Impacto de las bacteriemias nosocomiales en una unidad de cuidados intensivos neonatales

introduction. nosocomial infections have an important impact in newborns morbidity and mortality. an epidemiological surveillance program is the cornerstone to prevent them. the study's objectives were: 1) to develop a surveillance program (high risk nursery component), 2) to establish an epidemiological diagnosis in the neonatal intensive care unit, 3) to analyze bloodstream infections and 4) to evaluate the impact of bloodstream infections on mortality and their risk factors. population, material & methods. since april 1995 to november 2001 we studied prospectively all newborns with nosocomial infections, especially bloodstreams infections, according to nniss criteria. we used an analytic cohort study to evaluate the infection control program and a case-control study to analyze mortality and risk factors. epi6 v6.1 and spss programs were used for statistical analysis. results. no changes in nosocomial infections rates were found in the first four years of study (1995: 14.5‰; 1996: 14.9‰; 1997: 14.5‰; 1998: 12.4‰), with a significant decrease in the last three (1999: 5‰; 2000: 10.2‰; 2001: 10.8‰) (p= 0.00006). we studied 332 episodes, 39% (128/332) were bloodstream infections; 43 (34%) of them were associated with central lines and 85 (76%), without this association. since 1996, central line associated bloodstream infections rates decreased significantly (1995: 15.8‰; 1996: 6.2‰; 1997: 5.8‰; 1998: 2.8‰; 1999: 1.5‰; 2000: 4.3‰; 2001: 4.6‰) (p= 0.011). according with these, non central line-associated infections also showed a significant decrease during this period (1995: 4.5‰; 1996: 3.7‰; 1997: 5.8‰; 1998: 3.4‰; 1999: 1‰; 2000: 1.8‰ and 2001: 1.3‰) (p<0.001). in central line associated bloodstream infections, staphylococcus coagulase-negative and candida spp were prevalent. the most frequent microorganism in infections non-associated with central lines was klebsiella pneumoniae. the mortality rate of bloodstream infections was 21.3%, the attributable risk was 50%. m