Introduction Infertility affects 15–20% of couples, with 3% of Hungarian children conceived through in vitro fertilization (IVF). Predicting IVF outcomes before treatment initiation remains challenging. Emerging evidence suggests that the vaginal and seminal microbiota influence reproductive health by modulating local immunity, implantation, and gamete function. This study aimed to evaluate associations between baseline clinical, biochemical, and microbiological parameters and IVF outcomes. Material and methods We conducted a retrospective, single-center cohort study including 475 couples undergoing IVF with or without intracytoplasmic sperm injection at the University of Szeged (January 2022–December 2023). Data collection encompassed maternal demographics, reproductive history, baseline hormone levels, ovarian stimulation characteristics, and endometrial thickness. The results of microbiological cultures of vaginal discharge and semen samples, including Lactobacillus colonization, pathogen distribution, and the antibiotic resistance status of the pathogens, were recorded. The primary outcome was clinical pregnancy. Machine learning models, including support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGBoost) were applied, to explore the predictive value of combined clinical and microbial features for IVF outcome. Results Positive vaginal cultures were identified in 121 women (25.5%), most commonly Candida albicans (26%), Streptococcus agalactiae (17%), and Escherichia coli (17%) among the positive cases. Among 134 men (29%) with positive semen cultures, Enterococcus faecalis (46%) and Escherichia coli (23%) predominated. Single-organism growth predominated, and pathogen overlap between partners was rare (13 couples, 3%). Vaginal Lactobacillus presence negatively correlated with several pathogens, including Candida albicans , Streptococcus agalactiae , Gardnerella vaginalis , and Enterococcus faecalis . Clinical pregnancy rates were similar between women with positive and negative vaginal cultures (36% vs. 39%, n.s.). Machine learning analyses showed that maternal age remained the dominant predictor, while microbial data contributed modestly, with Lactobacillus consistently emerging as the most relevant microbial feature. Conclusions Vaginal and seminal microbial alterations are common among couples undergoing IVF, yet true pathogen overlap between partners was minimal. Lactobacillus colonization demonstrated a clear protective association, supporting its potential as a biomarker for reproductive success. Although microbial features alone are insufficient to predict IVF outcomes, integrating microbial profiles with established clinical parameters could inform personalized fertility management.
Vágvölgyi et al. (Wed,) studied this question.