Thirty-seven studies were identified, of which 30 were included in the analysis. No figures were given for the number of women.
Unopposed oestrogen more than doubles the risk of endometrial cancer. The overall summary relative risk (RR) was 2.3 for oestrogen users compared with nonusers (95% CI: 2.1, 2.5), and the risk increases with higher doses and greater duration of use: 10 or more years use of unopposed oestrogen is associated with an RR of 9.5 (95% CI: 7.4, 12.3). The risk of death from endometrial cancer also appears to be increased, but the analysis included only 4 studies and the CI is wide (RR, 2.7; 95% CI: 0.9, 8.0). Conjugated oestrogens are associated with significantly higher risk than synthetic ones (RR, conjugated, 2.5; 95% CI: 2.1, 2.9; RR, synthetic, 1.3; 95% CI: 1.1, 1.6). The risk diminishes with time since last use, but is still 2.3 (95% CI: 1.8, 3.1) 5 years after discontinuation of unopposed oestrogen. Interrupted use of oestrogen does not lead to lower risk than daily use. Risk of noninvasive cancer was increased most, but risks of advanced cancer and death were also increased.
The overall summary RR for endometrial cancer among women who took oestrogen plus progestin was 0.8 (95% CI: 0.6, 1.2) but the direction of the effect differed between cohort and case-control studies.
Heterogeneity between studies was substantially reduced when they were stratified by dose or duration of oestrogen use, suggesting that the effect of these two variables accounts for most of the variation between studies. Smaller studies gave higher risk estimates, suggesting some publication bias. However, the overall rank correlation of study size with RR (0.08) was not significant, and removal of small studies from the analysis did not significantly affect the results.