Vaginal progesterone gel for luteal phase support in IVF/ICSI cycles: a meta-analysis Nikolaos P. Polyzos, M.D.,a,b Christina I. Messini, M.D.,a Evangelos G. Papanikolaou, M.D., Ph.D.,c Davide Mauri, M.D.,b Spyridon Tzioras, M.D.,b Ahmed Badawy, M.D., Ph.D.,d and Ioannis E. Messinis, M.D., Ph.D.a a
Department of Obstetrics and Gynecology, University Hospital of Larissa, University of Thessalia, Larissa; b Sections of Obstetrics and Gynecology and Public Health, Panhellenic Association for Continual Medical Research, Athens; and c Assisted Reproduction Unit, Biogenesis Medical Centre, Thessaloniki, Greece; and d Department of Obstetrics and Gynecology, Mansoura University, Mansoura, Egypt
Objective: To investigate whether vaginal progesterone gel may result in similar or higher pregnancy rates compared with all other vaginal progesterone forms when used for luteal-phase support. Design: Meta-analysis of randomized controlled trials using odds ratios (OR) and 95% confidence intervals (CI). Patient(s): Infertile women undergoing IVF or ICSI. Intervention(s): Vaginal progesterone gel 90mg once or twice daily versus any other vaginal progesterone form for luteal phase support. Main Outcome Measure(s): Clinical pregnancy rates. Result(s): Seven randomized controlled trials, involving 2,447 patients, were included in the analysis. No difference was observed in the overall clinical pregnancy rate when comparing vaginal progesterone gel with any other vaginal progesterone form. Moreover, clinical pregnancy rates were similar in protocols using only GnRH agonists and when comparing vaginal gel with the traditional treatment of 200 mg 3 vaginal progesterone capsules. Conclusion(s): This meta-analysis provides solid evidence that no significant difference exists between vaginal gel and all other vaginal progesterone forms in terms of clinical pregnancy rates. (Fertil Steril 2010;94:2083–7. 2010 by American Society for Reproductive Medicine.) Key Words: Luteal phase support, progesterone vaginal gel, infertility, IVF, ART, meta-analysis
Luteal-phase defect has been reported in both GnRH agonist (1) and GnRH antagonist IVF/ICSI cycles (2). For this reason, use of medication for luteal-phase support (LPS) has been considered to be mandatory to ensure intact corpus luteum function and avoid any decrease in implantation and pregnancy rates. Human chorionic gonadotropin has been initially used for LPS and resulted in higher pregnancy rates compared with no treatment; nonetheless, the increased incidence of ovarian hyperstimulation syndrome discouraged its use (3). Progesterone (P) is at the moment the most widely used agent. Whereas IM injections resulted in significantly higher clinical pregnancy rates compared swith vaginal P (4), data from a more recent meta-analysis supported that vaginal P results in pregnancy rates similar to the intramuscular regimen (5). Therefore, taking into consideration the increased side effects of IM regimen and updated results suggesting equivalent pregnancy rates between IM and vaginal P, vaginal route administration appears to be the most appropriate approach at the moment. Various forms of vaginal P have been used for LPS to increase patients’ convenience. Several pioneering randomized controlled trials assessed clinical pregnancy and side effects rates between vaginal gel and all other P regimens. Nonetheless, The Practice Committee Received October 14, 2009; revised December 14, 2009; accepted December 21, 2009; published online February 19, 2010. N.P. has nothing to disclose. C.M. has nothing to disclose. E.P. has nothing to disclose. D.M. has nothing to disclose. S.T. has nothing to disclose. A.B. has nothing to disclose. I.M. has nothing to disclose. Reprint requests: Nikolaos P. Polyzos M.D., Department of Obstetrics and Gynecology, University Hospital of Larissa, University of Thessalia, 22 Papakiriazi Street, 41222 Larissa, Greece (FAX: þ3024210670096; E-mail:
[email protected]).
0015-0282/$36.00 doi:10.1016/j.fertnstert.2009.12.058
of the American Society for Reproductive Medicine in collaboration with the Society for Reproductive Endocrinology and Infertility in 2008 highlighted that even if treatment results with vaginal suppositories or capsules in doses ranging between 200 and 600 mg/day appear similar to 90 mg/day gel, investigations have been limited by relatively small study cohorts (6). We therefore performed a meta-analysis of all available randomized controlled trials comparing vaginal P capsules, suppositories, or pessaries versus vaginal gel as LPS in fresh IVF cycles to investigate any differences regarding clinical pregnancy rates.
MATERIAL AND METHODS Identification of Randomized Trials Two independent investigators searched the Cochrane Central Trials Registry and Medline without year or language restriction in August 2009. Results were compared and a consensus reached with the involvement of a third investigator. We used the searching algorithm (progestins OR progesterone OR progestagens OR crinone OR utrogestan OR utrogest OR cyclogest OR endometrin) AND (IVF OR in vitro fertilization OR icsi OR infertility OR luteal phase) with an array of terms suggestive of randomized controlled trials as recommended by the Cochrane Reviewers Handbook (7). The full strategy is available upon request. In addition, we tried to identify any previous systematic reviews of randomized trials in this field. We scrutinized the references of all eligible trials. Cross-searches were performed in Medline using the names of investigators who were lead authors in at least one eligible trial.
Eligibility Criteria We considered as eligible all randomized controlled trials comparing any form of vaginal P with vaginal P gel in women undergoing IVF or ICSI. We only considered fresh IVF cycles, and therefore trials including
Fertility and Sterility Vol. 94, No. 6, November 2010 Copyright ª2010 American Society for Reproductive Medicine, Published by Elsevier Inc.
2083
frozen-thawed IVF cycles were discarded from the analysis. Whenever patients were randomized to receive two different forms of vaginal P other than gel, the trial was discarded from the analysis.
FIGURE 1 Flowchart of trial selection. RCT ¼ randomized controlled trial.
Data Extraction Two independent investigators extracted the data from all eligible trials. Discrepancies were resolved by the involvement of another investigator. From each eligible trial we recorded for both arms the following items: authors’ names, journal and year of publication, country of origin, number of patients randomized per arm, and patients’ mean age at enrollment. Additionally we recorded the exact regimens used for luteal phase support (dosing and schedule), the IVF protocol used, the number of embryos transfered, and the exact time of the initiation of LPS. Finally, we recorded study design items, including whether there was a description of the mode of randomization, allocation concealment, and blinding.
Outcome Measures The main outcome was the number of clinical pregnancies per arm defined as at least the presence of gestational sac on ultrasound examination 4 weeks after embryo transfer. Secondary outcomes included the multiple gestation rate, miscarriage rate, and side effects rate.
Analysis For each eligible study group, we estimated the odds ratio (OR) for clinical pregnancy between the groups in comparison as well as the 95% confidence interval (CI). We analyzed the data to estimate the OR for achieving a clinical pregnancy when comparing microionized vaginal P gel versus any other vaginal P preparation. Whenever vaginal P gel was compared with two different doses of another vaginal formulation or two different forms other than vaginal gel, these patients’ groups were grouped together and we calculated the OR for the comparison of vaginal P gel versus any other vaginal preparation. In addition, to estimate the effect of P gel compared with the standard treatment modality for LPS, we pooled the ORs separately for the comparison of 90 mg once or twice daily vaginal P gel versus standard LPS treatment of 600 mg daily vaginal P capsules (200 mg 3). Between-study heterogeneity for the OR was evaluated using the c2 (Cochran Q) statistic test (8) and the I2 index (9).We then synthesized the data across studies using fixed-effects (Mantel–Haenszel) or random-effects (DerSimonian and Laird) modeling when between-study heterogeneity was present (8). Analyses were performed in Stata SE 10.0 (Stata Corp., College Station, TX). All P values are two tailed.
RESULTS Characteristics of Eligible Trials The electronic search yielded 2,357 items. After careful scrutiny, eight trials were considered to be potentially eligible (Fig. 1). One trial examined the comparison of vaginal P gel versus vaginal capsules in frozen-thawed cycles and was discarded from the analysis (10). Overall, seven trials involving 2,447 patients were considered to be eligible for our analysis. Four trials compared 90 mg once or twice daily daily dosage of vaginal P gel versus 600 mg daily vaginal P capsules (utrogestan or utrogest) (11–14), one trial versus 200, 400, or 600 mg utrogestan and 400 mg daily vaginal P pessaries (cyclogest) (15), one versus 100 or 200 mg daily vaginal P inserts (endometrin) (16), and one versus 800 mg daily vaginal P pessaries (cyclogest) (17). The median ages of patients enrolled ranged from 30.1 to 34.8 years Table 1. Six of the trials used GnRH agonist protocol for IVF/ICSI cycles (11–13, 15–17) and one either GnRH agonist or antagonist (14). In the majority of the trials, ovarian stimulation protocols included either hMG or rFSH, with the exception of one trial in which the authors did not specify the exact controlled ovarian hyperstimulation protocol used (15).
2084
Polyzos et al.
Polyzos. Vaginal progesterone gel in IVF/ICSI cycles. Fertil Steril 2010.
Four of the trials reported the number of embryos transfered: In three of them fewer than three embryos were transfered (11, 13, 16), and in one investigators might have transfered up to four day 2 or day 3 embryos according to the patients’ age (12). Finally, the day of the initiation of LPS was the day of the embryo transfer in three of the trials (11, 15, 17), 1 day after the oocyte retrieval in two trials (12, 16), and 1 day before ET in one trial (14).
Methodologic Quality of Eligible Trials Three of the trials provided an adequate randomization mode (11, 14, 17), and four an adequate mode for allocation concealment (11, 12, 16, 17). One of the trials was single blinded (assessor blinded) (16), and two of the trials recruited patients from more than one center (11, 16).
Vaginal Gel Versus Any Other Vaginal Preparation Cumulatively, 2,447 patients were randomized to receive either vaginal P gel or any other preparation for LPS in IVF/ICSI cycles. When pooling the data, the OR for achieving a clinical pregnancy was 1.05 (95% CI 0.88–1.25; P¼.575), suggesting that no difference exists (Figure 2).
Vaginal progesterone gel in IVF/ICSI cycles
Vol. 94, No. 6, November 2010
TABLE 1 Baseline characteristics and pregnancy outcomes of eligible trials. Author
Enroll year
Arms
Patients (n)
Age (y)
Stimulation protocol
Doody (16)
2005–2006
3 D3 2 D5
2004–2005
NR
