INTRODUCTION: Stroke is one of the leading causes of death. Belgian stroke mortality rates are only available from the death certificates notification. OBJECTIVES: To estimate stroke incidence, mortality and case fatality, to compare our figures with those available from the death certificates notification and from neighboring countries, and identifying factors influencing survival. METHODS: A nationwide prospective registration by 178 Belgian general practitioners recording all cerebrovascular events in 1998-1999. All events, including those for which medical care was immediately found in hospital, were included. RESULTS: The yearly age-and-gender-adjusted stroke attack rates were estimated at 185 cases per 100,000 inhabitants (95% CI = 169-202). After 12 months, 47% of the patients died, resulting in an estimated yearly stroke mortality rate of 88 per 100,000 inhabitants (95% CI = 73-105). The case fatality rate among men (39%; 95% CI = 29-51) was lower than among women (50%; 95% CI = 39-63) (p < 0.05). In a forward stepwise logistic regression diabetes, stroke history, hemorrhagic stroke, coma, swallow deficit and urinary incontinence had a negative influence on stroke survival. Treated hypertension had a beneficial effect on survival. CONCLUSIONS: Stroke mortality rates estimated by the sentinel practices are fairly comparable with those based on the death certificates notification. The stroke attack rates correspond with those of neighboring countries. The role of treated hypertension on the outcome of strokes remains controversial in this registration.