Mifepristone for uterine fibroids

Abstract

Background

Uterine fibroids are the most common benign uterine tumours present in women of reproductive age. Mifepristone (RU‐486) competitively binds and inhibitsprogesterone receptors. Studies have suggested that fibroid growth depends on the sexual steroids. Mifepristone has been shown to decrease fibroid size. This review summarises the effects of mifepristone treatment on fibroids and the associated adverse effects as described in randomised controlled trials.

Objectives

To determine the efficacy and safety of mifepristone for the management of uterine fibroids in pre‐menopausal women.

Search methods

We searched the specialised register of the Cochrane Menstrual Disorders and Subfertility (Cochrane Menstrual Disorders and subfertility Review Group), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2011, Issue 4), MEDLINE, EMBASE, PsycINFO, and CINAHL (to November 2011). We handsearched a number of journals, and searched reference lists, databases of ongoing trials and the Internet. There were no language restrictions.

Selection criteria

Only truly randomised controlled trials of mifepristone versus other forms of medical therapy or placebo in pre‐menopausal women with confirmed uterine fibroids were included.

Data collection and analysis

Four authors independently extracted data and assessed trial quality. Data were analysed using the Peto odds ratios (OR) for dichotomous data and the weighted mean differences for continuous data, with 95% confidence intervals (CI). Meta‐analyses were performed using the fixed‐effect model.

Main results

Three studies involving 112 participants were included. Comparison interventions included different dosages of mifepristone, placebo and vitamin B tablets. There is evidence that treatment with mifepristone relieves heavy menstrual bleeding compared with placebo (Peto OR 17.84; 95% CI 6.72 to 47.38; 2 RCTs, 77 women, I² = 0%). Three studies (Bagaria 2009; Engman 2009; Fiscella 2006) were included in the meta‐analysis of this comparison. There was no evidence of an effect of mifepristone on the fibroid volume (standardised mean difference (SMD) ‐0.02; 95% CI ‐0.38 to 0.41; 99 women). Two studies (Bagaria 2009; Fiscella 2006) were included in the meta‐analysis of this comparison. There was no evidence of an effect of mifepristone on uterine volume (mean difference (MD) ‐77.24; 95% CI ‐240.62 to 86.14; 72 women). The pooled data suggest an increased adverse event (abnormal endometrial histology) in the mifepristone group compared to placebo (OR 31.65; 95% CI 4.83 to 207.35; 2 RCTs; 54 women; I² = 0%). Only one study (Bagaria 2009) reported endometrial hyperplasia at the end of the therapy (12/19 women in the mifepristone group versus 0/16 in the placebo group; OR 55.0; 95% CI 2.86 to 105.67). Engman 2009 found a significantly higher rate of cystic glandular dilatation in women in the mifepristone group (5/8 women biopsied) compared with the placebo group (1/11 women biopsied) (OR 16.67; 95% CI 1.36 to 204.03). One study (Fiscella 2006) suggested significant improvements (P < 0.001) for specific quality of life outcomes.

Authors' conclusions

Mifepristone reduced heavy menstrual bleeding and improved fibroid‐specific quality of life. However, it was not found to reduce fibroid volume. Further well‐designed, adequately powered RCTs are needed before a recommendation can be made on the use of mifepristone for the treatment of uterine fibroids.

Plain language summary

Mifepristone for uterine fibroids

Uterine fibroids are also known as uterine leiomyoma,myoma or  fibromyoma and are non‐cancerous benign growths in the uterus. Fibroids are the most common benign tumours in females and are typically found during the middle and later reproductive years. Common symptoms include heavy bleeding, menstrual pain, pressure in lower abdomen, infertility or miscarriage. Fibroids can be treated with surgery using either myomectomy (removal of fibroids leaving the uterus in place) or hysterectomy (removal of uterus). Drugs such as mifepristone have been suggested as a therapeutic option. This review includes three trials and 112 women with uterine fibroids under mifepristone treatment. These clinical trials included a small number of participants and show limited methodological quality. The studies included in this review show that mifepristone had a moderate effect in relief of bleeding and showed an improvement in fibroid‐specific quality of life. Determination of the effects of mifepristone on uterine fibroid volume requires much larger trials to draw a confident conclusion for mifepristone in clinical use.

Background

Description of the condition

Uterine fibroids, also known as uterine leiomyomas or myomas, are almost always benign growths of smooth muscle layer of the uterus (Meniru 2001). Fibroids may grow as a single tumour or in clusters. A single fibroid can range in size from one to 20 cm or more (Marshall 1997).

Fibroids are the most common benign uterine tumours present in women of reproductive age. The incidence varies from 5.4% to 77% in women of reproductive age, depending on the method of diagnosis (Drinville 2007; Lethaby 2002).

The aetiology of uterine fibroids is unknown. However they are thought to be oestrogen and progesterone dependant because fibroids are known to shrink after either menopause or treatment with gonadotrophin‐releasing hormone agonists (GnRH) (which reduce oestrogen and progesterone levels) (Rein 1995; West 1993).

Most women with fibroids are asymptomatic. Notwithstanding this, women may complain of a lump or "pelvic fullness", caused by the size of the fibroids. The most common symptom is abnormal uterine bleeding, usually long and heavy menstrual periods. Fibroids may interfere with fertility because they can cause distortion of the uterus. Apart from the mass effect, the primary mechanism by which fibroids may cause subfertility is unknown (Hart 2001).

The typeof treatment approach should consider medical and social factors, age, parity,childbearing expectancies, extent and severity of symptoms, sizeand number of myomas, location of myomas, associated medicalconditions, proximity to menopauseand desire for uterine preservation (Wallach 2004).

Hysterectomy is the most common treatment when fibroids are symptomatic. Based on data from 1990 to 1997, the presence of uterine fibroids was the main indication for hysterectomy in the US (Farquhar 2002). Myomectomy, the removal of fibroids surgically without hysterectomy, is the second most common surgical treatment (Guarnaccia 2001).

The management of uterine myomas may involve the following  approaches (or a combination of them): expectant management, medicalmanagement (GnRH analogues, progestational compounds, anti‐progestins), uterine arteryembolisation, and other approaches (e.g. high‐frequency focused ultrasound,laser treatment, cryotherapy, or thermoablation) (Wallach 2004).

Even although fibroids are often diagnosed and treated, there remains considerable uncertainty and controversy among clinicians and women regarding their best management (Myers 2002).

Description of the intervention

Observational data suggested that treatment with mifepristone is associated with a reduction in uterine and fibroid size, pain, and bleeding.

This could be based on the fact that mifepristone competitively binds and antagonisesprogesterone receptors. A number of short‐term investigations have suggested that mifepristone is effective in inhibiting ovulation, inducing luteolysis and disrupting endometrial integrity when administered to normally cycling women (Garzo 1988; Liu 1987; http://archie.cochrane.org/sections/documents/view?version=DDFAED1B82E26AA200420E8618B064AF&format=REVMAN#REF‐Luukkainen‐1988#REF‐Luukkainen‐1988; Roseff 1990). The long‐term administration of 100 mg/day of mifepristone may induce absence of ovarian cycles and relief of pelvic pain in women with endometriosis (Kettel 1991). Mifepristone is thought to have an inhibitory effect on the growth of fibroids. Transient elevations in transaminases occurred in 4% of women, and endometrial hyperplasia was detected in 10 (28%) of 36 women screened by endometrial biopsy (Steinauer 2004). A spectrum of endometrial pathology induced by progesterone receptor modulators has been reported (Mutter 2008). Currently, mifepristone is approved only for medical abortion.

How the intervention might work

Progestogens alone have been reported to influence fibroid growth. These compounds were thought to produce a hypo‐oestrogenic effectby inhibiting gonadotrophin secretion and suppressing ovarianfunction (Istre 2007; Tiltman 1985; Vikhliaeva 1990).

Mifepristone may represent a viable alternative to GnRH analogues for pre‐operative application. Also, it could have an indication in peri‐menopausal women with large, symptomatic fibroids, who could be able to take this medication until menopause, when the fibroids typically decrease. Consequently, it would yield major savings in cost and morbidity in view of the large number of hysterectomies done for fibroids in peri‐menopausal women (Drinville 2007). Another possible application is in younger women with large fibroids who wish to retain their fertility. They may also benefit from continuous low‐dose mifepristone, until the time that they wish to conceive (Eisinger 2003).

Why it is important to do this review

Studies have suggested that fibroid growth depends on sex hormone steroids (Kawaguchi 1989). Studies have provided further biochemical, histological and clinical evidence that progesterone has a critical role in leiomyoma growth (Kim 2012). Mifepristone, a progesterone receptor modulator with primarily antagonistic properties, has been shown to decrease leiomyoma size in non‐RCTs (Murphy 1995; Yang 1996).

We have systematically reviewed the literature in an attempt to summarise the effects of mifepristone treatment on fibroids and the associated adverse effects as described in RCTs.

Objectives

To determine the efficacy and safety of mifepristone for the management of uterine fibroids in pre‐menopausal women.

Methods

Criteria for considering studies for this review

Types of studies

RCTs were included. We excluded quasi RCTs. We would have included cross‐over studies if they had pre‐cross‐over data available; however, no cross‐over studies were found.

Types of participants

Pre‐menopausal women with confirmed uterine fibroid diagnosed by clinical manifestation and physical signs, and confirmed by ultrasound scanning, computerised tomography (CT), magnetic resonance imaging (MRI), or a combination of at least two of the procedures. The diagnostic criteria produced by The International Federation of Gynecology and Obstetrics (FIGO 2001) or the International Gynecology and Obstetrics Association were used. Women with an intrauterine device (IUD), ectopic pregnancy, adrenal failure, haemorrhagic disorders, inherited porphyria, and anticoagulant or long‐term corticosteroid therapy were excluded.

Types of interventions

• Mifepristone compared to placebo.

• Mifepristone compared to no treatment.

• Mifepristone compared to other medical intervention (e.g. GnRH analogues, progestational compounds, anti‐progestins).

• Mifepristone compared to a surgical intervention (e.g. hysterectomy, myomectomy hysteroscopic or laparoscopic, high‐frequency focused ultrasound,laser treatment, cryotherapy, or thermoablation).

Types of outcome measures

Primary outcomes

• Relief of symptoms: abnormal uterine bleeding, pain and pressure measured by patient self reports or scales.

Secondary outcomes

• Reduction in fibroid size: measured by ultrasonography or MRI.

• Reduction in uterine volume measured by ultrasonography or MRI.

• Live birth.

• Pregnancy rate.

• Recurrence rate with the possibility of necessitating further additional therapy.

• Occurrence of adverse events.

• Cost‐effectiveness.

• Uterine Fibroid Symptoms ‐ Quality of Life.

Search methods for identification of studies

See the Menstrual Disorders and Subfertility Group methods used in reviews as stated in their module.

All reports that describe (or might describe) RCTs of mifepristone for uterine fibroids were sought.

1. The MDSG Specialised Register of controlled trials was searched by the Group's trial search co‐ordinator using the key words:

Keywords CONTAINS "Leiomyoma"or "leiomyomata"or"fibroids"or "myoma"or"myomas"or "myomata"or "Leiomyoma"or "leiomyomata"or "uterine fibroids"or "uterine leiomyomas"or"uterine myoma"or"uterine myomas" or Title CONTAINS "Leiomyoma"or "leiomyomata"or"fibroids"or "myoma"or"myomas"or "myomata"or "Leiomyoma"or "leiomyomata"or "uterine fibroids"or "uterine leiomyomas"or"uterine myoma"or"uterine myomas"

AND

Keywords CONTAINS "mifepristone" or "RU486" or Title CONTAINS "mifepristone" or "RU486"

This register also contains unpublished trial abstracts. These are found through the handsearching of 20 relevant journals and conference proceedings.

External referees who are experts in this field were asked to check the completeness of the search strategy, and to identify any additional, ongoing and planned trials.

No language restrictions were applied.

Electronic searches

We searched electronic databases using key words and the appropriate Cochrane highly sensitive search strategies for identifying randomised trials. The last search was conducted in November 2011 (Higgins 2011). The databases searched were: MEDLINE, EMBASE, CENTRAL and CINAHL. The full search strings for each database are available in the following appendices: Appendix 1; Appendix 2; Appendix 3; Appendix 4.

Searching other resources

Other strategies for locating studies included:

• contacting organisations and individuals working in the field,

• screening conference proceedings and reference list of review articles,

• searching clinical trial registers.

Data collection and analysis

Data was analysed using Review Manager 5.1 software (RevMan 2011).

Selection of studies

We used the criteria for study eligibility as described by the Cochrane MDSG entity (see Appendix 5).

The titles and abstracts of articles found in the search were screened by LO, who discarded clearly ineligible studies. The aim was to be overly inclusive to minimise the risk of losing relevant studies.

MT, AS and AR obtained copies of the full‐text articles and made copies for LO, in which details of the authors and institutions were struck out and the results sections removed.

At least two out of three review authors (LO, AS, MT), including two content experts, screened the studies for inclusion.

Both review authors independently assessed whether the studies met inclusion criteria. No disagreement needed to be resolved by discussion.

Further information was asked of the study authors when the information contained in papers was insufficient to make a decision about eligibility.

Data extraction and management

LO provided MT, AR, PS and AS with the results sections of the included studies and the four review authors independently extracted information using a prepared form (see Appendix 6).

Discrepancies were resolved by discussion.

For each included trial, information was collected regarding the location of the study, methods of the study (as per 'Risk of bias' assessment checklist), participants (age range, eligibility criteria), nature of the interventions and data relating to the outcomes specified above.

Assessment of risk of bias in included studies

The risk of bias of all studies deemed eligible for the review was assessed independently by four review authors (MT, AR, LO and PS), and discrepancies were resolved by discussion. A summary of the 'Risk of bias' assessment was presented as part of the Characteristics of included studies table.

Risk of bias was assessed using six domains (see below) and a judgement of either 'high', 'low' and 'unclear' was made. All judgements were fully described and presented as Figure 1 and Figure 2.

1.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

2.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

1. Sequence generation

Was the allocation sequence adequately generated?

2. Allocation concealment

Was allocation adequately concealed?

3. Blinding of participants, personnel and outcome assessors

Was knowledge of the allocated intervention adequately prevented during the study?

4. Incomplete outcome data

Were incomplete outcome data adequately addressed?

5. Selective outcome reporting

Are reports of the study free of suggestion of selective outcome reporting?

6. Other sources of bias

Was the study apparently free of other problems that could put it at a high risk of bias?

Measures of treatment effect

The data from the primary studies was entered into RevMan 5.1 software (RevMan 2011).

For binary (or dichotomous) outcomes, results for each study was expressed as Peto odds ratios (OR) with 95% confidence intervals (CI) and combined for meta‐analysis where appropriate.

For continuous outcome data, results from each study were expressed as a mean difference (MD) with 95% CI and combined for meta‐analysis using the weighted mean difference (WMD).

Unit of analysis issues

As clusters RCTs were not included, individuals were considered as the unit of analysis.

Dealing with missing data

The review authors contacted the lead authors of the trials where data clarification was required. This contact was made by email, telephone, or both.

Assessment of heterogeneity

Assessment of heterogeneity was possible when two or more primary studies were available for inclusion in a meta‐analysis.

Statistical analysis was performed in accordance with the guidelines developed by The Cochrane Collaboration (Higgins 2011).

Heterogeneity (variability) between the results of different studies was examined by:

•  visual inspection of the CIs. A poor overlap indicates heterogeneity,

• low P value (< 5%) or a large Chi² statistic relative to the degree of freedom indicates heterogeneity,

•  I² statistic > 50% suggests substantial heterogeneity.

Assessment of reporting biases

This review sought to expose the publication and related biases (PRB) by using alternative, robust search strategies including handsearching (Hopewell 2007a) and the Internet (e.g. Google and other search engines), comprehensive search of the grey literature (Hopewell 2007b), alternative sources of data or synthesised evidence, and by contacting experts and the research community.

Graphical, descriptive and analytical methods were used to detect, and mitigate, the problem. If sufficient trials were available, funnel plots were to be constructed to illustrate the effect size versus measure of precision. A visual inspection of the plot(s) would have been used to confirm the presence and magnitude of PRB (Song 2002). Further, complex statistical methods were not used to explore for PRB by plotting estimates against corresponding precision for each meta‐analysis as follows: Begg & Mazumdar's rank correlation test (Begg 1994), Egger's regression test (Egger 1997) and the Trim and Fill method (Duval 2000). These alternative methods are necessary because we anticipate that, as in most reviews, our meta‐analysis may include a small numbers of studies. Further, the asymmetry observed in the funnel plot may be due either to serious methodological flaws (Stuck 1998) or to the fact that the intervention is highly effective.

Data synthesis

The included studies were combined when appropriate using fixed‐effect models.

Where combining primary studies was not appropriate, they were summarised in a narrative form. This review shall be updated every two years or earlier if any new RCT evidence becomes available.

Subgroup analysis and investigation of heterogeneity

The presence or absence of heterogeneity was considered before pooling data from two or more trials.

The pre‐specified potential sources of heterogeneity were used to explore possible explanations of variation in effect between trials, and to guide interpretation of the findings.

If substantial heterogeneity was present, then the potential sources of heterogeneity that would have been explored were individual study risk of bias, dosage of mifepristone and differences between patient populations such as age or size of fibroid.

Steps to assess this would have included:

• considering if a meta‐analysis was appropriate,

• considering completing a subgroup analysis (age, dose),

• considering completing a meta‐regression analysis,

• considering ignoring the heterogeneity.

Where heterogeneity has been identified, we are aware of the limited value of an interpretation of the causes of it.

Sensitivity analysis

A sensitivity analysis was considered to explore the presence of any substantial heterogeneity and to confirm the results seen.

Sensitivity analyses would have been conducted for the primary review outcomes to determine whether the results were robust to decisions made during the review process. These analyses would exclude the following studies:

1. studies that do not clearly describe adequate procedures for allocation concealment and blinding,

2. studies with more than 10% of data missing or imputed for the primary outcomes.

Other sensitivity or subgroup analyses may have been conducted to investigate significant heterogeneity found during the review process. These would be interpreted with caution as they were not pre‐specified.

Results

Description of studies

Studies identified

The initial electronic searches identified 135 citations. After reading the titles and abstracts 126 were excluded because they were duplicates, non‐clinical studies, review articles, case reports, case series or had study objectives different from those of this review. Nine studies appeared to meet inclusion criteria but six studies were excluded after three review authors (LO, MT and AR) independently assessed the full articles. We identified three studies that met the inclusion criteria. We have shown the flow of reference retrieval in Figure 3.

3.

Study flow diagram.

Results of the search

The initial electronic search identified 135 citations. Search current date November 2011.

Included studies

Three RCTs were included in this review (Bagaria 2009; Engman 2009; Fiscella 2006). These RCTs reported random assignment of participants with uterine fibroids to mifepristone or placebo (Bagaria 2009; Fiscella 2006), or vitamin‐B tablets (Engman 2009) (see Characteristics of included studies). The trials were conducted in the US (Fiscella 2006), Sweden (Engman 2009) and India (Bagaria 2009). All trials were published in English.

Participants

The total number of participants was 112. Trials included pre‐menopausal women with symptomatic uterine fibroids diagnosed by vaginal and abdominal uterine ultrasonography. One trial reported baseline comparability between groups (Bagaria 2009), while there was a slight difference in free testosterone levels at baseline in one study (Engman 2009). In Fiscella 2006, the median body mass index and uterine volume differed between the control and intervention groups at baseline.

Interventions

One study (Engman 2009) compared mifepristone at a dose of 50 mg every other day with vitamin B tablets as placebo, while another (Bagaria 2009) compared mifepristone at a dose of 10 mg/day with placebo. Another study (Fiscella 2006) compared mifepristone at a dose of 5 mg/day with placebo tablets. Treatment duration was six months in one trial (Fiscella 2006) and three months in the other two trials (Bagaria 2009; Engman 2009).

Outcomes measured

See Characteristics of included studies.

All three studies assessed the effect of mifepristone on leiomyoma‐related symptoms such as menorrhagia, dysmenorrhoea, pelvic pressure, pelvic pain, low backache, rectal pressure, urinary frequency and dyspareunia. These symptoms were measured according to a visual analogue scale in one study (Bagaria 2009) and a 5‐point Likert scale was used in two trials (Engman 2009; Fiscella 2006).

Only one study (Fiscella 2006) assessed specific overall quality of life (QoL) using a Uterine Fibroid Symptom Quality of Life scale (Spies 2002). This study also included global health status (measured by the Medical Outcomes 36‐item Short Form (SF‐ 36) survey) and global pain (measured by the McGill Pain Questionnaire).

Fibroid volume and uterine volume were assessed by ultrasound evaluation (transvaginal or transabdominal) in all studies. Each study used different calculation methods. See Characteristics of included studies.

Blood loss was quantified using a pictorial blood loss assessment chart and calculating a menstrual blood loss (MBL) index in two trials (Bagaria 2009; Fiscella 2006) while a daily record of bleeding reported on a 5‐point Likert scale in one trial (Engman 2009).

All the included studies assessed haemoglobin levels and liver function tests, and all studies investigated side effects such as nausea, vomiting, diarrhoea, headache, fatigue, hot flushes and decreased libido.

Endometrial biopsy was done at baseline and after completion of the therapy in all included studies (Bagaria 2009; Engman 2009; Fiscella 2006).

None of the studies assessed live birth, pregnancy rate, recurrence rate with the possibility of necessitating further additional therapy, or cost‐effectiveness of treatment.

The bibliographic search did not identify studies comparing mifepristone with no treatment, other medical intervention (except for leuprolide acetate, which was excluded (Reinsch 1994)) or surgical intervention.

Excluded studies

Of the nine studies initially identified, six were excluded. Details are included in the Characteristics of excluded studies table.

Risk of bias in included studies

See Figure 1; Figure 2

Allocation

Bagaria 2009, Engman 2009 and Fiscella 2006 all had a low risk of bias with regard to method of allocation concealment (see risk of bias in Characteristics of included studies). In one study (Bagaria 2009) packets were randomised to contain either the drugs or the placebo, then participants as they were enrolled in the trial were assigned to receive a previously sequentially numbered packet of identical appearance. Fiscella 2006 used sequentially numbered, opaque, sealed envelopes and a pharmacy‐controlled randomisation, and Engman 2009 also used a central allocation and coded packets.

Sequence generation was by computer‐generated random tables/lists (Bagaria 2009; Engman 2009; Fiscella 2006).

Blinding

All the included studies stated that patients and study personnel were blinded to the treatment groups. However, the important changes in symptoms such as bleeding patterns or flushes could have affected the blinding. One study (Fiscella 2006) reported at the end of the trial that 19 of 20 (95%) women in the treatment group correctly guessed that they had been receiving mifepristone. Although no other study stated this information, it is possible that a similar situation may have been present in these studies because of the marked effect of the intervention on bleeding patterns. Although there is no risk of bias in the outcome measurements (fibroid volume and related symptoms: bleeding patterns) this statement is unclear for performance bias.

No study reported outcome assessor blinding.

Incomplete outcome data

Bagaria 2009 reported five women lost to follow‐up; however, no reasons for missing data were provided. Engman 2009 excluded two participants after randomisation owing to uncontrollable bleeding in one case and elevated serum follicle‐stimulating hormone (FSH) in the other. Neither case had a clinically relevant impact on the intervention effect estimate. Fiscella 2006 stated that all participants, including the three who provided only baseline measures and two more who withdrew later, were included in the analysis, but the method of imputation was not described.

Selective reporting

Protocols were available online for two of the three included studies (Engman 2009; Fiscella 2006). Engman 2009 reported all data as stated in the protocol; however Fiscella 2006 failed to report non‐significant symptom data. Bagaria 2009 had an unclear risk of reporting bias because the protocol was not available.

Other potential sources of bias

No other potential sources of bias were identified.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

Summary of findings 1. Mifepristone for uterine fibroids.

Patient or population: patients with uterine fibroids Settings: trials were conducted in the US (Fiscella 2006), Sweden (Engman 2009) and India (Bagaria 2009) Intervention: mefepristone
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Mifepristone
Relief of bleeding Pictorial bleeding charts Follow‐up: mean 3 to 6 months	Study population		OR 17.84 (6.72 to 47.38)	77 (2 studies)	⊕⊕⊝⊝ low1,2
	Moderate
Reduction in average fibroid volume Scale from: 0 to 100 Follow‐up: mean 3 to 6 months	The mean reduction in average fibroid volume in the control groups was 108 Mean	The mean reduction in average fibroid volume in the intervention groups was 0.02 standard deviations higher (0.38 lower to 0.41 higher)		99 (3 studies)	⊕⊕⊝⊝ low3,4
Reduction in uterine volume Scale from: 0 to 100 Follow‐up: mean 3 to 6 months	The mean reduction in uterine volume in the control groups was 272.86 Mean	The mean reduction in uterine volume in the intervention groups was 77.24 lower (240.62 lower to 86.14 higher)		72 (2 studies)	⊕⊕⊝⊝ low1,5
Adverse events Follow‐up: mean 3 to 6 months	Moderate		OR 31.65 (4.83 to 207.35)	54 (2 studies)	⊕⊕⊝⊝ low6,7
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Blinding (performance bias and detection bias) outcome assessors: probably not done. Investigator stated as blinded but unsure as to whether investigator is outcome assessor. Incomplete outcome data (attrition bias): no reasons for missing data provided ‐ "lost to follow‐up".[06:38:22 p.m.] Anggie Ramírez: authors used a non‐validated questionnaire. No explanation was provided
² Only 77 participants included
³ Blinding (performance bias and detection bias) outcome assessors: probably not done. Investigator stated as blinded but unsure as to whether investigator is outcome assessor. Incomplete outcome data (attrition bias): no reasons for missing data provided ‐ "lost to follow‐up". Used a non‐validated questionnaire. No explanation was provided. The questionnaire used for leiomyoma‐related symptoms is described by study authors as not validated and had low sensitivity. The study had baseline imbalance, such an imbalance could suggest failure of randomisation to balance known and unknown confounders factors. Owing to the marked effect in bleeding patterns and hot flushes, participants/study personnel could have been aware of treatment assignments
⁴ Only 99 participants with few outcomes
⁵ Only 72 participants with few outcomes
⁶ Blinding (performance bias and detection bias) outcome assessors: probably not done. Investigator stated as blinded but unsure as to whether investigator is outcome assessor. Incomplete outcome data (attrition bias): no reasons for missing data provided ‐ "lost to follow‐up". Owing to the marked effect in bleeding patterns and hot flushes, participants/study personnel could have been aware of treatment assignments
⁷ Only 54 participants with few outcomes

Summary of findings 2. Mifepristone (10 mg/day/3 months) compared to placebo for uterine fibroids.

Patient or population: patients with uterine fibroids Settings: India Intervention: mifepristone (10 mg/day) Comparison: placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Mifepristone (10 mg/day)
Leiomyoma volume (mL)1 Scale from: 0 to 100. Follow‐up: mean 3 months	The mean leiomyoma volume in the control groups was 118.3 ± 243 mL	The mean leiomyoma volume in the intervention groups was 0.10 lower (0.76 lower to 0.57 higher)		35 (1 study)	⊕⊕⊝⊝ low2,3
Uterine volume (mL)1. Scale from: 0 to 100. Follow‐up: mean 6 months	The mean uterine volume in the control groups was 281.1 ± 417 mL	The mean uterine volume in the intervention groups was 93.10 lower (317.07 lower to 130.87 higher)		35 (1 study)	⊕⊕⊝⊝ low2,3
Relief of bleeding dichotomic Follow‐up: mean 3 months			OR 27.02 (7.25 to 100.7)	35 (1 study)	⊕⊕⊝⊝ low2,4
Relief of pelvic pain dichotomic Follow‐up: mean 3 months	182 per 1000	318 per 1000 (78 to 722)	OR 2.10 (0.38 to 11.71)	26 (1 study)	⊕⊕⊝⊝ low2,3
Relief of dysmenorrhoea dichotomic Follow‐up: mean 3 months	182 per 1000	708 per 1000 (345 to 918)	OR 10.93 (2.37 to 50.51)	26 (1 study)	⊕⊕⊝⊝ low2,3
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Formula: 4/3 π W/2 x L/2 x T/2
² Blinding (performance bias and detection bias) outcome assessors: probably not done. Investigator stated as blinded but unsure as to whether investigator is outcome assessor. Incomplete outcome data (attrition bias): no reasons for missing data provided ‐ "lost to follow‐up"
³ Only 35 participants with few outcomes
⁴ No explanation was provided

Summary of findings 3. Mifepristone (50 mg/1 day after/3 months) compared to placebo for uterine fibroids.

Patient or population: patients with uterine fibroids Settings: Sweden Intervention: mifepristone (50 mg/1 day after/3 months) Comparison: placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Mifepristone (50 mg/1 day after/3 months)
Leiomyoma volume (mL). Scale from: 0 to 100. Follow‐up: mean 3 months	The mean leiomyoma volume in the control groups was 118 ± 59.26 mL	The mean leiomyoma volume in the intervention groups was 0.13 lower (0.89 lower to 0.63 higher)1		30 (1 study2)	⊕⊕⊝⊝ low3,4
Cystic glandular dilatation dichotomic Follow‐up: mean 3 months	91 per 1000	625 per 1000 (120 to 953)	OR 16.67 (1.36 to 204.03)	19 (1 study)	⊕⊕⊝⊝ low3,4,5
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ P = 0.02
² Volume was calculated using the formula for an ellipsoid, 0.523 x A x B x C
³ The questionnaire used for leiomyoma‐related symptoms is described by study authors as not validated and had low sensitivity. The study had baseline imbalance, such as imbalance could suggests failure of randomisation to balance known and unknown confounders factors. Owing to the marked effect in bleeding patterns and hot flushes, participants/study personnel could have been aware of treatment assignments.
⁴ Only 30 participants with few outcomes
⁵ CIs very wide

Summary of findings 4. Mifepristone (5 mg/day/6 months) versus placebo for uterine fibroids.

Patient or population: patients with uterine fibroids Settings: US Intervention: mifepristone (5 mg/day/6 months) Comparison: placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Mifepristone (5 mg/day/6 months)
Relief of bleeding dichotomic Follow‐up: mean 6 months			OR 2.22 (0.54 to 9.13)	42 (1 study)	⊕⊕⊝⊝ low1,2
Leiomyoma volume (mL) Follow‐up: mean 6 months	The mean leiomyoma volume in the control groups was 94.1 ± 113 mL	The mean leiomyoma volume in the intervention groups was 0.24 higher (0.40 lower to 0.89 higher)		37 (1 study)	⊕⊕⊝⊝ low1,2
Uterine volume (mL) Follow‐up: mean 6 months	The mean uterine volume in the control groups was 537.5 ± 291 mL	The mean uterine volume in the intervention groups was 59.20 lower (298 lower to 179.65 higher)		37 (1 study)	⊕⊕⊝⊝ low1,2
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹ Authors used a non‐validated questionnaire
² Only 30 participants included

The studies identified used different doses of mifepristone and were not easily comparable. Meta‐analyses for relief of bleeding, fibroid volume, uterine volume and adverse events were possible; however, the small number of participants limits the generalisability of the results.

Comparison of mifepristone versus placebo

The three included studies compared mifepristone with placebo (Bagaria 2009, Fiscella 2006) or vitamin B (Engman 2009).

1.1 Relief of bleeding (Analysis 1.1)

Data from two studies were included in the meta‐analysis (Bagaria 2009; Fiscella 2006). There is evidence that treatment with mifepristone relieved bleeding compared with placebo (Peto OR 17.84; 95% CI 6.72 to 47.38; two RCTs, 77 women, I² = 0%) (Analysis 1.1; Figure 4).

1.1. Analysis.

Comparison 1: Mifepristone compared with placebo, Outcome 1: Relief of bleeding

4.

Forest plot of comparison: 1 Mifepristone compared with placebo, outcome: 1.1 Relief of bleeding.

Engman 2009 presented the data as median and ranges, so the study was not included in the meta‐analysis. The authors reported that, during weeks 9 to 12 of treatment, women in the mifepristone group had one bleeding day (median, no range provided), while in the placebo group, women had three bleeding days (range 0 to 27 days; P < 0.001).

1.2 Relief of pain (Analysis 1.2)

We have subgrouped this outcome into two as set out below (Analysis 1.2; Figure 5).

1.2. Analysis.

Comparison 1: Mifepristone compared with placebo, Outcome 2: Relief of pain

5.

Forest plot of comparison: 1 Mifepristone compared with placebo, outcome: 1.2 Relief of pain.

1.2.1 Relief of dysmenorrhoea

Only Bagaria 2009 reported on this outcome. There was no evidence of an effect of treatment with mifepristone on dysmenorrhoea compared with placebo (Peto OR 2.10; 95% CI 0.38 to 11.71; one RCT, 26 women).

1.2.2 Relief of pelvic pain

Only Bagaria 2009 reported on this outcome in a form suitable for analysis. There was no evidence of an effect when comparing mifepristone with placebo (Peto OR 0.89; 95% CI 0.27 to 2.95; one RCT, 26 women).

Fiscella 2006 provided data on change in bleeding patterns and change in pain, but only as graphics; therefore, it was not possible to use numerical data.

1.3 Relief of pressure symptoms

Only Engman 2009 reported on bladder pressure, but not in a format suitable for analysis. Bagaria 2009 reported on a reduction in urinary complaints, which is not an outcome pre‐specified in this protocol.

1.4 Uterine Fibroid Symptom Quality of Life (Analysis 1.4)

One study (Fiscella 2006) used the UFS‐QoL, a new disease‐specific symptom and health‐related QoL questionnaire for fibroids (Spies 2002). The scale is 1 to 100, with higher scores indicating better QoL. The author has provided raw data for individual indicators of the Uterine Fibroid Symptoms UFS‐QoL measures. There is evidence of an effect of mifepristone on fibroid‐specific quality of life (MD 33.05, 95% CI 18.27, 47.83, 37 women, Analysis 1.4). By 6 months, the mean fibroid‐specific quality of life measures had improved by 135% in the mifepristone group, compared with a 41% improvement in the placebo arm of the trial. The study suggests significant improvements (P < 0.001) for specific fibroid‐related QoL, with aspects including concern, activities, energy and mood, control, self‐consciousness and sexual functioning.

1.4. Analysis.

Comparison 1: Mifepristone compared with placebo, Outcome 4: Uterine Fibroid Symptom Quality of Life

1.5 Reduction in fibroid volume (Analysis 1.5)

Three studies (Bagaria 2009; Engman 2009Fiscella 2006) were included in the meta‐analysis of this comparison. There was no evidence of an effect of mifepristone on the fibroid volume (SMD ‐0.02; 95% CI ‐0.38 to 0.41; 99 women) (Analysis 1.5).

1.5. Analysis.

Comparison 1: Mifepristone compared with placebo, Outcome 5: Reduction in average fibroid volume

1.6 Reduction in uterine volume (Analysis 1.6)

Two studies (Bagaria 2009; Fiscella 2006) were included in the meta‐analysis of this comparison. There was no evidence of an effect of mifepristone on uterine volume (MD ‐77.24; 95% CI ‐240.62 to 86.14; 72 women) (Analysis 1.6).

1.6. Analysis.

Comparison 1: Mifepristone compared with placebo, Outcome 6: Reduction in uterine volume

1.7 Live birth rate

None of the identified studies reported live birth rate.

1.8 Pregnancy rate

None of the identified studies reported pregnancy rate.

1.9 Recurrence rate

None of the studies identified reported recurrence rate necessitating additional therapy.

1.10 Adverse events (Analysis 1.7)

The pooled data show more women with endometrial change in the mifepristone group compared with placebo (OR 31.65; 95% CI 4.83 to 207.35; two RCTs, 54 women, I² = 0%) (Analysis 1.7). Only one study (Bagaria 2009) reported endometrial hyperplasia at the end of the therapy, 12/19 women in the mifepristone group compared with 0/16 in the placebo group (OR 55.0; 95% CI 2.86 to 105.67). There was no demonstrated atypia. Trial authors reported that all the patients with endometrial hyperplasia (12 patients) showed normal endometrium on histopathology specimens when they underwent surgical intervention. Engman 2009 also collected endometrial biopsies, without any evidence of hyperplasia or malignancy. They found a significantly higher rate of cystic glandular dilatation in women in the mifepristone group (5/8 women biopsied) compared with the placebo group (1/11 women biopsied) (OR 16.67; 95% CI 1.36 to 204.03). Fiscella 2006 also reported a higher rate of cystic glandular dilatation in the mifepristone group.

1.7. Analysis.

Comparison 1: Mifepristone compared with placebo, Outcome 7: Adverse events

No study showed liver function test alterations.

1.11 Cost‐effectiveness

No study reported this outcome.

Discussion

Summary of main results

This systematic review included three RCTs, with small samples, which resulted in limited overall quality. Therefore, the analysis generated an overall low quality of evidence. These results limit recommendations for the use of mifepristone in current clinical practice for the treatment of fibroids.
Three comparisons were meta‐analysed: relief of bleeding (two RCTs); reduction of average fibroid volume (three RCTs) and reduction of uterine volume (two RCTs). However, none of the three studies has enough subjects to rule out type II errors. There is significant reduction of bleeding but no significant reductions in uterine and fibroid volume. The reduction of bleeding is probably due to the progesterone receptor modulator effect of mifepristone and not by a direct effect of reducing fibroid size.

Fiscella 2006 reported significant improvements (P < 0.001) for specific measured aspects of QoL. Treatment with mifepristone was also associated with gains in energy and health status and concomitant reductions in fatigue and pain.

Available data suggest that, compared with placebo, mifepristone was better in relieving heavy menstrual bleeding. The available data suggest that mifepristone had no effect on the average size of fibroids and showed a non‐significant effect (P = 0.35) on reduction of uterine volume. Some improvements in QoL in symptomatic women were noted. Mifepristone increased the occurrence of endometrial changes found on biopsy. The significance of these changes has yet to be elucidated (Mutter 2008) as these appear to be different from the possibly pre‐malignant endometrial hyperplasia associated with unopposed oestrogen treatment. In addition, Narvekar et al showed that low‐dose mifepristone inhibits endometrial proliferation and upregulates androgen receptor expression (Narvekar 2004). The included trials had small numbers of participants and the possibility of type I and II errors cannot be excluded. This reduces the ability to influence clinical decision‐making. While there was no statistically significant heterogeneity, there remains the potential for clinical heterogeneity owing to the differences in doses used in the trials.

Overall completeness and applicability of evidence

There were only three small studies from three different countries with three different dose regimens identified and only one study considered QoL using the UFS‐QoL scale. The patients in the studies can be considered to be representative of the population of pre‐menopausal women with uterine fibroids.

Quality of the evidence

See: Table 1, Table 3 and Table 4. Tables were developed in GRADE PRO 2011).

According to the GRADE system used by the review authors, the quality of the evidence was rated as low.

Every trial included stated that efforts were made to blind the participants and the staff to the administration of the medication. However, the effect of mifepristone in diminishing or stopping heavy menstrual bleeding was likely to have been discernible, which may have had an effect on the blinding. Although it is unlikely that this fact affected objective measurements like fibroid or uterus size, incidence of endometrial hyperplasia and bleeding patterns, it may have had some influence in a differential way on subjective self‐reported outcomes related to QoL (e.g. UFS‐QoL scale, 5‐point Likert scale).

One of the studies (Fiscella 2006) showed differences in baseline characteristics (body mass index, uterine volume) that may have affected the measured outcomes.

Overall, few participants were lost to follow‐up and this was unlikely to influence the measure of effect.

Potential biases in the review process

The literature search was very exhaustive; therefore, the risk of publication bias is low. All important decisions regarding the inclusion and exclusion of the articles, as well as those related to the analysis, were made by discussion and consensus.

Agreements and disagreements with other studies or reviews

A Cochrane review (Liu 2009) compared the Huoxue Sanjie herbal decoction with mifepristone. Both interventions showed no significant difference in the disappearance of uterine fibroids, the number of patients with shrinking of uterine fibroids or the average uterine volume. The authors found that the decoction was less effective than mifepristone on reducing the average uterus size (MD 23.23 cm³; 95% CI 17.85 to 28.61).

Authors' conclusions

Implications for practice.

Mifepristone was shown to reduce heavy menstrual bleeding and improve fibroid‐specific quality of life in pre‐menopausal women with fibroids. The lack of reduction in fibroid volume suggests that at present this medication is not recommended as a therapy for uterine fibroids.

Implications for research.

Further well‐designed large RCTs are needed, including comparisons of mifepristone at different doses with placebo. In addition, studies comparing the efficacy of a number of progesterone receptor modulators including mifepristone, ulipristil and asoprisnil (among others) would provide important clinical evidence.

What's new

Date	Event	Description
19 January 2021	Review declared as stable	This Review has superseded because it has been merged with "Selective progesterone receptor modulators (Murji 2017) https://doi.org/10.1002/14651858.CD010770.pub2

History

Protocol first published: Issue 2, 2009
Review first published: Issue 8, 2012

Date	Event	Description
23 March 2012	New search has been performed	New secondary outcomes added: reduction in uterine volume and fibroid‐specific quality of life.
13 April 2008	Amended	converted to new review format
13 December 2006	New citation required and major changes	Substantive amendment

Appendices

Appendix 1. MEDLINE search strategy

<1950 to August Week 4 2009>
1 exp Leiomyoma/ (14375)
2 exp Fibroma/ (10225)
3 fibroid$.ti,ab,sh. (2647)
4 fibroma$.ti,ab,sh. (13514)
5 fibromyoma$.ti,ab,sh. (526)
6 (uter$ adj3 myoma$).ti,ab,sh. (2296)
7 Leiomyoma$.ti,ab,sh. (15778)
8 or/1‐7 (30634)
9 Mifepristone/ (4453)
10 Mifepristone.ti,ab,sh. (4822)
11 mifegyne.ti,ab,sh. (11)
12 mifeprex.ti,ab,sh. (9)
13 r38486.ti,ab,sh. (1)
14 ru‐38486.ti,ab,sh. (416)
15 ru‐486.ti,ab,sh. (1495)
16 or/9‐15 (5268)
17 8 and 16 (86)
18 randomized controlled trial.pt. (280165)
19 controlled clinical trial.pt. (80498)
20 randomized.ab. (189021)
21 placebo.ab. (115356)
22 cross‐over studies/ (24836)
23 (crossover or cross‐over or cross over).tw. (43875)
24 clinical trials as topic.sh. (146150)
25 randomly.ab. (136940)
26 trial.ti. (82239)
27 or/18‐26 (663847)
28 humans.sh. (10982011)
29 27 and 28 (598163)
30 29 and 17 (21)
31 from 30 keep 1‐21 (21)

Appendix 2. EMBASE search strategy

1980 to 2009 Week 35>
1 Leiomyoma/ (4671)
2 Fibroma/ (3071)
3 Uterus Myoma/ (5748)
4 fibroid$.tw. (2548)
5 fibroma$.tw. (4514)
6 fibromyoma$.tw. (159)
7 (uter$ adj3 myoma$).tw. (1248)
8 Leiomyoma.tw. (3873)
9 or/1‐8 (17497)
10 Mifepristone/ (6862)
11 Mifepristone$.tw. (1847)
12 Lunarette.tw. (1)
13 Mifegyne.tw. (160)
14 Mifeprex.tw. (81)
15 Ru 38486.tw. (808)
16 Ru 486.tw. (3259)
17 Ru38486.tw. (297)
18 Ru486.tw. (1540)
19 Ru 486 6.tw. (2)
20 or/10‐19 (7213)
21 Clinical Trial/ (553624)
22 Randomized Controlled Trial/ (172938)
23 exp randomization/ (27016)
24 Single Blind Procedure/ (8471)
25 Double Blind Procedure/ (73817)
26 Crossover Procedure/ (21683)
27 Placebo/ (130648)
28 Randomi?ed controlled trial$.tw. (34707)
29 Rct.tw. (2891)
30 random allocation.tw. (645)
31 randomly allocated.tw. (10460)
32 allocated randomly.tw. (1362)
33 (allocated adj2 random).tw. (565)
34 Single blind$.tw. (7642)
35 Double blind$.tw. (86500)
36 ((treble or triple) adj blind$).tw. (141)
37 placebo$.tw. (112693)
38 prospective study/ (85065)
39 or/21‐38 (726966)
40 case study/ (6306)
41 case report.tw. (122424)
42 abstract report/ or letter/ (508941)
43 or/40‐42 (635252)
44 39 not 43 (701604)
45 9 and 44 and 20 (90)
46 from 45 keep 1‐90 (90)

Appendix 3. CENTRAL search strategy

<3rd Quarter 2009>
1 exp Leiomyoma/ (283)
2 exp Fibroma/ (1)
3 fibroid$.ti,ab,sh. (178)
4 fibroma$.ti,ab,sh. (17)
5 fibromyoma$.ti,ab,sh. (11)
6 (uter$ adj3 myoma$).ti,ab,sh. (113)
7 Leiomyoma$.ti,ab,sh. (332)
8 or/1‐7 (473)
9 Mifepristone/ (278)
10 Mifepristone.ti,ab,sh. (376)
11 mifegyne.ti,ab,sh. (0)
12 mifeprex.ti,ab,sh. (0)
13 r38486.ti,ab,sh. (0)
14 ru‐38486.ti,ab,sh. (1)
15 ru‐486.ti,ab,sh. (102)
16 or/9‐15 (391)
17 8 and 16 (11)
18 from 17 keep 1‐11 (11)

Appendix 4. CINAHL search strategy

1 exp Leiomyoma/ 
2 exp Fibroma/ 
3 fibroid$.ti,ab,sh. 
4 fibroma$.ti,ab,sh. 
5 fibromyoma$.ti,ab,sh.
6 (uter$ adj3 myoma$).ti,ab,sh. 
7 Leiomyoma$.ti,ab,sh. 
8 or/1‐7 
9 Mifepristone/ 
10 Mifepristone.ti,ab,sh. 
11 mifegyne.ti,ab,sh. 
12 mifeprex.ti,ab,sh. 
13 r38486.ti,ab,sh. 
14 ru‐38486.ti,ab,sh. 
15 ru‐486.ti,ab,sh. 
16 or/9‐15 
17 8 and 16 
18 exp clinical trials/ 
19 Clinical trial.pt. 
20 (clinic$ adj trial$1).tw. 
21 ((singl$ or doubl$ or trebl$ or tripl$) adj (blind$3 or mask$3)).tw. 
22 Randomi?ed control$ trial$.tw. 
23 Random assignment/ 
24 Random$ allocat$.tw. 
25 Placebo$.tw. 
26 Placebos/ 
27 Quantitative studies/ 
28 Allocat$ random$.tw. 
29 or/18‐28 
30 29 and 17
31 from 30 keep 1‐2

Appendix 5. Study eligibility

Date
Extractor
Trial authors
Publication year
Journal
1) Design
Described as randomised? If no then exclude. If yes go to questions 2	Yes No Unclear
2) Participants
(a)	Yes No Unclear
(b)	Yes No Unclear
If 'no', exclude. Otherwise go to question (3).
3) Interventions
	Yes No Unclear
	Yes No Unclear
If 'no' to (a) or (b), exclude.
Final decision
Include (if all 'yes') Exclude (if any 'no') Unclear
Excluded or unclear because:
If 'unclear', action taken:

Appendix 6. Data extraction form

Date:
Extractor (initials):
Trial authors:
Year of publication:
Journal:
Study setting
(1) Participants
Inclusion criteria:	Exclusion criteria:
Median or mean age:	Ethnicity
Age range:	Gravidity
Were all treatment groups comparable at baseline:	Yes No Unclear
If no or unclear, describe any differences:
Notes:
2) Interventions	Tx 1	Tx2
Tx used
Formulation used
Route
Dose
Duration
Timing and frequency
Notes
(3) Outcomes
Further information:
Trialists contacted for more information:	yes	no
Address
Ph
Email
Data
Comments

Data and analyses

Comparison 1. Mifepristone compared with placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Relief of bleeding	2	77	Peto Odds Ratio (Peto, Fixed, 95% CI)	17.84 [6.72, 47.38]
1.2 Relief of pain	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Totals not selected
1.2.1 Relief of dysmenorrhoea	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Totals not selected
1.2.2 Relief of pelvic pain	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Totals not selected
1.4 Uterine Fibroid Symptom Quality of Life	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.5 Reduction in average fibroid volume	3	99	Std. Mean Difference (IV, Fixed, 95% CI)	0.02 [‐0.38, 0.41]
1.6 Reduction in uterine volume	2	72	Mean Difference (IV, Fixed, 95% CI)	‐77.24 [‐240.62, 86.14]
1.7 Adverse events	2	54	Odds Ratio (M‐H, Random, 95% CI)	27.45 [4.06, 185.56]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bagaria 2009.

Study characteristics
Methods	Study design: double‐blind RCT Total study duration: 3 months
Participants	Total number: 40 participants 20 randomised to IG and 20 randomised to CG Losses to follow‐up: IG: 1, CG: 4 35 participants (IG: 19; CG: 16) completed all 3 moths of follow‐up Setting: the gynaecological outpatient department of University College of Medical Sciences and Guru Teg Bahadur Hospital, Delhi Country: India Inclusion criteria: pre‐menopausal women with uterine fibroids. Specific inclusion criteria were not stated by the study authors Exclusion criteria: pregnancy or breast feeding major medical morbidity such as ovarian, cervical or uterine malignancy; presence of liver, respiratory (asthma), renal, heart disease, pelvic inflammatory disease or any other adnexal pathology patient necessitating early surgical intervention for uterine leiomyoma hormonal contraception or any hormonal therapy in the previous 3 months Age (mean ± SD): IG: 40.3 ± 6.8 years, CG: 41.1 ± 9.3 years Baseline imbalances: both groups were comparable for baseline parameters
Interventions	IG: mifepristone 10 mg/day, prepared from 200 mg tablet at the start of the study using lactose as filler, oral doses CG: placebo capsules containing lactose filler, oral doses
Outcomes	Leiomyoma volume (mL): assessments were performed by abdominal and vaginal uterine ultrasonography. Volumes were calculated using the Viscomi formula (4/3 π W/2 x L/2 x T/2) where W is uterine width, L is uterine length and T is uterine thickness Uterine volume (mL): assessments were performed by abdominal and vaginal uterine ultrasonography. Volumes were calculated using the formula: 4/3 π abc where a, b and c represent radii of the sphere in 3 dimensions Other symptoms: monthly assessment Severity of the symptoms was graded according to the visual analogue scale while quantification of blood loss was done using pictorial blood loss assessment chart: pelvic pain pelvic pressure bladder pressure low back pain rectal pain intercourse pain urinary frequency menstrual blood loss Pregnancy rates: the study did not address this outcome Haemoglobin levels (mg/dL): at the end of treatment Liver function test: at the end of treatment Adverse events: endometrial pathology: endometrial biopsy at baseline and 3‐months after treatment termination. Diagnostic criteria not stated amenorrhoea: number of participants who were amenorrhoeic at 3 months hot flushes: changes in prevalence at 3 months. How measures were taken was not stated by the study authors recurrence rate of symptoms: the study did not address this outcome others: prevalence of: nausea, vomiting, diarrhoea, headache, decreased libido, fatigue, weakness
Notes	The study was approved by the ethical committee of the department of University College of Medical Sciences and Guru Teg Bahadur Hospital Financial support was not stated
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random tables, randomising packets to contain either active drug or placebo
Allocation concealment (selection bias)	Low risk	Sequentially numbered drug packets of identical appearance
Blinding (performance bias and detection bias) Participants	Unclear risk	"The patients and the investigator were not aware of the drug being dispensed." Probably participants and study personnel could have guessed the group allocation, but the review authors judge that the main outcomes and the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns) are not likely to be influenced by lack of blinding Although there is no risk of bias in the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns), this statement is unclear for performance bias
Blinding (performance bias and detection bias) study personnel	Unclear risk	Although there is no risk of bias in the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns), this statement is unclear for performance bias
Blinding (performance bias and detection bias) outcome assessors	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	High risk	No reasons for missing data provided ‐ "lost to follow‐up"
Selective reporting (reporting bias)	Low risk	The published report includes all expected outcomes
Other bias	High risk	Imprecision, only 35 participants with few outcomes

Engman 2009.

Study characteristics
Methods	Study design: randomised, double‐blind, placebo‐controlled study Total study duration: 3 months
Participants	Total number: 30 14 randomised to IG and 16 randomised to CG 2 women dropped out during the course of the study (IG: 0, CG: 2) 28 participants completed all 3 months of follow‐up (IG: 14, CG: 14) Setting: the Karolinska University Hospital, Stockholm Country: Sweden Inclusion criteria: "Healthy, non‐pregnant women, referred for evaluation to the outpatient clinic owing to leiomyoma‐related problems indicating surgical intervention" (diagnostic criteria not stated) Exclusion criteria: pregnancy "any history of breast cancer or other malignancy, bleeding not possible to control with tranexamic acid and iron medication ‐ abnormal mammogram and breast biopsy ‐ adnexal abnormality, suspicion of leiomyosarcoma... abnormal FSH or LH levels or any other hormonal dysfunction of significance; laboratory findings that would give suspicion of blood, liver or renal dysfunction, abnormal pap smear at screening or any contraindications to the use of mifepristone" any steroid therapy in the previous 3 months Age (mean ± SD): IG: 40.8 ± 4.7 years, CG: 40.9 ± 7.6 years Baseline imbalances: groups were comparable for age, parity, BMI, endometrial double layer thickness at baseline
Interventions	IG: mifepristone 50 mg (one quarter of 200 mg Mifegyne), every other day with therapy starting on the first day of cycle, oral doses CG: vitamin B tablets (one quarter of TrioBe Recip), visually identical to mifepristone as placebo
Outcomes	Primary outcome: uterine leiomyoma size: measuring (by vaginal uterine ultrasonography) the 3 largest diameters (A, B and C) in 2 planes in approximately perpendicular x, y and z axis directions. Volume was calculated using the formula for an ellipsoid, 0.523 x A x B x C. Measurements were made or supervised every 4 weeks until the surgery by the same operator measurements for both dominant myoma volume and total myoma volume Secondary outcomes: uterine blood flow: using a flow velocity waveform according to the formula: PI:(S‐D)/TAMAX. PI and peak flow were measured in the uterine arteries, the peripheral and central leiomyoma vessels bleeding: assessed by using daily menstrual logs and pictorial bleeding charts. A monthly blood loss index was calculated from menstrual history by assigning values 1 to 4 to each day of spotting, indicating light, moderate and heavy flow, respectively, and then summing the results leiomyoma and general symptoms: assessed using a 5‐point Likert scale items (0 = no symptoms; 4 = very severe). The scores were registered weekly and summarised for every 4‐week period (range 0 to 16) and compared within and between the treatment groups at 4, 8 and 12 weeks of treatment local symptoms: pelvic pain pelvic pressure bladder pressure low back pain rectal pain urinary frequency intercourse pain and general hormone‐related symptoms including hot flushes, headache, nausea, vomiting, diarrhoea, mood fluctuation, libido, weakness and fatigue Adverse events: endometrial pathology: endometrial biopsy at baseline and 3 months were evaluated by an expert investigator blinded to treatment amenorrhoea: number of participants who were amenorrhoeic at 3 months liver function testing, other hormones and routine blood parameters: baseline and at the end of the study Recurrence rate of symptoms: not reported
Notes	The study was approved by the Ethics committee at Karolinska Institutet The study was supported by grants from the Swedish Research Council. Karolinska Institutet and Stockholm city county/Karolinska Institutet
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random computer‐generated list
Allocation concealment (selection bias)	Low risk	Medication was packed and coded by the pharmacy
Blinding (performance bias and detection bias) Participants	Unclear risk	"Patients and study staff were blinded to treatment group". Owing to the marked effect in bleeding patterns and hot flushes, participants/study personnel could have been aware of treatment assignments. The review authors judge that the objective outcomes of leiomyoma size and bleeding patterns are not likely to be influenced by lack of blinding. But subjective outcomes as hot flushes (which had a significant difference between groups) could have been affected by this awareness in this study Although there is no risk of bias in the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns), this statement is unclear for performance bias
Blinding (performance bias and detection bias) study personnel	Unclear risk	Although there is no risk of bias in the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns), this statement is unclear for performance bias
Blinding (performance bias and detection bias) outcome assessors	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for missing outcome data unlikely to be related to true outcome. Reasons for exclusion are uncontrollable bleeding in one case, elevated serum FSH in another
Selective reporting (reporting bias)	Low risk	The published reports include all expected outcomes. Protocol available on clinicaltrials.gov NCT00579475
Other bias	High risk	The questionnaire used for leiomyoma‐related symptoms is described by study authors as not validated and had low sensitivity

Fiscella 2006.

Study characteristics
Methods	Study design: randomised, double‐blind, placebo‐controlled study Total study duration: 6 months
Participants	Total number: 42 22 randomised to IG and 20 randomised to CG 5 women dropped out during the course of the study (IG: 2, CG: 3) 39 participants began the trial and participated for at least 1 month. 37 participants completed all 6 months of follow‐up 3 dropped out before 1 month of prescription (IG: 2, CG: 1). 1 dropped out after 1 month of prescription and 1 after 3 months of prescription (IG: 0, CG: 3) Setting: Departments of Family Medicine, Obstetrics and Gynecology, Biostatistics, and Community & Preventive Medicine, School of Medicine and Dentistry, University of Rochester, Rochester, New York Country: US Inclusion criteria: 18 years or older, pre‐menopausal women, with moderately severe leiomyoma‐related symptoms (> 39 on the Uterine Fibroid Symptom Quality of Life Symptom Severity Subscale) total uterine volume ≥ 160 mL, and at least 1 leiomyoma that was ≥ 2.5 cm measured by vaginal and abdominal ultrasound Exclusion criteria: pregnancy or intended to become pregnant in the next 6 months major medical morbidity or severe anaemia, active mental illness, elevated liver enzymes or substance abuse use of short‐acting hormones in the past 3 months, use of GnRH analogues or other long‐acting hormonal medications in the past 6 months Age (mean ± SD): IG: 48.8 ± 6.2 years, CG: 43.2 ± 4.7 years Baseline imbalances: "the two groups were well‐matched for baseline characteristics with the exception of body mass index and baseline uterine volume"
Interventions	IG: mifepristone 5 mg/day, oral doses CG: placebo capsules daily, identical in weight and appearance to mifepristone
Outcomes	Primary outcome: mean change in leiomyoma specific overall QoL (using Uterine Fibroid Symptom ‐ Quality of Life) scale 1 to 100, with higher scores indicating better QoL. The study suggests significant improvements (P < 0.001) for specific QoL. Treatment with mifepristone also was associated with significant gains in energy and health status and concomitant reductions in fatigue and pain. The author provided raw data for individual indicators of the QoL and reduction of the values of the Uterine Fibroid Symptoms ‐ Quality of Life measures Secondary outcomes: global health status: measured by the Medical Outcomes 36‐item Short Form [SF‐36] survey global pain: measured by McGill Pain Questionnaire each of these questionnaires was administered at baseline, 1 month, 3 months and 6 months, except the McGill Pain Questionnaire, which was assessed monthly upper and lower limits, and whether high or low score is good, was not stated by study authors bleeding: assessed by using daily menstrual logs and pictorial bleeding charts. A monthly blood loss index was calculated from menstrual history by assigning values 1 to 4 to each day of spotting, indicating light, moderate, and heavy flow, respectively, and then summing the results uterine size (mL): uterine volume and leiomyoma size were assessed by vaginal or abdominal ultrasonogram, or both, at baseline, 1 month, 3 months and 6 months. Volume changes were analysed. The uterus was measured in 3 planes and a total volume calculated pregnancy rates presence and intensity of likely leiomyoma symptoms: assessed monthly using a 5‐point Likert scale items: pelvic pain pelvic pressure bladder pressure low back pain rectal pain urinary frequency intercourse pain drug adverse effects, including hot flushes, headache, nausea, vomiting, mood swings, diarrhoea, decreased libido, weakness, fatigue and nervousness Adverse events: endometrial pathology: endometrial biopsy at baseline and 6 months amenorrhoea: number of participants who are amenorrhoeic at 6 months alterations in liver function testing: mean changes at 1, 3 and 6 months recurrence rate of symptoms: not reported
Notes	The study was approved by the University of Rochester Institutional Review Board. Women were paid for each study visit to defray their expenses The study was supported by funding from the National Institute for Child Health and Human Development. Athenium Laboratories (New York, NY) supplied the drug at cost
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Blocked randomisation, using a computer random number generator
Allocation concealment (selection bias)	Low risk	Sequentially numbered, opaque, sealed envelopes
Blinding (performance bias and detection bias) Participants	Unclear risk	"Capsules were indistinguishable in all aspects of their outward appearance. For each drug an identically matched placebo was available". However, study authors stated that 19 of 20 (95%) women in the treatment group correctly guessed that they had been receiving mifepristone because of the dramatic improvements in symptoms, including cessation of bleeding among women in the intervention group. The review authors judge that objective outcomes such as leiomyoma size and bleeding patterns are not likely to be influenced by lack of blinding. However, leiomyoma‐specific QoL (primary aim of this study) as well as global QoL could have been affected by awareness of treatment allocation Although there is no risk of bias in the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns), this statement is unclear for performance bias
Blinding (performance bias and detection bias) study personnel	Unclear risk	Although there is no risk of bias in the outcome measurements (leiomyoma volume and related symptoms: bleeding patterns), this statement is unclear for performance bias
Blinding (performance bias and detection bias) outcome assessors	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	"All 42 women who were randomised were included in the analysis, including the three who provided only baseline measures and two more who withdrew later" No reasons given for loss of outcome data Method of imputation not described
Selective reporting (reporting bias)	Unclear risk	Insufficient information to judge 'yes' or 'no'. Non‐significant symptom data has not been reported. Protocol available on http://clinical trials.gov and gives similar information to methods section (NCT00133705)
Other bias	High risk	The study had baseline imbalance: body mass index (IG: mean: 31.7 kg/m2, SD: 8.7 kg/m2, and CG: mean: 27.2 kg/m2, SD: 5.6 kg/m2) uterine volume (IG: mean: 719 mL, SD: 663 mL, CG and mean: 449 mL, SD: 236 mL risk of selection bias because the use of blocked randomisation in a blinded trial (small size) where the blinding is broken. Such as imbalance could suggest failure of randomisation to balance known and unknown confounders factors participants were permitted to use analgesics. Although analgesic use did not differ between groups, inappropriate administration of a co‐intervention could have biased the global pain outcome

BMI: body mass index; CG: control group; FSH: follicle‐stimulating hormone; GnRH: gonadotrophin‐releasing hormone; IG: intervention group; LH: luteinising hormone; QoL: quality of life; RCT: randomised controlled trial; SD: standard deviation.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Carbonell 2008	The study compared different dosages of mifepristone
Eisinger 2003	The study compared different dosages of mifepristone
Eisinger 2005	The study compared different dosages of mifepristone (companion to Eisinger 2003)
Reinsch 1994	Data were provided as percentages, in a useless fashion for analysis
Yang 1996	Non‐randomised design
Zeng 1998	Non‐randomised design

Differences between protocol and review

Addition of two secondary outcomes, reduction in uterine volume measured by ultrasonography or Magnetic Resonance Imaging (MRI) and Uterine Fibroid Symptom Quality of Life.

The protocol states that a random‐effects model would be used but this was changed to a fixed effects model after discussion with the statistician.

Contributions of authors

For the full review:

MT: lead author, conceived, designed, coordinated and wrote up the review

LO: co‐author, conceived and designed the review, participated in data extraction, management and interpretation, and in writing up the review

AS: co‐author, participated in data extraction, management and interpretation, and in writing up the review

AR: co‐author, participated in data extraction, management and interpretation, and in writing up the review

PS: co‐author, participated in data interpretation and in writing up the review

Sources of support

Internal sources

• IHCAI Foundation, Costa Rica

External sources

• No sources of support supplied

Declarations of interest

Peter Stone is a member of a group of clinicians previously responsible for the registration of mifepristone in New Zealand.

Stable (no update expected for reasons given in 'What's new')