Methods	RCT with randomisation of individual women. Randomisation was by a central telephone system at the Clinical Trials Unit using minimisation to ensure comparability
Participants	Inclusion criteria Women of less than 13 weeks’ gestation with a diagnosis of either incomplete miscarriage or early fetal/embryonic demise. Defined ICM as areas of mixed echogenicity within the uterine cavity with or without a disordered gestational sac. Early embryonic demise was defined as an intact gestational sac of greater than 20 mm mean diameter with no other internal structures and early fetal demise as a fetus over 6 mm crown rump length with no hart activity on transvaginal ultrasound scan. N = 1200 women. Incomplete miscarriage N = 274; early fetal demise N = 924. Exclusion criteria Women with severe haemorrhage or pain; pyrexia > 38°C; severe asthma; haemolytic disease or blood dyscrasias; current anticoagulation or systemic corticosteroid Rx; twin or higher order pregnancy; smoker aged > 35; inability to understand English.
Interventions	Intervention 1: vaginal misoprostol. 800 ug. N = 398 total; ICM = 90; IUFD = 308. Intervention 2: surgery - s. Suction curettage. N = 403 total; ICM = 92; IUFD = 310. Comparison: expectant care. N = 399 total; ICM = 92; IUFD = 306. All women were given a specific information sheet, 30 co-dydramol tablets, and an emergency telephone number
Outcomes	Primary outcome: gynaecological infection within 14 days of trial entry Secondary outcomes: antibiotics for presumed gynaecological infection within 14 days and within 8 weeks; duration of clinical symptoms (pain, additional analgesia, vaginal bleeding; days off work, days before return to usual daily activities); complications (fall in haemoglobin at 10-14 days, blood transfusion, unplanned consultations or admission within 14 days and within eight weeks); efficacy; psychological outcomes (depression and anxiety); and return to normal activity Unplanned curettage assessed at 2 weeks and 8 weeks.
Notes	Setting: early pregnancy assessment unit in 7 hospitals in UK. Results are reported by both IUFD and ICM. However, randomisation was not reported as stratified so there will be risk of bias in using data from the subgroups. Additional outcomes assessed but not pre-specified in the review: none.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Randomisation was by a central telephone system at the Clinical Trials Unit in Oxford. We used minimisation to ensure comparability between women with respect to participating centres, parity, type of miscarriage and gestation”
Allocation concealment (selection bias)	Low risk	Randomisation was by a central telephone system at the Clinical Trials Unit in Oxford, and although no specific information given on randomisation, clinical trials units generally use computer-generated random numbers list
Blinding (performance bias and detection bias) All outcomes	High risk	Not possible to blind women nor clinicians, because women given medical or surgical intervention were treatment in hospital and women in expectant arm were able to go home. Cannot blind surgery vs medical treatment or expectant care.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Loss of participants to follow up: loss immediately after randomisation: misoprostol = 0; surgery = 1; expectant care = 1. loss at 14 day outcomes; misoprostol = 9 ; surgery = 8; expectant care = 5. loss at 8 week outcomes: misoprostol = 3; surgery = 2; expectant care = 6. However, we do not know whether these women had ICM or IUFD, but at a maximum loss would be 10%. Exclusions after randomisation: In each of the surgical group and expectant care group, one woman with a viable pregnancy was excluded. Analysis was by ITT
Selective reporting (reporting bias)	Unclear risk	All important pre-specified outcomes were reported but we have not assessed the trial protocol
Other bias	Unclear risk	Study stopped early because struggling to recruit and not because of benefit, so bias unlikely. There were no important baseline differences between the 3 groups However, the randomisation was not reported as stratified by women with ICM and women with IUFD and so these may not have similar groups for comparison