Résumé
Resistance to antibiotics can lead to therapeutic failures of infectious diseases but also can slow down advances in intensive medicine or surgery. Among extremely resistant bacteria (XDR), carbapenemase-producing enterobacteria (CPE) are of particular concern because of their capacities to hydrolyze most of the β-lactam antibiotics including the last resort antibiotic carbapenems [1]. In France, blaOXA-48 and blaNDM are the most widespread carbapenemases. Another kind of XDR of concern are vancomycin-resistant Enterococci (VRE) associated with vanA gene [2]. Based on current knowledge, these 3 antimicrobial resistance genes (ARG) are emerging in France and spread through an epidemic mode. However, data on XDR carriage is patchy in general population because it is monitored only in high risk-hospitalized patients. New tools are needed to make surveillance of emerging resistances more comprehensive.During SARS-CoV-2 pandemic, efficiency of wastewater-based epidemiology (WBE) has been showed [3]. Because CPE and VRE are gastrointestinal bacteria, we propose WBE to improve the surveillance in human population. To achieve this goal, we developed a strategy to monitor blaOXA-48, blaNDM and vanA genes in wastewater from three hospital wards. Two wards were at high-risk for XDR carriage in patients. The last one was considered at low-risk and was used as a standard. blaCTX-M gene was also monitored as “endemic standard” because its endemic character both inside and outside the hospital. Genes’ quantifications were performed by qPCR and dPCR from hospital wastewater samples. CPE and VRE were cultured on selective media, enumerated and identified by MALDI-TOF-MS. Molecular signals were observed both for endemic (blaCTX-M) and emerging ARG (blaOXA-48, blaNDM and vanA) with space-time variations. It was supported by significant positive correlations between results obtained by qPCR and dPCR. For the low-risk ward, a stable baseline signal was observed, with quantities of genes significantly different from those of the 2 high-risk wards. This suggests the possibility to determine a basal state of ARG quantifications in hospital wastewater to set up alert thresholds. blaNDM and vanA had epidemic dynamics as expected. Furthermore, blaOXA-48 signal was either similar or superior to blaCTX-M signal in all the wards, suggesting an undervaluation of OXA-48 diffusion in France, or a possible amplification of the signal in the hydric bacterial community of wastewater pipes. Finally, there was no taxonomic overlap between resistant bacterial communities from patients and from wastewater. These results could be in favor of ARG exchanges between patients’ and waterborne bacteria.This study highlights the feasibility of AMR monitoring with WBE in a hospital. To validate this approach, further studies had to be performed considering ARG transfers and persistence.