The clinical records a years cohort of 280 newborn infants consecutively hospitalized for 48 h or more in our neonatal intensive-care unit (NICU) were reviewed. Information on the infants' conditions during the first 12h of life, and on the procedures used in the NICU, were collected. Statistical significance was tested by univariate analysis with the chi(2) test and by multivariate logistic regression analysis with the software program SPSS (Version 10). Over the one-year period reviewed, 90 hospital-acquired infections (HAIs) were contracted; 55 (19.6%) of infants had at least one infection during their stay. The overall in-hospital mortality was 7.1%, and mortality was higher in infants in whom at least one infection developed than in non-infected infants (12.7 vs. 5.8% P=0.13). Very low birthweight infants (VLBW<1,501 g) who had more severe clinical conditions on admission [clinical risk index for babies (CRIB) score >/=5] had an almost two-fold higher risk of contracting a HAI. In the multivariate regression analysis, the onset of a HAI was strongly associated with a low gestational age and the presence of an intravascular catheter. HAIs frequently complicate hospitalization in NICUs and are associated with increased mortality. Our findings also suggest that CRIB could be predictive for the risk of infection in VLBW infants.