Total antibiotic use in a state-wide area and resistance patterns in Brazilian hospitals: an ecologic study.

INTRODUCTION: Use of antibiotic and bacterial resistance is the result of a complex interaction not completely understood. OBJECTIVES: To evaluate the impact of entire antimicrobial use (community plus hospitals) on the incidence of bloodstream infections in intensive care units adjusted by socioeconomic factors, quality of healthcare, and access to the healthcare system. DESIGN: Ecologic study using a hierarchical spatial model. SETTING: Data obtained from 309 hospitals located in the state of São Paulo, Brazil from 2008 to 2011. PARTICIPANTS: Intensive care units located at participant hospitals. OUTCOME: Hospital acquired bloodstream infection caused by MDRO in ICU patients was our primary outcome and data were retrieved from São Paulo Health State Department. Socioeconomic and healthcare indexes data were obtained from IBGE (Brazilian Foundation in charge of national decennial census) and SEADE (São Paulo Planning and Development Department). Information on antimicrobial sales were obtained from IMS Brazil. We divided antibiotics into four different groups (1-4). RESULTS: We observed a direct association between the use of group 1 of antibiotics and the incidences of bloodstream infections caused by MRSA (1.12; 1.04-1.20), and CR-Acinetobacter sp. (1.19; 1.10-1.29). Groups 2 and 4 were directly associated to VRE (1.72; 1.13-2.39 and 2.22; 1.62-2.98, respectively). Group 2 was inversely associated to MRSA (0.87; 0.78-0.96) and CR-Acinetobacter sp. (0.79; 0.62-0.97). Group 3 was inversely associated to Pseudomonas aeruginosa (0.69; 0.45-0.98), MRSA (0.85; 0.72-0.97) and VRE (0.48; 0.21-0.84). No association was observed for third generation cephalosporin-resistant Klebsiella pneumoniae and Escherichia coli. CONCLUSIONS: The association between entire antibiotic use and resistance in ICU was poor and not consistent for all combinations of antimicrobial groups and pathogens even after adjusted by socioeconomic indexes. Selective pressure exerted at the community level seemed not to affect the incidences of MDRO infection observed in intensive care setting.