目的·比较宫腔镜检查后给予广谱抗生素与16S rRNA基因测序并针对性给予抗生素联合补充乳杆菌对于反复种植失败（recurrent implantation failure，RIF）合并慢性子宫内膜炎（chronic endometritis，CE）患者妊娠结局的影响。方法·回顾性分析2023年1月至2025年5月在上海交通大学医学院附属瑞金医院妇产科生殖医学中心就诊的676名排除已知因素（未行胚胎植入前非整倍体遗传学检测）的RIF患者（接受分泌中期16S rRNA测序或增殖期宫腔镜检查）的妊娠结局。27名患者16S rRNA检查结果提示乳杆菌占比<90%且单一病原体占比≥10%，纳入16S rRNA诊断CE组（16S rRNA组）；根据致病菌给予针对性抗生素治疗7~10 d后阴道给药补充乳杆菌10 d。65名患者经宫腔镜检查及CD138免疫组织化学染色确诊CE，纳入宫腔镜诊断CE组（宫腔镜组）；接受左氧氟沙星联合甲硝唑治疗14 d。所有患者于治疗后下一个月经周期开始接受冻融胚胎移植（frozen embryo transfer，FET）。使用13项影响妊娠结局的协变量将2组FET周期进行1∶4倾向性评分匹配。主要研究终点为活产率，次要研究终点包括临床妊娠率、人绒毛膜促性腺激素（human chorionic gonadotropin，HCG）阳性率、流产率、妊娠并发症及新生儿不良结局发生率。进行多因素Logistic回归分析评估与活产相关的潜在风险因素。结果·宫腔镜组65名患者共接受127个FET周期，16S rRNA组27名患者共接受33个FET周期；宫腔镜组47名患者（74个FET周期）与16S rRNA组22名患者（26个FET周期）完成匹配。匹配后，16S rRNA组活产率（61.5% vs 23.0%，P=0.001）、临床妊娠率（65.4% vs 33.8%，P=0.011）、HCG阳性率（73.1% vs 46.7%，P=0.038）均显著高于宫腔镜组；2组流产率、异位妊娠率、妊娠并发症及新生儿不良结局发生率差异均无统计学意义（均P>0.05）。多因素Logistic回归分析结果显示，宫腔镜组与更低的活产率相关（OR=0.230，95%CI 0.087~0.607，P=0.003）。此外，女方高龄、体质量指数（body mass index，BMI）增加、仅移植1个胚胎以及移植卵裂胚均与更低的活产率相关（均P<0.05）。结论·与宫腔镜诊断联合广谱抗生素治疗相比，基于16S rRNA测序的菌群检测及针对性抗生素治疗联合补充乳杆菌可更有效改善RIF合并CE患者的妊娠结局，并且未发现对于母胎安全的不良影响。

Objective ·To compare pregnancy outcomes between broad-spectrum antibiotics administered after hysteroscopy and 16S rRNA gene sequencing combined with targeted antibiotics and Lactobacillus supplementation in patients with recurrent implantation failure (RIF) and chronic endometritis (CE). Methods ·This retrospective study analyzed 676 RIF patients who underwent either a 16S rRNA test during the mid-secretory phase or hysteroscopy during the proliferative phase at the Reproductive Medicine Center, Department of Obstetrics and Gynecology, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, from January 2023 to May 2025, after excluding known factors (preimplantation genetic testing for aneuploidies was not performed). Twenty-seven patients with 16S rRNA results indicating Lactobacillus abundance <90% and single-pathogen abundance≥10% were included in the 16S rRNA-diagnosed CE group (16S rRNA group). These patients received targeted antibiotics based on the pathogens for 7‒10 d, followed by vaginal Lactobacillus supplementation for 10 d. Sixty-five patients were diagnosed through hysteroscopic examination and CD138 immunohistochemical testing and were included in the hysteroscopy-diagnosed CE group (hysteroscopy group). They received levofloxacin combined with metronidazole for 14 d. All patients commenced frozen embryo transfer (FET) cycles in the subsequent menstrual cycle after treatment. Propensity score matching (1:4) was performed on FET cycles from the two groups using 13 covariates affecting pregnancy outcomes. The primary endpoint was the live birth rate. The secondary endpoints included the clinical pregnancy rate, human chorionic gonadotropin (HCG)-positive rate, miscarriage rate, and the incidences of pregnancy complications and neonatal adverse outcomes. Multivariate Logistic regression analysis was performed to assess potential risk factors affecting live birth. Results ·In the hysteroscopy group, 65 patients underwent 127 FET cycles, while in the 16S rRNA group, 27 patients underwent 33 FET cycles. Among them, 47 patients (74 FET cycles) from the hysteroscopy group and 22 patients (26 FET cycles) from the 16S rRNA group were successfully matched. After matching, the 16S rRNA group had significantly higher live birth rate (61.5% vs 23.0%, P=0.001), clinical pregnancy rate (65.4% vs 33.8%, P=0.011), and HCG-positive rate (73.1% vs 46.7%, P=0.038) than the hysteroscopy group. There were no significant differences in miscarriage rate, ectopic pregnancy rate, or the incidences of pregnancy complications and neonatal adverse outcomes between the two groups (all P>0.05). Multivariate Logistic regression analysis showed that the hysteroscopy group was associated with a lower live birth rate (OR=0.230, 95%CI 0.087‒0.607, P=0.003). Additionally, advanced maternal age, increased body mass index (BMI), transfer of only one embryo, and transfer of cleavage-stage embryos were all associated with a lower live birth rate (all P<0.05). Conclusion ·Compared with broad-spectrum antibiotics following hysteroscopy, 16S rRNA sequencing-based microbiota testing combined with targeted antibiotic therapy and Lactobacillus supplementation more effectively improves pregnancy outcomes in RIF patients with CE, and no adverse effects on maternal and fetal safety have been found.