Vigilância epidemiológica e mortalidade de pacientes críticos com bacteremia por bacilos Gram-Negativos internados em hospital terciário de grande porte no Brasil

Abstract

Introduction - Bloodstream infections (BSI) are among the most frequent hospital infections, particularly in critically ill patients, with high epidemiological indicators of morbidity, mortality, and costs. The impact of the etiology of ICS by Gram-negative bacilli (GNB) multidrug-resistance (MDR) and its relationship with inappropriate initial empirical antibiotic therapy (IIEAT) and mortality are little known in Brazil due to the lack of data. Objectives - To determine the risk factors and prognosis, as well as the impact of IIEAT on the treatment and evolution of patients with BSI by GNB MDR. Material and methods - An observational, retrospective study was carried out with patients admitted to an adult intensive care unit (ICU), clinical-surgical, in a tertiary hospital, from January/2012 to December/2018, which resulted in a cohort of 270 patients, with the first episode of this infection, with positive blood culture by GNB. The search for demographic, clinical, and microbiological data, including the profile of resistance to antimicrobials, was carried out thorough research in the files of the Hospital Infection Control Service and patient records. Univariate and multivariate statistical analyzes were performed to assess mortality predictors for BSIs by MDR and IIEAT microorganisms. Results - In total, 145 (53.7%) of the 270 patients included in the study did not survive hospitalization at the unit. In this group, the oldest (58.8 years vs. 46.9 years) and the most severely ill (100.0% vs. 93.6%, p ≥ 0.05) stood out about those discharged, but without statistical differences. On the other hand, they showed as predictors of mortality: septic shock (OR= 2.5; CI= 1.3-4.9; p= 0.0051), immunocompromise (OR= 2.6; CI= 1.3- 5.2; p= 0.0066) and use of mechanical ventilation (OR= 28.0; CI= 6.3-123.6; p= <0.0001). However, infections by MDR microorganisms (68.2% vs. 64.8%), as well as the IIEAT prescription (67.5% vs. 74.4%) were not predictors of these patient’s non-survival. In the multivariate analysis, infection caused by MDR phenotypes presented as predictors: previous use of antibiotics (OR= 2.8; CI= 1.2-6.6; p= 0.0155), with emphasis on carbapenems (OR= 2,1; CI= 1.0-4.3; p= 0.0409), as well as the IIEAT (OR= 3.8; CI= 2.2-6.5; p= <0.0001). In contrast, previous exposure to antibiotics (OR= 2.2; CI= 1.0-4.9; p= 0.0370), etiology by non-fermenting MDR GNB samples (OR= 5.3; CI= 1.7-16.1; p= 0.0030) and tracheostomy (OR= 2.0; CI= 1.0-3.7; p= 0.0280) were predictive factors for IIEAT. Conclusions - The etiology of ICS by GNB MDR samples, especially non-fermenters, 15 were risk factors for IIEAT, as well as previous use of antibiotics, although they were not predictors of hospital mortality, which is significantly associated with the severity of the disease of the patient and the use of mechanical ventilation as an invasive procedure. Due to the scarcity of therapeutic options in the therapies of these infections, prevention and control measures and the development of new antibiotics are critical aspects to be considered.