In order to identify risk factors for typhoid fever in a highly endemic area, we undertook a case-control study in the Mekong delta, Viet Nam. Cases were 144 consecutive patients admitted to hospital with blood culture-confirmed typhoid fever. Two controls (1 in the hospital and 1 in the community) were chosen for each case. Standardized interviews were conducted with questions regarding recent contact with a typhoid fever patient, eating habits, hygiene and socio-economic level. Cases were more likely to have been in contact with a patient with typhoid fever than hospital controls (adjusted odds ratio (OR) = 5.2, 95% confidence interval (95% CI) 1.7-15.9) or community controls (adjusted OR = 11.9, 95% CI 2.3-60.7); 11% and 14% of typhoid fever cases (compared to hospital or community controls, respectively) were attributable to recent contact with a patient with this disease. These findings suggest that strategies directed towards the persons in contact with a patient might reduce the incidence of secondary cases of typhoid fever.