Acute Graft-Versus-Host Disease After Deceased-Donor Liver Transplantation: A Case Report

Introduction

Acute graft-versus-host disease (GVHD) is most commonly associated with hematopoietic stem cell transplantation but has also been reported, although infrequently, following solid organ transplantation, particularly liver transplantation (LT). In LT, GVHD occurs when donor-derived immunocompetent lymphocytes mount an immune response against host tissues. While the incidence of GVHD after LT is estimated at 0.1–2%, with mortality exceeding 70% according to large-scale analyses [1,2], early descriptions of this phenomenon date back to the seminal report by Burdick in 1988 following orthotopic LT [3], which helped establish GVHD as a recognizable LT complication and shaped subsequent efforts toward its diagnosis and treatment.

According to Billingham’s criteria, GVHD requires the presence of immunocompetent donor T cells, a histocompatibility mismatch, and impaired recipient immunity, which are conditions often encountered in LT [4,5].

GVHD after LT (LT-GVHD) is believed to be underdiagnosed, with clinical features that overlap with those of drug reactions or viral infections. In a large series by Smith et al, the incidence of GVHD was approximately 1%, with a mortality rate exceeding 90% despite aggressive therapy [6]. Risk factors included close human leukocyte antigen (HLA) matching, older recipient age, and donor-recipient age differences greater than 40 years. Most patients develop symptoms such as rash, fever, and pancytopenia within 2–6 weeks, and donor lymphoid chimerism has been demonstrated to be a diagnostic hallmark. The authors emphasized that strategies aimed at minimizing risk, such as avoiding closely matched HLA donors or tailoring immunosuppression in high-risk settings, can improve outcomes [6].

Multiple risk factors for LT-GVHD have been proposed, including advanced donor age, HLA disparity, and immunocompromised recipient status. Clinical manifestations typically involve the skin and gastrointestinal tract and hematopoietic suppression, often mimicking drug reactions or sepsis, leading to diagnostic delays. Diagnostic modalities include skin biopsy and peripheral blood short tandem repeat (STR)-based chimerism analysis, both of which have been validated as early tools for GVHD confirmation [7,8]. STR testing can also help distinguish between donor and recipient origins in cases of cytopenia [9].

GVHD presents a wide spectrum of manifestations after LT. Although the classic form involves simultaneous injuries to the skin, gastrointestinal tract, and liver, an isolated hepatic variant with predominant hepatic inflammation and minimal extrahepatic symptoms has been described. Histologically, bile duct epithelial injury is a hallmark feature of hepatic GVHD and includes nuclear pleomorphism, cytoplasmic vacuolization, and apoptosis. Ductopenia and portal fibrosis can occur in more advanced cases. Although liver involvement is usually accompanied by GVHD in other organs, hepatic variants may occur in isolation, complicating timely recognition. These distinctions are critical because histological overlap with drug-induced liver injury or infection often necessitates careful clinicopathologic correlation [10].

Most documented cases show symptom onset at 3 to 6 weeks after transplantation, with a median of 30 days (interquartile range, 21–42 days) [1]. In this report, we describe a patient in whom GVHD was histologically diagnosed on postoperative day (POD) 20, earlier than in most previously reported cases. This unusually early presentation underscores the importance of heightened clinical suspicion for GVHD as early as the third postoperative week, even before the classical triad of symptoms is fully established.

Furthermore, culture-negative sepsis has emerged as a leading cause of death in GVHD following LT. Murali et al reported that despite aggressive antimicrobial therapy, most patients with fatal GVHD have no identifiable pathogens, implicating immune dysregulation as a central driver of systemic failure [11].

We report this case not only because of the early histologic confirmation of GVHD on POD 20, one of the earliest described cases after deceased-donor LT (DDLT), but also because of the diagnostic clarity provided by skin biopsy, the absence of infection, and rapid progression despite timely corticosteroid therapy. This case provides valuable insights into the challenges of recognizing steroid-refractory GVHD early in the post-transplantation period and reinforces the urgency of developing more effective therapeutic strategies. This case is notable for the exceptionally early onset of biopsy-proven GVHD on postoperative day 20, highlighting the diagnostic challenge and the need for heightened clinical suspicion even during the early postoperative period.

Case Report

The patient was a 48-year-old Asian woman with end-stage liver disease due to primary biliary cholangitis diagnosed in January 2010 by liver biopsy. At that time, she presented with elevated liver enzyme levels and hyperbilirubinemia following the use of herbal medications after childbirth. Preoperative infection screening ruled out hepatitis B, C, and HIV infections. Herpes simplex virus (HSV) immunoglobulin M (IgM) and rubella IgG/IgM tests were negative. Anti-HAV (total) was positive, and anti-HAV IgM was negative, consistent with past exposure or immunization.

Serological testing revealed positive results for antimitochondrial and antinuclear antibodies. Autoimmune hepatitis was considered but ultimately excluded based on the simplified scoring system of the International Autoimmune Hepatitis Group, which yielded a score of 5, indicating a low likelihood of autoimmune hepatitis. She was treated with ursodeoxycholic acid, which resulted in partial biochemical improvement and continued regular hepatology follow-up until 2024.

Prior to transplantation, she remained under outpatient surveillance and maintained relatively stable liver function for several years. However, she eventually developed signs of hepatic decompensation, including refractory ascites. At the time of listing, her total bilirubin and albumin levels were 17.27 mg/dL and 3.0 g/dL, respectively, and her MELD score were 33. She was mentally alert but physically frail at the time of transplantation.

She underwent DDLT using a graft from a 38-year-old male ABO-compatible donor. At the time of donor liver procurement, frozen section analysis revealed 60% steatosis (macrovesicular: 40%; microvesicular: 20%), no portal fibrosis, and mild portal inflammation. The cold ischemia time was 3 hours 56 minutes. The graft weighed 1196 g with a recipient body weight of 47 kg, and the graft-to-recipient weight ratio was approximately 2.54%.

The inferior vena cava was intraoperatively reconstructed using the piggyback technique. End-to-end portal vein anastomosis was performed to the main portal vein of the recipient. Hepatic artery reconstruction was performed from the donor’s common hepatic artery to that of the recipient. Biliary anastomosis was performed using a duct-to-duct technique without stent placement.

Cross-matching was performed using complement-dependent cytotoxicity (CDC) assays with donor lymphocytes obtained during organ procurement. Anti-human T-cell globulin- and dithiothreitol (DTT)-treated assays were negative. Warm B-cell cross-matching was initially weakly positive (titer 1: 2), but became negative following DTT treatment, suggesting the presence of IgM-type antibodies without clinical significance. Despite positive flow cytometric cross-matching (T-cell mean fluorescence intensity (MFI) 2.94; B-cell MFI 9.94), these results were not considered clinically impactful because of the negative CDC results.

The early postoperative course was stable. The patient was extubated on POD 2 and was mentally alert. On POD 5, she was transferred from the intensive care unit to the general ward, where she resumed oral intake and was able to ambulate independently. Tacrolimus was adjusted to maintain a target trough level of approximately 8 ng/mL. By POD 15, liver function was well preserved, with total bilirubin, aspartate aminotransferase, and alanine aminotransferase levels of 1.23 mg/dL, 18 IU/L, and 39 IU/L, respectively, and international normalized ratio of 1.00.

On POD 16, the patient developed sudden-onset anemia (hemoglobin, 5.1 g/dL) without overt bleeding or hemolysis. Workup, including reticulocyte count, LDH, haptoglobin, and Coombs testing, was unremarkable, suggesting bone marrow suppression or immune-mediated cytopenia. Owing to worsening coagulopathy, a bone marrow biopsy could not be performed.

By POD 19, she developed high fever (39.2°C), generalized erythematous rash involving over 60% of the body surface area, and profuse watery diarrhea exceeding 1500 mL/day. Laboratory findings showed worsening leukopenia (white blood cell count, 1500/μL), neutropenia (absolute neutrophil count, 3.87), and thrombocytopenia (platelet count, 55 000/μL). Liver function also declined. Microbiological evaluation, including CMV/EBV polymerase chain reaction and bacterial culture, was negative. Blood cultures drawn on POD 13, 17, 22, and 26 remained negative after 5 days of incubation, consistent with culture-negative sepsis, implicating immune dysregulation.

A skin punch biopsy of the trunk on POD 20 revealed prominent vacuolar degeneration of the basal layer and scattered keratinocyte apoptosis. Superficial dermal lymphohistiocytic infiltrate was also noted, without evidence of eosinophilic infiltration or viral cytopathic changes. Given the histopathological features and clinical context, the findings favored GVHD, and differential diagnoses such as drug-induced dermatitis or erythema multiforme were considered less likely (Figure 1).

Treatment was initiated with high-dose intravenous methylprednisolone (1 g/day for 3 days), followed by tapering. Mycophenolate mofetil was discontinued because of leukopenia, and tacrolimus was adjusted to maintain trough levels of 12 to 15 ng/mL. Despite immunosuppressive intensification, there was no improvement; her fever and diarrhea persisted and the rash did not improve. These findings indicated steroid-refractory GVHD. The patient experienced progressive pancytopenia, renal dysfunction (creatine, 3.37 mg/dL), and hepatic deterioration. She ultimately developed signs of overwhelming sepsis and died of multiorgan failure on POD 30, despite aggressive supportive care.

Discussion

GVHD following LT is rare and associated with exceedingly poor outcomes. A recent meta-analysis [1] reported a median onset of 30 days after transplantation, pooled incidence of 1.2%, and overall mortality exceeding 70%. Our patient’s diagnosis on POD 20 is one of the earliest biopsy-confirmed cases of DDLT reported in the literature, emphasizing the need for early recognition. A clinical timeline summarizing the perioperative course and the onset of GVHD-related symptoms is presented in Figure 2.

This is compounded by the challenge of diagnosing sepsis in the absence of positive culture results. Recent multicenter data from Korea demonstrated that patients with culture-negative septic shock had significantly higher in-hospital mortality than those with culture-positive sepsis despite timely empirical antibiotic therapy. These findings support the notion that immune dysregulation, rather than microbial burden, is the principal driver of organ failure in GVHD-associated sepsis [12].

This case highlights several important considerations. First, isolated unexplained anemia preceding classical GVHD symptoms can be an early warning sign of marrow involvement. Similar clinical patterns and therapeutic dilemmas have been described in other multi-case studies, including a recent report of 6 patients with LT-GVHD [13]. Second, while first-line therapy remains high-dose corticosteroids, the responses are often incomplete. In our case, the rapid progression despite early immunosuppressive modulation underscores the need for adjunctive approaches.

Emerging therapies, including JAK inhibitors, anti-thymocyte globulin, and whole-liver intensity-modulated radiation therapy (IMRT), have shown promise in selected refractory cases [14,15]. For instance, symptomatic improvement with whole-liver IMRT has been reported, although fatal liver injury occurred in 1 patient despite treatment [14].

Our patient’s rapid clinical deterioration and profound pancytopenia precluded the initiation of second-line therapies such as JAK inhibitors, anti-thymocyte globulin, or IMRT. Moreover, the lack of response to corticosteroids and the early onset of multiorgan dysfunction significantly narrowed the window for escalation to experimental modalities.

Finally, timely histopathological confirmation via skin biopsy remains critical, as emphasized in recent case-based studies [8]. In our case, biopsy provided a definitive confirmation of GVHD and guided the prompt escalation of immunosuppression despite the absence of infectious causes. The histological findings observed in our patient’s skin biopsy, such as apoptotic keratinocytes and lymphocytic interface dermatitis, paralleled the histopathological characteristics described in liver tissue affected by GVHD, reinforcing the systemic and multiorgan nature of this disease [16].

Conclusions

This case is one of the earliest confirmed presentations of acute GVHD after DDLT with diagnosis on POD 20, and shows that even subtle hematological shifts, such as isolated anemia, can result in GVHD. Prompt clinical recognition, early biopsy, and decisive immunosuppressive modulation are critical, although outcomes remain poor in disseminated diseases. This unusually early presentation underscores the urgent need to develop more effective therapeutic strategies for steroid-refractory GVHD after liver transplantation.