BACKGROUND: Necrotizingenterocolitis (NEC) is a major health concern for premature infants and its patho-genesis remains poorly understood. The current mouse NEC model has not well been characterized. OBJECTIVES: In this study, we develop a simple mouse model of NEC and determine the role of several factors modulating human NEC (i.e., breast milk, birth weight, cesarean section and bacteria) on intestinal injury. METHODS: In a first experiment, pups born naturally and dam fed for <12 hours were gavaged with adult commensal bacteria or E. Fecalis, and exposed to hypoxia-cold stress-formula feeding, and compared with controls without bacteria inoculation. 72-hour mortality was recorded, and small intestines were examined histologically. In a second experiment, we compared the incidence of NEC in mice dam fed for <12 hours to those dam fed for 12 to 24 hours or delivered by cesarean section prior to being submitted to the NEC protocol. RESULTS: In pups inoculated with 10(7) CFU of a standardized preparation of adult commensal bacteria or 10(5) CFU of E. Fecalis, the incidence of severe NEC (>grade 2) was 70% and 37% respectively vs 6% in the controls (no bacteria)(p<0.05). In pups dam fed for 12 to 24 hours, NEC incidence was 44(+/-12)% lower vs those dam fed less than 12 hours (p<0.05). We did not find any difference in the NEC incidence between naturally-born pups dam fed for less than 12 hours and these born by cesarean section. The incidence of severe NEC was higher in pups with low birth weight. CONCLUSIONS: we have simplified and characterized a neonatal mouse NEC model that shares several risk factors with human NEC. Now that transgenic mice are available, this model will be useful to study the role played by specific proteins in vivo in NEC development.