Do households or hospitals have a larger contribution to carbapenemase-producing Enterobacteriaceae in municipal wastewater?

Wastewater-based surveillance is considered a valuable new surveillance tool for antibiotic resistant bacteria as it allows for easy screening of large numbers of individuals, in particular when measuring at municipal wastewater treatment plants (WWTPs). However, it has been questioned whether wastewater at municipal WWTPs primarily represents emissions of resistant bacteria from healthcare institutions such as hospitals. To gain insight in how well untreated wastewater (influent) at WWTPs reflects carriage of carbapenemase-producing Enterobacteriaceae (CPE) in the general population, this study investigated the relative contributions of hospitals and households to CPE in WWTP influents in the Netherlands. CPE concentrations were determined in wastewater of eight hospitals and influents of the associated municipal WWTPs using selective culture. For one WWTP, wastewater from five city districts, a nursing home and an asylum seekers' centre in its catchment was also analysed. Wastewater CPE loads (i.e. the total number of CPE per day) were calculated from CPE concentrations and daily water volumes, as was the proportional contribution of each individual wastewater source to the CPE load in WWTP influent. The relation between isolates from WWTPs and source wastewaters was confirmed using whole genome sequencing. Although CPE concentrations were higher in hospital wastewaters compared to municipal wastewater, hospitals generally explained less than 10% of CPE in WWTP influents. However, hospitals occasionally were a main contributor of specific CPE subtypes. Detection of CPE in wastewater from all five city districts confirmed the existence of multiple CPE sources in the general population. The minor contribution of hospitals to CPE loads in WWTP influents underscores the suitability of influents of municipal WWTP for use in wastewater-based surveillance. However, the presence of hospitals in a WWTP catchment area may occasionally skew the distribution of CPE genotypes, which, depending on the purpose of the surveillance, should be considered when designing a sampling scheme.