Antimicrobial resistance genes and modelling of treatment failure in bacterial vaginosis: clinical study of 289 symptomatic women

Clinical management of bacterial vaginosis (By) is difficult owing to inaccurate diagnostic tests, limited drug choices, and a high rate of recurrence. To our knowledge, there has not been a previous study of antimicrobial resistance (AMR) genes in community practice using next generation sequencing (NGS). A case control study (1 :1 age-matched with and without By) was undertaken in a series of 326 nongravid women of reproductive age with symptoms of BV to determine the prevalence of AMR genes. NGS was used to describe the complete vaginal microbiota and identify bacterial genes associated with resistance to: macrolides and/or lincosamides ermA, ermB, ermC, erM, ermTR and mefA; tetracyclines, beta-lactams, streptomycin, gentamicin and/or tobramycin acrA, acrB, mecA, tet, tetA, to/C and aac2; 5-nitroimadazoles - nim and nimB; and triazoles - cdr1 and mdr1. An evidence base was created to inform treatment decisions applicable to individual patients. AMR genes were identified in all drug classes: macrolides, 35.2 %; lincosamides, 35.6 %; tetracyclines, 21.8 %; aminoglycosides (streptomycin, gentamicin and tobramycin), 5.2 % each; 5-nitroimidazoles, 0.3 %; and triazoles, 18.7 %. There was more than a fourfold-higher frequency of AMR genes in pathogens from BV than from non-BV patients for macrolides (58.2 versus 12.3 %, respectively), lincosamides (58.9 versus 12.3 %) and tetracyclines (35.6 versus 8.0 %) (Fisher's exact test; all p<0.001). For each patient with BV, the spectrum of resistance genes was matched to the pathogens present. AMR genes were present in the majority of vaginal microbiomes of patients with symptoms of BV.