Background: In retrospective studies, outbreaks of NEC have often been related to specific pathogens such as E coli and K pneumonia. We performed a prospective study to investigate whether there is an association between NEC and specific pathogens. Methods: Between 4/93 and 3/97, we screened stools of all neonates of <36 wk admitted to our neonatal unit for bacteria in weekly intervals. Clinical and bacteriological data from each infant who developed NEC were compared to those from 2 control infants matched for gestational age (GA) and date of admission .Results: 19 infants developed definite NEC; 9 of these underwent laparotomy; 2 died. Mean GA at birth was 29.7 wk (SD 2.3) in cases and 29.8 wk(2.4) in controls. The only significant difference in the clinical course prior to NEC was a more severe stage of respiratory distress in the cases(Median 2 (0-4) vs. 0 (0-4), p<0.01). Within the last week prior to NEC, bacteria were identified in stools of all cases and 81% of controls (n.s.). There was no significant difference in the occurrence of specific pathogens or groups of pathogens in cases compared with controls (e.g. gram-neg. spp., 32% of cases, 32% of controls), despite the fact that 12 of 19 NEC occurred in 3 clusters of 3-14 wk each. Conclusion: Although we also observed a typical clustering of NEC cases, there were no specific bacteria associated with these outbreaks when data from cases were compared with those from matched controls.