Methods	RCT with randomisation of individual women.
Participants	Inclusion criteria Women with incomplete spontaneous miscarriage and IUFD, with only women with incomplete miscarriage being included in this review. Women aged > 18 years with positive pregnancy test and with one of the following: in transvaginal ultrasonography an inhomogeneous mass with a diameter of 15-50 mm in the uterine cavity (incomplete spontaneous abortion); empty amnion sac with a diameter of > 15 mm (anembryonic pregnancy ); or crown-rump length > 5 mm without signs of fetal heart function (missed abortion). All kinds of miscarriage were included (missed abortion; anembryonic pregnancies; incomplete spontaneous abortion). N = 19 women (98 were randomised of which 19 had incomplete miscarriage). Exclusion criteria Women with profuse bleeding; signs of endometritis, allergies to either drug; severe asthma, suspected cases of molar or extrauterine pregnancy.
Interventions	Intervention: oral mifepristone + vaginal misoprostol. Oral mifepristone (200 mg) + vaginal misoprostol (800 ug). N = 11. Comparison: surgery. Curettage. N = 8. Some women (mainly nulliparous) were given 400 ug vaginal misoprostol 2 hours before to ripen the cervix.
Outcomes	Complete abortion rate; bleeding; pain; satisfaction; complications including infection (clinical signs or elevated infection parameters in lab tests) treated with oral or IV antibiotics; continuous and heavy bleeding; blood transfusions; curettage for any reason; intense pain requiring admission
Notes	Setting: Oulu University Hospital, Finland. The 19 women with incomplete spontaneous miscarriage were part of a larger study of 98 women who had had various forms of miscarriage (incomplete spontaneous miscarriage; anembryonic pregnancy; missed miscarriage). Separate data were available from the authors for the women with incomplete spontaneous miscarriage. Of the 19 women, 16 were < 13 weeks’ gestation and 3 were between 13 and 23 weeks’ gestation. This information is held at the Cochrane Pregnancy and Childbirth Group Office
Risk of bias
Bias	Authors’ judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“…computer randomised program with the block length of 6.”
Allocation concealment (selection bias)	Low risk	“An independent consult performed the randomisation and assigned the randomisation list to a secretary, who made the numbered opaque envelopes for the study. …Allocation concealment was used to confirm that neither the clinician nor the patient knew the type of treatment in advance… After informed consent the next numbered envelope was opened to define the type of treatment of each patient.”
Blinding (performance bias and detection bias) All outcomes	High risk	Cannot blind women nor clinicians to the treatment because this study compared medical versus surgical treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Authors randomised 98 women (19 with ICM and 79 with IUFD) with 49 in each group. Of these, they reported on 48 in medical and 47 in surgical groups because 1 woman in the medical management group had an ERPC and in the surgical group one woman had an emergency ERPC and one had a spontaneous complete miscarriage. Of these women only 19 had incomplete miscarriage (the remainder had intrauterine deaths) and of these all appear to be accounted for in the analysis
Selective reporting (reporting bias)	Unclear risk	We did not assess the protocol, and additionally the authors did not report on bleeding; blood transfusions’
Other bias	Low risk	No apparent additional biases apparent.