Acute Graft Versus Host Disease Following Liver Transplantation: The Enemy Within

Authors


*Corresponding author: J. Andrew Bradley, jab52@cam.ac.uk

Abstract

This article reviews acute graft vs. host disease (GVHD) as a complication of orthotopic liver transplantation (OLT). The incidence, presentation, clinical course and outcome of GVHD after OLT are summarized and the pathogenesis is discussed, drawing parallels with GVHD after allogeneic haematopoietic stem cell transplantation. Risk factors for GVHD after OLT are examined and the potential role of donor lymphocyte macrochimerism in the recipient peripheral blood as a diagnostic aid for GVHD is discussed. Finally, treatment of GVHD after OLT is reviewed with particular emphasis on the potential role of some of the newer biological agents.

Introduction

Graft vs. host disease (GVHD) following orthotopic liver transplantation (OLT) is an uncommon complication but has a high mortality and poses a major diagnostic and therapeutic challenge. Cellular GVHD occurs when immunocompetent donor lymphocytes originating from the transplanted liver undergo activation and clonal expansion, allowing them to mount a destructive cellular immune response against recipient tissues. Although humoral GVHD is commonly seen after an ABO-mismatched liver transplant when donor-derived lymphocytes produce antibodies to red-cell antigens, it usually causes only mild and self-limiting haemolytic anaemia of little clinical importance (1). Cellular GVHD (hereafter referred to as GVHD), on the other hand, is directed against MHC (and possibly minor HC antigens) and often results in severe multisystem disease with a high mortality. The aim of this review was to provide an update on GVHD after OLT, focusing on the clinical aspects and pathogenesis of the condition, with a view to improving early recognition and management. Where appropriate, similarities and differences between GVHD after OLT and GVHD after allogeneic bone marrow or stem-cell transplantation (SCT) are highlighted.

Clinical Features of GVHD After OLT

Incidence

Graft vs. host disease after allogeneic SCT has been recognized as a common and serious clinical problem for more than three decades but it was not until 1988 that acute GVHD after OLT was first described (2). The patient, a 51-year-old male, was treated with a combination of antithymocyte globulin (ATG) and high-dose corticosteroids, and very fortuitously, in the light of subsequent reports, made a good recovery (2). There are now more than 30 case reports and two recently published series of GVHD after OLT in adults. The true incidence of GVHD after OLT, however, remains uncertain. From the two reported series, one describing 13 cases of GVHD from a total of 1082 patients after OLT (1.2%) and the other seven patients from a total series of 453 patients (1.5%), the incidence of GVHD after adult OLT can be estimated at approximately 1–2% (3,4). This estimate is considerably higher than that reported to the United Network for Organ Sharing (UNOS). UNOS registry data collected on 4472 children and 32 481 adults who underwent OLT between 1995 and 2002 reports an incidence of GVHD of only 0.1% in both groups (personal communication, UNOS registry, September 2003). A possible explanation for the low incidence of GVHD in the UNOS registry is that GVHD is under-reported and that symptoms of GVHD may be incorrectly attributed to other causes such as infection and adverse responses to drugs.

Presentation of GVHD

The symptoms of GVHD after OLT usually become apparent between 1 and 8 weeks after transplantation, often after an initial uneventful recovery from surgery and discharge from hospital (3). In both children and adults the presenting symptom of GVHD after OLT is usually an isolated fever or skin rash. In approximately 15% of reported cases the disease remains confined to the skin but in most patients GVHD rapidly progresses to become a multisystem disease involving the skin, gastrointestinal tract and haematopoietic tissues (4).

The skin rash has a predilection for the palms and soles and is initially maculopapular but may progress to bullae formation and desquamation. The most common gastrointestinal symptom is diarrhoea secondary to loss of absorptive function caused by lymphocyte infiltration and destruction of the intestinal mucosa. Unlike GVHD after allogeneic SCT, where the biliary epithelium is a major target resulting in impaired liver function, the transplanted liver is not a target for GVHD, as both it and the immunocompetent cells responsible for GVHD are of donor origin. Marked neutropaenia and thrombocytopaenia often occur several days after the initial presentation of GVHD and lead to life-threatening infection and haemorrhage.

Outcome

Although there are several case reports of adult patients recovering from GVHD after OLT (2,3,5–13), the reported mortality rate from the only two published series is very high at approximately 85% (3,4). The outcome is closely related to the pattern of GVHD. All of the seven reported adult cases where GVHD was confined to the skin recovered with treatment, whereas only six of the 44 (14%) reported adult cases with multisystem GVHD survived (Table 1). Death was usually attributable to overwhelming sepsis and often followed several weeks of supportive therapy. The outcome of GVHD is particularly poor for those patients who present with fever and 29 of the 30 (97%) reported adult cases died following presentation with fever (Table 1). The high mortality from GVHD after liver transplantation is comparable to that seen in patients who develop severe multisystem GVHD after allogeneic SCT (14). Recipient age does not appear to influence the outcome of adults with GVHD and the mean age of those who survive is the same as those with fatal disease (Table 1). It has been suggested that the outcome of GVHD after OLT in children may be better than in adults (15) and five of the eight reported cases of GVHD in children survived (15–17). However, a tendency towards preferential reporting of successful management of GVHD in the literature may obscure the true mortality.

Table 1.  Comparison of variables between patients surviving and those with a fatal outcome from graft vs. host disease after orthotopic liver transplantation in adults

Variable
Survivors
n = 13a
Fatal outcome
n = 38b

p-value*
  1. 1Data not available in three cases.

  2. 2Data not available in 10 cases.

  3. 3Data not available in 12 cases.

  4. *Chi-squared test.

  5. aData compiled from references (2–5,8–13).

  6. bData compiled from references (5,9,10,18–20,31,36,37,67–74). NS = not significant.

Mean age of recipient (years)48.7148.22NS
Skin rash alone7/13 (54%)0/38 (0%)0.001
Fever at presentation1/13 (8%)29/38 (76%)0.001
Subsequent fever2/13 (15%)33/38 (87%)0.001
Pancytopaenia at presentation2/13 (15%)14/38 (37%)NS
Subsequent pancytopaenia8/13 (62%)30/38 (79%)NS
Interval from presentation to treatment, median (range)12 days (3–37d)8 days (1–27 d)NS
Day of presentation post OLT, median (range)20 days (12–114 d)24 days (7–115 d)NS
Acute rejection episode before GVHD3/13 (23%)5/26 (19%)3NS

In those patients fortunate enough to survive the initial phase of acute GVHD after OLT, the disease may resolve completely (2,5,8) or it may take a protracted course. Symptoms of skin rash and diarrhoea typically continue in a relapsing and remitting pattern and the eventual outcome is usually sepsis and death (18–20). This latter manifestation of GVHD is probably best regarded as ‘protracted acute GVHD’ as the symptoms differ from those of chronic GVHD. Chronic GVHD is a well-recognized and distinct syndrome that affects 25–50% of patients who have undergone allogeneic SCT (14) but has only been reported in two patients following OLT (11,21). It is characterized by fibrosis in the skin and subcutaneous tissues causing contractures and scarring alopecia, and there is often involvement of the salivary and lachrymal glands producing dry eyes and mouth similar to that seen in autoimmune disease. Affected patients are prone to infection, and may eventually succumb to sepsis. The two reported cases of chronic GVHD after OLT both followed living-related liver donation from HLA homozygous parental donors. The patients developed persistent alopecia and chronic skin disease and in one case symptoms were still present 4 years after transplantation (11,21).

Pathogenesis of GVHD After OLT

The essential requirements for the development of GVHD were defined almost four decades ago, based on observations made in lethally irradiated rodents reconstituted with allogeneic stem cells (22). First, the graft must contain immunologically competent cells; second, the recipient must be recognized as foreign by the graft; and third, the recipient must be unable to reject the graft before it mounts an effective immune response. These basic requirements are equally applicable to GVHD after OLT and it is helpful to bear them in mind when conceptualizing treatment of GVHD. However, it should also be noted that the situation in HLA-mismatched cell and organ transplantation is often more complex and both graft rejection and GVHD may occur simultaneously. Recent studies have provided many additional new insights into the pathophysiology of GVHD, both in terms of the interactions between alloreactive T cells and host alloantigens and with respect to the profound dysregulation of the cytokine network that characterizes GVHD. The findings from such studies have emphasized the complexity of the GVHD syndrome and highlighted the importance of interactions between innate and adaptive immunity in the development of GVHD. They have also been important in identifying new strategies for treatment.

The pathogenesis of acute GVHD after allogeneic SCT can be conceptualized by a three-phase model that is also useful as a framework for considering the pathogenesis of GVHD after OLT (Figure 1; 23).

Figure 1.

Vicious circle of graft vs. host disease (GVHD). Tissue damage leads to production of pro-inflammatory cytokines augmented by intestinal translocation of LPS, which stimulates further production of pro-inflammatory cytokines from monocytes. In addition to causing further injury to local tissues, pro-inflammatory cytokines enhance recipient antigen-presenting cells (APCs) activity, promote development of donor alloreactive Th1 cells and activate nonspecific effector cells (NK cells and macrophages). The net result is amplification of target cell injury.

The 3-phase model: phase I

Phase I comprises the preparative regime (chemotherapy and/or radiotherapy), which is an integral component of SCT and is toxic to many tissues, particularly the gastro-intestinal tract. Pro-inflammatory cytokines such as TNF-α and IL-1 are released by recipient macrophages and epithelial cells in response to direct tissue injury or to translocation of bacterial products (e.g. lipopolysaccharide) resulting from mucosal injury of the intestine (24,25). The production of such pro-inflammatory mediators greatly facilitates the development and profoundly affects the course of acute GVHD. Although patients undergoing OLT do not usually receive preparative treatment of the type used in SCT, the surgical procedure and other factors that may be present in the recipient, such as infection, may give rise to a systemic or local inflammatory response. This may in turn promote the development of GVHD through release of inflammatory mediators that heighten the activity of host antigen-presenting cells (APCs) and promote proliferation of alloreactive donor T cells.

The 3-phase model: phase II

The second phase of the three-phase model is activation and proliferation of alloreactive donor T cells. The development of GVHD after organ transplantation is largely dependent on the number of immunocompetent T cells transferred within the graft. Graft vs. host disease is well recognized after transplantation of the liver and small bowel, where both grafts contain substantial numbers of lymphocytes, but it is very rare after transplantation of solid organs that are not associated with large amounts of lymphoid tissue, such as the heart and kidney (26). It is estimated that between 109 and 1010 donor lymphocytes remain in the portal tracts and the parenchyma of a liver graft after flushing with cold preservation solution (27). In addition, variable amounts of lymphoid tissue located around the porta hepatis are inevitably transferred along with the donor liver. The total lymphocyte load transferred with a liver graft is comparable in magnitude to that typically administered during SCT (14), although the impact of cold ischaemia on the overall functional ability of the hepatic lymphocyte population is not clear.

The molecular events that occur when donor alloreactive T cells interact with recipient APCs are well characterized, and include engagement of the αβ T-cell receptor complex with HLA/peptide, together with the interactions between multiple costimulatory molecules including CD40–CD154 and CD28–B7 (28). After activation, donor T cells express the high-affinity IL-2 receptor and, under the influence of regulatory cytokines such as IL-12, IFN-γ and IL-18, undergo IL-2-dependent clonal expansion (14). They differentiate predominantly along the Th1 pathway into effector T cells that secrete pro-inflammatory cytokines, including IL-2 and IFN-γ (23).

The 3-phase model: phase III

The third phase of GVHD is the effector phase. In mouse models, the development of acute GVHD is associated with a strongly polarized Th1 response characterized by heightened NK cell activity and antihost cytotoxic T lymphocytes (CTLs), whereas chronic GVHD is associated with a polarized Th2 response and autoantibody production (29,30). Acute GVHD in humans is also attributed predominantly to a Th1 response, although to regard it as a simple Th1 polarized response is an oversimplification (23). Cytotoxic T lymphocytes induce apoptosis in target cells through release of perforin or granzyme or through FasL–Fas interaction. Non-specific effector cells such as macrophages and NK cells recruited and activated by pro-inflammatory cytokines released from effector T cells or in response to lipopolysaccharide (LPS) and other inflammatory mediators also contribute to and amplify the tissue damage (23).

The net result is a vicious circle in which tissue damage caused by effector cells leads to further release of inflammatory mediators and cytokines and to increased activity of host APCs, which in turn heightens further the destructive alloreactive T-cell response.

Distinct phenotype of hepatic T cells

The phenotype of hepatic T cells differs markedly from that of peripheral T cells. Hepatic T cells have a CD4:CD8 ratio of 1:3.5 (vs. 2:1 in peripheral blood) and the percentage of hepatic CD3+ cells expressing γδ receptors or NK cell markers is very high (7–54% vs. <6% in peripheral blood) (27). In addition to mature lymphoid cells, the donor liver may contain stem cells with the potential for haematopoiesis and it is interesting to note that donor-derived haematopoiesis has been reported after OLT in two adult recipients with GVHD (3,31). However, in both cases the haematopoiesis observed was probably a result of rather than the cause of GVHD, as any alloreactive immature T cells of donor origin are likely to have been deleted in the recipient thymus.

To what extent the distinctive phenotype of the hepatic T-cell population impacts on the development of GVHD after OLT is not clear. It is perhaps surprising, given the large number of lymphocytes residing in the liver, that GVHD is not seen more frequently after OLT. This may, in part, reflect the functional effects of hepatic lymphocytes, as γδ T lymphocytes produce predominantly Th2 cytokines, which down-regulate Th1 responses and suppress GVHD (32). The rarity of chronic GVHD after OLT is also intriguing and cannot as yet be explained. It may be that recovery from acute GVHD, which occurs in a small minority of patients, is dependent on rejection of donor cells by recipient T cells, in which case chronic disease would not be anticipated.

Risk Factors in the Development of GVHD

The only two risk factors that have been clearly identified in the development of GVHD after OLT are the HLA match between donor and recipient and recipient age.

HLA

Although HLA matching is not undertaken for cadaveric OLT and does not improve graft survival, those recipients who by chance receive a well-matched graft appear to be at increased risk of GVHD (3,17,21). This may be because alloreactive hepatic lymphocytes from a well-matched graft are less likely to be rapidly destroyed in the recipient. In animal models of graft resistance following bone marrow transplantation, allogeneic lymphocytes are eliminated by the process of natural cytotoxicity, which is mediated predominantly by NK cells. Natural killer cells kill allogeneic cells unless they express appropriate HLA class I determinants which interact with inhibitory receptors (killer-cell immunoglobulin-like receptors or KIR) on the NK cell surface (33). If donor lymphocytes express shared HLA class I determinants with the recipient that engage effectively with KIR, they are protected from elimination and are able to populate the recipient and initiate GVHD. Unlike SCT, there is no preparative regime before OLT and the recipient T-cell-mediated response may be important in the elimination of allogeneic lymphocytes. In experimental animals GVHD is readily achieved by injecting parental lymphocytes into F1 recipients. Parental cells are not recognized as foreign by the recipient but they recognize and respond to recipient alloantigens. This situation is mirrored in living-donor OLT when a recipient receives a HLA-homozygous liver from one of their parents and it is associated with a particularly high risk of acute GVHD (21).

In a retrospective analysis of more than a thousand patients who had undergone OLT in a single centre, Smith et al. identified close HLA matching as a significant risk factor in the 12 patients who developed GVHD and concluded that multiple HLA class I mismatches protected against GVHD (3). In a retrospective analysis of 412 patients undergoing OLT in our own centre, the risk of GVHD in patients receiving a liver graft with 3–4 antigen mismatches at HLA-A and -B was 1%. In patients receiving a graft with only 0–1 antigen mismatches at HLA-A and -B, the risk of GVHD rose to 7.4%, and further increased to 12.5% if 0–1 DR mismatches were also present (Taylor et al., Cambridge, UK, September 2003, unpublished data.)

Recipient age

Recipient age is also a definite risk factor for developing GVHD after OLT in adults and this is in keeping with acute GVHD after allogeneic SCT where recipients older than 40 years are known to be at increased risk of severe GVHD (34). The two reported series of GVHD after adult OLT both suggest that GVHD is more common in older recipients (3,4). Smith et al. observed that recipients who were older than 65 years at the time of OLT were nine times more likely to develop GVHD than those aged less than 65 years. They also suggested that recipients of livers from donors more than 40 years younger were at greater risk of GVHD (3). The mean age of patients developing GVHD after OLT in our own centre was 59.3 years, compared with 48.1 years in patients who did not develop GVHD (p < 0.01; 4). The reason for the association between recipient age and GVHD after OLT is not clear but there are several potential explanations. The function of the immune system declines with age and this may impair the ability of the recipient to control the expansion of alloreactive T lymphocytes of donor origin. Colonization of the gastrointestinal tract with bacterial and viral pathogens may be increased in older recipients and this may prime donor lymphocytes that cross-react with target tissues in GVHD. Tissue repair slows with age, and intestinal GVHD may more readily lead to bacterial translocation and amplification of the GVHD response. Finally, recent evidence suggests that APCs from aged recipients stimulate a greater response in allogeneic cells (35). Interestingly, although older recipients are at increased risk of GVHD, the incidence of GVHD in children and adults reported to the UNOS registry is similar. The explanation for this apparent paradox is not clear and it is difficult to draw any firm conclusions, as the degree of underdiagnosis is not known in children.

Recipient immunocompromise

Recipient immune compromise before OLT has also been suggested as a risk factor for the development of GVHD (5), but there is no firm evidence for this in the two published series (3,4). In the series of seven patients with GVHD after OLT from our centre, five were transplanted for alcoholic cirrhosis (with or without hepatocellular carcinoma), one patient was retrospectively found to have bacterial septicaemia at the time of transplantation, and one patient was taking long-term azathioprine before OLT (4). Smith et al. found that cryptogenic cirrhosis was the only diagnosis encountered more frequently among patients with GVHD, though this series of 13 patients included a case of severe and rapidly fatal GVHD after OLT in a 15-year-old recipient with underlying combined immunodeficiency (3). Several of the published case reports of GVHD after OLT describe an episode of acute rejection before the onset of GVHD (2,7,10,11,36), arguing against the suggestion that recipient immunosuppression is a prerequisite for GVHD after OLT.

Macrochimerism in the Diagnosis of GVHD After OLT

The accurate diagnosis of acute GVHD after OLT has been hampered by the lack of a sensitive and specific diagnostic test. A skin biopsy showing epidermal dyskeratosis with epithelial cell necrosis is highly suggestive, but not pathognomonic. The first case report of GVHD after OLT identified the presence of cells expressing donor antigens in the peripheral blood and bone-marrow and commented on the identification of chimerism as an early diagnostic aid (2). However, although subsequent reports noted macrochimerism (where donor cells comprise >1% of circulating nucleated cells in the peripheral blood) in many patients with symptomatic GVHD, this was not a consistent finding (37,38) and the significance of microchimerism (where donor cells comprise <1% of circulating cells) is unclear (26). We evaluated the diagnostic potential of a low-sensitivity PCR-SSP technique and flow cytometry to detect peripheral blood chimerism in 33 OLT recipients who presented with symptoms suggestive of GVHD (4). Seven patients were found to have macrochimerism (with levels ranging from 4 to 50%) at a median of 5 weeks post OLT (range 2–8 weeks) and in all the diagnosis of GVHD was subsequently confirmed by histology or clinical course. Five of the seven patients with macrochimerism died. Twenty-six of the 33 patients did not have peripheral blood macrochimerism despite having symptoms consistent with GVHD; in all cases an alternative diagnosis was eventually established or recovery was rapid and spontaneous. Detection of macrochimerism may, in addition to confirming the diagnosis of established GVHD, be of value in the diagnosis of asymptomatic GVHD. Asymptomatic macrochimerism (25%) was reported by Joysey et al. in a patient who developed symptoms of GVHD 3 weeks later (10). However, it is not known if macrochimerism commonly precedes symptomatic GVHD.

Although macrochimerism is a useful diagnostic aid for patients with symptoms suggestive of GVHD, it is important to note that macrochimerism appears transiently in the majority of patients in the early postoperative period after OLT. Two prospective studies of donor lymphocyte chimerism after transplantation, one of 11 patients and one of 16 patients, showed that the presence of donor lymphocytes in the circulation peaks within the first week after transplantation when they comprise up to 24% of peripheral blood lymphocytes (9,13). Macrochimerism declines rapidly thereafter and is usually absent beyond 3–4 weeks post transplant.

Prevention of GVHD

Donor selection

Some transplant centres undertaking living donor OLT do not proceed if a potential parental donor is homozygous at all HLA loci because of the high risk of GVHD in this situation (21). Although GVHD is more common after cadaveric OLT when the donor and recipient have a close HLA match, it is not practicable to routinely allocate livers on the basis of HLA mismatch.

Depletion of donor lymphocytes

Complete depletion of donor T cells from allogeneic peripheral blood SCT eliminates the risk of GVHD (39,40) but is not desirable because it impairs engraftment and increases the relapse rate when SCT is performed for haematological malignancy. Depletion of T lymphocytes from the liver before transplantation would also eliminate the risk of GVHD. This could be achieved, at least in principal, by treating the cadaveric donor with antilymphocyte globulin (ALG) or by modifying the donor liver ex vivo either by irradiation or by perfusing it with lytic monoclonal antibodies directed against a lymphocyte cell-surface protein (41–43). However, whether these approaches can be justified is debatable, given the low incidence of GVHD after OLT. It is also conceivable that eliminating donor lymphocytes from the liver might have an adverse effect on graft survival (44). Recipients with the lowest levels of macrochimerism in the early post transplant period have a higher rate of acute rejection (9).

Reduced immunosuppresion

In recipients thought to be at particularly high risk of GVHD after OLT, either through close HLA matching or because of age-related risk factors, it has been suggested that using a less intensive immunosuppressive protocol (e.g. steroids and azathioprine only) combined with increased surveillance to detect acute rejection may be advantageous (3). The rationale for this approach is that in the absence of calcineurin blockade the recipient immune system is better able to reject donor lymphocytes, thereby preventing GVHD. Protecting high-risk patients from the serious consequences of GVHD, it is argued, justifies the substantial increase in acute rejection that is an inevitable consequence of this policy (3) Improved identification of those recipients at highest risk from GVHD would clearly lend weight to this approach.

Treatment of GVHD

The evidence base for selecting the most appropriate therapy for established acute GVHD after OLT is very limited and treatment is therefore largely empirical, although the extensive literature on the management of acute GVHD after SCT provides guidance. Surprisingly, the available data suggest that early treatment of GVHD after OLT may not necessarily affect the eventual outcome. Analysis of reported cases shows that the median time interval between presentation of GVHD and initiation of treatment is similar in survivors and non-survivors (12 and 8 days, respectively, Table 1).

Corticosteroids

First-line therapy for acute GVHD after SCT comprises corticosteroids and in approximately half of cases this produces long-term remission (23,45,46). The efficacy of corticosteroids in this setting is likely to be a result of their potent anti-inflammatory property as well as their lympholytic and immunosuppressive effects. There is general agreement that high-dose corticosteroids are also an important part of first-line treatment for acute GVHD after OLT (3,4).

Unfortunately, however, GVHD after OLT is less responsive to corticosteroids than GVHD after SCT. Most patients are therefore given additional immunosuppression, usually in the form of antibody preparations such as ALG, antithymocyte globulin (ATG) or OKT3, but there is no evidence that these agents alter the eventual outcome (2,5,10,31,36,37). This situation is similar to that for corticosteroid resistant GVHD after SCT where a durable response is rarely seen with second-line therapy, and for patients given ATG or OKT3 survival is less than 20% at 6 months (47–51). Administration of OKT3 results in a self-limited release of pro-inflammatory cytokines (including TNFα and IFNγ) in association with T-cell lysis. Because the pro-inflammatory cytokines produced may potentiate GVHD the use of OKT3 to treat GVHD after OLT cannot be recommended.

Interleukin-2 receptor antibodies

Monoclonal antibodies (mAbs) directed against CD25 (basiliximab and daclizumab) have recently been used to treat patients with GVHD after both SCT and OLT (6,8,52–54). The clonal expansion of host reactive CTL during the development of GVHD is dependent on IL-2 and targeting CD25 (IL-2Rα, the high-affinity IL-2 receptor), which is expressed predominantly on activated T cells (55,56), is conceptually attractive. There are three reported cases where adult GVHD after OLT was treated with anti-CD25 mAbs and in all GVHD resolved, despite the presence of severe multisystem GVHD at presentation (6,8). In GVHD after SCT the response rate following administration of anti-CD25 mAbs in corticosteroid-resistant disease is generally 40–50% though this is offset by the significantly increased risk of life-threatening infectious complications requiring intensive therapy (52–54).

Anti-TNFα (monoclonal antibody)

The central role played by TNFα in the pathogenesis of GVHD has made it a logical therapeutic target. Infliximab, an anti-TNFα mAb that neutralizes soluble TNFα and induces lysis of cells responsible for its production, has recently been used to treat patients with corticosteroid-resistant GVHD after SCT. In a series of four cases, infliximab led to complete resolution of gastrointestinal disease and significant improvement in skin and liver disease in three patients (57). However, in a larger series of 11 cases, the response to infliximab was poor and only two patients survived (18%) (58). This may to some extent reflect differences in the dose and timing of administration and further controlled trials are required before anti-TNFα can be recommended for the treatment of GVHD after OLT.

Reduction in immunosuppressive therapy

Reduction and even complete withdrawal of immunosuppression has been proposed as a treatment for GVHD after OLT, on the basis that it may allow the recipient immune system to reject alloreactive donor lymphocytes more effectively (7,15,16). Reducing immunosuppression poses the risk of concurrent graft rejection (7). Moreover, freeing alloreactive donor T cells from the constraints of immunosuppression may paradoxically lead to a worsening of GVHD and currently there is insufficient evidence to support this approach.

Control of infection

The importance of preventing and controlling intercurrent infection in acute GVHD after SCT is well recognized (14,59), and the same principles should be applied for the management of GVHD after OLT. Broad spectrum antibacterial and antifungal prophylaxis should be administered early in the disease process and consideration should be given to CMV prophylaxis, as GVHD is often accompanied by CMV infection (60,61). Most cases of GVHD after OLT develop pancytopaenia, which contributes to overwhelming infection and death. Growth factors, notably granulocyte colony stimulating factor (GCSF), have been administered in several reported cases of severe GVHD after OLT and led to substantial recovery of neutrophil counts (3). Administration of appropriate recombinant haematopoietic growth factors to treat profound neutropaenia appears a reasonable approach although it is not clear whether it influences eventual outcome.

Future prospects

Recent insights into the pathogenesis of GVHD after SCT have provided several new potential therapeutic targets. Biological agents that deplete or prevent activation of alloreactive T cells may have a role and mAbs or fusion proteins that block T-cell costimulatory activity have been shown to be effective at preventing the development of acute GVHD in animal models although they are less effective at treating established disease (62,63). In view of the importance of lipopolysaccharide (LPS) in initiating and potentiating GVHD, strategies that prevent LPS ligand/receptor interaction have stimulated much interest. Synthetic LPS antagonists have been shown to reduce GVHD in animal models and are being developed for clinical trial (64). Interleukin-12 is another potential target molecule. It is a key regulatory cytokine in the promotion of Th1 responses and in mouse models of acute GVHD; neutralizing IL-12 during the induction of acute GVHD prevents early mortality although it may produce chronic autoimmune disease (65). Finally, administration of keratinocyte growth factor (an epithelial tissue repair factor) reduces the severity of GVHD in mouse models by protecting against the intestinal mucosal damage that is important in the early pathogenesis of the disease (66).

Conclusion

Graft vs. host disease after OLT is an uncommon but serious complication and merits a high index of clinical suspicion, especially in older recipients and those with well-matched grafts. Confirmed GVHD should be treated promptly with corticosteroids and possibly anti-CD25 mAbs along with chemoprophylaxis to prevent infection. The prognosis remains poor, but better understanding of the pathophysiology of GVHD has provided a number of new therapeutic targets and new agents are currently entering clinical trials to assess their efficacy in treating GVHD after SCT.

Acknowledgement

The authors are grateful to Dr Michael Cecka at the UNOS for kindly providing data on the incidence of GVHD after OLT and Sarah Taranto.

Ancillary