Neonatal gut injury and infection rate: impact of surgical debridement on outcome

Infectious burden of gut injury (G-INJ) associated with necrotizing enterocolitis (NEC) or with spontaneous intestinal perforation (SIP) in neonates has not been ascertained. We sought to test the hypotheses that: (1) infants with G-INJ develop higher number of infections including non-concurrent infections than infants without G-INJ in a neonatal intensive care unit (NICU); (2) surgical debridement (DEB) of infants with severe G-INJ is associated with lower infectious morbidity and mortality. All infants admitted to the regional NICU from October 1991 to February 2003 were included in this prospective prevalence investigation of G-INJ and infections. Non-viable (<23 week gestational age) infants, infants with congenital anomalies, and those who developed NEC after SIP were excluded. Standard definitions of National Centers for Disease Control and Prevention were used for different categories of infections. Episodes of infections were classified as concurrent or non-concurrent (post G-INJ) based upon their timing in association with G-INJ. Infants with G-INJ associated with Bell stage II or higher NEC or with SIP were further stratified by DEB into two subgroups. A previously described 7-point clinical score was used to divide G-INJ into mild (0–2), moderate (3–5), and severe (6–7) categories. Surgical outcomes were determined by using χ² and logistic regression analyses. Data are expressed as mean ± SD or as odds ratio (OR) with 95% confidence intervals (CI); P<0.05 was considered significant. Of all 5,481 infants, 954 (17.4%) developed 1,734 episodes of infections. Prevalence of G-INJ was 4% (n=222); of these, 33% (n=73) underwent DEB. Infants with G-INJ had lower mean birth weight (1,414±766 vs. 2,153±104 g; P<0.0001) and lower mean gestational age (29.6±4.2 vs. 32.9±4.8 weeks; P<0.0001) than their peers (n=5,259). Controlling for birth weight and gestational age, odds for non-concurrent blood stream infections (BSIs) in G-INJ infants were higher (OR 13.98, CI 10.289–19.01, P<0.0001) than the remaining population without G-INJ. Forty-four percent of all episodes of fungemia, 32% of all episodes of BSIs occurred in G-INJ infants (P<0.0001). Within the G-INJ group, there were no demographic differences between the DEB and non-DEB infants. Controlling for severity of G-INJ, odds for non-concurrent BSIs (OR 3.45, CI 1.04–11.36, P<0.05) and for mortality (OR 3.35, CI 1–10, P<0.05) among non-DEB infants were higher than in DEB infants. Infants with G-INJ suffered from a disproportionate number of all blood-stream infections in our intensive care nursery. Infants with severe G-INJ whose management includes DEB are more likely to survive and to incur less infectious morbidity.