Comparison of Multiple Carbapenemase Tests Based on an Unbiased Colony Selection Method

Abstract

AbstractCarbapenemase-producing organisms (CPOs) present a major threat to public health, demanding precise diagnostic techniques their detection. Discrepancies among CPO tests have raised concerns, partly due to limitations in detecting bacterial diversity within host/specimen. We explored the impact of unbiased colony selection on carbapenemase testing and assessed its relevance on various tests. Based on “FirstAll” for unbiased colony selection to reduce bias, we compared modified carbapenem inactivation method/EDTA-modified carbapenem inactivation method (mCIM/eCIM), Carba5, the CPO panel, and multiplex PCR (M-PCR). Initially, we compared FirstAll to conventional colony selection for mCIM. Second, we used M-PCR as a reference, to evaluate test performance across seven CPO species. The results revealed that FirstAll selection improved carbapenemase detection, revising false-negative in 10.5% ofK. pneumoniaeisolates. In addition, 12.4% of CPOs tested positive for multiple carbapenemase genes. Both the Carba5 test and CPO panel showed suboptimal performance (sensitivity/specificity: Carba5 75.5%/89.0%, CPO panel 78.1%/74.0%). Carba5 test provided specific carbapenemase class assignments but CPO panel failed in 20.3% of cases. Carba5 test and the CPO panel results correlated well with ceftazidime-avibactam minimal inhibitory concentrations (MICs). Concordance for class A/D with MICs was 88.3% for Carba5 and 92.0% for the CPO panel; whereas for class B, it was 86.5% for Carba5 and 76.2% for the CPO panel. In conclusion, FirstAll as the unbiased colony selection impacted carbapenemase testing. With FirstAll, the diagnostic performance of either Carba5 or the CPO panel was compromised. The utilization of ceftazidime-avibactam guided by either the CPO panel or Carba5 was appropriate.ImportanceThe increasing carbapenemase-producing organisms (CPO) is concerning due to high mortality rates and limited treatment options. Precise testing for CPO is crucial not only for antibiotic treatments but also for infection control. However, discrepant results for an individual overtime or even intra-specimen are found in either phenotypic or genetic testing, posing considerable challenge in clinical management. Based on the colonization-infection model of CPO infections, there would be strain heterogeneity in an individual. On top of the heterogeneity, the single colony selection method in conventional CPO testing would be the source of discrepancy and bias. To test the hypothesis, we proposed FirstAll method as the unbiased colony selection method. We demonstrated that FirstAll corrected around 10% false-negative cases. Lower diagnostic performances of CPO tests were also found in comparison to previous related studies. The study revealed that colony selection would have considerable impacts on CPO testing.