NS, normal salineThe etiology of necrotizing enterocolitis has remained elusive despite numerous theories. It occurs primarily in infants born prior to 34 wk gestation (1, 2). These infants have intestinal ileus, abdominal distention, and bloody stools. Roentgenographic examination of the abdomen reveals distended loops of bowel and in the classic presentation, pneumatosis intestinalis (1, 3). Often the clinical course is fulminant with hemorrhagic necrosis of the bowel, metabolic acidosis, respiratory insufficiency, and sepsis progressing to death (4).While numerous epidemiologic associations have been suggested, only a few consistent underlying factors have been identified. These include prematurity (1, 5 ) , enteral feeding (6, 7), and enteric bacteria (3,8). Reports have also noted a clustering of cases suggesting an infectious agent, but investigations to define a causative bacteria or virus have as yet been inconclusive (6, 9). Numerous investigators have implicated compromised intestinal vascular supply resulting from birth asphyxia, respiratory distress, hypotension, and the use of umbilical arterial catheters. Unfortunately, recent studies have shown that many of these epidemiologic associations cannot be confirmed with careful analysis ( 10).Although prevailing theories suggest a primary vascular insult, many preterm infants seem to have an abrupt onset of necrotizing enterocolitis with no obvious predisposing factors. We ex- amined the intestinal contents of preterm infants with severe necrotizing enterocolitis. The low intraluminal pH and high protein content in the intestines of these neonates led us to examine the variables of pH and protein in a rabbit intestinal loop model.
MATERIALS A N D METHODSNronute.~. We obtained the intestinal contents at or adjacent to the site of perforation or necrosis in 17 infants requiring surgical removal of the intestine for advanced necrotizing enterocolitis. The intestinal contents were analyzed for blood, pH, carbohydrate, protein, and bacteria. The presence of blood was confirmed by the guaiac test. The presence of carbohydra1.e available for bacterial fermentation was initially determined by reducing substances and then quantitated by measurement of glucose after incubation with lactase (Nutritional Biochemicals). The protein content was determined by the Lowry technique and the primary protein present was estimated by acrylamide gel electrophoresis (1 1). The contents were cultured for aerobic and anaerobic organisms by routine bacteriologic techniques. The predominant lactose fermenting aerobic organism was isolated on MacConkey media.Rubhit model. A modification of the rabbit intestinal loop model of Kasai and Burrows provided a means for examining the variables of intraluminal pH and protein (12, 13). Sixteen New Zealand white weanling, 3-wk-old-male rabbits were fasted for 8 h. Blood from a cross-match compatible donor rabbit was drawn by syringe and the red blood cells were labeled with Chromiums1. A I-ml aliquot of blood was then injected into each rabb...
