Nosocomial bloodstream infections in neurosurgery: a 10-year analysis in a center with high antimicrobial drug-resistance prevalence.

Background: Data on nosocomial bloodstream infections (NBSI) in neurosurgery is limited. This study aimed to analyze the epidemiology, microbiology, outcome, and risk factors for death in neurosurgical patients with NBSI in a multidrug resistant setting.
Methods: Neurosurgical patients with a confirmed NBSI within the period 2003-2012 were retrospectively analyzed. NBSI was diagnosed when a pathogen was isolated from a blood sample obtained after the first 48 h of hospitalization. Patients' demographic, clinical, and microbiological data were recorded and analyzed using univariate and multivariate analysis.
Results: A total of 236 patients with NBSI were identified and 378 isolates were recovered from blood cultures. Incidence of NBSI was 4.3 infections/1000 bed-days. Gram-negative bacteria slightly predominated (54.5 %). The commonest bacteria were coagulase-negative staphylococci (CoNS, 26 %), Klebsiella pneumoniae (15.3 %), Pseudomonas aeruginosa (14.8 %), and Acinetobacter baumannii (13.2 %). Carbapenem resistance was found in 90 % of A. baumannii, in 66 % of P. aeruginosa, and in 22 % (2003-2007) to 77 % (2008-2012) of K. pneumoniae isolates (p < 0.05). Most CoNS and Staphylococcus aureus isolates (94 and 80 %, respectively) were methicillin-resistant. All Gram-negative isolates were sensitive to colistin and all Gram-positive isolates were sensitive to vancomycin and linezolid. Antimicrobial consumption decreased after 2007 (p < 0.05). Overall mortality was 50.4 %. In multivariate analysis, advanced age and stay in an Intermediate Care Unit (IMCU) were independent risk factors for in-hospital mortality (p < 0.05).
Conclusions: Overall, high incidence of NBSI and considerable resistance of Gram-positive and particularly Gram-negative bacteria were noted in neurosurgical patients. Mortality was high with advanced age and stay in IMCU being the most important death-related factors.