Final regression model for coronary artery calcification.

Backward elimination regression was used to develop multivariate models for coronary artery calcification (CAC) level. CAC was analyzed using the ln-agatston score in individuals with a score >0. Independent variables were divided by their standard deviations (shown in parentheses). The candidate starting variables were: age, gender, race/ethnicity, IL-6, BMI, systolic BP, use of BP lowering medication, smoking status, total-cholesterol, HDL-cholesterol, use of lipid lowering medication, type 2 diabetes status, CMV and H. pylori titers, and CD4+ memory cell proportions or, in separate analyses, CD4+ naive cell proportions. Only significant variables (p<0.05) were retained in the final model to obtain the model's R2. ns: non-significant.