Measles surveillance data in Blantyre, Malawi were reviewed for 1996-8 to describe the epidemiology of infection and to estimate vaccine efficacy (VE) by the screening method. A total of 674 measles cases were reported to the Blantyre District Health Office during this period. Age distribution showed that 108 (16.1%) of the cases were aged less than 1 year. The median age was 5 years. Eighty percent of the cases between 1 and 19 years had been previously vaccinated. VE was 68.6% (95% CI, 527-792) for children 12-23 months of age and 67.3% (95% CI, 48.3-79.3) for infants 9-11 months of age. Reasons for this low vaccine efficacy are discussed. Previous vaccination history was negatively associated with the risk for developing cough during measles infection (odds ratio (OR), 0.30; 95% CI, 0.09-0.91), diarrhoea (OR, 0.64; CI, 0.44-0.95) and pneumonia (OR, 0.40; CI, 0.25-0.62). Logistic regression analysis showed that pneumonia in adults was negatively associated with vaccination history. The passive surveillance system for measles in Malawi was useful to describe the epidemiology of measles.