Glucocorticoid supplementation during ovulation induction for assisted reproductive technology: a systematic review and meta-analysis

Abstract Background There is ongoing interest in glucocorticoid treatment during oocyte stimulation to treat infertility in women who have undergone Assisted Reproductive Technology (ART). Objective This meta-analysis was performed to evaluate the efficiency and safety of adjuvant glucocorticoid therapy on pregnancy outcomes in infertile women undergoing ART cycles. Study design A literature search was performed in PubMed, EMBASE, Web of Science, and the Cochrane Library up to December 2022. To assess the efficacy and safety of additional glucocorticoid treatment during ovulation induction in women who underwent IVF or ICSI treatment, only randomized controlled trials were included. Results Overall, glucocorticoid therapy during ovulation showed a nonsignificant effect of prednisolone improving the live birth rate (OR = 1.03, 95% CI [.75, 1.43], I2 = .0%, p = .84), abortion rate (OR = 1.14, 95% CI [.62, 2.08], I2 = 31%, p = .68), and implantation rate (OR = 1.1, 95% CI [.82, 1.5], I2 = 8%, p = .52) of infertile women compared to the control group. The present meta-analysis revealed that the clinical pregnancy rate per cycle tended to increase after glucocorticoid treatment (OR = 1.29, 95% CI [1.02, 1.63], I2 = 8%, p = .52). Conclusions The present meta-analysis suggested that ovarian stimulation prednisolone therapy does not significantly improve clinical outcomes in women undergoing IVF/ICSI. Although the results indicated that adjuvant glucocorticoid therapy during ovarian stimulation may increase the clinical pregnancy rate, subgroup analysis showed that it was affected by infertility factors, dose schedules, and length of treatment. Therefore, these results should be interpreted with caution.


Introduction
Infertility is a global health issue affecting approximately 48 million couples, and it is defined as the inability to develop clinical pregnancy after 1 year of unprotected sexual intercourse [1].Although IVF/ICSI has revolutionized the landscape of infertility treatment, it remains far from being a panacea, and the success rates over recent years have plateaued [2].Therefore, new advancements in the development of assisted reproductive technology are needed, including increasing the ovarian response, improving endometrial receptivity, and protecting against miscarriage [3,4].Patients are experiencing a poor ovarian response (POR) to ovarian stimulation (OS), leading to cycle cancelation and a reduced chance of live birth.
In addition, high serum progesterone during the early follicular phase is another concern for IVF outcomes in infertile women [5].
Glucocorticoids (GCs) are well-known as important stress regulators and anti-inflammatory agents, and their immunosuppressant function and clinical use have elicited substantial interest in the field of reproduction [6,7].The results of several studies have suggested that GCs enhance the IVF/ICSI pregnancy rate for women with positive autoantibodies [8,9] or women with unexplained repeated pregnancy loss [10,11].An increasing number of studies implicate the role of GCs with a positive effect on IVF/ICSI outcomes.However, little information is available concerning the effect of GCs treatment during the follicular phase on ART outcomes.Preliminary studies have reported the role of GCs during ovarian stimulation with an improved pregnancy rate and decreased cancelation rates [12,13].Some researchers have proposed an empirical strategy, suggesting that adrenal suppression by GCs may be beneficial in certain anovulatory patients [14].In women with polycystic ovarian disease (PCOS), Farahnaz et.al reported that the pregnancy rate is significantly higher in women receiving clomiphene citrate plus dexamethasone compared to women receiving letrozole alone [15], while Bider et.al demonstrated that GCs supplementation is unnecessary during gonadotropin therapy because the action of gonadotropins eliminates the ovulation disturbances observed in women with PCOS [16].Recent studies have reported that a lowdose oral dexamethasone treatment in women with high progesterone levels in the early proliferative phase sensitizes ovaries to gonadotropin stimulation, leading to the secretion of less progesterone, and women treated with dexamethasone showed a higher live birth rate than the controls [17,18].
There is ongoing interest in GCs treatment during oocyte stimulation to treat infertility in women who have undergone ART.However, the clinical evidence for using GCs in IVF and ICSI cycles during OS is limited, and the efficiency of a GCs protocol is urgent to be determined.The purpose of the present meta-analysis was to examine the effects and safety of GCs on pregnancy outcomes during ovulation induction in women undergoing IVF and ICSI.

Search strategy
We searched PubMed, EMBASE, Cochrane Library, and WOS (through 31 December 2022).There was no language or time limitation imposed in the literature search, including International Scientific Indexing conference proceedings and databases for the registration of ongoing and archived randomized controlled trials (RCTs).This systematic review and meta-analysis were prospectively designed and had been registered in PROSPERO (Registration number: CRD42023424715).Literature was filtered utilizing combinations of Medical Subject Headings (i.e.MeSH) and text words to generate two subsets of citations for GCs ("Dexamethasone" or "Adrenocorticotropic Hormone" or "corticosteroid") and ART (" ART"; "in vitro fertilization"; "IVF"; assisted reproductive technology; "ICSI"; "infertility").Both subsets were combined with "AND'" to generate a subset of citations relevant to our analysis.In addition, we manually searched the reference lists of all known major studies and review articles to identify cited articles not captured by the electronic search.
2.2.Criteria for considering studies for this review 2.2.1.The inclusion criteria were as follows: 1) published and unpublished parallel, randomized controlled trials (RCTs); 2) any cause of infertility in autobody-negative women who underwent IVF or ICSI for at least one year; and 3) trials that compared the efficacy between a control group and a group that received supplementary GCs during stimulation.
Two researchers evaluated the selected studies and extracted the data independently.Any disagreement was resolved by discussion.First, the ineligible literature was excluded by reading the title and abstract, and the full texts were then read to identify studies to be included.The following data were extracted from the literature: first author, year of publication, study design, sample size, participants, intervention protocol, results, and other characteristics.

Exclusion criteria
The exclusion criteria were as follows: 1) quasirandomized controlled trials; 2) studies of couples undergoing frozen embryo transfers or oocyte donation cycles; 3) studies in which the administration of adjuvant GCs was limited to the peri-implantation period.These interventions are the topic of another published Cochrane review about preimplantation [19].

Study quality assessment and assessment of risk of bias
The selected studies were independently reviewed by two authors (L.L. and L.T.Q) for the quality assessment of the included studies according to the Cochrane Collaboration's criteria (version 5.1.0,Available from www. cochrane-handbook.org), by which the generation of sequence allocation, allocation concealment, blinding, incomplete outcome data, selective reporting, and other bias are assessed.Any disagreements were resolved by discussion or by a third review author (Z.X.L.).Publication bias was evaluated by a funnel plot.Subgroup analysis was performed according to infertility factors, dose schedules, and length of treatment.Sensitivity analysis was employed to determine the effect of a single study on the overall estimation.In addition to subgroup analysis, metaregression for the included studies was conducted to identify factors for heterogeneity.

Outcome measures
The primary results included live birth rate (LBR) for all populations (birth of at least one live infant per embryo transfer) and incidence of multiple pregnancies.The secondary outcomes were as follows: effectiveness, including clinical pregnancy rate (CPR, diagnosed by the presence of gestational sacs as assessed by ultrasound at 7 weeks of gestation); and implantation rate per embryo transfer.We also assessed the safety of GCs during ovarian stimulation for maternal complications and neonatal outcomes as secondary outcomes.We recorded the following maternal complications: abortion rate (mostly defined as pregnancy loss between 12 weeks and 20 weeks of gestation), multiple pregnancy rate, ectopic pregnancy rate, ovarian hyperstimulation syndrome (OHSS), gestational diabetes mellitus (GDM), and pregnancy-induced hypertension (PIH).We also analyzed neonatal outcomes, including congenital anomalies, chromosomal aberrations, and different organ system malformations.

Statistical analysis
The meta-analysis was performed using Review Manager (RevMan; version 5.4).All of the outcomes were measured in terms of the Mantel-Haenszel odds ratio (OR) with 95% confidence intervals (CI) using a fixed effects model.p < .05 was considered statistically significant.The causes of heterogeneity were evaluated by subgroup analysis according to the different control group treatment protocols.Statistical heterogeneity was evaluated by the I 2 statistic, and I 2 < 50% was considered to indicate no significant heterogeneity.If the I 2 value was >50%, which showed significant heterogeneity, a random effects model was used.The causes of heterogeneity were then investigated using variations in features of infertility factors, interventions, and study design.We also conducted a sensitivity analysis by sequentially removing one study at a time to determine whether any study affected the overall results.

Study selection
The process of literature retrieval and research selection is shown in Figure 1.We identified 7191 related publications, and 134 of these were selected after the title and the abstract were read.Among these, 125 publications were excluded as follows: 64 studies without RCTs, 20 studies without GCs treatment, 15 studies that used glucocorticoids for ovulation induction protocols and not IVF/ICSI, 18 studies that began GCs administration only before implantation, and 7 studies that were meta-analyses or reviews.In addition, the study by Ashrafi (2007) was excluded even though it met the inclusion criteria because no pregnancy rate data had been made available and we were unable to contact the authors.Finally, while the remaining 9 RCTs were considered to meet the selection criteria and were included in this meta-analysis.The quality of the included RCTs was evaluated by the risk of bias using Revman software (online Supplement Figure 1)
The GCs treatment protocol for infertile women also varied among studies as follows: GCs were given on the first day of gonadotrophin administration until the day of hCG administration in four studies (Keay 2001, Fridstr€ om 1999, Rein 1996, Shanliu 2018); GCs were given daily throughout ovarian stimulation until the pregnancy test in four studies (Ando 1996, Kemeter 1986, Mohammadi 2018 and Kim 1997); and dexamethasone was given on the first day of stimulation until embryo transfer in one study (Bider 1996).
Among these studies, seven studies used long protocols with a gonadotropin-releasing hormone (GnRH) analog, and only the study by Kemeter 1986 used an older ovarian stimulation protocol of clomiphene in combination with gonadotropins.All participants enrolled in the present meta-analysis underwent IVF, except for patients in the study by Shanliu 2018, who reported the effect of low-dose dexamethasone on patients who underwent a first IVF/ICSI cycle.

Primary outcomes
Effectiveness.Outcome 1 live birth rate.The hetero- geneity was low with regard to the effect of GCs treatment for infertile women on the abortion rate (I 2 ¼ 0%).The OR for the live birth rate is shown in online Supplement Figure 2 (Analysis 1.1).There was low certainty in determining whether there was any difference between the groups in the live birth rate (odds ratio (OR) 1.03, 95% confidence interval (CI) .75 to 1.43; 5 RCTs, n ¼ 889, I 2 ¼ 0%).
Safety.Outcome 2 multiple pregnancy rate per pregnancy.One study reported MPR in women undergoing IVF when GCs were used in the study group (Kim 1997), but there were no differences between the two groups in the multiple pregnancy rate.

Secondary outcomes
Effectiveness.Outcome 3 clinical pregnancy rate.All included studies reported clinical pregnancy rates in 502 women undergoing IVF/ICSI.The results of the pooled data were included in the meta-analysis, which demonstrated that GCs treatment played a role in increasing clinical pregnancy rate in women who underwent IVF/ICSI per cycle compared to the placebo group (OR 1.29, 95% CI 1.02-1.63,p ¼ .04;fixed effects model, Analysis 1.2; online Supplement Figure 3).A random-effects model was used to perform a sensitivity analysis.(RR 1.28, 95% CI 1.01 to 1.62).
We conducted a subgroup analysis of infertility factors and medication time to further clarify the specific impact of GCs on the clinical pregnancy rate.The results confirmed a nonsignificant improvement in the clinical pregnancy rate following GCs therapy among women undergoing ART due to tubal factor infertility only (OR 1.24, 95% CI .7 to 2.19, p ¼ .46).However, it remained uncertain whether there was an effect of oocyte stimulation GCs administration on clinical pregnancy rates in a subgroup of women with multiple unexplained infertility factors undergoing ART, including endometriosis, PCOD, high DHEAS, and progesterone concentration (OR 1.31, 95% CI 1 to 1.68; 7 RCTs, n ¼ 1254; I 2 ¼ 22%; Analysis 1.3; Figure 2).In addition, the medication time of GCs administration differed among these studies.There was a certain therapeutic effect in women who received GCs from the first day of simulation until embryo transfer or pregnancy test (OR 1.54, 95% CI 1.05 to 2.26; 5 RCTs, n ¼ 992; I2 ¼ 0%; Analysis 1.4; Figure 2(A)), but there was a nonsignificant effect in women who received GCs on the first day of gonadotrophin administration until the day of The meta-analysis showed a nonsignificant difference in the implantation rate in a total of 1354 cycles of the six RCTs between the GCs treatment and control groups (I 2 ¼8%).Thus, it remained uncertain whether adjustive GCs treatment improved the implantation rate per couple compared to no GCs/placebo.(OR ¼ 1.1; 95% CI .82 to 1.5; Analysis 1.5; Figure 3(A)).statistically significant change in the miscarriage rate per pregnancy between the GCs treatment and control groups (OR 1, 95% CI .82 to 1.5; 6 RCTs, n ¼ 821; I 2 ¼ 8%; Analysis 1.6; Figure 3(B)).Due to the limited sample size, it remained uncertain whether ovarian stimulation GCs influenced miscarriage rates per couple compared to no GCs/placebo (Bider 1996;Kemeter 1986;Kim 1997; Shanliu 2018).The random-effects model confirmed this result.Outcome 6 incidence of ectopic pregnancies per cycle.The OR for ectopic pregnancies per cycle is shown in Figure 3(C).It remained uncertain whether administered GCs increased ectopic pregnancy rates per cycle compared to no GCs/placebo (OR 4.06, 95% CI .45 to 36.65; 3 RCTs, n ¼ 1967; I 2 ¼ 0%; Analysis 1.7; Figure 3(C)).Three studies reported the ectopic pregnancy rate in women undergoing IVF/ICSI when GCs were used during OS in the study group (Kim 1997;Kemeter 1986; Shanliu 2018).In the study by Kemeter 1986, a total of 156 infertile patients were allocated randomly to the corticosteroid treatment group (73 patients) and the control group (73 patients), and the data showed no significant difference between the two groups in terms of ectopic pregnancies (2/73 vs. 0/73, respectively).The second study conducted by Shanliu 2018 showed no difference in the ectopic pregnancy rate between the study (1/230) and control groups (0/229).Kim 1997 reported no ectopic pregnancy in either the GCs or placebo group.Pooling of the results of these studies showed a higher but not statistically significant likelihood of an ectopic pregnancy rate in the GCs treatment group.Outcome 7 incidence of ovarian hyperstimulation syndrome per couple.As only one RCT reported OHSS (Mohammadi 2018), a meta-analysis was not performed.The reported study did not show a significant reduction in the incidence of OHSS in patients receiving GCs compared to those receiving placebo (19.4% vs. 16.5%).Therefore, it remains uncertain whether stimulation GCs influenced the rate of incidence of OHSS compared to placebo.

Safety of neonates. Outcome 8 neonatal outcomes.
The study by ShanLiu et al. showed that the proportions of infants with low birthweight (1.5-2.499kg) and very low birthweight (<1.5 kg) or with low Apgar scores were comparable, and none of the differences demonstrated a statistical significance between the study group and the control group [17].Outcome 9 incidence of fetal abnormalities.One study evaluated the effect of corticosteroids on the incidence of fetal abnormalities in women undergoing ART cycles (Shanliu 2018) [17], but it was not possible to perform a meta-analysis due to the heterogeneity of GCs administration and duration.The study conducted by Shan Liu 2018 enrolled 459 patients undergoing IVF and showed kidney malformation in one fetus in the DEX group, while kidney malformation and polydactyly were observed in two fetuses in the control group; they also reported one infant congenital malformation (Fallot's tetralogy) in the control group.These were cumulative reproductive outcomes in one ovulation cycle in 2 years and the number of congenital malformations, not clinical outcomes of fresh cycles.

Discussion
We performed this systematic review and meta-analysis to summarize the efficacy of GCs treatment during ovulation induction in women undergoing IVF/ICSI.The studies included in our meta-analysis were randomized controlled trials of the effect of GCs treatment during ovulation induction on IVF/ICSI outcomes.The present meta-analysis suggested that there was no significant improvement in the live birth rate, miscarriage rate, ectopic pregnancy rates per cycle, or implantation rate per embryo transfer between GCs supplementation during ovarian stimulation for IVF or ICSI and the control group.This meta-analysis was not sufficiently powerful to confirm the effect of OS GCs on the live birth rate because the subgroup analysis reached the other outcome and the sample size was small.
In 2007, Boomsma et al. published the first metaanalysis, which included 13 studies and they found that GCs therapy did not significantly improve the live birth rate or pregnancy rate.In a subgroup analysis that only included fresh IVF cycles, they found a significant increase in pregnancy rate in the GCs treatment group (OR, 1.50; 95% CI, 1.05-2.13)[19].
An updated meta-analysis on the impact of GCs in ART comes from data based on a 2012 meta-analysis, which shows no clear evidence to prove that the use of GCs during transplantation can effectively improve pregnancy outcomes [27].Our research mainly focuses on the efficacy of GCs treatment during ovulation induction in women undergoing IVF/ICSI not only for the duration of implantation.The present meta-analysis revealed that the clinical pregnancy rate per cycle tended to increase after GCs treatment.Subgroup analysis showed that it was affected by infertility factors, dose schedules and length of treatment.Therefore, these results should be interpreted with caution.At present, individual GCs adjuvant treatment is based on each specific case according to the clinical needs and actual condition of the patient.Therefore, subgroup analysis may provide more certainty of the evidence for clinical application.
For several years, potential mechanisms by which GCs may affect ovarian function have been discussed.Reportedly, GCs exert a range of positive effects on oocyte maturation and ART outcomes by enhancing 11b-HSD1 activity in adrenal glands and altering cytokine levels [28,29].In addition, dexamethasone increases serum GH and follicular fluid IGF-1 levels [30], which synergistically act with follicle-stimulating hormone (FSH) in vitro to improve the ovarian response [31].There is also evidence of no beneficial effects on ovarian responsiveness and IVF-ET outcome after suppression of adrenal androgens with GCs [26].In addition, there is increasing concern about the potential negative effects of GCs on ART outcomes.However, the congenital malformations, mode of delivery, mean gestational age, premature births, and fetal abnormalities are similar between the two groups in cumulative reproductive outcome in one ovulation cycle in 2 years, and these reported perinatal outcomes raised concerns for potential adverse effects and safety of GCs [17,32].
The present study had several limitations.The present study included only RCTs that compared GCs with placebo or no treatment during OS.Thus, the present analysis was limited by the differences in the GCs administration protocol and the lack of an adequate sample size.Moreover, the live birth rate should have been the primary outcome for the present analysis, but it was not reported in several studies [33].Therefore, we utilized the clinical pregnancy rate, which provided a less reliable surrogate endpoint.These results cannot easily be extrapolated to the live birth rate due to the high rate of multiple pregnancies, ectopic pregnancies, abortions, and perinatal mortality.Furthermore, GCs, as crucial immunomodulators, may have a higher chance of contributing to immunological imbalance, justifying the better result in the clinical outcomes [34,35].Due to the limited sample size and high heterogeneity, we excluded autoantibody-positive women, indicating that additional studies with larger sample sizes are needed in the future.
In conclusion, GCs treatment during ovulation induction showed no significant differences in clinical outcomes in ART.Future research should focus on a full analysis of not only benefits but also potential risks of side effects.Due to the small size of the included studies, additional investigations are expected to demonstrate whether GCs adjuvant therapy has the potential to improve reproductive outcomes during ovarian stimulation at present.