Combined liver and small bowel transplantation (LSBT) is a life-saving procedure in children with intestinal failure and parenteral nutrition–related end-stage liver disease. Although the experience is too limited yet to draw definitive conclusions, the small bowel may be less prone to rejection in this combination. Nevertheless, most published series of transplanted patients focus on intestinal function and rejection, and little is known about the liver (1–3).
Centrilobular necrosis is not an uncommon finding after isolated liver transplantation. Differential diagnosis includes rejection, ischemia, viruses, toxics, and autoimmunity (4–12). Associated with inflammation, with or without portal involvement, it is increasingly recognized as a specific feature of acute rejection. The prognosis may be ominous with development of chronic rejection. Various determinants have been used, either acute or chronic centrilobular rejection, centrilobular necrosis, veno-occlusive disease, or venular stenosis.
We report here centrilobular necrosis in children who underwent combined LSBT in our institution in a study designed to assess the predictive factors, the possible causes, and the prognosis of this finding in the particular setting of LSBT.
PATIENTS AND METHODS
From November 1994 to August 2000, 20 children underwent 21 combined LSBT in our institution, all of them with pediatric donors. Fifteen children survived more than 1 month, with a follow-up of 3 months to 6 years. Their clinical characteristics are reported in Table 1. In six children, centrilobular necrosis (CLN) was recorded at least once on liver biopsy (group 1). They were compared with the nine other children without CLN (group 2). The operative technique on the liver changed throughout our experience: direct biliobiliary anastomosis for the first child, Roux-en-Y loop for the second and third children (patients 7 and 8), “en bloc” liver, bile duct, and intestine, including duodenum (“Omaha procedure”) (3) from then on. The right colon was transplanted in some children (3/6 in group 1, 6/9 in group 2). All patients had an ileal stoma, which was closed from 2 months to 2 years after transplantation.
The immunosuppression protocol included tacrolimus (target blood level 15–20 ng/ml for the first month patients 1–3 and 7–12, 20–25 ng/ml for the other ones, tapered thereafter), methylprednisolone then prednisone (2 mg/kg, tapered to 1 mg/kg at day 30, 0.5 mg/kg at day 90, discontinued every other day from the sixth month), azathioprine (2 mg/kg, discontinued at different times after transplantation). Patient 6 received anti-CD25 antibodies and no azathioprine, because of retransplantation and loss of the first graft from chronic rejection. All children received prophylactic antibiotics until day 7, trimethoprim-sulfamethoxazole three times a week from day 7, prophylactic acyclovir (patients 1–3) or ganciclovir, and antihypertensive treatment (nifedipine, labetalol when needed).
Protocol intestinal biopsies were performed from day 6, twice a week initially then according to clinical findings. Liver biopsy was performed only when clinically or biologically indicated.
The clinical events in the six patients in group 1 before the discovery of CLN (index biopsy, IB) were as follows.
Patient 1 experienced acute liver rejection with typical portal findings 1 month after LSBT. Control liver biopsy showed persistence of mild portal inflammation and bile plugs. She underwent surgery 2 months after LSBT for drainage of a biliary collection. Roux-en-Y anastomosis was performed. Liver biopsy (IB) was performed because liver tests did not normalize after surgery.
Patient 2 had completely normal liver tests until 22 months after LSBT, when cholestasis and cytolysis appeared after a severe episode of bacterial diarrhea (IB).
Patient 3 had arterial thrombosis and was reoperated on day 7, albeit unsuccessfully. Cytomegalovirus infection was diagnosed and treated from day 14. Acute liver rejection was diagnosed and treated on day 20. Liver tests were normal until 3 months after LSBT. A first biopsy disclosed portal inflammation without endotheliitis, mild lobular inflammation, and ductular proliferation. IB was performed 6 months after LSBT because of persistently elevated liver tests.
Patient 4 experienced reelevation of liver tests after initial normalization on day 19. Liver biopsy showed mild lobular inflammation. Liver tests never returned completely to normal. Small bowel rejection was diagnosed and treated from day 36. Liver tests worsened during varicella-zoster infection, and did not improved after acyclovir treatment (IB).
Patient 5 was diagnosed with generalized lymphoproliferative disease 1 month after LSBT. Immunosuppression was dramatically reduced, and treatment with anti-CD20 antibody instituted. Increase in liver tests began 6 weeks later, when lymphoproliferative disease was in remission (IB).
Patient 6 underwent retransplantation of liver and small bowel, 6 months after a first transplant for pseudoobstruction that failed because of chronic bowel rejection. Increase in liver enzymes on day 23 led to liver biopsy (IB). Minimal bowel rejection was diagnosed at the same time.
The biologic characteristics of the six children at time of the IB are recorded in Table 2.
All nine children of group 2 underwent at least one liver biopsy during follow-up, either because of increase in liver tests (patients 7, 8, 11, 12, and 15) or systematically during every reoperation.
Twenty-five liver biopsies performed in the six patients were independently reviewed by two pathologists (D.C. and J.C.F.). For each patient, four biopsies on average were studied (range, 1–8). The first posttransplant biopsy was obtained between days 8 and 17 in four patients, 1 and 2 months after LSBT in the last two patients.
For all biopsies, samples were fixed in alcoholic Bouin or 10% formalin solution, embedded in paraffin, and stained with hematoxylin and eosin (H&E), periodic-acid Schiff (PAS), Sirius red, Masson’s trichrome, Perls, and Gordon Sweet stains.
Ultrasonography was performed in all cases and controlled regularly throughout the evolution. Viral infections were systematically investigated: viral serologies for hepatitis A, B, and C, cytomegalovirus (CMV), Epstein-Barr virus (EBV), parvovirus; antigenemia for CMV, polymerase chain reaction for EBV and Herpes type 6, rapid detection of adenovirus and enterovirus in stools. Antinuclear, anti-smooth muscle, and anti-liver-kidney microsome autoantibodies were systematically searched.
For the six patients in group 1, the liver lesions observed were quite similar, with only a difference in intensity of inflammation. These lesions consisted of a mixed but predominantly mononuclear inflammation associated with edema and hepatocyte necrosis in centrilobular topography. A subendothelial inflammation was also focally present, either in the perivenular area or in the portal veins. In many cases, the inflammatory lesions were associated with fibrosis of terminal hepatic venules or perivenular fibrosis. A mild mononuclear portal inflammation was observed in five patients, focally associated with portal fibrosis. In only one patient were bile ducts damaged (Figs. 1–3).
In the nine patients of group 2, liver biopsy disclosed either a mild acute rejection with classic portal findings, without associated lobular inflammation (three patients, once each) or symptoms of biliary obstruction (one patient), or was normal (all other cases).
In the six patients in group 1, no viral infection or autoimmune disease was evidenced at the time of the first biopsy, or subsequently, to explain the histologic findings. Ultrasonography, which was routinely performed to assess blood flow, was always normal, except in patient 3 who underwent arteriography and reintervention for arterial thrombosis.
Treatment and Outcome
In patients 1, 3, and 4, abnormal liver enzymes were observed in the first weeks after LSBT. The first biopsies showed only mild inflammatory lesions without obvious biliary or venular tropism, and were not interpreted as rejection. In patients 1 and 3, CLN was seen for the first time 3 and 6 months after LSBT, and interpreted as possible azathioprine toxicity. Azathioprine was suspended. Because of persistent abnormalities, rejection was suspected and treated. In the other patients, rejection was diagnosed and treated earlier, as soon as CLN appeared. In all children, high-dose methylprednisolone was followed by initial normalization of liver tests. A follow-up biopsy was not systematically performed at this time.
Later on, recurrence of biologic abnormalities was observed after severe infections: herpes meningoencephalitis and EBV primary infection in patient 1, CMV primary infection in patient 5, respectively 18 and 8 months after discovery of CLN. A control liver biopsy at this time showed persistent lesions, albeit less intense.
In patients 1–3, liver tests worsened also when the dosage of prednisone was less than 0.5 mg/kg/d. Mycophenolate mofetil (MMF) was then added. In patient 4 MMF was introduced from the occurrence of CLN. From this time on, liver tests remained most of the time below twice the upper limit of normal. Target tacrolimus level was 8–10 ng/ml, prednisone 0.2–0.3 mg/kg on alternate days, and MMF 600 mg/1.73 m2. Patients 1–3 were switched back from MMF to azathioprine because of digestive intolerance, in all of them after more than 1 year of MMF. In patients 1 and 3, there was an increase in liver enzymes, transient in patient 3. In patient 1, liver biopsy was once more performed, showing persistent inflammatory lesions more intense than on the previous biopsy, and a significant degree of fibrosis: the dosage of prednisone was increased and MMF reintroduced. This child had also developed from the third year after LSBT a severe inflammatory colitis on her native colon, for which she was treated with sulfasalazine and antibiotics (colimycin and metronidazole) with incomplete improvement. Results of the last biologic and histologic control are reported in Table 3. Perivenular inflammation was improved except in patient 1, but was still present in all children who had biopsies. Fibrosis worsened on subsequent biopsies in three children (patients 1, 2, and 5).
Despite these histologic findings, none of the children presented during follow-up with symptoms such as ascites or jaundice, and the liver was not enlarged on clinical examination.
We compared these six children with the nine others who did not develop CLN. We compared the total ischemic time of the graft, the level of immunosuppression, the rate of associated transplantation of the colon, the rate of early reoperation (first 6 months) whatever the reason, the delay of weaning of parenteral nutrition, the timing for closure of ileal stoma, the rate of early bacterial or viral infection, and the rate of intestinal rejection and found no differences (Table 1). To focus on early severe events that could have triggered rejection, five children in group 2 were reoperated on during the first month, three had bacterial infection, and one, CMV infection during this same period of time. Two, however, presented an easy postoperative course with only one episode each of mild acute rejection, of the liver for one and the gut for the other one.
We report for the first time CLN in 6 of 20 consecutive combined LSBT. CLN is reported in 10–30% of adults (4,5,8,12) and 10% of children (10,15) after liver transplantation. The gradient of enzymatic activity and oxygenation along the axis of the hepatic lobule (zonation) is responsible for the particular location of necrosis. Hepatocytes located around the central vein (zone 3) work in low oxygen content and are thus extremely sensitive to ischemia and other aggravations such as acute biliary obstruction and azathioprine toxicity (10).
The initial Banff criteria for histologic diagnosis of liver rejection described centrilobular lesions as a possible reflection of obliterative vasculopathy (13), in conjunction with typical portal lesions of acute rejection (14). In the last update, there is more emphasis on the early phase of chronic rejection in the terminal hepatic venules, characterized as “alterations of the central venules and surrounding hepatic parenchyma”(16). The endothelium of the central vein may be a target for immune rejection (phlebitis). The local congestion can block the outflow, and the inflammatory component can spread from the subendothelium to the sinusoids (9,11). CLN has been increasingly recognized as a form of graft rejection carrying a poor prognosis, especially when associated with ductopenia (4,5,8,10–12,15). When there are no biliary lesions, the prognosis may be less severe (9,17).
Differential diagnosis includes vascular compromise, azathioprine or tacrolimus hepatotoxicity, viral hepatitis, and de novo autoimmune hepatitis (6,9,10,12,18). We did not systematically perform angiography (10), because we think that ultrasonography is a powerful tool to diagnose thrombosis amenable to surgery and that the finding of discrete lesions on peripheral vessels may serve to strengthen the diagnosis of vascular rejection. The liver lesions induced by azathioprine appear generally later than those of rejection, and inflammation and endotheliitis are less prominent (19). The histologic lesions described with tacrolimus include an extravasation of erythrocytes in Disse’s space, and only mild inflammation (18). In our patients liver tests tended to improve with increase in the dosage of tacrolimus, which would argue against a toxic mechanism. Viral hepatitis is important to exclude because the treatment (low immunosuppression) is opposite to that of rejection. Autoimmune hepatitis is rare and diagnosed on prominent portal inflammation and presence of autoantibodies. In cases of CLN because of rejection, the liver tests and histologic lesions improve only with increased immunosuppression.
Little is known about the liver in most reports on intestinal transplantation. One article only describes the pathology of liver allografts in combined transplantation (20): of 13 patients, one had similar lesions to our patients, and developed chronic rejection 1 year after transplantation.
We believe that in our patients this lesion is a reflection of chronic vascular rejection. Although we did not systematically perform angiography, we believe that only in one child arterial thrombosis may have contributed to the ischemic lesions. CLN was in some cases associated with portal lesions typical of rejection, and was responsive to increased immunosuppression. Its evolution was really chronic under heavy immunosuppression including a high level of daily prednisone. The addition of MMF seemed helpful to spare prednisone, but we were concerned about the digestive side effects (diarrhea) of the drug (21). We observed in some children the development of rather severe fibrosis, close to veno-occlusive disease (6,11), whereas in others the lesions remained inflammatory and mildly fibrotic.
We studied which characteristics of these children could explain this liver reaction. Epithelial dysplasia was an initial diagnosis more common in this group, but the numbers are too small to draw a conclusion. Older grafts seem more prone to chronic rejection (22), but all the donors were pediatric. The duration of parenteral nutrition after transplantation could have influenced the cholestatic reaction, but there were no obvious differences between the groups. The associated transplanted colon could favor bacterial translocation, and thus immune stimulation, but only half the children had received the colon, and five were among the nine with a normal liver. We can speculate though that in the first patient, the chronic inflammation owing to colitis could have triggered a constant immune stimulation. Early reoperation could also favor immune stimulation, but there was no difference between the two groups. Considering immunosuppression, the target level of tacrolimus was higher from transplantation for patient 5 in group 1, and for patient 12 in group 2; the dose of prednisone and azathioprine were the same in the whole series; and only one patient had received antiCD25 antibodies. Immunosuppression could thus not be an explanation. One child in both groups had been retransplanted. Azathioprine was a suspect in patients 1 and 3, but suspension of the therapy did not modify the course.
However, we observed that the first occurrence of CLN was noted in the weeks or few months after an acute event: biliary surgery, arterial thrombosis and CMV infection, bacterial diarrhea, varicella, lymphoproliferative disease, or retransplantation. Although we could not demonstrate it by the blood level of immune suppressive drugs, the effective immunosuppression at the tissue level could have been modified in some children, but not all.
The immunosuppression given to these children was far heavier than what is usual for an isolated liver graft. This total blockade of the immune system could have prevented the process of induction of tolerance: the latter requires activation of T and B lymphocytes (23,24) and is part of the normal reaction to the allograft. Why rejection occurred only in certain children, and why in the liver and not in the intestine as well, remains unexplained.
This raises two important questions: what is the prognosis of these lesions, some already severe a few years after transplantation; and what is the optimal protocol of immunosuppression to prevent such an injury (15) ?
1. Reyes J, Bueno J, Kocoshis S, et al. Current status of intestinal transplantation in children. J Pediatr Surg 1998; 33: 243.
2. Atkinson P, Chatzipetrou M, Tsaroucha A, Lehmann R, Tzakis A, Grant D. Small bowel transplantation in children. Pediatr Transpl 1997; 1: 111.
3. Langnas A, Shaw BW Jr, Antonson DL, et al. Preliminary experience with intestinal transplantation in infants and children. Pediatrics 1996; 97: 443.
4. Gomez R, Colina F, Moreno E, et al. Etiopathogenesis and prognosis of centrilobular necrosis in hepatic grafts. J Hepatol 1994; 21: 441.
5. Ludwig J, Gross JB Jr, Perkins JD, Moore SD. Persistent centrilobular necroses in hepatic allografts. Hum Pathol 1990; 21: 656.
6. Dhillon AP, Burroughs AK, Hudson M, Shah N, Rolles K, Scheuer PJ. Hepatic venular stenosis after orthotopic liver transplantation. Hepatology 1994; 19: 106.
7. Anand AC, Hubscher SG, Gunson BK, McMaster P, Neuberger JM. Timing, significance and prognosis of late acute liver allograft rejection. Transplantation 1995; 60: 1098.
8. Turlin B, Slapak GI, Hayllar KM, Heaton N, Williams R, Portmann B. Centrilobular necrosis after orthotopic liver transplantation: a longitudinal clinicopathologic study in 71 patients. Liver Transpl Surg 1995; 1: 285.
9. Tsamandas AC, Jain AB, Felekouras ES, Fung JJ, Demetris AJ, Lee RG. Central venulitis in the allograft liver. Transplantation 1997; 64: 252.
10. Allen KJ, Rand EB, Hart J, Whitington PF. Prognostic implications of centrilobular necrosis in pediatric liver transplant recipients. Transplantation 1998; 65: 692.
11. Sebagh M, Debette M, Samuel D, et al. “Silent” presentation of veno-occlusive disease after liver transplantation as part of the process of cellular rejection with endothelial predilection. Hepatology 1999; 30: 1144.
12. Nakazawa Y, Walker NI, Kerlin P, et al. Clinicopathological analysis of liver allograft biopsies with late centrilobular necrosis. Transplantation 2000; 69: 1599.
13. International Working Party. Terminology for hepatic allograft rejection. Hepatology 1995; 22: 648.
14. International Panel. Banff schema for grading liver allograft rejection: an international consensus document. Hepatology 1997; 25: 658.
15. Nakazawa Y, Jonsson JR, Walker NI, et al. Fibrous obliterative lesions of veins contribute to progressive fibrosis in chronic liver allograft rejection. Hepatology 2000; 32: 1240.
16. International panel. Update of the international Banff schema for liver allograft rejection: working recommendations for the histopathologic staging and reporting of chronic rejection. Hepatology 2000; 31: 792.
17. Martin S, Russo P, Dubois J, Alvarez F. Centrilobular fibrosis in long-term follow up of pediatric liver transplant recipients [Abstract]. J Pediatr Gastroenterol Nutr 2000; 31 (suppl 2): S195.
18. Hytiroglou P, Lee R, Sharma K, et al. FK506 versus cyclosporine as primary immunosuppressive agent for orthotopic liver allograft recipients: histologic and immunopathologic observations. Transplantation 1993; 56: 1389.
19. Sterneck M, Wiesner R, Ascher N, et al. Azathioprine hepatotoxicity after liver transplantation. Hepatology 1991; 14: 806.
20. Tsamandas AC, Furukawa H, Abu-Elmagd K, Todo S, Demetris AJ, Lee RJ. Liver allograft pathology in liver-small bowel or multivisceral recipients. Mod Pathol 1996; 9: 767.
21. Ducloux D, Ottignon Y, Semhoun-Ducloux S, et al. Mycophenolate mofetil-induced villous atrophy. Transplantation 1998; 66: 1115.
22. Waaga AM, Gasser M, Laskowski I, Tilney NL. Mechanisms of chronic rejection. Curr Opin Immunol 2000; 12: 522.
23. Li Y, Li XC, Zheng XX, Wells AD, Turka LA, Strom TB. Blocking both signal 1 and signal 2 of T cell activation prevents apoptosis of alloreactive T cells and induction of peripheral allograft tolerance. Nat Med 1999; 5: 1298.
24. Glynne R, Akkaraju S, Healy JI, Rayner J, Goodnow CC, Mack DH. How self-tolerance and the immunosuppressive drug FK506 prevent B cell mitogenesis. Nature 2000; 403: 672.