Laskarzewski P., J. A. Morrison, R. Horvitz, P. Khoury, K. Kelly, M. Mellles and C. J. Glueck (U. of Cincinnati, College of Medicine, Cincinnati, OH 45267). The relationship of parental history of myocardial infarction, hypertension, diabetes and stroke to coronary heart disease risk factors In their adult progeny. Am J Epidemiol 1981; 113: 290–306. Both univariate and multivariate analysis of covariance were used to assess the relationships of parental history of myocardial infarction (Ml), hypertension, stroke, and diabetes to coronary heart disease (CHD) risk factors of their 556 adult progeny (≧24 years old) including 200 white males, 257 white females, 30 black males, and 69 black females in the Princeton School study. The CHD risk factors considered were plasma cholesterol (CHOL), tri-glycerlde (TG), high and low density lipoprotein cholesterol (C-HDL, C-LDL), Quetelet index, and systolic and diastolic blood pressure (SBP and DBP). Fewer than half of the univariate comparisons appeared to be consistent with the hypothesis that positive family history is associated with a greater tendency toward CHD as measured by known risk factors. By univariate analysis, a history of paternal, but not maternal Ml was associated with higher C-LDL and TG in their offspring. There were also several significant univariate interactions of risk factors with age in progeny of fathers with Ml; the offsprings' increase in C-LDL and total CHOL with age was greater than that observed in subjects whose fathers did not have a Ml. By multlvariate analysis, neither paternal nor maternal history of Ml played a significant role relative to risk factor levels in their progeny. By multivariate analysis, maternal history of stroke was significant when all seven risk factors in their off-spring were viewed collectively, including higher CHOL, TG, C-LDL, and, in black kindreds, higher SBP and DBP. By multivariate analysis, maternal history of hypertension was significant; offspring of hypertensive mothers had Increased TG and Quetelet levels, lower C-HDL, and higher SBP and DBP. While parental diabetes had few significant univariate and no multivariate relationships with CHD risk factors in their progeny, parental history of diabetes had considerable interaction with other parental history information (parental stroke, Ml and hypertension). Using multivariate analysis, the existence of two affirmative answers to parental history questions did not cause a difference in their offspring's risk factors above that difference caused by the existence of one affirmative answer. In the multivariate model, the significance of interaction terms between covariables and main effects implies that the offspring's characteristics, such as race group, are more important in detecting differences in risk factors than one piece of additional information of parental history. Univariate and multivariate analysis of covariance may provide for somewhat different patterns of risk factor aggregations; the differing results can probably be attributed to the correlations which exist between the CHD risk factors themselves. In a cohort of adults, average age 40 years, while parental health history of Ml, hypertension, stroke and diabetes may facilitate identification of CHD risk factors in progeny, its utility will be circumscribed for primary care physicians who must rely on univariate comparisons.
Read full abstract