Classification and risk-factor analysis of infections in a surgical neonatal unit.
BACKGROUND/PURPOSE: Nosocomial infection may result in significant morbidity in surgical neonates. Traditionally, nosocomial infections are classified using time cut-off points. Gastrointestinal carriage of microorganisms has not been used as a criterion for classifying infection in surgical neonates. The aims of the study were to (1) determine the overall infection rate, (2) distinguish between nosocomial and community acquired infections using a 48-hour postadmission cutoff and the carrier state criterion, and (3) determine risk factors for clinical infection. METHODS: A 1-year prospective observational cohort study was undertaken in a regional neonatal surgical unit between 1997 and 1998. All infants residing for >/=3 days in the unit were included in the study (n = 167). Patient demographics, including illness severity (PRISM score), were recorded for all infants. Surveillance throat and rectal swabs were obtained on admission and twice weekly thereafter to determine carrier status. Carriage was defined as isolation of the same microorganism from at least 2 consecutive surveillance samples. Infective episodes were diagnosed if a clinical diagnosis of local or general inflammation was microbiologically proven. RESULTS: A total of 167 infants responsible for 174 admissions were studied. Median gestational age was 38 weeks (range, 24 to 42), median birth weight was 3 kg (range, 1 to 3.6), median age on admission was 8 days (range, 0 to 142), median length of hospital stay was 8 days (range, 3 to 95). The diagnoses were gastrointestinal disorders (n = 96), abdominal wall defects (n = 22), neural tube defects and hydrocephalus (n = 17), thoracic disorders (n = 16), urologic disorders (n = 12), and abdominal tumours (n = 4). Twenty-five infants had 33 episodes of infection giving an overall infection rate of 14.9%. The predominant infecting organism was Stapylococcus aureus (n = 11); others were enterococcus, coagulase negative staphylococcus, Candida spp, Gram-negative bacilli, and anaerobes. A total of 27 of 33 infective episodes were caused by microorganisms carried by the infants on admission (primary endogenous). Only 6 children had "true" nosocomial infections. Using a traditional 48 hour cutoff, 87% of the infections were classed as nosocomial. Birth weight, presence of central venous line, PRISM score, and length of stay were identified as significant risk factors for developing clinical infection. CONCLUSIONS: (1) Carriage allowed us to identify the true nosocomial infection rate (microorganisms acquired in the unit), which was only 18%. In contrast, using a traditional 48 hour cutoff, 87% of the infections would have been classed as nosocomial and warranted unnecessary cross-infection investigations. (2) The results of this study confirm that birth weight, illness severity (PRISM score), presence of central venous catheter, and length of hospital stay were independent risk factors associated with clinical infection in surgical neonates.