Mortality Impact of CLSI Carbapenem Breakpoint Changes in Enterobacteriaceae Bloodstream Infections: a Patient-level Analysis of Published Data

BACKGROUND: In 2010, the Clinical and Laboratory Standards Institute (CLSI) lowered susceptibility breakpoints for carbapenem agents, though few studies have evaluated the clinical impact of these changes. The objective of this study was to determine whether a clinical breakpoint exists for carbapenems in patients with Gram-negative Enterobacteriaceaebloodstream infections (GNBSIs). METHODS: Patient-level data from 4 publications (Rhodes JN 2016, Biehle LR 2015, Patel TS 2015, Esterly JS 2012) reporting outcomes for GNBSIs treated with carbapenems for ≥48 hours were compiled. Patients with an MIC ≥16 mg/L to both imipenem (IMI) and meropenem (MER) were excluded. Classification and Regression Tree (CART) analyses were used to determine optimal splits for carbapenem MIC with respect to 30-day all-cause mortality. Univariate and multivariate regression analyses were conducted using Stata 14. RESULTS: A total of 194 patients were extracted. Klebsiella pneumoniae was the most common pathogen (70.1%) followed by Escherichia coli(15.0%). Primary bacteremia/unknown and urinary tract were the most common sources of GNBSI (31.4% and 25.8%, respectively). MER and IMI MICs were available for 144 patients (74.2%) and 138 patients (71.1%), respectively. Carbapenem agent used was known for 141 patients, of which 94 received MER, 24 received ertapenem, 12 received doripenem, and 11 received IMI. CART analysis identified a significant difference in mortality between patients infected with organisms having MER MICs ≤ 1 mg/L (n = 10/121, 8.3%) vs. those with >1 mg/L (n = 7/21, 33.3%; P <0.01) regardless of carbapenem used. This breakpoint was also identified in the subgroup of patients with available MER MICs who were treated with MER (n = 5/64, 7.8% vs. 7/19, 36.8%; P < 0.01). In multivariate logistic regression, MER MIC > 1 mg/L was associated with increased odds of 30-day mortality after controlling for ICU admission in the any carbapenem treatment (OR 5.0, 95% CI 1.63-15.6; P < 0.01) and MER treated populations (OR 7.16, 1.88-27.3; P < 0.01). CONCLUSION: This pooled patient-level analysis of GNBSIs treated with carbapenems represents the largest of its kind to date. A significant increase in mortality was identified in patients with MER non-susceptible isolates as defined by the 2010 CLSI breakpoints. DISCLOSURES: T. Patel, Merck: Grant Investigator, Research grant. J. Esterly, Merck: Employee, Salary. E. B. Hirsch, Merck: Grant Investigator, Grant recipient. The Medicines Company: Speaker’s Bureau, Speaker honorarium