Background:
Hysterectomy and oophorectomy are among the most performed gynecologic surgeries worldwide. Evidence of the potential public health consequences of these common interventions, including cancer risk, has been inconsistent. Current clinical guidelines rarely consider long-term risks associated with these procedures. In the absence of randomized trials, observational data can be used to provide effect estimates that are directly relevant for public health.
Objective:
To estimate the effect of hysterectomy with and without ovarian conservation on all-cause and cancer-specific mortality among adult women.
Methods:
Using observational data from the Mexican Teachers’ Cohort, a large prospective study of women initiated in 2006-08 in Mexico, we specified and emulated a target trial of hysterectomy with and without ovarian conservation and 9-year all-cause and cancer-specific mortality. Baseline was defined at the first follow-up questionnaire in 2011, and eligible participants were cancer-free women who didn’t report hysterectomy before baseline. Incident hysterectomy with and without ovarian conservation were self-reported. The primary outcome was all-cause mortality. Mortality from all cancers, breast and colorectal cancer were secondary outcomes and were classified using ICD-10 codes from death certificates. Deaths and causes of death were identified through a probabilistic linkage algorithm with national death registries. For comparison with previous observations, we evaluated cardiovascular death. We used inverse probability weights and pooled logistic regression to estimate 9-year cumulative risks, risk ratios (RR), and risk differences (RD). We obtained 95% confidence intervals (95-%CI) from nonparametric bootstraps (500 replications).
Results:
Among 92,584 eligible women (84,391 without surgery; 5,347 hysterectomy with ovarian conservation; 2,846 hysterectomy without ovarian conservation), we identified 2,808 deaths during 9-years of follow-up. Mean age was 44.6 years (SD 7.8) among women without surgery, 49.2 years (SD 5.5) among those undergoing hysterectomy with ovarian conservation, and 50.3 years (SD 5.8) among those undergoing hysterectomy without ovarian conservation. Nine-year cumulative risks of all-cause mortality were 1.54% (95%-CI: 1.46–1.62) for no surgery, 1.72% (95%-CI: 1.37–2.07) for hysterectomy with ovarian conservation, and 2.14% (95%-CI: 1.61–2.68) for hysterectomy without ovarian conservation. Compared with women without surgery, hysterectomy without ovarian conservation showed a 30% higher all-cause mortality risk (RR=1.31; 95%-CI: 1.01, 1.60; RD=2.40 per 1,000; 95%-CI: 0.12, 4.77). Results for all outcomes for hysterectomy with ovarian conservation were null. Relative to no surgery, hysterectomy without ovarian conservation showed higher cancer mortality risk (RR=1.55, 95%-CI: 0.98, 2.16; RD=0.8 per 1,000; 95%-CI: -0.04, 1.96); however, this procedure showed a lower breast cancer mortality risk (RR = 0.51, 95%-CI: 0.01, 0.94; RD = −0.6 per 1,000; 95%-CI: -1.30, -0.06). In contrast, the corresponding estimate for colorectal cancer suggests that this procedure increases colorectal cancer mortality risk (RR=1.79, 95%-CI: 0.01–3.81; RD=0.30 per 1,000; 95%-CI: -0.44, 0.80); however, confidence intervals were wide. Compared to no surgery, the RR for hysterectomy without ovarian conservation for cardiovascular mortality was 1.79 (95%-CI: 0.97, 2.57) and the corresponding RD was 1.10 per 1,000 (95%-CI: -0.04, 2.41).
Conclusions/Implications:
Hysterectomy without ovarian conservation may increase all-cause mortality risk in Mexican women. This surgical procedure may increase overall cancer mortality; however, it may lower breast cancer mortality.