OBJECTIVE--To investigate the relation between mortality and treated systolic and diastolic blood pressures. DESIGN--Randomised double blind placebo controlled trial. Mortality in the two treatment groups was examined in thirds of treated systolic and diastolic blood pressures. PATIENTS--339 And 352 patients allocated to placebo and active treatment, respectively. The groups were similar at randomisation in sex ratio (70% women), mean age (71.5 years), blood pressure (182/101 mm Hg), and proportion of patients with cardiovascular complications (35%). MEASUREMENTS AND MAIN RESULTS--In the placebo group total mortality rose with increasing systolic pressure whereas it had a U shaped relation with diastolic pressure, the total lowest mortality being in patients in the middle third of the distribution of diastolic pressure. In the group given active treatment total mortality showed a U shaped relation with systolic pressure and an inverse association with treated diastolic pressure. In both groups cardiovascular and non-cardiovascular mortality followed the same trends as total mortality. The increased mortality in the lowest thirds of pressure was not associated with an increased proportion of patients with cardiovascular complications at randomisation or with a fall in diastolic pressure exceeding the median fall in pressure in each group. In contrast, patients in the lowest thirds of pressure showed greater decreases in body weight and haemoglobin concentration than those in the middle and upper thirds of pressure. CONCLUSIONS--In patients taking active treatment total mortality was increased in the lowest thirds of treated systolic and diastolic blood pressures. This increased mortality is not necessarily explained by an exaggerated reduction in pressure induced by drugs as for diastolic pressure a U shaped relation also existed during treatment with placebo. In addition, patients in the lowest thirds of systolic and diastolic pressures were characterised by decreases in body weight and haemoglobin concentration, and the patients in the lowest thirds of diastolic pressure taking active treatment also by an increased non-cardiovascular mortality, suggesting some deterioration of general health.