NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Ectopic Pregnancy and Miscarriage: Diagnosis and Initial Management in Early Pregnancy of Ectopic Pregnancy and Miscarriage. London: RCOG; 2012 Dec. (NICE Clinical Guidelines, No. 154.)

  • In April 2019 NICE updated its guideline on ectopic pregnancy and miscarriage. See the evidence reviews for the areas in which new recommendations were developed. The 2012 recommendations have been retained in the new guideline. This 2012 full guideline includes the evidence supporting the 2012 recommendations and has not been updated.

In April 2019 NICE updated its guideline on ectopic pregnancy and miscarriage. See the evidence reviews for the areas in which new recommendations were developed. The 2012 recommendations have been retained in the new guideline. This 2012 full guideline includes the evidence supporting the 2012 recommendations and has not been updated.

Cover of Ectopic Pregnancy and Miscarriage

Ectopic Pregnancy and Miscarriage: Diagnosis and Initial Management in Early Pregnancy of Ectopic Pregnancy and Miscarriage.

Show details

7Management of threatened miscarriage and miscarriage

7.1. Introduction

Threatened miscarriage is the most common complication of early pregnancy, occurring in approximately 20% of pregnant women before 20 weeks of gestation (Sotiriadis et al, 2004). Although many women who have threatened miscarriage go on to have a successful pregnancy, there is an increase in risk of miscarriage in the same pregnancy of 2.6 times and 17% of women with threatened miscarriage go on to have further complications in the same pregnancy. In the UK, it is estimated that around a quarter of a million pregnancies each year end in a miscarriage (The Miscarriage Association, 2011). This loss is associated with a significant amount of physical and psychological morbidity. This chapter presents evidence and guidance for clinically effective and cost-effective care for women with miscarriage considering both clinical and psychological outcomes.

7.2. Progesterone for threatened miscarriage

Review question

What is the effectiveness of progesterone in improving outcomes in women with threatened miscarriage?

Introduction

Progesterone is an essential hormone secreted by the corpus luteum that provides early pregnancy support until placental production takes over at 10 to 12 weeks of gestation. Historically, low levels of circulating progesterone have been linked to impending miscarriage and the presence of associated vaginal bleeding. It has been postulated, therefore, that a lack of progesterone is a cause of miscarriage rather than a secondary signal of failing pregnancy.

This review analyses the evidence from published studies where progesterone/progestogen supplementation has been introduced in pregnancies complicated by threatened miscarriage in the first trimester (presence of vaginal bleeding before 12+6 weeks of gestation). Various outcomes were examined to determine any detrimental effect or proven efficacy.

Description of included studies

Six studies were included in this review (Duan et al., 2010; El-Zibdeh et al., 2009; Pandian, 2009; Gerhard et al., 1987; Omar et al., 2005; Palagiano, et al. 2004).

Four studies are randomised trials (El-Zibdeh et al., 2009; Pandian, 2009; Gerhard et al., 1987; Palagiano, et al. 2004) and two are observational studies (Duan et al., 2010; Omar et al., 2005). One study was conducted in Jordan (El-Zibdeh et al., 2009), two in Malaysia (Pandian, 2009; Omar et al., 2005), one in Germany (Gerhard et al., 1987), one in China (Duan et al., 2010) and one in Italy (Palagiano et al., 2004).

Three included studies (Duan et al., 2010; Gerhard et al., 1987; Palagiano, et al. 2004) assessed the efficiency of progesterone administration in women with bleeding in early pregnancy and three other studies (El-Zibdeh et al., 2009; Pandian, 2009; Omar et al., 2005) evaluated the effect of progestogen (dydrogesterone) on pregnancy outcomes for threatened miscarriages. Route of administration varied, consisting of intramuscular administration in one study (Duan et al., 2010), oral in three studies (El-Zibdeh et al., 2009; Pandian, 2009; Omar et al., 2005) and vaginal pessary in two studies (Gerhard et al., 1987; Palagiano et al., 2004).

Evidence profile

Table 7.1. GRADE summary of findings for comparison of progesterone with no treatment or placebo.

Table 7.1

GRADE summary of findings for comparison of progesterone with no treatment or placebo.

Evidence statements

Term birth

One meta-analysis of two studies found no statistically significant difference in term birth in women who received progesterone/progestogen treatment compared with women who had no treatment. The evidence for this finding was of low quality.

Preterm birth

One meta-analysis of two studies and one further study found no statistically significant difference in preterm birth in women who received progesterone/progestogen treatment compared with women who had no treatment. The evidence for these findings was of low and very low quality.

Miscarriage

One meta-analysis of four studies found a reduced incidence of miscarriage in women who received progesterone/progestogen treatment compared with women who had no treatment. This finding was statistically significant and the evidence for this finding was of low quality. One further study found that incidence of miscarriage was lower in women who received progestogen treatment compared with women who had no treatment. This finding was statistically significant. When stratified sub-group analyses were conducted for women with vaginal bleeding, women with vaginal spotting, women with fetal heart activity, women with presence of yolk sac and women with regular intrauterine gestational sac the findings favoured the progesterone group but the difference was no longer statistically significant. The evidence for these findings was of very low quality.

Miscarriage (oral progesterone)

One meta-analysis of two studies found a reduced incidence of miscarriage in women who received oral progestogen treatment compared with women who had no treatment. This finding was statistically significance and the evidence for this finding was of low quality.

Miscarriage (vaginal progesterone)

One meta-analysis of two studies found no statistically significant difference in incidence of miscarriage in women who received vaginal progesterone treatment compared with women who had no treatment. The evidence for this finding was of low quality.

Pregnancy rate at 20 weeks

Two studies found that the rate of pregnancy at 20 weeks was higher in women who received progestogen treatment compared with women who had no treatment. This finding was statistically significant in both studies. The evidence for this finding was of low quality in one study and very low in the other. When stratified sub-group analyses were conducted for women with vaginal bleeding, women with vaginal spotting, women with fetal heart activity and women with presence of yolk sac the findings favoured progesterone treatment but the difference was no longer statistically significant. The evidence for these findings was of very low quality.

Placental abruption

One study found no statistically significant difference in incidence of placental abruption in women who received progesterone treatment compared with women who had no treatment. The evidence for this finding was of very low quality.

Hypertensive disorders in pregnancy

One meta-analysis of two studies and one further study found no statistically significant difference in incidence of hypertensive disorders of pregnancy in women who received progesterone/progestogen treatment compared with women who had no treatment. The evidence for this finding was of very low quality.

Gestational diabetes

One study found no statistically significant difference in incidence of gestational diabetes in women who received progesterone treatment compared with women who had no treatment. The evidence for this finding was of very low quality.

Intrahepatic cholestasis of pregnancy

One study found no statistically significant difference in intrahepatic cholestasis of pregnancy in women who received progesterone treatment compared with women who had no treatment. The evidence for this finding was of very low quality.

Pain score at the end of 5 day treatment

One study found that the mean pain score at the end of the 5 days of treatment was lower in women who received progesterone treatment compared with women who had placebo treatment. This finding was statistically significant. The evidence for this finding was of moderate quality.

Health economics

The cost-effectiveness analysis undertaken for this guideline suggested that progesterone for threatened miscarriage was cost effective when compared with no treatment. In the base case analysis progesterone dominated no treatment, producing an incremental quality adjusted life year (QALY) gain and a cost saving of £49. Progesterone saved costs because the savings from averted miscarriage more than offset treatment costs. Probabilistic sensitivity analysis found progesterone to be cost effective in 99.93% of Monte Carlo simulations. In sensitivity analysis when a much higher treatment cost of £200 was assumed, progesterone still had an 83% probability of being cost effective at a willingness to pay threshold of £20,000 per QALY. The model is described in more detail in Section 10.2.

Evidence to recommendations

Relative value placed on the outcomes considered

The guideline development group (GDG) agreed that the most important outcomes were the rate of term pregnancy, miscarriage and pregnancy rate beyond 20 weeks of gestation. It also recognised that the side effects associated with progesterone treatment is an important outcome. The group had hoped that there would be evidence regarding long-term outcomes of progesterone use, but none was reported in the included studies.

Consideration of clinical benefits and harms

The evidence from randomised controlled trials (RCT) showed that progesterone or dydrogesterone treatment and placebo or no treatment of miscarriage had a similar effect on both term and preterm birth. Progesterone or dydrogesterone treatment was significantly associated with fewer miscarriages, less severe pain and higher rate of pregnancy at 20 weeks of gestation in women with threatened miscarriages. A significant difference in favour of dydrogesterone for the pregnancy rate at 20 weeks was also shown in a small and underpowered observational study. However, when a pre-specified stratified analysis was performed based on vaginal bleeding, vaginal spotting and presence of yolk sac, the result was no longer statistically significant, though this could be due to the very small sample size.

The GDG acknowledged the importance of the chosen end point for each of the included studies. The studies chose different end points (some at 20 weeks of gestation, others at birth), but none adopted an end point beyond the birth and none included neonatal congenital abnormalities as an outcome. The GDG was concerned that routine administration of progesterone or dydrogesterone might interfere with ‘natural’ miscarriages that were associated with genetic abnormalities in the fetus. This, in turn, could result in an increase in the rate of neonatal abnormalities or later miscarriages.

The evidence showed that women receiving progesterone treatment and women receiving placebo or no treatment of threatened miscarriage had similar rates of pregnancy complications (gestational diabetes, hypertension in pregnancy and intrahepatic cholestasis of pregnancy).

The group noted that there was no evidence available for longer term outcomes such as developmental delay and incidence of congenital abnormalities. While they recognised that there was no evidence suggesting short-term harm, the GDG members felt that without evidence about the longer term effects of progesterone or dydrogesterone, they would be concerned about recommending its use. In particular, the group was concerned about the use of synthetic progestogens as they believed that these were more likely to be associated with poor long-term outcomes.

Separately from the issue of whether or not progesterone or dydrogesterone should be offered, the GDG agreed that there should be a follow-up procedure in place for women with a threatened miscarriage. The GDG members agreed that if the bleeding gets worse, women should be advised to return in order to receive further assessment. They also agreed that there should be a follow-up scan if the bleeding persists in order to determine whether or not the pregnancy is still viable. The group wished to strike a balance between offering women reassurance and ensuring that not all women would need to return for a scan. Ultimately, it was agreed that 14 days would be an appropriate time to wait before offering a follow-up scan.

Information giving and emotional support

The GDG recognised that for many women threatened miscarriage will be a stressful and difficult time. Healthcare professionals providing care for these women will need to provide accurate information and communicate this in a way that balances optimism with a degree of caution. As well as providing information about what to expect, women should be given details of support organisations and advice about what to do in an emergency. The GDG also recognised that the offer of a follow-up ultrasound scan for women in whom bleeding persists was an important source of support.

Consideration of health benefits and resource uses

The economic evaluation undertaken for this guideline demonstrated that progesterone or dydrogesterone treatment was likely to be a cost-effective treatment for women with threatened miscarriage as it not only reduces the threat of miscarriage but also produces net savings as a result of averted miscarriage. However, the GDG had reservations about some of the studies which informed the treatment effect size in the health economic model. Furthermore, in the UK it is not usual practice to offer progesterone or dydrogesterone for threatened miscarriage, and the group had concerns about recommending a significant change in practice based on such poor quality evidence.

Quality of evidence

The evidence available for this question was generally of low or very low quality. The few studies that were included generally had a low numbers of participants, and the GDG was concerned that even the randomised trials that were included were potentially subject to bias. Two of the trials were funded by manufacturers of progesterone, including one trial with a single author and which had a high loss to follow-up.

Overall, the GDG felt that the evidence was insufficient to recommend the use of progesterone or dydrogesterone. This was partly because there was no demonstrated significant difference in the rate of term birth, but mainly because of the GDG's concern about the lack of long-term safety data. The group felt strongly that further, high quality studies investigating both the efficacy and safety of progesterone and progestogens were needed, and decided that this was a priority area for research, particularly considering outcomes of term birth, late miscarriage and incidence of congenital abnormalities.

Recommendations

NumberRecommendation
48Advise a woman with vaginal bleeding and a confirmed intrauterine pregnancy with a fetal heartbeat that:
  • if her bleeding gets worse, or persists beyond 14 days, she should return for further assessment
  • if the bleeding stops, she should start or continue routine antenatal care.
NumberResearch recommendation
RR 5Are progesterone or progestogens effective in treating threatened miscarriage?
Why this is important
Approximately 20% of pregnancies miscarry in the first trimester and many women will experience some bleeding and/or pain in early pregnancy that does not cause miscarriage. In many countries, women with bleeding and/or pain will be treated with progesterone or progestogens to try and decrease the risk of miscarriage. The evidence for the effectiveness of this treatment has been inconclusive, but data from a meta-analysis of several small studies suggest that progestogens are better than placebo. However, there are theoretical risks to prescribing any treatment in pregnancy and for many practitioners this will be a major change in practice. The lack of strong evidence makes this a priority area for research.

A very large multicentre randomised controlled trial of women treated with either progesterone/progestogen or placebo should be conducted. The trial should be large enough so that it is sufficiently powered to detect differences in long-term outcomes. The population would be women with pain and bleeding and a spontaneous, confirmed, viable, singleton, intrauterine pregnancy between 6 and 12 weeks gestation. Progesterone/progestogen or placebo would be administered from when bleeding starts until the end of the 13th week. Pregnancy proceeding beyond the end of the first trimester might be the primary outcome. Live birth should also be measured, as well as pregnancy outcome, gestation at birth and presence of congenital abnormalities.

7.3. Expectant management compared with active treatment of miscarriage

Review question

How effective is expectant management of miscarriage compared with active treatment for improving women's clinical and psychological outcomes?

Introduction

Although historically miscarriages were often treated with a surgical procedure, there are now other less invasive options available, in the forms of medical treatment and expectant management. However, the ability of women to access each mode of management varies across England and Wales, a fact which could be attributed to uncertainty about their relative efficacy and risk of complications. Therefore, these reviews aimed to establish which treatment option is the most clinically effective and cost effective, recognising the importance of women's psychological outcomes. This includes establishing whether simply allowing the natural process of miscarriage to complete its course leads to any worse outcomes than if medical or surgical treatment are used.

Description of included studies

Twelve studies were included in this review of which one was a qualitative paper (Smith et al., 2006) and the other 11 were reports of seven RCTs (Blohm et al., 1997; Chipchase & James, 1997; Ngai et al., 2001; Nielsen & Hahlin, 1995; Nielsen et al., 1996; Nielsen et al., 1999; Shelley et al., 2005; Smith et al., 2009; Trinder et al., 2006; Wieringa-de Waard et al., 2002a; Wieringa-de Waard et al., 2002b).

The RCT papers report the outcomes and follow-up data of seven trials conducted in the UK (one trial reported on by Chipchase & James, 1997 and a second trial reported on by Smith et al., 2009 and Trinder et al., 2006), Australia (Shelley et al., 2005), Sweden (one trial reported on by Nielsen & Hahlin, 1995, Nielsen et al., 1996 and Blohm et al., 1997 and a second trial reported on by Nielsen et al., 1999), The Netherlands (one trial reported on by both Wieringa-de Waard et al., 2002a and Wieringa-de Waard et al., 2002b) and Hong Kong (Ngai et al., 2001). The qualitative study is the follow-up to an RCT conducted in the UK, including both participants and non-participants of the trial (Smith et al., 2006).

All studies compared expectant management with medical and/or surgical management of miscarriage (both of which in isolation or in combination were defined as ‘active’ by the GDG) and reported at least one priority outcome. The GDG felt that the experience of being in the placebo arm of a randomised controlled trial was not comparable to expectant management because women receiving a placebo are blinded to whether they are receiving an active mode of management, which may have an effect on outcomes. Therefore placebo controlled trials were not included for this comparison and were instead included for the review question on the appropriate dose of misoprostol and mifepristone (see Section 7.5).

The trials were all conducted in developed countries and their populations include women with missed miscarriages and/or women with ongoing miscarriages. One RCT (Wieringa-de Waard et al., 2002b) also included a group of women who did not accept randomisation but instead chose to be managed according to their preferences. They were followed up in exactly the same way as the randomised group and their outcomes were analysed separately and compared to the randomised group.

Evidence profile

Thirteen outcomes (grouped in seven broad categories) were chosen by the GDG as being of priority to inform recommendations.

Heterogeneity was low (under 60%) for all outcomes except for bleeding duration (two trials) where it was 86%; therefore for this outcome, the NCC-WCH technical team used a random effects model in the meta-analysis. The high heterogeneity could be the result of the fact that one of the trials compared expectant with medical management whereas the other compared expectant with surgical management.

Table 7.2. GRADE summary of findings for comparison of expectant management with active treatment.

Table 7.2

GRADE summary of findings for comparison of expectant management with active treatment.

Two studies undertook additional data collection and analysis as a follow-up to an RCT. One of these is a UK study that includes qualitative data collection and analysis (Smith et al, 2006); the other was conducted in the Netherlands and compares quality of life and anxiety scores for both randomised and not randomised women taking part in a trial (Wieringa-de Waard et al, 2002b). Findings from these studies are summarised in Table 7.3 with quotations used to illustrate key themes identified from the qualitative research.

Table 7.3. Additional findings evaluating expectant and active miscarriage management strategies.

Table 7.3

Additional findings evaluating expectant and active miscarriage management strategies.

Evidence statements

Need for unplanned intervention

One study found that the need for unplanned intervention was higher in women who received expectant management compared with women who received active treatment. This finding was statistically significant and the evidence for this outcome was of high quality.

Infection

One meta-analysis of seven studies did not find a statistically significant difference in infection for women who received expectant management compared with women who received active treatment. The evidence for this outcome was of high quality.

Gastrointestinal side effects

One study found that gastrointestinal side-effects were lower in women who received expectant management compared with women who received active treatment. This finding was statistically significant and the evidence for this outcome was of high quality.

Need for a blood transfusion

One study found that the need for a blood transfusion was higher in women who received expectant management compared with women who received active treatment. This finding was statistically significant and the evidence for this outcome was of high quality.

Duration of bleeding

One study found that the duration of bleeding was longer in women who received expectant management compared with women who received active treatment. This finding was statistically significant. The evidence for this finding was of moderate quality. One study found that the duration of bleeding was longer in women who received expectant management compared with women who received surgical management. This finding was statistically significant. However, the same study did not find a statistically significant difference in duration of bleeding in women who received expectant management compared with women who received medical management. The evidence for both of these findings was of high quality. One meta-analysis of two studies and one further study did not find a statistically significant difference in duration of bleeding in women who received expectant management compared with women who received active treatment. The evidence for this finding was of moderate quality. One study reported duration of bleeding in a manner that did not allow assessment of statistical significance.

Pain

Three studies did not find a statistically significant difference in the duration of pain for women who received expectant management compared with women who received active treatment. One further study reported this outcome in a manner that did not permit assessment of statistical significance. The evidence for this outcome was of moderate quality.

Two studies did not find a statistically significant difference in the severity of pain for women who received expectant management compared with women who received active treatment. The evidence for this outcome was of moderate quality.

Unplanned admissions

One study found that unplanned admissions were higher in women who received expectant management compared with women who received active treatment. This finding was statistically significant and the evidence for this outcome was of high quality.

Women's satisfaction

Two studies did not find a statistically significant difference in satisfaction for women who received expectant management compared with women who received active treatment. The evidence for this outcome was of moderate quality in one study and low quality in the other.

Anxiety

Three studies did not find a statistically significant difference in anxiety for women who received expectant management compared with women who received active treatment. The evidence for this outcome was of moderate quality in one study and low in the others.

Mental health

One study found that mental health scores were higher in women who received expectant management compared with women who received active treatment. This finding was statistically significant. One further study found no statistically significant difference in mental health scores between the two groups. The evidence for this outcome was of low quality.

Live birth rate in a subsequent pregnancy

One study did not find a statistically significant difference in the live birth rate in a subsequent pregnancy for women who received expectant management compared with women who received active treatment. The evidence for this outcome was of moderate quality.

Subsequent conception rate

Two studies did not find a statistically significant difference in subsequent conception rate for women who received expectant management compared with women who received active treatment. The evidence for this outcome was of low quality.

Emotional and psychological outcomes

One qualitative study found that when comparing expectant management and active treatment there was general consensus among women related to the following issues:

  • fear of intervention, especially anaesthetic, hospitalisation and surgery
  • desire to have a predictable course/end in terms of management and symptoms
  • need for more information in terms of bleeding and pain, details on the timing, duration and effects of interventions.

In contrast there were areas with great variation in responses:

  • appropriateness and need for intervention (to miscarry naturally compared with having something done to help expedite completion)
  • awareness of the event (benefit in experiencing the event and saying goodbye compared with avoiding consciousness of the miscarriage)
  • the ‘baby’ (seeing it and saying goodbye compared with being scared about what they might see and feeling responsible for the ‘baby's’ death in case of misdiagnosis)
  • degree and experiences of pain and bleeding and care received (lack of caring by staff compared with it being reassuring to be at home).

The evidence for these findings was of high quality.

One study found that mental health scores were worse in women who were randomised to active treatment compared with women who had chosen active treatment. This difference was statistically significant and the evidence for this outcome was of very low quality.

The same study did not find a statistically significant difference between women randomised to expectant management and women who chose expectant management for mental health score or anxiety. In addition, the study did not find a statistically significant difference between women who chose expectant management and women who chose active treatment for either mental health score or anxiety. The evidence for these outcomes was of very low quality.

Health economics

One good quality economic evaluation of the miscarriage treatment (MIST) trial (Petrou et al., 2006) undertaken in an English setting found that the mean cost of surgical management was £200 more expensive than medical management (95% confidence interval [CI] £122–£278) and that the mean cost of medical management was £273 more expensive than expectant management (95% CI £160–£376). There was more than a 50% chance that expectant management was the most cost-effective treatment if the decision maker was not prepared to spend more than £70,000 to prevent a single gynaecological infection.

Five studies of lower quality (Hughes et al., 1996; Graziosi et al., 2005; Niinimaki et al., 2009; Rocconi et al., 2005; You & Chung, 2005) generally agreed with the economic evaluation of the MIST trial in that they all found surgical management to be more expensive than medical management. Where the studies also considered expectant management (Rocconi et al., 2005; You & Chung, 2005) this was found to be the least expensive alternative apart from a decision tree artefact where expectant management cost exceeded that of medical management because the expected management decision could be overridden by a patient preference for surgical treatment (You & Chung, 2005). For a full description of the literature review see Section 10.3.

Evidence to recommendations

Relative value placed on the outcomes considered

The GDG felt that the most important clinical outcomes were the need for an unplanned intervention, the incidence of infection, gastrointestinal side-effects and the need for a blood transfusion. While it recognised that live birth rate is an important outcome, it did not feel that it was likely the evidence would be of a sufficient size to show a statistically significant difference. Women's emotional and psychological outcomes were also prioritised. The group also recognised that women's satisfaction with their care is an important outcome, but felt that this is a difficult outcome to capture in research as the women are not able to express their satisfaction with the care they received compared with care that they did not receive. Thus it is difficult to determine which management strategy women are likely to prefer.

Consideration of clinical benefits and harms

The evidence showed that expectant and active treatment of miscarriage were similar in terms of severity of pain experienced, anxiety, women's satisfaction with care, number of days in pain and the incidence of infection. Expectant treatment of miscarriage was associated with fewer gastro-intestinal side-effects, but women having expectant management of miscarriage had more days of bleeding and a significantly greater chance of needing a blood transfusion. Unplanned intervention was significantly greater in the expectant management group.

Consideration of health benefits and resource uses

The review of health economic literature clearly demonstrated that expectant management is the most cost-effective approach by a significant margin, even taking into account the additional care that some women may require as the result of emergency procedures and unplanned interventions. As a result, the GDG agreed that, for the majority of women, expectant management should be the first-line treatment that is offered. The GDG members felt that, for most women, expectant management would be an acceptable or even preferable alternative, as it negates the risk of intervening and accidentally terminating a viable pregnancy. In addition, they noted the almost universal fear of intervention that was reported in the qualitative study. However, the GDG also recognised that for some women, expectant management will be unacceptable. It recognised the importance of being able to offer a choice if this were the case and so agreed that medical management could be offered as an alternative as this was the next most cost-effective treatment.

The group also recognised that there may be some clinical reasons why expectant management would not be appropriate. In particular, it noted the increased rate of blood transfusion following expectant management. Based on their clinical experience, the GDG members felt that this could be due to an increase in severe blood loss at later gestation. Using their clinical judgement, they agreed that, for women at increased risk of haemorrhage, such as those late in the first trimester or women at increased risk of the effects of haemorrhage (such as women with coagulopathies), other management options (both medical and surgical) should be considered. They were aware that this might have cost implications in terms of treatment, but felt that this should be weighed against the increased likelihood of a blood transfusion and need for emergency care in this subset of women, and the risk of poor outcomes following haemorrhage. The group also recognised that for women with evidence of infection, expectant management would be inappropriate and so other management options (both medical and surgical) should be considered. Finally, the group agreed that women who have had a previous traumatic experience associated with pregnancy (such as previous miscarriage or antepartum haemorrhage) should be offered their choice of treatment. The GDG felt that these women would likely be particularly anxious about having to wait for up to two weeks and more at risk of resultant psychological morbidities.

The GDG noted that there was a failure rate associated with expectant management of miscarriage, which is captured in a rate of further intervention of over 35%. It also recognised that for some women, a lengthy period of expectant management would not be acceptable; therefore the group felt that it was appropriate to designate an endpoint, after which women in whom expectant management had thus far been unsuccessful could choose to undergo an intervention to expedite the process. From their clinical experience, the GDG members felt that the natural course of a miscarriage might generally be expected to occur over a period of 7 to 14 days from the start of bleeding, and therefore that women should be reviewed after this time. The decision about the precise timing of the appointment for review should be made by the woman and healthcare professional together based on individual circumstances. Women should be rescanned after 7 to 14 days unless the symptoms are improving or have resolved, in which case a pregnancy test after 3 weeks would be sufficient.

Quality of evidence

The evidence available for this question was generally of high or moderate quality. In addition, it was felt that the health economic literature was of sufficient quality that it would be used to inform the review, rather than needing to develop a bespoke model for the question. While the quality of evidence was not high for all outcomes, it was for the majority, including those on which the group placed most emphasis, and thus the group felt confident in using the evidence to develop its recommendations.

Information giving and emotional support

One of the key themes from the qualitative data analysed in the review was the fact that women wanted more information regarding what to expect during their treatment. As expectant management may be an unfamiliar concept to some women, the GDG felt that it was vital that they be informed about the process, including what to expect in terms of pain and bleeding, what pain-relieving measures they might use as well as what options would be available in the event that expectant management does not result in the completion of the miscarriage. The GDG also felt that it was important that women be given details of support organisations, so that they could access support and counselling services. The GDG recognised that there can be a lot of information to give to women and agreed that healthcare professionals should ensure that sufficient time is made available to discuss all of the relevant issues and that this be supported with written information for the women to take away. In some instances, this might also mean arranging an additional appointment.

Although the qualitative studies for this review were looking at women's experiences of different treatment options, the GDG agreed that the evidence highlighted the more general point that women react to complications and the loss of pregnancy in different ways. The GDG therefore wished to stress the importance of giving information in a sensitive way, taking into account each individual woman's emotional response.

7.4. Surgical management compared with medical management of miscarriage

Review question

How effective is surgical management of miscarriage compared with medical management for improving women's clinical and psychological outcomes?

Introduction

The standard procedure for managing miscarriage has been surgical management. Surgical completion of miscarriage carries a risk of complications and often requires a general anaesthetic. Medical management may provide an alternative treatment but needs to be studied in terms of maternal outcomes, including emotional and psychological outcomes, as well as pain, bleeding and adverse clinical events including re-admission and unplanned intervention.

Description of included studies

Twenty-seven studies were included in this review (Chung et al., 1999; Dabash et al. 2010; Dao et al., 2007; Davis et al., 2007; de Jonge et al., 1995; Demetroulis et al., 2001; Egarter et al., 1995; Fang et al., 2009; Graziosi et al., 2004; Graziosi et al., 2005a; Graziosi et al., 2005b; Harwood & Nansel, 2008; Hinshaw, 1997; Lee et al., 2001; Montesinos et al., 2011; Moodliar et al., 2005; Muffley et al., 2002; Niinimaki et al., 2006; Sahin et al., 2001; Shelley et al., 2005; Shwekerela et al., 2007; Smith et al., 2006; Smith et al., 2009; Tam et al., 2005; Taylor et al., 2011; Trinder et al., 2006; Zhang et al., 2005).

One study is a qualitative study which followed up both participants and non-participants of an RCT conducted in the UK (Smith et al., 2006). Another study is a partially randomised trial, conducted in the UK, which included both women who had chosen their method of management and those who had been randomised to medical or surgical management (Hinshaw, 1997). The remainder of the included studies report the outcomes and follow-up data of 16 randomised controlled trials, conducted in the UK (Demetroulis et al., 2001 and a second trial reported on by both Smith et al., 2009 and Trinder et al., 2006), Australia (Shelley et al., 2005), Austria (Egarter et al., 1995), Burkina Faso (Dao et al., 2007), China (Fang et al., 2009), Ecuador (Montesinos et al., 2011), Egypt (Dabash et al., 2010), Finland (Niinimaki et al., 2006), Ghana (Taylor et al., 2011), Hong Kong (one trial reported on by Chung et al., 1999, Lee et al., 2001 and Tam et al., 2005), the Netherlands (one trial reported on by Graziosi et al., 2004, Graziosi et al., 2005a and Graziosi et al., 2005b), South Africa (two trials: de Jonge et al., 1995; Moodliar et al., 2005), Tanzania (Shwekerela et al., 2007), Turkey (Sahin et al., 2001), and the USA (one trial reported on by Davis et al., 2007, Harwood & Nansel, 2008 and Zhang et al., 2005 and a second trial reported on by Muffley et al., 2002).

All studies compared medical and surgical management of miscarriage and reported at least one priority outcome. The trials were conducted in developed and developing countries, and their populations include women with missed miscarriages and/or women with ongoing miscarriages.

Evidence profile

Sixteen outcomes (grouped into eight broad categories) were prioritised by the GDG to inform recommendations. In outcomes with high heterogeneity (over 60%), the NCC-WCH technical team used a random effects model (the remaining outcomes used fixed effects models) and explored the heterogeneity with sensitivity analyses. Findings from these sensitivity analyses are described here.

Table 7.4. GRADE summary of findings for comparison of medical management with surgical management.

Table 7.4

GRADE summary of findings for comparison of medical management with surgical management.

Gastro-intestinal side-effects

Overall heterogeneity (89%) was not significantly reduced by considering: only trials in developed countries; trials where only drugs administered vaginally were used; only trials where at least one oral drug was used; or only trials where the surgical arm received general anaesthetic. The technical team also explored the effect of dosage, with and without mode of administration. None of the combinations tested reduced the heterogeneity to below 60%. The high heterogeneity could be as a result of the variety of medical regimens used, the combination of three gastro-intestinal side-effects into one overall outcome, or the fact that it was a patient-reported outcome.

Unplanned visits to a medical facility

Considering only developed countries reduced the heterogeneity from 64% to 25% and resulted in there being no significant difference between the number of unplanned visits in the medical and surgical arms. This could be as a result of different patterns of healthcare-seeking behaviour in developed and developing countries.

Satisfaction

Overall heterogeneity (77%) was not reduced by considering only developed countries or removal of the trials in which the surgical arm received only local or verbal anaesthesia. The high heterogeneity could be a result of factors not generally reported in the trials, for example variable waiting times for surgery (reported in one trial as a contributing factor to dissatisfaction in the surgical patients).

Two UK studies undertook additional data collection and analysis as a follow-up to RCTs. One study (Smith et al, 2006) used qualitative methods of data collection and analysis, the other used a simple descriptive survey with data reported as descriptive statistics (Hinshaw, 1997). Findings from these studies are summarised in Table 7.5 with quotations used to illustrate key themes identified from the qualitative research.

Table 7.5. Additional findings evaluating medical and surgical miscarriage management strategies.

Table 7.5

Additional findings evaluating medical and surgical miscarriage management strategies.

Additional analysis of surgical complications

The rate of surgical complications is not reported as an outcome in the GRADE table because it is intuitive that the rate will be higher in women randomised to the surgical arm (the vast majority of whom undergo surgery) when compared to women randomised to the medical arm (a minority of whom end up having surgery). However, the GDG was concerned that potential side-effects of surgery should not be overlooked, and therefore further analysis was done based on data on surgical complications that was reported in the studies.

Five trials (Chung et al., 1999; Egarter et al., 1995; Graziosi et al., 2004; Muffley et al., 2002; Trinder et al., 2006) reported the rates of surgical complications, which included uterine perforation, cervical laceration, haemorrhage and intrauterine synechia. The incidence of surgical complications was generally low, ranging from 2% to 8% among women randomised to surgery. The incidence was under 5% in four out of the five studies that reported it.

The potential for reducing the risk of complications through cervical priming prior to surgery (as hypothesised in Chung et al., 1999) could not be assessed, because none of the studies that assessed surgical complication rates reported priming.

The association of complication rate with type of miscarriage (for example missed or incomplete) could not be explored because three of the studies only included women with missed miscarriage, and in the remaining two studies over 75% of women had a missed miscarriage.

Evidence statements

Need for an unplanned intervention

One meta-analysis of 18 studies found that the need for an unplanned intervention was higher in women who received medical management compared with women who received surgical management. This finding was statistically significant and the evidence for this outcome was of high quality.

Infection

One meta-analysis of seven studies did not find a statistically significant difference in the incidence of infection for women who received medical management compared with women who received surgical management. The evidence for this outcome was of high quality.

Gastro-intestinal side-effects

One meta-analysis of 12 studies found that gastro-intestinal side-effects were higher in women who received medical management compared with women who received surgical management. This finding was statistically significant. The evidence for this outcome was of moderate quality.

Need for a blood transfusion

One meta-analysis of eight studies did not find a statistically significant difference in the need for a blood transfusion for women who received medical management compared with women who received surgical management. The evidence for this outcome was of high quality.

Duration of bleeding

One meta-analysis of five studies and three other studies found that duration of bleeding was longer in women who received medical management compared with women who received surgical management. This finding was statistically significant. The evidence for this finding was of high quality in the meta-analysis and one single study, and moderate quality in the other two studies. Three further studies did not find a statistically significant difference in duration of bleeding between the two groups. The evidence for this finding was low quality in two studies and very low quality in the third. One further study reported duration of bleeding in a manner that did not allow assessment of statistical significance.

Pain

One study found that the duration of pain was longer in women who received medical management compared with women who received surgical management. This finding was statistically significant and the evidence for this finding was of high quality. Five further studies did not find a statistically significant difference in duration of pain between the two groups. The evidence for this finding was of moderate quality in two studies, low in two other studies and very low in the fifth study. One further study reported duration of pain in a manner that did not allow assessment of statistical significance.

One meta-analysis of three studies found that the severity of pain on a ten-point scale was higher in women who received medical management compared with women who received surgical management. This finding was statistically significant. The evidence for this finding was of high quality. One further study did not find a statistically significant difference in severity of pain on a ten-point scale between the two groups. The evidence for this finding was of low quality. One further study reported severity of pain on a ten-point scale in a manner that did not allow assessment of statistical significance.

Two studies found that the severity of pain on a seven-point scale was lower in women who received medical management compared with women who received surgical management. This finding was statistically significant and the evidence for this outcome was of moderate quality.

Unplanned visits to a medical facility

One meta-analysis of five studies did not find a statistically significant difference in the unplanned visits to a medical facility for women who received medical management compared with women who received surgical management. The evidence for this outcome was of low quality.

Unplanned admissions

One study found that unplanned admissions were higher in women who received medical management compared with women who received surgical management. This finding was statistically significant and the evidence for this outcome was of high quality.

Satisfaction

One meta-analysis of nine studies did not find a statistically significant difference in the reported satisfaction for women who received medical management compared with women who received surgical management. The evidence for this outcome was of low quality.

Social function at 2 weeks

Three studies did not find a statistically significant difference in the social function score at 2 weeks for women who received medical management compared with women who received surgical management. The evidence for this outcome was of high quality.

Mental health at 2 weeks

One meta-analysis of two studies did not find a statistically significant difference in the mental health score at 2 weeks for women who received medical management compared with women who received surgical management. The evidence for this outcome was of low quality.

Live birth rate in a subsequent pregnancy

One meta-analysis of two studies and one further study did not find a statistically significant difference in the live birth rate in a subsequent pregnancy for women who received medical management compared with women who received surgical management. The evidence for this outcome was of moderate quality and very low quality respectively.

Women's preferences, emotional and psychological outcomes

One qualitative study found that there was general consensus among women regarding fear of intervention, a desire for their treatment to follow a predictable course and the need for more information. In contrast, there was wider variation in women's feelings about the appropriateness of intervention, awareness of the event, feelings about the ‘baby’, the degree of pain and bleeding experienced, and the care received. The evidence for these findings was of high quality.

One partially randomised trial provided very low quality evidence that the acceptability of medical management was lower than surgical management, but was unaffected by whether women were randomised to, or had chosen, their method of management. Women who chose surgical management stated timescale, issues of awareness, avoidance of pain/bleeding and perceived effectiveness as the reasons for their preference. Women who chose medical management stated avoidance of general anaesthesia/surgery, and the feeling that it was more natural and that they were in control, as the main reasons for their preference. However, the evidence for these findings was of very low quality.

Evidence to recommendations

Relative value placed on the outcomes considered

For this review, the GDG felt that both quantitative and qualitative evidence were equally valuable. Clinically, the group felt that the need for further intervention, requirement for a blood transfusion and side-effects were important, in addition to other outcomes with cost implications, such as need for admission. From their clinical experience, and the qualitative evidence, the GDG members noted that women's responses to miscarriage and their preferences for mode of management are highly variable, and therefore the views and experiences of women reported in the qualitative study were vital in informing their decision. In contrast, the group did not feel that satisfaction with treatment and psychological scores were particularly useful outcomes. What evidence there was showed little difference between the two treatment arms, but the group felt that these outcomes are often difficult to capture accurately in randomised populations, and are less informative than qualitative data that explores women's experiences of their mode of management.

Consideration of clinical benefits and harms

Medical management of miscarriage avoids the need for surgery in over 70% of women. The GDG felt that this would be an important consideration for women, as surgery often requires a general anaesthetic and has an associated risk of complications. The risk of surgical complications among women randomised to surgery ranged from 2% to 8%, and included uterine perforation, haemorrhage, cervical lacerations and synechia. However, medical management is also associated with a significantly higher rate of unplanned intervention and unplanned admission. Gastro-intestinal side-effects, such as vomiting, are higher with medical management of miscarriage but the risk of infection and haemorrhage requiring transfusion is similar for both forms of active management. From their experience, the GDG members felt that this small risk of blood transfusion may be an over-estimation as in current clinical practice few women receive one.

The results of the qualitative studies supported the GDG's view that individual women often have very different priorities and expectations of their treatment. They recognised that, while women may strongly wish to avoid undergoing surgery, for some women the predictability, promptness and high likelihood of success following surgical treatment would be an attractive option. The GDG felt that, as the majority of women would have undergone up to 2 weeks of expectant management as a first-line treatment, it was important that they be given a choice about how to proceed at that point. Although the health economics analysis showed that medical management was more cost effective than surgical management, the GDG noted that this was based on estimates of first-line treatment, and therefore could not directly be applied to women in whom expectant management had already failed. The group felt strongly that after a period of expectant management women should have the choice of how to proceed, and therefore recommended that a discussion of the options take place.

The evidence showed that outcomes such as duration of pain seemed comparable in the medical and surgical arms, although the tendency was that pain lasted longer and was more severe after medical management. Similarly, the duration of bleeding was very variable but was generally less in the surgical arm than the medical arm. From their clinical experience, the GDG members felt that women's experiences of pain and bleeding after miscarriage also tend to be extremely variable. The GDG noted that medical management of miscarriage seems to be more successful (in terms of avoiding surgical intervention) in women with incomplete or inevitable miscarriage when compared with those with a missed miscarriage. It also recognised that successful treatment was higher in studies that allowed longer follow-up before surgical intervention and where follow-up was clinical rather than ultrasound orientated. However, due to the differences between the studies, the GDG did not feel that the evidence was strong enough to make a recommendation that might supersede women's choice. Overall, it noted that medical management had both advantages (in terms of avoiding surgery) and disadvantages (in terms of the potential for increased pain and bleeding) and therefore the individual woman's preference and specific clinical situation should inform the choice of second-line management strategy.

Consideration of health benefits and resource uses

While both unplanned admissions and need for an unplanned intervention are higher after medical management of miscarriage compared to surgical treatment, the health economics analysis of first-line treatment options calculated that medical management was more cost-effective than surgical management due to the reduced cost of the initial treatment. However, the GDG felt that, having recommended expectant management (the most cost-effective option) as a first-line management strategy on the grounds of cost, or medical management if expectant management was unacceptable, it was appropriate that women then have a choice of all treatments if the first management strategy failed, particularly as the health economics was based on outcomes of first-line treatment. It noted that, for women in whom expectant management had not been successful, the success of medical management was likely to be reduced, and therefore the associated costs of unplanned interventions and admissions would be increased.

Quality of evidence

Much of the evidence for this review was of high or moderate quality. In particular, the GDG welcomed the inclusion of the MIST trial, a high quality randomised controlled trial conducted in the UK, with an associated qualitative study investigating women's views of different modes of miscarriage management. Because it explored women's experiences so comprehensively, and it was conducted in the UK, the GDG believed that it was likely to represent the spectrum of different views that women might have regarding their preferred treatment options. However, as it was only one study involving a small group of women, the GDG felt that it was important to recommend that further research be done to evaluate whether different modes of management impact on patient experience and longer term psychological and emotional outcomes.

Information giving and emotional support

The GDG noted that an overarching theme from the qualitative data was the fact that women wanted more information about what to expect, what their course of treatment would entail (including potential complications) and what support would be provided (both immediately and longer term). It was noted that a lack of information often led to uncertainty which could heighten women's anxiety. Therefore, the group felt that it was important that women were informed about the possible course of events following their chosen management course, including what to expect in terms of the duration and severity of bleeding and where and when to get help in an emergency. In addition, it was the experience of some of the group that women are often uncertain about what to expect in the recovery period, and that they therefore need to be given more information about this, including details of how to access counselling and other support services.

Other considerations

Choice of treatment is important for women and satisfaction is higher where women have been offered and exercised their choice. The GDG felt it important to support women's choice following a period of expectant management where this had not been successful. It recognised the potential for increased psychological sequelae if women were denied a choice after 7–14 days of expectant management, during which time they may have been continuing to bleed and desiring a prompt completion of the process. The GDG recognised that for women with greater difficulty in accessing health care (for example women with English as a second language, drug users, travellers or those living in a remote area), surgical management might be preferable as a second-line strategy due to the reduced need for unplanned intervention and unplanned admission. However, as the GDG was recommending that all treatment options should be discussed with women if their first-line treatment is not successful, it did not feel it was necessary to make any specific recommendations for these groups of women.

Recommendations

NumberRecommendation
49Use expectant management for 7–14 days as the first-line management strategy for women with a confirmed diagnosis of miscarriage. Explore management options other than expectant management if:
  • the woman is at increased risk of haemorrhage (for example, she is in the late first trimester) or
  • she has previous adverse and/or traumatic experience associated with pregnancy (for example, stillbirth, miscarriage or antepartum haemorrhage) or
  • she is at increased risk from the effects of haemorrhage (for example, if she has coagulopathies or is unable to have a blood transfusion) or
  • there is evidence of infection.
50Offer medical management to women with a confirmed diagnosis of miscarriage if expectant management is not acceptable to the woman.
51Explain what expectant management involves and that most women will need no further treatment. Also provide women with oral and written information about further treatment options.
52Give all women undergoing expectant management of miscarriage oral and written information about what to expect throughout the process, advice on pain relief and where and when to get help in an emergency.*
53If the resolution of bleeding and pain indicate that the miscarriage has completed during 7–14 days of expectant management, advise the woman to take a urine pregnancy test after 3 weeks, and to return for individualised care if it is positive.
54Offer a repeat scan if after the period of expectant management the bleeding and pain:
  • have not started (suggesting that the process of miscarriage has not begun) or
  • are persisting and/or increasing (suggesting incomplete miscarriage).
Discuss all treatment options (continued expectant management, medical management, and surgical management) with the woman to allow her to make an informed choice.
55Review the condition of a woman who opts for continued expectant management of miscarriage at a minimum of 14 days after the first follow-up appointment.
*

See also recommendation 3 for details of further information that should be provided.

NumberResearch recommendation
RR 6In women with confirmed miscarriage, does the type of management strategy (expectant, medical and surgical) impact on women's experience, including psychological and emotional outcomes?
Why this is important
The management of miscarriage in the UK has changed in many ways over the past 2 decades, particularly in the shift from inpatient to outpatient or day case care and the introduction of medical and expectant management as alternatives to surgery.

Despite these changes there is a lack of research into the effects of these different approaches from the woman's perspective, in particular their psychological and emotional impact. Miscarriage is distressing for most women, and the type of management itself might affect women's need for counselling, with a resulting cost to the NHS. Because of this it is an important area for research.

The deficiency in the literature could be addressed by a comparative study of women having the different management strategies (expectant, medical or surgical) and in a variety of clinical settings (for example, early pregnancy assessment unit, gynaecological ward or gynaecological emergency unit). The data collected could be both quantitative (using validated psychological health questionnaires) and qualitative (focusing particularly on women's experience of the particular type and setting of care).

7.5. Misoprostol and mifepristone for managing miscarriage

Review question

What is the most appropriate dose of misoprostol and mifepristone to provide for managing miscarriage?

Introduction

Medical management of miscarriage has been offered to women suffering miscarriage for a number of years with varying doses, timing and routes of administration of drugs being used. The reviews carried out for the purposes of this guideline aimed to ascertain the most appropriate and efficacious dose. The GDG considered the data for two groups of women – those having a missed miscarriage and those having an incomplete miscarriage – as some of the clinicians on the GDG reported a difference in treatment outcomes for these groups of women.

Description of included studies

Twenty-one studies were included in this review (Ayudhaya et al., 2006; Bagratee et al., 2004; Blanchard et al., 2004; Blohm et al., 2005; Creinin et al., 1997; Kovavisarach & Jamnansiri, 2005; Kovavisarach & Sathapanachai, 2002; Kushwah & Singh, 2009; Lelaidier et al., 1993; Lister et al., 2005; Ngoc et al., 2004; Ngoc et al., 2005; Pang et al., 2001; Paritakul & Phupong, 2011; Rita et al., 2006; Shah et al., 2010; Stockheim et al., 2006; Tang et al., 2003; Tang et al., 2006; Tanha et al., 2010; Wood & Brain, 2002).

All of the included studies were randomised controlled trials and were conducted in the UK (Bagratee et al., 2004), USA (Creinin et al., 1997; Lister et al., 2005), Canada (Wood & Brain, 2002), France (Lelaidier et al., 1993), Sweden (Blohm et al., 2005), Israel (Stockheim et al., 2006), Hong Kong (Pang et al., 2001; Tang et al., 2003; Tang et al., 2006), Thailand (Ayudhaya et al., 2006; Kovavisarach & Jamnansiri, 2005; Kovavisarach & Sathapanachai, 2002; Paritakul & Phupong, 2011), Vietnam (Blanchard et al., 2004; Ngoc et al., 2004; Ngoc et al., 2005), India (Kushwah & Singh, 2009; Rita et al., 2006), Pakistan (Shah et al., 2010) and Iran (Tanha et al., 2010).

The GDG decided that, for this review question, studies should only be included if they treated women with incomplete miscarriages and women with missed miscarriages as separate populations. Four studies only included women with incomplete miscarriages (Blanchard et al., 2004; Ngoc et al., 2005; Pang et al., 2001; Paritakul & Phupong, 2011) and two studies included women with both incomplete miscarriages and missed miscarriage, but reported at least one outcome separately for the two populations (Bagratee et al., 2004; Blohm et al., 2005). The remainder of the studies only included women with missed miscarriage.

All of the included studies evaluated the use of misoprostol and/or mifepristone for the management of first trimester miscarriage. One study compared the efficacy of misoprostol alone with a combined regimen of mifepristone and misoprostol (Stockheim et al., 2006). Four studies compared different dosages of misoprostol using the same route of administration, of which one evaluated vaginal misoprostol (Kovavisarach & Jamnansiri, 2005), one evaluated sublingual misoprostol (Tang et al., 2006) and two evaluated oral misoprostol (Blanchard et al., 2004; Ngoc et al., 2005). Nine studies compared misoprostol administered via different routes, of which three compared oral and sublingual administration (Ayudhaya et al., 2006; Kushwah & Singh, 2009; Paritakul & Phupong, 2011), three compared sublingual and vaginal administration (Shah et al., 2010; Tang et al., 2003; Tanha et al., 2010), and four compared oral and vaginal administration (Creinin et al., 1997; Ngoc et al., 2004; Pang et al., 2001; Rita et al., 2006). Six trials were placebo controlled, of which five evaluated misoprostol (Bagratee et al., 2004; Blohm et al., 2005; Kovavisarach & Sathapanachai, 2002; Lister et al., 2005; Wood & Brain, 2002) and one evaluated mifepristone (Lelaidier et al., 1993).

Evidence profile

The treatment regimens described in the GRADE tables and evidence statements detail the maximum number of doses that women could receive; some women did not receive repeat doses if expulsion had started. If multiple doses of misoprostol were given during the same visit, the interval between doses was between 3 and 6 hours, except in one trial whose participants also received supplemental doses after 12 hours (Kushwah & Singh, 2009). In four studies a repeat dose was given if expulsion had not occurred after 24 hours (Bagratee et al., 2004; Creinin et al., 1997; Lister et al., 2005; Wood & Brain, 2002) and in one further study the treatment regimen consisted of two doses administered 48 hours apart (Stockheim et al., 2006). For further details of study regimens please see evidence tables in Appendix H.

Table 7.6. GRADE summary of findings for comparison of vaginal misoprostol with placebo for the management of missed miscarriage.

Table 7.6

GRADE summary of findings for comparison of vaginal misoprostol with placebo for the management of missed miscarriage.

Table 7.7. GRADE summary of findings for comparison of mifepristone with placebo for the management of missed miscarriage.

Table 7.7

GRADE summary of findings for comparison of mifepristone with placebo for the management of missed miscarriage.

Table 7.8. GRADE summary of findings for comparison of mifepristone plus misoprostol with misoprostol only for the management of missed miscarriage.

Table 7.8

GRADE summary of findings for comparison of mifepristone plus misoprostol with misoprostol only for the management of missed miscarriage.

Table 7.9. GRADE summary of findings for comparison of vaginal misoprostol in different dosages for the management of missed miscarriage.

Table 7.9

GRADE summary of findings for comparison of vaginal misoprostol in different dosages for the management of missed miscarriage.

Table 7.10. GRADE summary of findings for comparison of sublingual misoprostol in different dosages for the management of missed miscarriage.

Table 7.10

GRADE summary of findings for comparison of sublingual misoprostol in different dosages for the management of missed miscarriage.

Table 7.11. GRADE summary of findings for comparison of oral misoprostol with sublingual misoprostol for the management of missed miscarriage.

Table 7.11

GRADE summary of findings for comparison of oral misoprostol with sublingual misoprostol for the management of missed miscarriage.

Table 7.12. GRADE summary of findings for comparison of sublingual misoprostol with vaginal misoprostol for the management of missed miscarriage.

Table 7.12

GRADE summary of findings for comparison of sublingual misoprostol with vaginal misoprostol for the management of missed miscarriage.

Table 7.13. GRADE summary of findings for comparison of oral misoprostol with vaginal misoprostol for the management of missed miscarriage.

Table 7.13

GRADE summary of findings for comparison of oral misoprostol with vaginal misoprostol for the management of missed miscarriage.

Table 7.14. GRADE summary of findings for comparison of vaginal misoprostol with placebo for the management of incomplete miscarriage.

Table 7.14

GRADE summary of findings for comparison of vaginal misoprostol with placebo for the management of incomplete miscarriage.

Table 7.15. GRADE summary of findings for comparison of oral misoprostol in different dosages for the management of incomplete miscarriage.

Table 7.15

GRADE summary of findings for comparison of oral misoprostol in different dosages for the management of incomplete miscarriage.

Table 7.16. GRADE summary of findings for comparison of oral misoprostol with vaginal misoprostol for the management of incomplete miscarriage.

Table 7.16

GRADE summary of findings for comparison of oral misoprostol with vaginal misoprostol for the management of incomplete miscarriage.

Table 7.17. GRADE summary of findings for comparison of oral misoprostol with sublingual misoprostol for the management of incomplete miscarriage.

Table 7.17

GRADE summary of findings for comparison of oral misoprostol with sublingual misoprostol for the management of incomplete miscarriage.

Evidence statements

Management of missed miscarriage: comparison of vaginal misoprostol and placebo

Success of medical treatment

One study found that the success of medical treatment was higher in women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. A meta-analysis of two studies also found that the success of medical treatment was higher in women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. These findings were statistically significant and the evidence for these findings was of high quality. However, another meta-analysis of two studies did not find a statistically significant difference in this outcome between women who received vaginal misoprostol (400 micrograms, one dose) and women who received a placebo. The evidence for this finding was of very low quality.

Need for further intervention

One study found that the need for further intervention was lower in women who received vaginal misoprostol (400 micrograms, 1 dose) compared with women who received a placebo. A second study found that the need for further intervention was lower in women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. A meta-analysis of two studies also found that the need for further intervention was lower in women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. These findings were statistically significant and the evidence for these outcomes was of high quality.

Unplanned visits to a medical facility

One study did not find a statistically significant difference in the unplanned visits to a medical facility for women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of moderate quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea and/or vomiting for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in the incidence of nausea for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. One further study did not find a statistically significant difference in the incidence of nausea for women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. The quality of evidence for this outcome was low in one study and moderate in the other.

One study did not find a statistically significant difference in the severity of nausea for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. The evidence for this outcome was of low quality.

One study did not find a statistically significant difference in the incidence of vomiting for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. One further study did not find a statistically significant difference in the incidence of nausea for women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. The quality of evidence for this outcome was low in one study and moderate in the other.

One study did not find a statistically significant difference in the severity of vomiting for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. The evidence for this outcome was of low quality.

One study did not find a statistically significant difference in the incidence of diarrhoea for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. Another study did not find a statistically significant difference in the incidence of diarrhoea for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. One further study did not find a statistically significant difference in the incidence of nausea for women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. The quality of evidence for this outcome was very low in one study, low in the second study and moderate in the third.

One study did not find a statistically significant difference in the severity of diarrhoea for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. The evidence for this outcome was of low quality.

One study reported the incidence of any gastrointestinal side-effects in a manner that did not permit statistical comparison between the two arms. The evidence for this outcome was of low quality.

One study did not find a statistically significant difference in the incidence of fever for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in the incidence of infection for women who received vaginal misoprostol (400 micrograms) compared with women who received a placebo. The evidence for this outcome was of low quality.

One study did not find a statistically significant difference in the incidence of pelvic inflammatory disease for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of low quality.

Duration of bleeding

One study did not find a statistically significant difference in the duration of bleeding for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of moderate quality.

Pain

One study did not find a statistically significant difference in the incidence of menstrual cramping for women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of moderate quality.

One study found that the incidence of lower abdominal pain was higher in women who received vaginal misoprostol (400 micrograms, one dose) compared with women who received a placebo. This finding was statistically significant. The evidence for this outcome was of low quality.

One study found that the severity of pain was higher in women who received vaginal misoprostol (400 micrograms, one dose) compared with women who received a placebo. This finding was statistically significant and the evidence for this finding was of low quality. However, another study did not find a statistically significant difference in this outcome between women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) and women who received a placebo. A further study did not find a statistically significant difference in this outcome between women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) and women who received a placebo. The evidence for this finding was of moderate quality.

Satisfaction

One study did not find a statistically significant difference in the incidence of satisfaction for women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the satisfaction score for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of moderate quality.

Management of missed miscarriage: comparison of mifepristone and placebo

Success of medical treatment

One study found that the success of medical treatment was higher in women who received mifepristone compared with women who received a placebo. This finding was statistically significant and the evidence for this outcome was of low quality.

Need for further intervention

One study found that the need for further intervention was lower in women who received mifepristone compared with women who received a placebo. This finding was statistically significant and the evidence for this outcome was of low quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of endometritis for women who received mifepristone compared with women who received a placebo. The evidence for this outcome was of low quality.

Pain

One study did not find a statistically significant difference in the incidence of pain for women who received mifepristone compared with women who received a placebo. The evidence for this outcome was of low quality.

Management of missed miscarriage: comparison of a combined regimen of mifepristone plus misoprostol and misoprostol only

Success of medical treatment

One study did not find a statistically significant difference in the success of medical treatment for women who received mifepristone plus misoprostol compared with women who received misoprostol only. The evidence for this outcome was of moderate quality.

Need for further intervention

One study did not find a statistically significant difference in the need for further intervention for women who received mifepristone plus misoprostol compared with women who received misoprostol only. The evidence for this outcome was of moderate quality.

Management of missed miscarriage: comparison of vaginal misoprostol dosages

Success of medical treatment

One study found that the success of medical treatment was higher in women who received 800 micrograms of vaginal misoprostol compared with women who received 600 micrograms of vaginal misoprostol. This finding was statistically significant. The evidence for this outcome was of high quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea for women who received 800 micrograms of vaginal misoprostol compared with women who received 600 micrograms of vaginal misoprostol. The evidence for this outcome was of moderate quality.

There were no events in either arm for the outcome of incidence of vomiting. The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the incidence of diarrhoea for women who received 800 micrograms of vaginal misoprostol compared with women who received 600 micrograms of vaginal misoprostol. The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the incidence of fever for women who received 800 micrograms of vaginal misoprostol compared with women who received 600 micrograms of vaginal misoprostol. The evidence for this outcome was of moderate quality.

Pain

One study found that the incidence of pain was higher in women who received 800 micrograms of vaginal misoprostol compared with women who received 600 micrograms of vaginal misoprostol. This finding was statistically significant. The evidence for this outcome was of high quality.

Management of missed miscarriage: comparison of sublingual misoprostol dosages

Success of medical treatment

One study did not find a statistically significant difference in the success of medical treatment for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The evidence for this outcome was of moderate quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea on day 1 or days 2 to 9 for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The evidence for these outcomes was of low quality.

One study did not find a statistically significant difference in the incidence of vomiting on day 1 or days 2 to 9 for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The evidence for these outcomes was of low quality.

One study did not find a statistically significant difference in the incidence of diarrhoea on day 1 for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The same study found that the incidence of diarrhoea on days 2 to 9 after treatment was higher in women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9) compared with women who received a single dose of sublingual misoprostol (600 micrograms). The evidence for these outcomes was of moderate quality.

One study did not find a statistically significant difference in the incidence of fever on day 1 for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The same study reported no events in either arm for the incidence of fever on days 2 to 9. The evidence for these outcomes was of low quality.

One study did not find a statistically significant difference in the incidence of chills and rigour on day 1 for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The same study reported no events in either arm for the incidence of chills and rigour on days 2 to 9. The evidence for these outcomes was of low quality.

Duration of bleeding

One study did not find a statistically significant difference in the duration of bleeding for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The evidence for this outcome was of moderate quality.

Pain

One study did not find a statistically significant difference in the incidence of pain on day 1 or days 2 to 9 for women who received a single dose of sublingual misoprostol (600 micrograms) compared with women who received an extended course of sublingual misoprostol (600 micrograms on day 1 and then 400 micrograms daily on days 2 to 9). The evidence for these outcomes was of moderate quality.

Management of missed miscarriage: comparison of oral and sublingual misoprostol

Success of medical treatment

One study did not find a statistically significant difference in the success of medical treatment for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (400 micrograms, six doses). Another study did not find a statistically significant difference in the success of medical treatment for women who received oral misoprostol compared with women who receive the same dose of sublingual misoprostol (200 mg of mifepristone plus 600 micrograms of misoprostol, with three supplemental doses of 400 micrograms of misoprostol after 12 hours). The quality of evidence for this outcome was low quality in one study and moderate in the other.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea or vomiting for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (400 micrograms, six doses). The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the incidence of nausea for women who received oral misoprostol compared with women who receive the same dose of sublingual misoprostol (200 mg of mifepristone plus 600 micrograms of misoprostol, with three supplemental doses of 400 micrograms of misoprostol after 12 hours). The evidence for this outcome was of low quality.

One study found that the incidence of vomiting was higher in women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (200 mg of mifepristone plus 600 micrograms of misoprostol, with three supplemental doses of 400 micrograms of misoprostol after 12 hours). This finding was statistically significant. The evidence for this outcome was of moderate quality.

One study found that the incidence of fever was higher in women who received sublingual misoprostol compared with women who received the same dose of oral misoprostol (400 micrograms, six doses). This finding was statistically significant and the evidence for this finding was of high quality. However, another study found that the incidence of fever was higher in women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (200 mg of mifepristone plus 600 micrograms of misoprostol, with three supplemental doses of 400 micrograms of misoprostol after 12 hours). This finding was statistically significant and the evidence for this finding was of moderate quality.

One study did not find a statistically significant difference in the incidence of chills for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (400 micrograms, six doses). The evidence for this outcome was of moderate quality.

Pain

One study found that the incidence of pain was higher in women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (regimen for both study groups: 200 mg of mifepristone plus 600 micrograms of misoprostol, with three supplemental doses of 400 micrograms of misoprostol after 12 hours). This finding was statistically significant and the evidence for this finding was of moderate quality. However, another study found that there was no statistically significant difference in the incidence of pain in women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (400 micrograms, six doses). The evidence for this finding was of moderate quality.

Satisfaction

One study found that the reported incidence of satisfaction was higher in women who received sublingual misoprostol compared with women who received the same dose of oral misoprostol (regimen for both study groups: 200 mg of mifepristone plus 600 micrograms of misoprostol, with three supplemental doses of 400 micrograms of misoprostol after 12 hours). This finding was statistically significant and the evidence for this finding was of moderate quality.

Management of missed miscarriage: comparison of sublingual and vaginal misoprostol

Success of medical treatment

One meta-analysis of two studies did not find a statistically significant difference in the success of medical treatment for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, five doses in one study and maximum not reported in the other). A second study did not find a statistically significant difference in the success of medical treatment for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The quality of evidence for this finding was very low in the meta-analysis and moderate in the other study.

Need for further intervention

One meta-analysis of two studies did not find a statistically significant difference in the need for further intervention for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, five doses in one study and maximum not reported in the other). A second study did not find a statistically significant difference in the need for further intervention for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The quality of evidence for this finding was very low in the meta-analysis and low in the other study.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, five doses). A second study did not find a statistically significant difference in the incidence of nausea for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The quality of evidence for this finding was low in one study and moderate in the other.

One study did not find a statistically significant difference in the incidence of vomiting for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, maximum number of doses not reported). A second study did not find a statistically significant difference in the incidence of vomiting for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The quality of evidence for this finding was low in one study and moderate in the other.

One study found that the incidence of diarrhoea was higher in women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, maximum number of doses not reported). A second study found that the incidence of diarrhoea was higher in women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). These findings were statistically significant. The quality of evidence for this finding was moderate in one study and high in the other.

One study found that the incidence of fever was higher in women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, maximum number of doses not reported). This finding was statistically significant and the evidence for this finding was of moderate quality. However, another study did not find a statistically significant difference in this outcome between women who received sublingual misoprostol and women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The evidence for this finding was of moderate quality.

One study did not find a statistically significant difference in the incidence of chills or shivering for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, five doses). A second study did not find a statistically significant difference in the incidence of chills or shivering for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The quality of evidence for this finding was low in one study and moderate in the other.

Duration of bleeding

One study did not find a statistically significant difference in the duration of bleeding for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The evidence for this outcome was of high quality.

Pain

One study found that the incidence of cramps was higher in women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, maximum number of doses not reported). This finding was statistically significant and the evidence for this finding was of moderate quality.

One study found that the incidence of severe pain was higher in women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, maximum number of doses not reported). This finding was statistically significant and the evidence for this finding was of moderate quality.

One study reported 100% incidence of lower abdominal pain in women who received sublingual misoprostol and women who received the same dose of vaginal misoprostol (600 micrograms, three doses). The evidence for this outcome was of high quality.

Satisfaction

One meta-analysis of two studies did not find a statistically significant difference in the reported incidence of satisfaction for women who received sublingual misoprostol compared with women who received the same dose of vaginal misoprostol (400 micrograms, five doses in one study and maximum not reported in the other). The evidence for this outcome was of very low quality.

Management of missed miscarriage: comparison of oral and vaginal misoprostol

Success of medical treatment

One study found that the success of medical treatment was higher in women who received vaginal misoprostol (600 micrograms, two doses) compared with women who received oral misoprostol (400 micrograms, three doses). A second study found that the success of medical treatment was higher in women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) compared with women who received oral misoprostol (400 micrograms, repeat after 24 hours). These findings were statistically significant. The quality of evidence for this finding was moderate in one study and very low in the other. One further study did not find a statistically significant difference in this outcome between women who received the same dose of vaginal and oral misoprostol (800 micrograms, one dose). The evidence for this finding was of moderate quality.

Need for further intervention

One study found that the need for a further intervention was higher in women who received oral misoprostol (400 micrograms, three doses) compared with women who received vaginal misoprostol (600 micrograms, two doses). This finding was statistically significant and the evidence for this finding was of moderate quality. One further study did not find a statistically significant difference in this outcome between women who received the same dose of vaginal and oral misoprostol (800 micrograms, one dose). The evidence for this finding was of low quality.

Admission to a medical facility

One study did not find a statistically significant difference in the incidence of admission to a medical facility for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, one dose). The evidence for this outcome was of very low quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea for women who received oral misoprostol (400 micrograms, three doses) compared with women who received vaginal misoprostol (600 micrograms, two doses). A second study did not find a statistically significant difference in the incidence of nausea for women who received oral misoprostol (400 micrograms, one dose) compared with women who received vaginal misoprostol (800 micrograms, one dose). The quality of evidence for this outcome was low in one study and very low in the other.

One study found that the incidence of vomiting was higher in women who received vaginal misoprostol (800 micrograms, one dose) compared with women who received oral misoprostol (800 micrograms, one dose). This finding was statistically significant and the evidence for this finding was of high quality. However, another study found that there was no statistically significant difference in this outcome between women who received vaginal misoprostol (600 micrograms, two doses) and oral misoprostol (400 micrograms, three doses). A further study found that there was no statistically significant difference this outcome between women who received vaginal misoprostol (800 micrograms, repeat after 24 hours) and women who received oral misoprostol (400 micrograms, repeat after 24 hours). The quality of evidence for this finding was low in one study and very low in the other.

One study did not find a statistically significant difference in the incidence of diarrhoea for women who received oral misoprostol (400 micrograms, three doses) compared with women who received vaginal misoprostol (600 micrograms, two doses). A second study did not find a statistically significant difference in the incidence of diarrhoea for women who received oral misoprostol (400 micrograms, one dose) compared with women who received vaginal misoprostol (800 micrograms, one dose). Another study did not find a statistically significant difference in the incidence of diarrhoea for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, one dose). The quality of evidence for this outcome was low in one study, very low in the second study and moderate in the third study.

One study did not find a statistically significant difference in the incidence of hyperpyrexia for women who received oral misoprostol (400 micrograms, three doses) compared with women who received vaginal misoprostol (600 micrograms, two doses). The evidence for this outcome was of low quality.

One study did not find a statistically significant difference in the incidence of fever or chills for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, one dose). The evidence for this outcome was of moderate quality.

Duration of bleeding

One study did not find a statistically significant difference in the duration of bleeding for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, one dose). The evidence for this finding was of moderate quality. One further study reported duration of bleeding in a manner that did not permit a comparison between the two arms. The evidence for this finding was of very low quality.

Pain

One study did not find a statistically significant difference in the incidence of pain for women who received oral misoprostol (400 micrograms, three doses) compared with women who received vaginal misoprostol (600 micrograms, two doses). A second study did not find a statistically significant difference in the incidence of pain for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, one dose). The quality of evidence for this outcome was low in one study and high in the other.

One study did not find a statistically significant difference in the severity of pain for women who received oral misoprostol (400 micrograms, one dose) compared with women who received vaginal misoprostol (800 micrograms, one dose). The evidence for this outcome was of very low quality.

Satisfaction

One study did not find a statistically significant difference in the reported incidence of satisfaction for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, one dose). The evidence for this outcome was of high quality.

Management of incomplete miscarriage: comparison of vaginal misoprostol and placebo

Success of medical treatment

One study did not find a statistically significant difference in the success of medical treatment for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of low quality.

Need for further intervention

One study did not find a statistically significant difference in the need for further intervention for women who received vaginal misoprostol (400 micrograms, one dose) compared with women who received a placebo. A second study did not find a statistically significant difference in the need for further intervention for women who received vaginal misoprostol (600 micrograms, repeat after 24 hours) compared with women who received a placebo. The evidence for this outcome was of low quality.

Management of incomplete miscarriage: comparison of oral misoprostol dosages

Success of medical treatment

One meta-analysis of two studies did not find a statistically significant difference in the success of medical treatment for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of moderate quality.

Need for further intervention

One meta-analysis of two studies did not find a statistically significant difference in the need for further intervention for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of low quality.

Adverse effects

One meta-analysis of two studies did not find a statistically significant difference in the incidence of nausea for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of very low quality.

One meta-analysis of two studies did not find a statistically significant difference in the incidence of vomiting for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of low quality.

One study found that the incidence of diarrhoea was higher in women who received two 600 microgram doses of oral misoprostol compared with women who received a single 600 microgram dose of oral misoprostol. This difference was statistically significant. The evidence for this finding was of moderate quality.

One meta-analysis of two studies did not find a statistically significant difference in the incidence of fever or chills for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of low quality.

Duration of bleeding

Two studies did not find a statistically significant difference in the duration of heavy bleeding for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The quality of evidence for this outcome was moderate in one study and low in the other.

Two studies did not find a statistically significant difference in the duration of normal bleeding for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The quality of evidence for this outcome was moderate in one study and low in the other.

Two studies did not find a statistically significant difference in the duration of light bleeding for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The quality of evidence for this outcome was moderate in one study and low in the other.

Pain

One meta-analysis of two studies did not find a statistically significant difference in the incidence of pain for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of moderate quality.

Two studies did not find a statistically significant difference in the severity of pain for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of low quality.

Satisfaction

One meta-analysis of two studies did not find a statistically significant difference in the reported incidence of satisfaction for women who received two 600 microgram doses of oral misoprostol compared with women who received a single dose of 600 microgram oral misoprostol. The evidence for this outcome was of moderate quality.

Management of incomplete miscarriage: comparison of oral and vaginal misoprostol

Success of medical treatment

One study did not find a statistically significant difference in the success of medical treatment for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). The evidence for this outcome was of very low quality.

Need for further intervention

One study did not find a statistically significant difference in the need for further intervention for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). The evidence for this outcome was of very low quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). The evidence for this outcome was of low quality.

One study did not find a statistically significant difference in the incidence of vomiting for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). The evidence for this outcome was of low quality.

One study found that the incidence of diarrhoea was higher in women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). This finding was statistically significant and the evidence for this finding was of moderate quality.

One study did not find a statistically significant difference in the incidence of fever for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). The evidence for this outcome was of low quality.

Duration of bleeding

One study did not find a statistically significant difference in the duration of bleeding for women who received oral misoprostol compared with women who received the same dose of vaginal misoprostol (800 micrograms, two doses). The evidence for this outcome was of moderate quality.

Pain

One study found that the duration of pelvic pain was longer in women who received vaginal misoprostol compared with women who received the same dose of oral misoprostol (800 micrograms, two doses). This finding was statistically significant and the evidence for this finding was of moderate quality.

Management of incomplete miscarriage: comparison of oral and sublingual misoprostol

Success of medical treatment

One study did not find a statistically significant difference in the success of medical treatment for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of low quality.

Need for further intervention

One study did not find a statistically significant difference in the need for further intervention for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of low quality.

Adverse effects

One study did not find a statistically significant difference in the incidence of nausea for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

One study reported no incidences of vomiting in women who received oral misoprostol and women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the incidence of diarrhoea for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the incidence of fever/chills for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

Incidence of heavy bleeding

One study reported no incidences of heavy bleeding in women who received oral misoprostol and women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

Pain

One study did not find a statistically significant difference in the incidence of pain/cramps for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

One study did not find a statistically significant difference in the severity of pain for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of high quality.

Satisfaction

One study did not find a statistically significant difference in the incidence of reported satisfaction for women who received oral misoprostol compared with women who received the same dose of sublingual misoprostol (600 micrograms). The evidence for this outcome was of moderate quality.

Evidence to recommendations

Relative value placed on the outcomes considered

The primary outcome measures for this question were the success rate of the treatment and the need for further interventions. The secondary outcomes were the various side-effects associated with treatment. Of these, the group felt that pain was less informative than the others, as all women would be likely to experience a degree of pain as a natural consequence of the miscarriage. It was also evident from the placebo trials that gastro-intestinal symptoms may accompany miscarriage.

Consideration of clinical benefits and harms

When considering the appropriate dose and mode of administration, the group considered the main priority to be the efficacy of the treatment. The evidence compared a number of different doses and regimens. For women with missed miscarriage, a single dose of 800 micrograms (oral or vaginal) of misoprostol was the most effective overall. For women with an incomplete miscarriage, the evidence suggested that a single dose of 600 micrograms of misoprostol was effective, but the group recognised that units might prefer to use 800 micrograms for alignment of protocols.

For women with missed miscarriage, the group noted that, used at the same dose, both vaginal and oral routes of administration had similar effectiveness, and that both were more effective than sublingual administration. The majority of side-effects did not show a difference by route. There was contradictory evidence regarding diarrhoea: one study showed more diarrhoea associated with a vaginal route of administration than with an oral route, but two further studies showed no statistically significant difference and this matched the GDG members' clinical experience. The evidence suggested that there were more side-effects from sub-lingual administration than with vaginal administration. In addition, the GDG members recognised from their own clinical experience that there can often be difficulties with sub-lingual administration as women are expected to hold up to four relatively large tablets under their tongue for a long period of time.

For women with an incomplete miscarriage, one study found there was a significantly lower incidence of diarrhoea with vaginal administration compared with oral administration. A second study showed no significant differences between oral and sub-lingual administration for all outcomes.

The group felt that an additional benefit of vaginal administration is that if women vomit after receiving medication orally, this can interfere with the absorption of the drug and it can be difficult to determine if the dose should be repeated.

Overall, the GDG agreed that vaginal misoprostol is the preferred treatment. However, it recognised that it is important to take into account women's preferences and thus agreed that the oral route would be appropriate if vaginal administration was not acceptable to the woman.

Consideration of health benefits and resource uses

The group recognised that there was very little evidence which compared the efficacy of misoprostol in combination with mifepristone compared with misoprostol alone for the treatment of miscarriage. However, what evidence there was suggested that there was no difference in effectiveness. Given the large cost of mifepristone compared with misoprostol1, the group agreed that mifepristone should not be used routinely in the management of miscarriage. It was recognised that mifepristone is currently used in UK practice for this indication and that it would therefore be helpful to conduct further research. The GDG agreed that a trial was needed to determine definitively whether the addition of mifepristone improves the success rate of medical management, the results of which could then be used to evaluate whether it is a cost-effective treatment.

Quality of evidence

The evidence for this question was drawn from RCTs; however, the quality was mixed and ranged from high to very low. Generally, the group felt that the evidence was of sufficient quality to make recommendations, although it recognised that it would have been helpful to have had further evidence related to the efficacy of mifepristone and misoprostol in combination.

Information giving and emotional support

The group stressed the importance of providing women with information about the treatment, what to expect as the miscarriage progresses, the potential side-effects of treatment and the next steps if the treatment is ineffective. It felt that informing women about the likely length and extent of bleeding was a particular consideration, as this could be particularly distressing or worrying.

Other considerations

The GDG discussed the fact that the majority of women receiving medical treatment would now be receiving it as second-line treatment, as a result of recommendations about first-line expectant management made in Section 7.4. All of the studies evaluated medical management as first-line treatment. However, the group noted that the women participating in the studies were likely to have presented at different points in the course of their miscarriage, depending on their symptoms and ability to access health care. For example, women with no symptoms might have sought care, and been diagnosed with a missed miscarriage, weeks after the miscarriage had actually occurred, and therefore would be reasonably comparable to women presenting earlier who had then been expectantly managed for a period of time. Despite this, the group accepted that the treatment success rates reported in the trials might be higher than would be expected in women in whom expectant management had already failed, and that this would be a consideration for women choosing their preferred second-line treatment. However, it also felt that the relative efficacy of different routes of administration would be unlikely to vary according to whether the treatment was first or second line, and, therefore, that the results of the comparisons reported in the trials could safely be extrapolated to women actively choosing medical management after a period of expectant management.

The GDG recognised that there are a number of side-effects associated with miscarriage itself and with the drugs used in its management, and that clinicians should treat these side-effects at the same time as providing miscarriage management.

The group agreed that, for women with a missed miscarriage, if the treatment is ineffective (that is, bleeding has not started) after 24 hours, there should be a clinical review, which could just be a telephone conversation rather than a formal appointment. The group also recognised the importance of having a follow-up process in place to ensure that there is no molar or ectopic pregnancy. The group agreed that all women should be advised to check that a urine pregnancy test is negative 3 weeks after receiving medical treatment to determine if it has been effective, and that they should return for clinical review if the test is still positive.

While misoprostol is commonly used to treat miscarriage, it is not currently licensed for use for this indication and so women's consent should be obtained before it is used.

Recommendations

NumberRecommendation
56Do not offer mifepristone as a treatment for missed or incomplete miscarriage.
57Offer vaginal misoprostol for the medical treatment of missed or incomplete miscarriage. Oral administration is an acceptable alternative if this is the woman's preference.*
58For women with a missed miscarriage, use a single dose of 800 micrograms of misoprostol.*
59Advise the woman that if bleeding has not started 24 hours after treatment, she should contact her healthcare professional to determine ongoing individualised care.
60For women with an incomplete miscarriage, use a single dose of 600 micrograms of misoprostol. (800 micrograms can be used as an alternative to allow alignment of treatment protocols for both missed and incomplete miscarriage.)*
61Offer all women receiving medical management of miscarriage pain relief and anti-emetics as needed.
62Inform women undergoing medical management of miscarriage about what to expect throughout the process, including the length and extent of bleeding and the potential side effects of treatment including pain, diarrhoea and vomiting.
63Advise women to take a urine pregnancy test 3 weeks after medical management of miscarriage unless they experience worsening symptoms, in which case advise them to return to the healthcare professional responsible for providing their medical management.
64Advise women with a positive urine pregnancy test after 3 weeks to return for a review by a healthcare professional to ensure that there is no molar or ectopic pregnancy.
*

Although this use is common in UK clinical practice, at the time of publication (December 2012), misoprostol did not have a UK marketing authorisation for this indication. The prescriber should follow relevant professional guidance, taking full responsibility for the decision. Informed consent should be obtained and documented. See the General Medical Council's Good practice in prescribing medicines – guidance for doctors for further information.

NumberResearch recommendation
RR 7Is the combination of mifepristone and misoprostol more effective than misoprostol alone in the medical management of miscarriage?

7.6. Setting for surgical management of miscarriage

Review question

What is the effectiveness of surgical management of miscarriage in an outpatient (office) setting compared with any other setting for improving women's clinical and psychological outcomes?

Introduction

Historically women diagnosed with a miscarriage who opted for surgical management would undergo the procedure in a traditional theatre setting under a general anaesthetic. Recent advances in surgical techniques and equipment have seen the introduction of surgical procedures being performed without a general anaesthetic (but with some other form of anaesthesia, analgesia or sedation) in an outpatient setting. This review sought to identify which settings are associated with better outcomes.

Description of included studies

Four studies were included in this review. Three studies were conducted in the USA (Blumenthal et al., 1994; Dalton et al., 2006; Edwards et al., 2007) and one in South Africa (De Jonge et al., 1994).

One retrospective observational study (Edwards et al., 2007) assessed and compared the efficiency, post procedure quality of life and acceptability of manual vacuum aspiration (MVA) performed in an outpatient setting with electric vacuum aspiration (EVA) performed in a hospital in-patient setting in women experiencing a first-trimester miscarriage. EVA was performed under either general anaesthesia, ‘monitored anaesthesia care’ (MAC) or spinal anaesthesia. Analgesia for MVA was provided with a paracervical block of lidocaine.

One randomised clinical trial (De Jonge et al., 1994) assessed evacuation under systemic analgesia (fentanyl and midazolam) in a treatment room compared with evacuation under general anaesthesia in an operating theatre.

One prospective observational study (Dalton et al., 2006) examined women's satisfaction with an office-based surgical procedure for early pregnancy failure and compared the resource use and cost between office and operating room management. Anaesthesia for the MVA performed in an office setting consisted of oral lorazepam, ibuprofen and/or propoxyphene napsylate, with paracervical block. EVA was performed under anaesthesia that included intravenous sedation, regional anaesthesia or general anaesthesia.

One quasi experimental study (Blumenthal et al., 1994) also examined the cost effectiveness of performing manual vacuum aspiration curettage (MVAC) either in an emergency room or in a labour room as an alternative to the traditional suction curettage (SC) in the operating room. For the SC procedure, sedation was achieved with a combination of short-acting benzodiazepines and narcotics and MVAC was performed under systemic analgesia (fentanyl and midazolam). None of the women had general anaesthesia in this study.

Evidence profile

The evidence from all four studies is presented in one profile.

Table 7.18. GRADE summary of findings for comparison of operating room with office setting for surgical management of miscarriage.

Table 7.18

GRADE summary of findings for comparison of operating room with office setting for surgical management of miscarriage.

Evidence statements

Evidence was identified from four studies that reported efficiency of MVA performed as an outpatient compared with EVA performed in an operating room in women experiencing a first-trimester miscarriage.

One study found lower pain severity within 48 hours of treatment in women following EVA in an operating room compared to MVA in an office setting. This finding was statistically significant and the evidence for this outcome was of very low quality.

Two studies study found longer waiting times from admission to procedure in women undergoing EVA in an operating room compared with MVA in office setting. This finding was statistically significant and the evidence for this outcome was of moderate and very low quality.

One study found a higher proportion of women receiving a blood transfusion following EVA in an operating room compared with MVA in an office setting. This finding was statistically significant and the evidence for this outcome was of low quality.

One study found higher mean blood loss in women following EVA in an operating room compared with MVA in an office setting. This finding was statistically significant and the evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in the proportion of women who presented to an emergency department on the same day of treatment following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in the proportion of women who reported passing tissue within 48 hours of treatment following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in success rate 30 days after treatment in women following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in the proportion of women who developed fever after treatment following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in maximum total satisfaction score in women following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference in post procedure infection in women following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

One study did not find a statistically significant difference of need for re-evacuation in women following EVA in an operating room compared with MVA in an office setting. The evidence for this outcome was of very low quality.

Evidence to recommendations

Relative value placed on the outcomes considered

The GDG considered the success rate of the treatment to be the most important outcome for this question. In addition, the group felt that need for an emergency hospital visit was also an important measure as an indicator of the comparative safety of the procedure in the different settings.

Consideration of clinical benefits and harms

The evidence showed that the median waiting time, the number of women requiring a blood transfusion, and the mean blood loss were all lower in an outpatient setting. The group recognised that although women's pain severity scores 48 hours after treatment were significantly lower in an inpatient setting, the average scores were relatively low for both groups. The group noted that the rate of blood transfusions in the study which reported this outcome was particularly high in both arms and that this evidence was unlikely to be applicable to a UK setting. In addition, the group noted that the study which reported the mean blood loss outcome had more than twice as many women in one arm as in the other and that this was likely to have affected the results.

Consideration of health benefits and resource uses

It was recognised that there would be an initial cost for units to set up the facility for performing manual vacuum aspiration as an outpatient procedure. However, once these costs have been met, it is likely that an outpatient setting would be cost effective, given the reduced time for conducting the procedure.

Quality of evidence

The evidence was generally of very low quality and so the group did not feel able to make a strong recommendation that all surgical management should routinely be conducted as an outpatient procedure. However, the group did feel that the evidence about reduced waiting times justified a recommendation that units should be able to offer surgical management as an outpatient procedure in order to provide women with a choice.

Information giving and emotional support

The GDG felt that it was important that women were given appropriate information about the different treatment options and what to expect during the procedure, in order that they could make an informed choice about their treatment. In addition, it agreed that women should be provided with information about what to expect during the recovery period.

Other considerations

The group recognised that some women will prefer to have the procedure conducted under general rather than local anaesthetic (that is, in a theatre setting). In addition, it recognised that at later gestations it may not always be clinically appropriate to offer the procedure without a general anaesthetic. As a result, the GDG did not feel it appropriate to recommend that all surgical procedures be conducted as outpatient procedures.

Recommendations

NumberRecommendation
65Where clinically appropriate, offer women undergoing a miscarriage a choice of:
  • manual vacuum aspiration under local anaesthetic in an outpatient or clinic setting or
  • surgical management in a theatre under general anaesthetic.
66Provide oral and written information to all women undergoing surgical management of miscarriage about the treatment options available and what to expect during and after the procedure.*
*

See also recommendation 3 for details of further information that should be provided.

Footnotes

1

A pack of three 200 mg mifepristone tablets costs approximately £50, whilst a pack of sixty 200 microgram misoprostol tablets costs approximately £10.

Copyright © 2012, National Collaborating Centre for Women's and Children's Health.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

Bookshelf ID: NBK132781

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (3.1M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...