Hemorrhagic bowel necrosis associated with acute digitalis poisoning in an infant.

In infants and children, acute cardiac glycoside (CG) intoxication, resulting from therapeutic or nontherapeutic ingestion is rare but life-threatening. Manifestations of CG poisoning are classically divided into cardiac and extracardiac. Cardiac manifestations include many kinds of arrhythmia and/or conduction defect (1). The extracardiac manifestations comprise electrolyte disturbances, central nervous system (CNS) disorders and gastrointestinal problems. The most common gastrointestinal problems are nausea, vomiting, anorexia, weight loss, failure to thrive and abdominal pain. Although some symptoms are centrally mediated, transient mesenteric ischemia may play a role in several. Although severe mesenteric ischemia associated with CG intoxication is a well-known cause of bowel necrosis in adults (2,3), to our knowledge, this condition has never been reported in children. We report an infant with mesenteric hemorrhagic necrosis following acute digoxin poisoning, whose clinical and radiologic signs mimicked necrotizing enterocolitis (NEC). CASE REPORT A 2 month old female was admitted with a 12 hour history of drowsiness, vomiting and mottled skin. Born at term after an uneventful pregnancy, the child had cystic fibrosis and congenital heart disease, characterized by ventricular septal defect, patent foramen ovale, and patent ductus arteriosus. Daily medications included pancreatic enzymes supplements, antibiotics, and digoxin (0.01 mg/kg/day). The baby was breast fed for 15 days, and switched to formula thereafter. On arrival, the infant was somnolent, pale, and afebrile. On physical examination, she had tachypnea (84 breath/minute) and mild chest retractions. Cardiac evaluation showed tachycardia (144 beats/minute) and a pansystolic grade III murmur. The abdomen was distended but nontender. The remainder of the examination was unremarkable. Serum electrolytes revealed a potassium level of 5.7 mEq/L. An electrocardiogram (ECG) demonstrated a complete atrioventricular dissociation, with a good ventricular rhythm. On further history, it was discovered that, due to a misunderstanding of the directions, parents had been giving the child tenfold (0.1 mg/kg/day) the prescribed dosage of digoxin, for the previous week. Digoxin intoxication was confirmed by high serum digoxin level (4.5 ng/mL). Digoxin-specific Fab antibody fragments (Digibind®, Glaxo Wellcome, Inc., Research Triangle Park, NC) were promptly administered intravenously. Serum potassium levels returned within normal limits (3.4 mEq/L), and a repeat ECG, performed shortly thereafter, demonstrated resolution of the dysrhythmia. Nevertheless, the baby became lethargic, dehydrated and respiratory distress worsened. The abdomen became more distended and tympanic with no bowel sounds on auscultation. Abdominal x-ray showed signs suggesting NEC (Fig. 1).FIG. 1.: Plain abdominal x-ray. Bowel loops are dilated with linear (arrow heads) and bubbly air collections in intestinal wall. Note branching appearance of portal vein gas (arrows).A nasogastric tube was passed which drained conspicuous milky material. In a few hours, respiratory distress developed requiring intubation and mechanical ventilation. The patient became anuric. Continuous infusion of dopamine (5 μg/kg/min) and intravenous antibiotics were given, and total parenteral nutrition was started. Abnormal laboratory tests included: white blood cells 2,300/mm3, blood urea nitrogen 68 mg/dl, creatinine 1.6 mg/dL, aspartate aminotransferase (AST) 85 UI/L, alanine aminotransferase (ALT) 83 U/L, amylase 462 UI/L. Blood gas analysis showed pH 7.26, bicarbonates 19.7 mEq/L, PaCO2 43 mmHg, bases excess –6.3. The full septic screen was normal. Resuscitation measures restored blood pressure and urine output. On day 2 after admission, a repeat abdominal x-ray suggested bowel perforation, and the baby was sent to surgery. At laparotomy, mild bloodstained peritoneal fluid and generalized small bowel ischemia, without evidence of perforation were found. Intraoperative intestinal decompression was performed, draining bloody alimentary contents. The postoperative course was stormy, with persistence of hepatic, renal, and pancreatic disorders for about one week. There was a partial wound dehiscence with evisceration and extensive bowel necrosis. Conservative treatment was initiated with daily curettage and toilet of the open wound at the bedside; the necrotic small bowel was removed and specimens sent for histology (Fig. 2).FIG. 2.: Histological examination of resected small bowel. On left side, mucosa is hemorrhagic and necrotic. Both submucosal arteries and veins (*) (**) are engorged, with fibrinoid thrombi in the latter, partially occluding the vascular lumen. On right side note disarrangement of muscular fibers with edema of muscularis. Chronic inflammatory cells invade mucosa and submucosa. Inset: note edema and massive necrosis of intestinal villi. (Hematoxylin & eosin stain; ×20)Following stabilization, the patient underwent a second-look operation. The large bowel was healthy, whereas only 30 cm of small bowel appeared viable. On postoperative day 1, the patient developed severe bradycardia (heart rate 40 beats/min), respiratory acidosis (pH 7.13, PaCO2 75,5 mmHg) regardless of mechanical ventilation, and oliguria. Laboratory tests showed Hb 3.9 g/dL, total bilirubin 4.92 mg/dL, AST 2730 UI/L, ALT 950 UI/L, gamma-glutamyltransferase 97 UI/L, C reactive protein 2.27 mg/dL. Blood derivatives and continuous infusion of inotropic agents were attempted but the clinical conditions progressively worsened, and the infant died on postoperative day 2. The parents did not consent to postmortem examination. DISCUSSION We present an infant with acute digoxin intoxication complicated by hemorrhagic bowel necrosis. The relationship between plasma CG concentration and clinical manifestations of toxicity is not clearly established in children. Based on a review of collected series (1), the accepted lower limit for toxic symptoms is 2 ng/mL for children, and 3.5 ng/mL for neonates and small infants. However, in a retrospective review of cases of pediatric CG intoxication at three academic institutions (4), symptoms of toxicity have been reported with a serum digoxin level as low as 0.2 ng/mL. Toxic effects are more likely to occur in neonates and small infants than in older children probably due to the higher blood and tissue digoxin concentrations and reduced renal clearance of CG in younger infants (1). In our case, the course was complicated by massive hemorrhagic bowel necrosis. Mesenteric ischemia is considered the cause of abdominal pain in children with acute CG intoxication (1). However, to our knowledge, no previous cases of hemorrhagic bowel necrosis associated with acute CG intoxication have been documented during childhood. Only one case of pediatric CG poisoning associated with severe bowel involvement has been described in a 3 day old boy who died of NEC, acute renal failure and progressive sepsis nine days after the administration of digoxin-specific Fab antibody fragments (5). No histology of involved bowel was provided. Our patient also had clinical and radiologic signs suggesting NEC (Fig. 1). However, final histologic examination of resected bowel specimens was more suggestive of hemorrhagic bowel necrosis, secondary to nonocclusive mesenteric ischemia (NOMI) (Fig. 2). This clinical entity has been previously reported in infants operated on for hypoplastic left heart syndrome (6). Due to strikingly overlapping clinical and radiologic findings, NOMI can be erroneously identified as NEC (6). Some demographic, clinical, and pathologic features may distinguish NEC from NOMI. NEC is a disease typical of prematurity which develops shortly after oral feeding, usually infant formula, is introduced (7), whereas NOMI occurs mainly in older infants and there is no clear relationship to the feeding pattern (6). Probably, NEC and NOMI share circulatory ischemia of gastrointestinal tract as an important pathogenic factor. However, the primary underlying disease in NOMI results in significant hypoxemia and circulatory failure, which is clearly associated with a much more diffuse systemic ischemic process. Conversely, in NEC the ischemic process is less important, the initial bacterial colonization of the newborn intestine playing a pivotal role (7). Therefore, NOMI has no specific pattern of distribution and diffuse intestinal disease and involvement of other solid organs including liver, spleen, kidney, and pancreas is common (6). Conversely, NEC is predominantly located in the distal ileum and colon and diffuse intestinal disease is rare. At microscopy, NEC presents coagulation necrosis as the dominant feature with signs of acute inflammation in about 80% of patients (8), while NOMI displays exclusively hemorrhagic necrosis without coagulation (2,3,6). Our patient was born at term, breast-fed, she presented with signs of involvement of multiple organs, and the disease was diffuse with pathologic findings of exclusively hemorrhagic necrosis. Present findings suggest NOMI rather than NEC. The relationship between acute CG intoxication and NOMI is well established in adults (2,3). The pathophysiological pathway is probably the result both of a direct constrictor effect on the vascular smooth muscle and an indirect effect through an increase in alpha-adrenergic tone (9). The intensity and location of the vasoconstrictor effect of CG on the mesenteric arteries is probably unpredictable, ranging from simple abdominal pain (1) to massive hemorrhagic bowel necrosis as reported in adults (2,3) and in our patient. In patients with acute CG intoxication complicated by NOMI, bowel ischemia has been reported to improve following administration of digoxin-specific Fab antibody fragments (10). In our patient, the small bowel damage did not recover despite administration of this drug, and evolved to massive hemorrhagic necrosis. Intestinal hemorrhagic necrosis may be followed by distant tissue damage, resulting in a multiple organ dysfunction syndrome (MODS) (6). Following intestinal damage, our patient presented clinical and laboratory findings consistent with MODS according to Wilkinson's criteria for pediatric patients (11). After the second operation, a more severe MODS led to death, which was compatible with the “two-hit” model for MODS proposed by Meakins (12). According to this model, once the host inflammatory response has been primed, a second insult can reactivate the process in a more amplified form (12). Therefore, in childhood, any acute CG intoxication associated with an abdominal complaint should be assessed to rule out NOMI. If NOMI is suspected, immediate aggressive resuscitation measures must be commenced to prevent further progression of bowel ischemia and possibly MODS. General treatment guidelines (6) include broad-spectrum antibiotic therapy, bowel rest, and total parenteral nutrition. Serial x-rays of the abdomen should be obtained to monitor the evolution of bowel conditions and detect any early sign of bowel perforation. Surgery should be reserved for cases with evidence of bowel necrosis or perforation and refractory to medical treatment. In conclusion, the present report demonstrates that NOMI, an uncommon, but often catastrophic, complication in acute CG intoxication, may occur even in infants. In pediatric acute CG intoxication, abdominal pain should not be overlooked. Physicians should be aware that abdominal symptoms may be the prodromic signs of an evolving intestinal ischemia, which may progress towards MODS.