Comparison of 30-day crude mortality rates of different antibiotic therapies and identification of independent risk factors for 30-day crude mortality in bloodstream infections with non-KPC carbapenemase producing Klebsiella pneumoniae: Importance of appropriate therapy and combination therapy

Abstract Background. The best available antibiotic treatment against carbapenemase-producing Klebsiella pneumoniae (CPKp) has not been clearly defined particularly for OXA-48 carbapenemase producing strains. We compared different antibiotic regimens for 30-day crude mortality rates of bloodstream infections (BSIs) caused by these bacteria. Methods. This is a retrospective cohort study including all non-KPC CPKp BSIs treated between 01.01.2014 and 31.07.2018. Antibiotic therapy was appropriate if initiated within 5 days from the onset of BSI and including at least an active drug with an adequate dosage. Identification of bacteria were performed by automated systems and in vitro susceptibility testing was evaluated by EUCAST breakpoints. MICs of carbapenems and other antibiotics were confirmed by E-test. Colistin MIC was confimed by broth microdilution. Tigecycline MICs were evaluated according to US FDA criteria. Molecular detection of carbapenemases were performed with PCR for blaKPC, blaNDM, blaVIM, blaOXA-48. Results. Of the 174 CRKp BSIs occuring during study period, 129 patients met all the inclusion criteria. Susceptibility to antibiotics was as follows: colistin (51.9%), tigecycline (67.4%), gentamicine (38.0%), amikacin (43.0%), meropenem (32.6%), ertapenem (3.1%), ciprofloxacin (18.6%), trimethoprim-sulfamethoxazole (24.0%), ceftazidime (7.8%) and cefepime (7.8%), ceftriaxone (7.0%). 85.3% and 3.1% of isolates harbored OXA-48 and NDM genes, respectively. In addition, 9.3% of the isolates carried OXA-48 and NDM concomitantly and 2.3% encoded OXA-48 and VIM together. Inappropriate therapy was associated with higher 30 day crude mortality rate than appropriate therapy (adjusted OR: 4.29; 95% CI, 1.77-10.39; p=0.001). However, there were no significant difference between monotherapy and combination therapy arms in terms of 30-day crude mortality rate (adjusted OR: 2.2; 95% CI, 0.45-10.6; P=0.32). In multivariate logistic regression analysis, Increment mortality score (adjusted OR: 1.27; 95% CI, 1.14-1.42; P <0.001) and interaction between inappropriate therapy and colistin susceptibility (OR: 4.04; 95% CI: 1.62-10.02; P=0.003) were identified as independent risk factors for 30 day all cause mortality. Conclusions. Appropriate regimens were associated with lower mortality rates than inappropriate therapy in BSI with CPKp possesing OXA-48, dominantly. On the other hand, combination of two or more active agents and monotherapy had similar efficacy in terms of 30-day crude mortality. Colistin resistance and inappropriate therapy were not associated with increased mortality by themselves, however their interaction increased 30-day mortality rate significantly. Thus, high rate of colistin resistance among CPKp isolates remains as a significant cause of mortality in our setting