Combined Bowel-Liver Transplantation in an Infant with Microvillous Inclusion Disease

Microvillous inclusion disease is an uncommon congenital disorder thought to be transmitted on an autosomal recessive basis (1). It typically manifests itself as an intractable secretory diarrhea, usually within the first weeks of life. The diagnosis is confirmed on electron microscopy of intestinal mucosal biopsies, which show characteristic intracytoplasmic vesicles containing microvilli, whereas the enterocyte apical membrane lacks microvilli. Both the small and large bowels are involved. No treatment is currently available, and patients are supported by total parenteral nutrition (TPN) plus replacement of their massive intestinal losses (2). The vast majority of affected children eventually die of septic complications or hepatic insufficiency stemming from TPN-induced cholestasis. We report the case of a boy with microvillous inclusion disease and cirrhosis successfully treated by combined bowel and liver transplantation. CASE REPORT The patient was born at term to healthy, nonconsanguineous French Canadian parents after a normal pregnancy and uncomplicated delivery; he weighed 3,350 g. He required 2 days of phototherapy for jaundice due to an ABO incompatibility and was then discharged. At 1 week of age, he was readmitted to Ste-Justine Hospital for vomiting, dehydration, and metabolic acidosis. Despite rehydration with i.v. fluids and treatment of a urinary tract infection, his metabolic acidosis persisted. Screening for metabolic diseases or a cogenital infection showed negative results. Once rehydrated, the infant was noted to have six to seven large-volume, viscous, and malodorous stools per day. Upper and lower endoscopies were macroscopically unremarkable, but histology of the duodenum confirmed severe villous atrophy with normal crypts and no abnormal inflammatory infiltrate. Light microscopy of esophageal, antral, and colonic biopsies were normal. Electron microscopy confirmed the diagnosis of microvillous inclusion disease, with a nearly complete lack of microvilli on the apical membrane of the enterocyte. Characteristic vesicles containing microvilli were noted within the enterocytic cytoplasm of duodenal as well as colonic biopsies. Brush-border disaccharidases were uniformly diminished. Beginning at the age of 2 weeks, the patient was completely dependent upon TPN. His secretory diarrhea persisted despite withholding of enteral feeds, necessitating 8 mEq/kg/day of sodium and 170 ml/kg/day of intravenous fluids. In order to maintain oral stimulation, he was allowed to drink only Pedialyte® in small quantities (10 ml). He nevertheless had 30-70 g/kg of enteric losses daily. Attempts to increase calories enterally (Pregestimil®, Neocate®) resulted in an exacerbation of diarrhea with carbohydrate malabsorption (stool pH <5, reducing substance positive). Growth retardation and mild developmental delay became evident. The mild, mixed hyperbilirubinemia (total 50 and direct 16 μmol/L) was initially attributed to an ABO incompatibility. However, cholestasis and elevated transaminases progressed rapidly with TPN. At 4 months, a hepatic biopsy showed the presence of cholestasis with mild ductular proliferation, ballooning of hepatocytes, and giant cell transformation, suggestive of TPN-induced liver dysfunction. Other causes of cholestasis were excluded. At the age of 7 months, in view of hepatic insufficiency with coagulopathy and progressive cholestasis, in association with complete gut failure, he underwent a combined orthotopic bowel-liver transplant. The grafts were obtained from a 5-month-old ABO-compatible, cytomegalovirus-negative donor. Our patient weighed 5.3 kg at the time of transplant, which was performed at the Children's Hospital of Western Ontario. The native bowel was resected from just distal to the ligament of Treitz to the rectum. Donor proximal jejunum was anastomosed to the recipient's proximal jejunum and the transverse colon to the native rectum. An ileal vent (Bishop-Koop chimney) was made in order to provide access to the bowel graft for biopsies. The liver was transplanted with a Roux-en-Y biliary anastomosis. The early postoperative period was complicated by a perforation proximal to the distal ileal anastomosis of the Bishop-Koop chimney 2 weeks after transplantation. It was thought to be due to a vascular injury at the time of transplantation and/or a complication of the forceps biopsy. The distal part of the terminal ileum and proximal ascending colon were excised, and another ileostomy with a colonic mucus fistula was performed. Immunosuppression was initiated with FK506 (Tacrolimus) as a continuous i.v. infusion. Enteral FK506 was added 3 weeks later, and the i.v. form was discontinued 6 weeks posttransplant. Doses were adjusted to attain levels between 20 and 45 ng/ml, as determined by IMA automated whole blood FK506 assay using a microparticle enzyme immunoassay (Abbott Laboratories, Mississauga, Ontario, Canada). The mean (±SD), FK506 level over the first 3 months was 30.6 ± 9.9 ng/ml. Solumedrol was also given i.v. from the first day and tapered subsequently to 5 mg prednisone orally per day. The patient also received OKT3 (2.5 mg/kg/day) for 3 days, with the first dose at the time of surgery. One unit of donor blood was also administered intraoperatively. Intestinal biopsies were initially monitored twice weekly, then weekly, and then if rejection was suspected. Stomal output was followed daily, and weekly Tc 99-DTPA scans were carried out to monitor intestinal permeability. Liver function was followed with biochemical markers, which always remained normal. Only one mild episode of bowel rejection without liver involvement occurred 3 months posttrasplant. Stomal output had increased to 800 ml/day at a time when difficulties were encountered in trying to attain therapeutic FK506 levels by enteral administration. The rejection was easily controlled by reinstituting intravenous FK506 while maintaining the prednisone at 5 mg per day. Significant side effects from FK506 were encountered early on. Recurrent anasarca was noted, including airway edema that prevented weaning from the ventilator for the first 4 weeks. Pleural and pericardial effusions were also problematic. Arterial hypertension was controlled with maximal doses of nifedipine. Other adverse effects of FK506 were renal insufficiency (blood urea nitrogen 17 μmol/L, creatinine 89 μmol/L), hyperkalemia due to toxic renal tubulopathy, glucose intolerance (without insulin requirement), hypoalbuminemia, and lymphopenia. Four months posttransplant, another episode of pericardial effusion and hemodynamic failure ensued, requiring vasopressor agents. It was accompanied by oliguria, hypoalbuminemia, and peripheral and airway edema with respiratory failure necessitating mechanical ventilation for 7 days. An EKG showed signs of biventricular hypertrophy but no ST segment changes. Cardiac rhythm remained normal, as were serum CPK values. Echocardiography showed the presence of a pericardial effusion and concentric hypertrophic cardiomyopathy. After exclusion of an infectious etiology or rejection, the cardiomyopathy and pericardial effusions were suspected to be related to FK506 toxicity. The FK506 dose at the time of this episode was 0.5 mg i.v.t.i.d., and serum levels were in the range of 30-40 ng/ml. Symptoms completely regressed upon cessation of the drug. Further immunosuppression was administered using antilymphocyte globulin and azathioprine for the first 10 days. The patient was then switched to a more readily absorbable form of cyclosporin A (Neoral) while continuing azathioprine and prednisone, without further problems. Enteral nutrition was initiated with a protein hydrolysate formula (Alimentum) at 3 weeks, and solids were added 5 months posttransplant. Intestinal transit time was initially slowed using codeine and loperamide. Steatorrhea of 14.4 g/day at 5 months and 7.5 g/day at 8 months posttransplant were noted, representing 30 and 23%, respectively, of long-chain triglycerides ingested. Nevertheless, he experienced a sustained weight gain with an oral caloric intake of 120-190 kcal/kg/day. He was on full enteral feeds from the sixth postoperative week. One episode of diarrhea due to Clostridium difficile occurred, accompanied by stool volumes of 300-500 g/day. Rejection was excluded histologically, and the diarrhea responded to oral metronidazole. Persistent prerenal insufficiency (urinary density ≥1.025, urea 18.8 μmol/L, urinary sodium <5 mEq/L, glomerular filtration rate 113 ml/min) and metabolic acidosis requiring bicarbonate supplements were attributed to prolonged intestinal losses of electrolytes and water. Stoma closure was therefore performed 8 months posttransplant. The renal insufficiency then promptly corrected without alteration of cyclosporine levels, and the bicarbonate losses gradually diminished. Aside from two episodes of septicemia in the immediate posttransplant period (central catheter and intestinal origin), the only other infectious event was seroconversion to Epstein-Barr virus during the first months posttransplant. No lymphoproliferative syndrome was encountered. At 22 months posttransplant, the patient is tolerating a normal diet for age without any requirements for tube feedings. His immunosuppression has been lessened to 2 mg/kg t.i.d. Neoral (blood cyclosporine levels of 80 ng/ml), 0.6 mg/kg/day azathioprine, and 0.3 mg/kg/day prednisone. He is active and thriving, with excellent weight gain and growth velocities (both >95% for age over the past 6 months) and completely normal development. His weight and height have progressed from well below the third percentile before the transplant to the 15th percentile at 26 months. DISCUSSION Microvillous inclusion disease is a relatively rare disorder, represented in only five of the 35 infants <3 months of age hospitalized with a diagnosis of severe, TPN-dependent intractable diarrhea at Ste-Justine Hospital over the past two decades (3). The pathogenesis of this disease remains unknown, and the only treatment available is nutritional support by central venous catheters. Our patient did not have a significant clinical response to a long-acting somatostatin analogue, as had been previously reported (4). The typical clinical course for most affected infants is that of a severe secretory diarrhea requiring home TPN. Patients are intolerant of enteral feedings (1,4), as was the case with our patient. Eventually, most patients die of TPN-related liver failure, sepsis, or complications related to central lines, including the lack of venous access sites (1,2). Indeed our previous three patients all died, the last two while waiting for a small-bowel donor. Advancements in immunosuppression, particularly the use of FK506, have recently allowed small-bowel transplantation to become a realistic option, with good survival of both patients and grafts (5-9). Recently, Oliva et al. reported the first successful intestinal transplant for microvillous inclusion disease in a patient 2.5 years old. (8). Our patient, at 7 months, was one of the youngest children ever to have undergone a whole-bowel transplant. No significant problems with rejection were encountered, but severe toxicity with the immunosuppressive drug FK506 caused major treatment difficulties. They included concentric hypertrophic cardiomyopathy with pericardial effusions and anasarca. The second episode resulted in life-threatening cardiorespiratory failure, leading to discontinuation of FK506. Adequate immunosuppression was achieved by switching to Neoral, a more readily absorbable microemulsion formulation of cyclosporin A. Although these adverse effects have not been reported with FK506 in humans (9-17), we have recently observed this problem in four other bowel and/or liver transplant patients (18). The pathophysiology of the FK506-induced cardiomyopathy is not known. Vasculitis limited to the heart has been reported in dogs (19), and accelerated artherosclerotic changes of the coronary vessels have been observed in rabbits (20). Although no cardiac biopsy was done, a dose-related cardiotoxicity has been recently reported with long-term parenteral administration of FK506 in rabbits (21). As in our patient, recovery was seen after drug discontinuation. No elevation of creatinine kinase was noted, nor were other signs of myocarditis. It thus seems likely that the cardiac problem was not due to either an inflammatory or an ischemic process. The absence of toxic FK506 blood levels was further evidenced by the absence of serious renal dysfunction or signs of neurotoxicity. From the fifth posttransplant month onward, our patient's immunosuppression thus consisted of Neoral (cyclosporine levels of 200 ng/ml initially, now down to 80 ng/ml), azathioprin and prednisone. Fortunately, no episodes of rejection have ensued >22 months posttransplant. Perhaps the time of discontinuing FK506 therapy was critical, since most rejection episodes appear to happen during the first 3 months after bowel transplant. Another very important and encouraging result from the experience with this patient was his very rapid adaptation to and acceptance of oral feeds. This finding is in distinction to the other recently reported case of a patient who required long-term jejunostomy feeding because of a disinterest in eating (8). Although it is true that our patient underwent transplantation at a much earlier age, intensive intervention on the part of the parents was likely instrumental in maintaining the baby's interest in drinking and eating. In conclusion, combined bowel-liver transplantation should be considered as a potentially life-saving therapy for microvillous inclusion disease in association with liver failure, even in infants <1 year of age. Although the toxic effects of FK506 compelled us to discontinue its use, long-term graft survival was possible with cyclosporin A in combination with azathioprine and low-dose steroids once the early, high-risk period posttransplantation had passed. Acknowledgment: Drs. Seidman and Paradis are supported by Chercheur-Boursier Clinicien research scholarship awards from the Fonds de Recherche en Santé du Québec. The authors thank Dr. P. Brochu for help in discussions on the liver histology and D. Lachapelle for expert secretarial assistance.