Geographic Variation in Diagnostic Performance of Amsel Criteria and Nugent Score for Vaginal Dysbiosis Defined by 16S rRNA Gene Sequencing
Zhu, M.; Mtshali, A.; Mzobe, G.; Magini, N.; Mitchev, N.; Khan, A.; Cooley, B.; Murthy, M.; Lewis, L.; Xu, J.; Shih, J. B.; Elsherbini, J.; Kama, A.; Mafunda, N.; Chetty, C.; Vermeren, L.; Passmore, J.-A.; Happel, A.-U.; Kwon, D. S.; Symul, L.; Mitchell, C. M.
Show abstract
Vaginal dysbiosis (VD), characterized by low abundance of vaginal lactobacilli and increased bacterial community diversity, is implicated in multiple adverse reproductive outcomes and is an emerging target for preventive interventions, including live biotherapeutic products (LBPs). The most common clinical presentation of VD is bacterial vaginosis (BV), but at least half of people are asymptomatic. We investigated how two commonly used diagnostic criteria for BV, namely Amsel and Nugent, align with 16s rRNA gene sequencing-defined community state types (CSTs) demonstrating VD. We analyzed screening specimens from a Phase 1b randomized trial of LBP conducted at two sites (CAPRISA, South Africa; MGH, USA), as well as a single follow-up visit from enrolled participants. Using sequencing-based CST as the reference and multinomial mixed-effects logistic models, we evaluated the association of Amsel BV and Nugent BV with CST IV (including subtypes IV-A and IV-B) and tested for site-specific effects. Amsel BV was significantly associated with CST IV-A, and IV-B; however, the strength of association was significantly diminished at CAPRISA compared to MGH, pointing to site-specific assessment differences or underlying biological variation. Nugent BV yielded stronger associations with CST IV-A, and IV-B and showed no evidence of a site-specific interaction, indicating consistent performance across sites. These findings indicate that diagnostic performance for VD varies by framework: Amsel criteria are susceptible to geographical site effects, whereas Nugent score demonstrates stronger and more site-agnostic associations. For clinical studies targeting VD, Nugent scoring and/or sequencing-based approaches should be prioritized for VD endpoint definition and stratification.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Evaluation of microbiome enrichment and host DNA depletion in human vaginal samples using Oxford Nanopore's adaptive sequencing 93%
- Antibiotics and the developing intestinal microbiome, metabolome and inflammatory environment: a randomized trial of preterm infants 92%
- New insights into the cultivability of human milk microbiota from ingestion to digestion and implications for its immunomodulatory properties 92%
Similar papers in this journal
- Evaluation of DNA extraction protocols from liquid-based cytology specimens for studying cervical microbiota 93%
- Variant calling for cpn60 barcode sequence-based microbiome profiling 92%
- Acceptability and efficacy of vaginal self-sampling for genital infection and bacterial vaginosis: A large, cross-sectional, non-inferiority trial 92%
Similar papers in this journal
- The vaginal microbiota of pregnant women varies with gestational age, maternal age, and parity 95%
- Microbiota of the pregnant mouse: characterization of the bacterial communities in the oral cavity, lung, intestine, and vagina through culture and DNA sequencing 93%
- Taxonomic and functional shifts in the perinatal gut microbiome of rhesus macaques 93%
Similar papers in this journal
- Temporal Dynamics of the Vaginal Microbiome and Host Immune Markers Before, During, and After Metronidazole Treatment for Bacterial Vaginosis 97%
- Longitudinal profiling of the macaque vaginal microbiome reveals similarities to diverse human vaginal communities: implications for use as a pre-clinical model for bacterial vaginosis 95%
- Lactobacillus iners dominates the vaginal microbiota of healthy Italian women of reproductive age. 94%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.