J-Shaped Relation Between Blood Pressure and Stroke
To the Editor:
In their interesting study on the J-shaped relation between blood pressure and stroke, Vokó and colleagues1 found an excess risk of stroke associated with diastolic blood pressure (BP) levels <65 mm Hg in treated hypertensive subjects but not in a group of untreated subjects. On this basis, they hypothesized a harmful effect of excessive BP reduction. The study was not a direct comparison of the prognostic value of pretreatment and in-treatment diastolic BP in the same subjects. By contrast, the authors examined stroke rate in a group of treated hypertensive subjects and in an untreated control group principally composed of normotensive subjects. Therefore, for progressively lower diastolic BPs, pulse pressure (PP) may have been higher in the treated hypertensive group than in the untreated control group. Given the potent prognostic value of PP, such an imbalance could have conditioned the greater stroke risk among the subjects with a diastolic BP <65 mm Hg than among those with a diastolic BP 65 to 74 mm Hg in the treated hypertensive group but not in the untreated group. To clarify this point, the authors should (a) not limit data presentation to the relative risks, but include a table with diastolic BP and PP in each category; (b) compare PP between treated and untreated subjects for each stratum of diastolic BP; and (c) examine PP among the potential determinants of stroke risk. If the J-shaped relation between diastolic BP and stroke risk was due to excessive BP lowering, such a relation should not be noted in an untreated control group with the same levels of diastolic BP and PP. Up to completion of this analysis, we remain skeptical about the reappearance of the J-shaped curve for effect of antihypertensive treatment and suspect that the results of this study are due to the potential confounding effect of PP and improper selection of the control group.