Quantifying the impact of genetic determinants of antibiotic resistance on bacterial lineage dynamics

The dynamics of antimicrobial resistance in bacteria are informed by the fitness advantages conferred by genetic determinants of resistance in the presence of antibiotic pressure and the potential fitness costs in its absence. However, frameworks for quantitative estimates of real-world fitness impact have been lacking, given multidrug resistance, multiple pathways to resistance, and uncertainty around drug exposures. Here, we addressed these challenges through analysis of genome sequences from clinical isolates of Neisseria gonorrhoeae collected over 20 years from across the United States, together with national data on antibiotic treatment. Using a hierarchical Bayesian phylodynamic model, we quantified the contributions of resistance determinants to strain success. Resistance mutations had a fitness benefit when the cognate antibiotic was in use but did not always incur a fitness cost otherwise. Two fluoroquinolone-resistance conferring mutations at the same site in gyrA had divergent fitness impact after fluoroquinolones were no longer used for treatment, findings supported by in vitro competition experiments. Fitness costs were alleviated by loss of costly resistance determinants and counterbalanced by gain of new fitness-conferring resistance determinants. Quantifying the extent to which the resistance determinants explained each lineage’s dynamics highlighted gaps and pointed to opportunities for investigation into other genetic and environmental drivers. This work thus establishes a method for linking pathogen genomics and antibiotic use patterns to quantify the fitness impact of resistance determinants and the factors shaping ecological trends.