A Multicenter, Prospective, Observational Study to Determine Predictive Factors for Multidrug-Resistant Pneumonia in Critically Ill Adults: The DEFINE Study

Objective: This study was conducted to describe the prevalence, epidemiology, and clinical outcomes of multidrug-resistant (MDR) organism (MDRO) pneumonia in critically ill patients. Methods: A multicenter, prospective, observational study of patients admitted to 60 intensive care units (ICUs), from 34 hospitals, in the United States from November to December 2016. Adults (> 18 yrs) receiving antimicrobial therapy at least 5 days for pneumonia were included. Patients were classified into two categories, with or without MDRO, and subcategorized by pneumonia type. Measurements and main results: Demographics, medication histories, and health care exposure were collected during ICU admission and compared using t test and chi-square tests. Multivariate logistic regression was used to determine predictive factors for MDRO pneumonia and hospital mortality. Of 652 patients, 92 patients (14.1%) developed MDR pneumonia. Predictors of MDRO pneumonia were acid suppression therapy within the previous 90 days (odds ratio [OR] 1.88 [1.14–3.09]; p=0.013), mechanical ventilation (OR 1.96 [1.14–3.35]; p<0.001), and history of MDRO infection (OR 4.74 [2.21–10.18]; p<0.001). Appropriate initial antimicrobial selection occurred in 58 patients (63%) with MDRO pneumonia compared to 464 patients (82.7%) in patients without MDRO pneumonia (p<0.001). MDRO pneumonia was not associated with hospital mortality (18.5% vs 17.6%, p=0.087). Conclusions: In a broad cohort of critically ill patients, MDRO pneumonia is infrequent, and associated with factors describing the intensity of health care provided. Presence of MDRO pneumonia is not associated with hospital mortality. Further study is needed to clarify risk factors for multidrug-resistant pneumonia in critically ill patients.