Pretransplant Management and Small Bowel-Liver Transplantation in an Infant with Microvillus Inclusion Disease

Abstract

Microvillus inclusion disease (MID), or microvillus atrophy, is a rare, generalized enteropathy first described in 1978 (1), that affects both small and large bowel. Apart from some late-onset variants of MID, most affected children have manifestations of abundant, intractable, watery diarrhea from the first days of life, which persists at stable rates(50-300 ml/kg per day) despite the child's receiving nothing by mouth(2). Very severe malabsorption and exacerbation of diarrhea by oral feeding (2,3) enforce total parenteral nutrition. Despite intravenous replacement of enteric losses of fluids and electrolytes, there is a high mortality rate in the first year of life. Metabolic acidosis and electrolyte imbalances are contributing factors, as well as sepsis and liver failure related to early cholestatic liver disease. Recently, two children with MID were reported (4,5) to have undergone successful intestinal transplantation(including colectomy of the disease-affected large bowel). In this article, we present the first child with MID who underwent successful combined liver and small bowel transplantation without removal of the major part of the recipient's large bowel. CASE REPORT This male infant was born at term as the eighth child of healthy, consanguineous Arab parents (first cousins) after an uneventful pregnancy with no history of polyhydramnios and an uncomplicated delivery. He weighed 3900 g at birth and was initially breast fed. Three siblings, one boy and two girls, had died after a history of vomiting and diarrhea at the age of 15 days, 1 month, and 12 months, respectively. In the third child, the diagnosis of MID was established from typical electron microscopic findings(6) in a duodenal biopsy specimen. In the current patient's second week of life the onset of urine-like diarrhea with additional mucus was recognized. An initial attempt at enteral feeding with a peptide-containing commercial formula led to marked steatorrhea and severe metabolic acidosis. Therefore, total parenteral nutrition was initiated. Even without receiving food, the patient's daily stool volume ranged between 130 and 225 ml/kg per day. The fecal electrolyte composition revealed the typical pattern of secretory diarrhea (Na+, 105 mmol/l; Cl-, 74 mmol/l; K+, 12.1 mmol/l; stool osmolarity, 281 mOsm/kg; fecal osmotic gap, 47 mmol/l; and stool pH, 8). Examination of peroral jejunal biopsy specimens revealed severe villous atrophy and minor crypt hyperplasia. Periodic acid-Schiff (PAS) staining showed reduction of the usual brush border with positive staining material present within the apical cytoplasm of enterocytes. The PAS abnormality extended from the upper crypt region onto the surface epithelium. Transmission electron microscopy showed the surface epithelium with variably shortened microvilli, increased lysosomal bodies, and increased secretory granules, similar to that typically seen in MID (2). Microvillus inclusions (1-3,6) were seen in some cells but were not widespread. All disaccharidase enzyme activities were abnormally low. Parenterally administered fluids were increased to keep the sodium:fluid ratio (50 mmol/l) constant to an amount of 270 ml/kg per day to achieve a urine production of at least 4 ml/kg per hour. Under these conditions, the kidneys compensated for enteral loss of bicarbonate, and the patient did not have further episodes of metabolic acidosis, although no bases were substituted. Renal sodium excretion normalized (2 mmol/kg per day) from very low initial levels (0.02 mmol/kg per day), and plasma renin activity stayed in the normal range. From the age of 3 to 9 months, the patient thrived, going from the 3rd to the 50th weight percentile, with an average daily weight gain of 25 g. He demonstrated a normal growth rate, with his body length in the 10th percentile, in keeping with the height of his parents. Nevertheless, at the age of 9 months the decision was made to put him on the waiting list for combined liver and small bowel transplantation, and he was transferred to the University of Nebraska Medical Center. Liver transplantation was considered because the patient's liver function had steadily deteriorated. Cholelithiasis was demonstrated by ultrasound from 4 months of age. Since he was 7 months old, his spleen had been palpable just below the costal margin, indicating some portal hypertension. His serum bilirubin had climbed to 6 mg/dl at 9 months and to 8.5 mg/dl (conjugated bilirubin, 6.8 mg/dl) at 12 months of age. A specimen obtained in liver biopsy at 9 months of age showed evidence of total parenteral nutrition-induced cholestasis and mild portal fibrosis, with early bridging. At 11 months of age, he had recurrent ascites, along with very low serum albumin levels. At the time of transplantation (12 months of age), histologic analysis of the excised liver clearly showed cirrhosis, bridging fibrosis, cholestasis, and features consistent with chronic total parenteral nutrition injury. His gallbladder showed cholelithiasis and signs of chronic cholecystitis. The general condition of the severely icteric patient was stable, but liver protein synthesis had further deteriorated (prealbumin, 17.3 mg/dl-normal range, 19-47 mg/dl; prothrombin time 15.2 sec-normal range, 11-13.8 sec) with compensatory increased immunoglobulin (IgG 1718 mg/dl; normal range, 453-916 mg/dl). After the performance of a percutaneous liver biopsy at 9 months of age, the patient had severe intraperitoneal bleed requiring surgery. An increased risk of bleeding was not suspected at that time because his platelet count(130 × 103/µl), prothrombin time (13.1 sec) and partial thromboplastin time (25.4 sec) were within normal ranges. Life-threatening hemorrhagic shock developed, with respiratory failure, ischemic hepatitis, and complete renal failure related to acute tubular necrosis that needed hemofiltration for 1 week. With the onset of symptoms of shock, it was interesting to observe that stool volume decreased rapidly, and the patient had 1 week with no stool at all, despite fluid overload before the onset of hemofiltration. After complete recovery until the time of transplantation, his daily stool volume stayed low (10 to 30 ml/kg per day). There was no correlation of the 24-hour stool volume with the intake of such drugs as H2 antagonists, that are thought to have some influence on the amount of secretions into the gastrointestinal tract. Indeed, after ranitidine, which was given during the period of intensive care, was discontinued, the daily stool volume did not increase again. Therefore, some enteral feedings were introduced through a nasogastric tube (1 ml/kg per hour of half-strength Neocate [Scientific Hospital Supplies Inc., Gaithersburg, MD, U.S.A.], leading to an increase in stool volume nearly equivalent to the amount of administered formula. Parenteral nutrition could not be reduced, because it was necessary to infuse a total amount of energy of 356 kJ/kg (85 kcal/kg) per day before and after the reduction of the daily stool volume to maintain a normal growth rate. The demands for intravenous fluid replacements decreased significantly, and the total amount of parenteral fluids administered daily was no more than 90 to 120 ml/kg in contrast to the requirements before the hemorrhagic shock (270 ml/kg per day). Metabolic acidosis or oliguria did not occur. Despite the dramatic decrease in stool volume, the small intestinal mucosa showed no morphologic improvement. Microvillus inclusions were present in the surface epithelium, and no differences in the distribution of the PAS stain could be recognized before and after the shock (Fig. 1B). Of note, evaluation of the colonic mucosa of the excised ascending colon at the time of transplantation and 3 months after the shock (the only colonic material available) did not reveal an obvious increase in PAS staining in the apical cytoplasm. Besides some patchy areas, usually at crypt openings, with some suggestion of PAS stain within the cytoplasm, brush-border staining on the surface epithelium appeared normal (Fig. 1A). The staining pattern was different in comparison with that of colonic sections of other MID patients available at the University Department of Paediatric Gastroenterology of the Royal Free Hospital in London that showed an accumulation of PAS-positive material within the epithelial cells and did not show a brush-border stain (Fig. 1C illustrates a representative biopsy from another case of microvillus atrophy). Electron microscopy was not performed on the colonic samples.FIG. 1: Periodic acid-Schiff-(PAS) stained sections. (A) Colonic mucosa from the patient. Surface epithelium showing brush-border staining (arrows; original magnification,× 850). (B) Jejunal mucosa from the patient taken at the same time as the colonic sample. Note accumulation of PAS-positive material within apical cytoplasm (arrows). The luminal PAS-positive material is mucus (original magnification, × 850). (C) Colonic mucosa from another case of microvillus atrophy showing PAS-positive material within epithelial cells (arrows; original magnification, × 850). In comparison to A, there is no brush border stain.When the patient was 12 months old, a combined orthotopic liver and small bowel transplantation was performed at the University of Nebraska with a graft from an AB0-compatible, cytomegalovirus-negative pediatric donor, using previously described techniques (7). Briefly, the small bowel from approximately 8 cm distal to the ligament of Treitz and the right colon of the patient were removed. An enteric anastomosis was performed between the jejunum of the transplanted bowel and the jejunum of the native bowel in an end-to-side fashion. An end-to-side ileocolonic anastomosis was subsequently performed between the donor ileum and the recipient colon. A loop ileostomy was fashioned to provide access to the distal bowel graft for performance of mucosal biopsy. Oral tacrolimus and oral prednisone were used for immunosuppression, as described previously (7). Oral prednisone was preceded by a 5-day steroid bolus protocol with intravenous methyl prednisolone (7). Tacrolimus levels were maintained at 25 ng/mL based on the IMX method (Abbott Laboratories, Chicago, IL, U.S.A.) for the first 4 weeks, then gradually lowered to 20 ng/mL and maintained between 15 and 20 ng/mL 3 months after transplantation. Well-known side effects of tacrolimus (8,9) were observed in the patient, including chronic renal insufficiency requiring oral furosemide as a permanent medication. For this reason, the patient was sensitive to fluid overloads that quickly caused airway edema, especially in the first 10 weeks after transplantation. Arterial hypertension was treated with enalapril and amlodipine. Four weeks after the transplant, mild biventricular hypertrophy was observed in two-dimensional echocardiographic measurements without hemodynamic alterations. After slightly lowering tacrolimus levels, no progression occurred. The patient had four episodes of mild bowel rejection, 8 days, 32 days, 48 days, and 3.5 months after receiving the transplant, which were treated with a 5-day steroid bolus protocol (7). During the first episode, the patient showed clinical signs of an ileus and some capillary leakage. Endoscopic findings of erosion, peeling of the mucosal surface, and histologic changes consistent with acute rejection (10) were restricted to the proximal allograft. During the second and third episodes of rejection, the patient did not show the suspicious clinical signs or visual endoscopic abnormalities. These were detected in routine analysis of biopsy specimens removed from the distal bowel graft biweekly during the first 6 weeks. In both cases, minimal to mild intraepithelial inflammation was present in some crypts. The fourth episode of rejection with active cryptitis in the distal ileum occurred 1 week after adenovirus-associated colitis developed and was signalled by increased ileostomy output. Histologic examination of graft specimens out of these episodes showed good villous architecture with no distinguishable abnormalities. The early postoperative course was complicated by two perforations of the proximal graft and then by septic peritonitis and intraperitoneal abscess formation, which demanded surgical interventions and lavage. Enteral nutrition was initiated with an amino acid formula (half-strength Neocate) at 0.5 ml/kg per hour, 6.5 weeks after transplantation. Two months after transplantation, a complex enteral feeding formula (PediaSure With Fiber; Ross, Columbus, OH, U.S.A.), was given without difficulty. Ten weeks after receiving an intestinal transplant, the patient was receiving full enteral feeding through his gastrostomy at a volume of approximately 95 ml/kg per day, administered continuously for a period of 24 hours. Introduction of oral loperamide reduced ileostomy output from 50 ml/kg per day to 30 ml/kg per day, and intravenous replacements of fluids could be discontinued. His weight was maintained between the 25th and 50th percentiles for age in the following months, and his general condition was excellent. Nevertheless, he remained dependent on continuous drip feedings because he refused to take significant amounts of oral food. Beside ileostomy output, the patient had very few anal stools of normal consistency. The ileostomy was taken down 6 months after the transplantation. No adverse effects were noted from the retention of the major part of the recipient's large bowel. The child had two to four stools per day of pulpy consistency. The patient died 190 days after transplantation. From a completely healthy status, fever and diarrhea developed at home, and the patient was admitted 24 hours later with severe dehydration. No bacterial or viral pathogen could be identified in stool and blood cultures. Gangrene and perforation of the intestinal graft were detected; and despite removal of about half of the graft, the patient died within 2 days after the onset of symptoms of the consequences of protracted shock. Histologic evaluation of the removed graft showed ischemic necrosis but no signs of allograft rejection or viral cytopathic effects. Several vascular structures belonging to larger, deeper arteries showed evidence of previously unrecognized transplant arteriopathy(11), with concentric narrowing of the vessel lumen. This was thought to have caused a diminished, although normally sufficient, blood supply to the intestinal graft. Under conditions of severe dehydration, this would explain the infarction. It would also explain why only segments of the graft and not native parts of the small or large intestine were ischemic. DISCUSSION Microvillus inclusion disease is a well-defined disorder within the syndrome of intractable diarrhea of infancy. Despite parenteral fluid and electrolyte replacement, the mortality rate is high in reported cases(2,3), and failure to thrive is common(3). One contributor to this bad outcome seems to be the metabolic acidosis that can easily occur in these patients under conditions of increased fluid demands (3,5,12). The condition requires permanent exogenous administration of such anionic bases as bicarbonate or acetate for maintaining normal blood pH. Assuming that the kidneys can compensate for intestinal losses of bicarbonate when the glomerular filtration rate is high enough, we adopted the concept of mere replacement of intestinal losses and further increased intravenous fluids at a constant sodium concentration, until acidosis permanently disappeared, which was the case at a urine production of 4 ml/kg hour (96 ml/kg per day). The reduction of intravenous fluids resulted in the reappearance of metabolic acidosis when urine production decreased to less than 3.5 ml/kg per hour. Under this therapeutic regimen, the patient thrived. The impressive diarrhea in MID with its high sodium concentration is still to some extent unexplained. The glucose-Na+ contransport in MID has normal features (13,14), and the direct and indirect constitutive exocytotic pathways appear to be intact(15). The diarrhea was assumed to result from an imbalance between decreased absorption and unaltered secretion(13). The latter explanation would be in good agreement with an achieved partial reduction of the daily stool output by administering a long-acting somatostatin analogue, as observed by Couper et al.(12), in that there is evidence that somatostatin inhibits secretion of gastrointestinal hormones that mediate fluid secretion in the gastrointestinal tract (16-18). However, other investigators have not observed a therapeutic benefit(5,19,20), and the best result obtained with somatostatin has been the reduction of stool output by one half(2,12). It is interesting that diarrhea in MID-in contrast to other secretory diarrheas such as congenital chloride diarrhea (21) and diarrhea related to deficient sodium-hydrogen exchange (22)-seems not to start antepartum in general. Polyhydramnios has not been described in any MID patient (2,3) except for one premature born boy who had dihydropyrimidinase deficiency as well (20). Furthermore, in most cases, the onset of diarrhea occurs from several hours to as long as 60 days after birth (2). In the patient reported here, diarrhea began in the second week of life, and his meconium appeared to be normal. It is possible that his diarrhea was not noticed because of its urine-like quality. However, there seems to be some event after birth that switches the intestinal tract irreversibly to producing secretory diarrhea. It is not known whether the enteropathy, which has been described in all cases of MID, is present from the beginning or whether its appearance coincides with the onset of diarrhea. It is of particular interest in this patient that hemorrhagic shock, severe enough to cause such tissue damage as acute tubular necrosis and ischemic hepatitis with consecutive need for cell division and cell differentiation to achieve tissue regeneration, led to a complete of diarrhea for a period and to a output than that seen It be significant that the PAS stain of the colonic mucosa had a appearance 3 months after the shock. However, because no colonic specimens were before the onset of symptoms of shock, it is to whether the PAS stain was in the colon before this as described in other cases It is possible that the obvious reduction in diarrhea after recovery from shock was related to a gain in function of his native colon as a of a normal mucosal differentiation during the course of tissue the colon have to the and the reduction have been caused by decreased we that the to reduction in daily stool volume after recovery from the shock was caused by secretion into the small with no on mucosal there was no in the mucosal appearance of the small intestine after the episode of shock. There was no obvious in because it was not possible to the total amount of and any enteral feeding led to an increase in stool there has been a of transplantation distal and small and large in a with MID there are two cases of successful intestinal transplantation in In both cases, the recipient's colon was removed. were between the ileum and 4 cm of colon of the recipient and between the colon and the of the patient is the first case of intestinal transplantation of MID in which the major part of the native colon was only the ascending colon was removed. The normal colonic mucosal architecture is to the of this with MID patients intestinal transplantation show whether the recipient's colon is to be The course of patient protracted of can be for intestinal graft The case illustrates that transplant can be present when the general condition of the patient is the growth rate is and no clinical signs such as or are In no in mucosal histologic appearance were seen in specimens obtained at a biopsy 1 week before the of the is thought to be a of intestinal transplantation, but this completely in case that its is and that when it a significant risk for graft failure and patient patient his of life after transplantation, and parenteral nutrition and fluid replacements could be discontinued. with intestinal transplantation in at the University of Nebraska good patient and graft for of liver and small bowel in part to the of tacrolimus as an (7). are not intestinal transplantation to the best for children with the of the disease is other