The evidence base for physiotherapy in myalgic encephalomyelitis/chronic fatigue syndrome when considering post-exertional malaise: a systematic review and narrative synthesis

Key points

1. 0118 RCTs (2320 participants total) were included, but only one RCT enrolled patients meeting criteria that require post-exertional malaise (ME), and that study found no sustained benefit.
2. 02Interventions appearing effective for broadly defined chronic fatigue (e.g., graded exercise therapy) showed weaker or no effects in patients diagnosed with stricter CFS or ME criteria.
3. 03Effect sizes on fatigue and physical function were modest, and post-treatment scores remained far below population norms even when statistically significant improvements occurred.
4. 04Objective outcome measures (fitness tests, activity monitoring, employment) rarely showed significant improvement even when self-reported fatigue scores did.
5. 05Patient surveys outside the reviewed RCTs indicate that more than half of ME patients who undertook graded exercise therapy reported negative outcomes, while activity pacing had the lowest negative response rate.

How it was conducted

Design

Systematic review and narrative synthesis of randomized controlled trials

Databases searched

PubMed, CINAHL, PEDro; RCTs published 2000-2020

Included studies

18 RCTs with 2320 total participants

Diagnostic subgroups

3 RCTs using Oxford criteria (no PEM required - CF); 14 RCTs using CDC-94/Fukuda or NICE criteria (PEM optional - CFS); 1 RCT using Canadian Consensus Criteria (PEM required - ME)

Interventions covered

Graded exercise therapy, activity pacing, qigong, yoga, orthostatic training, pain education, multidisciplinary rehabilitation, self-management education

Quality assessment

PEDro scale; scores ranged from 5 to 8, median 6; 15 studies rated high quality, 4 fair

What they found

• Only 1 of 18 RCTs enrolled ME patients (PEM required by CCC criteria); that self-management and pacing program showed short-term benefit for self-efficacy (SES p<0.05) but not physical functioning (SF-36 ns), and no outcomes were sustained at long-term follow-up.
• In 3 CF trials (Oxford criteria, no PEM required), GET reduced fatigue: mean CFQ pre 28.4, post 22.7; SF-36-PF pre 34.6, post 46.2; effects sustained to 1-year follow-up in the PACE trial (CFQ p<0.01; SF-36-PF p<0.01).
• In 14 CFS trials (PEM optional), mean CFQ pre 28.7, post 18.4; SF-36-PF pre 41.8, post 46.7; however physical functioning improvements were absent or negative in most studies.
• Post-treatment SF-36-PF scores across groups averaged approximately 47, which remains below the norm for adults aged 75-84 years (mean SF-36-PF of 58) and far below the norm for working-age adults (approximately 90).
• In the PACE trial (Oxford criteria, n=641), GET improved fatigue (CFQ p<0.01) and physical function (SF-36-PF p<0.01) at 1 year, but both GET and APT failed to reduce employment loss or increase fitness; walking improvements were small; at 2-year follow-up differences between groups were not significant.
• In the GES trial (NICE criteria, n=211), guided graded exercise self-help improved fatigue at short term (CFQ p<0.001; SF-36-PF p<0.01), but approximately a quarter of participants in the intervention group reported deterioration in physical functioning (SF-36-PF reduction of 10 points).
• Qigong trials in CFS patients (3 RCTs, n=154, 150, 70) showed post-treatment fatigue reduction (CFQ p<0.001) and no adverse events, with one yoga RCT (n=30) explicitly noting absence of PEM after practice.
• In the Nùñez multidisciplinary rehabilitation trial (CFS, n=120), SF-36 bodily pain favored the control group at 12 months (p<0.05); the authors noted the intervention may have been harmful to some participants.

Limitations

• Only one RCT enrolled ME patients who met diagnostic criteria requiring post-exertional malaise, making it impossible to draw conclusions about physiotherapy for this specific group.
• Most RCTs relied on subjective patient-reported outcome measures; objective measures such as fitness tests and activity monitoring rarely showed significant or clinically meaningful improvement even when self-report measures did.
• Heterogeneity of comparison groups, sample sizes, follow-up durations, and diagnostic criteria prevented meta-analysis and limited cross-study comparisons.
• Adverse event reporting was inconsistent or absent in many trials, making it difficult to quantify harm; PEM occurrence was measured as an outcome in only 2 of 18 RCTs.

Why it matters

For patients

Patients with ME who experience post-exertional malaise should be cautious about graded exercise therapy, as there is no evidence it helps and good reason to think it may worsen symptoms; activity pacing and self-management education are safer approaches focused on staying within energy limits.

For clinicians

Physiotherapists should not extrapolate evidence from broad chronic fatigue trials to ME patients diagnosed with criteria requiring post-exertional malaise, and should prioritise symptom management, pacing guidance, and harm avoidance over exercise-based rehabilitation.

For readers

This review highlights a critical gap in ME research: almost all physiotherapy trials enrolled patients without a firm requirement for post-exertional malaise, so the existing evidence base cannot be assumed to apply to the most symptomatic and diagnostically distinct ME population.