Pulse Pressure and Risk of Cardiovascular Disease
- James D. Neaton, PhD
-
Division of Biostatistics
School of Public Health
University of Minnesota
Minneapolis
- Michael Domanski, MD
-
Clinical Trials Group
National Heart, Lung, and Blood Institute
Bethesda, Md
- Jeremiah Stamler, MD
-
Department of Preventive Medicine
Northwestern University Medical School
Chicago, Ill
for the MRFIT Research Group
Since this article does not have an abstract, we have provided the first 150 words of the full text.
In Reply: Drs Port and Jennerich note that the coefficients we cited for the Cox model for both SBP and DBP are inconsistent with the Figure. These results represent 2 approaches to the study of blood pressure and risk of cardiovascular mortality, one based on a regression model which assumes a log-linear relationship with blood pressure, and one which classifies participants by blood pressure levels, and which makes no assumption about log-linearity. Both approaches are valid. The apparent inconsistency arises in part because the Cox coefficients cited are for all men and the Figure gives rates for 2 separate age groups, and in part because numbers of deaths and participants in the high SBP and low DBP categories (the discordant group) are much smaller than numbers in the high SBP and high DBP groups (concordant group). Thus, the positive association dominates when a log-linear model is fitted. Perhaps we failed …
