Drug-Resistant Bacteria Slip Through South African Wastewater Plants Into Rivers

Antibiotic resistance is often described as a silent pandemic, and one of its quietest highways may be the treated water flowing out of wastewater treatment plants. A new study from Gauteng, South Africa, has now provided one of the most detailed snapshots to date of how multidrug-resistant bacteria move through two municipal treatment works and end up in the rivers that receive their effluent. The findings, published in Frontiers in Water, suggest that conventional activated sludge treatment — the workhorse technology used across South Africa and much of the world — cannot be relied upon to stop resistant microbes at the plant boundary.

The research team, led by Nkgameleng K. Mokhonoana and Ilunga Kamika of the University of South Africa’s Institute for Nanotechnology and Water Sustainability, focused on two wastewater treatment plants in Johannesburg serving dense urban and hospital catchments. WWTP1 treats domestic sewage through an activated sludge process followed by chlorination, with a design capacity of 150,000 kilolitres per day. WWTP2 is considerably larger, at 405,000 kilolitres per day, and adds maturation ponds before final chlorination. That seemingly small engineering difference would prove decisive in the results.

Between June and late October 2023, the researchers collected 72 water samples in triplicate from four locations at each plant: the raw influent, the treated effluent, and river sites upstream and downstream of the discharge points. Alongside the microbiology, they measured core physicochemical parameters in situ — pH, temperature, total dissolved solids and dissolved oxygen — using a calibrated multiparameter probe. On paper, the water chemistry looked reassuring. Effluent pH ranged from 7.25 to 7.51 and temperatures from 16.73 to 22.33 degrees Celsius, both within South African regulatory discharge limits, while total dissolved solids stayed well below the 450 ppm cap. Statistical analysis found no significant differences in these parameters between the plants or across sampling points.

The biology told a far more troubling story. Using membrane filtration and selective culture media, the team enumerated presumptive Enterococcus and Staphylococcus species, two genera of Gram-positive bacteria that include clinically important pathogens such as Enterococcus faecium and Staphylococcus aureus. Concentrations reached between 6.07 and 7.42 log CFU/mL for Enterococcus and 6.58 to 7.26 log CFU/mL for Staphylococcus. The removal performance of the two plants diverged dramatically. WWTP2 achieved reductions of 89.5 to 90.5 percent, whereas WWTP1 managed only between 0.68 and 10.76 percent — meaning that in some campaigns its effluent carried bacterial loads barely lower than its incoming raw sewage. Downstream river concentrations generally exceeded those upstream.

Why such a stark difference between two plants using broadly similar processes? The researchers point to the maturation ponds at WWTP2, which provide an additional stage of biological and physical treatment that appears to substantially reduce bacterial loads before discharge. WWTP1’s persistently high effluent counts, exceeding 1,000 CFU/mL across all seasons, hint at disinfection inefficiencies, though the team notes that chlorine residuals were not measured to confirm this directly. The context matters too: South Africa’s 2023 Green Drop Report classified 334 of the 855 audited wastewater treatment works — 39 percent — as being in a critical state requiring urgent intervention, amid ageing infrastructure, inadequate maintenance and vandalism.

To probe the resistance threat, the team grew 54 presumptive E. faecium and S. aureus isolates and tested them against nine antibiotics using the Kirby–Bauer disk diffusion method, following Clinical and Laboratory Standards Institute guidelines. An isolate was classed as multidrug-resistant if it was non-susceptible to at least one agent in three or more antimicrobial classes. Six isolates met that criterion. Enterococcus isolates showed the strongest resistance to tetracycline across all sampling periods, with resistance also observed to azithromycin and teicoplanin — the latter a last-resort glycopeptide antibiotic. Staphylococcus isolates were most often resistant to tetracycline, clarithromycin and azithromycin, though they remained fully susceptible to ampicillin and vancomycin. Multiple antibiotic resistance indices ranged from 0.13 to 0.71, with the highest values found in Enterococcus isolates — strikingly, including one recovered from the upstream river at WWTP1, suggesting pre-existing environmental contamination independent of the plant’s discharge.

The most technically revealing part of the study came from whole-genome sequencing. Six multidrug-resistant isolates — five E. faecium and one S. aureus — were sequenced on an Illumina MiSeq, with reads assembled and screened against the ResFinder and Comprehensive Antibiotic Resistance Database. Multilocus sequence typing assigned the E. faecium isolates to four sequence types: ST1052, ST1058, ST94 and ST32, while the S. aureus isolate belonged to ST243. ST32 and ST94 have previously been reported in livestock in Ghana and linked to resistance against erythromycin, ciprofloxacin and tetracycline. ST243, meanwhile, has been identified in dairy cows in South Africa’s Free State province associated with macrolide and tetracycline resistance — a signal that animal and environmental reservoirs may be connected. ST1052 and ST1058 appear to be rare or underreported lineages in South Africa.

The gene inventory was sobering. The E. faecium isolates carried tetracycline resistance genes tetL and tetM — both commonly associated with mobile genetic elements that can hop between bacteria — along with msr(C), erm(B) and lsa(A), which together confer resistance to macrolides, lincosamides, streptogramins and pleuromutilins. The S. aureus genome harboured norA and norR, genes involved in fluoroquinolone resistance, alongside additional determinants flagged by the CARD database, including FosB for fosfomycin resistance. Virulence profiling added a further layer of concern: the S. aureus isolate carried 55 virulence genes, including toxin genes such as lukF-PV and lukS-PV, which encode cytotoxins associated with leukocyte destruction and tissue necrosis, and hla and hld, implicated in skin and soft tissue infections. A single E. faecium isolate carried 24 virulence genes, including cylA, previously associated with human infections, plus genes for adhesion, biofilm formation and capsule production.

The authors are careful to caveat their conclusions. Resistant bacteria were detected upstream of the plants, so wastewater discharge cannot be confirmed as the sole source of riverine contamination — illegal dumping and sewage leaks in the catchment likely contribute too. Only six isolates were sequenced due to cost constraints, and the study did not determine whether the resistance genes sit on chromosomes or mobile elements, leaving their horizontal transfer potential unresolved. The researchers also note discrepancies between phenotypic resistance and detected genes, which can arise from differences in gene expression, database limitations or the stringent screening thresholds used. Detection of a gene, they emphasise, does not prove it is being expressed.

Even with those limitations, the overall picture is hard to dismiss: clinically relevant multidrug-resistant bacteria, carrying a battery of resistance and virulence genes, are passing through at least one major South African treatment plant and reaching the rivers downstream. The findings underscore what global surveillance bodies, including the World Health Organization’s GLASS programme — which estimates that one in six bacterial infections worldwide now shows resistance to commonly used antibiotics — have long warned: the environment is a critical reservoir in the antimicrobial resistance crisis. The authors call for routine monitoring of effluents and receiving waters, longitudinal sampling with metagenomic approaches, and source-tracking to disentangle plant contributions from background contamination. As conventional activated sludge systems continue to discharge into the same rivers that supply communities downstream, the study makes a clear case that treatment technology designed decades ago to remove organic matter must now be rethought with resistance in mind.

Subject of Research: Antibiotic-resistant bacteria and resistance genes in South African wastewater treatment plants and receiving rivers

Article Title: Antibiotic resistance in South African wastewater: a snapshot from two treatment plants and their receiving rivers

Article References: Mokhonoana, N. K., Msagati, T. A. M., Malefetse, T. J., Chokoe, K., & Kamika, I. (2026). Antibiotic resistance in South African wastewater: a snapshot from two treatment plants and their receiving rivers. Frontiers in Water, 8, Article 1955602. https://doi.org/10.3389/frwa.2026.1955602

Image Credits: AI Generated

DOI: 10.3389/frwa.2026.1955602

Keywords: antibiotic resistance, wastewater treatment, Enterococcus faecium, Staphylococcus aureus, whole-genome sequencing, South Africa, antimicrobial resistance genes, rivers, activated sludge, multidrug resistance, public health, water quality