Obesity and hypertension are often found in the same patients, particularly in elderly women. However, few data on the joint impact of these two conditions in women and the elderly are available. In the current study of 800 elderly hypertensive patients randomly assigned to active treatment or placebo, the initial mean body mass index (BMI) was 26.7 kg/m2 in 560 women and 25.7 kg/m2 in 240 men. During the trial, total mortality and cardiovascular and noncardiovascular terminating events were highest in the patients at the leanest BMI quintile. The association between BMI and cardiovascular end points was U-shaped, whereas noncardiovascular mortality decreased with increasing BMI. The results in the women were similar to those in the total group. The U-shaped relation was confirmed with Cox's proportional hazards model, controlling for age, gender, systolic blood pressure, hemoglobin, serum cholesterol, blood glucose, and cardiovascular complications at entry. The BMI level with the lowest risk was 28 to 29 kg/m2 for total mortality and cardiovascular terminating events, 26 to 27 kg/m2 for cardiovascular mortality, and 31 to 32 kg/m2 for noncardiovascular mortality. BMI did not modify the favorable effects of drug treatment. There was no evidence that obesity would protect elderly hypertensive men or women from cardiovascular complications.