The etiology of NEC is multifactorial. Outcome depends on the severity of bowel disease and perforation or gangrene are difficult to predict. The systemic inflammatory response is probably implicated in the pathophysiology of this entity. The increase in circulating mediators might affect bone marrow resulting in altered number of released immature inflammatory cells. It is reasonable to speculate that the more severe cases of NEC will probably have the greater inflammatory response and bone marrow alteration. The present study was designed to evaluate the relationship between laboratory evidence of inflammation and severity of NEC, including perforation. All 25 patients admitted to the NICU at the University hospital from 1/1/1995 through 30/6/96 who had a diagnosis of NEC were included. The following indicators of inflammation were evaluated daily in the days prior to and for 7 days after the diagnosis of NEC: WBC, Immature/mature ratio (IM/M) and eosinophil and platelet counts. Mean birth weight was 1200 g (GA: 27-36 wks.). Patients were classified into Medical Management (MM) or Surgical Treatment (ST) according to treatment requirements. Twenty one patients needed only MM; 4 required ST for perforation. Clinical signs, known risk factors and age at diagnosis(<7 days) were similar in both groups. Blood cultures were negative in all patients. In the ST group, 3 of the neonates had never been fed, all had platelet counts of < 50.000, neutropenia with altered IM/M ratio and increased eosinophil counts (> 10%) 48 hs prior to perforation; 1 patient died. In the MM group, only 1 patient had neutropenia; none had altered IM/M ratios, decreased platelet counts or increased eosinophils. These data show that markers for inflammation are significantly altered 48 hs before perforation. We conclude that in neonates with NEC the presence of more than one of the following: neutropenia, low platelet counts and increased eosinophils may be helpful in predicting which infants will eventually require surgery.