Randomized Trial of Very Early Medication Abortion.

Background: Medication abortion, with a combination of mifepristone and misoprostol, is highly effective and safe. However, there is insufficient evidence on efficacy and safety at very early gestations before a pregnancy can be visualized with ultrasonography.

Methods: We conducted a multicenter, noninferiority, randomized, controlled trial involving women requesting medication abortion at up to 42 days of gestation with an unconfirmed intrauterine pregnancy on ultrasound examination (visualized as an empty cavity or a sac-like structure without a yolk sac or embryonic pole). Participants were randomly assigned to either immediate start of abortion (early-start group) or standard-care treatment delayed until intrauterine pregnancy was confirmed (standard group). The primary outcome was complete abortion. The noninferiority margin was set at 3.0 percentage points for the absolute between-group difference.

Results: In total, 1504 women were included at 26 sites in nine countries and were randomly assigned to the early-start group (754 participants) or the standard group (750 participants). In an intention-to-treat analysis, a complete abortion occurred in 676 of 710 participants (95.2%) in the early-start group and in 656 of 688 (95.3%) in the standard group; the absolute between-group difference was -0.1 percentage points (95% confidence interval, -2.4 to 2.1). Ectopic pregnancies occurred in 10 of 741 participants (1.3%) in the early-start group and in 6 of 724 (0.8%) in the standard group, with one rupture before diagnosis (early-start group). Serious adverse events occurred in 12 of 737 participants (1.6%) in the early-start group and in 5 of 718 (0.7%) in the standard group (P = 0.10); the majority were uncomplicated hospitalizations for treatment of ectopic pregnancy or incomplete abortion.

Conclusions: Medication abortion before confirmed intrauterine pregnancy was noninferior to standard, delayed treatment with respect to complete abortion. (Funded by the Swedish Research Council and others; VEMA EudraCT number, 2018-003675-35; ClinicalTrials.gov number, NCT03989869.).