Exposure to doxycycline increases risk of carrying a broad range of enteric antimicrobial resistance determinants in an elderly cohort

Objectives

High rates of antibiotic prescription in residential aged care are likely to promote enteric carriage of antibiotic-resistant pathogens and increase the risk of antibiotic treatment failure. Despite their importance, relationships between antibiotic exposures and patterns of enteric resistance carriage in this population remain poorly understood.

Methods

We conducted a cross-sectional metagenomic cohort analysis of stool samples from residents of five long-term aged-care facilities in South Australia. Taxonomic composition was determined, and enteric carriage of antibiotic resistance genes (ARGs) was identified and quantified against the Comprehensive Antibiotic Resistance Database. Both the detection and abundance of stool taxa and ARGs were related to antibiotic exposures up to 12 months prior. Factors associated with the abundance of ARGs of high clinical concern were identified.

Results

Stool samples were provided by 164 participants (median age: 88 years, IQR 81–93; 72% female). Sixty-one percent (n = 100) of participants were prescribed antibiotics at least once in the prior 12 months (median prescriptions: 4, range: 1–52), most commonly a penicillin (n = 55, 33.5%), cephalosporin (n = 53, 32.3%), diaminopyrimidine (trimethoprim) (n = 36, 22%), or tetracycline (doxycycline) (n = 21, 12.8%). More than 1100 unique ARGs, conferring resistance to 38 antibiotic classes, were identified, including 20 ARGs of high clinical concern. Multivariate logistic regression showed doxycycline exposure to be the greatest risk factor for high ARG abundance (adjusted odds ratio [aOR]=14.8, q<0.001) and a significant contributor to inter-class selection, particularly for ARGs relating to penicillins (aOR=3.1, q=0.0004) and cephalosporins (aOR=3.4, q=0.003). High enteric ARG abundance was associated with the number of separate antibiotic exposures (aOR: 6.4, q<0.001), exposures within the prior 30 days (aOR: 4.6, q=0.008) and prior 30–100 days (aOR: 2.6, q=0.008), high duration of antibiotic exposure (aOR: 7.9, q<0.001), and exposure to 3 or more antibiotic classes (aOR: 7.4, q<0.001). Carriage of one or more ARGs of high clinical concern was identified in 99% of participants (n = 162, median: 3, IQR: 2–4), involving 11 ARGs conferring resistance to aminoglycosides, four to beta-lactams, one to glycopeptides, three to fluoroquinolones, and one to oxazolidinones. Carriage of ARGs of high clinical concern was positively associated with exposure to doxycycline (aminoglycoside, fluoroquinolone, and oxazolidinone ARGs) and trimethoprim (fluoroquinolone and beta-lactam ARGs). Analysis of doxycycline impact on microbiota composition suggested that observed resistome changes arose principally through direct ARG selection, rather than through the antibiotic depletion of sensitive bacterial populations.

Conclusions

The gut microbiome of aged care residents is a major reservoir of antibiotic resistance. As a critical antibiotic in medical practice, a comprehensive understanding of the impact of doxycycline exposure on the gut resistome is paramount for informed antibiotic use, particularly in an evolving landscape of prophylactic applications. Near-universal asymptomatic carriage of clinically critical resistance determinants is highly concerning and reinforces the urgent need for improved management of antibiotic use in long-term aged care.