To the Editor: A recent editorial1 and article2 continue to promulgate and link the unproven concepts that patients with chronic fatigue syndrome (CFS) are “deconditioned” and exercise is beneficial in treatment. The cited study by Fulcher and White3 is open to opposite conclusions, depending on their use of the outcome descriptor “better”. If the term is restricted to “much better” and “very much better”, then, as cited by Lloyd,1 16 of 29 people with CFS rated themselves as “better” after a graded exercise program, compared with only 8 of 30 in the control group who completed a flexibility treatment regimen. However, if the “better” descriptor combines “a little better”, “much better” and “very much better”, which is the interpretation used by Wallman et al,2 then the scores for the exercise versus flexibility groups are not different, being 27 of 29 and 26 of 30, respectively, agreeing with the conclusion of Wallman et al.2

Whichever interpretation is applied, any beneficial effect of the graded exercise program in people with CFS in these studies must be independent of any training effect or change in level of “conditioning”, as this was reported in one study,2 but not in the other.3

A fundamental flaw with most exercise studies in CFS is the use of submaximal or symptom-limited tests, which provide notoriously misleading data when compared with maximal exercise testing procedures.4,5 Wallman et al2 correctly identify maximal oxygen consumption as the “gold standard” measure of exercise capacity, yet such measurements were not made in the three articles they cited. When such procedures are applied, the exercise capacity of people with CFS is not significantly different from either measured or age-predicted values for healthy sedentary people.6 Wallman et al2 suggested that maximal testing procedures could favour the recruitment of “more robust or healthier” patients and provide misleading information. In the first place this is denied by the study of Sargent et al,6 in which the illness status reported by patients who completed the maximal tests was similar to that in previous CFS studies. In the second place, the maximal test protocol chosen for a given population should be designed to exclude any influence of fatigue on the metabolic measurements. This is confirmed by the results from the study cited,6 in which the metabolic measurements met the published criteria of a maximal test.4,5

In summary, patients with CFS are not “deconditioned”. Neither their muscle strength nor their exercise capacity is different from that of other sedentary members of the community (> 70%). We remain unaware of any incontrovertible evidence that the various “exercise training” programs suggested in previous articles improve either the physiological or clinical status of people with CFS.

To the Editor: The claim in Lloyd’s editorial1 that “the criteria for diagnosis are well accepted internationally” ignores the recent publication of the Canadian consensus guidelines for the diagnosis and management of myalgic encephalomyelitis/chronic fatigue syndrome,2 which were sponsored by Health Canada and written by an international group of well published researchers. The Canadian definition of chronic fatigue syndrome (CFS) requires the concurrent presence for six months of fatigue, post-exertional fatigue, sleep dysfunction, pain (including headaches) and neurological/cognitive manifestations, as well as at least one symptom from two of autonomic, neuroendocrine and immune manifestation categories (pp 12–13). These requirements add clinical specificity to the Fukuda criteria and exclude subjects who may have chronic fatigue for other reasons, such as psychiatric disorder without multiple physical symptoms.

Lloyd refers to the “recent refinements to improve reliability” in the revision of the research case definition by Reeves et al.3 The SPHERE screening instrument recommended by that article was designed for psychiatric screening in primary care. It arbitrarily classifies people with multiple physical symptoms, often severe in degree and associated with major disability, as having somatisation disorder. This is akin to subclassifying people with severe multiple sclerosis as having somatoform disorder and those with fewer and less severe symptoms as the “core” multiple sclerosis group, a finding which is not supported by the evidence.

Conclusions from the article by Wallman et al4 cannot be generalised to the severely ill. Recruitment was from “notices placed in medical surgeries and by advertisements in local newspapers”. Patients with severe CFS, who can barely venture outside their homes and are often too ill to read, would be unlikely to participate. Loblay, Chair of the Royal Australasian College of Physicians Working Group for CFS Clinical Practice Guidelines, urges caution about generalising from exercise studies, which never include people with severe CFS: “All these studies involve people willing and able to participate. The people who find it makes them feel lousy drop out.”5

Lloyd asserts exercise is no longer a question (“. . . graded physical exercise should become a cornerstone of the management approach for patients with CFS”). To promote such a strong, unqualified message to busy general practitioners who may be unfamiliar with the range of severity in CFS risks serious harm to patients.

In reply: Scroop and Burnet correctly identify the vagaries of the necessarily subjective measurement of outcomes in intervention studies of chronic fatigue syndrome (CFS). Given that muscle strength, endurance and recovery are essentially normal in patients with CFS,1 rather than become too focused on the best approach to measurement of exercise capacity the key issue is whether patients benefit in terms of self-reported symptom severity or functional status.

The weight of evidence indicates that graded physical exercise does provide such benefits. Whether this occurs via improvements in aerobic fitness or via the well-recognised psychological and social benefits of exercise is something of a side-issue.

Stein and Hunter draw attention to the recently published Canadian consensus guidelines for the diagnosis and management of myalgic encephalomyelitis/CFS. Although this document may provide a welcome recognition for Canadian patients with the disorder, unlike the Australian guidelines,2 it is devoid of an evidence base for the recommendations. Sadly, rather than “add[ing] clinical specificity”, it is also highly likely that the modified diagnostic criteria fall into the trap of preferentially identifying patients with somatisation disorder,3 as such individuals often report large numbers of unexplained symptoms, and hence the addition of 20 or more symptoms to the diagnostic criteria may well bias towards inclusion of such patients.

Stein and Hunter are incorrect in the assertion that SPHERE was designed for psychiatric screening in primary care, as the instrument arose out of our studies in CFS specifically seeking to identify clinically significant fatigue states.4

I support the recommendation about caution in generalising from existing published data regarding graded exercise to patients who are severely ill, as such patients are indeed likely to be under-represented in published studies. Nevertheless, it is noteworthy that the recommendations made in the Canadian document cited by Stein and Hunter also clearly support the notion of graded physical exercise: “Patients should gently and gradually increase their level of activity.” Thus, rather than leave the severely affected to continue to “barely venture outside their homes”, I would recommend a carefully designed graded exercise program in the home, with a goal of improving functional performance sufficiently to escape those confines.