Exploring massage benefits for arthritis of the knee (embark)

Abstract

Purpose: The purpose of the study was to verify the efficacy of an 8-week course of manualized Swedish massage, at an optimized dose, as well as to assess a maintenance dose, for reducing pain and increasing function compared to an active (light touch bodywork) control, and passive (usual care) control, in 222 adults with osteoarthritis (OA) of the knee. Detailed, protocolized cost-accounting was also an aim of the study. Although a formal cost-effectiveness study was beyond the scope of this research, this study allowed for the collection of robust data, using validated questionnaires and diaries to capture relevant cost, benefit, and quality-of-life data to be used in a future, formal cost-effectiveness trial. Methods: Participants were recruited from three sites: Duke, Rutgers, and Yale. Participants had radiographically-confirmed knee OA, met the American College of Rheumatology criteria for OA, and were experiencing moderate pain. Participants were randomly assigned to intervention (total body Swedish massage) or one of two control groups (light touch bodywork [LT] or usual care [UC]), with participants in the massage and LT groups undergoing study measures at baseline and after treatment (8 weeks), as well as at 16, 24, 36, and 52 weeks. The participants in the UC group underwent study measures at baseline, and at 8, 16, 24, 32, 40 and 48 weeks. All participants in the massage and LT arms remaining in the study at 8 weeks were randomly assigned (50%) to maintenance dosing (60-minute session every 2 weeks starting at week 10), or (50%) UC control, in their respective treatment arm in a factorial pattern. Measures included questionnaires, time to walk 50 feet, knee range of motion, health care cost expenditure and pain medication and adverse event logs. The primary outcome was the WOMAC global score. Values are presented as predicted means with 95% confidence intervals adjusted for stratification variables (site and BMI[<30/30+]) unless otherwise stated. Results: The study sample consisted of 222 patients aged 35 to 89 years with a mean age of 63 years, generally obese, non-Hispanic white, and female. Baseline characteristics were similar in the three therapy groups at baseline and the five therapy groups at week 8. In the intent-to-treat analysis of WOMAC global score change pre- versus post-intervention (baseline to week 8), patients in the massage group improved 17.70 points; patients in the LT group improved 9.54 points; and patients in the UC group improved 8.15 points. The between-group difference-in-change for massage versus UC was −9.55 points (95% confidence interval [CI], −14.66–−4.45, p = 0.001); for massage versus LT was −8.16 points (95% CI, −13.50–−2.81, p = 0.003); and for LT versus UC was −1.40 (95% CI, −6.81–4.01, p = 0.611). In the modified intent-to-treat analysis of WOMAC score change pre- versus post-maintenance therapy (week 8 to week 24), patients in the massage group without maintenance therapy worsened 8.68 points, patients in the massage group with maintenance therapy improved 0.89 points, patients in the LT group without maintenance therapy improved 4.60 points, patients in the LT with maintenance therapy improved 10.64 points in the LT group with maintenance therapy, and the UC group improved 1.45 points. The between group difference-in-change for massage with versus without maintenance therapy, 9.82 points, was significant at 24 weeks (95% CI, 1.80–17.85). In analysis of long-term change across the four maintenance groups (baseline to week 24 and baseline to week 52), patient groups all improved, but omnibus tests of group effect were not significant. All previous effect analyses were repeated for the following secondary outcomes: range of motion, visual analogue scale, timed walk, pain interference, and the three WOMAC subscales. Conclusions: At 8 weeks, subjects in the massage group saw a clinically significant greater reduction in WOMAC global scores than the LT and UC groups, reaffirming the results of our previous studies. In the massage group, maintenance sustained, but did not enhance, massage effects - large initial improvements were maintained among those in maintenance treatment and partially lost among those in UC. In the LT group, modest initial improvements (non-clinically significant) were furthered among those in maintenance treatment and maintained in the UC group. By 52 weeks, improvement was seen in all groups, but differences in improvement between the groups were no longer significant. We are still examining what participants did over the course of the year that resulted in improvement in all groups. Further study of the role of maintenance massage therapy past 8 weeks is warranted.