A case-control study of leukemia at a naval nuclear shipyard.

A matched case-control study was conducted of 53 leukemia deaths and of 212 controls within a previously studied cohort of 24,545 on-shore workers employed between January 1, 1952 and August 15, 1977 at the Portsmouth (New Hampshire) Naval Shipyard. The study sought to ascertain a priori whether there was an association between leukemia deaths and occupational exposure to either ionizing radiation or organic solvents. To obtain information on individual exposures, radiation dose histories and detailed work histories by job and shop were evaluated for each subject. No statistically significant associations were found either between ionizing radiation or presumed solvent exposure and myelogenous or lymphatic leukemia. However, when specific job categories and shops were examined without benefit of a priori hypotheses, two occupations, electrician and welder, were found to be associated with leukemia. For electricians, the Mantel-Haenszel odds ratio (ORMH) was significantly elevated for all leukemias (ORMH = 3.00, 95% confidence interval (CI) = 1.29-6.98), particularly for lymphatic leukemia (ORMH = 6.00, 95% CI = 1.47-24.45). For welders, the odds ratio was not significantly elevated for all leukemias (ORMH = 2.25, 95% CI = 0.92-5.53), but was significantly elevated for myeloid leukemia (ORMH = 3.83, 95% CI = 1.28-11.46). These findings persisted when potential confounders were adjusted by means of a conditional logistic regression model.