Lack of true J-shaped curve in systolic pressure

Age- and sex-adjusted mortality rates (per 1000 patient-years, bars show the 95 percent confidence intervals) according to achieved systolic pressure in active treatment and control groups in a meta-analysis of seven randomized clinical trials of hypertensive patients. The number of events is shown below each bar. Among treated patients, cardiovascular mortality initially falls at achieved systolic pressures below 180 mmHg and then rises again at low systolic pressures (upper right panel). However, a similar relationship is seen with noncardiovascular mortality (bottom right panel) and in the control groups (left panels). Thus, the increase in mortality at low diastolic pressures probably reflects underlying poor health rather than an adverse effect of antihypertensive therapy.