CFS subjects were recruited from notices placed in medical surgeries and by advertisements in local newspapers. Eighty-two patients (aged 16–74 years) applied to participate. Before participation, written confirmation of a CFS diagnosis, as defined by Fukuda et al,11 was required from each subject’s doctor. Fourteen applicants were excluded because of alternative diagnoses or failure to provide written confirmation of diagnosis. The remaining 68 patients were randomised (by an independent investigator) to a graded exercise or a relaxation/flexibility intervention. Six subjects withdrew from baseline testing for reasons not associated with the study, and a further subject was excluded because her body mass index (44 kg/m2) prevented her from participating in the exercise test. This left 32 CFS subjects in the graded exercise group (27 women, 5 men), and 29 CFS subjects in the relaxation/flexibility group (20 women, 9 men) (Box 1).

Based on an assumption of moderate effect size (0.4–0.6), a sample size of at least 30 subjects was needed in each intervention group to detect an effect at an α of 0.05 with 80% power.12

All subjects were assessed for depression as defined by the Diagnostic and statistical manual for mental disorders13 using a computerised version of the Composite International Diagnostic Interview.14

For the 4 weeks before and 4 weeks after the intervention, subjects attended weekly testing sessions at a university laboratory, on the same weekday at the same time. These assessment results were averaged for each 4-week period. The same exercise physiologist conducted all tests.

Resting heart rate (HR) and blood pressure (BP) were first recorded, then subjects were assessed on the Aerobic Power Index test,15 a submaximal cycle test that has been shown to be reliable in a CFS population (intraclass correlation coefficient, 0.97).16 This exercise test starts at low power and increases in small power increments until an individual target heart rate (THR: 220 – age × 0.75) is reached.15 The power output adjusted for body weight (W·kg-1) that coincided with a subject’s THR was determined through interpolation techniques described by Telford and Minikin.17 If a subject failed to reach the THR, the W·kg-1 achieved during the last full minute of exercise was recorded.

Ratings of perceived exertion (RPE), as measured by the Borg scale,18 were recorded 55 seconds into each minute of the exercise test. Oxygen consumption was analysed continuously during the exercise test by a metabolic cart consisting of a computerised online system. Blood lactate was measured before and 3 minutes after the exercise test.

A similar interpolation procedure to that used to calculate power output at THR was used to determine RPE, oxygen uptake and respiratory exchange ratio values that coincided with each individual’s THR. RPE values that equated to the subject’s THR were divided by the final workload (W·kg-1) achieved either at or before THR. This ratio represents perceived exertion sensed at the end of the exercise test adjusted for the actual work done.

Weekly energy expenditure was determined using the Older Adult Exercise Status Inventory.19 This scale excludes activities involving sitting or lying down.

Mental and physical fatigue was assessed using a 14-item self-rating scale,20 and anxiety and depression were rated by the Hospital Anxiety and Depression Scale.21

Cognitive function was tested using a computerised version of the modified Stroop Colour Word test. This test has been used previously to assess attentional function in CFS.22-24 We used two levels of difficulty based on speed of presentation.

On completion of the 12-week intervention, subjects indicated whether they felt the program had benefited them, using a global impression scale25 from 1 (“Very much better”) to 7 (“Very much worse”).

Initial exercise duration was between 5 and 15 minutes, and intensity was based on the mean HR value achieved midpoint during the submaximal exercise tests. Graded exercise consisted of an aerobic activity that used the major large muscles of the body. Subjects could choose walking, cycling or swimming. Subjects were instructed to exercise every second day, unless they had a relapse. If this occurred, or if symptoms became worse, the next exercise session was shortened or cancelled. Subsequent exercise sessions were reduced to a length that the subject felt was manageable. This form of exercise, which allows for flexibility in exercise routines, is known as pacing.26

Each subject was supplied with a small laminated Borg scale, and an HR monitor to help them reach and maintain their required HR goals. Subjects rated the effort of each exercise session and recorded their exercise details in a diary.

Subjects were contacted by phone every second week over the 12 weeks to review their progress and to determine their exercise regimen for the following fortnight.

The relaxation/flexibility group served as an attention-placebo control. Subjects were required to listen to a relaxation tape, and perform selected stretching exercises every second day for 12 weeks. All subjects kept a diary recording their relaxation/flexibility sessions.

Subjects were contacted by phone every second week to review their progress and to discuss the flexibility regimen for the following fortnight. Subjects participating in relaxation/flexibility were specifically requested not to participate in any extra physical activity while they were enrolled in the study.

The exercise physiologist attempted to spend the same amount of time on the phone with all subjects in both therapy groups.

Preliminary screening of the data indicated that all continuous variables were normally distributed. Analyses were conducted using 2 (group) × 2 (time) mixed-model ANOVAs, with repeated measures on the time factor.27 Significant interaction effects were examined further by post hoc comparison of group differences using F tests. Differences in ability to reach target HR during the exercise test were assessed using a Mann–Whitney non-parametric test. Results for the self-rated clinical global impression change score were collapsed in a 2 × 2 contingency table and compared using a χ2 test corrected for continuity. All statistical procedures were performed using SPSS.28

This project was approved by the University of Western Australia Human Research Ethics Committee.

Physiological results are shown in Box 2 and Box 3. Repeated-measures ANOVAs showed significant treatment by time interactions for all variables except resting diastolic BP.

We observed a significant improvement in resting systolic BP (P = 0.018) and in W·kg-1 (P = 0.019) in the exercise group. Improvement in W·kg-1 was reflected by a significant increase in net lactate production (P = 0.036), as well as by the number of exercise tests on which THR was achieved by the exercise group during post-intervention testing (P = 0.030) (Box 4). Oxygen uptake values were 9.6% higher after the intervention in the exercise group compared with an 8.9% decline in the relaxation/flexibility group, but the difference in final values for the groups was not significant. Of interest, ratings of perceived exertion were lower after the exercise intervention (P = 0.013). Activity levels increased in the graded exercise group, although the final levels did not differ between the groups (Box 2).

Mean psychological scores are shown in Box 5. Post-hoc analysis showed that, while baseline depression scores were similar in both groups, scores were significantly lower in the exercise group following the intervention (P = 0.027). While there were no other significant post-intervention differences, follow-up tests showed a significant improvement in mental fatigue over time in the exercise group (P < 0.001).

On the modified Stroop Colour Word test, there were no significant differences between the groups before the intervention, but a significant difference in favour of the graded exercise group after the intervention on the more difficult level of this test (P = 0.029) (Box 5).

Box 6 shows the scores on the self-rated clinical global impression change scale. Although there was no significant difference between the two groups (χ2 = 1.46; df = 1; P = 0.227), a greater percentage of subjects in the exercise group rated themselves as being better (exercise group, 29/32 [91%]; relaxation group, 22/29 [76%]).