Wastewaters can be analyzed to generate population-level data for public health surveillance, such as antibiotic resistance monitoring. To provide representative data for the contributing population, bacterial isolates collected from wastewater should originate from different individuals and not be distorted by a selection pressure in the wastewater. Here we use Escherichia coli diversity as a proxy for representativeness when comparing grab and composite sampling at a major municipal wastewater treatment plant influent and an untreated hospital effluent in Gothenburg, Sweden. All municipal samples showed high E. coli diversity irrespective of the sampling method. In contrast, a marked increase in diversity was seen for composite compared to grab samples from the hospital effluent. Virtual resampling also showed the value of collecting fewer isolates on multiple occasions rather than many isolates from a single sample. Time-kill tests where individual E. coli strains were exposed to sterile-filtered hospital wastewater showed rapid killing of antibiotic-susceptible strains and significant selection of multi-resistant strains when incubated at 20 °C, an effect which could be avoided at 4 °C. In conclusion, depending on the wastewater collection site, both sampling method and collection/storage temperature could significantly impact the representativeness of the wastewater sample.

The spread of antibiotic resistance among bacterial pathogens is to a large extent mediated by mobile antibiotic resistance genes (ARGs). The prevalence and geographic distribution of several newly discovered ARGs, as well as some clinically important ARGs conferring resistance to last resort antibiotics, are largely unknown. Targeted analysis of wastewater samples could allow estimations of carriage in the population connected to the sewers as well as release to the environment. Here we quantified ARGs conferring resistance to linezolid (optrA and cfr(A)) and colistin (mcr-1, -2, -3, -4 and -5) and the recently discovered gar (aminoglycoside ARG) and sul4 (sulphonamide ARG) in raw hospital and municipal wastewater as well as treated municipal wastewater during five years in a low antibiotic resistance prevalence setting (Gothenburg, Sweden). Additionally, variations in bacterial composition of the wastewaters characterized by 16S rRNA sequencing were related to the variations of the ARGs in an attempt to reveal if the presence of known or suspected bacterial host taxa could explain the presence of the ARGs in wastewater. The mcr-1, mcr-3, mcr-4, mcr-5, sul4 and gar genes were detected regularly in all types of wastewater samples while optrA and cfr(A) were detected only in hospital wastewater. The most abundant genes were mcr-3 and mcr-5, especially in municipal wastewater. The detection of optrA was restricted to a peak during one year. Most of the ARGs correlated with taxa previously described as bacterial hosts and associated with humans. Although some of the tentative hosts may include bacteria also thriving in wastewater environments, detection of the ARGs in the wastewaters could reflect their presence in the gut flora of the contributing populations. If so, they could already today or in the near future hinder treatment of bacterial infections in a setting where they currently are rarely targeted/detected during clinical surveillance.

There is a risk that residues of antibiotics and other antimicrobials in hospital and municipal wastewaters could select for resistant bacteria. Still, direct experimental evidence for selection is lacking. Here, we investigated if effluent from a large Swedish hospital, as well as influent and effluent from the connected municipal wastewater treatment plant (WWTP) select for antibiotic resistant Escherichia coli in three controlled experimental setups. Exposure of sterile-filtered hospital effluent to a planktonic mix of 149 different E. coli wastewater isolates showed a strong selection of multi-resistant strains. Accordingly, exposure to a complex wastewater community selected for strains resistant to several antibiotic classes. Exposing individual strains with variable resistance patterns revealed a rapid bactericidal effect of hospital effluent on susceptible, but not multi-resistant E. coli. No selection was observed after exposure to WWTP effluent, while exposure to WWTP influent indicated a small selective effect for ceftazidime and cefadroxil resistant strains, and only in the E. coli mix assay. An analysis of commonly used antibiotics and non-antibiotic pharmaceuticals in combination with growth and resistance pattern of individual E. coli isolates suggested a possible contribution of ciprofloxacin and β-lactams to the selection by hospital effluent. However, more research is needed to clarify the contribution from different selective agents. While this study does not indicate selection by the studied WWTP effluent, there is some indications of selective effects by municipal influent on β-lactam-resistant strains. Such effects may be more pronounced in countries with higher antibiotic use than Sweden. Despite the limited antibiotic use in Sweden, the hospital effluent strongly and consistently selected for multi-resistance, indicating widespread risks. Hence, there is an urgent need for further evaluation of risks for resistance selection in hospital sewers, as well as for strategies to remove selective agents and resistant bacteria.

Horizontal gene transfer (HGT) plays an important role in the dissemination of antibiotic resistance genes. In sewer systems, human-associated and environmental bacteria are mixed together and exposed to many substances known to increase HGT, including various antibacterial compounds. In wastewaters, those substances are most often detected below concentrations known to induce HGT individually. Still, it is possible that such wastewaters induce HGT, for example via mixture effects. Here, a panel of antibiotics, biocides and other pharmaceuticals was measured in filter-sterilized municipal and hospital wastewater samples from Gothenburg, Sweden. The effects on HGT of the chemical mixtures in these samples were investigated by exposing a complex bacterial donor community together with a GFP-tagged E. coli recipient strain. Recipients that captured sulfonamide resistance-conferring mobile genetic elements (MGEs) from the bacterial community were enumerated and characterized by replicon typing, antibiotic susceptibility testing and long read sequencing. While exposure to municipal wastewater did not result in any detectable change in HGT rates, exposure to hospital wastewater was associated with an increase in the proportion of recipients that acquired sulfonamide resistance but also a drastic decrease in the total number of recipients. Although, concentrations were generally higher in hospital than municipal wastewater, none of the measured substances could individually explain the observed effects of hospital wastewater. The great majority of the MGEs captured were IncN plasmids, and resistance to several antibiotics was co-transferred in most cases. Taken together, the data show no evidence that chemicals present in the studied municipal wastewater induce HGT. Still, the increased relative abundance of transconjugants after exposure to hospital wastewater could have implications for the risks of both emergence and transmission of resistant bacteria.

Carbapenemase-producing Enterobacterales (CPE) constitute a significant threat to healthcare systems. Continuous surveillance is important for the management and early warning of these bacteria. Sewage monitoring has been suggested as a possible resource-efficient complement to traditional clinical surveillance. It should not least be suitable for rare forms of resistance since a single sewage sample contains bacteria from a large number of individuals. Here, the value of sewage monitoring in early warning of CPE was assessed at the Sahlgrenska University Hospital in Gothenburg, Sweden, a setting with low prevalence of CPE. Twenty composite hospital sewage samples were collected during a two-year period. Carbapenemase genes in the complex samples were analyzed by quantitative PCR and the CPE loads were assessed through cultures on CPE-selective agar followed by species determination as well as phenotypic and genotypic tests targeting carbapenemases of presumed CPE. The findings were related to CPE detected in hospitalized patients. A subset of CPE isolates from sewage and patients were subjected to whole genome sequencing. For three of the investigated carbapenemase genes, bla_NDM, bla_OXA-48-like and bla_KPC, there was concordance between gene levels and abundance of corresponding CPE in sewage. For the other two analyzed genes, bla_VIM and bla_IMP, there was no such concordance, most likely due to the presence of those genes in non-Enterobacterales populating the sewage samples. In line with the detection of OXA-48-like- and NDM-producing CPE in sewage, these were also the most commonly detected CPE in patients. NDM-producing CPE were detected on a single occasion in sewage and isolated strains were shown to match strains detected in a patient. A marked peak in CPE producing OXA-48-like enzymes was observed in sewage during a few months. When levels started to increase there were no known cases of such CPE at the hospital but soon after a few cases were detected in samples from patients. The OXA-48-like-producing CPE from sewage and patients represented different strains, but they carried similar bla_OXA-48-like-harbouring mobile genetic elements. In conclusion, sewage analyses show both promise and limitations as a complement to traditional clinical resistance surveillance for early warning of rare forms of resistance. Further evaluation and careful interpretation are needed to fully assess the value of such a sewage monitoring system.

Antibiotic resistance presents a serious and still growing threat to human health. Environmental exposure levels required to select for resistance are unknown for most antibiotics. Here, we evaluated different experimental approaches and ways to interpret effect measures, in order to identify what concentration of trimethoprim that are likely to select for resistance in aquatic environments. When grown in complex biofilms, selection for resistant E. coli increased at 100 µg/L, whereas there was only a non-significant trend with regards to changes in taxonomic composition within the tested range (0-100 µg/L). Planktonic co-culturing of 149 different E. coli strains isolated from sewage again confirmed selection at 100 µg/L. Finally, pairwise competition experiments were performed with engineered E. coli strains carrying different trimethoprim resistance genes (dfr) and their sensitive counterparts. While strains with introduced resistance genes grew slower than the sensitive ones at 0 and 10 µg/L, a significant reduction in cost was found already at 10 µg/L. Defining lowest effect concentrations by comparing proportion of resistant strains to sensitive ones at the same time point, rather than to their initial ratios, will reflect the advantage a resistance factor can bring, while ignoring exposure-independent fitness costs. As costs are likely to be highly dependent on the specific environmental and genetic contexts, the former approach might be more suitable as a basis for defining exposure limits with the intention to prevent selection for resistance. Based on the present and other studies, we propose that 1 µg/L would be a reasonably protective exposure limit for trimethoprim in aquatic environments.

IntroductionThe occurrence of antibiotic resistance in faecal bacteria in sewage is likely to reflect the current local clinical resistance situation.AimThis observational study investigated the relationship between Escherichia coli resistance rates in sewage and clinical samples representing the same human populations.MethodsE. coli were isolated from eight hospital (n = 721 isolates) and six municipal (n = 531 isolates) sewage samples, over 1 year in Gothenburg, Sweden. An inexpensive broth screening method was validated against disk diffusion and applied to determine resistance against 11 antibiotics in sewage isolates. Resistance data on E. coli isolated from clinical samples from corresponding local hospital and primary care patients were collected during the same year and compared with those of the sewage isolates by linear regression.ResultsE. coli resistance rates derived from hospital sewage and hospital patients strongly correlated (r2 = 0.95 for urine and 0.89 for blood samples), as did resistance rates in E. coli from municipal sewage and primary care urine samples (r2 = 0.82). Resistance rates in hospital sewage isolates were close to those in hospital clinical isolates while resistance rates in municipal sewage isolates were about half of those measured in primary care isolates. Resistance rates in municipal sewage isolates were more stable between sampling occasions than those from hospital sewage.ConclusionOur findings provide support for development of a low-cost, sewage-based surveillance system for antibiotic resistance in E. coli, which could complement current monitoring systems and provide clinically relevant antibiotic resistance data for countries and regions where surveillance is lacking.