A Cochrane review published September 21, 2026 pooled 87 randomized trials covering about 36,000 women and found that low-dose mifepristone prevented pregnancy better than levonorgestrel, the drug sold over the counter in the United States as Plan B. Women who took under 25 mg of mifepristone had about 27% fewer pregnancies (risk ratio 0.73), and the reviewers rated that the highest grade of certainty they give.
Mifepristone is not approved in the US for emergency contraception at any dose, so the option this review ranks highest is one you cannot buy here. What it does tell you is that the pill most people reach for on the drugstore shelf is not the strongest one a clinician has. Women’s health research runs thin in a lot of places, and creatine hit the same gap around the menstrual cycle.
Key takeaways
- Low-dose mifepristone led to about 27% fewer pregnancies than levonorgestrel, rated high-certainty evidence.
- Mifepristone caused less nausea and vomiting, but was more likely to delay the next period.
- The doses tested were small, under 25 mg in the low-dose arm and 25 mg to 50 mg in the mid-dose arm, and mifepristone is not approved in the US for emergency contraception.
What the study found
The review was led by researchers at Oregon Health & Science University and published in the Cochrane Database of Systematic Reviews. Participants were women seeking emergency contraception within five days of unprotected sex.
Against levonorgestrel, low-dose mifepristone (under 25 mg) cut pregnancies by about 27% (risk ratio 0.73, 95% CI 0.59 to 0.90) across 14 studies and 8,752 participants. The true benefit is very likely between 10% and 41% fewer pregnancies. Cochrane graded this high-certainty, meaning further research is very unlikely to change the result.
Mid-dose mifepristone (25 mg to 50 mg) showed about 33% fewer pregnancies (risk ratio 0.67, 95% CI 0.49 to 0.91) across 27 studies and 6,052 participants, at moderate certainty. Head to head, the mid dose beat the low dose only slightly, right at the edge of no difference (risk ratio 0.77, 95% CI 0.59 to 1.00). Against the older Yuzpe regimen, which combines estrogen and progestin, mifepristone showed about 86% fewer pregnancies (risk ratio 0.14, 95% CI 0.05 to 0.41), based on three studies.
The tradeoff was consistent. Low-dose mifepristone raised the chance of a delayed period by about 52% (risk ratio 1.52, 95% CI 1.11 to 2.08). A late period after emergency contraception looks exactly like the thing you were trying to prevent, which the authors flag as a real source of stress.
Dr. Kumar’s take
Two things deserve more attention than the press release gave them.
First, the dose. The mifepristone studied here for emergency contraception is small: under 25 mg in the low-dose arm, and 25 mg to 50 mg in the mid-dose arm. Those are the only doses this review tested. The lead author’s own disclosure on the release says it plainly: she prescribes emergency contraception but does not prescribe mifepristone for it, because it is not approved for that use in the USA.
Second, the geography. Seventy-nine of the 87 studies were run in China, with four in the UK, two in Cuba, and two multi-country. The authors say this may limit how well the side-effect and satisfaction findings apply elsewhere. Pregnancy prevention is a hard endpoint and probably holds; how a drug feels to take is shaped by the care setting around it.
There is also a gap worth naming. The review set out to include trials comparing mifepristone with ulipristal acetate, the prescription pill sold in the US as Ella. The breakdown of the 87 included studies lists no such comparison, and the copper IUD comparison rested on two trials rated very low certainty. So this review tells you mifepristone beats levonorgestrel. It does not tell you how it stacks up against the two options an American clinician can actually offer.
What it means for you
The usable finding is not “get mifepristone.” It is that levonorgestrel, the default drugstore choice, is beatable, and a clinician has other tools.
If you or someone you are advising might need emergency contraception, ask a doctor or pharmacist now, before the clock is running, how the over-the-counter pill compares with prescription ulipristal or a copper IUD for that person. Body weight, days elapsed, and other medications all change the answer, and this review settles none of them.
Nothing here should push anyone away from levonorgestrel if that is what is available. It works. It is simply not the best-performing pill in the evidence, and most people buying it have never been told there is a comparison to have. The gap between what evidence shows and what reaches the person deciding turns up all over medicine, including in whether a pill has to be real to help you.