CORRESPONDENCE

the very strong correlation be- tween the patients’ symptomatic improvement and their belief (al- beit erroneous) at the end of the trial that they had received active immunotherapy-we suggest that the patient group found it similarly hard. Indeed, we are tempted to suggest that this is a clear benefit for cognitive therapy!

We agree with several other con- clusions reached by the authors. We have received extensive ad- verse criticism of our program un- der the misguided apprehension that graded activity would be harm- ful in CFS. We are delighted that Lloyd and colleagues have shown these fears to be groundless. Second, we agree that our study populations may well be different. Six of our patients in the pilot study needed inpatient treatment be- cause they were partially or com- pletely restricted to a wheelchair.’ Our impression, following helpful conversations with Drs. Lloyd and Hickie, is that we may see a more disabled group of patients with higher rates of psychiatric disorder than they do. We also believe that the differences in disability stem from the fact that Australian re- searchers, led by Lloyd and col- leagues have convincingly demon- strated that the fatigue in CFS is not due to muscle dysfunction. The opposite view, which we consider both erroneous and dysfunctional, remains widespread in our coun- try. Finally, we agree without reser- vation that the benefits of CBT in our open study may still have re- flected nonspecific effects of treat- ment, likely to be important in a condition with such a high placebo response as CFS.

We are now addressing this is- sue in a trial comparing 12 sessions of CBT with an equal amount of a control treatment-namely, relax- ation therapy.

In conclusion, the authors state that their findings do not support the hypothesis that CFS could be adequately treated by CBT alone.

We do not feel that this study is an adequate test of CBT in CFS. We also doubt that either a specific cause or a specific treatment will ever be found in a heterogeneous condition such as CFS.

Trudie Chalder Alicia Deale

Simon Wessely Academic Department of

Psychological Medicine Isaac Marks

Institute of Psychiatry King’s College School of Medicine and Dentistry

London, United Kingdom

1. Lloyd AR, Hickre I, Brockman A, et al. lmmuno logic and psychologic therapy for patients wrth chronic fatigue syndrome: a doubleblind placebo controlled trial. Am J Med. 1993;94:197-203. 2. Butler S, Chalder 1, Ron M, Wessely S. Cognitive behavior therapy in the chronic fatigue syndrome. J Neural Neurosurg Psychiatry. 1991;54:153-158. 3. Chalder T, Butler S, Wessely S. Inpatient management of chronrc fabgue syndrome. Behav- ioral Psychotherapy. In press. 4. Wessely S, Butler S, Chalder S, David A. The cognitive behavioral management of the post-viral fatigue syndrome. In: Jenkins R, Mowbray J, eds. The Post-viral Fatigue Syndrome (M.E.). Chichester UK: John Wiley; 1990:305-334. 5. Lloyd A, Gandevia S, Hales J. Muscle perform- ance, voluntary activation, twitch properties and perceived effort in normal subjects and patients with the chronic fatrgue syndrome. Brain. 1991; 114:85-98.

Manuscript submitted March 9, 1993, and accepted May 10, 1993.

To the Editor: Lloyd et al’ concluded that pa-

tients with chronic fatigue syn- drome (CFS) did not respond to psychological therapy. Specifically they claim to have found that cog- nitive behavioral treatment (CBT) was no more effective than ordi- nary medical follow-up. An exam- ination of the design of this study, however, indicates that only lii- ited conclusions can be drawn.

Cognitive behavioral treatment is not a specific therapy but rather a category of treatments which can be administered in differing forms and doses. There are strong argu- ments to suggest that CBT may be effective in patients with CFS and a preliminary study suggested it to

be effective in at least a proportion of patients.”

The form of CBT evaluated by Lloyd et al was based on the as- sumption that the principal factor perpetuating illness in CFS is a simple phobia of activity. In our experience, this model is inade- quate. We have found that patients’ belief that their symptoms are a re- sult of physical disease and require medical treatment is also an im- portant factor perpetuating illness’ by preventing the patient working on psychological and social prob- lems inhibiting recovery. Not only did the patients belief in physical disease appear to be unchallenged in this study, it may even have been unwittingly confirmed by the si- multaneous administration of im- munoglobulin injections. Although the majority of patients were de- pressed, the therapy did not specif- ically address this problem. Fur- thermore, the “dose” of CBT was far too small (6 sessions of less than 1 hour for patients who had been ill for up to 28 years) to real- istically expect much effect,

The results obtained by the study are in keeping with these criticisms. The therapy failed even to achieve an increase in the pa- tient’s level of activity (suggesting noncompliance with the behav- ioral program) and improvement was correlated with the patient’s belief that they had received active immunotherapy (which suggests patients believed this to be a more appropriate treatment for their physical illness). The investigators are to be congratulated for at- tempting a controlled evaluation of psychological therapy for CFS. Unfortunately the type and dose of therapy given was inadequate. The role of CBT in this condition re- mains to be evaluated.

Michael Sharpe Wa)rneford H0spita.b

Oxford, United Kingdom

1. Lloyd AR, Hrckre I, Brockman A, et al. Immuno- logic and psychologic therapy for patients with chronic fatrgue syndrome: a double-blind placebo-

420 April 1995 The American Journal of Medicine@ Volume 98

CORRESPONDENCE

controlled trial. Am J Med. 1993;94: 197-203. 2. Sharpe M. Nonpharmacological treatment. In: Chronic Fatigue Syndrome, ClBA Symposium 173. Chichester, U.K.: Wiley; 1993. 3. Butler S, Chalder T, Ron ME, et al. Cognitive behavior therapy in chronic fatigue syndrome. J Neural Neurosurg Psychiatry. 1991;54: 153-158. 4. Sharpe MC, Hawton KE, Seagraott V, et al. Pa- tients who present with fatigue: a follow up of refer- rals to an infectious diseases clinic. BMJ. 1992; 305:147-152.

Manuscript submitted March 23, 1993, and accepted May 10, 1993.

The Reply: The criticisms raised by Chalder

et al and Sharpe need to be exam- ined within their proper context. Both groups have been strong ad- vocates, largely on theoretical grounds, of nonpharmacological treatments for patients with CFS.‘,’ Both have placed great emphasis on the initial study by the former group.:’ Although that “trial” con- sisted initially of 50 patients, only 32 (64%) actually accepted treat- ment. Outcome data was reported for treated patients only. Five pa- tients withdrew after commencing treatment. Paradoxically for a “trial of CBT,” 20 patients also received tricyclic antidepressants. Six of the patients were also hospitalized for a “period between three to eight weeks”. Eventually, 46% (23/50) of the original cohort reported them- selves “better” or “much better”, while only 18% [S/SO) of the pa- tients in this trial of mixed antide- pressant therapy/CBT/hospitaliza- tion met the authors’ criteria for complete recovery. As such, this study was described by the inves- tigators as an evaluation of “stan- dard clinical management”. In their own words it was “thus, subject to all the deficiencies of a non-blind, non-randomized study.”

Our study did not simply assess the efficacy of either (immunologic or psychologic) therapy at the end of the active treatment, but rather at follow-up after a 3-month inter- vention-free period. This created a strict-but clinically meaningful- outcome criterion for both thera-

pies being evaluated. Both corre- spondents suggest our cognitive- behavioral treatment consisted of too few sessions. In the study of Butler et al3 the mean number of therapist hours per treated patient was 7.5. After correction for the ac- tual number of enrolled subjects, this drops to 4.8 hours per patient. In our study, therapists spent 4 to 6 hours with each patient (a l-hour initial assessment, plus 5 sessions of 30 to 60 minutes). The duration of treatment (10 weeks) was as long as that encountered in most CBT programs for psychological disorders. It does remain possible that an increased “dose” of CBT may have efficacy for some pa- tients with CFS.

Our model for CBT in patients with CFS, like that described by Butler et al,” drew heavily on the re- lated field of CBT for patients with chronic pain. The program was de- signed by a psychiatrist (I.H.) who has extensive experience with pa- tients with CFS and with cognitive- behavioral approaches to the treat- ment of depression. The treatment was administered by experienced psychiatrists. Treatment plans for each patient were discussed by the 3 psychiatrists and were individu- alized to take into account the de- gree of disability and psychopath- olog~ in each case. A detailed manual was presented to the pa- tient outlining both the rationale for treatment and the proposed method (see Appendix of the arti- cle). A family member was enlisted to help supervise the program. An individually designed exercise pro- gram was recommended for each patient. Patients beliefs regarding key attitudes such as the possible negative effects of physical activity were challenged. A model of com- mencing and ceasing activity ac- cording to normal social cues, rather than according to the onset of fatigue, was repeatedly dis- cussed. Progress was monitored at each session. Clearly the therapy administered in our trial consti-

tutes more than adequate CBT, and the lack of efficacy cannot simply be attributed to perceived “defic iencies” with our treatment package.

Both correspondents adhere to a model of CBT where it is neces- sary for patients to trade one un- substantiated dogma (“that their illness is entirely physical”) for an- other (that their disability is deter- mined largely by “psychological or social problems”). For example, Sharpe castigates us for not chal- lenging “the patients’ belief in physical illness” and “unwittingly confirming” this illness attribution by administering immunotherapy. Within our model such dualism is fiiy rejected in favor of an eclec- tic approach that ls consistent with available data, plausible and ac- ceptable to patients (unlike the ap- proach of Butler et al). In tying CBT to immunologic therapy, we maximized patient presentation for psychological intervention. Our approach simply seeks to ex- pand the patient’s view of those factors contributing to their ongo- ing disability (ie, emphasizing psy- chosocial as well as physical fac- tors). Consistent with our wider view, patients accepted two differ- ent types of treatment within a properly constructed trial.

Chalder et al take issue with our use of hours of activity as an out- come variable, suggesting that the lack of change in the mean value on this measure at the 3-month fol- low-up should be viewed as an in- dication of noncompliance. Given that reduced activity is one of the principal measures of disability in this disorder, we decided that treat- ment efficacy should be judged by this criterion in addition to Karnofsky performance scores and subjective reports of fatigue and other variables. Had we also relied simply on an assessment at the end of the treatment, with a purely sub- jective measure of the patients “ability to work” or “ability to par- ticipate in physical tasks,” we too

April 1995 The American Journal of Medicine@ Volume 98 421