The idea that chronic fatigue has an infectious origin has become popular, but the main evidence for such an association has come from retrospective case-control studies, which are subject to ascertainment bias. We report a prospective study of the outcome of clinically diagnosed infections in patients presenting to UK general practitioners. Questionnaires assessing fatigue and psychiatric morbidity were sent to all patients aged 18-45 years in the study practices. The prevalence of chronic fatigue and chronic fatigue syndrome was then ascertained among 1199 people aged 18-45 who presented to the general practitioners with symptomatic infections and in 1167 people who attended the surgeries for other reasons. 84% were followed up at 6 months. 9.9% of cases and 11.7% of controls reported chronic fatigue (odds ratio 1.0 [95% CI 0.6-1.1]). There were no differences in the proportions who met various criteria for chronic fatigue syndrome. No effect of infection was noted when we excluded subjects who reported fatigue or psychological morbidity at the baseline screening. The strongest independent predictors of postinfectious fatigue were fatigue assessed before presentation with clinical infection (3.0 [1.9-4.7]) and psychological distress before presentation (1.8 [1.2-2.9]) and at presentation with the acute infection (1.8 [1.1-2.8]). There was no effect of sex or social class. Our study shows no evidence that common infective episodes in primary care are related to the onset of chronic fatigue or chronic fatigue syndrome.