Multiple‑Antibiotic Resistance (MAR) Index, Not Integron Prevalence, Predicts Ciprofloxacin Resistance in Community Ampicillin-Resistant Escherichia coli from Turkiye

Background Extended spectrum β-lactamase producing (ESBL) Escherichia coli is becoming more common in the community globally, but it is still unclear what factors contribute to multidrug resistance (MDR) outside of hospitals. To measure the correlation between the multiple antibiotic resistance (MAR) index, ciprofloxacin resistance, and class 1 integron carriage in the Turkish community of ampicillin-resistant E. coli . Methods Sixty-five ampicillin-resistant E. coli were found. The VITEK 2 system was used to determine susceptibility to 13 antibiotics. Phylogroups, integrase genes, bla genes, and plasmid-mediated quinolone resistance (PMQR) genes were all identified by PCR. The binary resistance matrix was subjected to ward hierarchical clustering (Euclidean distance). Firth penalized multivariable logistic regression in R was used to evaluate the predictors of ciprofloxacin resistance; variables with univariable p  < 0.10 were placed in a backward model. Results The greatest resistance was to ciprofloxacin (63.1%) and trimethoprim-sulfamethoxazole, while no carbapenem resistance was detected. The bla _CTX−M−1 alleles (93.7%) dominated the ESBL phenotype, which was present in 29.2% of isolates. 37 out of 65 isolates (56.9%) had class 1 integrons, which were found to contain primarily of aadA1, aadA5, dfrA7 -like and dfrA17 . With a median MAR index of 0.23 (IQR 0.07–0.46), 53.8% of the population was above the ecological risk threshold of 0.20. The only independent predictor of ciprofloxacin resistance in the Firth model was the MAR index ( β  = 3.53 ± 0.90; OR = 34.2, 95% CI = 4.0–292; p  < 0.001). Following false discovery rate correction, the genes for integron carriage, bla _CTX−M , and PMQR were not significant ( q  > 0.40). Due to sparse fully susceptible rows, the unpenalized model showed quasi separation with an extreme OR (2.1 × 10⁹). Three phenotypic clusters (high, intermediate, and low MAR) were resolved by Ward clustering; however, integrons were dispersed equally among the clusters ( χ² = 2.3, p  = 0.31). Conclusion Both ciprofloxacin resistance and phenotypic clustering in community ampicillin-resistant E. coli are explained by the MAR index rather than integron carriage. Integrons by themselves are therefore inadequate indicators of the burden of MDR in a community, and holistic resistance metrics perform better than single genetic markers.