Hypertensive patients over the age of 60 years were admitted to a double-blind placebo-controlled trial. Patients in the actively treated group received a combined potassium-losing and -sparing diuretic (triamterene 50 mg plus hydrochlorothiazide 25 mg; n = 416); this dose could be doubled and methyldopa (up to 2 g, daily) was added in 35% of patients when blood pressure remained high. The placebo group (n = 424) received matching capsules and tablets. Adverse effects were assessed in the double-blind period of the trial by calculating the incidence of abnormal biochemical results, investigator reports of diseases and prescriptions of concomitant therapy and a self-administered symptom questionnaire completed by patients. In 1000 hypertensive subjects over 60 years of age, 1 year of active treatment would prevent 11 fatal cardiac events, 6 fatal and 11 non-fatal strokes and 8 cases of severe congestive heart failure. No unexpected adverse treatment effects were observed. A significant excess incidence rate (per 1000 person years) was found in the active group compared with placebo for: (1) impaired renal function, a serum creatinine greater than 180 mumol/l (2.0 mg/dl); (2) mild hypokalaemia, a serum potassium less than 3.5 mmol/l; (3) reports of gout; and (4) an elevated serum uric acid greater than 0.52 mmol/l in men or greater than 0.46 in women. Elevated blood sugar and prescriptions for hypoglycaemic drugs tended to be more frequent in the actively treated group, but this difference was not statistically significant. In both groups, there was a low incidence (less than 7 per 1000 person years) of anaemia and depression and diseases of the liver, gall bladder or pancreas.(ABSTRACT TRUNCATED AT 250 WORDS)