Chronic fatigue syndrome (CFS) is defined as persistent or relapsing fatigue that has occurred for at least 6 months, is not alleviated by rest, and causes substantial reduction in activities. The fatigue cannot be explained by medical or psychiatric conditions and must be accompanied by at least 4 of 8 specified symptoms (unusual post exertional fatigue, impaired memory or concentration, unrefreshing sleep, headaches, muscle pain, joint pain, sore throat, and tender cervical nodes) 1. There is considerable discrepancy in results between studies from different institutions; so as yet, there are no characteristic signs or laboratory markers of CFS and its pathophysiology has not been elucidated 2.

This lack of diagnostic signs or laboratory markers notwithstanding, many manifestations of CFS resemble those of musculoskeletal and infectious diseases 3. In large part, the illnesses caused by these diseases reflect immune system activation and there is evidence for immune system dysfunction in some cases of CFS. In particular, antinuclear antibodies (ANA) and other common autoantibodies have been evaluated in people with CFS: unfortunately, with variable results. For example, one study found that 52% of tertiary care- CFS referral-patients had antibodies to nuclear envelope antigens 4 while another study found the same ANA antibody rates in both CFS and controls 5. Recently, investigators reported that antibodies to the human muscarinic cholinergic receptor 1 may provide a biologic explanation for the cognitive impairment observed in people with CFS 6. This lack of consensus between studies may in large-part reflect recruitment bias associated with studies of persons enrolled from tertiary referral clinics combined with imprecise evaluation of the illness and inadequate or inappropriate control populations.

We had the opportunity to measure the associations of common autoantibodies and autoantibodies to neuronal cell antigens and CFS in two case control studies; one of primary care patients with CFS who were identified by a physician surveillance network; the other a study of people with CFS identified from the community. The physician surveillance study was conducted 1988 through 1993 in Atlanta, Georgia 7 and the community study 1997 through 2000 and identified subjects with CFS from the general population of Wichita, Kansas 8.

Both studies rigorously classified people as CFS and controls in both studies were enrolled to represent the general population and matched to cases by sex, race, and age 910. The hypothesis of the present study is that the appearance of cell-specific autoimmune antibodies may define subsets of CFS and give clues to the etiology and pathogenesis and may help to explain the neurocognitive symptoms experienced by CFS patients. Secondarily, we wished to evaluate the extent to which patients with CFS who were receiving primary medical care treatment for CFS were similar to people with CFS in the community.

Although women predominated in both study groups other demographic and clinical characteristics differed and reflected basic differences between patients with CFS who obtain medical care and those in the general population (most of whom have not seen a physician) (Table 1). Of note, CFS cases from physician surveillance were somewhat younger than those identified in the population (mean 39 and 46 years, respectively) and controls were similarly different: those recruited in the physician study had been ill about half as long as those in the community (69 and 128 months, respectively) and were more likely to report sudden onset CFS (36.4%) than those in the general population with CFS (13.5%).

CFS is a complex, debilitating illness, which is characterized by at least 6 months of severe persistent or relapsing fatigue and a group of characteristic but nonspecific symptoms. Despite more than a two decades of extensive research, no diagnostic tests exist, and effective control and prevention remain elusive because the cause and pathophysiology of CFS remain unknown. CFS is clinically similar to several rheumatic autoimmune disorders that can be diagnosed and characterized by autoantibody profiles. For this reason, we conducted an exhaustive evaluation of 11 ubiquitous nuclear and cellular autoantigens in addition to two neuronal specific antigens.

The serum samples tested in this study were collected from a physician surveillance study conducted in Atlanta 15 and a population-based community study in Wichita 8. The physician surveillance study was conducted over 7 years and used approximately 70% of all primary care physicians in Atlanta. All patients were carefully evaluated for unexplained unwellness and fatigue. The community study was a random digit dial survey of 90,000 people (25% of the Wichita, Kansas population). All CFS cases were medically and psychiatrically evaluated and rigorously classified as CFS, other unexplained unwellness, or medically/psychiatrically explained unwellness. The serum evaluated in this study includes carefully evaluated CFS subjects, matched controls and non-fatigued controls from the community. Therefore, the results from this study should be applicable to similarly designed studies.

Very few studies have evaluated the presence of autoantibodies in people with CFS. Those that have tested for the same autoantibodies report discordant results. Konstantinov et al, 4 found high rates of antinuclear antibodies (ANA) in CFS patients while Skowera et al, 5 found no difference in the rate of ANA between CFS patients and controls. One explanation for these discrepancies could be a technical one where laboratories used different reagents and methods resulting in discordant results. Another explanation could be that the CFS subjects were evaluated differently. Rigor in evaluating CFS patients and applying the case definition 1 is pivotal and undoubtedly accounts for much of the variation. The results presented also show little evidence for autoantibodies to ubiquitous nuclear and cellular autoantibodies.

The findings of this study hint that evaluation of certain autoantibodies may give clues to etiology and ongoing pathology in subsets of CFS subjects. A few of the physician surveillance CFS cases from the youngest age category had autoantibodies to ssDNA. Autoantibodies to ssDNA have been associated with both viral and bacterial infection 1617. Interestingly, CFS subjects who describe a sudden onset to their illness often report flu-like illness. The fact that antibodies to ssDNA were detected only in this age group may reflect an immune response to infection commonly affecting this age group, such as infectious mononucleosis from Epstein Barr Virus infection.

There was a higher prevalence of autoantibodies to MAP2 in the non-fatigued men in the community study compared to the non-fatigued women. The significance of this finding is not known but highlights the importance of carefully stratifying and controlling for factors that could affect interpretation of results. There did seem to be a slight association of MAP2 autoantibodies with duration of CFS illness. Among CFS subjects in both study populations, those who had been sick longer had higher rates of autoantibodies than those that report shorter duration of illness. MAP2 is a neuron specific cytoskeleton protein. Autoantibodies to MAP2 have been demonstrated in patients with neuropsychiatric systemic lupus erythematosus 14. While no lesions or loss of central nervous system function is reported in people with CFS, loss of memory, concentration and cognitive impairment are common complaints. Future studies will attempt to associate assessment of these parameters with the presence of MAP2 autoantibodies.

There was no evidence of higher rates of the common autoantibodies in people with CFS. However, certain subsets of CFS subjects that had higher rates of antibodies to microtubule-associated protein 2 (MAP2) and ssDNA. Autoantibodies to specific host cell antigens may be a useful approach to stratify CFS subjects and provide clues to CFS etiologies.

The author(s) declare that they have no competing interests.

SDV was instrumental in the design of the experimental approach, analysis, presentation, discussion of these data and manuscript preparation. WCR contributed to the design of the experimental approach, and to the manuscript preparation. Both authors read and approved the final manuscript.