Introduction

There are several conditions where physicians' and patients' views diverge, especially regarding etiology, treatment, and the relevance of psychogenic factors. The best-known examples are fibromyalgia and irritable bowel syndrome. In the case of chronic fatigue syndrome (CFS), an attempt was made in 2001 to reconcile different viewpoints by means of a working party that included patients and their advocates, along with physicians and medical researchers ¹. These attempts were only partially successful -- common ground was limited to a general statement about the effects of the illness on the patient and the family, the fact that it can occur at any age, and that it affects both sexes.

Despite differences of opinion about etiology, a recent review of treatments published in JAMA has helped understanding of what may work, and what won't. For this review, the authors considered CFS to mean all illnesses, including myalgic encephalitis (ME), chronic Epstein-Barr syndrome, and postviral fatigue syndrome (PVFS), with symptom complexes similar to CFS -- namely fatigue, headaches, sleep disturbances, difficulties in concentration, and muscle pain, with physical and mental fatigue lasting at least 6 months as the defining characteristic.

Method

The authors searched 19 specialist databases for published or unpublished studies on interventions for the treatment of CFS in adults or children. Controlled trials - randomized or nonrandomized -- were selected according to established inclusion criteria; trials including patients with fibromyalgia were not selected for analysis. There were 36 randomized (10 with a cross-over design) and 8 non-randomized trials included in the final analysis. After data extraction, a 'validity score' was assigned to each trial, based on generally accepted validity criteria for clinical trials; the maximum score was 20.

Studies were classified into those showing any (or some) effect, an overall effect, or no effect of treatment. A total of 38 different outcome measures were evaluated; these fell into 5 categories; psychological, physical, quality-of-life, and physiological outcomes, and use of resources.

An 'any effect' study was defined as one with any outcome measure showing a significant difference between the intervention and control groups. An 'overall effect' study had more than one outcome measure significantly different between treatment groups. And 'no effect' meant that there were no significant differences between the groups for any outcome measures.

The 30-odd different interventions were grouped into 6 different categories: behavioral, immunological, pharmacological, nutritional supplements, complementary/alternative medicine (CAM), and other treatments.

Results

A total of 2,801 participants were included in the 44 trials analyzed. Ages ranged from 11 to 87 years, the proportion of women was 71% (range 19% - 100%), and the duration of illness varied from 27 days to 34 years.

Of the 44 trials, 29 (66%) showed some effect of the treatment, 18 (41%) an overall beneficial effect of treatment, while one showed a negative effect of the intervention.

The distribution of the individual study results and the average validity score are given in the following table according to the type of intervention:

Type of Rx.	No. Studies	'Any Effect'	'Overall Effect'	Validity Score
Behavioral	9	7/9	6/9	15
Immunologic	10	8/10	4/10	12
Pharmacologic	13	4/13	2/13	13
Supplements	6	4/6	2/6	12
CAM	3	2/3	2/3	5
Other types	3	3/3	2/3	3

Behavioral treatment, which came from better-validated trials and had a good overall effect (66%), consisted chiefly of graded exercise therapy or cognitive behavioral therapy. Cognitive behavioral therapy studies recorded the highest drop-out rate -- 19% -- with graded exercise therapy second (18%).

Immunological therapy consisted mostly of immunoglobulin treatment, with one acceptable trial each using gamma-globulin, alpha-interferon, Ampligen®, or terfenadine (now unavailable); results were less convincing than those for behavioral treatment, and side effects occurred in some participants.

Pharmacological interventions, which included hydrocortisone, fludrocortisone, or one of a number of other drugs, were, for the most part, very disappointing, and carried a risk of side effects. Two studies of essential fatty acid supplementation reported some beneficial effects, and magnesium was found to be effective in one study; these 3 studies were fairly well validated. CAM treatment didn't provide any convincing benefits -- although 2 of the 3 studies had an overall effect (massage in one, osteopathy in the other), these studies were not well validated. The 'other treatment' study results were also not compelling.

Conclusions

It seems fairly clear graded exercise therapy and cognitive behavioral therapy represent the approaches that have been best investigated, and that have the best chance of a beneficial effect. However, further studies are required to establish standard outcome measures, to allow adequate comparison of other therapies. CFS is a fluctuating disease, so that follow-up rates should be long enough to establish a successful treatment, rather than a temporary respite in the course of the disease.

The wide variety of interventions revealed in this analysis may reflect the uncertainty of the etiology, but it also indicates the dissatisfaction of patients with available treatments. Although most interventions were without side effects, the relatively high dropout rates with the behavioral therapy approaches suggest patients may have been unhappy with therapy directed at a likely psychogenic cause. All these factors make CFS a difficult disease to study, but any positive efficacy in such a debilitating condition carries its own rewards, and should spur further research.