Phenotypic Associations in Resistance to Medically Important Antibiotics Among Highly Resistant Bacteria Isolated from Cambodian Dairy Farms

Abstract

Antimicrobial resistance (AMR) is recognized by global health authorities as one of the foremost public health and development challenges confronting humanity today, leading the WHO to establish categories of medically important antibiotics for humans and animals. On farms with a high burden of infection, inappropriate antibiotic use, mixed bacterial infections, and environmental contamination with antibiotics from animal sources might accelerate the development of antibiotic resistance, particularly from a One Health Approach. Therefore, this study was conducted to determine correlations in AMR among antibiotics using bacteria isolated from dairy farms with a high burden of intramammary infections. In addition, antibiotic resistance percentages were compared between a smallholder dairy farm and an intensive dairy farm in Cambodia, where the dairy industry has been established for only a few years. Bacterial isolates were collected from two dairy farms: Farm A, a medium-sized local farm with approximately 30 milking cows, and Farm B, a large intensive farm with approximately 480 milking purebred Holstein-Friesian milking cows. All bacteria were tested for antimicrobial susceptibility using a disk diffusion test with 17 antibiotics from nine classes. Classifications based on CLSI interpretive criteria were distinguished from those based on study-defined, extrapolated inhibition-zone thresholds, designated expected AMR (AMR-E). The latter classifications were exploratory and were not validated measures of resistance. The antimicrobials included antibiotics from classes important in human medicine: human-used only (HU), highest-priority critically important antimicrobials (HPCIA), critically important antimicrobials (CIAs), and highly important antimicrobials (HIAs). Fisher’s exact test was used to compare AMR between farms (p < 0.05), and Spearman’s correlation adjusted by a Benjamini–Hochberg (Q < 0.05) was used to evaluate associations among antimicrobial susceptibility classifications and between these classifications and multidrug resistance (MDR) status. The results showed that the distribution of bacteria differed between the two farms. For antibiotics used only in humans, no imipenem-resistant Enterobacterales were detected, whereas three Enterobacterales isolates were classified as resistant to meropenem. Differences between farms varied by bacterial group and antimicrobial agent. Many significant correlations in AMR were found among antibiotics, especially those within the same class. Different correlations among antibiotics were found across bacterial groups, including Enterobacterales, staphylococci, enterococci, and streptococci. Phenotypic associations were observed among antimicrobial susceptibility classifications, particularly for agents within the same class. These associations do not establish shared resistance mechanisms or predict treatment outcomes. Findings based on extrapolated thresholds require validation using appropriate susceptibility testing methods.