Towards unraveling antimicrobial resistance dynamics: A longitudinal exploration of rectal swab metagenomes

The increasing prevalence of antimicrobial resistance (AMR) poses significant challenges in clinical settings. In particular, early screening and detection of colonization by multidrug-resistant organisms (MDROs) in patients at admission is crucial. In this context, the clinical use of metagenomics (mNGS) holds promise for fast and untargeted diagnostic methods. Here, we aimed to evaluate the long-term stability of the rectal microbiome and the diagnostic accuracy of mNGS in comparison to culture and whole-genome sequencing (WGS) of MDROs. We analyzed rectal swabs from 26 patients with two consecutive admissions over a four-year period. The detected antimicrobial resistance genes and assembled metagenomes were compared to those obtained via classical culture-based antimicrobial susceptibility testing and WGS of isolated MDROs. Our results showed that the rectal microbiome is variable during the two timepoints, with a β-diversity greater in magnitude than what is currently known for the gut microbiome, highlighting the variability in the niche. Nevertheless, we also observed strong co-occurrence of taxa, suggesting that the rectal swab microbiome is also a regulated niche with cooperative biotic interactions. In total, we isolated and sequenced 6 MDROs from 6 patients at individual timepoints. Almost all AMR genes from the genomes of the isolates (median: 100%, range: 84.6-100%) could be detected by mNGS of the rectal swabs. Thus, in patients with positive cultures, we could not detect the isolated MDRO species or associated AMR genes at all screening visits. In addition, we detected AMR genes and pathogenic species in patients with negative cultures. In conclusion, our study showed that, in principle, mNGS of rectal swabs can detect clinically relevant AMR profiles. However, the cooccurrence of AMR genes and pathogenic species does not always correlate with culture-based diagnostic results but rather indicates a potential risk of horizontal AMR gene transfer. However, it is unclear whether the observed discrepancies are due to transient or locally confined colonization of MDROs, limits of detection, or variability of the sampling method and specimens.