Abstract
Objectives This study aimed to evaluate the performance of routinely used phenotypic tests to detect β-lactamase production in isolates coproducing multiple β-lactamase types. Methods Commonly used phenotypic tests for the detection of extended spectrum β-lactamases (ESBL), AmpC β-lactamase, and carbapenemases were compared with detection and sequencing of β-lactamase genes (as the reference test) in 176 uropathogenic Enterobacteriaceae coproducing multiple β-lactamases from two hospitals in the Western Province of Sri Lanka. Results Majority of the isolates (147/176, 83.5%) carried β-lactamase genes with (90/147, 61%) harboring multiple genes. The Clinical and Laboratory Standards Institute screening method using cefotaxime (sensitivity [Se], 97; specificity [Sp], 93; accuracy [Ac], 94) and ceftriaxone (Se, 97; Sp, 91; Ac, 93) was the most effective to detect ESBLs. The modified double disc synergy test (Se, 98; Sp, 98; Ac, 97) and combined disc test (Se, 94; Sp, 98; Ac, 96) showed good specificity for confirmation of ESBLs. Cefoxitin resistance (Se, 97; Sp, 73; Ac, 85) and the AmpC disc test (Se, 96; Sp, 82; Ac, 86) were sensitive to detect AmpC β-lactamase producers coproducing other β-lactamases but showed low specificity, probably due to coproduction of carbapenemases. Meropenem was useful to screen for New Delhi metallo β-lactamases and OXA-48-like carbapenemases (Se, 97; Sp, 96; Ac, 96). The modified carbapenem inactivation method showed excellent performance (Se, 97; Sp, 98; Ac, 97) in identifying production of both types of carbapenemases and was able to distinguish this from carbapenem resistance due to potential mutations in the porin gene. Conclusion Microbiology laboratories that are still depend on phenotypic tests should utilize tests that are compatible with the types of β-lactamase prevalent in the region and those that are least affected by coexisting resistance mechanisms.