Abstract
The MAGENTA pragmatic parallel groups randomized controlled trial compared graded exercise therapy (GET) with activity management (AM) in treating paediatric myalgic encephalomyelitis or chronic fatigue syndrome (ME/CFS). Children aged 8-17 years with mild/moderate ME/CFS and presenting to NHS specialist paediatric services were allocated at random to either individualised flexible treatment focussing on physical activity (GET, 123 participants) or on managing cognitive, school and social activity (AM, 118 participants) delivered by NHS therapists. The primary outcome was the self-reported short-form 36 physical function subscale (SF-36-PFS) after 6 months, with higher scores indicating better functioning. After 6 months, data were available for 201 (83%) participants who received a mean of 3.9 (GET) or 4.6 (AM) treatment sessions. Comparing participants with measured outcomes in their allocated groups, the mean SF-36-PFS score changed from 54.8 (standard deviation 23.7) to 55.7 (23.3) for GET and from 55.5 (23.1) to 57.7 (26.0) for AM giving an adjusted difference in means of -2.02 (95% confidence interval -7.75, 2.70). One hundred thirty-five participants completed the mean SF-36-PFS at 12 months, and whilst further improvement was observed, the difference between the study groups remained consistent with chance. The two study groups showed similar changes on most of the secondary outcome measures: Chalder Fatigue, Hospital Anxiety and Depression Scale: Depression, proportion of full-time school attended, a visual analogue pain scale, participant-rated change and accelerometer measured physical activity, whether at the 6-month or 12-month assessment. There was an isolated finding of some evidence of an improvement in anxiety in those allocated to GET, as measured by the Hospital Anxiety and Depression Scale at 6 months, with the 12-month assessment, and the Spence Children's Anxiety scale being aligned with that finding. There was weak evidence of a greater risk of deterioration with GET (27%) than with AM (17%; p = 0.069). At conventional UK cost per QALY thresholds, the probability that GET is more cost-effective than AM ranged from 18 to 21%. Whilst completion of the SF-36-PFS, Chalder Fatigue Scale and EQ-5D-Y was good at the 6-month assessment point, it was less satisfactory for other measures, and for all measures at the 12-month assessment. Conclusion: There was no evidence that GET was more effective or cost-effective than AM in this setting, with very limited improvement in either study group evident by the 6-month or 12-month assessment points. Trial registration: The study protocol was registered at www.isrctn.com (3rd September 2015; ISRCTN 23962803) before the start of enrolment to the initial feasibility phase.