Nosocomial infection in a neonatal intensive care unit and its prevention with selective intestinal decolonization. A multivariant evaluation of infection reduction.

A prospective cohort study in a neonatal intensive care unit (ICU) was carried out to evaluate whether the incidence of infection in neonates receiving intestinal decolonization was reduced in comparison to those who did not. This study was performed after controling possible confounding infection risk factors. A total of 536 babies were screened in our ICU during the 27-month study period. Neonates were admitted to the ICU for different reasons: low weight, respiratory distress syndrome, acute fetal suffering, surgery, etc. The doctor in charge decided whether the baby should be decolonized or not, so this experimental study was non-random. Thus more of the babies with a greater risk of infection were decolonized more often than the other babies who were not so much at risk. In this study, babies were classified by type of decolonization given: a well-performed Selective Intestinal Decolonization (SID) was done (early and with three oral drugs: E polymyxin, tobramycin and nystatin): 10.8% of the babies; Incorrect SID (was begun late and/or less than three drugs were used): 16.7% of the babies; and Without SID (72.9%). Total nosocomial infection (NI) was 11.2%, catheter-associated sepsis was 42% of the total NI. When the NI incidence was directly compared among groups, it was lower in the group without SID, but infants with decolonization initially had more infection risk factor than the first group. For this reason, multiple logistic regression was used in order to stratify factors by infection probability, and correcting the existing bias.(ABSTRACT TRUNCATED AT 250 WORDS)