Colonization dynamics of extended-spectrum beta-lactamase-producing Enterobacterales in the gut of Malawian adults

Articles https://doi.org/10.1038/s41564-022-01216-7 Colonization dynamics of extended-spectrum beta-lactamase-producing Enterobacterales in the gut of Malawian adults Joseph M. Lewis 1,2,3,4 ✉, Madalitso Mphasa1, Rachel Banda1, Mathew A. Beale4, Eva Heinz 2, Jane Mallewa5, Christopher Jewell6, Brian Faragher2, Nicholas R. Thomson Nicholas A. Feasey 1,2,7 4,7 and Drug-resistant bacteria of the order Enterobacterales which produce extended-spectrum beta-lactamase enzymes (ESBL-Enterobacterales, ESBL-E) are global priority pathogens. Antimicrobial stewardship interventions proposed to curb their spread include shorter courses of antimicrobials to reduce selection pressure but individual-level acquisition and selection dynamics are poorly understood. We sampled stool of 425 adults (aged 16–76 years) in Blantyre, Malawi, over 6 months and used multistate modelling and whole-genome sequencing to understand colonization dynamics of ESBL-E. Models suggest a prolonged effect of antimicrobials such that truncating an antimicrobial course at 2 days has a limited effect in reducing colonization. Genomic analysis shows largely indistinguishable diversity of healthcare-associated and community-acquired isolates, hence some apparent acquisition of ESBL-E during hospitalization may instead represent selection from a patient’s microbiota by antimicrobial exposure. Our approach could help guide stewardship protocols; interventions that aim to review and truncate courses of unneeded antimicrobials may be of limited use in preventing ESBL-E colonization. A ntimicrobials are one of the most successful therapies available to modern medicine but the spread of antimicrobial resistance (AMR) is a threat to their effective use. Considerable global effort is being directed at antimicrobial stewardship programmes which the World Health Organization considers a key tool in reducing AMR1. Antimicrobial stewardship at the individual level often emphasizes rationalization of antimicrobials through narrowing their spectrum of action as soon as possible after commencement of broad empiric antimicrobial therapy in severely unwell individuals. The time frame (for example, 48 h) for this is typically pragmatically selected to match likely availability of diagnostic test results. Rationalization of therapy is partly based on the assumption that it will reduce emergence of AMR but the mechanism by which antimicrobial exposure acts at the individual level to promote colonization and/or infection with resistant pathogens, and the dynamics of colonization and decolonization, are not well understood2–5. Improved understanding of the dynamics of individual-level AMR-acquisition under antimicrobial pressure can therefore inform the design of stewardship protocols. One setting in which antimicrobial stewardship is a considerable challenge is in the treatment of severe febrile illness in the low- and middle-income countries of sub-Saharan Africa (sSA). In Blantyre, Malawi, for example, as in much of sSA, limited availability of diagnostics results in prolonged courses of broad-spectrum antimicrobials—mainly ceftriaxone, a third-generation cephalosporin (3GC) antibiotic6—for severe febrile illness. Ceftriaxone has been extensively used since its introduction to the Malawian national formulary in 20057 but this has been associated with an increase in 3GC resistance8, particularly in bacteria of the order Enterobacterales. This is mainly mediated by extended-spectrum beta-lactamase (ESBL) enzymes8–10. ESBL-producing Enterobacterales (henceforth ESBL-E) are an increasing public health challenge throughout much of sSA11,12 and often have few or no locally available treatment options; in Blantyre, 91% of invasive Klebsiella pneumoniae are now 3GC resistant8 and strategies to reduce ESBL-E infections are needed. Gut mucosal colonization with ESBL-E is thought to precede invasive infection, is common across sSA and has often been found to be associated with prior hospitalization and/or antimicrobial exposure12,13. An improved mechanistic understanding of colonization dynamics following these exposures therefore has the potential to inform evidence-based interventions to reduce colonization and hence opportunity for transmission. Here, we present the results from a clinical study of longitudinal ESBL-E carriage in Blantyre, Malawi, sampling adults as they pass through the hospital and are exposed to antimicrobials. We use multistate modelling14 and whole-genome sequencing as a high-resolution bacterial typing tool to describe and understand the dynamics of ESBL-E colonization. Results Antimicrobial exposure drives increase in ESBL-E prevalence. Between 19 February 2017 and 2 October 2018, we recruited 425 adults: (1) 225 patients with sepsis and antimicrobial exposure, admitted to Queen Elizabeth Central Hospital (QECH), Blantyre; (2) 100 antimicrobial-unexposed inpatients and (3) 100 antimicrobial-unexposed community participants (Table 1). There were 1,631 study visits, with successful stool or rectal swab collection at 1,417/1,631 (87%) visits; missing samples were equally distributed across all study arms and visits (Fig. 1 and Supplementary Fig. 1). At least one ESBL-E species was cultured in 723/1,417 Malawi-Liverpool Wellcome Clinical Research Programme, Blantyre, Malawi. 2Liverpool School of Tropical Medicine, Liverpool, UK. 3University of Liverpool, Liverpool, UK. 4Wellcome Sanger Institute, Hinxton, UK. 5Kamuzu University of Health Sciences, Blantyre, Malawi. 6University of Lancaster, Lancaster, UK. 7London School of Hygiene and Tropical Medicine, London, UK. ✉e-mail: 1 Nature Microbiology | VOL 7 | October 2022 | 1593–1604 | www.nature.com/naturemicrobiology 1593 Articles Nature Microbiology Table 1 | Baseline characteristics of included participants Sepsis, receiving antibiotics (n = 225) Inpatient, not receiving antibiotics (n = 100) Community, not receiving antibiotics (n = 100) P Total (n = 425) Age (yr) 35.9 (27.8–43.5) 40.4 (29.1–48.3) 32.5 (24.0–38.4) <0.001 35.6 (26.9–43.9) Male 114/225 (51%) 51/100 (51%) 40/100 (40%) 0.163 205/425 (48%) HIV-positive 143/225 (64%) 12/100 (12%) 18/100 (18%) <0.001 173/425 (41%) HIV-negative 70/225 (31%) 77/100 (77%) 22/100 (22%) 169/425 (40%) HIV unknown 12/225 (5%) 11/100 (11%) 60/100 (60%) 83/425 (20%) Current CPT 98/141 (70%) 5/12 (42%) 7/18 (39%) 0.013 110/171 (64%) Current ART 117/143 (82%) 9/12 (75%) 18/18 (100%) 0.082 144/173 (83%) Months on ART 28.7 (3.7–72.6) 35.1 (2.9–79.8) 31.5 (13.0–79.9) 0.698 29.5 (3.8-72.8) Antibiotics within 28 db 60/225 (27%) 0/100 (0%) 0/100 (0%) <0.001 60/425 (14%) Hospitalized within 28 d 18/225 (8%) 1/100 (1%) 0/100 (0%) <0.001 19/425 (4%) Current TB treatment 10/225 (4%) 0/100 (0%) 4/100 (4%) 0.083 14/425 (3%) Number of adults 2.0 (2.0–3.0) 3.0 (2.0–4.0) 2.0 (2.0–4.0) 0.907 3.0 (2.0–4.0) Number of children 2.0 (1.0–3.0) 2.0 (1.0–3.0) 2.0 ( (...truncated)