Diet, blood pressure, and multicollinearity.

Abstract

Recent reports of an inverse association between dietary calcium intake and hypertension stimulated this analysis of the relationship of blood pressure to more than 20 dietary factors among a group of 8000 Japanese men in Hawaii. Reported intakes of potassium, calcium, protein, and milk were all inversely associated with blood pressure levels when examined one at a time while controlling for other risk factors. Alcohol intake was directly associated with blood pressure, and was treated as a confounding variable in the analysis. The association of potassium intake with blood pressure was relatively stronger than the associations for other nutrients, but the intake of potassium was so highly correlated with intakes of calcium, milk, and protein that it was not statistically possible to identify the independent association of potassium and blood pressure. Calcium intake was strongly correlated with milk and potassium intakes, and only calcium from dairy sources was associated with blood pressure. These data thus indicate that several dietary factors are inversely related to blood pressure levels independently of other risk factors such as age, body mass, and alcohol intake. The high degree of intercorrelation (multicollinearity) among these dietary factors, however, indicates that the independent role of any specific nutrient cannot be conclusively separated from the possible effects of other nutrients in this type of study.