Management of miscarriage: expectant, medical, or surgical? Results of randomised controlled trial (miscarriage treatment (MIST) trial)

Eligibility criteria

Women with a pregnancy of less than 13 weeks' gestation who had been diagnosed as having either an incomplete miscarriage or early fetal/embryonic demise were eligible. Exclusion criteria were severe haemorrhage or pain, pyrexia above 37.5°C, severe asthma, haemolytic disease or blood dyscrasias, current anticoagulation or systemic corticosteroid treatment, twin or higher order pregnancy, smoker aged over 35, and inability to understand written English.

We defined early embryonic demise as an intact gestation sac of greater than 20 mm mean diameter with no other internal structures and early fetal demise as a fetus of over 6 mm crown-rump length with no heart activity on transvaginal ultrasound scan.18 We defined an incomplete miscarriage as areas of mixed echogenicity within the uterine cavity with or without a disordered gestation sac.

Randomisation and blinding

Randomisation was by a central telephone system at the Clinical Trials Services Unit, Oxford. We used minimisation to ensure comparability between women with respect to participating centre, parity, type of miscarriage, and gestation.

Interventions

We gave all women a specific information sheet, 30 co-dydramol tablets, and an emergency telephone number. Women in the expectant management arm were allowed home with no intervention. In the medical management arm, women with an incomplete miscarriage were admitted to hospital and given a single vaginal dose of 800 μg misoprostol.19 20 Women with early fetal or embryonic demise were pre-treated with a single oral dose of 200 mg mifepristone,21 then admitted to hospital 24-48 hours later for a single vaginal dose of 800 μg misoprostol. A surgical evacuation of retained products of conception was offered if expulsion of retained products had not started within eight hours of insertion of misoprostol. Women in the surgical management arm were admitted for surgical suction curettage under general anaesthesia. No centre used prophylactic antibiotics at the time of curettage.

In all three groups, blood was taken for full blood count. Rhesus negative women were offered 250 IU of anti-D irrespective of their allocated management. A follow-up appointment was arranged 10-14 days after trial entry for a transvaginal ultrasound scan, full blood count, consultation with the study nurse, and examination by a gynaecologist if symptoms of infection were present. Retained products of conception were diagnosed if areas of mixed echogenicity within the uterine cavity were seen. A surgical curettage was offered if retained products of conception were present. Clinical symptoms were also taken into account; individual doctors in the early pregnancy clinics made the decision to offer curettage, in association with the women.

Study hypothesis

Our study hypothesis was that a clinically important difference would exist in documented gynaecological infection within 14 days of expectant or medical management compared with surgical management of first trimester miscarriage.

Outcomes

The primary outcome was documented gynaecological infection within 14 days of trial entry, defined as two or more of purulent vaginal discharge, pyrexia > 38.0°C, tenderness over the uterus on abdominal examination, and a white cell count above 15x10⁹/l. Secondary outcomes were treatment with antibiotics for presumed gynaecological infection within 14 days and within eight 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 admissions within 14 days and within eight weeks); efficacy—we defined a successful outcome as no unplanned surgical curettage within eight weeks; psychological outcomes (depression and anxiety on the hospital anxiety and depression scale at eight weeks); and return to normal activity (standard UK SF-36).

Sample size estimation

On the basis of the one published trial before the MIST trial started, we expected the incidence of the primary outcome in the standard care group (surgical management) to be 10%.6 To detect a 50% lower incidence of this outcome in the surgical group, compared with the expectant or medical management group, we needed to recruit 474 women to each group, giving a total sample size of 1422 women. This sample size would have 80% power to detect the treatment effect significant at the 5% level.

Statistical methods

We used SPSS to analyse data. We express the differences between groups as risk differences with 95% confidence intervals.

Trial management

A research fellow (JT) based at one of the centres coordinated the day to day activity of the seven participating centres. Randomisation, data management, and analyses were done at the National Perinatal Epidemiology Unit, Oxford. A multidisciplinary steering committee oversaw the trial. We established an independent data monitoring committee, which met annually during the period of recruitment to review interim analyses; its terms of reference stated that interim results should not be revealed to the steering committee unless a strong reason to alter the protocol or stop the trial emerged.

Participant recruitment, flow, and follow-up

Of 3905 women attending the early pregnancy clinics, we recruited and randomised 1200 (31%) women (figure); 1620 women refused trial entry and were offered routine surgical management; 1085 women were not eligible for entry to the study. The number of women recruited to the trial was lower than that needed to meet the original sample size calculation. Recruitment was slower than anticipated and despite an additional 33 months of recruitment, we recruited a total of 1200 women. Two women recruited to the trial were subsequently found to have a viable pregnancy. No important baseline differences were evident between the three groups (table 1).

Flow of patients through MIST trial

• Download figure
• Open in new tab
• Download powerpoint

Of the 402 women randomised to surgical management, 356 (89%) had a surgical curettage. Among the 46 women who did not have elective curettage, 30 (8%) miscarried before admission and 16 (4%) declined to have curettage after randomisation; however, 12 (3%) subsequently had curettage. In 12 (3%) women in the medical group, the miscarriage occurred spontaneously before admission (2 (0.5%) subsequently had curettage).

Gynaecological infection

We found no difference in the primary outcome measure—that is, the incidence of infection within the first 14 days—between the expectant group and the surgical group or between the medical group and the surgical group—surgical group 3% (12/402), expectant group 3% (11/398), medical group 2% (9/398) (table 2). The incidence of infection defined as prescription of antibiotic for presumed gynaecological infection within the first 14 days was significantly lower in the expectant group (17/398) compared with the surgical group (34/402) (risk difference 4%, 95% confidence interval 1% to 8%). The incidence of this outcome in the medical group (31/398) was not significantly different from that in the surgical group (risk difference 1%, − 3% to 5%). We found no significant difference at eight weeks for either group compared with the surgical group.

Unplanned hospital consultation and admissions

The number of unplanned hospital consultations (without admission), both within the first 14 days after randomisation and within the first eight weeks, was similar for all management groups. However, the number of unplanned hospital admissions was significantly higher in the expectant group (196, 49%) than in the surgical group (32, 8%) (risk difference − 41%, − 47% to − 36%) and was also higher in the medical group (72, 18%) (− 10%, − 15% to − 6%).

Unplanned curettage

Table 3 shows the number of women who had unplanned surgical curettage and the indications. One hundred and forty two (36%) women randomised to medical management had surgical curettage compared with 22 (5%) women who had unplanned curettage in the surgical group (risk difference − 30%, − 35% to − 25%). Of 142 women in the medical group who had curettage, the indication for this was failure of the medical management protocol in 90 (23%). A further 52 (13%) women had an unplanned curettage, 11 (3%) as an emergency procedure before anticipated admission.

The type of miscarriage was a factor in the success of management. Twenty (6%) women with early fetal demise who were allocated to the surgical group had an unplanned curettage, compared with a total of 116 (38%) of the medical group (risk difference − 31%, − 37% to − 25%) and 154 (50%) of the expectant group (− 44%, − 50% to − 37%) (table 4). Of women with an incomplete miscarriage, 2 (2%) in the surgical group had an unplanned curettage compared with a total of 26 (29%) in the medical group (risk difference − 27%, − 37% to − 17%) and 23 (25%) in the expectant group (− 23, − 33% to − 13%) (table 5).

Clinical symptoms, complications, and side effects

Cessation of bleeding after randomisation was significantly earlier in the surgical group than in the medical group (P = 0.0004) and the expectant group (P < 0.0001), but this did not affect haemoglobin concentration and haematocrit at 10-14 days. However, 7 (2%) women randomised to expectant management and 4 (1%) women randomised to medical management had a blood transfusion. No women randomised to surgical management had transfusions.

We found no significant difference in vomiting, diarrhoea, or pain (when assessed by the medical staff) in the medical group compared with the surgical group. We used the outcome of “extra analgesia” as a proxy measure to evaluate the need for analgesia by all three groups in the outpatient setting. The expectant management group received significantly more analgesia (table 2).

We found no difference between the three groups in the incidence of surgical complications in the women who had surgery. Surgical complications were reported in 2% (9/402), 1% (4/398), and 1% (4/398) of women allocated to receive surgical, expectant, and medical management respectively.

Psychological and activities outcomes

In all three groups the median time to return to usual daily activities was two days. Sick leave was also similar in all three groups—the median was nine days in the surgical group, eight days in the expectant group, and nine days in the medical group. No differences existed in anxiety or depression scores on the hospital anxiety and depression questionnaire six to eight weeks after miscarriage. Similarly, no differences existed in activities of daily living on any of the eight subscales of the UK SF-36.