Genotypic determinants of fluoroquinolone and macrolide resistance in Neisseria gonorrhoeae.

Background High rates of antimicrobial resistance (AMR) in Neisseria gonorrhoeae hinder effective treatment, but molecular AMR diagnostics may help address the challenge. This study aimed to appraise the literature for resistance-associated genotypic markers linked to fluoroquinolones and macrolides, to identify and review their use in diagnostics. Medline and EMBASE databases were searched and data pooled to evaluate associations between genotype and phenotypic resistance. The minimum inhibitory concentration (MIC) cut-offs were ≤ 0.06 mg L-1 for non-resistance to ciprofloxacin and ≤ 0.5 mg L-1 for non-resistance to azithromycin. Diagnostic accuracy estimates were limited by data availability and reporting. It was found that: 1) S91 and D95 mutations in the GyrA protein independently predicted ciprofloxacin resistance and, used together, gave 98.6% (95% confidence interval (CI) 98.0-99.0%) sensitivity and 91.4% (95%CI 88.6-93.7%) specificity; 2) the number of 23S rRNA gene alleles with C2611T or A2059G mutations was highly correlated with azithromycin resistance, with mutation in any allele giving a sensitivity and specificity of 66.1% (95%CI 62.1-70.0%) and 98.9% (95%CI 97.5-99.5%) respectively. Estimated negative (NPV) and positive predictive values (PPV) for a 23S rRNA diagnostic were 98.6% (95%CI 96.8-99.4%) and 71.5% (95%CI 68.0-74.8%) respectively; 3) mutation at amino acid position G45 in the MtrR protein independently predicted azithromycin resistance; however, when combined with 23S rRNA, did not improve the PPV or NPV. Viable candidates for markers of resistance detection for incorporation into diagnostics were demonstrated. Such tests may enhance antibiotic stewardship and treatment options.