Conventional GnRH antagonist protocols versus long GnRH agonist protocol in IVF/ICSI cycles of polycystic ovary syndrome women: a systematic review and meta-analysis

Abstract

Gonadotropin-releasing hormone (GnRH) analogues are commonly used in clinical practice to prevent premature luteinizing hormone (LH) surge during In-Vitro Fertilization/ Intra-Cytoplasmic Sperm Injection (IVF/ICSI) cycles. This review aimed to summarize the available evidence comparing the effects of conventional GnRH antagonist protocols, the most commonly used GnRH antagonist protocols, and GnRH agonist protocols on IVF/ICSI outcomes in women with polycystic ovary syndrome (PCOS). A comprehensive electronic search was carried out in Pubmed, Cochrane CENTRAL, Scopus, Web of Science, CINAHL, TRIP, ClinicalTrials.gov and ISRCTN registry from inception until 24 November 2020 without any language or date restrictions. In addition, reference lists of eligible studies and previous meta-analyses were hand-searched to identify relevant studies. Eligible randomized controlled trials were those designed to compare the effects of conventional GnRH antagonist protocols and GnRH agonist protocols on IVF/ICSI outcomes in PCOS subjects. The Cochrane ROB 2.0 tool was used to assess the risk of bias of each study, and the GRADE assessment was used to evaluate the overall quality of evidence. Data synthesis and analyses were done using Review Manager 5.3 with the assistance of Revman Web. A random-effects model was used for all meta-analysis. Dichotomous outcomes were reported as Relative Risk (RR) and continuous outcomes as Weighted Mean Difference (WMD), both with 95% CIs. The primary outcomes were Live birth rate, Ongoing pregnancy rate, and Ovarian hyperstimulation syndrome (OHSS) rate. Other IVF outcomes were considered secondary outcomes. We included ten studies with 1214 randomized PCOS women. Using GnRH antagonist protocols led to a significantly lower OHSS rate (RR = 0.58; 95% CI: [0.44 to 0.77], P = 0.0002), shorter stimulation duration (WMD = - 0.91; 95% CI: [-1.45 to - 0.37] day, P = 0.0009), lower gonadotropin consumption (WMD = - 221.36; 95% CI: [- 332.28 to - 110.45] IU, P < 0.0001), lower E2 levels on hCG day (WMD = - 259.21; 95% CI: [- 485.81 to - 32.60] pg/ml, P = 0.02), thinner endometrial thickness on hCG day (WMD = - 0.73; 95% CI: [- 1.17 to - 0.29] mm, P = 0.001), and lower number of retrieved oocytes (WMD = - 1.82; 95% CI: [- 3.48 to - 0.15] oocytes, P = 0.03). However, no significant differences in live birth rate, ongoing pregnancy rate, clinical pregnancy rate, multiple pregnancy rate, miscarriage rate and cycle cancellation rate were seen between the GnRH antagonist protocols and the long GnRH agonist one. Although more cycles were cancelled due to poor ovarian response in the GnRH antagonist protocol (RR = 4.63; 95% CI: [1.49 to 14.41], P = 0.008), similar rates of cancellation due to risk of OHSS were noticed in both groups. The differences in IVF/ICSI outcomes may arise from the different patterns of gonadotropins suppression that the GnRH analogues exhibit during the early follicular phase of IVF/ICSI cycles and the divergent direct impacts of these analogues on ovaries and endometrial receptivity. The main evidence limitation was Imprecision. Conventional GnRH antagonist protocols represent a safer and more cost-effective treatment choice for PCOS women undergoing IVF/ICSI cycles than the standard long GnRH agonist protocol without compromising the IVF/ICSI clinical outcomes. The study had no sources of financial support and was prospectively registered at PROSPERO (International Prospective Register of Systematic Reviews) under registration number (CRD42021242476).

Figure 1

Flow diagram selection process.

Figure 2

Forest Plot: Live Birth Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 3

Forest Plot: Ongoing Pregnancy Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 4

Forest Plot: OHSS Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 5

Forest Plot: OHSS Rate Per Randomized Woman (Per OHSS Grade). (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 6

Forest Plot: Clinical Pregnancy Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 7

Forest Plot: Multiple Pregnancy Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 8

Forest Plot: Miscarriage Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 9

Forest Plot: Cycle Cancellation Rate Per Randomized Woman. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 10

Forest Plot: Cycle Cancellation Rate Per Randomized Woman (Per Cause of Cancellation). (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 11

Forest Plot: Stimulation Duration. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 12

Forest Plot: Gonadotropin Dose. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 13

Forest Plot: E2 Levels on hCG Day. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 14

Forest Plot: Endometrial Thickness on hCG Day. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 15

Forest Plot: Number of Retrieved Oocytes. (A) Bias arising from the randomization process; (B) Bias due to deviations from intended interventions; (C) Bias due to missing outcome data;(D) Bias in measurement of the outcome; (E) Bias in selection of the reported result and (F) overall bias.

Figure 16

Funnel Plot: Stimulation duration.