High blood pressure (BP) defines a prognostically heterogeneous group. Because BP varies according to time, setting and means of observation, it has been postulated that BP reactivity might better predict cardiovascular disease (CVD) than does unidimensional measurements. To assess BP reactivity, the difference between pretreatment nurse (RN) and physician (MD) diastolic BP (DBP)--systematically recorded in that order--or MD-RN DBP, was obtained in 1737 previously untreated patients with sustained, RN BP greater than or equal to 160 and/or 95 mmHg. Patients stratified by tertiles of MD-RN DBP [(I) less than or equal to - 3, (II) -2 to 3 and (III) greater than or equal to 4 mmHg] were similar by sex, race, age, body mass index, cholesterol, electrocardiography, prior CVD, smoking and pretreatment or attained in-treatment BPs. During 14 years of followup, myocardial infarction (MI) incidence per 1000/year were, tertile I (3.2), II (3.7), III (7.6) (relative risk = 2.4, III vs I + II, p less than 0.05), whereas stroke incidence and non-CVD mortality were evenly distributed. By Cox survival analysis, controlling for other entry characteristics only age, sex and DBP reactivity remained predictive (p less than or equal to 0.03) of MI or total CVD. Thus, BP reactivity, probably a centrally-mediated phenomenon, identifies a subgroup of hypertensives with an increased propensity for MI despite successful BP control.
