Emergency contraception uses medications or intrauterine devices to prevent pregnancy after unprotected intercourse. Clinicians need evidence on comparative effectiveness, safety, and acceptability of available options. This Cochrane review assessed whether mifepristone prevents pregnancy more effectively or with different harms than other emergency contraception methods.
The review searched CENTRAL, MEDLINE, Embase, two additional databases, and two trial registers, with the last search date 28 February 2025. Eligible studies were randomized controlled trials (RCTs) comparing mifepristone to oral emergency contraception pills (including levonorgestrel, the Yuzpe regimen, and ulipristal acetate) or to copper or levonorgestrel intrauterine devices (IUDs). Non-randomized studies of interventions (NRSIs) were sought to capture rare harms or minimal clinically important differences, but none provided additional outcome data beyond the included RCTs. Participants were women seeking emergency contraception within five days of unprotected intercourse. The review used Cochrane's RoB 1 tool for risk of bias and GRADE to assess certainty of evidence. Critical outcome: number of pregnancies. Important outcomes: any side effects, changes in menstrual timing (early or delayed menses), and acceptability measured as treatment satisfaction.
A total of 87 randomized trials with approximately 36,000 participants were included. Geographically, 79 studies were conducted in China, four in the UK, two in Cuba, and two were multi-country. Comparisons included 41 trials of mifepristone versus levonorgestrel, three versus the Yuzpe regimen, 42 comparing different mifepristone doses, and two trials versus copper IUD. Non-randomized studies identified were excluded because they did not add unique outcome data. Many trials had incomplete reporting for key risk-of-bias domains.
The review reports results by comparison and by dose where applicable. Certainty of evidence across outcomes ranged from very low to high, with downgrades primarily for unclear reporting and risk of bias. Sensitivity analyses limited to trials at low overall risk of bias produced similar effect estimates to the primary analyses.
Comparisons were stratified by mifepristone dose.
Mid-dose mifepristone (25–50 mg) versus levonorgestrel: Mid-dose mifepristone likely results in fewer pregnancies (risk ratio [RR] 0.67, 95% CI 0.49 to 0.91; I2 = 0%; 27 studies, 6052 participants) and likely reduces the occurrence of any side effects (RR 0.55, 95% CI 0.39 to 0.76; I2 = 72%; 17 studies, 4350 participants). Mid-dose mifepristone probably leads to fewer instances of early menses (RR 0.71, 95% CI 0.49 to 1.03; 7 studies, 1324 participants) but increases the risk of delayed menses (RR 1.29, 95% CI 1.01 to 1.64; I2 = 25%; 17 studies, 3615 participants). Certainty for these outcomes was typically moderate.
Low-dose mifepristone (<25 mg) versus levonorgestrel: Low-dose mifepristone reduced pregnancy compared to levonorgestrel (RR 0.73, 95% CI 0.59 to 0.90; I2 = 0%; 14 studies, 8752 participants) with high-certainty evidence. It likely causes a large reduction in any side effects (RR 0.26, 95% CI 0.17 to 0.38; I2 = 0%; 3 studies, 609 participants; moderate certainty). Low-dose mifepristone reduced the incidence of early menses (RR 0.45, 95% CI 0.35 to 0.58; I2 = 0%; 6 studies, 1898 participants) but slightly increased delayed menses (RR 1.52, 95% CI 1.11 to 2.08; I2 = 50%; 10 studies, 7618 participants), both with high-certainty evidence. One study reported slightly higher treatment satisfaction with low-dose mifepristone (RR 1.03, 95% CI 0.99 to 1.07; 1 study, 724 participants; moderate certainty). Many trials did not report acceptability outcomes.
Across three studies (2144 participants), mifepristone (any dose) probably reduces pregnancies compared with the Yuzpe regimen (RR 0.14, 95% CI 0.05 to 0.41; I2 = 0%; moderate-certainty evidence). Mifepristone also probably results in a large reduction in any side effects compared with Yuzpe (RR 0.89, 95% CI 0.83 to 0.95; I2 = 97%; 2 studies, 1693 participants; moderate-certainty evidence), notably fewer nausea and vomiting events. Mifepristone increases the risk of delayed menses substantially (RR 2.83, 95% CI 2.31 to 3.47; I2 = 85%; 3 studies, 1912 participants; high-certainty evidence). One study reported that mifepristone likely increases treatment satisfaction compared with Yuzpe (RR 1.17, 95% CI 1.10 to 1.25; 1 study, 615 participants; moderate certainty). No trials reported early menses for this comparison.
Only two trials compared mifepristone with copper IUD. Neither trial reported data on side effects or treatment satisfaction. Evidence for pregnancies and altered menses was of very low certainty due to risk of bias, imprecision, and few events; therefore, reliable conclusions cannot be drawn from these data.
Comparing mid-dose (25–50 mg) with low-dose (<25 mg) mifepristone across multiple trials: mid-dose results in a slight reduction in pregnancies (RR 0.77, 95% CI 0.59 to 1.00; I2 = 0%; 26 studies, 12,358 participants) and a slight increase in delayed menses (RR 1.32, 95% CI 1.15 to 1.50; I2 = 53%; 22 studies, 11,733 participants), both with high-certainty evidence. There may be little to no difference in the occurrence of any side effect or early menses between doses; those outcomes were rated with low certainty. Treatment satisfaction was not reported in these dose-comparison trials.
The certainty of evidence across outcomes ranged from very low to high, with common reasons for downgrading including unclear reporting of random sequence generation, allocation concealment, blinding, and incomplete outcome data. A large proportion of trials were conducted in China, which may limit generalizability for side effects and satisfaction. Substantial missing data affected several risk-of-bias domains. Sensitivity analyses restricted to low risk-of-bias trials did not materially change effect estimates.
Both mid-dose and low-dose mifepristone are probably more effective than levonorgestrel for emergency contraception. Mifepristone is likely more effective than the Yuzpe regimen. Nausea and vomiting are less common with mifepristone than with Yuzpe. A delay in menses is probably the main adverse effect of mifepristone; higher doses appear associated with a greater frequency of delayed menses than lower doses. Evidence comparing mifepristone with copper IUD is insufficient and of very low certainty.
This Cochrane review received partial funding from the World Health Organization (WHO). The review is registered with PROSPERO (CRD42024591428). Copyright © 2026 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.