Secondary Logo

Journal Logo

Contents: Original Research

Pretreatment With Mifepristone Compared With Misoprostol Alone for Delivery After Fetal Death Between 14 and 28 Weeks of Gestation

A Randomized Controlled Trial

Allanson, Emma R. PhD; Copson, Sean MBBS; Spilsbury, Katrina PhD; Criddle, Sonya RM; Jennings, Belinda RM; Doherty, Dorota A. PhD; Wong, Antonia M. MClinPharm; Dickinson, Jan E. MD

Author Information
doi: 10.1097/AOG.0000000000004344
  • Free
  • Journal Club

The use of mifepristone across the gestational spectrum has increased since the first clinical trials in the 1980s investigating its use as an abortifacient in early pregnancy.1 Mifepristone (a competitive progesterone antagonist) primes the myometrium and cervix to respond to prostaglandins and is, therefore, used in combination with a prostaglandin analogue (eg, misoprostol). Mifepristone has been safely and efficaciously used in medical termination of pregnancy in the first and second trimesters.2,3

Several randomized trials have demonstrated the usefulness of mifepristone before prostaglandin use for second trimester termination where the fetus is alive. Mifepristone priming results in a shortened induction to abortion interval and a reduced dosage of prostaglandins required to achieve delivery.3–6 There are fewer data available on the role of mifepristone in termination of pregnancy after fetal death, and the relevance of blocking the effect of residual progesterone in a nonviable pregnancy is unclear. Two randomized controlled trials completed since commencement of this study demonstrated a reduction in the time to delivery after misoprostol administration with preceding mifepristone priming in the setting of fetal death in the second and third trimesters.7,8 The randomized trial of Chaudhuri and Datta7 investigated women with a fetal death after 20 weeks of gestation (median gestational age 32 weeks) using a regimen of 50–100 micrograms of vaginal misoprostol every 6 hours for a maximum of four doses and reported a significant reduction in delivery interval in those women randomized to mifepristone, compared with placebo. The recent trial of Bracken et al8 used 200 micrograms of buccal misoprostol every 3 hours for a maximum of 16 doses in pregnancies at 14–28 weeks of gestation, with a significant reduction in the fetal expulsion time for the women receiving mifepristone priming. There are also several cohort studies supporting the use of misoprostol after mifepristone priming in the setting of fetal death after the first trimester.

Our institution typically has used vaginal misoprostol after mifepristone priming following a randomized trial9 and wished to extend this experience to women with a fetal death. Therefore, to assess the efficacy of mifepristone, we conducted a randomized double blind placebo-controlled trial that compared misoprostol alone with misoprostol after mifepristone priming for the termination of a pregnancy after fetal death in women between 14–28 weeks of gestation. Our primary study hypothesis was that pretreatment with mifepristone would significantly decrease the delivery interval after prostaglandin commencement.

METHODS

From April 10, 2013, to December 10, 2013, women with a fetal death between 14–24 weeks of gestation who were admitted to King Edward Memorial Hospital in Perth, Western Australia were screened for enrollment in this randomized trial. The initial gestational age inclusion criteria were determined by usual practice in our unit for the management of a fetal death at less than 24 weeks of gestation; however, owing to slow recruitment, a protocol amendment was made and approved on the December 10, 2013, such that women at up to 28 weeks of gestation were eligible to be recruited until the trial end on the November 23, 2016. A sample size of 232 women was planned; however, we encountered a significant decline in women presenting to our hospital with a fetal death. It is likely that this was the result of changes in local protocols at other units in Western Australia, which increased their capacity to care for these women, where they previously would have been referred to our unit. The majority of women who presented to our hospital and met inclusion criteria were recruited; however, a pragmatic decision was made to cease the trial at 66 participants, given the slow accrual time to reach this number. All participants provided written consent. The trial was prospectively registered with ANZCTR (ACTRN12612000884808) and institutional ethics approval was obtained on March 6, 2012 (Women and Newborn Health Service Human Research Ethics Committee, 1977EW).

The primary aim of this trial was to compare the time interval from commencement of misoprostol to delivery of the fetus between the two study regimens. Secondary aims were to assess the incidence of maternal side effects, including postpartum blood loss, placental retention rates, the need for manual removal of placenta, the need for blood transfusion, the incidence of febrile illness, and the need for antiemetic and analgesic treatment. Women's satisfaction was also assessed. A secondary arm of this trial (with the results to be published separately) was to assess the circulating progesterone and estrogen levels in the trial patients as well as a cohort of women undergoing induced abortion of pregnancy with a live fetus. The aim of this was twofold: to consider whether the efficacy of mifepristone is related to the concentration of progesterone in the setting of fetal death, and to analyze the usual range of both progesterone and estrogen levels in pregnancy.

Women with gestational ages of 14 0/7 to 28 6/7 weeks and a confirmed fetal death in utero were invited to participate in the trial. If the gestational age was unknown, an ultrasonogram-confirmed crown–rump length of 85 mm was considered as the lower limit to allow inclusion. Women were excluded if they were taking corticosteroids (see below), had a documented allergy to misoprostol or mifepristone, had a history of more than three previous cesarean deliveries, presented with ruptured membranes, presented in spontaneous labor, or were non–English-speaking or were unable to provide written consent. Mifepristone is contraindicated in patients on long-term corticosteroids in Australia (as determined by the Australian Therapeutic Drug Administration), and this is written in the manufacturer guidelines in our country. We recognize that in other countries mifepristone may be “offered with precaution.”10

After obtaining written consent, one of the study investigators allocated a unique study participant number to the woman. This was sequential and documented in the study data set, as well as in the patient notes and in the inpatient medication chart. The medication chart was then provided to the pharmacy. Before the study commenced, the pharmacist (A.M.W.) had used a random sampling program to allocate participants to the case and control groups to the unique study identifiers, with 1:1 randomization. On receiving the medication chart, the trial pharmacist used the unique study participant number to provide the trial drug, ensuring that the study investigators, the participants, and the staff who administered the treatment remained blinded to the patient's allocated study group.

Women were randomized to receive either 200 mg of mifepristone or placebo 24–48 hours before misoprostol. The mifepristone was placed in capsules identical to the placebo capsules. The women were directly observed swallowing the trial medication. The women were discharged with safety advice and then admitted to the hospital 24–48 hours later for standard management of termination of pregnancy after a fetal death. According to standard unit protocols, women between 14 and 24 weeks of gestation received 400 micrograms of misoprostol vaginally every 6 hours, and women between 24 and 28 weeks of gestation were given 200 micrograms vaginally every 4 hours. Women who had not delivered within 24 hours of misoprostol administration were reviewed by a consultant obstetrician for decision on the requirement for additional misoprostol.

In all groups, management of the placenta after delivery of the fetus followed the standard hospital protocol, with 10 units of oxytocin administered intramuscularly after delivery of the fetus. If expulsion of the placenta did not occur within 60 minutes of delivery (or if heavy vaginal bleeding occurred before this time) then manual removal of the placenta in the operating room was undertaken.

Hemoglobin and hematocrit were measured before administration of the trial drug and in the 24 hours after delivery. Maternal pulse, blood pressure, and temperature were recorded every 3 hours during the termination. A visual analogue scale was used at 3-hour intervals to assess nausea and pain. Estradiol and progesterone levels before administration of the trial drug were assessed on maternal serum. Objective documentation of vomiting and diarrhea was made by the nursing staff. The nature, frequency, and dosage of analgesia were recorded. At time of discharge, patients completed a four question visual analogue questionnaire (used in previous similar randomized trials conducted by our group9) with answers obtained using a visual analogue ruler scaled from 0 to 10, with 0 perceived as “much better than expected” and 10 as “much worse than expected”, for the following questions: What did you think of the procedure? How would you rate your pain during the procedure? Would you recommend this method of termination to a friend in a similar situation? How much control did you feel you had?

The trial was conducted in accordance with Good Clinical Practice. A Clinical Trials Notification application was approved by the Therapeutic Goods Administration.

Based on a median duration of labor with misoprostol alone in the second trimester of 13 (IQR, 11–22) hours,11 a sample size of 116 per group was calculated to achieve 80% power to detect a 30% reduction in the median duration of labor (equivalent to 9.1 hours) at a 0.05 significance level while using a two-sided log-rank test (PASS 2008). Equality of categorical and normally distributed continuous variables by treatment group were assessed using two-sided Fisher exact tests and t tests, respectively, on all available data for each variable. Where normality was in doubt, medians and interquartile range were reported and Wilcoxon rank-sum tests were performed. Because two women in each arm delivered before misoprostol, we were unable to undertake an intention-to-treat analysis given the primary outcome of time from misoprostol to delivery. Time to event analysis was performed for the outcome events of admission, misoprostol dose, delivery, and discharge. No censoring occurred because all included patients experienced the event. Kaplan-Meier graphs depicted survivor functions over time for each outcome event stratified by treatment groups. Cox models with the treatment group as the sole covariate were used to estimate the relative hazard rate of experiencing each outcome event. Effect modification by gestation and parity were tested by inclusion of interaction terms in the time from misoprostol to delivery Cox models. All models were tested for violation of the proportional hazards assumption. Stata 15 was used for analysis, and P<.05 was considered statistically significant.

RESULTS

Sixty-six women were randomized to placebo or treatment. The flow diagram for patient accrual is shown in Figure 1. Eleven women declined to participate in the trial, 19 women were unable to provide consent (most often due to being non–English speaking), and 15 women were unable to be recruited (for example attended out of hours). One woman withdrew after randomization but before treatment started, and another woman delivered between randomization and treatment. A further four women started treatment but experienced the delivery outcome before misoprostol was used. Baseline characteristics did not differ between groups and are shown in Table 1.

Fig. 1.
Fig. 1.:
Flow diagram for patient accrual.Allanson. Mifepristone for Pregnancy Termination for Fetal Death. Obstet Gynecol 2021.
Table 1.
Table 1.:
Baseline Characteristics

Women in the treatment group had fewer doses of misoprostol (2.1 vs 3.4, P=.002) and less total misoprostol (a mean of 767.7 micrograms vs 1,181.8 micrograms, P=.003). The median time from misoprostol to delivery in the placebo group was 10.5 hours, compared with 6.8 hours in the treatment group (hazard ratio 2.41 95% CI 1.39–4.17, P=.002) (Fig. 2). The effect of mifepristone on time from misoprostol to delivery did not seem to vary among women with gestational ages of less than 20 weeks compared with those with 20 weeks or more (P=.795). However, owing to the relatively small study sample size, we lacked statistical power to detect any small differences in time to delivery. Parous women in the mifepristone group had a significantly shorter time to delivery after misoprostol compared with parous women in the placebo group (hazard ratio 2.81 95% CI 1.412–5.57, P=.003). There was no evidence that time to delivery after misoprostol varied between mifepristone and placebo in nulliparous women; however, the numbers of women included in this subgroup analysis is small (n=22).

Fig. 2.
Fig. 2.:
Survivor function of time from misoprostol to delivery by treatment group. Placebo: median time to delivery 10.5 (interquartile range [IQR] 8.0–15.0). Mifepristone: median time to delivery 6.8 (IQR 5.3–10.6). Log-rank test P=.001.Allanson. Mifepristone for Pregnancy Termination for Fetal Death. Obstet Gynecol 2021.

There was no difference in the time from administration of mifepristone or placebo to delivery (median of 40.1 hours in the placebo group vs 42.3 hours in the treatment group). Nor was there any difference in the time from either admission to the hospital to discharge (median of 29.5 hours in the placebo group vs 28.3 hours in the treatment group), or from the commencement of misoprostol to discharge (median of 27.8 hours in the placebo group vs 26.3 hours in the treatment group).

Four women in the placebo group and no women in the mifepristone group took longer than 24 hours after misoprostol to deliver. Their gestational ages ranged from 16 to 22.1 weeks, and the total doses of misoprostol received were between 6 and 10. There were no apparent differences between these women and those who delivered in less than 24 hours.

Although not significant, women in the placebo group were more likely not to receive oxytocics in the third stage, and the reasons for this were not recorded. There was, however, no difference in the length of third stage, estimated blood loss, predelivery or postdelivery hemoglobin and hematocrit, or temperature at any time point. One woman in the mifepristone arm required a blood transfusion. Of the five women readmitted, four were for retained products of conception requiring suction curettage in the operating room (four in the placebo group and one in the mifepristone group) and one was for endometritis requiring intravenous antibiotics (in the mifepristone group). Only the patient with endometritis was admitted for longer than a day (Table 2). There was no significant difference between the treatment and placebo groups at the time of trial enrollment in the mean estradiol levels (15,418.1 pmol/L vs 9,860.5 pmol/L, P=.137) or mean progesterone levels (166.6 nmol/L vs 101.0 nmol/L, P=.164) at the time of enrollment in the trial.

Table 2.
Table 2.:
Maternal Complications

Satisfaction scores (from a scale of 0–10 with 0 perceived as “much better than expected” and 10 as “much worse than expected”) differed between the two groups for question of “what did you think of the procedure,” with women in the placebo group rating it worse than women in the mifepristone group (mean score 7.8 vs 6.3, P=.035). There was no difference in the other three questions about perception (Table 3).

Table 3.
Table 3.:
Women's Perceptions of the Procedure

DISCUSSION

This randomized trial demonstrates that the use of mifepristone before misoprostol administration for termination of pregnancy after fetal death between 14 and 28 weeks of gestation significantly reduces the time from the commencement of misoprostol to delivery, with no difference in adverse maternal events. Despite the early termination of our trial, we achieved our planned outcome of a 30% reduction in the median duration of labor. Parity (vs nulliparity) but not gestational age affected the time to delivery in this study. Unlike other randomized trials on the use of mifepristone in termination of pregnancy,9 our trial did not show any effect of gestational age on the primary outcome. It may be that this is a consequence of the small numbers of women in each gestational age group.

Our trial adds to the data of the two existing randomized controlled trials.7,8 Chauduri and Datta randomized 110 women with fetal death after more than 20 weeks of gestation (including those in the third trimester) to 200 mg of mifepristone followed by a misoprostol regimen of 100 micrograms vaginally every 6 hours in women at a gestational age of less than 26 weeks and 50 micrograms vaginally every 6 hours in women at a gestational age of more than 26 weeks. Despite a differing misoprostol regimen, they also found a reduction in time to delivery of the fetus after the commencement of misoprostol (16.3 hours compared, with 9.8 hours when mifepristone used, P<.001).7 A further randomized trial was published in 2020 by Bracken et al8 that used 200 micrograms of buccal misoprostol every 3 hours after mifepristone pretreatment at a gestational age range of 14–28 weeks. The authors reported completion of the delivery process by 48 hours in 82.2% of women in the mifepristone arm, compared with 81.4% of women in the placebo trial arm (P=.887); however, the median duration of delivery was significantly shorter in the mifepristone arm (7 hours vs 12 hours, P<.001). In a cohort of 96 patients, Wagaarachchi et al12 found that mifepristone before misoprostol for induction after a fetal death after 24 weeks of gestation resulted in an average time-to-delivery of 8.5 hours, although there was no control group. A 2007 retrospective study that compared mifepristone and misoprostol with misoprostol alone (with a variety of doses) in fetal death found a reduction in time to delivery in the mifepristone group only among those with gestational ages between 21–25 weeks.13 In their retrospective cohort study, Fyfe and Murray14 report a shorter duration of labor with the use of mifepristone before induction after 20 weeks of gestation, although it should be noted that only 20% of patients in the mifepristone treatment group had experienced a fetal death before induction. In addition to this, a nonblinded trial conducted in Nepal showed that the use of mifepristone after a fetal death from 20 weeks of gestation onward decreased the dose of misoprostol required, but not the total time from commencement of misoprostol to delivery.15 Two further prospective but nonblinded randomized trials in women with gestational ages of more than 28 weeks and with a fetal death demonstrated a shorter time to delivery and fewer doses of misoprostol needed after use of mifepristone.16,17

Before conducting this trial, given the existing data for the use of mifepristone at similar gestational ages in pregnancies with a live fetus and considering the antiprogesterone action of mifepristone, we hypothesized that the potential change in maternal serum progesterone levels after the occurrence of fetal death may affect the efficacy of mifepristone. However, we found no difference in maternal serum progesterone levels between the placebo and mifepristone groups, and our findings reflect those found in similar trials with live fetuses. In their double blinded trial, Ngoc et al4 randomized 260 women with a live fetus between 14 and 21 weeks of gestation to 200 mg mifepristone or placebo, followed by 400 micrograms misoprostol, buccally, every 3 hours and showed a reduction in the median interval to delivery in the treatment (8.1 hours vs 10.6 hours, P<.001). Dabash et al5 also used 200 mg mifepristone or placebo before 400 micrograms misoprostol buccally every 3 hours in women with a live fetus between 14 and 21 weeks of gestation and found a reduction in delivery 10.4±6.6 hours in the group that received mifepristone, compared with 20.6±9.7 hours in the misoprostol alone group (P<.001).

Although there was no difference in the maternal complications between the two groups, the risk and incidence of complications associated with medical termination of pregnancy in the second trimester must be an important component of clinical decision making. Our findings of a 29.7% retained placenta rate, a 26.5% need for manual removal of placenta, and an18.3% incidence of blood loss greater than 500 mL are more than what was found by Dickinson et al in women who were undergoing termination between 14 and 24 weeks of gestation9 and need to be considered when counselling patients. It is, however, plausible that the shortened time to delivery with the sequential use of mifepristone followed by misoprostol could result in fewer maternal complications; although this was not borne out in our data, this may be a function of smaller than planned recruitment and should not be completely discounted.

Although there was an improvement in the time from misoprostol to delivery in the treatment group, there were no significant differences between the two groups in either the total amount of time spent in hospital or the time from commencing misoprostol to leaving the hospital. In our unit, this is likely a function of most admissions historically being overnight (which is borne out in the admissions time in both groups being just longer than 24 hours) to allow for the perinatal care processes after terminating a pregnancy (pastoral care, the taking of memento photos and prints for parents, and time spent with the baby). However, the significant reduction in delivery time (as well as the difference in the number of women taking up to 24 hours to deliver) would plausibly allow for termination of pregnancy in cases such as those described here to be done as a day case, without the need for overnight hospital admission, as long as one could complete all of the other processes either during that time or in an outpatient capacity.

Women who received misoprostol had a better opinion of the procedure than those who received placebo, although neither rate it well. This is a significant finding, although we are limited in our ability to comment on the reasons both the ratings and the differences were observed in our trial. We appreciate the clinical circumstance of an unexpected fetal death may have contributed significantly to this. Moreover, some evidence suggests that women’s experience of a medical (as opposed to surgical) termination of pregnancy worsens with increasing gestational age,18 and this may be reflected here in the scores in both groups of women. It is also plausible that the relatively short time to delivery in the mifepristone group goes some way to account for the improved perception of the procedure in these women. Perception scores in all four categories in both groups of women were worse in our trial compared with the sequential mifepristone misoprostol trial in the second trimester of Dickinson et al,9 although with the notable difference of planned medical termination of pregnancy in that trial compared with the occurrence of unexpected fetal death in our group of women. The fact that women in our trial did not have a particularly good impression of control nor did they want to recommend the procedure to a friend may be a function of the clinical scenario that these women were experiencing.

This trial has several strengths. Randomization, allocation concealment and blinding were maintained throughout the trial. The gestational age range included is greater than any other single trial of mifepristone that includes women in the second trimester with a fetal death. Moreover, the mifepristone dose used is the same as the other randomized trial in fetal death in the second trimester,7 adding to the robustness of evidence for its use in this clinical scenario. We are, however, limited by unexpected slow recruitment and the need to terminate our trial before achieving the planned study size. Equally, the need for a protocol amendment to increase capacity for recruitment is potentially confounding given the different treatment regimen applied to patients who were at 24 weeks of gestation or more; however, this is likely reflective of real-world practice. Associations or lack thereof with the independent variables of gestational age and parity must, therefore, be interpreted with caution owing to the small numbers of women in these analyses.

Authors' Data Sharing Statement

  • Will individual participant data be available (including data dictionaries)? Yes, on request.
  • What data in particular will be shared? Deidentified, blinded data that underly the results reported in this article.
  • What other documents will be available? Study protocol is available atanzctr.org.au
  • When will data be available (start and end dates)? Beginning immediately and ending 24 months following article publication.
  • By what access criteria will data be shared (including with whom, for what types of analyses, and by what mechanism)? Following approval by a relevant institutional review board of a methodologically sound proposal, data will be available directly from the lead author to investigators at academic institutions.

REFERENCES

1. Rodger MW, Baird DT. Induction of therapeutic abortion in early pregnancy with mifepristone in combination with prostaglandin pessary. Lancet 1987;2:1415–8. doi: 10.1016/s0140-6736(87)91126-3
2. Kulier R, Kapp N, Gülmezoglu AM, Hofmeyr GJ, Cheng L, Campana A. Medical methods for first trimester abortion. The Cochrane Database of Systematic Reviews 2011, Issue 1. Art. No.: CD002855. doi: 10.1002/14651858.CD002855.pub4
3. Wildschut H, Both MI, Medema S, Thomee E, Wildhagen MF, Kapp N. Medical methods for mid-trimester termination of pregnancy. The Cochrane Database of Systematic Reviews 2011, Issue 11. Art. No.: CD005216. doi: 10.1002/14651858.CD005216.pub2
4. Ngoc NT, Shochet T, Raghavan S, Blum J, Nga NT, Minh NT, et al. Mifepristone and misoprostol compared with misoprostol alone for second-trimester abortion: a randomized controlled trial. Obstet Gynecol 2011;118:601–8. doi: 10.1097/AOG.0b013e318227214e
5. Dabash R, Chelli H, Hajri S, Shochet T, Raghavan S, Winikoff B. A double-blind randomized controlled trial of mifepristone or placebo before buccal misoprostol for abortion at 14-21 weeks of pregnancy. Int J Gynaecol Obstet 2015;130:40–4. doi: 10.1016/j.ijgo.2015.02.023
6. El-Refaey H, Templeton A. Pregnancy: induction of abortion in the second trimester by a combination of misoprostol and mifepristone: a randomized comparison between two misoprostol regimens. Hum Reprod 1995;10:475–8. doi: 10.1093/oxfordjournals.humrep.a135965
7. Chaudhuri P, Datta S. Mifepristone and misoprostol compared with misoprostol alone for induction of labor in intrauterine fetal death: a randomized trial. J Obstet Gynaecol Res 2015;41:1884–90. doi: 10.1111/jog.12815
8. Bracken H, Ngoc NTN, Ha DQ, Paredes NR, Quyet VB, Linh NTH, et al. Mifepristone pretreatment followed by misoprostol 200 mcg buccal for the medical management of intrauterine fetal death at 14-28 weeks: a randomized, placebo-controlled, double blind trial. Contraception 2020;102:7–12. doi: 10.1016/j.contraception.2020.02.007
9. Dickinson JE, Jennings BG, Doherty DA. Mifepristone and oral, vaginal, or sublingual misoprostol for second-trimester abortion: a randomized controlled trial. Obstet Gynecol 2014;123:1162–8. doi: 10.1097/aog.0000000000000290
10. Davey A. Mifepristone and prostaglandin for termination of pregnancy: contraindications for use, reasons and rationale. Contraception 2006;74:16–20. doi: 10.1016/j.contraception.2006.03.003
11. Dickinson JE, Evans SF. The optimization of intravaginal misoprostol dosing schedules in second-trimester pregnancy termination. Am J Obstet Gynecol 2002;186:470–4. doi: 10.1067/mob.2002.121085
12. Wagaarachchi PT, Ashok PW, Narvekar NN, Smith NC, Templeton A. Medical management of late intrauterine death using a combination of mifepristone and misoprostol. BJOG 2002;109:443–7. doi: 10.1111/j.1471-0528.2002.01238.x
13. Väyrynen W, Heikinheimo O, Nuutila M. Misoprostol-only versus mifepristone plus misoprostol in induction of labor following intrauterine fetal death. Acta Obstet Gynecol Scand 2007;86:701–5. doi: 10.1080/00016340701379853
14. Fyfe R, Murray H. Comparison of induction of labour regimes for termination of pregnancy, with and without mifepristone, from 20 to 41 weeks gestation. Aust N Z J Obstet Gynaecol 2017;57:604–8. doi: 10.1111/ajo.12648
15. Agrawal A, Basnet P, Thakur A, Rizal P, Rai R. Induction of labor using misoprostol with or without mifepristone in intrauterine death. JNMA J Nepal Med Assoc 2014;52:785–90.
16. Panda S, Jha V, Singh S. Role of combination OF mifepristone and misoprostol verses misoprostol alone in induction of labour in late intrauterine fetal death: a prospective study. J Fam Reprod Health 2013;7:177–9.
17. Sharma D, Singhal SR, Paul A. Comparison of mifepristone combination with misoprostol and misoprostol alone in the management of intrauterine death. Taiwan J Obstet Gynecol 2011;50:322–5. doi: 10.1016/j.tjog.2011.07.00
18. Robson SC, Kelly T, Howel D, Deverill M, Hewison J, Lie ML, et al. Randomised preference trial of medical versus surgical termination of pregnancy less than 14 weeks' gestation (TOPS). Health Technol Assess 2009;13:1–124, iii-iv. doi: 10.3310/hta13530
Figure
Figure

Supplemental Digital Content

© 2021 by the American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.