The PACE Trial Controversy
The PACE trial (White et al., 2011) was the largest and most expensive ME/CFS treatment study ever conducted (5 million, n=641), comparing graded exercise therapy (GET), cognitive behavioral therapy (CBT), adaptive pacing therapy (APT), and specialist medical care (SMC). Its conclusions—and their subsequent dismantling—constitute the most significant controversy in ME/CFS research history.
1 What the Trial Claimed
The original publication in The Lancet reported that GET and CBT were “moderately effective” treatments that improved fatigue and physical function compared to APT and SMC. A subsequent “recovery” analysis claimed that 22% of patients in the GET and CBT arms had “recovered.”
2 Criticisms of Methodology
Extensive criticism has identified fundamental methodological problems:
- Mid-trial outcome changes: The threshold for “improvement” on the Chalder Fatigue Scale was lowered from \(\leq\) 18 to \(\leq\) 22, and for physical function from \(\geq\) 85 to \(\geq\) 60 on the SF-36. Under the revised thresholds, 13% of participants met “recovery” criteria at baseline—before any treatment (Wilshire et al. 2018)
- Subjective outcomes only: All primary outcomes were self-reported questionnaires administered by non-blinded researchers. Objective measures (six-minute walk test, step-count actigraphy, employment status, welfare benefit claims) showed no significant improvement (Geraghty and Blease 2019)
- Non-blinded design: Participants and therapists knew which treatment was being administered, and outcome assessors were not independent. In a trial using subjective outcomes, this is a critical source of bias
- Oxford criteria: The trial used the broadest diagnostic criteria, likely including patients without ME/CFS
- Lack of PEM monitoring: No systematic assessment of post-exertional malaise—the defining feature of the disease under study
3 Reanalysis Results
After a protracted legal battle for data access under the Freedom of Information Act, independent reanalysis by Wilshire et al. (2018) found:
- Using the trial’s original (pre-change) outcome thresholds, recovery rates dropped from 22% to 7%—not significantly different from the control groups
- Step-count actigraphy showed no significant improvement in any group
- Six-minute walk distance improvements were clinically trivial (approximately 20 meters)
- Long-term follow-up showed convergence of all groups, suggesting no durable benefit (ME/CFS Science 2021)
4 Long-Term Follow-Up: No Sustained Benefits
Long-term follow-up data from the PACE trial and the related GETSET trial (graded exercise self-help) provide the strongest evidence against sustained benefit from GET and CBT (ME/CFS Science 2021). At long-term follow-up, all treatment groups—including the control groups—converged to similar outcome levels. This convergence pattern indicates that any short-term improvements attributed to GET and CBT were not durable; patients in the active treatment arms did not maintain advantages over those who received specialist medical care alone. The pattern is consistent with regression to the mean and non-specific effects (therapeutic attention, expectation) rather than genuine disease modification.
5 Impact on Clinical Guidelines
The PACE trial’s legacy has been profound and ultimately self-correcting. NICE (UK, 2021) withdrew its recommendation for GET and de-emphasized CBT, explicitly citing the evidence for harm (National Institute for Health and Care Excellence 2021). The CDC (US, 2022) similarly removed exercise therapy recommendations. However, clinical practice in many countries has been slow to follow guideline changes, and GET continues to be prescribed—particularly in countries where ME/CFS recognition lags (see Section Exercise Therapy Debates). The persistence of PACE-derived recommendations was starkly illustrated in Australia, where the Royal Australian College of General Practitioners (RACGP) published a Handbook of Non-Drug Interventions (HANDI) recommendation for “incremental physical activity” in ME/CFS as late as 2024—years after NICE and CDC withdrew equivalent guidance. Stallard et al. (2026) published a formal critique in the Australian Journal of General Practice, demonstrating that an AGREE II appraisal scored the HANDI guideline at only 2% for development rigor (Stallard et al. 2026). The critique identified the guideline’s dependence on the PACE trial and an outdated Cochrane review, documented patient survey data reporting 54–74% harm rates from GET, and called for immediate withdrawal pending the National Health and Medical Research Council (NHMRC) review of ME/CFS guidelines. An invited counterargument by Ewald (Ewald 2026) defended HANDI as a non-drug intervention summary rather than a clinical guideline and cited moderate GRADE evidence from randomized controlled trials, while acknowledging that ME/CFS pathophysiology remains unknown. The exchange illustrates the ongoing tension between legacy evidence syntheses that predate PACE reanalysis and the post-2021 international consensus against GET.
6 Lessons for Future Research
The PACE trial illustrates how subjective outcomes, non-blinded design, mid-trial protocol changes, and broad case definitions can produce misleading results in ME/CFS research. Future trials should require: objective outcome measures (two-day CPET, actigraphy, biomarkers), PEM-requiring diagnostic criteria, independent outcome assessment, pre-registered protocols with locked outcome thresholds, and systematic PEM monitoring as a safety endpoint. (See Subjective Outcomes in Unblinded Trials for a full methodological analysis of how subjective outcomes interact with blinding failures.)