Background: Data on determinants of persistence of major depressive episodes (MDE) are inconsistent due to methodological shortcomings of the studies involved. Aims: To examine determinants of persistence of MDE in subjects from the general population (N=250) with new episodes of DSM-III-R major depression. Method: The Netherlands Mental Health Survey and Incidence Study is a prospective epidemiologic survey in the adult population (N=7076), using the Composite International Diagnostic Interview (CIDI). A broad range of potential determinants was assessed. Results: Determinants of persistence were severity of the index episode, longer duration of previous episodes, (chronic) physical illness and lack of social support. A recurrent episode predicted shorter duration. Limitations: Follow-up was only 2 years and duration of episode was retrospectively assessed. Conclusions: Just as in clinical populations, illness-related factors seem to be the strongest predictors of persistence of MDE. A thorough assessment of each depressed patient on the predictors of persistence is advisable.