Antibiotic resistance poses human health risks, and there are concerns about the effect of environmental antibiotic residues in the selection and spread of antibiotic resistance. The aim of this study was to identify antibiotic residue levels that are likely to select for resistance and relative contributions from different aquatic sources, of various aquatic environmental compartments of the WHO Western Pacific region (WPR) and the WHO South-East Asia region (SEAR), including in China and India.

A systematic review of empirical studies that measured antibiotic concentrations in aquatic environments, published between 2006 and 2019, and a probabilistic environmental hazard assessments approach, were used to identify antibiotic concentrations that are likely to select for resistance in various aquatic environmental compartments of the WPR and SEAR, including in China and India. The assessment involved the use of measured environmental concentrations and predicted no-effect concentrations (PNECs).

The systematic review found 218 relevant studies of 5230 screened from the WPR and 22 relevant studies of 2625 screened from the SEAR; some of these relevant studies were largely from China (n=168) and India (n=15). 92 antibiotics in the WPR and 45 in the SEAR were detected in various aquatic compartments. Antibiotic concentrations that most likely exceeded PNECs (0–100%) were observed in wastewater, and influents and effluents of wastewater treatment plants. Antibiotic concentrations that most likely exceeded PNECs were also observed in aquatic environmental compartments. The highest risk for the development of resistance was in tap or drinking water of the WPR and China for ciprofloxacin (62·5%). The relative contributions of potential sources of antibiotic contamination in waterways, such as hospitals, municipals, livestock, and pharmaceutical manufacturing, was determined for each antibiotic.

The concentrations of antibiotic residues found in wastewater and wastewater treatment plants of the WPR and SEAR make them potential hotspots for the development of antibiotic resistance, which creates human health risks from environmental exposure via drinking water. These findings can help decision makers to target risk reduction measures against environmental residues of priority antibiotics in high-risk sites, and help to focus research efforts in these world regions.

Antibiotic resistance has become a major threat to human health and the environment. However, the prevalence of environmental antibiotic residues and their potential role in the selection and spread of antibiotic resistance globally is uncertain. Prevalence and effects of antibiotic residues in waterways have been emphasised as an important area of research needed to define the risks to human health. These risks are more threatening in low-income and middle-income countries, including in the WHO Western Pacific region (WPR) and South-East Asia region (SEAR), where regulations and infrastructure to reduce antibiotic residues in waterways are often weak compared to high-income countries. The build-up of environmental antibiotic residues has been measured in several countries, but previous work has not considered how this could affect the spread of antibiotic resistance. Previous work has derived predicted no effect concentrations of various antibiotics but has not meaningfully applied these to real world situations.

Using data drawn from a systematic review of empirical studies from the WPR and SEAR, including China and India, we present a quantitative assessment of antibiotic concentrations in various aquatic environmental compartments (wastewater, wastewater treatment plants, river water, lake water, drinking water, groundwater, sea water, and other water compartments), identify the antibiotic residue levels that are likely to select for resistance to different antibiotic types, and estimate the relative contributions of potential sources of antibiotic contamination in waterways, such as hospitals, municipals, livestock, and pharmaceutical manufacturing.Our findings indicate that wastewater and wastewater treatment plants serve as potential hotspots for the development of antibiotic resistance in aquatic environments of the WPR and SEAR. Antibiotic residues also appear to pose a potential human health risk from environmental exposure via drinking water.

Our findings could be used by decision makers to undertake well directed actions towards monitoring and controlling antibiotic residues. Actions could be targeted to wherever there appears to be a high risk of development of resistance within the aquatic environments (ie, wastewater, wastewater treatment plants, and drinking water) of the WPR and SEAR. Our findings could assist policy makers in more efficiently allocating resources, which is especially important for resource-poor settings. Our findings might also help in prioritising future areas of research on antibiotics and antibiotic resistance in aquatic environments.

Eligibility criteria were specified a priori as original research articles of any quantitative design related to the subject matter. Studies were included if they were: original; reported antibiotics or antibacterials employed for systemic use in humans or animals; reported any type, group, or class of antibiotics; collected any type of water samples from the environment, such as river water, lake water, drinking water, groundwater, sea water, other water compartments (eg, surface water, canal, pond, stream, surrounding aquaculture, reservoir drainage, estuary, nearshore, and offshore), wastewater (eg, from municipals, hospitals, or pharmaceutical manufacturing sources), and influents and effluents of wastewater treatment plants; and measured antibiotic residue concentrations in one or several of the aquatic environments. Included studies had to have been done in the WPR, or the SEAR, and have been published from 2006 to 2019. Grey literature (eg, conference abstracts and dissertations) was included in the search. There were no language restrictions in the literature search. Studies were not included if they only reported antibiotic residues in non-water samples (eg, soil, sediment, manure, and vegetables).

The first screening stage (relevance screening) consisted of an evaluation of titles and abstracts of all records retrieved, to retain only those relevant to the review question. The full texts (inclusive of tables, figures, and supplementary data) of eligible papers were assessed according to the predefined inclusion and exclusion criteria in the second screening stage. The reference lists of identified studies were also searched. If there was a need for further clarification, the authors were contacted to clarify the issue by email.

Data on antibiotic concentrations, for reported measured antibiotics, were extracted from the papers and clustered by each aquatic environmental compartmental media or system to include the following items: author, year, journal, location, country, compartment, chemical, concentration in ng/L (minimum, maximum, mean, and median), removal efficiency for wastewater treatment plants, detection limit, instrumentation, sample size or sampling notes, notes, and full citation.

Title and abstract screening, full-text review, data extraction, and risk of bias assessment were carried out by a reviewer (NH) and random samples of all stages were independently reviewed by another reviewer (AJT). Disagreements were resolved through discussion. Where the two reviewers could not reach an agreement, a third reviewer arbitrated (CSL).

After the data were extracted and compiled from the primary literature, they were used for probabilistic environmental hazard assessments. The likelihood of the environmental occurrence of each antibiotic exceeding the thresholds for the development of antibiotic resistance, in various proportions of exposure in various aquatic environmental compartments of the WPR, the SEAR, China, and India, were estimated.

Probabilistic environmental hazard assessment models were formulated with the environmental exposure distributions of the maximum measured environmental concentrations of antibiotics for each water compartment. As maximum, mean, and median values were not consistently reported in the studies or literature, and to conservatively maximise estimates of exposure, maximum measured environmental concentrations of antibiotics were used for hazard assessments of aquatic environmental compartments. Based on available data (the reported maximum measured environmental concentrations of antibiotics), at least five maximum measured environmental concentrations of an antibiotic for each compartment in the WPR and in China, and at least three maximum measured environmental concentrations of an antibiotic for each compartment of the SEAR and India were used.

The probabilistic environmental hazard assessment does not generate a single point estimate, but rather produces a likelihood and range that a particular exposure and effect will occur. Accordingly, this type of assessment allows the risk assessors or managers (eg, decision makers, regulatory agencies, or industries) to do the assessment, independent of most value judgements, to predict the likelihood that a certain level of protection would be attained. For instance, the assessors might require that the environmental concentration associated with the level of protection be exceeded only 5% of the time (alternatively, 95% of all exposures [water concentrations] would be expected to be equal to or less than the required environmental concentration). Using these centile levels depending on the resources available to the risk managers and the capacity and applicability, appropriate action is taken. The assessors then do not work in a void but have some understanding of what resistance reduction their actions are likely to have.

The probabilistic environmental hazard assessment also allows the analysis of variability (refers to the heterogeneity and diversity) and uncertainty (refers to imperfect knowledge or a lack of a precise knowledge) to be incorporated into exposure or hazard assessments.

The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

The systematic review of studies reported between Jan 2, 2006, and Dec 2, 2019, measuring antibiotic residue concentrations in the aquatic environments, included 218 relevant studies of 5230 screened from the WPR and 22 relevant studies of 2625 screened from the SEAR; some of these relevant studies have largely been from China (n=168) and India (n=15; , ). In the WPR, 92 antibiotics were detected, and in the SEAR, 45 antibiotics were detected. The overall risk of bias in the included studies was low. The most updated and precise analytical procedure, liquid chromatography coupled with tandem mass spectrometry, was used in most published studies. The number of publications reporting the occurrence of antibiotic residues in the aquatic environments of the WPR, the SEAR, China, and India over time and by country, and on the detection frequency and distribution of antibiotics in the various aquatic environmental compartments of the WPR and SEAR, is in . We found that human and veterinary antibiotic residues were detected in diverse aqueous environmental compartments of the WPR and SEAR. Antibiotics from various classes, such as fluoroquinolones, macrolides, tetracyclines, β-lactams, lincosamides, sulfonamides, amphenicols, glycopeptides, and aminoglycosides, were reported in wastewater (eg, from municipals, hospitals, and pharmaceutical manufacturing sources), influents and effluents of wastewater treatment plants, and receiving aquatic environments of the WPR and SEAR.

We made a quantitative assessment of the environmental and human health risks (eg, percentage exceedance of PNECs and proportions of exposure) due to all the reported antibiotic residues in the environmental aquatic compartments of the WPR, the SEAR, China, and India. Antibiotic residues were ubiquitous, and residual concentrations of some antibiotics exceeded the thresholds for the development of resistance in various proportions of exposure in various aquatic environments, including surface water, groundwater, wastewater, and wastewater treatment plants. Wastewater and wastewater treatment plants appeared to serve as the main sources for the development of antibiotic resistance in these regions. The relative effect of various contributions of antibiotic contamination in the waterways, such as hospitals, municipals, livestock, and pharmaceutical manufacturing sources, was determined for each antibiotic. Antibiotic residues in these waters appear to pose a potential environmental and human health risk of promoting the emergence and development of antibiotic resistance. The assessments in this Article can aid in developing priority actions to prevent and control antibiotic resistance in the environment. As such, this has the potential to assist policy makers in efficiently allocating resources, which is especially important for resource-poor settings in the WPR and SEAR.

Overall, results from the probabilistic environmental hazard assessment indicated that some antibiotics need to be considered priority antibiotics for preventive and control measures and for further environmental and human health research. Residues of these antibiotics were common in most aqueous compartments of the environment (eg, wastewater, influents and effluents of wastewater treatment plants, and receiving aquatic environments) of the WPR and SEAR and therefore pose high risk of environmental and human health risk from the development of antibiotic resistance. These antibiotics include fluoroquinolones (ciprofloxacin, norfloxacin, ofloxacin, and enrofloxacin), macrolides (azithromycin, clarithromycin, erythromycin, and roxithromycin), and tetracyclines (tetracycline and oxytetracycline) in the aquatic environments of the WPR and China, and fluoroquinolones (ciprofloxacin, ofloxacin, and levofloxacin) in the aquatic environments of the SEAR and India, as per the current evidence. This list might grow further with time and generation of further evidence.

Our findings can be used by decision makers to undertake well directed actions towards monitoring and controlling antibiotic residues, such as adequate waste management and wastewater treatment measures, and One Health integrated surveillance. Antibiotics can be targeted to wherever there appears to be a high risk of the development of resistance within the aquatic environments of the WPR and SEAR. The data here could inform risk managers and assessors in prioritising the antibiotic pollutants and environments that might require intervention, and therefore assist policy makers in efficiently allocating resources, which is especially important for resource-poor settings. Furthermore, the results can help in prioritising future areas of research on antibiotics and antibiotic resistance in aquatic environments.

There were some important limitations to our study. The outcomes might be biased to some degree due to non-availability of data from some of the countries in the region. There was a scarcity of data on environmental occurrence of some antibiotics or compartments. The geographical representativeness of the available scientific literature and of our literature search might be affected by publication bias and language bias. There is a lack of data on PNECs for the development of resistance for some important antibiotics, such as chlortetracycline and most sulfonamides (although the European Committee for Antimicrobial Susceptibility Testing database has data on minimum inhibitory concentrations on most clinically relevant species, most environmental bacterial species cannot be cultivated and thus have unknown minimum inhibitory concentrations) and PNECs rely on extrapolation of in-vitro bacterial susceptibility data to field conditions with complex microbial communities under different antibiotic exposures. Variability or heterogeneity should also be considered—eg, sites, seasons, rural and urban areas, upstream and downstream conditions, water quality, characteristics of the aquatic compartment, sampling and analytical methods or errors, and legislation for the use and discharge of antibiotics. Studies on minimal selective concentrations in complex microbial communities are scarce and further experimental validation is necessary to evaluate how well the PNECs estimate the potential for antibiotic resistance development and selection. Uncertainty and variability have the potential to result in overestimates or underestimates of the risk. For example, many experimentally derived PNECs are greater than these modelled PNECs, so risks might be overestimated. Although the PNECs used are generally considered most protective in terms of resistance, PNECs are still sometimes higher than PNECs derived using eco-toxicological assays; therefore, environmental risk might be underestimated in some cases. Furthermore, we choose to use the maximum measured concentrations of antibiotic residues, which could bias our results toward overestimating the risk of exposure and antimicrobial resistance development. Studies that we did not have full access to were not included, introducing a degree of selection bias to the review. However, the use of a large number of included studies (218 from the WPR and 22 from the SEAR) in the analysis might mitigate this problem.

In conclusion, antibiotics were present in various aquatic compartments of the environment (wastewater, influents and effluents of wastewater treatment plants, river water, lake water, drinking water, groundwater, sea water, and other water compartments) of the WPR and SEAR, including China and India. Results from the probabilistic environmental hazard assessment indicated that residual concentrations of particular antibiotics in many aquatic environmental compartments exceeded the thresholds for development of resistance. The highest risks for development of antibiotic resistance were observed in wastewater and wastewater treatment plants, which might serve as hot spots for the development of resistance in aquatic environments of the WPR and SEAR. Antibiotic residues appear to pose a potential human health risk from environmental exposure via drinking water. This study presents the likelihood of the environmental occurrence of each antibiotic exceeding thresholds for development of antibiotic resistance, in various proportions of exposure in various aquatic environmental compartments of the WPR and SEAR, and the relative effect of various contributions of antibiotic contamination in waterways, such as hospitals, municipals, livestock, and pharmaceutical manufacturing sources. This can aid in developing effective preventive and control measures to combat antibiotic resistance, and to protect environmental and human health.

NH, AJT, and CSL conceived and designed the study. NH developed the methods, and extracted and analysed the data. Random samples of all stages were reviewed by another reviewer (AJT). NH drafted the first version of the manuscript and revised it based on suggestions from AJT and CSL. All authors have read and approved the final version of the manuscript. All authors verified the underlying data and had full access to all the data in the study. All authors had final responsibility for the decision to submit for publication.

Analyses of all antibiotics in the various aquatic environmental compartments of the WPR, the SEAR, China, and India are publicly available online at .

We thank Karolinska Institutet University Library for the assistance in databases selection and search terms, and Ouyang Wenwei for his contribution in advising biostatistics.