In emergency contraception a drug or IUD is used to prevent pregnancy shortly after unprotected intercourse. Except for some Western-European countries and China, emergency contraception is largely under-utilised worldwide. In many developing countries lack of access to emergency contraception may subject women to unsafe abortions, which contribute significantly to maternal mortality and morbidity. Currently, several interventions (IUD, the Yuzpe regimen, levonorgestrel, mifepristone, danazol and some combination regimens) are available for emergency contraception. Information on the comparative efficacy, safety and convenience of these methods is crucial for reproductive health care providers and the women they serve.To determine which emergency contraceptive method following unprotected intercourse is the most effective, safe and convenient to prevent pregnancy.The search included the Cochrane Controlled Trials Register, Popline, MEDLINE, Chinese biomedical databases and UNDP/UNFPA/WHO/World Bank Special Programme on Human Reproduction (HRP) emergency contraception database (July 2003). Content experts and pharmaceutical companies were contacted.Randomised controlled trials and controlled clinical trials including women attending services for emergency contraception following a single act of unprotected intercourse were eligible.Data on outcomes and trial characteristics were extracted in duplicate and independently by two reviewers. Quality assessment was also done by two reviewers independently. Meta-analysis results are expressed as relative risk (RR) using a fixed-effects model with 95% confidence interval (CI). In the presence of significant heterogeneity a random-effect model was applied.Forty-eight trials with 33110 women were included. Most trials were conducted in China (37/48). Levonorgestrel is more effective than the Yuzpe regimen in preventing pregnancy (2 trials, RR: 0.51; 95% CI: 0.31 to 0.83). Single dose (1.5 mg) administration seems to have similar effectiveness as the standard 12 hours apart split-dose (0.75 mg twice) of levonorgestrel (2 trials, RR: 0.77, 95% CI: 0.45 to 1.30). Levonorgestrel has similar effectiveness to mid-dose (8 trials, RR: 1.64; 95% CI: 0.82 to 3.25) or low-dose (7 trials, RR: 1.38; 95% CI: 0.93 to 2.05) mifepristone. Low-dose (=< 10 mg) mifepristone is similarly effective as mid doses (25-50 mg) when only high quality trials are considered. Delay in the onset of subsequent menses is the main unwanted effect of mifepristone and seems to be dose-related. The Yuzpe regimen can be used when levonorgestrel and mifepristone are not available. Half-dose Yuzpe with single administration is associated with fewer side-effects but it is not clear whether it is as effective as the standard Yuzpe regimen (RR: 1.41; 95% CI: 0.76 to 2.61).Levonorgestrel 1.5 mg (two split doses or a single dose) and low and mid-doses (25-50 mg) of mifepristone offer high efficacy with an acceptable side-effect profile. Single dose simplifies the use of levonorgestrel for emergency contraception without an increase in side-effects. However, mifepristone might delay the following menstruation, which could increase anxiety, particularly in higher doses. The Yuzpe regimen could be used if levonorgestrel or mifepristone are not available. The intrauterine device (IUD) is another effective emergency contraceptive, and can be kept for ongoing contraception.