Twenty-seven trials (n=3,177) were included; the number of participants in each trial ranged from 20 to 611.
The authors reported that 3 studies were double-blind, 16 adequately described the method of randomisation, 15 adequately described allocation concealment and 26 adequately described withdrawals.
Seven studies compared medical with expectant management (n=513). Medical management was almost three times more likely than expectant management to induce complete evacuation of conception products (RD 49.7%, 95% CI: 28.3, 71.1, p<0.001) with an NNT of 2. However, there was significant heterogeneity between the studies (p<0.001). Patient satisfaction could not be compared.
Five studies compared surgical with expectant management (n=545). There was no significant difference in complete evacuation of the uterine cavity between the two management strategies, although the results favoured surgical management (overall RD 16.2%, 95% CI: -2.5, 35.0, p=0.09). There was significant heterogeneity between the studies (p<0.001). Based on 2 studies with significant heterogeneity (p=0.02), there was no significant difference in the rate of patient satisfaction.
Nine studies compared surgical with medical management (n=1,376). Surgical management was more likely than medical management to induce complete evacuation of conception products (RD 32.8%, 95% CI: 14.4, 51.1; p<0.001) with an NNT of 3. However, there was significant heterogeneity between the studies (p<0.001). Patient satisfaction did not significantly differ between surgical and medical management.
Four studies compared vaginal with oral misoprostol (n=496). These medical treatments did not significantly differ in rates of successful treatment (RD 8.3%, CIs not reported, p=0.29). There was significant heterogeneity between the studies.
In addition, one study compared different doses of oral misoprostol, one compared misoprostol with dinoprostone, and one compared methotrexate plus misoprostol with misoprostol alone (overall n=247).
Subgroup analyses did not change directions of significance for any of the comparisons. No differences in major complications or adverse events were reported for any of the comparisons.