A Cochrane review included 51 studies. Misoprostol (400-600 µg given vaginally or sublingually), gemeprost, mifepristone (200 or 600 mg), prostaglandin E and F2α demonstrated larger cervical preparation effects than placebo. Misoprostol 400 µg was nonsignificantly more effective than gemeprost 1mg in preparing the cervix (mean difference 0.53, 95% CI 0.03 to 1.04; 2 trials, n=154) and was associated with fewer gastrointestinal side-effects (RR 0.35, 95% CI 0.18 to 0.68; 1 trial, n=564). Mifepristone, 200 mg given 24 hours prior to the procedure, had a greater cervical ripening effect than misoprostol, 600 µg orally or 800 µg vaginally (mean difference -0.79, 95% CI -1.29 to -0.30; 2 trials, n=90). For vaginal administration of misoprostol, administration 2 hours prior was less effective than administration 3 hours prior to the abortion. Compared to oral misoprostol administration, the vaginal route was associated with significantly greater initial cervical dilation and lower rates of side-effects; however, sublingual administration 2-3 hours prior to the procedure demonstrated cervical effects superior to vaginal administration. When Compared to day-prior laminaria tents, 200 or 400 µg vaginal misoprostol showed no differences in the need for further mechanical dilation or length of the procedure; similarly, the osmotic dilators Lamicel and Dilapan showed no differences in cervical ripening when compared to gemeprost, although gemeprost had cervical effects which were superior to laminaria tents. Older prostaglandin regimens (sulprostone, prostaglandin E2 and F2α) were associated with high rates of gastrointestinal side-effects and unplanned pregnancy expulsions.