- Email: [email protected]
Progesterone luteal support after ovulation induction and intrauterine insemination: an updated systematic review and meta-analysis
ORIGINAL ARTICLE: INFERTILITY
Progesterone luteal support after ovulation induction and intrauterine insemination: an updated systematic review and meta-analysis Katherine A. Green, M.D.,a Jessica R. Zolton, D.O.,a Sophia M. V. Schermerhorn, B.S.,b Terrence D. Lewis, M.D., Ph.D.,a,d Mae W. Healy, D.O.,a,d Nancy Terry, M.L.S.,c Alan H. DeCherney, M.D.,a and Micah J. Hill, D.O.a,d a Program in Reproductive and Adult Endocrinology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland; b Uniformed Services University of the Health Sciences, Bethesda, Maryland; c National Institutes of Health Library, Bethesda, Maryland; and d Walter Reed National Military Medical Center, Bethesda, Maryland
Objective: To evaluate the effect of progesterone (P) for luteal phase support after ovulation induction (OI) and intrauterine insemination (IUI). Design: An updated systematic review and meta-analysis. Setting: Not applicable. Patient(s): Patients undergoing OI-IUI for infertility. Intervention(s): Exogenous P luteal support after OI-IUI. Main Outcome Measure(s): Live birth. Result(s): Eleven trials were identified that met inclusion criteria and constituted 2,842 patients undergoing 4,065 cycles, more than doubling the sample size from the previous meta-analysis. In patients receiving gonadotropins for OI, clinical pregnancy (relative risk [RR] 1.56, 95% confidence interval [CI] 1.21–2.02) and live birth (RR 1.77, 95% CI 1.30–2.42) were more likely in P supplemented patients. These findings persisted in analysis of live birth per IUI cycle (RR 1.59, 95% CI 1.24–2.04). There were no data on live birth in clomiphene citrate or clomiphene plus gonadotropin cycles. There was no benefit on clinical pregnancy with P support for patients who underwent OI with clomiphene (RR 0.85, 95% CI 0.52–1.41) or clomiphene plus gonadotropins (RR 1.26, 95% CI 0.90–1.76). Conclusion(s): Progesterone luteal phase support is beneficial to patients undergoing ovulation induction with gonadotropins in IUI cycles. The number needed to treat is 11 patients to have one additional live birth. Progesterone support did not benefit patients undergoing ovulation induction with clomiphene citrate or clomiphene plus gonadotropins. (Fertil SterilÒ 2017;-:-–-. Ó2017 by American Society for Reproductive Medicine.) Key Words: Intrauterine insemination, live birth, luteal support, ovulation induction, progesterone Discuss: You can discuss this article with its authors and with other ASRM members at https://www.fertstertdialog.com/users/ 16110-fertility-and-sterility/posts/14467-23524
Received December 6, 2016; revised January 4, 2017; accepted January 16, 2017. K.A.G. has nothing to disclose. J.R.Z. has nothing to disclose. S.M.V.S. has nothing to disclose. T.D.L. has nothing to disclose. M.W.H. has nothing to disclose. N.T. has nothing to disclose. A.H.D. has nothing to disclose. M.J.H. has nothing to disclose. Supported, in part, by the Program in Reproductive and Adult Endocrinology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Bethesda, MD. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of Defense or the US Government. Reprint requests: Katherine A. Green, M.D., National Institutes of Health, Eunice Kennedy Shriver National Institute of Child Health and Human Development, Program in Reproductive and Adult Endocrinology, 10 CRC, Room 1E-3140, 10 Center Drive, MSC 1109, Bethesda, MD 20892-1109 (E-mail: [email protected]). Fertility and Sterility® Vol. -, No. -, - 2017 0015-0282/$36.00 Copyright ©2017 Published by Elsevier Inc. on behalf of the American Society for Reproductive Medicine http://dx.doi.org/10.1016/j.fertnstert.2017.01.011 VOL. - NO. - / - 2017
S
uccessful implantation requires synchrony between a competent blastocyst and a receptive secretory phase endometrium (1). As estrogen rises during the follicular phase and a dominant follicle emerges, the mid-cycle LH (2) peaks and ovulation occurs (2). The pulsatile LH secretion stimulates the corpus luteum to produce P, which induces endometrial secretory transformation and promotes receptivity (3). Fertility treatments may interfere with the luteal phase via several mechanisms. Ovulation induction (OI) may 1
ORIGINAL ARTICLE: INFERTILITY result in a premature rise of P and alterations in endometrial receptivity (4, 5). Furthermore, supraphysiologic E2 elevation from ovarian stimulation may cause pituitary downregulation and alterations in luteal phase LH secretion (6, 7). Supraphysiologic E2 levels are often associated with multifollicular development during assisted reproductive technology (ART) (8), whereas only one to two dominant follicles may be achieved during OI and IUI. The use of GnRH analogs for pituitary down-regulation and mechanical disruption of follicles during oocyte aspiration may further affect luteal function (9, 10). Luteal support with exogenous P after ART is routine because it is associated with higher pregnancy and live birth rates (11, 12); however, there is no consensus on the use of P after OI-IUI. In 2013 our group published a systematic review and meta-analysis of randomized, controlled trials (RCTs) to evaluate the effect of luteal phase P support after OI-IUI (13). This analysis concluded that luteal phase support was beneficial in gonadotropin IUI cycles but not in clomiphene citrate (CC) IUI cycles. The endogenous rise of LH as a result of CC may provide further stimulation for the developing corpus luteum, which in turn will favorably impact the luteal phase (14). In contrast, stimulation with gonadotropins directly may result in negative feedback of E2 at the hypothalamus and decreased LH release, similar to ART cycles (15). In the past 3 years there have been several new RCTs published on this topic. The total number of subjects and papers has doubled since the prior review, necessitating new statistical analysis of the published data. The goal of this study was to review these recent publications and perform an updated meta-analysis to determine the impact of P supplementation after OI-IUI.
MATERIALS AND METHODS Study Design This is an updated systematic review and meta-analysis of RCTs evaluating exogenous P luteal support during the luteal phase after OI-IUI (13). The objective of this study was to identify eligible RCTs that have been published since our previous meta-analysis in 2013 and update the systematic review and meta-analysis.
Literature Search PubMed and Embase literature searches were performed for published RCTs evaluating P luteal supplementation vs. no luteal support after OI-IUI. Our previous meta-analysis conducted a literature search through January 8, 2013 and identified five studies meeting inclusion criteria (13). An updated literature search was performed, limited to publications from January 1, 2013 to the date of search execution, which occurred on September 12, 2016. The search used specific key words and database indexing terminology (available online as Supplemental Addendum).
Study Selection Study selection was performed according to the same criteria as in the previous meta-analysis (13). Only published RCTs 2
that compared exogenous P during the luteal phase after OI-IUI vs. no P were included. Any type of ovulation induction was allowed, including CC, exogenous gonadotropins, hCG, aromatase inhibitors, or a combination of these medications. All types of exogenous P were permitted, including oral, IM, or vaginal formulations. Publication in any language was allowed. Exclusion criteria included nonrandomization, timed intercourse cycles, natural cycles, publication as abstract only, book chapters, or review articles. The literature search identified 31 new publications, which were independently reviewed by three investigators (K.A.G., J.R.Z., and M.J.H.) to identify eligible studies. Of the 31 abstracts reviewed, 25 studies were excluded on the basis of abstract data indicating failure to meet inclusion criteria. Six full-text articles were reviewed for inclusion and exclusion criteria, all of which met inclusion criteria. There were no disagreements among the three reviewing investigators regarding the studies eligible for inclusion. Study quality and the potential for bias within each study was evaluated, considering randomization method, concealment of allocation, blinding of providers and patients, and flow of patients through the randomization, treatment, and outcome stages.
Data Collection Data from studies that met inclusion criteria were extracted independently by two investigators (K.A.G. and J.R.Z.). Clinical pregnancy and live birth data were extracted from intentto-treat results. When intent-to-treat results were not reported, data were calculated as intent-to-treat by making the denominator the number of patients enrolled, instead of the reported per-protocol results. Continuous data were extracted in the form of mean and SD. Additional extracted data included author, year of publication, journal, country of origin, randomization method, sample size, number of patients randomized, number of cycles performed, method of ovulation induction, type of P support, duration of P support, method of ovulation triggering, trial registry, and conflicts of interest. The primary outcome was live birth per patient. Secondary outcomes included clinical pregnancy per patient, clinical pregnancy per cycle, and live birth per cycle. Sensitivity analyses were performed excluding trials that allowed individual subjects to cross over into both treatment and control arms over multiple cycles.
Data Synthesis Data for synthesis were obtained from intent-to-treat results when reported. Primary analyses were performed using perpatient data, and additional analyses were performed using per-cycle data. Heterogeneity was evaluated using the Q test and I2 index values and reported for each outcome as P value and percentage, respectively. Random-effects models were used when studies had clinical heterogeneity in the ovulation induction method (i.e., medication type) used or when the I2 index was >50% (16). A fixed-effect model was used when the same type of ovulation induction method was used in all studies and the I2 index was <50%. Sensitivity analyses VOL. - NO. - / - 2017
VOL. - NO. - / - 2017
TABLE 1 Study characteristics of trials meeting inclusion in the systematic review. Country of study
Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahimi et al. 2010 (27) Maher 2011 (25)
Turkey
Unexplained infertility
75 IU FSH starting cycle day 3
Belgium
Unexplained infertility (implicit) Unexplained infertility
50 mg CC cycle days 3–7
Agha-Hosseini et al. 2012 (26)
Iran
Aali et al. 2013 (17)
Iran
Romero Nieto et al. 2014 (19)
Spain
Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18)
Turkey
Karadag et al. 2016 (21)
Turkey
Peeraer et al. 2016 (22)
Belgium
Iran Egypt
Iran
Patients
Unexplained Infertility (implicit), anovulatory infertility, at least one patent fallopian tube, post-wash total motile sperm R1 106 Unexplained infertility (implicit) Age <40 y, FSH<12 IU/L, sperm concentration R10 106/mL Unexplained infertility (implicit), post-wash total motile sperm R3 106 Unexplained infertility Unexplained or anovulatory infertility, at least one patent fallopian tube, mild male factor Unexplained infertility
Unexplained infertility, minimal-mild endometriosis, post-wash total motile sperm R5 106
Green. Progesterone luteal support after IUI. Fertil Steril 2017.
Ovarian stimulation
50 mg CC BID cycle days 3–7 plus 75 IU hMG cycle days 7–9 75 IU FSH starting days 2–5
Either 50 mg CC BID or letrozole 5 mg daily cycle days 3–7, with or without 75 IU hMG days 3–7 75–150 IU hMG starting on cycle day 3 or 100 mg CC on cycle days 3–7 plus hMG on days 8 and 9 Urinary or recombinant gonadotropins starting on cycle days 2–4
P supplementation Crinone 8% post-IUI day 2 through 12 wk Utrogestan 200 mg in 3 separate doses post-IUI day 1 through 7 wk Cyclogest 400 mg daily post-IUI day 2 through 10 wk Crinone 8% post-IUI day 1 through 14 d
Ovulation triggering
No. of cycles allowed
10,000 U hCG
Up to 3
5,000 U hCG
1
5,000 U hCG
Up to 3
10,000 U hCG
Up to 6, patients alternated treatment groups each cycle
Cyclogest 400 mg daily post-IUI day 1 through 14 d
10,000 U hCG
1
Cyclogest 400 mg daily post-IUI day 1 through 10 d
10,000 U hCG
1
Progesterone 200 mg vaginally daily post-IUI day 1 until 8 wk
6,500 IU recombinant hCG (rhCG) (Ovitrelle) 10,000 IU hCG
Up to 4, patients re-randomized each cycle
10,000 U hCG
1
75 IU FSH starting on cycle day 3; dose modified based on response 100 mg CC on cycle days 3–7 plus 75 IU hMG on cycle days 7–10 (longer based on response)
Crinone 8% on day of IUI until 12 wk Progesterone 400 mg vaginally BID post-IUI day 2 through 8 wk
Up to 3
Arm 1) 100 mg CC on cycle days 3–7 Arm 2) 75 IU FSH starting on cycle days 2–4 37.5–75 IU FSH starting on cycle days 2–3; increased by 37 IU on cycle day 7–10 if no response
Crinone 8% post-IUI day 1 until 10 wk
6,500 IU rhCG (Ovitrelle)
1
Crinone 8% post-IUI day 1 for 15 d
rhCG (Ovitrelle; dose not specified)
1
3
Fertility and Sterility®
Authors (reference)
ORIGINAL ARTICLE: INFERTILITY using random-effects models were performed in all cases in which fixed-effect models were the primary analysis, to ensure that the model type did not change the outcome. In all cases, random-effects models gave similar results to the fixed-effect models without altering the statistical significance. Risk differences between groups were calculated using Mantel-Haenszel estimates. When significant risk differences existed between groups, numbers needed to treat (NNTs) to have one additional live birth were calculated from the Mantel-Haenszel risk difference estimates. Bias was assessed at the study level using a qualitative review assessing randomization, concealment, blinding, and patient flow. Funnel plots were constructed to assess publication bias at the outcome level. Dichotomous outcome data were reported as relative risks (RRs) with 95% confidence intervals (CIs). Continuous data were synthesized using weighted means with 95% CIs. A priori subgroup analyses were planned to compare the types of P support and methods of ovulation induction. Sensitivity analyses were planned a priori if included trials were found to be at high risk of bias at either the study or outcome level. Exclusion sensitivity analysis was performed on each trial to determine whether its inclusion changed the statistical findings of each model. In no case did exclusion of any study change the findings. Data collection was performed in Excel (Microsoft Office 2007), and statistical analysis was performed using Mix 2.0 Pro (Bax L: MIX 2.0. Professional software for metaanalysis in Excel version 2.0.1.4. BiostatXL, 2011; http:// www.meta-analysis-made-easy.com).
RESULTS Studies Included for Systematic Review and MetaAnalysis A total of 31 abstracts were identified, and six full-text articles were reviewed, all of which met full inclusion criteria (17–22). By combining these six trials with the five studies previously identified in our 2013 meta-analysis (13), this updated meta-analysis of 11 RCTs comprised 2,842 patients and 4,065 OI-IUI cycles (17–27) (Supplemental Fig. 1). All 11 trials described inclusion criteria consistent with unexplained infertility. Specific inclusion criteria and study characteristics are outlined in Table 1. Five studies used exogenous gonadotropins for ovulation induction in all patients (19, 20, 22, 23, 25). Four of these studies specified the use of recombinant FSH for ovulation induction (20, 22, 23, 25), and one study did not list a specific medication beyond describing the use of urinary or recombinant gonadotropins (19). One study described two treatment arms, one of which consisted of patients who received only recombinant FSH and a second arm that included patients who underwent stimulation with CC (21). One trial used either hMG alone or CC plus 2 days of hMG (17). Two additional studies used CC plus hMG for all patients (18, 27). One trial used either CC or letrozole in all patients, with some patients also receiving hMG (26), and one study used only CC (24). Ovulation triggering was performed with either 5,000 IU or 10,000 IU of hCG in eight studies (17,18,20,23–27). Two trials used 6,500 IU of recombinant hCG (19, 21), and one study used 4
recombinant hCG but did not specify dose (22). All studies used vaginal routes of P luteal support, with variation in the type and dose of P as well as therapy duration (Table 1). None of the studies used oral or IM routes of P supplementation.
Assessment of Bias Risk The 11 included trials were individually assessed for risk of bias (Supplemental Table 1). The randomization process was clearly described in all studies. Six trials used computergenerated randomization sequences in a non–crossover design (18, 20, 21, 23, 24, 26). Two trials also used computer-generated randomization sequence; however, patients were randomized to their first cycle of treatment and then either alternated treatment groups in each subsequent treatment cycle (25) or were re-randomized in subsequent treatment cycles (19). These two studies were excluded in a sensitivity analysis removing studies that allow subjects to cross over between study groups. Two studies reported block randomization of patients (17, 22), and one trial reported sequential randomization of patients (27). Three studies described attempts to control for allocation concealment (18, 22, 23). Three trials reported a single-blind study design, with blinding of the providers (17, 22, 23). One trial described a double-blind study design with blinding of both providers and patients and the use of a placebo control (18). None of the studies documented blinding of outcomes data. Nine trials adequately reported on the flow of patients through the study and were at low risk of incomplete data reporting (17,19–26). One study disclosed an author's affiliation with a pharmaceutical company (22). There were no pharmaceutical affiliations or support disclosed in any of the other trials. There were no differences in female age or duration of infertility between the two randomized groups in any of the studies; however, one study did not compare the duration of infertility between the two groups (19) (Supplemental Table 2). One study reported significantly fewer follicles R17 mm on the day of hCG administration in the P-supplemented group compared with the control group (1.2 0.3 vs. 1.3 0.4 follicles, respectively, P¼ .02) (24). A separate study (21) reported a significantly higher mean body mass index for patients in the P-supported group compared with the control group in the gonadotropin OI-IUI arm of the study (24.7 2.4 kg/m2 vs. 23.6 2.3 kg/m2, respectively, P¼ .021). Funnel plot analysis did not provide evidence of any publication bias (Supplemental Fig. 2).
Overall Meta-Analysis Results Ten studies reported clinical pregnancy on a per-patient basis, totaling 2,444 patients undergoing OI-IUI (Table 2) (17,18,20– 27). In random-effects modeling, clinical pregnancy was increased with P support (RR 1.37, 95% CI 1.15–1.64, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .65) and the I2 index (I2 value ¼ 0%, 95% CI 0%–62%). Five studies reported live birth on a per-patient basis, totaling 1,027 patients undergoing OI-IUI (20, 22, 23, 25, VOL. - NO. - / - 2017
Fertility and Sterility®
TABLE 2 Primary pregnancy outcomes in the 1 included randomized, controlled trials reported on a per-patient basis. Authors (reference) Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahimi et al. 2010 (27) Maher 2011 (25) Agha-Hosseini et al. 2012 (26) Aali et al. 2013 (17) Romero Nieto et al. 2014 (outcomes listed per cycle) (19) Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18) Karadag et al. 2016 (21)
Peeraer et al. 2016 (22)
Group P Control P Control P Control P Control P Control P Control P Control P Control P Control P (CC arm) Control (CC arm) P (Gnd arm) Control (Gnd arm) P Control
Patients (n) 109 105 243 225 98 102 37 34 150 150 99 97 398e 398e 71 78 127 126 50 50 50 50 202 191
Cycles (n) 223 204 243 225 252 259 132 126 148 142 99 97 449 444 71 78 127 126 50 50 50 50 202 191
Positive hCG (%) NR NR NR NR NR NR NR NR 28.7 20.7 16.2 14.4 NR NR NR NR 30.8 22.2 10 16 20 18 NR NR
P value NR NR NR NR NR NS NR NR NS NR NR NR
Clinical pregnancy (%) a
39.4 23.8a 7c 8.4c 30.6a 25.5a 54.9a 35.2a 24d 13.3d 12.1a 9.3a 13.8a 11a 22.5a 14.1a 15.8f 12.7f 10d 12d 20d 16d 16.8d 11d
P value .01 NS NS .02 < .05 NS NS NS NS NS NS NS
Live birth (%) b
35.8 18.1b NR NR 19.4b 14.7b 35.2b 9.8b NR NR NR NR 10.2b 8.3b 19.7b 14.1b NR NR NR NR NR NR 14.9g 9.4g
P value .003 NR NS < .001 NR NR NS NS NR NR NR NS
Note: Gnd ¼ gonadotropin; NR ¼ not reported; NS ¼ not significant. a Clinical pregnancy defined as the presence of gestational sac on ultrasound. b Live birth defined as an infant alive 1 week after birth. c Values reported as ongoing pregnancy, defined as pregnancy developing beyond 12 weeks of gestation. d Clinical pregnancy defined as the presence of fetal cardiac activity on ultrasound. e Total number of patients enrolled in study (patients were re-randomized in subsequent cycles). f Clinical pregnancy not defined. g Live birth defined as live birth of a child after 24 weeks of gestation. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
27). In random-effects modeling, live birth was increased with P support (RR 1.76, 95% CI 1.29–2.40, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .73) and the I2 index (I2 value ¼ 0%, 95% CI 0%–84%). All 11 studies reported on clinical pregnancy on a percycle basis, totaling 4,065 OI-IUI cycles (Table 2). In random-effects modeling, clinical pregnancy was increased with P support (RR 1.34, 95% CI 1.15–1.57, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .75) and the I2 index (I2 value ¼ 0%, 95% CI 0%– 60%). Five studies reported live birth on a per-cycle basis without treatment group crossover, totaling 1,950 cycles undergoing OI-IUI (19, 20, 22, 23, 27). In random-effects modeling, live birth was increased with P support (RR 1.49, 95% CI 1.15–1.94, P¼ .002). Minimal heterogeneity was suggested by the results of the Q test (P¼ .59) and the I2 index (I2 value ¼ 0%, 95% CI 0%–79%).
Subgroup Analyses Gonadotropins. Five studies evaluating gonadotropin OI-IUI reported clinical pregnancy on a per-patient basis, totaling 927 patients (Fig. 1A) (20–23,25). In fixed-effect modeling, clinical pregnancy was increased with P support (RR 1.56, 95% CI 1.21–2.02, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .98) and the I2 index VOL. - NO. - / - 2017
(I2 value ¼ 0%, 95% CI 0%–79%). Four studies reported live birth on a per-patient basis, totaling 827 patients undergoing gonadotropin OI-IUI (Fig. 2A) (20, 22, 23, 25). In fixed-effect modeling, live birth was increased with P support (RR 1.77, 95% CI 1.30–2.42, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .73) and the I2 index (I2 value ¼ 0%, 95% CI 0%–84%). Six studies reported on clinical pregnancy per cycle with gonadotropin OI, totaling 2,220 OI-IUI cycles (19–23,25). In fixed-effect modeling, clinical pregnancy was increased with P support (RR 1.44, 95% CI 1.18–1.75, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .93) and the I2 index (I2 value ¼ 0%, 95% CI 0%–74%). Five studies reported live birth on a per-cycle basis, totaling 2,120 gonadotropin OI-IUI cycles (Fig. 2B) (19, 20, 22, 23, 25). In fixed-effect modeling, live birth was increased with P support (RR 1.59, 95% CI 1.24–2.04, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .21) and the I2 index (I2 value ¼ 0%, 95% CI 0%–85%). The risk difference in live birth rate per patient in gonadotropin OI-IUI cycles was 9.7% higher in P-supported patients (risk difference 9.7%, 95% CI 4.7%–14.7%). The number of patients needed to treat with P support in gonadotropin OIIUI cycles is 11 to have 1 additional live birth (NNT 95% CI 7–22 patients). The risk difference in live birth rate per cycle with gonadotropin OI-IUI was 5.4% higher in P-supported 5
ORIGINAL ARTICLE: INFERTILITY
FIGURE 1
A
Gonadotropins clinical pregnancy per patient
Author
Sample size
Measure (CI)
Weight %
P value
Erdem
214
1.66 (1.1; 2.51)
34.47%
0.02
Maher
71
1.61 (0.77; 3.35)
11.29%
0.2
Peeraer
393
1.53 (0.92; 2.54)
29.22%
0.1
Seckin
149
1.6 (0.8; 3.21)
14.19%
0.19
Karadag
100
1.25 (0.54; 2.9)
10.83%
0.6
Synthesis
927
1.56 (1.21; 2.02)
100%
<0.001 0.5
B
Favors no luteal support
Gonadotropins + clomiphene clinical pregnancy per patient
Author
Sample size
Measure (CI)
Weight %
P value
Agha-Hosenni
66
1.54 (0.65; 3.65)
13.28%
0.32
Hossein
253
1.24 (0.67; 2.28)
33.53%
0.49
Ebrahimi
200
1.2 (0.77; 1.88)
53.19%
0.42
Synthesis
519
1.26 (0.9; 1.76)
100%
0.17 0.5
C
Favors no luteal support
Clomiphene clinical pregnancy per patient
Author
Sample size
Measure (CI)
Weight %
P value
Kyrou
468
0.83 (0.44; 1.55)
66.36%
0.56
Agha-Hosenni
38
1 (0.29; 3.43)
13.45%
1
Karadag
100
0.83 (0.27; 2.55)
20.18%
0.75
Synthesis
606
0.85 (0.52; 1.41)
100%
0.53
0.25
1 RR
1 RR
0.5
Favors no luteal support
2
4
Favors luteal support
2
4
Favors luteal support
1 RR
2
4
Favors luteal support
Forrest plot of (A) clinical pregnancy with gonadotropins, (B) clinical pregnancy with CC plus gonadotropins, (C) clinical pregnancy with CC. All data reported on a per-patient basis. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
cycles than nonsupported cycles (risk difference 5.4%, 95% CI 2.7%–8.1%). The number of IUI cycles needed to treat with P support in gonadotropin OI-IUI cycles is 19 to have 1 additional live birth (NNT 95% CI 13–37 cycles). There is level 1 evidence that P luteal support increases live birth in gonadotropin OI-IUI cycles. CC plus gonadotropins. Three studies evaluating CC plus gonadotropin OI-IUI reported clinical pregnancy on a perpatient basis, totaling 519 patients undergoing OI-IUI (Fig. 1B) (18, 26, 27). In fixed-effect modeling, clinical pregnancy was unchanged with P support (RR 1.26, 95% CI 0.90–1.76, P¼ .17). Minimal heterogeneity was suggested by the results of the Q test (P¼ .87) and the I2 index (I2 value ¼ 0%, 95% CI 0%–89%). None of these three studies reported on live birth. Three studies evaluating CC plus gonadotropin OI-IUI reported clinical pregnancy on a per-cycle basis, totaling 830 cycles (18, 26, 27). In fixed-effect modeling, clinical pregnancy was unchanged with P support (RR 1.25, 95% CI 6
0.88–1.77, P¼ .21). Minimal heterogeneity was suggested by the results of the Q test (P¼ .83) and the I2 index (I2 value ¼ 0%, 95% CI 0%–89%). None of these three studies reported on live birth. The combined clinical pregnancy rate per patient was 23.1% in the P-supported group vs. 18.5% in the control groups. If this is the true difference in clinical pregnancy in clomiphene plus gonadotropin OI cycles, 2,502 subjects would be needed to detect an advantage in clinical pregnancy with P support. The current literature lacks the sample size to detect such a difference. There is insufficient evidence that P support is beneficial in clomiphene plus gonadotropin OI-IUI cycles. Letrozole plus gonadotropins. A single study evaluated letrozole plus gonadotropin OI-IUI (26). In 90 patients, the clinical pregnancy rate was 25.5% with P vs. 14.8% without P (P¼ .43). This study did not report on live birth. There is insufficient evidence to evaluate the effect of P luteal support in letrozole plus gonadotropin OI-IUI cycles. VOL. - NO. - / - 2017
Fertility and Sterility®
FIGURE 2
A Gonadotropins live birth per patient Author
Sample size
Measure (CI)
Weight %
P value
Erdem
214
1.98 (1.23; 3.19)
42.46%
0.01
Maher
71
3.06 (0.92; 10.2)
6.71%
0.07
Peeraer
393
1.58 (0.91; 2.73)
32.14%
0.1
Seckin
149
1.4 (0.68; 2.88)
18.68%
0.36
Synthesis
827
1.77 (1.3; 2.42)
100%
<0.001
0.5
B Gonadotropins live birth per cycle
Favors no luteal support
Author
Sample size
Measure (CI)
Weight %
P value
Erdem
427
1.88 (1.12; 3.14)
22.98%
0.02
Maher
258
3.41 (1.53; 7.6)
9.46%
0
Peeraer
393
1.58 (0.91; 2.73)
20.12%
0.1
Seckin
149
1.4 (0.68; 2.88)
11.69%
0.36
Romero Nieto
893
1.23 (0.81; 1.86)
35.75%
0.33
Synthesis
2,120
1.59 (1.24; 2.04)
100%
<0.001
0.5 Favors no luteal support
1 RR
1 RR
2
4
8
16
Favors luteal support
2
4
8
Favors luteal support
Forrest plot of (A) live birth per patient with gonadotropins, and (B) live birth per cycle with gonadotropins. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
Clomiphene citrate. Three studies evaluating CC OI reported clinical pregnancy on a per-patient basis, totaling 606 patients OI-IUI cycles (Fig. 1C) (21, 24, 26). In fixed-effect modeling, clinical pregnancy was unchanged with P support (RR 0.85, 95% CI 0.52–1.41, P¼ .53). Minimal heterogeneity was suggested by the results of the Q test (P¼ .97) and the I2 index (I2 value ¼ 0%, 95% CI 0%–89%). None of the studies on clomiphene reported on live birth outcomes. The combined clinical pregnancy rate per patient was 8.3% in the P-supported group vs. 9.8% in the control groups. If the true clinical pregnancy rate with clomiphene OI-IUI is 10% without P support, 1,372 subjects would be needed to detect an increase in 5% in pregnancy with P support. The current literature lacks the sample size to detect such a difference. Furthermore, these three studies suggest that such a difference in unlikely to exist. There is no evidence that P support is beneficial in CC OI-IUI cycles.
Sensitivity analysis. Sensitivity analyses were performed excluding studies that allowed subjects to cross over from one treatment group to another during the study period. In a fixed-effect model, sensitivity analysis of live birth in gonadotropin cycles of three studies including 765 patients without group cross-over (20, 22, 23), live birth was increased with P support (RR 1.69, 95% CI 1.23–2.34, P¼ .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .73) and the I2 index (I2 value ¼ 0%, 95% CI 0%–90%). The effect of P support on live birth in gonadotropin cycles was similar in analysis of per patient non–cross-over reporting (RR 1.69), per patient allowing cross-over reporting (RR 1.77), and per cycle reporting (RR 1.62).
Letrozole. A single study evaluated letrozole OI-IUI (26). In 96 patients, the clinical pregnancy rate was 23.2% with P vs. 12.5% without P (P¼ .28). This study did not report on live birth. There is insufficient evidence to evaluate the effect of P luteal support in letrozole OI-IUI cycles.
This is the largest systematic review and meta-analysis to date of RCTs evaluating clinical pregnancy and live birth after OIIUI. Our previous meta-analysis from 2013 demonstrated a benefit of exogenous P after gonadotropin OI-IUI (13). The number of studies and subjects available for analysis has
VOL. - NO. - / - 2017
DISCUSSION
7
ORIGINAL ARTICLE: INFERTILITY more than doubled in the past 3 years. Such an increase in the available data necessitated a reassessment of the prior metaanalysis. This updated meta-analysis included 11 RCTs and found similar results, confirming that P supplementation improves clinical pregnancy and live birth in gonadotropin OIIUI cycles. There does not seem to be a benefit of exogenous P in CC OI-IUI cycles. There is insufficient evidence that P support improves outcomes in OI-IUI cycles using letrozole or CC plus gonadotropins. Embryo implantation and the maintenance of pregnancy require a receptive endometrium that has undergone differentiation into the secretory phase under the direct influence of P (1, 28). Insufficient P in the luteal phase may result in inadequate secretory changes of the endometrium and subsequent embryo implantation failure (27). It is well established that luteal phase support is necessary in ART cycles given the inhibition of normal pulsatile LH release from the pituitary gland in the setting of supraphysiologic serum hormone levels, which results in insufficient P production from corpus lutea (6, 7). The use of luteal support in the form of exogenous hCG or P have been proposed, though P is favored given the risks of ovarian hyperstimulation syndrome with hCG (11). Compared with ART, ovulation induction for IUI is generally associated with the development of fewer follicles, lower E2 levels, the absence of exogenous pituitary down-regulation, and the lack of mechanical disruption of the corpus luteum that occurs during oocyte retrieval. Therefore, it is less clear whether corpus luteum function and P release are compromised by ovulation induction, and data regarding the utility of P supplementation after OI-IUI have been conflicting. Prior studies have been limited by small sample sizes, and most were not powered to detect statistically significant differences in clinical pregnancy or live birth rates between patients who received exogenous P after IUI and those who did not (17–27). The observed benefit of P luteal support in gonadotropin OI-IUI cycles but not in clomiphene cycles in this meta-analysis may be explained by differences in drug mechanism of action and physiologic effects. Clomiphene citrate initiates its ovulation-inducing effect at the level of the hypothalamus, competitively binding to estrogen receptors and preventing negative feedback of E2. This lack of negative feedback increases GnRH pulse frequency, stimulating the release of LH and FSH from the anterior pituitary (29, 30). Increased LH may enhance corpus luteum function and increase both E2 and P in the luteal phase (31–33). Conversely, gonadotropins directly stimulate the ovaries to produce E2, which results in negative feedback at the levels of the hypothalamus and pituitary (34, 35). This may disrupt the normal pulsatile release of LH from the pituitary and impair P secretion from the corpus luteum (36, 37). A shortened luteal phase has been demonstrated after OI with gonadotropins and has been associated with insufficient luteal phase serum P levels (23, 38). Therefore, there is biologic plausibility for the benefit of exogenous P for luteal support in gonadotropin OI-IUI cycles. Seckin et al. (20) reported a higher clinical pregnancy rate with luteal support when multifollicular development occurred. However, Peeraer et al. (22) were unable to confirm this association. 8
The studies in this meta-analysis that report follicle numbers typically had average cohorts of approximately two follicles per patient, and the gonadotropin dose was low (75–150 IU) in most studies. The majority of studies did not provide E2 levels and did not perform regression analysis to determine whether higher E2 levels were associated with a reduction in pregnancy in nonsupplemented cohorts. Although the disruption of pulsatile P release in gonadotropin OI-IUI is biologically plausible, this meta-analysis cannot prove that the mechanism of improved live birth with P luteal support is related to higher numbers of follicles or higher E2 levels disrupting the luteal phase after gonadotropin OI-IUI. It is also possible that P or other sex steroids could inhibit LH in the luteal phase or that multifollicular development could be associated with endometrial effects detrimental to pregnancy (23). Given the relatively low cost and side effect profile of vaginal P, we recommend supplementing P in all patients undergoing gonadotropin OI-IUI. There is no level 1 evidence to support measuring serum P levels in the luteal phase and only supplementing P in patients below a certain threshold. The main strength of this study is the large number of analyzed RCTs evaluating the benefit of P luteal supplementation after OI-IUI. Analyses were performed using intentto-treat data when available, and primary outcomes were reported per patient, which provides the most valuable information for clinical decision making and counseling purposes. Sensitivity analyses on a per-cycle basis showed similar results. Two studies allowed patient cross-over between study arms and failed to appropriately account for this in their statistical analyses (19, 25). However, sensitivity analyses excluding these two studies had minimal effect on the meta-analysis results. This study evaluated not only clinical pregnancy but also live birth, which is the main outcome of interest in studies evaluating the impact of an intervention on reproductive outcomes after fertility treatment. Analytic strengths and limitations of this meta-analysis are similar to those of our previous meta-analysis and were previously described in detail (13). Briefly, strengths include the performance of multiple sensitivity analyses, which controls for interstudy variance in the number of cycles per patient and ensures results were not impacted by unit of analysis bias (39). Furthermore, both random-effects and fixed-effect models were used to account for clinical and statistical heterogeneity between studies (16). The clinical heterogeneity among the 11 studies (ovulation induction agent, medication dose and duration, type and dose of hCG for ovulation trigger, and the timing, duration, and specific formulation of the vaginal P support) is a limitation. However, after random-effects modeling, secondary sensitivity analyses, and subgroup analyses, the study results remained significant, supporting the conclusion that P luteal support improves clinical outcomes after gonadotropin OI-IUI. It is unlikely that the drug used for ovulation trigger affected the observed differences, because all studies used a form of hCG, and it has been shown that urinary and recombinant hCG have equivalent clinical outcomes when used for ovulation trigger in IVF (40). The studies evaluating gonadotropin OI-IUI were almost exclusively composed of patients with a diagnosis of unexplained infertility, and most studies had VOL. - NO. - / - 2017
Fertility and Sterility® an average of two follicles per patient. Therefore, it is not likely that there were differences in ovarian stimulation goals (ovulation induction vs. superovulation) that affected the conclusion in the gonadotropin analysis. Although gonadotropin cycles are associated with a higher risk of multiple pregnancy (41), which could impact results if the multiple pregnancy rate differed between groups, no study in this meta-analysis reported a higher multiple gestation rate in the P-supported cohorts. The small numbers of patients who underwent OI-IUI with either CC plus gonadotropins or letrozole is a limitation of this study. As a result, there are insufficient data to form conclusions regarding the utility of luteal phase P support in these cycles. Limitations of these data have recently been reviewed by Quass and Hansen (42). The current data are clinically heterogeneous, lack placebo control, lack double-blind design, and arise from single-center study designs. This raises the potential need for large, double-blinded, placebo-controlled, multicenter trials to remove potential sources of bias that exist in current data. In conclusion, level 1 RCT data demonstrate that exogenous P during the luteal phase improves both clinical pregnancy and live birth rates in patients undergoing gonadotropin ovulation induction and IUI. These data strengthen the concept that there may be physiologic differences in luteal function after ovulation induction with gonadotropins compared with CC and that the former group benefits from P supplementation. The data do not demonstrate improved outcomes with P luteal support in CC or letrozole ovulation induction cycles.
9. 10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
REFERENCES 1.
2. 3.
4.
5.
6.
7.
8.
Large MJ, DeMayo FJ. The regulation of embryo implantation and endometrial decidualization by progesterone receptor signaling. Mol Cell Endocrinol 2012;358:155–65. de Carvalho JF, Caleiro MT. Primary antiphospholipid syndrome and thyroid involvement. J Clin Rheumatol 2010;16:164–7. Young SL. Oestrogen and progesterone action on endometrium: a translational approach to understanding endometrial receptivity. Reprod Biomed Online 2013;27:497–505. Mutlu MF, Erdem M, Erdem A, Mutlu I, Guler I, Demirdag E. The impact of premature progesterone rise on the outcome of intrauterine insemination cycles with controlled ovarian hyperstimulation in unexplained infertility. Eur J Obstet Gynecol Reprod Biol 2016;203:44–8. Palomino WA, Fuentes A, Gonzalez RR, Gabler F, Boric MA, Vega M, et al. Differential expression of endometrial integrins and progesterone receptor during the window of implantation in normo-ovulatory women treated with clomiphene citrate. Fertil Steril 2005;83:587–93. Nippoldt TB, Reame NE, Kelch RP, Marshall JC. The roles of estradiol and progesterone in decreasing luteinizing hormone pulse frequency in the luteal phase of the menstrual cycle. J Clin Endocrinol Metab 1989;69:67–76. Beckers NG, Macklon NS, Eijkemans MJ, Ludwig M, Felberbaum RE, Diedrich K, et al. Nonsupplemented luteal phase characteristics after the administration of recombinant human chorionic gonadotropin, recombinant luteinizing hormone, or gonadotropin-releasing hormone (GnRH) agonist to induce final oocyte maturation in in vitro fertilization patients after ovarian stimulation with recombinant follicle-stimulating hormone and GnRH antagonist cotreatment. J Clin Endocrinol Metab 2003;88:4186–92. Messinis IE, Templeton AA. Endocrine and follicle characteristics of cycles with and without endogenous luteinizing hormone surges during superovulation induction with pulsatile follicle-stimulating hormone. Hum Reprod 1987;2:11–6.
VOL. - NO. - / - 2017
20.
21.
22.
23.
24.
25.
26.
27.
Messinis IE. Ovarian feedback, mechanism of action and possible clinical implications. Hum Reprod Update 2006;12:557–71. Fauser BC, Devroey P. Reproductive biology and IVF: ovarian stimulation and luteal phase consequences. Trends Endocrinol Metab 2003;14:236–42. van der Linden M, Buckingham K, Farquhar C, Kremer JA, Metwally M. Luteal phase support for assisted reproduction cycles. Cochrane Database Syst Rev 2015:CD009154. Practice Committee of the American Society for Reproductive Medicine. Progesterone supplementation during the luteal phase and in early pregnancy in the treatment of infertility: an educational bulletin. Fertil Steril 2008;90: S150–3. Hill MJ, Whitcomb BW, Lewis TD, Wu M, Terry N, DeCherney AH, et al. Progesterone luteal support after ovulation induction and intrauterine insemination: a systematic review and meta-analysis. Fertil Steril 2013;100: 1373–80. Tavaniotou A, Albano C, Smitz J, Devroey P. Effect of clomiphene citrate on follicular and luteal phase luteinizing hormone concentrations in in vitro fertilization cycles stimulated with gonadotropins and gonadotropinreleasing hormone antagonist. Fertil Steril 2002;77:733–7. Tavaniotou A, Albano C, Smitz J, Devroey P. Comparison of LH concentrations in the early and mid-luteal phase in IVF cycles after treatment with HMG alone or in association with the GnRH antagonist Cetrorelix. Hum Reprod 2001;16:663–7. Borenstein M, Hedges LV, Higgins JP, Rothstein HR. A basic introduction to fixed-effect and random-effect models for meta-analysis. Res Synth Methods 2010;1:97–111. Aali BS, Ebrahimipour S, Medhdizadeh S. The effectiveness of luteal phase support with cyclogest in ovarian stimulated intra uterine insemination cycles: a randomized controlled trial. Iran J Reprod Med 2013; 11:309–14. Hossein Rashidi B, Davari Tanha F, Rahmanpour H, Ghazizadeh M. Luteal phase support in the intrauterine insemination (IUI) cycles: a randomized double blind, placebo controlled study. J Fam Reprod Health 2014;8:149– 53. ~ozRomero Nieto MI, Lorente Gonzalez J, Arjona-Berral JE, Del Mun Villanueva M, Castelo-Branco C. Luteal phase support with progesterone in intrauterine insemination: a prospective randomized study. Gynecol Endocrinol 2014;30:197–201. Seckin B, Turkcapar F, Yildiz Y, Senturk B, Yilmaz N, Gulerman C. Effect of luteal phase support with vaginal progesterone in intrauterine insemination cycles with regard to follicular response: a prospective randomized study. J Reprod Med 2014;59:260–6. Karadag B, Dilbaz B, Karcaaltincaba D, Sahin EG, Ercan F, Karasu Y, et al. The effect of luteal-phase support with vaginal progesterone on pregnancy rates in gonadotropin and clomiphene citrate/intra-uterine insemination cycles in unexplained infertility: a prospective randomised study. J Obstet Gynaecol 2016;36:794–9. Peeraer K, D’Hooghe T, Laurent P, Pelckmans S, Delvigne A, Laenen A, et al. Impact of luteal phase support with vaginal progesterone on the clinical pregnancy rate in intrauterine insemination cycles stimulated with gonadotropins: a randomized multicenter study. Fertil Steril 2016;106:1490–5. Erdem A, Erdem M, Atmaca S, Guler I. Impact of luteal phase support on pregnancy rates in intrauterine insemination cycles: a prospective randomized study. Fertil Steril 2009;91:2508–13. Kyrou D, Fatemi HM, Tournaye H, Devroey P. Luteal phase support in normoovulatory women stimulated with clomiphene citrate for intrauterine insemination: need or habit? Hum Reprod 2010;25:2501–6. Maher MA. Luteal phase support may improve pregnancy outcomes during intrauterine insemination cycles. Eur J Obstet Gynecol Reprod Biol 2011; 157:57–62. Agha-Hosseini M, Rahmani M, Alleyassin A, Safdarian L, Sarvi F. The effect of progesterone supplementation on pregnancy rates in controlled ovarian stimulation and intrauterine insemination cycles: a randomized prospective trial. Eur J Obstet Gynecol Reprod Biol 2012;165:249–53. Ebrahimi M, Asbagh FA, Dervish S. The effect of luteal phase support on pregnancy rates of the stimulated intrauterine insemination cycles in couples with unexplained infertility. Int J Fertil Steril 2010;4:51–6.
9
ORIGINAL ARTICLE: INFERTILITY 28.
29.
30.
31.
32.
33. 34. 35.
10
Johannisson E, Parker RA, Landgren BM, Diczfalusy E. Morphometric analysis of the human endometrium in relation to peripheral hormone levels. Fertil Steril 1982;38:564–71. Dickey RP, Vorys N, Stevens VC, Besch PK, Hamwi GJ, Ullery JC. Observations on the mechanism of action of clomiphene (MRL-41). Fertil Steril 1965;16: 485–94. Kerin JF, Liu JH, Phillipou G, Yen SS. Evidence for a hypothalamic site of action of clomiphene citrate in women. J Clin Endocrinol Metab 1985; 61:265–8. Hammond MG, Talbert LM. Clomiphene citrate therapy of infertile women with low luteal phase progesterone levels. Obstet Gynecol 1982;59:275–9. Fukuma K, Fukushima T, Matsuo I, Mimori H, Maeyama M. A graduated regimen of clomiphene citrate: its correlation to glycogen content of the endometrium and serum levels of estradiol and progesterone in infertile patients at the midluteal phase. Fertil Steril 1983;39:780–4. Shirai E, Iizuka R, Notake Y. Clomiphene citrate and its effects upon ovulation and estrogen. Fertil Steril 1972;23:331–8. Tavaniotou A, Smitz J, Bourgain C, Devroey P. Ovulation induction disrupts luteal phase function. Ann N Y Acad Sci 2001;943:55–63. Tavaniotou A, Albano C, Smitz J, Devroey P. Impact of ovarian stimulation on corpus luteum function and embryonic implantation. J Reprod Immunol 2002;55:123–30.
36.
37.
38.
39.
40.
41.
42.
Rossmanith WG, Laughlin GA, Mortola JF, Johnson ML, Veldhuis JD, Yen SS. Pulsatile cosecretion of estradiol and progesterone by the midluteal phase corpus luteum: temporal link to luteinizing hormone pulses. J Clin Endocrinol Metab 1990;70:990–5. Rossmanith WG, Laughlin GA, Mortola JF, Yen SS. Secretory dynamics of oestradiol (E2) and progesterone (P4) during periods of relative pituitary LH quiescence in the midluteal phase of the menstrual cycle. Clin Endocrinol (Oxf) 1990;32:13–23. Olson JL, Rebar RW, Schreiber JR, Vaitukaitis JL. Shortened luteal phase after ovulation induction with human menopausal gonadotropin and human chorionic gonadotropin. Fertil Steril 1983;39:284–91. van der Poel N, Farquhar C, Abou-Setta AM, Benschop L, Heineman MJ. Soft versus firm catheters for intrauterine insemination. Cochrane Database Syst Rev 2010:CD006225. Youssef MA, Abou-Setta AM, Lam WS. Recombinant versus urinary human chorionic gonadotrophin for final oocyte maturation triggering in IVF and ICSI cycles. Cochrane Database Syst Rev 2016:CD003719. Diamond MP, Legro RS, Coutifaris C, Alvero R, Robinson RD, Casson P, et al. Letrozole, gonadotropin, or clomiphene for unexplained infertility. N Engl J Med 2015;373:1230–40. Quaas AM, Hansen KR. The role of steroid hormone supplementation in non-assisted reproductive technology treatments for unexplained infertility. Fertil Steril 2016;106:1600–7.
VOL. - NO. - / - 2017
Fertility and Sterility®
SUPPLEMENTAL ADDENDUM: DETAILED SEARCH STRATEGY The search strategy consisted of the following phrases: (((((progesterone[tiab] OR progesterone[mh])) AND (luteal [tiab] OR luteal phase[mh])) AND (‘‘intrauterine insemination’’[tiab] OR IUI[tiab] OR insemination, artificial[mh] OR ovulation induction[mh] OR ‘‘ovarian stimulation’’[tiab] OR (ovar*[tiab] AND stimulat*[tiab]) OR ovulat*[tiab]))) AND
VOL. - NO. - / - 2017
(randomized controlled trial[pt] OR randomized controlled trial[mh] OR single-blind method[mh] OR double-blind method[mh] OR random allocation[mh] OR random*[tiab] OR ‘‘single blind’’[tiab] OR ‘‘double blind’’[tiab] OR placebo [tiab] OR ‘‘randomized controlled trial’’[tiab] OR controlled clinical trial[pt] OR ‘‘controlled clinical trial’’[tiab] OR (doubl*[tiab] AND blind*[tiab]) OR (singl*[tiab] AND blind* [tiab])).
10.e1
Assessment of bias in the included RCTs, including assessments of randomization, allocation concealment, blinding, data reporting, and declaration of conflicts of interest. Authors (reference) Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahimi et al. 2010 (27) Maher 2011 (25)
Agha-Hosseini et al. 2012 (26) Aali et al. 2013 (17) Romero Nieto et al. 2014 (19)
Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18) Karadag et al. 2016 (21) Peeraer et al. 2016 (22)
Randomization
Allocation concealment
Software-generated random Single researcher sequence aware, treating providers blinded Software-generated random None sequence Sequential randomization None Software-generated random None sequence; patients alternated treatment arms every subsequent cycle Software-generated random None sequence Block randomization (four None patients per block) Software-generated random None sequence; patients were re-randomized in subsequent treatment cycles Software-generated random None sequence Software-generated random Yes sequence Software-generated random None sequence Software-generated block Yes randomization (10 patients per block)
Green. Progesterone luteal support after IUI. Fertil Steril 2017.
Blinding of Blinding of Incomplete participants outcomes data reporting
Analysis type
Conflicts of interest or pharma sponsorship
Trial registry
None
None
Yes
Intent to treat
None
Not stated
None
None
Yes
Not stated
Yes
None None
None None
None Yes
Intent to treat and per protocol Per protocol Per protocol
None Not stated
Yes Not stated
None
None
Yes
Per protocol
Not stated
Yes
None
None
Yes
Intent to treat
None
Yes
None
None
Yes
Intent to treat but re-randomized with each cycle
None
Not stated
None
None
Yes
Intent to treat
Not stated
Not stated
Yes
None
None
Intent to treat
None
Not stated
None
None
Yes
Intent to treat
None
Not stated
None
None
Yes
Intent to treat and per protocol
One author is vice president Yes of Global Medical Affairs Fertility at Merck
ORIGINAL ARTICLE: INFERTILITY
10.e2
SUPPLEMENTAL TABLE 1
VOL. - NO. - / - 2017
VOL. - NO. - / - 2017
SUPPLEMENTAL TABLE 2 Fertility history, ovarian reserve, and stimulation results reported between the studies. Authors (reference) Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahami et al. 2010 (27) Maher 2011 (25) Agha-Hosseini et al. 2012 (26) Aali et al. 2013 (17) Romero Nieto et al. 2014 (19) Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18) Karadag et al. 2016 (21)
Peeraer et al. 2016 (22)
Group
Age (y)
P Control P Control P Control P Control P Control P Control P Control P Control P Control P (CC arm) Control (CC arm) P (Gnd arm) Control (Gnd arm) P Control
30 4.8 29.7 4.3 32.1 3.6 32.2 3.9 27.9 3.3 28.4 4.1 NR NR 27.4 3.7 26.8 3.6 28b 27.9b 32.9 3.4 33.1 3.3 27.8 5.2 28.9 5.5 27.53 4.23 28.11 4.62 26.3 4.23 25.7 4.2
Green. Progesterone luteal support after IUI. Fertil Steril 2017.
NS NS NS NR NS NS NS NS NS NS
Primary infertility (%) 64.2 63.8 71.4 69.6 NR NR NR NR NR NR 89.9 88.7 100 100 88.7 96.2 NR NR 66 48
P value NS NS NR NR NR NS NS NS NR NS
Day-3 FSH (IU/L) 7.3 2.7 7 2.7 5.4 3.1 5.1 2.1 6.1 2.7 6.7 2.0 NR NR 5.4 3.1 5.1 2.1 6.9 2.4 7.2 2.1 NR NR 6.7 2.2 6.8 2.3 5.98 2.33 6.38 2.56 6.4 1.7 5.8 1.6
P value NS NS NS NR NS NS NR NS NS NS
Peak E2 (pg/mL) NR NR 513 248 504 306 NR NR NR NR NR NR NR NR NR NR 597.7 323.4 670.4 434.5 NR NR NR NR
P value NR NS NR NR NR NR NR NS NR NR
Dominant folliclesa 1.6 0.6 1.5 0.9 1.2 0.3 1.3 0.4 2.0 0.7 2.2 0.8 NR NR 1.6 0.7 1.4 0.8 NR NR 1.4c 1.4c 2 1.3 2.3 1.8 2.21 1.29 2.43 1.31 NR NR
P value NS .02 NS NR NS NR NR NS NS NR
28.5 3.6 27.8 3.7
NS
54 58
NS
7.1 4.2 6.8 2.5
NS
NR NR
NR
NR NR
NR
31 3.97 31.5 3.8
NS
71 63
NS
NR NR
NR
289 (266–311)d 311 (288–334)d
NS
1.2 0.1 1.2 0.1
NS
10.e3
Fertility and Sterility®
Note: Data listed as mean SD unless otherwise stated. a Defined as follicles over 16-18 mm. b Range not reported. c SD not reported. d Mean (95% CI).
P value
ORIGINAL ARTICLE: INFERTILITY
Eligibility
Screening
Identification
SUPPLEMENTAL FIGURE 1 63 records identified through database searching
4 additional records identified through other sources
56 records after duplicates removed
56 records screened
15 full-text articles assessed for eligibility
Included
11 studies included in qualitative synthesis
41 records excluded
4 full-text articles excluded: 1 article was an abstract reported in an included paper 3 articles did not have a progesterone luteal support arm
11 studies included in quantitative synthesis (meta-analysis)*
PRISMA (Preferred Reporting Items for Systematic Reviews and MetaAnalyses) flow diagram of study selection. *Eleven studies were included in per-cycle analyses and ten studies were included in perpatient analyses (Romero Nieto et al 2014 (19) was excluded from per-patient analyses due to re-randomization of patients in subsequent cycles). Green. Progesterone luteal support after IUI. Fertil Steril 2017.
10.e4
VOL. - NO. - / - 2017
Fertility and Sterility®
SUPPLEMENTAL FIGURE 2
Selectivity funnel plot of all studies included in the analysis for assessment of publication bias. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
VOL. - NO. - / - 2017
10.e5
Progesterone luteal support after ovulation induction and intrauterine insemination: an updated systematic review and meta-analysis Katherine A. Green, M.D.,a Jessica R. Zolton, D.O.,a Sophia M. V. Schermerhorn, B.S.,b Terrence D. Lewis, M.D., Ph.D.,a,d Mae W. Healy, D.O.,a,d Nancy Terry, M.L.S.,c Alan H. DeCherney, M.D.,a and Micah J. Hill, D.O.a,d a Program in Reproductive and Adult Endocrinology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland; b Uniformed Services University of the Health Sciences, Bethesda, Maryland; c National Institutes of Health Library, Bethesda, Maryland; and d Walter Reed National Military Medical Center, Bethesda, Maryland
Objective: To evaluate the effect of progesterone (P) for luteal phase support after ovulation induction (OI) and intrauterine insemination (IUI). Design: An updated systematic review and meta-analysis. Setting: Not applicable. Patient(s): Patients undergoing OI-IUI for infertility. Intervention(s): Exogenous P luteal support after OI-IUI. Main Outcome Measure(s): Live birth. Result(s): Eleven trials were identified that met inclusion criteria and constituted 2,842 patients undergoing 4,065 cycles, more than doubling the sample size from the previous meta-analysis. In patients receiving gonadotropins for OI, clinical pregnancy (relative risk [RR] 1.56, 95% confidence interval [CI] 1.21–2.02) and live birth (RR 1.77, 95% CI 1.30–2.42) were more likely in P supplemented patients. These findings persisted in analysis of live birth per IUI cycle (RR 1.59, 95% CI 1.24–2.04). There were no data on live birth in clomiphene citrate or clomiphene plus gonadotropin cycles. There was no benefit on clinical pregnancy with P support for patients who underwent OI with clomiphene (RR 0.85, 95% CI 0.52–1.41) or clomiphene plus gonadotropins (RR 1.26, 95% CI 0.90–1.76). Conclusion(s): Progesterone luteal phase support is beneficial to patients undergoing ovulation induction with gonadotropins in IUI cycles. The number needed to treat is 11 patients to have one additional live birth. Progesterone support did not benefit patients undergoing ovulation induction with clomiphene citrate or clomiphene plus gonadotropins. (Fertil SterilÒ 2017;-:-–-. Ó2017 by American Society for Reproductive Medicine.) Key Words: Intrauterine insemination, live birth, luteal support, ovulation induction, progesterone Discuss: You can discuss this article with its authors and with other ASRM members at https://www.fertstertdialog.com/users/ 16110-fertility-and-sterility/posts/14467-23524
Received December 6, 2016; revised January 4, 2017; accepted January 16, 2017. K.A.G. has nothing to disclose. J.R.Z. has nothing to disclose. S.M.V.S. has nothing to disclose. T.D.L. has nothing to disclose. M.W.H. has nothing to disclose. N.T. has nothing to disclose. A.H.D. has nothing to disclose. M.J.H. has nothing to disclose. Supported, in part, by the Program in Reproductive and Adult Endocrinology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Bethesda, MD. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of Defense or the US Government. Reprint requests: Katherine A. Green, M.D., National Institutes of Health, Eunice Kennedy Shriver National Institute of Child Health and Human Development, Program in Reproductive and Adult Endocrinology, 10 CRC, Room 1E-3140, 10 Center Drive, MSC 1109, Bethesda, MD 20892-1109 (E-mail: [email protected]). Fertility and Sterility® Vol. -, No. -, - 2017 0015-0282/$36.00 Copyright ©2017 Published by Elsevier Inc. on behalf of the American Society for Reproductive Medicine http://dx.doi.org/10.1016/j.fertnstert.2017.01.011 VOL. - NO. - / - 2017
S
uccessful implantation requires synchrony between a competent blastocyst and a receptive secretory phase endometrium (1). As estrogen rises during the follicular phase and a dominant follicle emerges, the mid-cycle LH (2) peaks and ovulation occurs (2). The pulsatile LH secretion stimulates the corpus luteum to produce P, which induces endometrial secretory transformation and promotes receptivity (3). Fertility treatments may interfere with the luteal phase via several mechanisms. Ovulation induction (OI) may 1
ORIGINAL ARTICLE: INFERTILITY result in a premature rise of P and alterations in endometrial receptivity (4, 5). Furthermore, supraphysiologic E2 elevation from ovarian stimulation may cause pituitary downregulation and alterations in luteal phase LH secretion (6, 7). Supraphysiologic E2 levels are often associated with multifollicular development during assisted reproductive technology (ART) (8), whereas only one to two dominant follicles may be achieved during OI and IUI. The use of GnRH analogs for pituitary down-regulation and mechanical disruption of follicles during oocyte aspiration may further affect luteal function (9, 10). Luteal support with exogenous P after ART is routine because it is associated with higher pregnancy and live birth rates (11, 12); however, there is no consensus on the use of P after OI-IUI. In 2013 our group published a systematic review and meta-analysis of randomized, controlled trials (RCTs) to evaluate the effect of luteal phase P support after OI-IUI (13). This analysis concluded that luteal phase support was beneficial in gonadotropin IUI cycles but not in clomiphene citrate (CC) IUI cycles. The endogenous rise of LH as a result of CC may provide further stimulation for the developing corpus luteum, which in turn will favorably impact the luteal phase (14). In contrast, stimulation with gonadotropins directly may result in negative feedback of E2 at the hypothalamus and decreased LH release, similar to ART cycles (15). In the past 3 years there have been several new RCTs published on this topic. The total number of subjects and papers has doubled since the prior review, necessitating new statistical analysis of the published data. The goal of this study was to review these recent publications and perform an updated meta-analysis to determine the impact of P supplementation after OI-IUI.
MATERIALS AND METHODS Study Design This is an updated systematic review and meta-analysis of RCTs evaluating exogenous P luteal support during the luteal phase after OI-IUI (13). The objective of this study was to identify eligible RCTs that have been published since our previous meta-analysis in 2013 and update the systematic review and meta-analysis.
Literature Search PubMed and Embase literature searches were performed for published RCTs evaluating P luteal supplementation vs. no luteal support after OI-IUI. Our previous meta-analysis conducted a literature search through January 8, 2013 and identified five studies meeting inclusion criteria (13). An updated literature search was performed, limited to publications from January 1, 2013 to the date of search execution, which occurred on September 12, 2016. The search used specific key words and database indexing terminology (available online as Supplemental Addendum).
Study Selection Study selection was performed according to the same criteria as in the previous meta-analysis (13). Only published RCTs 2
that compared exogenous P during the luteal phase after OI-IUI vs. no P were included. Any type of ovulation induction was allowed, including CC, exogenous gonadotropins, hCG, aromatase inhibitors, or a combination of these medications. All types of exogenous P were permitted, including oral, IM, or vaginal formulations. Publication in any language was allowed. Exclusion criteria included nonrandomization, timed intercourse cycles, natural cycles, publication as abstract only, book chapters, or review articles. The literature search identified 31 new publications, which were independently reviewed by three investigators (K.A.G., J.R.Z., and M.J.H.) to identify eligible studies. Of the 31 abstracts reviewed, 25 studies were excluded on the basis of abstract data indicating failure to meet inclusion criteria. Six full-text articles were reviewed for inclusion and exclusion criteria, all of which met inclusion criteria. There were no disagreements among the three reviewing investigators regarding the studies eligible for inclusion. Study quality and the potential for bias within each study was evaluated, considering randomization method, concealment of allocation, blinding of providers and patients, and flow of patients through the randomization, treatment, and outcome stages.
Data Collection Data from studies that met inclusion criteria were extracted independently by two investigators (K.A.G. and J.R.Z.). Clinical pregnancy and live birth data were extracted from intentto-treat results. When intent-to-treat results were not reported, data were calculated as intent-to-treat by making the denominator the number of patients enrolled, instead of the reported per-protocol results. Continuous data were extracted in the form of mean and SD. Additional extracted data included author, year of publication, journal, country of origin, randomization method, sample size, number of patients randomized, number of cycles performed, method of ovulation induction, type of P support, duration of P support, method of ovulation triggering, trial registry, and conflicts of interest. The primary outcome was live birth per patient. Secondary outcomes included clinical pregnancy per patient, clinical pregnancy per cycle, and live birth per cycle. Sensitivity analyses were performed excluding trials that allowed individual subjects to cross over into both treatment and control arms over multiple cycles.
Data Synthesis Data for synthesis were obtained from intent-to-treat results when reported. Primary analyses were performed using perpatient data, and additional analyses were performed using per-cycle data. Heterogeneity was evaluated using the Q test and I2 index values and reported for each outcome as P value and percentage, respectively. Random-effects models were used when studies had clinical heterogeneity in the ovulation induction method (i.e., medication type) used or when the I2 index was >50% (16). A fixed-effect model was used when the same type of ovulation induction method was used in all studies and the I2 index was <50%. Sensitivity analyses VOL. - NO. - / - 2017
VOL. - NO. - / - 2017
TABLE 1 Study characteristics of trials meeting inclusion in the systematic review. Country of study
Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahimi et al. 2010 (27) Maher 2011 (25)
Turkey
Unexplained infertility
75 IU FSH starting cycle day 3
Belgium
Unexplained infertility (implicit) Unexplained infertility
50 mg CC cycle days 3–7
Agha-Hosseini et al. 2012 (26)
Iran
Aali et al. 2013 (17)
Iran
Romero Nieto et al. 2014 (19)
Spain
Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18)
Turkey
Karadag et al. 2016 (21)
Turkey
Peeraer et al. 2016 (22)
Belgium
Iran Egypt
Iran
Patients
Unexplained Infertility (implicit), anovulatory infertility, at least one patent fallopian tube, post-wash total motile sperm R1 106 Unexplained infertility (implicit) Age <40 y, FSH<12 IU/L, sperm concentration R10 106/mL Unexplained infertility (implicit), post-wash total motile sperm R3 106 Unexplained infertility Unexplained or anovulatory infertility, at least one patent fallopian tube, mild male factor Unexplained infertility
Unexplained infertility, minimal-mild endometriosis, post-wash total motile sperm R5 106
Green. Progesterone luteal support after IUI. Fertil Steril 2017.
Ovarian stimulation
50 mg CC BID cycle days 3–7 plus 75 IU hMG cycle days 7–9 75 IU FSH starting days 2–5
Either 50 mg CC BID or letrozole 5 mg daily cycle days 3–7, with or without 75 IU hMG days 3–7 75–150 IU hMG starting on cycle day 3 or 100 mg CC on cycle days 3–7 plus hMG on days 8 and 9 Urinary or recombinant gonadotropins starting on cycle days 2–4
P supplementation Crinone 8% post-IUI day 2 through 12 wk Utrogestan 200 mg in 3 separate doses post-IUI day 1 through 7 wk Cyclogest 400 mg daily post-IUI day 2 through 10 wk Crinone 8% post-IUI day 1 through 14 d
Ovulation triggering
No. of cycles allowed
10,000 U hCG
Up to 3
5,000 U hCG
1
5,000 U hCG
Up to 3
10,000 U hCG
Up to 6, patients alternated treatment groups each cycle
Cyclogest 400 mg daily post-IUI day 1 through 14 d
10,000 U hCG
1
Cyclogest 400 mg daily post-IUI day 1 through 10 d
10,000 U hCG
1
Progesterone 200 mg vaginally daily post-IUI day 1 until 8 wk
6,500 IU recombinant hCG (rhCG) (Ovitrelle) 10,000 IU hCG
Up to 4, patients re-randomized each cycle
10,000 U hCG
1
75 IU FSH starting on cycle day 3; dose modified based on response 100 mg CC on cycle days 3–7 plus 75 IU hMG on cycle days 7–10 (longer based on response)
Crinone 8% on day of IUI until 12 wk Progesterone 400 mg vaginally BID post-IUI day 2 through 8 wk
Up to 3
Arm 1) 100 mg CC on cycle days 3–7 Arm 2) 75 IU FSH starting on cycle days 2–4 37.5–75 IU FSH starting on cycle days 2–3; increased by 37 IU on cycle day 7–10 if no response
Crinone 8% post-IUI day 1 until 10 wk
6,500 IU rhCG (Ovitrelle)
1
Crinone 8% post-IUI day 1 for 15 d
rhCG (Ovitrelle; dose not specified)
1
3
Fertility and Sterility®
Authors (reference)
ORIGINAL ARTICLE: INFERTILITY using random-effects models were performed in all cases in which fixed-effect models were the primary analysis, to ensure that the model type did not change the outcome. In all cases, random-effects models gave similar results to the fixed-effect models without altering the statistical significance. Risk differences between groups were calculated using Mantel-Haenszel estimates. When significant risk differences existed between groups, numbers needed to treat (NNTs) to have one additional live birth were calculated from the Mantel-Haenszel risk difference estimates. Bias was assessed at the study level using a qualitative review assessing randomization, concealment, blinding, and patient flow. Funnel plots were constructed to assess publication bias at the outcome level. Dichotomous outcome data were reported as relative risks (RRs) with 95% confidence intervals (CIs). Continuous data were synthesized using weighted means with 95% CIs. A priori subgroup analyses were planned to compare the types of P support and methods of ovulation induction. Sensitivity analyses were planned a priori if included trials were found to be at high risk of bias at either the study or outcome level. Exclusion sensitivity analysis was performed on each trial to determine whether its inclusion changed the statistical findings of each model. In no case did exclusion of any study change the findings. Data collection was performed in Excel (Microsoft Office 2007), and statistical analysis was performed using Mix 2.0 Pro (Bax L: MIX 2.0. Professional software for metaanalysis in Excel version 2.0.1.4. BiostatXL, 2011; http:// www.meta-analysis-made-easy.com).
RESULTS Studies Included for Systematic Review and MetaAnalysis A total of 31 abstracts were identified, and six full-text articles were reviewed, all of which met full inclusion criteria (17–22). By combining these six trials with the five studies previously identified in our 2013 meta-analysis (13), this updated meta-analysis of 11 RCTs comprised 2,842 patients and 4,065 OI-IUI cycles (17–27) (Supplemental Fig. 1). All 11 trials described inclusion criteria consistent with unexplained infertility. Specific inclusion criteria and study characteristics are outlined in Table 1. Five studies used exogenous gonadotropins for ovulation induction in all patients (19, 20, 22, 23, 25). Four of these studies specified the use of recombinant FSH for ovulation induction (20, 22, 23, 25), and one study did not list a specific medication beyond describing the use of urinary or recombinant gonadotropins (19). One study described two treatment arms, one of which consisted of patients who received only recombinant FSH and a second arm that included patients who underwent stimulation with CC (21). One trial used either hMG alone or CC plus 2 days of hMG (17). Two additional studies used CC plus hMG for all patients (18, 27). One trial used either CC or letrozole in all patients, with some patients also receiving hMG (26), and one study used only CC (24). Ovulation triggering was performed with either 5,000 IU or 10,000 IU of hCG in eight studies (17,18,20,23–27). Two trials used 6,500 IU of recombinant hCG (19, 21), and one study used 4
recombinant hCG but did not specify dose (22). All studies used vaginal routes of P luteal support, with variation in the type and dose of P as well as therapy duration (Table 1). None of the studies used oral or IM routes of P supplementation.
Assessment of Bias Risk The 11 included trials were individually assessed for risk of bias (Supplemental Table 1). The randomization process was clearly described in all studies. Six trials used computergenerated randomization sequences in a non–crossover design (18, 20, 21, 23, 24, 26). Two trials also used computer-generated randomization sequence; however, patients were randomized to their first cycle of treatment and then either alternated treatment groups in each subsequent treatment cycle (25) or were re-randomized in subsequent treatment cycles (19). These two studies were excluded in a sensitivity analysis removing studies that allow subjects to cross over between study groups. Two studies reported block randomization of patients (17, 22), and one trial reported sequential randomization of patients (27). Three studies described attempts to control for allocation concealment (18, 22, 23). Three trials reported a single-blind study design, with blinding of the providers (17, 22, 23). One trial described a double-blind study design with blinding of both providers and patients and the use of a placebo control (18). None of the studies documented blinding of outcomes data. Nine trials adequately reported on the flow of patients through the study and were at low risk of incomplete data reporting (17,19–26). One study disclosed an author's affiliation with a pharmaceutical company (22). There were no pharmaceutical affiliations or support disclosed in any of the other trials. There were no differences in female age or duration of infertility between the two randomized groups in any of the studies; however, one study did not compare the duration of infertility between the two groups (19) (Supplemental Table 2). One study reported significantly fewer follicles R17 mm on the day of hCG administration in the P-supplemented group compared with the control group (1.2 0.3 vs. 1.3 0.4 follicles, respectively, P¼ .02) (24). A separate study (21) reported a significantly higher mean body mass index for patients in the P-supported group compared with the control group in the gonadotropin OI-IUI arm of the study (24.7 2.4 kg/m2 vs. 23.6 2.3 kg/m2, respectively, P¼ .021). Funnel plot analysis did not provide evidence of any publication bias (Supplemental Fig. 2).
Overall Meta-Analysis Results Ten studies reported clinical pregnancy on a per-patient basis, totaling 2,444 patients undergoing OI-IUI (Table 2) (17,18,20– 27). In random-effects modeling, clinical pregnancy was increased with P support (RR 1.37, 95% CI 1.15–1.64, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .65) and the I2 index (I2 value ¼ 0%, 95% CI 0%–62%). Five studies reported live birth on a per-patient basis, totaling 1,027 patients undergoing OI-IUI (20, 22, 23, 25, VOL. - NO. - / - 2017
Fertility and Sterility®
TABLE 2 Primary pregnancy outcomes in the 1 included randomized, controlled trials reported on a per-patient basis. Authors (reference) Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahimi et al. 2010 (27) Maher 2011 (25) Agha-Hosseini et al. 2012 (26) Aali et al. 2013 (17) Romero Nieto et al. 2014 (outcomes listed per cycle) (19) Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18) Karadag et al. 2016 (21)
Peeraer et al. 2016 (22)
Group P Control P Control P Control P Control P Control P Control P Control P Control P Control P (CC arm) Control (CC arm) P (Gnd arm) Control (Gnd arm) P Control
Patients (n) 109 105 243 225 98 102 37 34 150 150 99 97 398e 398e 71 78 127 126 50 50 50 50 202 191
Cycles (n) 223 204 243 225 252 259 132 126 148 142 99 97 449 444 71 78 127 126 50 50 50 50 202 191
Positive hCG (%) NR NR NR NR NR NR NR NR 28.7 20.7 16.2 14.4 NR NR NR NR 30.8 22.2 10 16 20 18 NR NR
P value NR NR NR NR NR NS NR NR NS NR NR NR
Clinical pregnancy (%) a
39.4 23.8a 7c 8.4c 30.6a 25.5a 54.9a 35.2a 24d 13.3d 12.1a 9.3a 13.8a 11a 22.5a 14.1a 15.8f 12.7f 10d 12d 20d 16d 16.8d 11d
P value .01 NS NS .02 < .05 NS NS NS NS NS NS NS
Live birth (%) b
35.8 18.1b NR NR 19.4b 14.7b 35.2b 9.8b NR NR NR NR 10.2b 8.3b 19.7b 14.1b NR NR NR NR NR NR 14.9g 9.4g
P value .003 NR NS < .001 NR NR NS NS NR NR NR NS
Note: Gnd ¼ gonadotropin; NR ¼ not reported; NS ¼ not significant. a Clinical pregnancy defined as the presence of gestational sac on ultrasound. b Live birth defined as an infant alive 1 week after birth. c Values reported as ongoing pregnancy, defined as pregnancy developing beyond 12 weeks of gestation. d Clinical pregnancy defined as the presence of fetal cardiac activity on ultrasound. e Total number of patients enrolled in study (patients were re-randomized in subsequent cycles). f Clinical pregnancy not defined. g Live birth defined as live birth of a child after 24 weeks of gestation. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
27). In random-effects modeling, live birth was increased with P support (RR 1.76, 95% CI 1.29–2.40, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .73) and the I2 index (I2 value ¼ 0%, 95% CI 0%–84%). All 11 studies reported on clinical pregnancy on a percycle basis, totaling 4,065 OI-IUI cycles (Table 2). In random-effects modeling, clinical pregnancy was increased with P support (RR 1.34, 95% CI 1.15–1.57, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .75) and the I2 index (I2 value ¼ 0%, 95% CI 0%– 60%). Five studies reported live birth on a per-cycle basis without treatment group crossover, totaling 1,950 cycles undergoing OI-IUI (19, 20, 22, 23, 27). In random-effects modeling, live birth was increased with P support (RR 1.49, 95% CI 1.15–1.94, P¼ .002). Minimal heterogeneity was suggested by the results of the Q test (P¼ .59) and the I2 index (I2 value ¼ 0%, 95% CI 0%–79%).
Subgroup Analyses Gonadotropins. Five studies evaluating gonadotropin OI-IUI reported clinical pregnancy on a per-patient basis, totaling 927 patients (Fig. 1A) (20–23,25). In fixed-effect modeling, clinical pregnancy was increased with P support (RR 1.56, 95% CI 1.21–2.02, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .98) and the I2 index VOL. - NO. - / - 2017
(I2 value ¼ 0%, 95% CI 0%–79%). Four studies reported live birth on a per-patient basis, totaling 827 patients undergoing gonadotropin OI-IUI (Fig. 2A) (20, 22, 23, 25). In fixed-effect modeling, live birth was increased with P support (RR 1.77, 95% CI 1.30–2.42, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .73) and the I2 index (I2 value ¼ 0%, 95% CI 0%–84%). Six studies reported on clinical pregnancy per cycle with gonadotropin OI, totaling 2,220 OI-IUI cycles (19–23,25). In fixed-effect modeling, clinical pregnancy was increased with P support (RR 1.44, 95% CI 1.18–1.75, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .93) and the I2 index (I2 value ¼ 0%, 95% CI 0%–74%). Five studies reported live birth on a per-cycle basis, totaling 2,120 gonadotropin OI-IUI cycles (Fig. 2B) (19, 20, 22, 23, 25). In fixed-effect modeling, live birth was increased with P support (RR 1.59, 95% CI 1.24–2.04, P< .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .21) and the I2 index (I2 value ¼ 0%, 95% CI 0%–85%). The risk difference in live birth rate per patient in gonadotropin OI-IUI cycles was 9.7% higher in P-supported patients (risk difference 9.7%, 95% CI 4.7%–14.7%). The number of patients needed to treat with P support in gonadotropin OIIUI cycles is 11 to have 1 additional live birth (NNT 95% CI 7–22 patients). The risk difference in live birth rate per cycle with gonadotropin OI-IUI was 5.4% higher in P-supported 5
ORIGINAL ARTICLE: INFERTILITY
FIGURE 1
A
Gonadotropins clinical pregnancy per patient
Author
Sample size
Measure (CI)
Weight %
P value
Erdem
214
1.66 (1.1; 2.51)
34.47%
0.02
Maher
71
1.61 (0.77; 3.35)
11.29%
0.2
Peeraer
393
1.53 (0.92; 2.54)
29.22%
0.1
Seckin
149
1.6 (0.8; 3.21)
14.19%
0.19
Karadag
100
1.25 (0.54; 2.9)
10.83%
0.6
Synthesis
927
1.56 (1.21; 2.02)
100%
<0.001 0.5
B
Favors no luteal support
Gonadotropins + clomiphene clinical pregnancy per patient
Author
Sample size
Measure (CI)
Weight %
P value
Agha-Hosenni
66
1.54 (0.65; 3.65)
13.28%
0.32
Hossein
253
1.24 (0.67; 2.28)
33.53%
0.49
Ebrahimi
200
1.2 (0.77; 1.88)
53.19%
0.42
Synthesis
519
1.26 (0.9; 1.76)
100%
0.17 0.5
C
Favors no luteal support
Clomiphene clinical pregnancy per patient
Author
Sample size
Measure (CI)
Weight %
P value
Kyrou
468
0.83 (0.44; 1.55)
66.36%
0.56
Agha-Hosenni
38
1 (0.29; 3.43)
13.45%
1
Karadag
100
0.83 (0.27; 2.55)
20.18%
0.75
Synthesis
606
0.85 (0.52; 1.41)
100%
0.53
0.25
1 RR
1 RR
0.5
Favors no luteal support
2
4
Favors luteal support
2
4
Favors luteal support
1 RR
2
4
Favors luteal support
Forrest plot of (A) clinical pregnancy with gonadotropins, (B) clinical pregnancy with CC plus gonadotropins, (C) clinical pregnancy with CC. All data reported on a per-patient basis. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
cycles than nonsupported cycles (risk difference 5.4%, 95% CI 2.7%–8.1%). The number of IUI cycles needed to treat with P support in gonadotropin OI-IUI cycles is 19 to have 1 additional live birth (NNT 95% CI 13–37 cycles). There is level 1 evidence that P luteal support increases live birth in gonadotropin OI-IUI cycles. CC plus gonadotropins. Three studies evaluating CC plus gonadotropin OI-IUI reported clinical pregnancy on a perpatient basis, totaling 519 patients undergoing OI-IUI (Fig. 1B) (18, 26, 27). In fixed-effect modeling, clinical pregnancy was unchanged with P support (RR 1.26, 95% CI 0.90–1.76, P¼ .17). Minimal heterogeneity was suggested by the results of the Q test (P¼ .87) and the I2 index (I2 value ¼ 0%, 95% CI 0%–89%). None of these three studies reported on live birth. Three studies evaluating CC plus gonadotropin OI-IUI reported clinical pregnancy on a per-cycle basis, totaling 830 cycles (18, 26, 27). In fixed-effect modeling, clinical pregnancy was unchanged with P support (RR 1.25, 95% CI 6
0.88–1.77, P¼ .21). Minimal heterogeneity was suggested by the results of the Q test (P¼ .83) and the I2 index (I2 value ¼ 0%, 95% CI 0%–89%). None of these three studies reported on live birth. The combined clinical pregnancy rate per patient was 23.1% in the P-supported group vs. 18.5% in the control groups. If this is the true difference in clinical pregnancy in clomiphene plus gonadotropin OI cycles, 2,502 subjects would be needed to detect an advantage in clinical pregnancy with P support. The current literature lacks the sample size to detect such a difference. There is insufficient evidence that P support is beneficial in clomiphene plus gonadotropin OI-IUI cycles. Letrozole plus gonadotropins. A single study evaluated letrozole plus gonadotropin OI-IUI (26). In 90 patients, the clinical pregnancy rate was 25.5% with P vs. 14.8% without P (P¼ .43). This study did not report on live birth. There is insufficient evidence to evaluate the effect of P luteal support in letrozole plus gonadotropin OI-IUI cycles. VOL. - NO. - / - 2017
Fertility and Sterility®
FIGURE 2
A Gonadotropins live birth per patient Author
Sample size
Measure (CI)
Weight %
P value
Erdem
214
1.98 (1.23; 3.19)
42.46%
0.01
Maher
71
3.06 (0.92; 10.2)
6.71%
0.07
Peeraer
393
1.58 (0.91; 2.73)
32.14%
0.1
Seckin
149
1.4 (0.68; 2.88)
18.68%
0.36
Synthesis
827
1.77 (1.3; 2.42)
100%
<0.001
0.5
B Gonadotropins live birth per cycle
Favors no luteal support
Author
Sample size
Measure (CI)
Weight %
P value
Erdem
427
1.88 (1.12; 3.14)
22.98%
0.02
Maher
258
3.41 (1.53; 7.6)
9.46%
0
Peeraer
393
1.58 (0.91; 2.73)
20.12%
0.1
Seckin
149
1.4 (0.68; 2.88)
11.69%
0.36
Romero Nieto
893
1.23 (0.81; 1.86)
35.75%
0.33
Synthesis
2,120
1.59 (1.24; 2.04)
100%
<0.001
0.5 Favors no luteal support
1 RR
1 RR
2
4
8
16
Favors luteal support
2
4
8
Favors luteal support
Forrest plot of (A) live birth per patient with gonadotropins, and (B) live birth per cycle with gonadotropins. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
Clomiphene citrate. Three studies evaluating CC OI reported clinical pregnancy on a per-patient basis, totaling 606 patients OI-IUI cycles (Fig. 1C) (21, 24, 26). In fixed-effect modeling, clinical pregnancy was unchanged with P support (RR 0.85, 95% CI 0.52–1.41, P¼ .53). Minimal heterogeneity was suggested by the results of the Q test (P¼ .97) and the I2 index (I2 value ¼ 0%, 95% CI 0%–89%). None of the studies on clomiphene reported on live birth outcomes. The combined clinical pregnancy rate per patient was 8.3% in the P-supported group vs. 9.8% in the control groups. If the true clinical pregnancy rate with clomiphene OI-IUI is 10% without P support, 1,372 subjects would be needed to detect an increase in 5% in pregnancy with P support. The current literature lacks the sample size to detect such a difference. Furthermore, these three studies suggest that such a difference in unlikely to exist. There is no evidence that P support is beneficial in CC OI-IUI cycles.
Sensitivity analysis. Sensitivity analyses were performed excluding studies that allowed subjects to cross over from one treatment group to another during the study period. In a fixed-effect model, sensitivity analysis of live birth in gonadotropin cycles of three studies including 765 patients without group cross-over (20, 22, 23), live birth was increased with P support (RR 1.69, 95% CI 1.23–2.34, P¼ .001). Minimal heterogeneity was suggested by the results of the Q test (P¼ .73) and the I2 index (I2 value ¼ 0%, 95% CI 0%–90%). The effect of P support on live birth in gonadotropin cycles was similar in analysis of per patient non–cross-over reporting (RR 1.69), per patient allowing cross-over reporting (RR 1.77), and per cycle reporting (RR 1.62).
Letrozole. A single study evaluated letrozole OI-IUI (26). In 96 patients, the clinical pregnancy rate was 23.2% with P vs. 12.5% without P (P¼ .28). This study did not report on live birth. There is insufficient evidence to evaluate the effect of P luteal support in letrozole OI-IUI cycles.
This is the largest systematic review and meta-analysis to date of RCTs evaluating clinical pregnancy and live birth after OIIUI. Our previous meta-analysis from 2013 demonstrated a benefit of exogenous P after gonadotropin OI-IUI (13). The number of studies and subjects available for analysis has
VOL. - NO. - / - 2017
DISCUSSION
7
ORIGINAL ARTICLE: INFERTILITY more than doubled in the past 3 years. Such an increase in the available data necessitated a reassessment of the prior metaanalysis. This updated meta-analysis included 11 RCTs and found similar results, confirming that P supplementation improves clinical pregnancy and live birth in gonadotropin OIIUI cycles. There does not seem to be a benefit of exogenous P in CC OI-IUI cycles. There is insufficient evidence that P support improves outcomes in OI-IUI cycles using letrozole or CC plus gonadotropins. Embryo implantation and the maintenance of pregnancy require a receptive endometrium that has undergone differentiation into the secretory phase under the direct influence of P (1, 28). Insufficient P in the luteal phase may result in inadequate secretory changes of the endometrium and subsequent embryo implantation failure (27). It is well established that luteal phase support is necessary in ART cycles given the inhibition of normal pulsatile LH release from the pituitary gland in the setting of supraphysiologic serum hormone levels, which results in insufficient P production from corpus lutea (6, 7). The use of luteal support in the form of exogenous hCG or P have been proposed, though P is favored given the risks of ovarian hyperstimulation syndrome with hCG (11). Compared with ART, ovulation induction for IUI is generally associated with the development of fewer follicles, lower E2 levels, the absence of exogenous pituitary down-regulation, and the lack of mechanical disruption of the corpus luteum that occurs during oocyte retrieval. Therefore, it is less clear whether corpus luteum function and P release are compromised by ovulation induction, and data regarding the utility of P supplementation after OI-IUI have been conflicting. Prior studies have been limited by small sample sizes, and most were not powered to detect statistically significant differences in clinical pregnancy or live birth rates between patients who received exogenous P after IUI and those who did not (17–27). The observed benefit of P luteal support in gonadotropin OI-IUI cycles but not in clomiphene cycles in this meta-analysis may be explained by differences in drug mechanism of action and physiologic effects. Clomiphene citrate initiates its ovulation-inducing effect at the level of the hypothalamus, competitively binding to estrogen receptors and preventing negative feedback of E2. This lack of negative feedback increases GnRH pulse frequency, stimulating the release of LH and FSH from the anterior pituitary (29, 30). Increased LH may enhance corpus luteum function and increase both E2 and P in the luteal phase (31–33). Conversely, gonadotropins directly stimulate the ovaries to produce E2, which results in negative feedback at the levels of the hypothalamus and pituitary (34, 35). This may disrupt the normal pulsatile release of LH from the pituitary and impair P secretion from the corpus luteum (36, 37). A shortened luteal phase has been demonstrated after OI with gonadotropins and has been associated with insufficient luteal phase serum P levels (23, 38). Therefore, there is biologic plausibility for the benefit of exogenous P for luteal support in gonadotropin OI-IUI cycles. Seckin et al. (20) reported a higher clinical pregnancy rate with luteal support when multifollicular development occurred. However, Peeraer et al. (22) were unable to confirm this association. 8
The studies in this meta-analysis that report follicle numbers typically had average cohorts of approximately two follicles per patient, and the gonadotropin dose was low (75–150 IU) in most studies. The majority of studies did not provide E2 levels and did not perform regression analysis to determine whether higher E2 levels were associated with a reduction in pregnancy in nonsupplemented cohorts. Although the disruption of pulsatile P release in gonadotropin OI-IUI is biologically plausible, this meta-analysis cannot prove that the mechanism of improved live birth with P luteal support is related to higher numbers of follicles or higher E2 levels disrupting the luteal phase after gonadotropin OI-IUI. It is also possible that P or other sex steroids could inhibit LH in the luteal phase or that multifollicular development could be associated with endometrial effects detrimental to pregnancy (23). Given the relatively low cost and side effect profile of vaginal P, we recommend supplementing P in all patients undergoing gonadotropin OI-IUI. There is no level 1 evidence to support measuring serum P levels in the luteal phase and only supplementing P in patients below a certain threshold. The main strength of this study is the large number of analyzed RCTs evaluating the benefit of P luteal supplementation after OI-IUI. Analyses were performed using intentto-treat data when available, and primary outcomes were reported per patient, which provides the most valuable information for clinical decision making and counseling purposes. Sensitivity analyses on a per-cycle basis showed similar results. Two studies allowed patient cross-over between study arms and failed to appropriately account for this in their statistical analyses (19, 25). However, sensitivity analyses excluding these two studies had minimal effect on the meta-analysis results. This study evaluated not only clinical pregnancy but also live birth, which is the main outcome of interest in studies evaluating the impact of an intervention on reproductive outcomes after fertility treatment. Analytic strengths and limitations of this meta-analysis are similar to those of our previous meta-analysis and were previously described in detail (13). Briefly, strengths include the performance of multiple sensitivity analyses, which controls for interstudy variance in the number of cycles per patient and ensures results were not impacted by unit of analysis bias (39). Furthermore, both random-effects and fixed-effect models were used to account for clinical and statistical heterogeneity between studies (16). The clinical heterogeneity among the 11 studies (ovulation induction agent, medication dose and duration, type and dose of hCG for ovulation trigger, and the timing, duration, and specific formulation of the vaginal P support) is a limitation. However, after random-effects modeling, secondary sensitivity analyses, and subgroup analyses, the study results remained significant, supporting the conclusion that P luteal support improves clinical outcomes after gonadotropin OI-IUI. It is unlikely that the drug used for ovulation trigger affected the observed differences, because all studies used a form of hCG, and it has been shown that urinary and recombinant hCG have equivalent clinical outcomes when used for ovulation trigger in IVF (40). The studies evaluating gonadotropin OI-IUI were almost exclusively composed of patients with a diagnosis of unexplained infertility, and most studies had VOL. - NO. - / - 2017
Fertility and Sterility® an average of two follicles per patient. Therefore, it is not likely that there were differences in ovarian stimulation goals (ovulation induction vs. superovulation) that affected the conclusion in the gonadotropin analysis. Although gonadotropin cycles are associated with a higher risk of multiple pregnancy (41), which could impact results if the multiple pregnancy rate differed between groups, no study in this meta-analysis reported a higher multiple gestation rate in the P-supported cohorts. The small numbers of patients who underwent OI-IUI with either CC plus gonadotropins or letrozole is a limitation of this study. As a result, there are insufficient data to form conclusions regarding the utility of luteal phase P support in these cycles. Limitations of these data have recently been reviewed by Quass and Hansen (42). The current data are clinically heterogeneous, lack placebo control, lack double-blind design, and arise from single-center study designs. This raises the potential need for large, double-blinded, placebo-controlled, multicenter trials to remove potential sources of bias that exist in current data. In conclusion, level 1 RCT data demonstrate that exogenous P during the luteal phase improves both clinical pregnancy and live birth rates in patients undergoing gonadotropin ovulation induction and IUI. These data strengthen the concept that there may be physiologic differences in luteal function after ovulation induction with gonadotropins compared with CC and that the former group benefits from P supplementation. The data do not demonstrate improved outcomes with P luteal support in CC or letrozole ovulation induction cycles.
9. 10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
REFERENCES 1.
2. 3.
4.
5.
6.
7.
8.
Large MJ, DeMayo FJ. The regulation of embryo implantation and endometrial decidualization by progesterone receptor signaling. Mol Cell Endocrinol 2012;358:155–65. de Carvalho JF, Caleiro MT. Primary antiphospholipid syndrome and thyroid involvement. J Clin Rheumatol 2010;16:164–7. Young SL. Oestrogen and progesterone action on endometrium: a translational approach to understanding endometrial receptivity. Reprod Biomed Online 2013;27:497–505. Mutlu MF, Erdem M, Erdem A, Mutlu I, Guler I, Demirdag E. The impact of premature progesterone rise on the outcome of intrauterine insemination cycles with controlled ovarian hyperstimulation in unexplained infertility. Eur J Obstet Gynecol Reprod Biol 2016;203:44–8. Palomino WA, Fuentes A, Gonzalez RR, Gabler F, Boric MA, Vega M, et al. Differential expression of endometrial integrins and progesterone receptor during the window of implantation in normo-ovulatory women treated with clomiphene citrate. Fertil Steril 2005;83:587–93. Nippoldt TB, Reame NE, Kelch RP, Marshall JC. The roles of estradiol and progesterone in decreasing luteinizing hormone pulse frequency in the luteal phase of the menstrual cycle. J Clin Endocrinol Metab 1989;69:67–76. Beckers NG, Macklon NS, Eijkemans MJ, Ludwig M, Felberbaum RE, Diedrich K, et al. Nonsupplemented luteal phase characteristics after the administration of recombinant human chorionic gonadotropin, recombinant luteinizing hormone, or gonadotropin-releasing hormone (GnRH) agonist to induce final oocyte maturation in in vitro fertilization patients after ovarian stimulation with recombinant follicle-stimulating hormone and GnRH antagonist cotreatment. J Clin Endocrinol Metab 2003;88:4186–92. Messinis IE, Templeton AA. Endocrine and follicle characteristics of cycles with and without endogenous luteinizing hormone surges during superovulation induction with pulsatile follicle-stimulating hormone. Hum Reprod 1987;2:11–6.
VOL. - NO. - / - 2017
20.
21.
22.
23.
24.
25.
26.
27.
Messinis IE. Ovarian feedback, mechanism of action and possible clinical implications. Hum Reprod Update 2006;12:557–71. Fauser BC, Devroey P. Reproductive biology and IVF: ovarian stimulation and luteal phase consequences. Trends Endocrinol Metab 2003;14:236–42. van der Linden M, Buckingham K, Farquhar C, Kremer JA, Metwally M. Luteal phase support for assisted reproduction cycles. Cochrane Database Syst Rev 2015:CD009154. Practice Committee of the American Society for Reproductive Medicine. Progesterone supplementation during the luteal phase and in early pregnancy in the treatment of infertility: an educational bulletin. Fertil Steril 2008;90: S150–3. Hill MJ, Whitcomb BW, Lewis TD, Wu M, Terry N, DeCherney AH, et al. Progesterone luteal support after ovulation induction and intrauterine insemination: a systematic review and meta-analysis. Fertil Steril 2013;100: 1373–80. Tavaniotou A, Albano C, Smitz J, Devroey P. Effect of clomiphene citrate on follicular and luteal phase luteinizing hormone concentrations in in vitro fertilization cycles stimulated with gonadotropins and gonadotropinreleasing hormone antagonist. Fertil Steril 2002;77:733–7. Tavaniotou A, Albano C, Smitz J, Devroey P. Comparison of LH concentrations in the early and mid-luteal phase in IVF cycles after treatment with HMG alone or in association with the GnRH antagonist Cetrorelix. Hum Reprod 2001;16:663–7. Borenstein M, Hedges LV, Higgins JP, Rothstein HR. A basic introduction to fixed-effect and random-effect models for meta-analysis. Res Synth Methods 2010;1:97–111. Aali BS, Ebrahimipour S, Medhdizadeh S. The effectiveness of luteal phase support with cyclogest in ovarian stimulated intra uterine insemination cycles: a randomized controlled trial. Iran J Reprod Med 2013; 11:309–14. Hossein Rashidi B, Davari Tanha F, Rahmanpour H, Ghazizadeh M. Luteal phase support in the intrauterine insemination (IUI) cycles: a randomized double blind, placebo controlled study. J Fam Reprod Health 2014;8:149– 53. ~ozRomero Nieto MI, Lorente Gonzalez J, Arjona-Berral JE, Del Mun Villanueva M, Castelo-Branco C. Luteal phase support with progesterone in intrauterine insemination: a prospective randomized study. Gynecol Endocrinol 2014;30:197–201. Seckin B, Turkcapar F, Yildiz Y, Senturk B, Yilmaz N, Gulerman C. Effect of luteal phase support with vaginal progesterone in intrauterine insemination cycles with regard to follicular response: a prospective randomized study. J Reprod Med 2014;59:260–6. Karadag B, Dilbaz B, Karcaaltincaba D, Sahin EG, Ercan F, Karasu Y, et al. The effect of luteal-phase support with vaginal progesterone on pregnancy rates in gonadotropin and clomiphene citrate/intra-uterine insemination cycles in unexplained infertility: a prospective randomised study. J Obstet Gynaecol 2016;36:794–9. Peeraer K, D’Hooghe T, Laurent P, Pelckmans S, Delvigne A, Laenen A, et al. Impact of luteal phase support with vaginal progesterone on the clinical pregnancy rate in intrauterine insemination cycles stimulated with gonadotropins: a randomized multicenter study. Fertil Steril 2016;106:1490–5. Erdem A, Erdem M, Atmaca S, Guler I. Impact of luteal phase support on pregnancy rates in intrauterine insemination cycles: a prospective randomized study. Fertil Steril 2009;91:2508–13. Kyrou D, Fatemi HM, Tournaye H, Devroey P. Luteal phase support in normoovulatory women stimulated with clomiphene citrate for intrauterine insemination: need or habit? Hum Reprod 2010;25:2501–6. Maher MA. Luteal phase support may improve pregnancy outcomes during intrauterine insemination cycles. Eur J Obstet Gynecol Reprod Biol 2011; 157:57–62. Agha-Hosseini M, Rahmani M, Alleyassin A, Safdarian L, Sarvi F. The effect of progesterone supplementation on pregnancy rates in controlled ovarian stimulation and intrauterine insemination cycles: a randomized prospective trial. Eur J Obstet Gynecol Reprod Biol 2012;165:249–53. Ebrahimi M, Asbagh FA, Dervish S. The effect of luteal phase support on pregnancy rates of the stimulated intrauterine insemination cycles in couples with unexplained infertility. Int J Fertil Steril 2010;4:51–6.
9
ORIGINAL ARTICLE: INFERTILITY 28.
29.
30.
31.
32.
33. 34. 35.
10
Johannisson E, Parker RA, Landgren BM, Diczfalusy E. Morphometric analysis of the human endometrium in relation to peripheral hormone levels. Fertil Steril 1982;38:564–71. Dickey RP, Vorys N, Stevens VC, Besch PK, Hamwi GJ, Ullery JC. Observations on the mechanism of action of clomiphene (MRL-41). Fertil Steril 1965;16: 485–94. Kerin JF, Liu JH, Phillipou G, Yen SS. Evidence for a hypothalamic site of action of clomiphene citrate in women. J Clin Endocrinol Metab 1985; 61:265–8. Hammond MG, Talbert LM. Clomiphene citrate therapy of infertile women with low luteal phase progesterone levels. Obstet Gynecol 1982;59:275–9. Fukuma K, Fukushima T, Matsuo I, Mimori H, Maeyama M. A graduated regimen of clomiphene citrate: its correlation to glycogen content of the endometrium and serum levels of estradiol and progesterone in infertile patients at the midluteal phase. Fertil Steril 1983;39:780–4. Shirai E, Iizuka R, Notake Y. Clomiphene citrate and its effects upon ovulation and estrogen. Fertil Steril 1972;23:331–8. Tavaniotou A, Smitz J, Bourgain C, Devroey P. Ovulation induction disrupts luteal phase function. Ann N Y Acad Sci 2001;943:55–63. Tavaniotou A, Albano C, Smitz J, Devroey P. Impact of ovarian stimulation on corpus luteum function and embryonic implantation. J Reprod Immunol 2002;55:123–30.
36.
37.
38.
39.
40.
41.
42.
Rossmanith WG, Laughlin GA, Mortola JF, Johnson ML, Veldhuis JD, Yen SS. Pulsatile cosecretion of estradiol and progesterone by the midluteal phase corpus luteum: temporal link to luteinizing hormone pulses. J Clin Endocrinol Metab 1990;70:990–5. Rossmanith WG, Laughlin GA, Mortola JF, Yen SS. Secretory dynamics of oestradiol (E2) and progesterone (P4) during periods of relative pituitary LH quiescence in the midluteal phase of the menstrual cycle. Clin Endocrinol (Oxf) 1990;32:13–23. Olson JL, Rebar RW, Schreiber JR, Vaitukaitis JL. Shortened luteal phase after ovulation induction with human menopausal gonadotropin and human chorionic gonadotropin. Fertil Steril 1983;39:284–91. van der Poel N, Farquhar C, Abou-Setta AM, Benschop L, Heineman MJ. Soft versus firm catheters for intrauterine insemination. Cochrane Database Syst Rev 2010:CD006225. Youssef MA, Abou-Setta AM, Lam WS. Recombinant versus urinary human chorionic gonadotrophin for final oocyte maturation triggering in IVF and ICSI cycles. Cochrane Database Syst Rev 2016:CD003719. Diamond MP, Legro RS, Coutifaris C, Alvero R, Robinson RD, Casson P, et al. Letrozole, gonadotropin, or clomiphene for unexplained infertility. N Engl J Med 2015;373:1230–40. Quaas AM, Hansen KR. The role of steroid hormone supplementation in non-assisted reproductive technology treatments for unexplained infertility. Fertil Steril 2016;106:1600–7.
VOL. - NO. - / - 2017
Fertility and Sterility®
SUPPLEMENTAL ADDENDUM: DETAILED SEARCH STRATEGY The search strategy consisted of the following phrases: (((((progesterone[tiab] OR progesterone[mh])) AND (luteal [tiab] OR luteal phase[mh])) AND (‘‘intrauterine insemination’’[tiab] OR IUI[tiab] OR insemination, artificial[mh] OR ovulation induction[mh] OR ‘‘ovarian stimulation’’[tiab] OR (ovar*[tiab] AND stimulat*[tiab]) OR ovulat*[tiab]))) AND
VOL. - NO. - / - 2017
(randomized controlled trial[pt] OR randomized controlled trial[mh] OR single-blind method[mh] OR double-blind method[mh] OR random allocation[mh] OR random*[tiab] OR ‘‘single blind’’[tiab] OR ‘‘double blind’’[tiab] OR placebo [tiab] OR ‘‘randomized controlled trial’’[tiab] OR controlled clinical trial[pt] OR ‘‘controlled clinical trial’’[tiab] OR (doubl*[tiab] AND blind*[tiab]) OR (singl*[tiab] AND blind* [tiab])).
10.e1
Assessment of bias in the included RCTs, including assessments of randomization, allocation concealment, blinding, data reporting, and declaration of conflicts of interest. Authors (reference) Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahimi et al. 2010 (27) Maher 2011 (25)
Agha-Hosseini et al. 2012 (26) Aali et al. 2013 (17) Romero Nieto et al. 2014 (19)
Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18) Karadag et al. 2016 (21) Peeraer et al. 2016 (22)
Randomization
Allocation concealment
Software-generated random Single researcher sequence aware, treating providers blinded Software-generated random None sequence Sequential randomization None Software-generated random None sequence; patients alternated treatment arms every subsequent cycle Software-generated random None sequence Block randomization (four None patients per block) Software-generated random None sequence; patients were re-randomized in subsequent treatment cycles Software-generated random None sequence Software-generated random Yes sequence Software-generated random None sequence Software-generated block Yes randomization (10 patients per block)
Green. Progesterone luteal support after IUI. Fertil Steril 2017.
Blinding of Blinding of Incomplete participants outcomes data reporting
Analysis type
Conflicts of interest or pharma sponsorship
Trial registry
None
None
Yes
Intent to treat
None
Not stated
None
None
Yes
Not stated
Yes
None None
None None
None Yes
Intent to treat and per protocol Per protocol Per protocol
None Not stated
Yes Not stated
None
None
Yes
Per protocol
Not stated
Yes
None
None
Yes
Intent to treat
None
Yes
None
None
Yes
Intent to treat but re-randomized with each cycle
None
Not stated
None
None
Yes
Intent to treat
Not stated
Not stated
Yes
None
None
Intent to treat
None
Not stated
None
None
Yes
Intent to treat
None
Not stated
None
None
Yes
Intent to treat and per protocol
One author is vice president Yes of Global Medical Affairs Fertility at Merck
ORIGINAL ARTICLE: INFERTILITY
10.e2
SUPPLEMENTAL TABLE 1
VOL. - NO. - / - 2017
VOL. - NO. - / - 2017
SUPPLEMENTAL TABLE 2 Fertility history, ovarian reserve, and stimulation results reported between the studies. Authors (reference) Erdem et al. 2009 (23) Kyrou et al. 2010 (24) Ebrahami et al. 2010 (27) Maher 2011 (25) Agha-Hosseini et al. 2012 (26) Aali et al. 2013 (17) Romero Nieto et al. 2014 (19) Seckin et al. 2014 (20) Hossein Rashidi et al. 2014 (18) Karadag et al. 2016 (21)
Peeraer et al. 2016 (22)
Group
Age (y)
P Control P Control P Control P Control P Control P Control P Control P Control P Control P (CC arm) Control (CC arm) P (Gnd arm) Control (Gnd arm) P Control
30 4.8 29.7 4.3 32.1 3.6 32.2 3.9 27.9 3.3 28.4 4.1 NR NR 27.4 3.7 26.8 3.6 28b 27.9b 32.9 3.4 33.1 3.3 27.8 5.2 28.9 5.5 27.53 4.23 28.11 4.62 26.3 4.23 25.7 4.2
Green. Progesterone luteal support after IUI. Fertil Steril 2017.
NS NS NS NR NS NS NS NS NS NS
Primary infertility (%) 64.2 63.8 71.4 69.6 NR NR NR NR NR NR 89.9 88.7 100 100 88.7 96.2 NR NR 66 48
P value NS NS NR NR NR NS NS NS NR NS
Day-3 FSH (IU/L) 7.3 2.7 7 2.7 5.4 3.1 5.1 2.1 6.1 2.7 6.7 2.0 NR NR 5.4 3.1 5.1 2.1 6.9 2.4 7.2 2.1 NR NR 6.7 2.2 6.8 2.3 5.98 2.33 6.38 2.56 6.4 1.7 5.8 1.6
P value NS NS NS NR NS NS NR NS NS NS
Peak E2 (pg/mL) NR NR 513 248 504 306 NR NR NR NR NR NR NR NR NR NR 597.7 323.4 670.4 434.5 NR NR NR NR
P value NR NS NR NR NR NR NR NS NR NR
Dominant folliclesa 1.6 0.6 1.5 0.9 1.2 0.3 1.3 0.4 2.0 0.7 2.2 0.8 NR NR 1.6 0.7 1.4 0.8 NR NR 1.4c 1.4c 2 1.3 2.3 1.8 2.21 1.29 2.43 1.31 NR NR
P value NS .02 NS NR NS NR NR NS NS NR
28.5 3.6 27.8 3.7
NS
54 58
NS
7.1 4.2 6.8 2.5
NS
NR NR
NR
NR NR
NR
31 3.97 31.5 3.8
NS
71 63
NS
NR NR
NR
289 (266–311)d 311 (288–334)d
NS
1.2 0.1 1.2 0.1
NS
10.e3
Fertility and Sterility®
Note: Data listed as mean SD unless otherwise stated. a Defined as follicles over 16-18 mm. b Range not reported. c SD not reported. d Mean (95% CI).
P value
ORIGINAL ARTICLE: INFERTILITY
Eligibility
Screening
Identification
SUPPLEMENTAL FIGURE 1 63 records identified through database searching
4 additional records identified through other sources
56 records after duplicates removed
56 records screened
15 full-text articles assessed for eligibility
Included
11 studies included in qualitative synthesis
41 records excluded
4 full-text articles excluded: 1 article was an abstract reported in an included paper 3 articles did not have a progesterone luteal support arm
11 studies included in quantitative synthesis (meta-analysis)*
PRISMA (Preferred Reporting Items for Systematic Reviews and MetaAnalyses) flow diagram of study selection. *Eleven studies were included in per-cycle analyses and ten studies were included in perpatient analyses (Romero Nieto et al 2014 (19) was excluded from per-patient analyses due to re-randomization of patients in subsequent cycles). Green. Progesterone luteal support after IUI. Fertil Steril 2017.
10.e4
VOL. - NO. - / - 2017
Fertility and Sterility®
SUPPLEMENTAL FIGURE 2
Selectivity funnel plot of all studies included in the analysis for assessment of publication bias. Green. Progesterone luteal support after IUI. Fertil Steril 2017.
VOL. - NO. - / - 2017
10.e5