OBJECTIVE--To study the association between coxsackie B virus infection and the postviral fatigue syndrome and to assess the immunological abnormalities associated with the syndrome. DESIGN--Case-control study of patients with the postviral fatigue syndrome referred by local general practitioners over one year. SETTING--General practitioner referrals in Dunbartonshire, Scotland. PATIENTS--254 Patients referred with the postviral fatigue syndrome (exhaustion, myalgia, and other symptoms referable to postviral fatigue syndrome of fairly recent onset--that is, several months) and age and sex matched controls obtained from same general practitioner; 11 patients were rejected because of wrong diagnoses, resolution of symptoms, and refusal to participate, leaving 243 patients and matched controls. MAIN OUTCOME MEASURES--Detailed questionnaire (patients and controls) and clinical examination (patients) and blind analysis of blood sample at entry and after six months for determination of coxsackie B virus IgM and IgG antibodies and other variables (including lymphocyte protein synthesis, lymphocyte subsets, and immune complexes). RESULTS--Percentage positive rates for coxsackie B virus IgM at entry were 24.4% for patients and 22.6% for controls and for coxsackie B virus IgG 56.2% and 55.3% respectively; there were no significant differences between different categories of patients according to clinical likelihood of the syndrome nor any predictive value in a fourfold rise or fall in the coxsackie B virus IgG titre in patients between entry and review at six months. The rates of positive antibody test results in patients and controls showed a strong seasonal variation. Of the numerous immunological tests performed, only a few detected significant abnormalities; in particular the mean value for immune complex concentration was much higher in 35 patients and 35 controls compared with the normal range and mean value for total IgM was also raised in 227 patients and 35 controls compared with the normal range. CONCLUSIONS--Serological tests available for detecting coxsackie B virus antibodies do not help diagnose the postviral fatigue syndrome. Percentage positive rates of the antibodies in patients simply reflect the background in the population as probably do the raised concentrations of total IgM and immune complexes.