Interplay between strain fitness and transmission frequency determines prevalence of antimicrobial resistance

Abstract

The steep rise of infections caused by bacteria that are resistant to antimicrobial agents threatens global health. However, the association between antimicrobial use and the prevalence of resistance is not straightforward. Therefore, it is necessary to quantify the importance of additional factors that affect this relationship. We theoretically explore how the prevalence of resistance is affected by the combination of three factors: antimicrobial use, bacterial transmission, and fitness cost of resistance. We present a model that combines within-host, between-hosts and between-populations dynamics, built upon the competitive Lotka-Volterra equations. We developed the model in a manner that allows future experimental validation of the findings with single isolates in the laboratory. Each host may carry two strains (susceptible and resistant) that represent the host’s commensal microbiome and are not the target of the antimicrobial treatment. The model simulates a population of hosts who are treated periodically with antibiotics and transmit bacteria to each other. We show that bacterial transmission results in strain co-existence. Transmission disseminates resistant bacteria in the population, increasing the levels of resistance. Counterintuitively, when the cost of resistance is low, high transmission frequencies reduce resistance prevalence. Transmission between host populations leads to more similar resistance levels, increasing the susceptibility of the population with higher antimicrobial use. Overall, our results indicate that the interplay between bacterial transmission and strain fitness affects the prevalence of resistance in a non-linear way. We then place our results within the context of ecological theory, particularly on temporal niche partitioning and metapopulation rescue, and we formulate testable experimental predictions for future research.

Transmission of antimicrobial resistance in the gut microbiome of gregarious cockroaches: the importance of interaction between antibiotic exposed and non-exposed populations

“…This is in agreement with previous ecological models showing that both the size and the frequency of a disturbance are important ( 81 ) but warrants further studies to show the relative importance and interaction in AMR ecology. This observation is, however, potentially important since it might suggest that if we reduce large transmission events then the normal microbiome might have sufficient resilience to absorb transmission, as recently suggested ( 82 ) and in line with our recent modeling approach ( 9 ). It was also apparent that when untreated microbiomes were frequently disturbed with small numbers of treated donors, their bacterial diversity was more variable.…”

“…It was also apparent that when untreated microbiomes were frequently disturbed with small numbers of treated donors, their bacterial diversity was more variable. The smaller magnitude of interactions at each event meant that only a few of the untreated recipients were impacted by the transfer; they were able to replenish their microbiome through interaction with the other untreated individuals during the intervals between transfers ( 9 , 83 , 84 ). In contrast, this was not possible in the single transfer event, where the majority of untreated recipients came into contact with the donors, resulting in the establishment of similar microbiomes between them.…”

“…Experiments on the transmission of AMR are lacking despite the need to validate the findings of several theoretical models ( 8 , 9 ). It is logistically difficult, time-consuming, and expensive to execute such experiments, especially in vivo .…”

“…While mathematical models have been valuable tools for predicting the potential trajectories and outcomes of AMR selection and dissemination, there remain major gaps in our understanding when it comes to actual, real-world transmission events ( 8 ). Experimental investigations into AMR transmission via microbiomes across hosts have been scant, not least because of the inherent challenges in setting up such studies ( 8 , 9 ). There is, however, a pressing need for tractable, in vivo models to test hypotheses and validate mathematical predictions.…”