Incidence and prevalence of antimicrobial resistance in outpatients with cancer: a multicentre, retrospective, cohort study

Background

Infections are the second leading cause of death in patients with cancer and are often caused by resistant bacteria. However, the frequency of antimicrobial resistance (AMR) in outpatients with cancer is not well understood. We aimed to compare the frequency of AMR bacterial pathogens in outpatients with and without cancer.

Methods

This retrospective cohort study evaluated antimicrobial susceptibility of bacteria isolated from adults (aged ≥18 years) with and without cancer seeking care in 198 outpatient health-care settings in the USA. Data were collected using the BD Insights Research Database. Patients who were not prescribed cancer medications or not admitted to an inpatient cancer unit in the predefined period were categorised as patients without cancer. Patients were included in the cancer cohort if they received medication solely or sometimes indicated for cancer. Data on gender and race or ethnicity were not collected. Non-duplicate and non-contaminant pathogens collected from various samples (ie, blood, intra-abdominal, respiratory, urine, skin or wound, and other) in outpatients were used to assess the coprimary outcomes: overall and source-specific proportions of non-susceptible pathogen isolates with corresponding AMR odds ratios (ORs); and rates of AMR pathogens per 1000 isolates with corresponding AMR incidence rate ratio (IRR) in patients with and without cancer.

Findings

Data were collected between April 1, 2018, and Dec 31, 2022. 53 006 (3·2%) of 1 655 594 pathogens identified were from 27 421 patients with cancer and 1 602 588 (96·8%) were from 928 128 patients without cancer. For Pseudomonas aeruginosa, carbapenem non-susceptibility was higher in pathogen isolates from patients with cancer (816 [14·4%] of 5683) than patients without cancer (10 709 [11·3%] 94 419; OR 1·22 [95% CI 1·13–1·32]). For Enterobacterales, fluoroquinolone non-susceptibility was higher in pathogen isolates from patients with cancer (8662 [28·0%] of 30 867) than patients without cancer (238 479 [21·8%] of 1 095 996; OR 1·44 [1·40–1·47]), as was carbapenem non-susceptibility (472 [1·5%] of 30 867 vs 9165 [0·8%] of 1 095 996; OR 1·89 [1·72–2·07]), multidrug-resistant pathogens (2672 [8·7%] of 30 867 vs 48 962 [4·5%] of 1 095 996; OR 2·03 [1·95–2·11]), and extended-spectrum β-lactamase producers (4343 [16·5%] of 26 327 vs 93 977 [9·4%] of 996 853; OR 1·96 [1·90–2·03]). For Staphylococcus aureus, meticillin resistance was higher in pathogen isolates from patients with cancer (4747 [53·0%] of 8959) than patients without cancer (129 291 [48·3%] of 267 520; OR 1·20 [1·15–1·25]). For Enterococcus spp, vancomycin resistance was higher in pathogen isolates from patients with cancer (1329 [18·6%] of 7145) than patients without cancer (12 333 [9·1%] of 135 772]; ORR 2·20 [2·06–2·34). The rates and corresponding IRRs of AMR pathogens per 1000 isolates was also higher in patients with cancer compared with patients without cancer, particularly for carbapenem non-susceptible P aeruginosa (IRR 2·06 [1·91–2·21]) and vancomycin-resistant enterococci (IRR 3·06 [2·89–3·24]). For all comparisons, p<0·0001.

Interpretation

AMR proportions and IRRs for most key pathogens were up to three-times higher in isolates from outpatients with cancer than those without cancer, highlighting the need for enhanced surveillance, infection prevention, and timely diagnostic stewardship to improve antibiotic prescribing in this population.

Funding

AMR Action Fund.