Increased doses of vaginal progesterone for the prevention of preterm birth in twin pregnancies: a randomised controlled double-blind multicentre trial

Authors


Prof Vicente Serra, Unidad de Medicina Materno-Fetal, Instituto Valenciano de Infertilidad, Universidad de Valencia, Plaza de la Policía Local 3, Valencia 46015, Spain. Email vicente.serra-serra@uv.es

Abstract

Objective  Previous trials have shown little benefit for preventing preterm birth in twin pregnancies using 90–200 mg of daily vaginal natural progesterone. Higher doses have not been tested. Our aim was to determine the efficacy and safety of two different daily doses of vaginal natural progesterone (200 and 400 mg), compared with placebo, for preventing preterm birth in unselected twin pregnancies.

Design  Randomised controlled double-blind multicentre trial (1:1:1).

Setting  The study was carried out in five university centres from Valencia, Murcia and Alicante (Spain).

Population  Women with dichorionic diamniotic twin pregnancies.

Methods  The women self-inserted two vaginal pessaries daily, containing placebo (n = 96), 200 mg of natural progesterone (n = 97) or 400 mg of natural progesterone (n = 97), from 20 to 34 weeks of gestation or delivery. Randomisation was performed by an external centre. Data were analysed on an intention-to-treat basis.

Main outcome measure  Preterm birth rate.

Results  The baseline characteristics for placebo and progesterone groups were similar. Comparison of the three groups and analysis of progesterone-treated versus untreated women showed similar pregnancy and neonatal outcomes. The proportion of preterm and very preterm births, low birthweight, perinatal mortality and neonatal morbidity showed no differences between the three groups. Similar results were also obtained when comparing the 200- versus 400-mg progesterone groups. No serious adverse effects were encountered.

Conclusions  Vaginal progesterone therapy was generally well tolerated, but failed to prevent preterm births in unselected dichorionic diamniotic twin pregnancies. The 400-mg progesterone dose offered no advantages over the 200-mg regimen.

Introduction

Multiple pregnancies are becoming increasingly common because of the expanded use of fertility treatments and older maternal ages at childbirth. The incidence of preterm delivery is about 7–11% of all pregnant women, and around 50% in twin pregnancies.1–3 Preterm birth remains a major cause of perinatal mortality and morbidity. Despite the advances in neonatal care during the last decades, preterm birth remains the major cause of handicaps in children born without congenital anomalies.

Treatment of preterm labour with tocolysis has not been successful in improving infant outcome. The administration of progestogens has been proposed as a strategy to prevent preterm birth.4 Encouraging results with progestogen supplementation have been demonstrated in women with singleton pregnancies at high risk of preterm delivery,5–7 particularly women with a midtrimester short cervix,5,7 or in women with a documented history of a previous spontaneous birth at <34 weeks of gestation.8,9

In contrast, recent trials have not shown any significant effect of progestogens on the rate of preterm delivery in unselected multiple pregnancies.10–14 Two different progestogens have been investigated in twin pregnancies: intramuscular 17 α-hydroxyprogesterone caproate (17P) and vaginal micronised natural progesterone. Although the dosage of the former is quite uniform (i.e. weekly 250 mg 17P im injections),10,13–16 different daily formulations of the latter have been proposed: 90 mg natural progesterone vaginal gel,11 100 mg natural progesterone vaginal pessary,17 and 200 mg natural progesterone vaginal pessary.5,18 To the best of our knowledge, higher doses of vaginal natural progesterone have not been tested in twin pregnancies.

We designed this randomised controlled double-blind trial (RCT) to test the hypothesis that twin pregnancies might need increased doses of vaginal natural progesterone to exert a protective effect on the known high risk of preterm birth. The aim of the trial was to determine the efficacy and safety of two different daily doses of vaginal natural progesterone (200 and 400 mg), compared with placebo, for preventing preterm birth in unselected twin pregnancies.

Methods

Trial design

This was a multicentre (over five sites), controlled, double-blind, phase-III trial, with balanced randomisation into three parallel groups (allocation ratio 1:1:1).

Primary and secondary outcomes

The primary clinical outcome measure was the preterm birth rate (<37 weeks of gestation). Secondary clinical outcome measures were: early preterm birth rate (<32 weeks of gestation); need for tocolytic treatment; rate of preterm premature rupture of membranes (<37 weeks of gestation); cervical length measurements at 20, 24 and 28 weeks of gestation; and perinatal mortality and morbidity. The safety outcome measures were the local tolerance to the treatment and the number of serious systemic adverse effects.

Inclusion and exclusion criteria

Inclusion criteria were: maternal age ≥ 18 years; dichorionic diamniotic twin pregnancy diagnosed by ultrasound; and written informed consent.

Exclusion criteria were: singleton pregnancies; monochorionic twin pregnancies; triplets or higher order multiple pregnancies; elective cervical cerclage prior to 14 weeks of gestation; history of hepatic problems or gestational cholestasis; abnormal liver enzymes; abnormal kidney function; local allergy to micronised natural progesterone; allergy to peanuts (because of the excipient used in the vaginal pessaries); recurrent vaginal bleeding; recurrent vaginal infections; fetal anomalies diagnosed by ultrasound; alcohol or illicit drug consumption; and smoking ≥ 10 cigarrettes/day.

Sample size estimation

The only work that had explored the use of vaginal progesterone for preventing preterm birth at the time at which we designed the study was that of da Fonseca et al.7 Based on that work, the sample size was calculated assuming a 50% reduction of the preterm birth rate, with an α error of 0.05 and a β error of 20% (i.e. with a power of 80%), and with maximum follow-up losses of 25%. This meant that we needed a minimum of 82 women with complete follow-up in each group.

Ethics

The protocol of the study was approved by the local ethics committees of each of the five participating centres. Signed informed consent was obtained from all participating women at the time of recruitment. The study was registered with the European Clinical trials database (EudraCT 2004-004136-31). We followed the CONSORT statement for reporting the results of the study.19

Randomisation

Participants were individually randomised to one of the three parallel groups (placebo, 200 mg progesterone and 400 mg progesterone). An external monitoring centre (Investiganet, Madrid, Spain) provided a randomisation code number for each pregnant woman with the corresponding medication packs. Randomisation was performed by computer (SPSS Random Number Generator; SPSS Inc., Chicago, IL, USA), using a randomisation sequence 1:1:1 ratio (blocks of nine, with no stratification). All study personnel and participants were blinded to treatment assignment for the duration of the trial.

Participants

Eligible women were recruited at the 11–13 weeks visit. If they had been treated with vaginal progesterone during the first trimester, the treatment was stopped at that visit. Randomisation and onset of the allocated treatment was performed at 20 weeks of gestation, after the fetal anomaly scan. As we had decided to analyse the data on an intention-to-treat basis, a total of 290 women with complete follow-up were finally included in the study (Figure 1).

Figure 1.

 Randomisation flow chart. *Reasons for discontinuing the study: woman’s withdrawal (n = 22); vaginal spotting or bleeding (n = 5); repeated vaginal candidiasis (n = 3); or local intolerance to treatment pessaries (n = 1).

Our study population was recruited and followed-up in five university centres: 94 women in the University Hospital Arrixaca (Murcia, Spain); 90 in the Valencian University Institute of Infertility (Valencia, Spain); 50 in the University Hospital La Fe (Valencia, Spain); 38 in the University Hospital Dr. Peset (Valencia, Spain); and 18 in the University General Hospital (Alicante, Spain). Recruitment started in December 2005 and ended in January 2008. Women were followed-up until delivery and neonates until they reached the age of 12 months.

Interventions

Women self-inserted daily, at bedtime, two vaginal pessaries containing placebo, a total of 200 mg of micronised natural progesterone or a total of 400 mg of micronised natural progesterone (Figure 1), from 20 to 34 weeks of gestation or delivery.

Vaginal pessaries were specially manufactured for this RCT (Effik Laboratories, Madrid, Spain) under the supervision of the external monitoring centre. The two daily vaginal pessaries (one blue and the other one white) had identical external appearance and weight in all three groups, but with different fillings: two placebo pessaries (placebo group), one placebo pessary and one progeffik® 200 pessary (200-mg progesterone group), and two progeffik® 200 pessaries (400-mg progesterone group). The placebo pessaries were filled only with the excipient of the progeffik® pessaries: peanut oil and soy lecithin.

Follow-up visits of the trial were performed at 24, 28, 32 and 34 weeks of gestation. The medication was given at each visit by the hospital pharmacy department, and the pregnant woman was instructed to return the empty blister packs at each follow-up visit to check for treatment compliance. The whole trial was monitored by the same external centre (Investiganet). Treatment compliance was defined as the proportion of days that each woman spent under the allocated treatment from the total period prescribed (e.g. from 20 to 34 weeks of gestation or delivery). The investigators were obliged to report to the external monitoring centre any adverse reaction or potential side effect that occurred during the trial. The assigned treatment was discontinued if women had any of the following complications before 34 weeks of gestation: a bleeding placenta praevia; a premature rupture of the membranes; or intrahepatic cholestasis. It was previously established that the trial would stop if any severe adverse effect occurred. A severe adverse effect was defined as any untoward medical occurrence that at any dose results in death, is life-threatening, requires inpatient hospitalisation or the prolongation of existing hospitalization, results in persistent or significant disability/incapacity, is a congenital anomaly/birth defect, or requires intervention to prevent permanent impairment or damage.

Cervical length was routinely measured by transvaginal ultrasound at 20, 24 and 28 weeks of gestation. Although not included in the trial protocol, two of the investigators (VS and AlP) also studied uterine and umbilical artery Doppler waveforms at 20, 24, 28 and 32 weeks of gestation in a subgroup of 86 women (placebo group, n = 30; 200-mg progesterone group, n = 25; and 400-mg progesterone group: n = 31). Women were managed according to routine clinical protocols of each participating centre. The use of other progestogens was not permitted throughout the trial.

Perinatal outcomes were compared between groups. Short-term neonatal morbidity included: respiratory distress syndrome; pneumonia; early-onset sepsis; seizures; grade III–IV intraventricular haemorrhage; stage II–III necrotising enterocolitis; and/or patent ductus arteriosus. Long-term neonatal morbidity included: broncho-pulmonary dysplasia; periventricular leukomalacia; and/or severe retinopathy of prematurity.

Statistical methods

Statistical analysis was performed on an intention-to-treat basis. All analyses were conducted using spss 15 (SPSS Inc). Numerical data are shown as means ± standard deviations. Comparisons between the three groups were performed using an anova test, followed by analysis of the least significant difference, where appropriate. Comparison between progesterone-treated and untreated women was performed using an unpaired Student’s t-test. Categorical data were compared using a chi-square analysis with Yates’ correction and Fisher’s exact test, where appropriate. A Kaplan–Meier survival analysis was performed to evaluate the rate of undelivered women throughout gestation in each group. P < 0.05 was considered as statistically significant.

Results

Baseline epidemiological and clinical characteristics of the study subjects were similar between groups (Table 1). No cases of abnormal uterine or umbilical Doppler waveforms were found at the onset of the trial in any of the women studied.

Table 1. Baseline characteristics of the study subjects
 Placebo (n = 96)200 mg of progesterone (n = 97)400 mg of progesterone (n = 97)Total progesterone (n = 194)
  1. Results expressed as means ± standard deviations or n (%).

  2. Comparison between groups: P = NS.

  3. Data given in parenthesis are expressed as percentage.

  4. *Women who smoked <10 cigarrettes/day.

  5. **Doppler analysis performed only in a subgroup of 86 women.

Maternal age (years)33.3 ± 5.233.5 ± 4.633.5 ± 4.033.5 ± 4.38
Pre-pregnancy body-mass index (kg/m2)25.6 ± 5.025.7 ± 4.225.3 ± 4.025.5 ± 4.2
Nulliparity75 (78.1)68 (70.1)79 (81.4)147 (75.8)
Natural conception4 (4.2)3 (3.1)1 (1.0)4 (2.1)
Ovarian stimulation32 (33.3)35 (36.1)42 (43.3)77 (39.7)
In vitro fertilisation (IVF)/intracytoplasmic sperm injection (ICSI) 60 (62.5)59 (60.8)54 (55.7)113 (58.3)
Previous preterm delivery1 (1.0)1 (1.0)1 (1.0)2 (1.0)
Smoking during pregnancy*21 (21.9)12 (12.4)18 (18.6)30 (15.5)
Amniocentesis10 (10.4)10 (10.3)3 (3.1)13 (6.7)
Average cervical length at 20 weeks of gestation (mm)43.5 ± 7.642.1 ± 7.742.4 ± 7.942.27 ± 7.8
Cervical length < 25 mm at 20 weeks of gestation1 (1.0)2 (2.1)2 (2.1)4 (2.1)
Average umbilical artery resistance index at 20 weeks of gestation**0.74 ± 0.040.72 ± 0.050.73 ± 0.050.73 ± 0.05
Average uterine artery pulsatility index at 20 weeks of gestation*1.18 ± 0.401.26 ± 0.201.34 ± 0.041.31 ± 0.31

The average treatment compliance was 85.1% in the placebo group, 81.9% in the 200-mg progesterone group and 77.5% in the 400-mg progesterone group. The differences between groups were not statistically significant.

Local tolerance to vaginal progesterone treatment was good. There was only one withdrawal as a result of local intolerance (non-infectious vaginitis with persistent vaginal itching) after 39 days of treatment in the 200-mg progesterone group. No severe adverse effects were found throughout the trial in any of the participating women.

Cervical lengths at 24 and 28 weeks of gestation were similar among the three groups. Moreover, there were no differences between groups in the individual changes of cervical length between 20 and 28 weeks of gestation.

The average gestational age at delivery was 36 weeks of gestation in all groups (Table 2). There were no significant differences between groups in the rate of preterm births or very preterm births. Spontaneous preterm births were also similar in the three groups. Survival analysis of the proportion of women delivered throughout gestation showed no significant differences between groups (Figure 2).

Table 2. Pregnancy outcome according to treatment group
 Placebo (n = 96)200 mg of progesterone (n = 97)400 mg of progesterone (n = 97)Total progesterone (n = 194)
  1. Results expressed as means ± standard deviations or n (%).

  2. Comparison between groups: P = NS.

  3. Data given in parenthesis are expressed as percentage.

  4. *Atosiban, indomethacin or nifedipine, depending on the protocol of each centre.

  5. **Cases with preterm premature rupture of membranes (PROM) occurred between 31 and 36 weeks of gestation.

  6. ***Intrahepatic cholestasis was defined as generalised pruritus (not secondary to skin lesions) with elevated aspartate aminotransferase (AST) & alanine aminotransferase (ALT).

Gestational age at delivery (weeks): 36 ± 2.636 ± 2.236 ± 2.836 ± 0.3
<37 weeks47 (49.0)48 (49.5)44 (45.4)92 (47.4)
<34 weeks13 (13.5)13 (13.4)10 (10.3)23 (11.9)
<32 weeks6 (6.3)3 (3.1)7 (7.2)10 (5.2)
<28 weeks1 (1.0)1 (1.0)3 (3.1)4 (2.1)
Medically indicated deliveries at <37 weeks of gestation 21 (21.9)19 (19.6)16 (16.5)35 (18.0)
Fetal death of a single co-twin 2 (2.1)03 (3.1)3 (1.5)
Tocolytic therapy* 17 (17.7)16 (16.5)17 (17.5)33 (17.0)
Steroid treatment for fetal lung maturation 10 (10.4)17 (17.5)16 (16.5)33 (17.0)
Preterm prelabour rupture of membranes** 3 (3.1)4 (4.1)1 (1.0)5 (2.6)
Hypertensive disorders 3 (3.1)5 (5.2)6 (6.2)11 (5.7)
Gestational diabetes 5 (5.21)2 (2.1)4 (4.12)6 (3.1)
Generalised pruritus 2 (2.1)9 (9.3)7 (7.2)16 (8.2)
Intrahepatic cholestasis*** 01 (1.0)5 (5.2)6 (3.1)
Caesarean delivery 68 (70.8)61 (62.9)59 (60.8)120 (61.9)
Figure 2.

 Survival analysis to evaluate the rate of undelivered women throughout gestation in each group. Kaplan–Meier test, = NS.

The incidence of pregnancy-related complications was also similar in the three groups (Table 2). No cases of chorioamnionitis were detected in any of the three groups; however, there was a non-significant trend towards a higher incidence of generalised pruritus and intrahepatic cholestasis among women treated with progesterone (with this trend being dose-dependent for intrahepatic cholestasis). In the subgroup of women studied, treatment with progesterone did not have any effect on uterine or umbilical circulation at 24 and 28 weeks of gestation.

There were five unexplained fetal deaths of a single co-twin (Table 2). Three of the fetal deaths were the first co-twin and two were the second co-twin. Both fetal deaths in the placebo group occurred at 24 weeks of gestation. The three fetal deaths in the 400-mg progesterone group occurred antepartum, at 24, 26 and 36 weeks of gestation, respectively. We could not find any reason to link the three fetal deaths of a single co-twin that occurred in the 400-mg progesterone group with the progesterone treatment itself.

There was a similar high incidence of caesarean deliveries in all groups (Table 2). There were no differences in the mean birthweight among groups (Table 3). The proportion of low birthweights (<2500 g) and very low birthweights (<1500 g) was also similar between groups. However, the neonatal Rohrer ponderal index (weight/height3) was slightly lower in the 400-mg progesterone group compared with the other two groups (< 0.05). There were no differences between groups in the 5-minute Apgar score or in the umbilical artery pH at birth. Progesterone-treated and untreated babies had similar neonatal morbidity. No differences were found between groups in the duration of the stay in the neonatal intensive care unit and the number of days requiring mechanical ventilation.

Table 3. Neonatal outcome according to treatment group
 Placebo (n = 190)200 mg progesterone (n = 194)400 mg progesterone (n = 191)Total progesterone (n = 385)
  1. Results expressed as means ± standard deviations or n (%).

  2. Comparison between groups: P = NS, except for *P < 0.05.

  3. Data given in parenthesis are expressed as percentage.

  4. **Umbilical artery pH at delivery was only available in 107 neonates from the placebo group, 106 from the 200-mg progesterone group and 108 from the 400-mg progesterone group.

Birthweight (g) 2322.4 ± 502.22396.3 ± 463.62335.9 ± 555.52366.1 ± 511.9
<2500 g117 (61.6)104 (53.6)113 (59.2)217 (56.4)
<1500 g13 (6.8)9 (4.6)13 (6.8)22 (5.7)
<10th centile14 (7.4)13 (6.7)21 (11.0)34 (8.8)
Neonatal Rohrer ponderal index (g/cm3) 2.40 ± 0.332.41 ± 2.62.33 ± 0.28*2.37 ± 0.26
5-minute Apgar score < 7 5 (2.6)1 (0.5)5 (2.6)6 (1.6)
Umbilical artery pH < 7.20** 7 (3.7)7 (3.6)12 (6.3)19 (4.9)
Major congenital malformation 7 (3.7)2 (1.0)3 (1.6)5 (1.3)
Admission to neonatal intensive care unit 28 (14.7)21 (10.8)19 (9.9)40 (10.4)
Mechanical ventilation 15 (7.9)9 (4.6)13 (6.8)22 (5.7)
Neonatal death 3 (1.6)05 (2.6)5 (1.3)
Short-term neonatal morbidity 27 (14.2)24 (12.4)31 (16.2)55 (14.3)
Long-term neonatal morbidity 2 (1.1)2 (1.0)02 (0.5)

There were six early neonatal deaths and two late neonatal deaths (Table 3). Both co-twins died in the neonatal period in one case from the placebo group and in one case from the 400-mg progesterone group. The remaining neonatal deaths were of a single co-twin. The five neonatal deaths in the 400-mg progesterone group were the result of complications related to extreme prematurity and/or very low birthweight, except for one that occurred in a baby born at 38 weeks of gestation with hypoplastic left heart syndrome.

Discussion

In singleton pregnancies, progestogen supplements have only helped to prevent preterm deliveries and improved neonatal outcomes in high-risk women (i.e. in women who have a short mid-trimester cervix or experienced a previous early preterm birth).5,6,15 For this reason, it was presumed that twin pregnancies could also benefit from this prophylactic measure, in view of the known high incidence of preterm birth in this group of women.1–3 However, previous trials using different progestogens in multiple pregnancies have failed to decrease the incidence of preterm birth in these women.10–14,16 Our trial aimed to explore the possibility of using a higher dose of natural micronised vaginal progesterone (400 mg/day) than those previously administered (90–200 mg/day),5,11,17,18 but we also failed to demonstrate any benefit in unselected twin pregnancies.

We have to admit that our study has some limitations.

  •  Our sample size estimation was based on the best available evidence at that time, which came from an early study performed in singleton pregnancies that reported a 50% reduction in the preterm birth rate using natural progesterone.7 Subsequent trials of natural progesterone in twin pregnancies have not demonstrated such a benefit.11
  •  We have studied an unselected population of twin pregnancies assuming that all of them are a high-risk group for preterm birth; however, this is probably incorrect because an important percentage of multiple preterm births are not spontaneous, but are medically indicated for other maternal or fetal reasons.3
  •  Our study population included a very low proportion of women with other risk factors for preterm birth (i.e. previous history of preterm birth or mid-trimester short cervix).
  •  When we designed the trial, we decided to analyse the data on an intention-to-treat basis. Although there were no differences in the treatment compliance rate between the three groups studied, a small percentage of women in each group did not complete the study (= NS).

The side effects of progestogens are poorly reported, except that intramuscular 17P appears to raise the risk for gestational diabetes,20 and oral natural progesterone, given in very high doses (600–1200 mg/day), may induce liver dysfunction.21,22 In our trial, no serious systemic adverse effects were encountered, the local tolerance was good in all but one woman and the rate of gestational problems was not higher in progesterone-treated women, except for a small increase in the incidence of generalised pruritus and a dose-dependent, non-significant, trend towards a higher incidence of intrahepatic cholestasis. There was also a slightly lower neonatal ponderal index in the 400-mg progesterone group (Table 2). Although it is known that progesterone down-regulates insulin-like growth factor 1 (IGF-1) mRNA and protein expression in cultured uterine leiomyoma cells,23 it seems unlikely that this subtle effect of 400 mg of progesterone on the fetal growth rate of twins could be attributed to a potential IGF-1 down-regulation, because the proportion of neonates with birthweights <2500 g and <1500 g in the 400-mg progesterone group was similar to the other two groups studied.

In our trial, the stillbirth and neonatal mortality rates were not increased in progesterone-treated women. These rates were similar to those described previously.24–26 We could not find any relationship between the three fetal deaths of a single co-twin in our series and the progesterone treatment itself.

Although the collective results of previous studies indicate that prophylactic progestogens neither reduce the rate of preterm birth nor improve neonatal outcome in unselected twin pregnancies,10–13,16,26,27 it might be possible that progestogens may benefit a subset of women with a twin pregnancy and some other risk factors for preterm birth. Three secondary analyses of randomised trials showed that intramuscular 17P and vaginal natural progesterone did not decrease the risk of preterm birth in twin pregnancies with either a short cervix or a previous history of preterm birth.5,15,28 However, a recent meta-analysis showed that vaginal progesterone was associated with a non-significant reduction in the rate of preterm birth, but effectively reduced the composite neonatal morbidity/mortality of women with twin gestations and a short cervix.29 Nevertheless, it is difficult to draw conclusions because the number of women with twin pregnancies analysed in all of these studies was relatively small and the definition of a short cervix was not uniform.5,15,28,29 In our trial, we had a very low incidence of women with a previous preterm delivery and/or a short cervix (Table 1).

The reasons why progestogen supplementation is effective for decreasing the preterm birth rate in high-risk singleton pregnancies,5,28,29 but has no effect in high-risk twin pregnancies,15,28,29 remains unknown. Although it has been suggested that the mechanisms underlying the parturition process in twins may differ from those in singletons,28 this has not yet been proven. The different sensitivities to progestogen supplementation depending on the risk factors also suggest that there may be more than one pathway leading to spontaneous preterm labour. Potential influencing factors could be the increased distension of the uterine muscle or the different ratio of placental/myometrial surface area in twins,30 compared with singleton pregnancies. However, these hypotheses only highlight that many details remain missing about the process of human parturition.31

In summary, comparison of the three groups studied (placebo versus 200 mg versus 400 mg of progesterone) and analysis of progesterone-treated versus untreated women showed similar pregnancy and neonatal outcomes. Although both progesterone doses were generally well tolerated, the 400-mg progesterone regimen had a subtle increase in side effects, and did not offer any advantage over the 200-mg progesterone dose. Both progesterone doses failed to prevent preterm birth in unselected dichorionic diamniotic twin pregnancies. We agree that current evidence does not support the routine use of progesterone in women with twin pregnancies.8

Disclosure of interests

All authors confirm no conflicts of interest with regards to the data reported here.

Contribution to authorship

VS conceived the idea, performed the literature search, designed the study, applied for ethical approval and trial registration, coordinated the whole trial, recruited women, collected data, assisted with manuscript preparation and authored the article. AlP coordinated the trial in his centre, performed the literature search, applied for ethical approval, recruited women, collected data, assisted with statistical analysis and manuscript preparation, and co-authored the article. JM, CL, JB, RG and MS recruited women, collected data, and co-authored the article. JJP coordinated the trial in his centre, applied for ethical approval, and co-authored and reviewed the article. IA coordinated the trial in her centre, recruited women, collected data and co-authored the article. JCME coordinated the trial in his centre, applied for ethical approval, collected data and co-authored the article. AnP conceived the idea, designed the study, and co-authored and reviewed the article.

Details of ethics approval

Ethical approval was granted by the Valencian University Institute of Infertility Clinical Research Ethics Committee (reference number: VLC-VS-0405-507-23) on 4 July 2005, after obtaining the approval of the local ethics committees of each of the five participating centres. Clinical trial authorisation was granted by the European Medicines and Healthcare Products Regulatory Agency (EudraCT 2004-004136-31). The trial was registered in this European database (https://www.clinicaltrialsregister.eu/ctr-search/search?query=eudract_number%3A2004-004136-31), and with a few months delay was also registered in ClinicalTrials.gov (NCT00480402). Trial registry numbers: EudraCT, 2004-004136-31; ClinicalTrials.gov, NCT00480402.

Funding

The trial was funded by grant EF489-2004/1 from Laboratorios Effik S.A. (Madrid, Spain).

Acknowledgements

We would like to thank the clinical staff of the five participating centres for clinical assistance during the trial. We also thank Dr Alfredo Perales-Puchalt (University Hospital La Fe, Valencia, Spain) for statistical assistance.

Ancillary