Luteal phase support strategies in ovarian stimulation–intrauterine insemination cycles: a systematic review and network meta-analysis of randomized controlled trials

Background Luteal phase support (LPS) is widely used in ovarian stimulation–intrauterine insemination (OS-IUI) cycles, but the comparative effectiveness of different LPS strategies remains uncertain. This study aimed to compare the effects of available LPS strategies on reproductive outcomes following OS-IUI. Methods PubMed, Embase, Web of Science, the Cochrane Library, Scopus, and ClinicalTrials.gov were searched from inception to August 7, 2026 for randomized controlled trials (RCTs) evaluating pharmacological LPS after OS-IUI. The primary outcome was clinical pregnancy, and secondary outcomes included ongoing pregnancy, live birth, biochemical pregnancy, multiple pregnancy, and miscarriage. Frequentist random-effects network meta-analysis was performed for outcomes forming connected evidence networks, while conventional pairwise meta-analysis was used when network analysis was not feasible. Treatment effects were expressed as risk ratios (RRs) with 95% confidence intervals (CIs). Risk of bias was assessed using the revised Cochrane Risk of Bias tool for randomized trials (RoB 2), and certainty of evidence was evaluated using the Confidence in Network Meta-Analysis (CINeMA) framework or the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach, as appropriate. Results 22 RCTs involving 4,994 women were included. For clinical pregnancy, 21 RCTs involving 4,547 women contributed to a network of 8 LPS strategies. The results showed that vaginal progesterone (VP) was associated with a significantly higher clinical pregnancy rate than no luteal phase support/placebo (NP) (RR = 1.27, 95% CI 1.09-1.50; Moderate certainty). For live birth, 4 RCTs involving 956 women showed that VP was associated with a higher live birth rate than NP (RR = 1.61, 95% CI 1.19–2.18; moderate certainty). Besides, VP combined with a GnRH agonist was associated with a lower multiple pregnancy rate than VP alone (RR = 0.29, 95% CI 0.10–0.85; very low certainty). No clear differences among LPS strategies were observed for ongoing pregnancy, biochemical pregnancy, or miscarriage. Conclusion In OS-IUI cycles, VP was associated with higher clinical pregnancy and live birth rates than NP. Compared with VP combined with a GnRH agonist, VP was associated with a higher multiple pregnancy rate, although the certainty of evidence was very low and this finding should be interpreted cautiously. Further well-designed RCTs are needed to confirm these findings and determine the optimal LPS strategy.