BackgroundCarbapenem-resistant Enterobacteriaceae (CRE) are a growing threat globally. Many CRE organisms carry plasmids that produce a carbapenemase enzyme (CP-CRE). Early detection of CP-CRE and aggressive infection prevention and control (IPC) measures are necessary to reduce transmission of resistant organisms and their plasmids. Many hospital laboratories do not test CRE for carbapenemase production. In Apr ‘18, the Philadelphia Department for Public Health (PDPH) established mandatory CRE reporting including isolate submission for mechanism testing at public health laboratories.MethodsWe analyzed trends of carbapenem resistance in CP and non-CP-CRE. We calculated sensitivity (Sen), specificity (Spe), positive and negative predictive values (PPV, NPV) at each level of carbapenem resistance to develop a predictive model with goal of effectively discriminating CP and non-CP CRE. ROC curves were plotted.ResultsFrom June,’18-Feb,’20, 351 CRE had genetic mechanism testing. 192 (54.7%) were Klebsiella pneumoniae, 53 (15.1%) were Enterobacter cloacae, 52 (14.8%) were E. coli, and the remaining 54 (15.4%) other Enterobacteriaceae.186 (53.0%) had a recorded minimum inhibitory concentration (MIC) for ertapenem, 191 (54.4%) for meropenem, 116 (33.0%) for imipenem, and 9 (2.6%) for doripenem. Doripenem was not further analyzed. The odds of being CP-CRE increased with increasing MIC. One standard dilution increase in MIC was associated with the following odds ratios (OR) of being CP-CRE: ertapenem OR 1.43 (95% CI: 1.11–1.83), imipenem OR 2.52 (95% CI: 1.7–3.793), meropenem, OR 2.20 (95% CI: 1.72–2.81). For each carbapenem, we calculated Sen, Spe, PPV, and NPV values for five MIC cut points; 0.5, 1, 2, 4, and 8 ug/ml, non-CP CRE defined as ≤ MIC cut point, CP-CRE defined as > cut point.Figure 1. ROC Curves Table 1. Sensitivity, Specificity, PPV, NPV of MIC Cut Points ConclusionThere were no MIC cut points that performed well in both Sen and Spe, however, for the purposes of IPC, correctly classifying CP CRE (Sen) is more important than correctly classifying non-CP CRE (Spe). MIC cut points that performed best were imipenem MIC of 1 ug/ml and meropenem MIC of 1 ug/ml, with Sen 98%, 94% respectively, PPV 85%, 84% respectively and NPV 88%, 76% respectively. When further testing is not available, MIC cut points may be used to infer CP production.Disclosures All Authors: No reported disclosures
Read full abstract