Abstract:

Background: Colistin (Polymyxin) antibiotics are reserve antibiotics for the treatment of severe and invasive infections caused by Gram-negative bacteria, and they are used as alternative antimicrobial therapeutic options to Carbapenem-resistant Gram-negative bacteria (CarbR-GNB) infections. Antimicrobial resistance (AMR) surveillance and research data in Tanzania are showing progressive increase in CarbR-GNB notably among Pseudomonas aeruginosa and Acinetobacter spp. Surprisingly, while Colistin is a reserved antibiotic in the human sector, it is widely used as a first line antimicrobial agent in veterinary sector. This calls for an urgent need to evaluate the burden of Colistin resistance (ColR) among CarbR-GNB isolates using a One-Health approach to guide holistic and specific interventions.

Objective: To determine ColR among CarbR-GNB bacteria isolated from human (patients), domestic animals (livestock and poultry) and environment (sewage and river water) in Mwanza, Tanzania.

Methods: A cross-sectional study was conducted in Mwanza city from January to August 2024 involving 15,196 clinical samples (urine, blood, pus, sputum, body fluids and stool /rectal swabs) from patients attending Bugando Medical Centre in the on-going hospital wide AMR Surveillance. In addition, 427 domestic animals (211 rectal samples from cattle and 216 cloaca samples from poultry) and 484 environmental premises samples (sewage water from waste water treatment plant and water from different areas along River Mirongo) were also collected from June to August 2024. Samples were cultured in plain MacConkey agar (for human samples) and MacConkey agar supplemented with 2 μg/mL cefotaxime (MCA-C) (for animals, water samples and stool samples from human). Bacterial identification was performed using conventional biochemical tests and antimicrobial susceptibility testing (AST) to determine phenotypic CarbR and ColR using Kirby Bauer’s disc diffusion method, ColR resistant strains were selectively isolated in Muller Hinton agar containing 2 μg/mL. The ColR strains were subjected to multiplex polymerase chain reaction for eight mobile ColR genes (mcr-1 to 9).


Results: A total of 290 CarbR GNB were isolated, and the proportions in the three interfaces were 7.6% (229/3008) in humans, 8.7% (37/427) in domestic animals, and 5.0% (24/484) in the environmental premises. The CarbR Acinetobacter baumanii and Pseudomonas aeruginosa were predominant isolates in patients, whereas, CarbR Escherichia coli was predominant across three interfaces. The overall proportion of ColR among CarbR-GNB was 25.8% (75/290); (95%CI: 21.1-31.2%). The distribution of 75 ColR isolates in patients, livestock, poultry and environmental premises were 24.9% (57/229), 33.3% (7/21), 50.0 % (8/16), and 12.5% (3/24), respectively, and the ColR was significantly more in poultry (OR: 95% CI = 7.0 (1.2-48.8), p-value=0.01). The most common ColR isolates were Acinetobacter baumanii and Escherichia coli in patients, and Stenotrophomonas maltophilia and Pseudomonas aeruginosa in domestic animals. Multiplex PCR based typeability of mcr genes was 14.7 % (11/75), and demonstrated the presence of mcr-3, mcr- 7, mcr-8 and mcr-9 across human and domestic animals’ interfaces, mcr-1 was demonstrated in human and mcr-4 in animal isolates and no mcr genes were detected among environmental isolates.

Conclusion: The overall prevalence of CarbR in the three interfaces was low (5.0% to 9.0%), however, a quarter of CarbR GNB isolates displayed ColR phenotypically. The carbapenemcolistin co-resistance was significantly higher in poultry. There was low typeability of the mcr genes using multiplex PCR technique. The findings underscore an urgent need to strengthen OneHealth AMR surveillance systems in Mwanza, Tanzania with specific focus on the resistance to the two last resort antibiotics for GNB. Elucidating alternative ColR mechanisms using robust molecular typing methods like whole genome sequencing is also recommended.