The rates and routes of transmission of multidrug resistant Klebsiella clones and genes into the clinic from environmental sources.

The Klebsiella group is highly diverse both genetically and ecologically, being commonly recovered from humans, livestock, plants, soil, water, and wild animals. Many species are opportunistic pathogens, and can harbour diverse classes of antimicrobial resistance (AMR) genes. K. pneumoniae is responsible for a high public-health burden, due in part to the rapid spread of health-care associated clones that are non-susceptible to carbapenems. Klebsiella thus represents a highly pertinent taxon for assessing the risk to public health posed by animal and environmental reservoirs. Here we report an analysis of 6548 samples and 3,482 genome sequences representing 15 Klebsiella species sampled over a 15-month period from a wide range of clinical, community, animal and environmental settings in and around the city of Pavia, in the northern Italian region of Lombardy. Despite carbapenem-resistant clones circulating at a high frequency in the hospitals, we find no genotypic or phenotypic evidence for non-susceptibility to carbapenems outside of the clinical environment. The non-random distribution of species and strains across sources point to ecological barriers that are likely to limit AMR transmission. Although we find evidence for occasional transmission between settings, hierarchical modelling and intervention analysis suggests that direct transmission from the multiple non-human (animal and environmental) sources included in our sample accounts for less than 1% of hospital disease, with the vast majority of clinical cases originating from other humans.

Due to the Covid pandemic data production in the project got delayed by 18 months and consequently all the planned main publications are correspondingly delayed. Data production was finalized on August 6, 2021, but the analyses for future publications will continue into 2022. The main scientific findings of the project so far are reported in the preprint https://www.biorxiv.org/content/10.1101/2021.08.05.455249v1 We found evidence for strong ecological barriers delimiting transmission of Klebsiella clones between animals, environment and human health care, which has implications for informing future One Health policy for some of the most urgent AMR priority pathogens classified by the WHO.

Klebsiella pneumoniae (Kp) is a leading cause of multidrug resistant hospital-acquired infections globally, and is responsible for an increasing public health burden in the community. Whilst there is an urgent need to mitigate the emergence and spread of Kp, this will only be possible with a detailed understanding of the forces underpinning the genetic and ecological diversity of this ubiquitous organism. Specifically, a broad ecological context is required to identify the sources of emergent community and health-care associated infection from the many and varied niches encompassing the environment, which will in turn inform on targeted surveillance and intervention policies. In order to address this, our primary methodology will be first to sample from multiple clinical, community, agricultural and environmental settings within a well-defined geographic locale, in and around a single town, Pavia, in Northern Italy. We will combine these samples with epidemiologically matched samples from elsewhere in Europe. We will then utilise whole genome sequencing (WGS) to characterise entire Kp communities from each setting. Pathogen WGS is being increasingly used for outbreak investigation, but these studies are based on single colony sequencing followed by phylogenetic analysis of the core genome. Excepting cases where the transmission chains are very short (eg from a cow to a farmer), the dynamics of transmission within the environment, and between the environment and the clinic, are far more intractable than within a single health care setting. Complex transmission chains, incomplete sampling, multiple co-colonising and co-transmitting strains, rapid movement of MGEs, and the extreme diversity of Kp will all limit the utility even very large datasets based on individual genome sequences for inferring broad transmission dynamics at this scale.