Relationships between serum lipids and subsequent mortality in an insured population.

OBJECTIVES To investigate relationships between serum lipids and all-cause mortality in an insured population issued policies at standard rates using contemporary analytic techniques. BACKGROUND As with other preferred risk factors, differences in lipid profiles can be an important contributor to morbidity and mortality endpoints in the general and insured population. To better understand the independent contribution of different lipid profiles to mortality, contemporary researchers commonly employ multivariate statistical techniques. Studies on insured populations utilizing multivariate statistical procedures are not commonly published and are employed in this study of an insured cohort. METHODS We studied internal data on direct and reinsurance business issued at standard rates, or rated only for abnormal lipids, between 1975 and 2000 followed through anniversaries in 2001, or prior termination, with an average duration of 6 years per policy. Cox proportional hazards models were used to study the multivariate relationship between various lipid values (cholesterol, cholesterol/HDL ratio, triglycerides) and mortality. RESULTS Five hundred twenty deaths were observed during follow-up. Mortality was noted to vary by differing levels of all 3 of the predictor variables of interest. The relationship between an abnormal lipid profile and mortality was strongest in males and nonsmokers. CONCLUSIONS In this insured cohort, changes in lipid profiles are associated with significant differentials in mortality.