Seven publications describing three randomised controlled trials were included in the review (22,800 women). One trial was rated as having low risks of bias for all domains, except for the domain of other bias (rated as high risk). The other two trials were rated as having low risks of bias for randomisation, allocation concealment and blinding, and high risks of bias for incomplete outcome data, selective reporting, and other bias.
One trial (13,175 women) found that the overall relative risks were statistically significantly higher among the tamoxifen group than the control group for endometrial cancer (RR 2.46, 95% CI 1.35 to 4.48) and pulmonary embolism (RR 3.01, 95% CI 1.20 to 1.58). Another trial (7,154 women) found a statistically significantly higher risk of deep vein thrombosis among tamoxifen participants (RR 1.68, 95% CI 1.13 to 2.51). No other statistically significant differences between the tamoxifen and control groups were found in the individual trials.
Subgroup analyses of women younger than 50 years (two trials) showed that there were no statistically significant differences between tamoxifen and control groups for the risks of endometrial cancer (RR 1.19, 95% CI 0.53 to 2.65) and pulmonary embolism (RR 1.16, 95% CI 0.55 to 2.43). A statistically significantly higher risk of deep vein thrombosis was observed among tamoxifen groups in the active phase of treatment (RR 2.30, 95% CI 1.23 to 4.31), but by the follow-up phase this difference was no longer statistically significant (RR 1.00, 95% CI 0.38 to 2.67).
Further results were reported in the review.