Correspondence Cognitive behaviour therapy for chronic fatigue syndrome -------------------------------------------------------- bron: The Lancet datum: 21 juli 2001 Sir--I have deep concerns about Judith Prins and colleagues' report (March 17, p 841).1 They selected patients by use of only the major criteria of the Centers for Disease Control definition,2 which includes patients with chronic fatigue or idiopathic chronic fatigue, but not those with chronic fatigue syndrome (CFS). They cannot, therefore, claim to have done a trial on or apply their results to CFS patients. Cognitive behaviour therapy (CBT) was administered as 16 1-h sessions over 8 months, which differs from the common pattern.3 95% of the audiotaped sessions were not assessed for adequacy. The 23 patients who withdrew from the CBT group should have been included for comparison at months 8 and 14. The 28% rate of drop out is too high to allow a final outcome analysis, and the withdrawal rates were not evenly distributed across the three groups. Since the number of withdrawn patients (23 of 82) was similar to the number of patients who benefited from intervention in this group (29 of 82), the primary outcome analysis was misleading. Thus, modified intention to treat analysis should have been done. If this was applied, the benefit of CBT would lose statistical advantage over the natural course in terms of self-reported improvement at the end of 14 months. Furthermore, at 8 months, Prins and colleagues showed data for 83 patients in their assessment of fatigue but only for 71 and 74 patients for Karnofsky score and self-rated improvement, respectively (in the report's table 4), in the same CBT group. Also, the numbers of patients per group analysed at the end of trial (58, 62, and 76 in the report's table 4) differs from those who completed the trial (55, 61, and 70, figure 2). It is unclear who administered the assessments for CIS fatigue, Karnofsky score, and so on. Functional assessments should be administered by physicians or people who are unaware of the patients' treatment assignments.4 The seemingly open design of the trial has serious limitations because of bias towards the psychological model of CFS. The investigators misdiagnosed at least seven of 278 patients who were specifically screened for this trial (error rate 2·5%). In my own practice, at least 40% of all physician-referred patients with a misdiagnosis of CFS have other medical disorders, most commonly depression. Indeed, the response rate of chronic fatigue patients to CBT in Prins and colleagues' trial is similar to that reported for depressed patients to CBT monotherapy.4 Did the trialists unwittingly choose patients with chronic depression and fatigue (rather than CFS), especially since no screening was done for depression? Finally, even if about 50% of CFS patients were assumed to truly benefit from appropriately administered CBT (although the trial does not provide any credible evidence for this proportion), I fail to understand how this proves CFS to be a functional somatic syndrome as claimed. Although CBT is useful, for example, in depression,4 there is a neurobiological model of depression and, also, of chronic fatigue.5 Successful pharmacotherapy of depression has led to new lines of research that would not be conceivable on a pure psychosocial paradigm. The case for CFS might not be very different. Diluting this difficulty would be a poor science. Abhijit Chaudhuri Department of Neurology, Institute of Neurological Sciences, Southern General Hospital, Glasgow G51 4TF, UK 1 Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 2001; 357: 841-47. [Text] 2 Fukuda K, Strauss SE, Hickie I, et al. The chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med 1994; 121: 953-59. [PubMed] 3 McCullough JP. Therapist manual for cognitive behavioural analysis system of psychotherapy. Richmond: Virginia Commonwealth University, 1995. 4 Keller MB, McCullough JP, Klein DN, et al. A comparison of nefazodone, the cognitive behavioural-analysis system of psychotherapy, and their combination for the treatment of chronic depression. N Engl J Med 2000; 342: 1462-70. [PubMed] 5 Chaudhuri A, Behan PO. Fatigue and basal ganglia. J Neurol Sci 2000; 179: 34-42. [PubMed] --------------------------------------------------------------------------- Sir--The title of Judith Prins and colleagues' study1 suggests they studied patients with CFS. However, only patients with the symptoms of chronic fatigue were studied. According to the Centers for Disease Control and Prevention (CDC) criteria for CFS four of eight symptoms must be present.2 A syndrome is an accepted combination of symptoms and signs, and defines the group of patients that are studied. By neglecting the CDC criteria, Prins and colleagues' findings cannot be compared with those from other studies of patients meeting those criteria. Another disadvantage of the study is that measurements were restricted to 8 or 14 months, which might have introduced bias into the results. The explanation of the missing data is insufficient, and a new calculation with estimated data, including a worst case scenario, would be worthwile. Such calculations will probably decrease the apparent efficacy of CBT. Information about the number of patients with increasing fatigue during the study would also be valuable. At 8 months, the number of improved patients by the behavioural therapy minus those improving without therapy was 17 for CIS fatigue, 20 for Karnofsky score, and 19 for self-rated improvement. At 14 months these numbers were seven, 11, and five, respectively, accompanied by striking increases in the p values. The results are, therefore, more interesting after a long period. *Ruud C W Vermeulen, Hans R Scholte, P Dick Bezemer *CFS Research Centre Amsterdam, Waalstraat 25, NL-1078 BR Amsterdam, Netherlands; Department of Biochemistry, Erasmus University Rotterdam; and Department of Clinical Epidemiology and Biostatistics, Vrije Universiteit, Amsterdam (e-mail:rcwvermeulen@intermedclin.com) 1 Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 2001; 357: 841-47. [Text] 2 Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. The chronic fatigue syndrome: a comprehensive approach to its definition and study--International Chronic Fatigue Syndrome Study Group. Ann Intern Med 1994; 121: 953-59. [PubMed] ---------------------------------------------------------------------------- Sir--Although Judith Prins and colleagues1 report various benefits from CBT in CFS, their methods and interpretation of results raise two important questions. First, is this type of expensive and difficult to obtain hospital-based treatment any more effective than active management in a primary-care situation? This comparison is important but is not measured by Prins and colleagues or in any other CBT trial involving CFS patients. Family physicians and members of primary-care health teams who are competent at assessing and managing CFS patients are capable of providing advice about the need to balance rest with appropriate increases in activity, symptomatic relief for pain, sleep disturbances and so on, guidance on education and employment, social benefits, &c, and access to providers of disability services and social support. All these factors are relevant to the recovery process in CFS, and it is astonishing that the only forms of outcome comparison done by Prins and colleagues were groups who received non-directive support from social workers, or who were left to follow a natural course with no interventions. This continued failure to obtain information on the outcome of patients receiving active management in primary care has important logistical and cost implications for service providers. The minimum prevalence of CFS is around 0·4% of adults (ie, up to 200 000) in the UK; there are too few therapists to take on this extra workload. Even if there were enough, the cost to the National Health Service, at more than UKŁ1000 per course of CBT, would be prohibitive. Second, Prins and colleagues conclude that CBT could be successfully and effectively administered by therapists with no previous experience in CBT (did they mean CFS?). This finding is in contrast to results obtained from a questionnaire on treatment outcomes.2 Of 2338 respondents, 285 had received CBT. Only 21 of 285 (7·4%) found CBT helpful, whereas 191 of 285 (67%) reported no change. 26% of respondents thought their disorder worsened after CBT. This survey supports the view that CFS is a heterogeneous illness in clinical presentation and possible cause. Although the purely psychosomatic explanation favoured by Prins and colleagues might apply to a subgroup of patients with CFS, other cases with neurological,3 muscular,4 and endocrine5 abnormalities cannot be adequately explained by this model. Underlying organic pathology might help to explain why those who do benefit from CBT improve only slightly, with objective evidence of more substantial and sustained recovery (ie, return to normal employment) being rare. Charles Shepherd ME Association, Stanford le Hope, Essex SS17 OAH, UK 1 Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 200l; 357: 841-47. [Text] 2 The severely affected: a survey of members of Action for ME. London: Action for ME, 200l. 3 Keenan PA. Brain MRI abnormalities exist in chronic fatigue syndrome. J Neurol Sci l999; 171: 1-2. [PubMed] 4 Lane R. Chronic fatigue syndrome: is it physical? J Neurol Neurosurg Psychiatry 2000; 69: 289. 5 Scott LV, Teh J, Reznek R, Martin A, Sohaib A, Dinan TG. Small adrenal glands in chronic fatigue syndrome. Psychoneuroendocrinology l999; 24: 759-68. [PubMed] ---------------------------------------------------------------------------- Sir--After hearing a presentation on Judith Prins and colleagues' study1 at the American Association for Chronic Fatigue Syndrome (AACFS) International Conference, Seattle, in January, 2001, I find disturbing the lack of full disclosure. At the AACFS, a question was asked about the length of benefit for CBT. The presenter stated that the natural course and CBT groups did not differ significantly 3 years after treatment. Second, there seems to be undeclared sampling bias: patients had to be sufficiently healthy to handle 1·5 h travelling time. With CFS, this stipulation causes a strong sampling basis to the healthiest patients. Around one patient failed to complete the study for every 1·4 patients completing CBT. I was disappointed that there was no analysis of factors between these two groups. Statistical experience suggests that CBT benefits might apply only to a subgroup. Last, and probably most importantly, Prins and colleagues did not show the effectiveness of CBT for patients with CFS because they did not include trials against the three treatments that have reported far greater results: ribonuclease-L therapy; anticoagulant therapy (Hemnex Laboratories report a 75% frequency of hereditary coagulation defects in CFS patients and a 90% incidence of hypercoagulation); and antibiotic treatment. Inclusion of these studies in the trial would have confirmed or rebutted the conclusions of Wessely and colleagues.2 I am working as a medical statistician with a local physician in a study of 14 patients with CFS and fibromyalgia selected for hypercoagulation testing. 100% have two or more coagulation assays at least 2 SD outside the normal range, and more than 90% have hereditary coagulation defects. Treatment by low-dose heparin or enoxapain has been effective, especially when combined with alternating antibiotic treatment after 2 months of heparin. We are setting up a rigorous study to investigate these anecdotal findings. I am in favour of CBT or equivalent as an adjunct to other biological treatments because of the apparent relation between cognitive behaviour, stress, adrenalin, and the families of chronic infections associated with CFS. I believe that the temporal benefits of CBT are real, but without treatment of organic issues, these gains will be lost over time. Kenneth M Lassesen PO Box 170, Kingston, WA 98346, USA (e-mail:KenL@exmsft.com) 1 Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 200l; 357: 841-47. [Text] 2 Reid S, Chalder T, Cleare A, Hotopf M, Wessely S. Chronic fatigue syndrome. BMJ 2000; 320: 292-96. [PubMed] ---------------------------------------------------------------------------- Sir--Judith Prins and colleagues' report1 leaves the clear impression that there is a powerful case for the provision of CBT as a specific therapy for CFS. However, careful assessment of published studies suggests that this impression is not evidence-based. The initial review of reports by Best and Stevens for the UK Research and Development Directorate, South and West, in 1996, showed five studies with limited methods from which little about efficacy could be concluded. Subsequently, a Cochrane review to 19982 could identify only three randomised controlled trials with acceptable methods: one supported3 and one did not support4 the use of CBT, although neither controlled for therapist exposure, and the third5 used a relaxation control. Thus, with Prins and colleagues' study, there are two randomised controlled trials in which CBT has been compared (in a pooled total of 113 patients) with an active, although not indistinguishable, intervention (110 patients). The two trials show significant benefits on the primary outcome measures, but conclusions about efficacy must be tentative in view of the paucity of trials; the small number of patients involved; the difficulties inherent in comparing CBT--which included a graded exercise component in both trials--with control interventions, such as relaxation or group support; and, importantly, the potential effect of publication bias. No conclusions can be made about the effectiveness of group CBT (the only cost-effective option in the long term) or the generalisability of CBT to the various subgroups of patients, such as those with severe or long-term disability. We suggest that the state of current evidence could be more rigorously described as follows: in the absence of any available medical treatment for CFS at present, these two trials together provide a small amount of evidence that CBT (or an equivalent beneficial patient-therapist encounter given on a one-to-one basis) can improve, but not cure, some physical symptoms in some members of the subgroup of CFS patients well enough to attend an outpatient clinic. This description is neither crisp nor appealing, but it more closely represents the true evidential picture. *Vance A Spence, Neil C Abbot *University Department of Medicine, Ninewells Hospital, Dundee DD1 9SY, UK; and Department of Epidemiology and Public Health, University of Leicester, Leicester (e-mail:vance.spence@ntlworld.com) 1 Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 200l; 357: 841-47. [Text] 2 Price JR, Couper J. Cognitive behaviour therapy for adults with chronic fatigue syndrome (Cochrane Review). In: The Cochrane Library, Issue 2. Oxford: Update Software, 2000. 3 Sharpe M, Hawton K, Simkin S, et al. Cognitive behaviour therapy for the chronic fatigue syndrome: a randomized controlled trial. BMJ 1996; 312: 22-26. [PubMed] 4 Lloyd AR, Hickie I, Brockman A, et al. Immunologic and psychologic therapy for patients with chronic fatigue syndrome: a double-blind, placebo-controlled trial. Am J Med 1993; 94: 197-203. [PubMed] 5 Deale A, Chalder T, Marks I, Wessely S. Cognitive behaviour therapy for chronic fatigue syndrome: a randomized controlled trial. Am J Psychiatry 1997; 154: 408-14. [PubMed] ---------------------------------------------------------------------------- Sir--Judith Prins and colleagues1 purport that psychological treatment improves physical symptoms of CFS. As someone recovered from CFS thanks to an old remedy for Addison's disease2 and who has written about the growing overlap between these two disorders, which share at least 40 clinical features,2,3 I judge their conclusion implausible. I do not believe that abnormalities that underlie both CFS and Addison's disease, namely, hypocortisolism, impaired adrenal cortical function, reduced adrenal gland size, and antibodies to the adrenal gland2,3 can be improved by CBT. All the physical and neuropsychological symptoms listed in the criteria for CFS can also be found in Addison's disease.2,3 Prins and co-workers, by arbitrarily ignoring such physical symptoms as lymphadenopathy, myalgia, arthralgia, sore throat, and postexertional malaise, might have biased the results of their study. Without those physical symptoms, it is hard to distinguish CFS reliably from depression. They might, therefore, have inadvertently included several patients with depression in their study. Patients with depression, however, are more likely than those with CFS to respond to psychological treatments, which could have biased Prins and colleagues' results. The same bias probably occurred in the studies that reported CBT to be effective for patients meeting the Oxford criteria for CFS,4 because these criteria ignore the physical symptoms that distinguish CFS from depression. To date, no evidence shows that CBT is effective for patients fulfilling the original criteria for CFS. In view of the impressive overlap of CFS with Addison's disease,2,3 it is hardly surprising that hydrocortisone5 and fludrocortisone,2 which are routinely used in combination by patients with Addison's disease, have proven beneficial for patients with CFS too, even though the two steroids were studied separately.3 It can reasonably be predicted that the combined administration of hydrocortisone and fludrocortisone will prove to be the best treatment for patients with CFS.2-5 Riccardo Baschetti CP 1011, 35100 Padua, Italy (e-mail:baschetti@shineline.it) 1 Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 2001; 357: 841-47. [Text] 2 Baschetti R. Orthostatic hypotension and chronic fatigue syndrome. JAMA 2001; 285: 1441. 3 Baschetti R. Chronic fatigue syndrome: a form of Addison's disease. J Intern Med 2000; 247: 737-39. [PubMed] 4 Baschetti R. Fibromyalgia, chronic fatigue syndrome, and Addison disease. Arch Intern Med 1999; 159: 2481. 5 Baschetti R. Hydrocortisone and chronic fatigue syndrome. Lancet 1999; 353: 1618. ---------------------------------------------------------------------------- Authors' reply Sir--We did not select relatively healthy patients with chronic fatigue without additional symptoms rather than patients with CFS. Apparently our statement that we did not include the CDC criterion of foura additional symptoms has led to misunderstanding. We show that the number of additional symptoms is dependent on the assessment method, and fatigue severity did not differ between patients who fulfilled the CDC criteria and patients who did not.1 We did assess additional symptoms at baseline. In our sample, 252 patients had the diagnosis CFS and 18 idiopathic chronic fatigue. Cut-off scores for fatigue severity and functional impairment guaranteed that all these patients were severely disabled. Our choice of control conditions is questioned. Support groups and natural course were closest to the health-care situation of CFS patients in the Netherlands. For the intention-to-treat population we used the normal definition--all patients randomised. Since the variables we analysed were different from baseline values, the actual population did not include patients with missing values at 8 months, 14 months, or both. Unlike Abhijit Chaudhuri suggests, we included patients who withdrew from treatment in the analyses. There is no discrepancy between numbers of patients--perhaps his undergraduate students will notice. Before the trial, we decided not to use imputation techniques to estimate missing data, since bias can also result from use of such data. Contrary to Ruud Vermeulen and colleagues' opinion, we provided enough circumstantial evidence to substantiate our results and explain the missing data. We used proper methods and the results clearly show that patients receiving CBT improved more, in masked and unmasked functional assessments, than patients in the non-CBT groups. Like others, we are interested in the long-term results of CBT. The first results on lasting benefits of CBT after 5 years are promising.2 At the end of our study, CBT was offered to all patients in the control conditions, and long-term comparison between CBT and natural course is impossible. This was not reported differently at the AACFS. Unlike Charles Shepherd, Chaudhuri, and Kenneth Lassessen suggested, the use of a psychological model does not preclude neurobiological components. There is still little consistent evidence for underlying organic pathology in CFS. Contrary to Vance Spence and Neil Abbot's suggestion, individual CBT and graded exercise have been proven helpful for CFS patients in six out of seven randomised controlled trials.3,4 The two treatments are comparable, since there is no CBT without graded exercise and no graded exercise without cognitive therapy. We agree that further identification of subgroups of CFS patients benefiting from CBT is needed. Also, less expensive treatments in primary care deserve attention in future trials. Riccardo Baschetti reiterates his well known statements on Addison's disease and CFS. Before we diagnose a patient as CFS, a thorough medical examination is done to rule out somatic disorders such as Addison's disease. In addition, a structured clinical interview for the revised DSM III criteria was done to prevent confounding of CFS and psychiatric diagnoses. The findings in CFS about subtle changes in the hypothalamus-pituitary-adrenal axis (of which pathogenetically the importance is unknown), have led to two randomised controlled trials, from which we and the investigators of these trials do not derive Baschetti's conclusion that steroids are treatment of choice.5 *J B Prins, G Bleijenberg, Th M deBoo, J W M van der Meer Departments of *Medical Psychology, Epidemiology and Biostatistics, and Internal Medicine, University Medical Centre, PO Box 9101, 6500 HB Nijmegen, Netherlands 1 Swanink CMA, Vercoulen JHMM, Bleijenberg G, Fennis JFM, Galama JMD, van der Meer JWM. Chronic fatigue syndrome: a clinical and laboratory study with a well-matched control group. J Int Med 1995; 237: 499-506. [PubMed] 2 Deale A, Husain K, Chalder T, Wessely S. Cognitive behaviour therapy versus relaxation for chronic fatigue syndrome: outcome at five year follow-up: final report for South Thames Research and Development Project Grant Scheme (London Region). 3 Powell P, Bentall RP, Nye Fi, Edwards RHT. Randomised controlled trial of patient education to encourage graded exercise in chronic fatigue syndrome. BMJ 2001; 322: 1-5. [PubMed] 4 Reid S, Chalder T, Cleare A, Hotopf M, Wessely S. Extracts from "Clinical Evidence": chronic fatigue syndrome. BMJ 2000; 320: 292-96. [PubMed] 5 Cleare AJ, Heap E, Malhi GS, Wessely S, O'Keane V, Miell J. Low-dose hydrocortisone in chronic fatigue syndrome: a randomised crossover trial. Lancet 1999; 353: 455-58. [Text]