Drug-resistant bacteria found in the air of a Johannesburg clinic, UJ study shows

Every Gram-negative bacterium that University of Johannesburg researchers grew from the air inside a Johannesburg clinic was resistant to several classes of antibiotics, and four in five resisted colistin, a last-resort drug kept for infections when little else works.

The study, by Lee Heine, Robyn Theresa Dodds and Atheesha Singh of UJ's Water and Health Research Centre, was published online in the journal Indoor Air in August. The clinic is not named; press reports describe it as being in a crowded, high-rise part of central Johannesburg.

The authors say only a few studies worldwide have looked at clinic or hospital air as a source of drug-resistant germs, and none of them in South Africa.

How the air was tested

In February and March 2025 the team sampled the air in 40 rooms and areas of the clinic between 06:00 and 08:30, before the day's patients arrived. They covered treatment rooms, bathrooms, entrances, passages, reception and common areas, a clinician's room and a file room. Each sample drew a cubic metre of air onto growth plates. Three outdoor spots were sampled for comparison.

Bacteria that grew were identified and tested against a panel of antibiotics. Moulds were identified by DNA sequencing.

The early hour matters for what the sample is. It is air in the building before the day's patients arrived, not air taken during a busy clinic. Forty rooms and areas, one cubic metre each, indoors, plus three outdoor spots. The outdoor samples are the comparison. They are not a second clinic.

What grew

All 77 Gram-negative isolates tested were multidrug-resistant, and three (3.90%) were extensively drug-resistant. Of 54 Gram-positive isolates, 52 (96.30%) were multidrug-resistant and three (5.56%) extensively drug-resistant.

The bacteria included members of the group doctors call ESKAPEE, which cause many hard-to-treat hospital infections: Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter species and E. coli.

Among the Gram-negative bacteria, 85.71% resisted cefuroxime and 67.53% ceftazidime, both commonly used first-line antibiotics. Resistance to the carbapenems, a reserve class, was lower but present: 31.17% for ertapenem and 25.97% for meropenem. And 80.52% resisted colistin.

Among the staphylococci and similar bacteria, 71.79% were resistant to oxacillin, the marker for methicillin resistance. Only 5.13% of that group resisted vancomycin, the usual treatment for MRSA. Of the 15 enterococci and related bacteria tested, seven were vancomycin-resistant.

The 80.52% is the colistin figure, on the Gram-negative isolates. It is not a claim that four in five of every germ in the building resists colistin. It is four in five of the 77 Gram-negative isolates the team grew and tested. Three of those 77, 3.90%, were extensively drug-resistant. On the Gram-positive side, 52 of 54 were multidrug-resistant. Seven of 15 enterococci and related bacteria were vancomycin-resistant. The oxacillin figure, 71.79%, is the methicillin-resistance marker in the staphylococci group. The vancomycin figure in that same group is 5.13%.

The team also grew 77 moulds, of which 22 were genetically distinct. All but three of the identified fungi were pathogens or opportunistic pathogens, including several Aspergillus species. Several of the moulds are signs of excess damp in a building, the authors note.

Outdoors, by contrast, only three bacterial respiratory pathogen isolates and one opportunistic pathogen were found.

The indoor and outdoor counts are not like for like in size. Indoors the team tested 77 Gram-negative isolates and 54 Gram-positive ones, and grew 77 moulds. Outdoors they found three bacterial respiratory pathogen isolates and one opportunistic pathogen. The paper's contrast is that the resistant load was inside.

The risk for patients

Heine said the findings were a concern in a country with high HIV and TB rates, which raise the risk of lower respiratory tract infections. In patients who are already immunocompromised, she said, resistant strains "are likely to spread to the blood stream and cause sepsis".

There are guidelines for water and for food surfaces, Heine noted, but none setting how many bacteria and fungi are acceptable in indoor air. "Monitoring bacterial and fungal air quality should be mandatory in healthcare facilities," she said in material released by the university.

The paper recommends that the clinic review its infection prevention and control programme, and that the national infection prevention and control framework and national air quality guidelines be updated to include routine air sampling with limits for healthcare settings.

That recommendation is for a limit that does not exist yet. Heine's point is that water and food surfaces have standards and indoor air in a clinic does not. The paper asks for the national framework and the air-quality guidelines to add routine sampling and limits. It does not itself set the number of bacteria per cubic metre that should be allowed. The growth media were selective, so the team could not estimate that total.

What the study does not show

It is one clinic, sampled over two months, which brings a seasonal bias. The growth media were selective, so the team could not estimate total bacterial counts per cubic metre, and money ran short to sequence every mould. Finding resistant bacteria in the air shows a possible source of exposure, the paper says, not that any patient was infected.

The clinic stays unnamed. February and March 2025 are the two months. Exposure is what the samples show. Infection of a named patient is what they do not show.