We investigated the association between retirement and primary cardiac arrest (PCA) in 126 male cases and controls, 25-75 years of age, without prior heart disease or comorbidity. After adjustment for age alone, retirement was not associated with an increased risk of PCA, (OR = 1.1; 95% confidence intervals = 0.5, 2.4). This lack of association was not uniform across age strata, however. In 10 of 19 discordant pairs 60 or more years of age, the control subject had been retired; in all seven discordant pairs under 60, the case had been retired (lower 95% CI of the relative risk = 1.9).