Serum cholesterol elasticities in relation to macro- and micronutrient balances

Abstract

The application of the isometric log-ratio (ilr) transformation to dietary data leads to the definition of nutrient balances that represent new variables which can be included in regression models as covariates. However, their effects on the response variable are not easy to quantify. We then propose borrowing the concept of elasticity from econometrics as an appealing means of improving the interpretability of the ilr-related coefficients. Using data from an Italian population-based study, a multiple linear regression model of the serum total-/HDL-cholesterol ratio on nutrient balances was fitted and the estimated coefficients were used to derive elasticities. The elasticities measure the relative changes in the cholesterol ratio in response to the relative changes in specific nutrient ratios, while keeping unchanged the proportional relationships between the other dietary elements. In a comprehensive overview of the diet, this alternative approach to dietary data analysis allows isocaloric analysis and may provide interesting new insights.