For decades, decision to treat hypertensive people through blood pressure lowering drugs has been based on blood pressure level. Taking into account the level of other risk factors has been advocated as a mean to rationalize and individualize these prescriptions, but physicians seem to be reluctant to apply this concept in their daily practice. Using an individual patient data base of randomized controlled trials in hypertension, we explored the impact of two strategies, one based on blood pressure level, the other on predicted risk, in terms of the proportion of clinical events that have been avoided. Two outcomes were explored: stroke, and major cardiovascular events (including stroke, myocardial infarction and cardiovascular death).

We divided the population of five trials (28 997 participants) in quartiles, either of blood pressure, or of predicted risk, and compared the number of events avoided in treating participants above each quartile threshold to the number of events avoidable in treating the whole population. The underlying assumption is that no event would have been avoided if participants were not treated. For stroke, the reduction by 50% of the number of individuals taking antihypertensive drugs would avoid 50% of the avoidable events if this strategy was purely random, 64% if this strategy was based on a threshold of blood pressure only, 75% if it was based on a threshold of predicted risk. The corresponding pattern for major cardiovascular events is 50%, 61%, 77%.

This is the first time that the interest of targeting blood pressure lowering drug prescription towards high risk people is illustrated with appropriate data. This example could be reproducible on other similar data bases.