Abstract 1489 Poster Session III, Monday, 5/3 (poster 41)

Background: Published studies continue to emphasize that NEC is a persistent cause of NICU morbidity and mortality. Its causes remain unknown. Rates of NEC for very premature infants differ among NICUs. We thought that comparing NEC rates among NICUs might elucidate risk factors for NEC.

Methods: 1482 first admissions with gestational age (GA)<32 weeks to 6 perinatal centers in New England (10/94-6/96) were prospectively abstracted for NEC, medical and demographic risk factors, and items from the admission Score for Acute Neonatal Physiology (SNAP). Diagnosis of NEC required pneumatosis. Deaths within the first 48 hours of life were excluded. We analyzed 82 cases of NEC. Factors with univariate association were entered in stages (prenatal, perinatal, and neonatal risk factors) into logistic regressions predicting medical, surgical or any NEC. NICUs were represented by dummy variables with the lowest incidence site (A) as reference.

Results: Overall NEC incidence was 5.2%. We found that most known and suspected risk factors for NEC were insignificant: patent ductus arteriosus, platelet counts <100K, hemotocrit>60, Apgar <5, mean blood pressure<25, pH<7.20, SNAP score >20, male, small for gestational age, antenatal steroids, steroids for bronchopulmonary dysplasia, umbilical arterial catheters and enteral feeds on day 3. The only predictive factors for NEC were GA, and narcotic use on day 3 (OR 7, CI 2-18) which was specific for surgical NEC only. After control for all risk factors, odds ratio for NEC still varied greater than 5 fold among sites.

Conclusions: NEC may be less frequent than previously reported. Conventional markers for NEC have little explanatory power except for GA. We found higher rates of surgical NEC in patients on narcotics on day of life 3. The absence of conventional predictors suggests important, unmeasured biologic events (i.e. epidemic infections) that lead to NEC. Reasons for high NEC rates at sites E and F are unknown. A review of practices at sites A, E and F might illuminate some new risk factors. (Table)

Table 1 No caption available
Full size table