Methodologic evaluation of incidence rates for hypertension: calculated for Pittsburgh's MRFIT usual care men.

Methodologic limitations in the calculation of incidence rates for hypertension and in the assessment of associations between rates of change in attributes and the development of hypertension were examined for Pittsburgh's MRFIT Usual Care population. In a group of normotensive middle-aged men at high risk for the development of coronary heart disease, the incidence of hypertension ranged from 6.2 to 15.5 per 100 man years, depending on the definition of hypertension. A simulation study showed that the magnitude of these incidence rates could be explained by random variation in the measurement of blood pressure. Controlling for these random effects, the development of hypertension was associated with high normal baseline diastolic blood pressure and with a reported past diagnosis of hypertension, but not with age or baseline body mass index. Finally, a method, estimating the relative risk associating weight gain and the development of hypertension, was presented.