Acute graft-versus-host disease (GVHD) occurs when histoincompatible lymphocytes from an allogenic donor engraft in a transplant recipient. Although the incidence of GVHD after liver transplantation is low (1%), the mortality rate is very high at approximately 85%.1 It is estimated that more than 1010 donor lymphocytes remain in the portal tracts and the parenchyma of a donor liver graft after flushing with cold preservative solution,2, 3 which is greater than the number of lymphocytes that are administered during bone marrow transplantation (BMT).4 The donor lymphocytes colonize the recipient, recognize the host tissue antigens as foreign, and react against the host tissue. GVHD typically occurs 1–6 weeks after an orthotopic liver transplantation (OLT) and presents with fever, skin rash, diarrhea, and/or pancytopenia.1 Unlike GVHD after BMT where the biliary epithelium is a major target, the transplanted liver is not a target for GVHD, as the liver allograft and the immunocompetent cells responsible for GVHD are both of donor origin. Marked neutropenia and thrombocytopenia often occur several days after the initial presentation of GVHD and frequently lead to life-threatening infection or hemorrhage.
We reported 11 cases of GVHD among 1082 liver transplants performed at our institution.1 Treatment typically consisted of increasing immunosuppression, use of antibody preparations such as OKT3 or antithymocyte globulin to eliminate the donor lymphocytes, and supporting myelopoiesis by use of cytokines. However, 10 of the 11 patients died, and cause of death was primarily the result of native bone marrow failure resulting in fatal sepsis. The results from other reported cases were similar.5, 6 The last unsuccessful case in 2001 prompted us to reevaluate our therapeutic approach to this group of patients. We reasoned that we needed to protect the recipient's bone marrow function by allowing the recipient's immune system to reject the donor lymphocytes. Thus, we decided to aggressively use single tandem repeat (STR) testing to confirm the diagnosis early and immediately dramatically reduce immunosuppression as soon as a diagnosis of GVHD is established. This report describes the results of this approach in 3 patients with established or suspected GVHD after liver transplantation.
A 58-year-old Caucasian male received an OLT for end-stage liver disease due to primary sclerosing cholangitis. The donor was a 24-year-old Caucasian male. The patient's HLA type was A1 B8, 37 Cw*06XX, 07XX Bw4, 6 DR1, 17 DQ2, 5 and the donor type was HLA A1 B8 Bw6 Cw*07XX DR4, 17 DQ 2, 8. The lymphocyte cross-match was negative against both T cells and B cells. Both the donor and recipient were blood type A. The patient received immunosuppression consisting of tacrolimus (blood levels maintained between 10–15 ng/dL), mycophenolate mofetil 1000 mg p.o. twice daily, and corticosteroids. He developed moderate acute cellular rejection on postoperative day (POD) 5, which was successfully treated with a corticosteroid recycle. The patient developed an unrelated anastomotic bile leak, which required operative revision on POD 15. At 10 weeks after transplantation, the patient developed a generalized skin rash (Figure 1). The patient did not have a fever or diarrhea, and his white blood cell (WBC) count was 5400 cells/mm3. His platelet counts dropped from 176,000 cells/mm3 to 33,000 cells/mm3. A skin biopsy showed histological features consistent with GVHD and infiltrating donor lymphocytes (Figure 2). STR testing of the CD3+ cells in his peripheral blood showed 7% donor cells. His skin biopsy was also tested by STR and showed 10% donor DNA. This was considered to be laboratory confirmation of the clinical diagnosis of GVHD. At the time of the GVHD diagnosis, the patient's immunosuppression regimen included tacrolimus (blood levels of 14.1 ng/dL) and prednisolone 5 mg p.o. q.d. After careful discussion with the patient and his family, full consent was obtained from the patient for using this therapeutic approach of withdrawal of immunosuppression. His immunosuppression was stopped, and during the next week his blood tacrolimus levels were <2 ng/dL. After 7 days, the patient developed biopsy-proven mild acute cellular rejection of the liver, his rash improved, and the STR testing showed the CD3 donor cells were reduced to 0% (Figure 3). The acute rejection responded to increased tacrolimus levels and a corticosteroid recycle; mycophenolate mofetil was restarted. The patient's STR testing was monitored on a weekly basis. During the 14th week after transplantation, the patient developed a low-grade fever (100°F), and STR testing showed that his donor cells were again detectable (6% CD3+ donor cells). However, his platelet counts and WBC counts remained stable. A clinical diagnosis of relapsing GVHD was made, and tacrolimus was discontinued. Immunosuppression included only prednisolone 20 mg q.d. During the 19th week after transplantation, the patient again developed mild acute cellular rejection, and STR testing showed the CD3+ donor cell levels were reduced to 1.8% and 0% over the next 2 weeks (Figure 3). The acute rejection was treated with 3 alternate-day pulses of 1 g hydrocortisone. The acute rejection diminished. After that, tacrolimus was restarted and levels were maintained between 5–10 ng/dL. The patient developed a re-stricture of the bile duct, which was surgically corrected. After more than a year of follow-up, he has no symptoms of GVHD and continues to maintain undetectable levels of CD3+ donor cells on STR testing.
A 53-year-old Caucasian male received an OLT for end-stage liver disease due to primary sclerosing cholangitis. The donor was a 50-year-old Caucasian male. The patient's HLA type was A1, 2 B8 Bw6; DR17 DQ2 and the donor type was HLA A1, 30 B18, 63 Bw 4,6 DR13, 17 DQ2, 6. The lymphocyte cross-match was negative against both T cells and B cells. Both the donor and recipient were blood type B. The patient's immunosuppression regimen consisted of tacrolimus (blood levels maintained between 10–15 ng/dL), mycophenolate mofetil 1000 mg p.o. bid, and corticosteroids. On POD 7, mild acute cellular rejection was confirmed by liver biopsy, and it was successfully treated with a steroid recycle. STR testing showed 20% CD3+ donor cells on POD 7, and the CD3+ donor cells persisted between 8% and 14% until the 16th week (Figure 4). Because he did not show clinical signs of GVHD during this time, no intervention was made. By the 18th week, however, he developed fever and neutropenia without a rash or diarrhea. Viral studies for cytomegalovirus were negative. Drug-induced neutropenia was also clinically excluded. STR testing then showed 11% donor CD3+ and CD8+ cells. Despite the absence of the characteristic rash or diarrhea of GVHD, the clinical diagnosis of GVHD was made based on the chronically elevated level of CD3+ and CD8+ cells and the presence of fever and neutropenia. Immunosuppression was reduced. The serum tacrolimus levels decreased to 2 ng/dL. After 10 days, he developed biopsy-proven mild acute rejection. The STR testing at the 20th postoperative week showed the donor CD3+ and CD8+ donor cells had disappeared from the peripheral blood. The acute rejection was treated only by increasing the dose of tacrolimus to maintain blood level between 10–15 ng/dL. The clinical symptoms and fever improved, the patient's WBC counts started to increase, and at the 24th week, the WBC counts normalized. After 6 months of follow-up, the patient is healthy with good liver allograft function; he has had no relapse of GVHD and continues to maintain undetectable levels of CD3 donor cells on STR testing.
A 62-year-old Japanese male received a liver transplant for end-stage liver disease due to chronic hepatitis B infection with hepatocellular carcinoma. The donor was an 8-year-old Hispanic male. The recipient HLA type was A2 B51, 61 Bw4, 6 DR8, 14 DQ 5, 8. The donor HLA type was A11, 24 B8, 62 Bw6 DR4, 8 DQ 4, 8. The lymphocyte cross-match was negative against both T cells and B cells. Both the donor and recipient were blood type A. He received immunosuppression consisting of cyclosporine, rapamycin, and corticosteroids. In the 2nd week after transplantation, the patient developed a generalized rash. A skin biopsy was suggestive of GVHD (Figure 5) and STR testing revealed 26.5% CD3+ donor cells (Figure 6). The following day, he developed diarrhea and neutropenia (WBC 1500 cells/mm3). A clinical diagnosis of GVHD was entertained, and cyclosporine and rapamycin were stopped. Despite stopping the immunosuppression, the neutropenia and thrombocytopenia progressed until no neutrophils could be detected, and the patient was dependent on platelet transfusions. The rash increased and diarrhea became worse. The STR testing revealed that the donor cells had increased, peaking at 84% CD3+ donor cells (Figure 6). A bone marrow biopsy was performed, and it showed a hypoplastic marrow with few hematopoietic cells. Because the patient was deteriorating with severe thrombocytopenia and neutropenia despite withdrawing immunosuppression and very few remaining lymphocytes were of recipient origin, a decision was made at this juncture to treat this patient's GVHD similar to GVHD in a BMT recipient by again increasing immunosuppression. Antithymocyte globulin (Thymoglobulin from IMTIX-Sangstat, Lyon, France) was started at a high dose of 1.5 mg/kg body weight. After 10 days of Thymoglobulin therapy, the patient showed a transient clinical response and the rash improved; however, the diarrhea persisted. The neutrophil count recovered into the normal range. The STR testing revealed the level of CD3+ donor cells had decreased to 10% (Figure 6). However, within a week of stopping Thymoglobulin, the patient had a relapse; the STR testing showing the level of CD3+ donor cells again increased to 57% (Figure 6). A second course of Thymoglobulin was initiated. There was a response to the treatment. The donor CD3 cell counts decreased (Figure 6), but the patient developed continuous diarrhea with exfoliation of the colonic mucosa, bacterial sepsis, and cytomegalovirus infection. The patient died on POD 63 from uncontrolled sepsis.
GVHD, graft-versus-host disease; BMT, bone marrow transplantation; OLT, orthotopic liver transplantation; STR, short tandem repeat.
Management of GVHD after liver transplantation, unlike GVHD following BMT, has not been standardized. The main difficulty in evaluating the efficacy of treatment modalities has been the low incidence of the disease, which means that we are left to review individual case reports or small groups of cases that have received widely varying treatments. Most treatments focus on increasing immunosuppression, usually in the form of antibody preparations such as antithymocyte globulin (ATG) or OKT3 as in treating conventional GVHD in a BMT recipient. However, several reports5–10 including our own prior experience1 did not demonstrate any survival benefit with this therapy. In our previous review of published cases,1 there were 8 cases where the patients were taken off their calcineurin inhibitor; however, in 3 of those cases the patients were then put on antilymphocyte antibody treatments. Three of the 5 patients in the remaining cases recovered. The case described by Lehner et al.,11 illustrates the confusion that surrounds whether to increase or decrease immunosuppression. In their patient, GVHD was first treated by starting ATG, 100 mg/day for 5 days.11 With this treatment, the patient exhibited a decrease in donor lymphocytes but developed severe infections including pneumonia. The authors also state that “in consideration of the precarious situation of the patient” the immunosuppression was withdrawn for 8 days, after which GVHD improved.
What is novel about our approach in these 3 cases is that it represents a prospective decision that the initial treatment of all patients with GVHD should be to withdraw immunosuppression. This is coupled with efforts to make earlier diagnosis of GVHD by aggressive use of HLA typing and STR markers. This approach worked in 2 of our 3 cases but did not work in 1 patient with early onset disease and high levels of donor chimerism.
The patient in Case 1 demonstrated rapid rejection of donor CD3+ cells (Figure 3) and mild liver allograft rejection, but developed a relapse after immunosuppression was restarted to treat the rejection. We may have treated the rejection too aggressively and quickly, not allowing complete destruction of the allograft donor lymphocytes. The second rejection was gently treated with a low level immunosuppression, allowing a complete rejection of donor lymphocytes. This patient presented primarily with a skin rash, and although the histopathologic features of a skin rash can be very suggestive of GVHD,12 they are not sufficient to rule out other causes. In this case, the diagnosis was firmly established by finding both persistent donor lymphocyte chimerism in the peripheral blood and infiltrating cells in the skin biopsy of donor origin. It should be noted that the donor was homozygous for the most common Caucasian HLA haplotype A1 B8 DR17 DQ2 and the recipient shared this haplotype. This 1-way histocompatibility was undoubtedly a risk factor for his development of GVHD.
The diagnosis of GVHD in the patient in Case 2 was not clear cut, because he did not develop a skin rash or diarrhea. However, the persistence of donor cells at >10% for 18 weeks after transplant was so unusual that it supports the diagnosis of GVHD. In our previous series of 11 cases of GVHD following liver transplantation, 2 cases presented with neutropenia and fever without a skin rash or diarrhea. In those cases, the patients' pancytopenia became profound, and both patients died of sepsis. Therefore, we felt it was important to treat this patient early. In addition, there was no other explanation for neutropenia; there was no evidence of a viral illness, and the patient was not receiving any drugs known to have a significant risk of neutropenia. Thus, we strongly suspected this was a case of GVHD.
The patient in Case 3 did not respond to the withdrawal of immunosuppression. Bone marrow failure developed, which resulted in overwhelming sepsis and death. A bone marrow transplant was considered, but the patient became septic and was unsuitable for BMT. The donor was more than 40 years younger than the recipient, which we have previously identified as a risk factor for the development of GVHD.1
There are 3 important differences between patients in Case 1 and 2 and the patient in Case 3. The presentation of acute GVHD occurred early in Case 3, during the 2nd week after transplantation, versus onset of GVHD at 2 months and 4 months after transplantation in Cases 1 and 2, respectively. Also, in Cases 1 and 2 the patients were treated for an episode of acute graft rejection. It is possible the late presentation of acute GVHD and the acute rejection episode were signs of the partial control of engraftment by the patient's immune system, which made it more likely withdrawal of immune suppression would be effective. It is possible these are also signs that these patients might have recovered without withdrawal of immunosuppression; however, such spontaneous recovery has been relatively rare. The third difference is that the patient in Case 3 had a larger percentage of donor cells (27%) at the time withdrawal of immunosuppression was started than did the patients in Cases 1 and 2. The case described by Lehner et al.11 is similar to the 2 cases we describe that responded well to withdrawal of immunosuppression in that the patient had a relatively late onset of GVHD (POD 32) and relatively low levels of donor chimerism (7.5%). We suggest that these patients may respond more favorably than those with early onset and high levels of donor chimerism.
We found the STR testing of CD3+ donor cells was useful and reflected the clinical response to the treatment. STR testing made it possible to see that the patient in Case 3 failed to respond to withdrawal of immunosuppression. This allowed us to alter the therapeutic strategy and initiate Thymoglobulin administration. The GVHD responded well to Thymoglobulin despite the profound degree of neutropenia and bone marrow aplasia but, as has been our previous experience, the patient developed severe sepsis, resulting in his death. Thus, when reduced immunosuppression does not result in donor cell clearance, increased immunosuppression may allow time for donor cell engraftment of the marrow (which is rare) or to initiate other therapeutic approaches such as BMT or immunotherapy.
The treatment of acute GVHD following liver transplantation continues to be a therapeutic challenge. There are conflicting opinions on whether to increase or decrease immunosuppression in the face of GVHD following liver transplantation. In these 3 case reports, we chose to withdraw immunosuppression upon diagnosis of GVHD and allow the patient's immune system to reject the allograft donor T cells (the effector cells for GVHD). The treatment response was followed closely by clinical features as well as STR testing for CD3+ donor lymphocytes. These case reports suggest this approach may be successful for some patients, in particular those with late-onset GVHD (>8 weeks) or low levels of donor cells (<20%). Larger, possibly multicenter studies are needed to collect enough data to reach firm conclusions regarding the most effective treatment for this devastating complication of liver transplantation.