# Uncovering antimicrobial resistance structures in Staphylococcus spp. from companion animals: latent class analysis of isolates from dogs

## TL;DR

This study identifies hidden patterns of antibiotic resistance in Staphylococcus bacteria from dogs in Peru, revealing subgroups that could impact treatment effectiveness.

## Contribution

The study introduces latent class analysis to uncover novel resistance structures in canine Staphylococcus isolates beyond standard MDR classifications.

## Key findings

- High resistance to β-lactams (99.7%) and moderate resistance to other classes were observed in Staphylococcus isolates from dogs.
- Latent class analysis identified four distinct resistance phenotypes among canine isolates, including one with extensive resistance to multiple antibiotic classes.
- The findings suggest the need for improved antimicrobial stewardship and surveillance in veterinary medicine to address emerging resistance patterns.

## Abstract

Antimicrobial resistance (AMR) in companion animals poses therapeutic challenges and potential zoonotic risks. Staphylococcus spp., a leading cause of skin and ear infections in dogs and cats, shows high levels of resistance to different families of antibiotics. This study aimed to describe AMR patterns in Staphylococcus spp. isolated from companion animals in Lima, Peru, and to identify latent resistance phenotypes among canine isolates using latent class analysis (LCA). We analyzed 2,159 isolates (2,035 from dogs and 124 from cats) collected between 2021 and 2024, assessing resistance to seven antimicrobial classes according to CLSI guidelines. High resistance was found against β-lactams (99.7%), followed by sulfonamides (41.6%) and tetracyclines (38.5%), while glycopeptide resistance was rare (0.14%). Overall, MDR prevalence was 49.9%, reaching 52.5% in isolates from dogs. LCA applied to canine isolates identified four latent phenotypes: Phenotype 1 (50.3%) combined β-lactam resistance with moderate resistance to macrolides, aminoglycosides, tetracyclines, and sulfonamides; Phenotype 2 (28.2%) showed β-lactam resistance with low resistance to other classes; Phenotype 3 (13.6%) included isolates resistant to β-lactams and fluoroquinolones; and Phenotype 4 (7.9%) displayed extensive resistance to β-lactams, fluoroquinolones, and tetracyclines. LCA revealed hidden resistance structures beyond conventional MDR definitions, identifying subgroups with clinically relevant resistance combinations that may reduce antimicrobial effectiveness. These latent phenotypes have direct therapeutic relevance, as they reveal resistance patterns that may compromise empirical treatment choices in small animal practice. Overall, our findings highlight the need to strengthen AMR surveillance and antimicrobial stewardship within a One Health framework in veterinary medicine.