ABSTRACT
OBJECTIVE: This study was designed to identify risk factors for nosocomial infections among infants admitted into eight neonatal intensive care units in Colombia. Knowledge of modifiable risk factors could be used to guide the design of interventions to prevent the problem. STUDY DESIGN: Data were collected prospectively from eight neonatal units. Nosocomial infection was defined as culture-proven infection diagnosed after 72 hours of hospitalization, resulting in treatment with antibiotics for >3 days. Associations were expressed as odds ratios. Logistic regression was used to adjust for potential confounders. RESULTS: From a total of 1504 eligible infants, 80 were treated for 127 episodes of nosocomial infection. Logistic regression analysis identified the combined exposure to postnatal steroids and H2-blockers, and use of oral gastric tubes for enteral nutrition as risk factors significantly associated with nosocomial infection. CONCLUSION: Nosocomial infections in Colombian neonatal intensive care units were associated with modifiable risk factors including use of postnatal steroids and H2-blockers.
Subject(s)
Cross Infection/etiology , Intensive Care Units, Neonatal , Colombia/epidemiology , Cross Infection/epidemiology , Histamine H2 Antagonists/adverse effects , Humans , Infant, Newborn , Intensive Care Units, Neonatal/statistics & numerical data , Intubation/adverse effects , Logistic Models , Prospective Studies , Risk Factors , Steroids/adverse effects , StomachABSTRACT
OBJECTIVE: The epidemiology of nosocomial infections (NI) in neonatal intensive care units in developing countries has been poorly studied. We conducted a prospective study in selected neonatal units in Colombia, SA, to describe the incidence rate, causative organisms, and interinstitutional differences. STUDY DESIGN: Data were collected prospectively from February 20 to August 30, 2001 from eight neonatal units. NI was defined as culture-proven infection diagnosed after 72 h of hospitalization, resulting in treatment with antibiotics for >3 days. Linear regression models were used to describe associations between institutional variables and NI rates. RESULTS: A total of 1504 infants were hospitalized for more than 72 h, and therefore, at risk for NI. Of all, 127 infections were reported among 80 patients (5.3%). The incidence density rate was 6.2 per 1000 patient-days. Bloodstream infections accounted for 78% of NIs. Gram-negative organisms predominated over gram-positive organisms (55 vs 38%) and were prevalent in infants < or =2000 g (54%). The most common pathogens were Staphylococcus epidermidis (26%) and Klebsiella pneumonia (12%). CONCLUSION: Gram-negative organisms predominate in Colombia among infants <2000 g. The emergence of gram-negative organisms and their associated risk factors requires further study.