers occurring during “control periods” (i.e., periods without pain exacerbation) and “hazard periods” (i.e., periods immediately preceding the pain exacerbation). Pain exacerbation was defined as an increase of 100 units in a subject’s WOMAC knee pain score (VAS 0-500) over the follow-up from his/her mildest pain score reported at the baseline visit. We collected data on potential triggers, including psychosocial factors, on 1 day prior, 2 days prior, and 3-7 days prior to the index dates (i.e., date of pain exacerbation for hazard period, and date of data assessment for control periods). We assessed daily mood (Negative/ Positive) using the Profile of Mood States (PANAS) and pain coping in the previous 30 days using the Pain Coping Inventory. We examined the relation of psychological factors to the risk of pain exacerbation using the conditional logistic regression model. Results: Of 267 participants (women: 61%, mean age: 62 years, mean BMI: 29.8 kg/m2) recruited in the study, 160 subjects experienced at least one episode of knee pain exacerbation. Of them 46 subjects had their pain coping assessed 30-days apart between some case periods and control periods. Higher negative affect (mood items e.g., distressed, irritable, nervous) and passive coping strategies (e.g. of items. I restrict my social activities; I focus on the location and intensity of pain) were significantly associated with increased risk of flares (Table). In contrast higher positive affect (mood items e.g., excited, proud, inspired) score and active coping strategies (e.g. of items. I stay busy or active, I clear my mind of bothersome thoughts) trended to an association with a protective effect from pain exacerbation.