David A Clark

Necrotizing Enterocolitis: Intraluminal Biochemistry in Human Neonates and a Rabbit Model

ABSTRACT: The intestinal contents of 17 neonates with necrotizing enterocolitis were analyzed for pH, carbohydrate, protein, and bacteria. The intraluminal pH was <5.0 (16/17). Sufficient carbohydrate and bacteria capable of fermenting the carbohydrate to organic acids were found. The intraluminal protein content was >5 g/dl. The variables of acid and protein were then examined in a rabbit intestinal loop model. The hemorrhagic response in individual loops was measured using Cr51 tagged red blood cells such that the microliters of blood per centimeter intestine could be determined. Loops with organic acid and protein had significantly (p < 0.01) more intramural blood than control loops. Organic acid (possibly generated by bacterial mixed acid fermentation of carbohydrate) in the presence of protein promotes intramural hemorrhage similar to that seen in neonates with necrotizing enterocolitis.

Continuous gastric pH measurement in young and older healthy preterm infants receiving formula and clear liquid feedings.

Gastric pH was recorded with an intragastric pH electrode for 12 h in two groups of healthy, preterm infants with similar birth weights (range 1.4 to 2.0 kg). Group I infants (n = 13) were less than 7 days old and Group II infants (n = 10) were 7-15 days old. Infants were fed three formula feedings and one clear liquid feeding during the study. In Group I, mean gastric pH measured at 15-min intervals was above 4.0 for 3 h after either feeding. In Group II mean gastric pH was lower particularly after clear liquid feedings, where it remained below pH 4.0 for the entire 3-h postprandial period. The percent of monitored time at gastric pH less than 4.0 was low in Group I--15.2 +/- 4.2% and 20.6 +/- 6.4% after formula and clear liquid, respectively. The percent time was greater in Group II--42.7 +/- 8.0% and 61.9 +/- 7.3% after formula and clear liquid, respectively. In the younger preterm infant, gastric pH does not appear sufficiently low to support peptic activity.

Differences in Thromboxane Production between Neonatal and Adult Platelets in Response to Arachidonic Acid and Epinephrine

Summary: In this study, we have investigated the possible role of the proaggregatory arachidonic acid (AA) metabolite thromboxane, in the impaired function of neonatal platelets. In platelet-rich plasma thromboxane production (measured by radioimmunoassay of thromboxane B2) was not different between neonates and adults when stimulated by thrombin (at 0.1 or 1.0 U/ml) or collagen (70 μ g/ml) although neonatal platelets produced decreased thromboxane (TBX2) postepinephrine stimulation. In response to 1 U/ml thrombin, adult and neonatal platelet-rich plasmas produced mean values of 3.41 ± 0.35 (SEM) and 3.11 ± 0.49 pmol of TXB2/106 platelets, respectively. Production of TXB2 in response to 0.1 U/ml thrombin was not dissimilar between neonates (1.01 ± 0.46 pmol) and adults (1.04 ± 0.38 pmol). When collagen was used as the aggregating agent, TXB2/ production was also not significantly different with values of 2.44 ± 0.48 and 1.90 ± 0.46 pmol/106 platelets produced by adult and neonatal platelet-rich plasma, respectively. In response to 200 μ M epinephrine, adult platelets produced 1.03 ± 0.39 pmol TXB2/106 platelets while neonatal platelet TXB2 production was significantly decreased (0.15 ± 0.04; P < 0.05). Thromboxane production in response to AA, however, was markedly elevated in neonatal platelet-rich plasma. When 200 and 400 μ M concentrations of AA were used as the aggregating stimuli, neonatal platelet rich plasma produced 3.17 ± 0.77 and 8.0 ± 1.47 pmol TXB2/106 platelets, respectively. These values were significantly elevated P < 0.02 and < 0.005) when compared to mean values of 0.41 ± 0.10 and 3.32 ± 0.15 pmol in adult platelet-rich plasma. This elevated thromboxane production was not, however, inherent in neonatal platelets since when washed platelets were studied, results were reversed. Adult platelets produced more thromboxane at all doses of AA evaluated. These results suggest that the elevated response to exogenous AA observed in neonatal platelet-rich plasma results from as yet undetermined plasma factors. The reported deficiencies in platelet function in the newborn clearly do not result from deficient thromboxane production poststimulation with the physiologic aggregating agents collagen and thrombin. Moreover, our study introduces a new and possibly important difference between adult and neonatal plasma, namely, the differential response to exogenous arachidonic acid.

Bupivacaine alters red blood cell properties: a possible explanation for neonatal jaundice associated with maternal anesthesia.

Cord blood was incubated with lidocaine, mepivacaine, bupivacaine, or buffer and red blood cell filterability was determined. Only bupivacaine at either 1 or 2 micrograms/ml prolonged filterability by an average of 58 to 65% over red cells treated with buffer alone. Tritiated bupivacaine was bound to a greater extent to red cell ghosts from cord blood (24.6 +/- 5.8%) than to adult red cell ghosts (14.6 +/- 2.6%). Finally, we determined red cell survival in 13-day-old rats injected with bupivacaine or buffer. At 2 h after injection, buffer-treated animals had a red cell survival of 96.9 +/- 3.3%, whereas 2-h survival was reduced to 82.6 +/- 8.7% for the animals injected with bupivacaine. Our results suggest that the neonatal jaundice associated with maternal anesthesia, especially bupivacaine, may be related to the observations that these agents cross the placenta, bind to the red cell membrane and reduce its filterability, resulting in shortened red cell survival.

Interaction between Bovine Casein and V. Cholerae Enterotoxin in the Rabbit Ileal Loop

Summary: Secretory IgA from human breast milk neutralizes cholera enterotoxin in the rabbit ileal loop system. No similar protection by purified bovine milk proteins could be demonstrated; however, one bovine milk protein, casein, had a deleterious effect on intestine exposed to very small quantities of enterotoxin. Highly purified cholera toxin (10 or 100 ng) was incubated with bovine protein solutions for 60 min at 37°C. One-milliliter aliquots were then injected into prepared rabbit intestine loops. The animals were sacrificed at 18 h and the intestinal loop contents were aspirated, and a volume to length of loop ratio (V/L) was determined. The activity of 100 ng of toxin was not enhanced by the majority of bovine milk proteins, but bovine casein caused a 14-40% increase in the fluid production (V/L of casein + toxin versus toxin, 1.05 versus 0.92 and 1.82 versus 1.30). All of the bovine proteins but casein inhibited the action of low dose enterotoxin. Bovinec asein caused a 78-90% increase in fluid production by loops exposed to a suboptimal toxin dose (10 ng) (V/L of casein + toxin versus toxin, 1.12 versus 0.63 and 0.95 versus 0.50). Virtually all of this enhancement of enterotoxin fluid response resided in the purified alpha-casein fraction.

Hydrops fetalis attributable to intrauterine disseminated intravascular coagulation.

Correspondence to: David A. Clark, M.D., Assistant Professor of Pediatrics, SUNY, Upstate Medical Center, 750 East Adams Street, Syracuse, NY 13210. Received for publication March, 1980; revised July, 1980 and accepted August, 1980. &dquo; r-% P RESUMED intrauterine disseminated intrar vascular coagulation (DIC) has been reported in association with multiple pregnancy with a dead twin fetus.’-’ None of these previous descriptions has included a liveborn infant presenting with hydrops fetalis.

NECROTIZING ENTEROCOLITIS |[lpar]|NEC|[rpar]|, A RADIOGRAPHIC AND PATHOLOGIC MODEL

NEC in neonates is suspected by abdominal distension, feeding intolerance and Guaiac + stools but is first clinically confirmed by abnormal findings on radiographs of the abdomen. Using our previously described rabbit intestinal loop model for NEC, we prepared 30 cm loops in 12 weanling rabbits into which a suspension of 3% lactose, 1% purified bovine casein and a Gram negative bacteria capable of the mixed acid fermentation were injected. The blood supply was carefully preserved. Radiographs of the abdomen were taken 6 or 18 hrs. later. Following the radiographs the rabbits were opened to examine the intestines.Radiographic findings included free abdominal air, portal air and pneumatosls intestinalis, with at least one finding in each rabbit. On gross examination distension, mucosal thinning and hemorrhage were commonly seen in the six hour experiment whereas the intestine in the rabbits sacrificed at 18 hours were more friable and necrotic. Microscopy revealed mucosal disruption with villus destruction and a fibinopuruient exudate. We conclude that the radiographic and pathologic findings in babies with NEC can be duplicated in a rabbit intestinal loop model. As bacteria ferment the carbohydrate, organic acid is produced which in the presence of a protein disrupts the intestinal mucosa. The gas generated in this reaction may dissect into the wall of the intestine. (pneumatosis intestinalis),may enter the blood stream(portal air),or may escape through a perforation to become free abdominal gas.

410 THE INTERACTION OF BOVINE CASEIN AND ENTEROTOXIN

Secretory IgA in human breast milk can neutralize cholera enterotoxin in the rabbit ileal loop system. No similar protection by separate milk proteins could be demonstrated but one milk protein, casein, had a deleterious effect on intestine exposed to very small quantities of enterotoxin.Rabbit intestinal loops were prepared by Sacks modification of the method of Kasai and Burrows. Highly purified V. Cholerae toxin (10 or 100 ng) was incubated with 10 mg of bovine protein solutions (lactalbumin, casein, lactoglobulin A, lactoglobulin B, gamma globulin) for 60 min. at 37° C. One ml aliquots were then injected into the rabbit ileal loops. Eighteen hours later the animals were sacrificed. The intestinal loop contents were aspirated and a volume/length of loop ratio was determined.The activity of 100 ng of toxin was not affected by the majority of bovine milk proteins, but bovine casein caused a 40% increase in the fluid production (p<.001). When the milk proteins were added to a low dose of enterotoxin (10 ng), all but casein nonspecifically inhibited the action of toxin. Bovine casein caused a 78% increase in fluid production by loops exposed to suboptimal toxin dose (10 ng). The exaggerated intestinal fluid response to low dose toxin and casein may result from mucosal damage similar to that caused by the higher dose of toxin alone.

PROTECTION AGAINST INTESTINAL REPERFUSION INJURY BY DEXAMETHASONE (DM)

Studies have implicated oxygen radicals in the etiology of necrotizing enterocolitis (NEC). DM, a membrane stabilizer, may protect against oxidant injury of cells. We assessed the role of DM in gut ischemia and reperfusion, which produces experimental NEC. Laparotomy under anesthesia was performed in 16 rabbits. In each, 4 intestinal loops, 5 cm long, were prepared. Only 2 of these loops were rendered ischemic for 5 min., the other 2 loops serving as non-ischemic controls. 8 rabbits were pre-treated with IV DM (0.1 mg/kg); in the other 8, Intraluminal (IL) DM (1 mg) was injected into 2 loops while the other 2 were injected with saline (S). The animals were sacrificed 4 hrs. after surgery. All ileal loops were fixed and examined histologically by a single pathologist unaware of group assignment. Histological changes were graded.The non-ischemic loops were normal. DM (IV & IL) was fully protective against reperfusion injury, with two exceptions. These dramatic results suggest that DM may prevent experimental intestinal necrosis.

EFFECT OF HISTAMINE (H) RECEPTOR BLOCKADE ON REPERFUSION INJURY IN THE RABBIT INTESTINE

Mast cells are abundant in the intestine and release H, an autocoid which has been implicated in the etiology of necrotizing intestinal disorders. We investigated the effect of histamine receptor blockers on reperfusion injury in the rabbit intestine. 25 rabbits underwent laparotomy under anesthesia. 8 were pre-treated with IV Cimetidine (C), an H2 blocker, (25 mg/kg), 9 with IV diphenhydramine (D), an H2 blocker, (5 mg/kg), and 8 received no drug. In each, 4 intestinal loops were prepared: 2 loops were rendered ischemic for 5 mins., the other 2 loops serving as nonischemic controls. The animals were sacrificed 4 hrs. after surgery. All Intestinal loops were fixed and examined histologically by a single pathologist unaware of group assignment. Histological changes were graded.Irrespective of ischemia, the intestinal loops in rabbits pre-treated with C and D demonstrated mucosal necrosis when compared to untreated animals (p<0.001, chi-square). These data suggest that blockade of both H1 and H2 receptor sites potentiated mucosal injury.