BACKGROUND: The vaginal microbiome may influence the success of medically assisted reproduction (MAR), including IVF with or without ICSI, and IUI. However, findings across studies remain inconsistent, and previous meta-analyses have largely relied on binary microbiome classifications or pairwise comparisons. The relative associations between individual vaginal community state types (CSTs) and pregnancy outcomes therefore remain uncertain. OBJECTIVE AND RATIONALE: To compare the associations between individual CSTs and pregnancy following MAR by integrating direct and indirect evidence in a systematic review and network meta-analysis (NMA) and to determine whether specific CSTs were associated with lower odds of pregnancy. SEARCH METHODS: PubMed, Embase, Web of Science, and the Cochrane Library were searched from database inception to 2 April 2026. Eligible longitudinal studies assessed vaginal microbiome composition using sequencing-based profiling before embryo transfer or IUI and reported pregnancy following MAR. Cross-sectional studies, studies without CST classification, or pregnancy outcome data, and studies based solely on post-treatment sampling were excluded. Two reviewers independently performed study selection and risk-of-bias assessment. Odds ratios (ORs) with 95% CIs were synthesized across CSTs using NMA. Twelve longitudinal studies including 2289 women met the inclusion criteria. The prospectively registered review was reported in accordance with PRISMA and the PRISMA extension for NMA. OUTCOMES: The NMA allowed integration of direct and indirect evidence across six CST categories. Compared with Lactobacillus (L.) crispatus, Lactobacillus jensenii-dominated CSTs (OR 0.42, 95% CI 0.19-0.92), and low-lactobacillus community CSTs (OR 0.59, 95% CI 0.38-0.91) were associated with significantly lower odds of pregnancy. No statistically significant differences were observed for other lactobacilli, including Lactobacillus gasseri and Lactobacillus iners, compared with L. crispatus. Tests found no evidence of overall or loop inconsistency in the primary network. Subgroup analyses showed broadly similar patterns but were limited by sparse data, imprecision, and local inconsistency in some subgroup networks. Interpretation was further limited by clinical and methodological heterogeneity, observational study designs, residual confounding, and sparse evidence for several CST comparisons. Live-birth evidence was particularly limited: only five studies reported a total of 57 live births. Transitivity could not be formally assessed because potential effect modifiers were not consistently reported for individual CSTs. WIDER IMPLICATIONS: The findings indicate that general assumption of Lactobacillus-dominated CSTs being beneficial to MAR-treatment effects should be regarded with caution, as these may differ in their associations with pregnancy, depending on the specific Lactobacillus dominance. However, the available evidence is insufficient to support clinical implementation of CST profiling. Larger, standardized prospective longitudinal studies with CST-specific reporting, adjusted estimates, repeated microbiome sampling, and live-birth outcomes are required to determine whether CST profiling has clinically useful prognostic value and whether microbiome-targeted interventions improve reproductive outcomes. REGISTRATION NUMBER: PROSPERO CRD420251066092.