A retrospective cohort study of vasectomy and prostate cancer in US men
E. Giovannucci, T. D. Tosteson, F. E. Speizer, A. Ascherio, M. P. Vessey and G. A. Colditz
Channing Laboratory, Department of Medicine, Harvard Medical School, Boston, MA 02115.
OBJECTIVE--To examine the relationship between vasectomy and prostate
cancer. DESIGN--Retrospective cohort study. SETTING--Husbands of nurses in
11 large states in the United States. PARTICIPANTS--In questionnaires
administered in 1976 and 1978, 14,607 married female participants of the
Nurses' Health Study reported vasectomy as the couple's form of
contraception. In 1989, we contacted these 14,607 women and 14,607
age-matched participants whose husbands had not had a vasectomy prior to
1978 to ascertain their husband's disease status between 1976 and 1989.
MAIN OUTCOME MEASURE--Diagnosis of prostate cancer. RESULTS--During the
study period, from 1976 through 1989, 96 new cases of prostate cancer were
diagnosed in the cohort. Vasectomy was associated with an increased risk of
prostate cancer (age-adjusted relative risk, 1.56; 95% confidence interval,
1.03 to 2.37; P = .04). The relative risk of prostate cancer increased over
time since vasectomy. Among men who had their vasectomy 20 or more years in
the past, the relative risk of prostate cancer was 1.89 (95% confidence
interval, 1.14 to 3.14; P = .005), and after excluding stage A and B cases,
the relative risk was 2.06 (95% confidence interval, 0.95 to 4.43; P =
.07). This elevated risk of prostate cancer among vasectomized men
persisted when we adjusted for smoking, alcohol consumption, educational
level, body mass index, and geographical area of residence.
CONCLUSIONS--These results support evidence from other epidemiologic
studies that vasectomy increases risk of prostate cancer. A biological
mechanism whereby vasectomy influences the rate of prostate cancer may be
related to a diminished secretory rate of prostatic fluid following
vasectomy, or, alternatively, to the postvasectomy immune response to sperm
antigens, which may cross-react with tumor-associated antigens and suppress
tumor immunosurveillance mechanisms.
