Using questionnaire information provided by 4,385 vasectomized and 13,155 matched, non-vasectomized men, we found no significant differences between them for a large number of symptoms and diseases, including those of the cardiovascular system. After taking into account differences in their marital status, past smoking habits, and physical activity at work, significant statistical associations of vasectomy with joint pain or swelling, back trouble, and a history of kidney or bladder infection persisted. Our data are inconsistent with the occurrence of large increases in the risks of many important diseases in vasectomized humans. (Am J Public Health 1982; 72:476-480.) Reports that vasectomized monkeys have more severe atherosclerosis than non-vasectomized monkeys has caused concern about the possibility that vasectomy may have unanticipated health effects in humans' 2. We investigated this possibility using computer-stored information from health questionnaires completed during the period from 1977 through 1980 by members of a large Northern California prepaid medical care program who had multiphasic health check-ups.
Materials and MethodsMedical Care Program taking multiphasic health check-ups in Oakland or in San Francisco has been computer stored. The information provided on these questionnaires about vasectomies, personal habits, symptoms of illness, and the histories of certain diseases constitute the basic data for this study.In the period from 1977 through 1980, 4,392 male examinees indicated on the health questionnaires that they had had vasectomies. We matched each vasectomized man with three men who did not indicate on the questionnaires that they had had vasectomies. The matching variables were age, race, and date of the health check-up. For 98 per cent of the vasectomized men, we found three "exact" matches (birth year was plus or minus one year, date of check-up was plus or minus one month, and race was same). To find matches for the remaining two per cent of vasectomized men, we relaxed the matching criteria by stages. For seven vasectomized men, three matches could not be found. These men were excluded, leaving 4,385 vasectomized and 13,155 matched, non-vasectomized men for the analyses.We used x2 tests for independent samples to assess statistical significance. As a measure of strength of association, we present here the ratio of the prevalence rates of the symptoms and illnesses in men with and without vasectomies. We used the test-based method to calculated 95 per cent confidence intervals for the rate ratios. We used stratification to handle confounding.
