Surgical procedures associated with risk of ovarian cancer.

BACKGROUND This historical cohort study was conducted to examine the relationship between gynaecological surgery and ovarian cancer risk. METHODS Women were included if they had had tubal ligation, hysterectomy, or unilateral ovariectomy in Ontario between March 1979 and April 1993. The cohort was linked to the Ontario Cancer Registry and the Ontario mortality file. Person-years in the cohort were accumulated until death, the removal of both ovaries, a diagnosis of ovarian cancer, or the end of the study period 31 December 1993. Observed cancers were compared to expected based on Ontario age- and calendar period-specific incidence rates. RESULTS For tubal ligation and hysterectomy, fewer ovarian cancers were observed than were expected by age, calendar year of procedure, and length of follow-up; the observed/expected ratios were generally statistically significant. In contrast, no protective effect was evident for unilateral ovariectomy; in fact statistically significant excess cancers were seen in early follow-up periods. Observed/expected ratios were nearly identical and somewhat protective for the two strata defined by whether or not the ovaries were visualized. Disruption of the ovarian pathway conferred a protective effect, while no disruption significantly increased risk. CONCLUSIONS The data do not support screening bias although short-term follow-up data indicate the possibility of detection bias. The long-term follow-up data, as well as the data on pathway disruption, are consistent with the hypothesis that the surgical procedures themselves may produce a protective effect against ovarian cancer, through alteration of the hormonal environment and/or by physical destruction of a carcinogen's route to the ovary.