Multidrug-resistant Gram-negative Bacteremia in Cancer Patients: Development of a Clinical Score for Risk Assessment. (4th October 2017)

• Record Type: Journal Article
• Authors: Carena, Alberto; Laborde, Ana; Rossi, Inés Roccia; Guerrini, Graciela; Jordán, Rosana; Valledor, Alejandra; Nenna, Andrea; Costantini, Patricia; Dictar, Miguel; Caeiro, Juan Pablo; Ibañez, María Luz González; Vizcarra, Pilar; Palacios, Cynthia Jimena; Pinoni, Victoria; Luck, Martín; Racioppi, Agustina; Poletta, Fernando; Herrera, Fabián
• Is Part Of: Open forum infectious diseases. Volume 4(2017)Supplement 1
• Page Start: S710
• Page End: S710
• Publication Date: 2017-10-04
• Subjects: Communicable diseases -- Periodicals; Medical microbiology -- Periodicals; Infection -- Periodicals; 616.9
• DOI: 10.1093/ofid/ofx163.1906
• Languages: English
• ISSNs: 2328-8957

Abstract

Abstract: Background: Multidrug-resistant Gram-negative bacteremias (MDR-GNB) have become a significant problem in several countries and have been reported as one of the major emerging causes of severe and fatal infections in cancer patients. The objective of this study was to evaluate risk factors for MDR-GNB in cancer and stem cell transplant (SCT) patients and develop a clinical score for risk assessment of MDR-GNB. Methods: Prospective multicenter study. Episodes of GNB in adult cancer and SCT patients were included in 10 centers, from July 2014 to November 2016. To evaluate risk factors for MDR-GNB, variables with P < 0.10 in univariate analysis were included in a logistic regression model for multivariate analysis. The predictive performance of the model and score were assessed using sensitivity, specificity, negative and positive predictive value, and area under the ROC curve, with value of 1.0 indicating perfect prediction. We evaluated the posttest probability of the different score values to predict MDR-GNB. Results: Three hundred and ninety-four episodes of GNB were included, being 168 (42.3%) MDR-GNB. We identified six risk factors: previous antibiotic treatment, previous intensive care unit hospitalization, seven or more days of hospitalization until bacteremia, neutropenia during bacteremia, severe mucositis as a clinical source and colonization or previous infection with MDR-GNB. Using a three-point cut-off value, the score had a sensitivity of 58.3% (95% CI 49–67.3%), a specificity of 83.2% (CI 95% 76.7–88.6%), a positive predictive value of 72.4% (95% CI 64.4–79.2%), and a negative predictive value of 72.6% (95% CI 67.9–76.8%). The predictive performance of the score was satisfactory (median AUROC 0.78, 95% CI 0.73–0.83). Compared with episodes of GNB that have two or fewer risk factors, those with three or more have an OR for MDR-GNB of 6.96 (95% CI, 4–12). In episodes with none of the risk factors, the negative likelihood ratio was 0.1 and the posttest probability for MDR-GNB was 6.8%. Conclusion: The development of this score represents a potential tool for routine clinical management, which should be validated in prospective studies. Disclosures: All authors: No reported disclosures.