Summary

Background: The degree to which serum total cholesterol predicts cariovascular disease is uncertain. While most authors have placed TC among the most powerful risk indicators of CVD, some have claimed that it predicted CVD in women only, or even not at all.

Objective: To determine the predictive value of serum total cholesterol relative to diabetes, smoking, systolic blood pressure and body mass index (kg/m2), for cardiovascular disease mortality in 3,461 occupationally active Israeli males.

Methods: A prospective follow-up was carried out for the years 1987-1998 to determine the effect of age, smoking habits, a history of diabetes, SBP, BMI and TC, at entry, on CVD mortality.

Results: There were 84 CVD deaths during a total of 37,174 person-years follow up. The hazard ratios (95% confidence intervals) for CVD mortality with respect to variables at entry were: diabetes 5.2 (2.1-13.2), age 2.2 (1.7-2.9), smoking 1.3 (1.0-1.8), SBP 1.4 (1.1-2.0), TC 1.5 (1.0-2.1) and BMI 1.2 (0.7-2.2). Among non-obese, non-diabetic, normotensive subjects the hazard ratio of TC adjusted for age and smoking was 1.16 (1.09-1.22) per 10 mg/dl. In the remaining subjects it was 1.04 (0.98-1.12) only. There was a significant interaction between TC and diabetes, hypertension or obesity (P=0.003).

Conclusions: In this population of Israeli males we found an interaction between TC and other risk indicators for CVD. Confirmation is required for the unexpected finding that the predictive value of TC for CVD mortality among non-diabetic, non-obese and normotensive subjects exceeded that among subjects with either of these risk factors.