Cholesterol and cancer in a population of male civil service workers.

Cancer incidence and mortality were ascertained in a cohort of 1910 male participants of the Albany Cardiovascular Health Center (CVHC). The New York State Cancer Registry, vital records files, CVHC follow-up records, New York State Retirement System files, and New York State Department of Motor Vehicles driver's license files were used. Serum cholesterol measurements as well as values for other exposure variables were obtained from records of medical examinations which began in 1953-1954. The study cohort was divided into two groups, based on initial serum cholesterol measurement (less than or equal to 190 mg/100 ml and less than or equal to 190 mg/100 ml). For total cancers, both incidence and mortality were similar in these groups. For digestive cancer, both incidence and mortality were slightly lower in the less than or equal to 190 mg/100 ml group. The deficit was not statistically significant. For respiratory cancer, relative risk and rate ratio estimates were in the range of 1.4-1.7 for incidence and mortality. The excess risk in the less than or equal to 190 mg/100 ml group was of borderline statistical significance. The association was concentrated in the lowest cholesterol quintile rather than suggesting a strong dose-response relationship. The estimates were not found to be confounded by cigarette smoking, body mass index, education or age. A reduction in the crude rate ratio estimate from 1.5 to 1.2 was observed when early cases were excluded, suggesting that part of the observed excess may be due to preclinical cancer.