Abstract 1568

After detection of the first five cases, in the first semester of 1997, a prospective cohort study was conducted, from August/1997 to July/1998, in a 26 beds NICU, to determine the risk factors for ESBLKp colonization or infection. All neonates were followed from admission to discharge. Screening for ESBLKp colonization was performed by culturing rectal specimens on days 3 and 7 after admission and once a week, thereafter. Bacterial isolates were identified by conventional biochemical tests and by the Vitek system: susceptibility tests were performed using the Vitek system; ESBL production was detected by the double disk (DD) test using standard ceftazidime, cefotaxime, ceftriaxone, aztreonam and amoxicillin-clavulonic acid disks. Chromosomal DNA of infection cases was analyzed by pulsed field gel electrophoresis (PFGE) after digestion with Xbal. The statistical tests used included the likelihood-ratio chi-square and Fisher's exact. During the study period a total of 145 newborns (8 cases of infection and 137 colonizations), out of 292 admitted to the NICU, acquired ESBLKp during hospitalization. The DD test was positive for all isolates. Highly related PFGE patterns were detected among the isolates tested, indicating a clonal relationship among them. The risk factors identified were: length of hospital stay > 7 days (RR=3.53; CI 95% 2.84-4.40; p<0,0001), birth weight < 1501g (RR=1.64; CI 95% 1.31-2.06; p=0.001), antibiotic use for more than three days (RR=1.77; CI 95% 1.42-2.20; p < 0.0001); use of central line device (RR=2.67; CI 95% 2.26-3.15; p<0.0001) and mechanical ventilation (RR=2.30; CI 95% 2.0-2.64; p<0.0001). Among children who used antibiotics (n=99), third generation cephalosporins had an association with ESBLKp acquisition (RR=1.89; CI 95% 1.54-2.32; p=0.0014). These risk factors were shown to be strongly associated with ESBLKp colonization or infection.