BackgroundLittle is known about the epidemiology of bulimia nervosa outside clinical settings. We report the incidence, prevalence and outcomes of bulimia nervosa using for the first time a nationwide study design.MethodTo assess the incidence and natural course and outcomes of DSM-IV bulimia nervosa among women from the general population, women (n=2881) from the 1975–79 birth cohorts of Finnish twins were screened for lifetime eating disorders using a two-stage procedure consisting of a questionnaire screen and the Structured Clinical Interview for DSM-IV (SCID). Clinical recovery was defined as 1-year abstinence from bingeing and purging combined with a body mass index (BMI) ⩾19 kg/m².ResultsThe lifetime prevalence of DSM-IV bulimia nervosa was 2.3%; 76% of the women suffered from its purging subtype and 24% from the non-purging subtype. The incidence rate of bulimia nervosa was 300/100000 person-years at the peak age of incidence, 16–20 years, and 150/100000 at 10–24 years. The 5-year clinical recovery rate was 55.0%. Less than a third of the cases had been detected by health-care professionals; detection did not influence outcome. After clinical recovery from bulimia nervosa, the mean levels of residual psychological symptoms gradually decreased over time but many women continued to experience significantly more body image problems and psychosomatic symptoms than never-ill women.ConclusionsFew women with bulimia nervosa are recognized in health-care settings. Symptoms of bulimia are relatively long-standing, and recovery is gradual. Many clinically recovered women experience residual psychological symptoms after attaining abstinence from bingeing and purging.