Adjustment of the apparent benefits of treatment on stroke risk in the MRC mild hypertension trial using data from the placebo-treated group.

Clinical trials show that drug treatment significantly decreases stroke risk in hypertension. The benefit as measured in clinical trials may be affected by changes of blood pressure at entry and by departure from randomised treatment, but the magnitude of such effects is disputed. We have assessed benefit from reduction of stroke using data from the MRC Trial of mild to moderate hypertension, taking these factors into account, and have studied the likely effect of recent guidelines. The original analysis suggested that 850 patients needed treatment for 1 year to prevent one stroke. Under the more conservative of two assumptions made about the effect of treatment, this falls to 695 patients when allowance is made for reduction of stroke in placebo group patients withdrawn and actively treated, to 680 patients when allowance is made for the fall in blood pressure after entry, and to 556 patients with allowance for both. When benefit is assessed in patients whose entry diastolic blood pressure was > or = 100 mm Hg, 557 patients require treatment annually per stroke saved and this is decreased to 360 patients when allowance is made for withdrawal and active treatment of placebo group patients. These results suggest that benefit from reduction of stroke was underestimated in the MRC trial and that this is likely to be present in most trials. Changes to diagnostic criteria for hypertension in new management guidelines are likely to have significant effects on the number of patients treated per stroke prevented.