Tom Kindlon , Ellen M Goudsmit
DOI: 10.2340/16501977-0493

Sir,
Given there is no formal system to report adverse reactions to non-pharmacological interventions such as graded exercise therapy (GET) for chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME), other sources of data need to be considered when evaluating safety. As noted by Clark & White, a large survey conducted in 2001 by the charity
Action for ME found that 50% of patients who received graded exercise felt worse (1, 2). They also referred to a subsequent study by the same group suggesting that many patients might not have been treated by experienced therapists (3). However, the sample was small and, as in all surveys, therapist competence was not assessed.
A review of all the surveys conducted to date not only supports the view that a significant proportion of patients experience adverse reactions following GET, but also that it is premature to attribute those reactions to practitioner inexperience or inadequate training (1, 4). For example, the results of a recent survey conducted by the ME Association showed that of the 906 individuals who had received GET, 33. 1% felt “much worse” and 23. 4% judged themselves to be “slightly worse” (4). Similarly, a survey of patients who had been treated in the previous 3 years, i. e. following the refinement of the protocol as discussed by Clark & White, revealed that 34% of the 722 who had tried GET perceived themselves to be worse (5).
Without details of the training of the therapist and their fidelity to the treatment manual, one can only speculate about the factors associated with poor outcome. Nijs et al. (6) discussed some of the possible reasons. However, there are additional factors that deserve consideration when evaluating the efficacy and safety of GET. Firstly, the survey results may reflect, at least in part, the experiences of patients receiving treatment in a clinical setting. As has been shown in studies on other interventions, the outcomes documented in routine practice may be more realistic than those obtained in randomized controlled trials (7). Secondly, many patients may not be able to complete graded activity schedules for various reasons, including ongoing pathology. For instance, Black & McCully (8) used an accelerometer to measure activity levels before, during and after a 4-week “training period” consistent with GET. They documented an increase in activity counts lasting between 4 and 10 days, and this was associated with higher scores for pain and fatigue. The inability to sustain target activity levels was also noted by Friedberg (9), who followed the progress of one patient during 26 sessions of GET. He recorded a 10. 6% decrease in mean weekly step counts, leading Friedberg to speculate that the subjective measures of improvement might have been the result of activity substitution and a corresponding reduction in perceived stress.
Finally, we were surprised that neither of the letters cited the research by White et al. (10). This elegant study supports the growing evidence of abnormal metabolic and immunological reactions to exercise in subsets with CFS. Although their sample was small, White et al. found elevated concentrations of the pro-inflammatory cytokine tumour necrosis factor-alpha at time-points of 3 h and 3 days after exercise. In addition, they documented increased levels of the anti-inflammatory cytokine transforming growth factor-beta after normal exertion. We therefore concur with Nijs et al. (6) as well as other researchers, that GET may not be appropriate for all patients with CFS and that pacing may provide a useful, acceptable and safe alternative (6, 11, 12).