A 47-year-old Asian man with no history of smoking or alcohol consumption was referred due to progressive jaundice potentially necessitating liver transplantation. His symptoms initially appeared as yellowish skin and generalized itchiness in early May 2022. His medical history was unremarkable except for recent vaccinations. He had received three COVID-19 vaccines: the ChAdOx1 nCoV-19 (AstraZeneca) on July 26, 2021 and December 29, 2021, and mRNA-1273 SARS-CoV-2 (Moderna) on April 26, 2022. Notably, he observed tea-colored urine on May 10, 2022, 2 weeks after the third vaccination. No other symptoms of discomfort were reported. He sought medical attention at a local hospital, where tests revealed elevated serum total bilirubin levels (3.9 mg/dL on May 20, increasing to 32 mg/dL on May 27), abnormal liver enzyme levels (aspartate transaminase [AST] at 234 U/L; alanine transaminase [ALT] at 542 U/L on May 20), and declining renal function (creatinine [Cre] at 2.81 mg/dL on June 13). During subsequent follow-up visits, his total bilirubin and creatinine levels rose to 55 mg/dL and 3.93 mg/dL, respectively, on July 4. Due to this progressive deterioration, he was admitted to the hospital on July 4, 2022. Extensive evaluations ruled out any history of prior hepatitis A, B, or C infections, and the patient had no risk factors for autoimmune diseases, such as a positive family history, current medication use, or recent infections. Abdominal computed tomography (CT) displayed no signs of cholelithiasis, biliary tract dilatation, cirrhosis, or portal hypertension. On July 8, 2022 (day 0), he was referred to our transplant center.

Upon admission, the patient exhibited yellowish skin and sclera, but he did not have clay-colored stools. Laboratory analyses revealed cholestatic liver injury with the following results: total bilirubin of 64.2 mg/dL, ALT of 46 U/L, AST of 38 U/L, alkaline phosphatase (ALP) of 361 U/L, international normalized ratio of 0.95, and Cre of 4.8 mg/dL. The fractional excretion of urea was 54.9%, which suggested intrinsic renal disease. The patient had adequate urine output of approximately 3000 mL/day during his hospital stay. An abdominal ultrasound performed on day 4 showed no cholelithiasis, dilatation of the common bile duct, normal pancreatic parenchyma, and bilaterally enlarged kidneys (15 cm in length, reference range: 9–12 cm). Hemogram results in the early clinical course indicated leukocytosis and thrombocytosis (Fig. 1A). Laboratory data on day 6 revealed worsening kidney function, with a Cre level of 6 mg/dL and BUN level of 106 mg/dL (Fig. 1B). Infective surveys were negative for SARS-CoV-2 RNA PCR and nucleocapsid antibody, hepatitis E immunoglobulin M (IgM), anti-HIV antibody, and leptospiral antibody–IgM. Previous but not recent exposure to cytomegalovirus, Epstein-Barr virus, and herpes simplex virus-1 was noted. Autoimmune surveys conducted via immunofluorescence assay were positive for anti-smooth muscle antibody (1:40) but negative for antinuclear antibody, anti-neutrophil cytoplasmic antibody, anti-mitochondrial antibody, and anti-liver–kidney microsomal antibody. The patient’s immunoglobulin G levels were within the normal range. An inflammatory profile revealed significantly elevated ferritin level (9015.9 ng/mL; normal range: 24–336 ng/mL), slightly elevated C-reactive protein (CRP) levels (0.97 mg/dL), and an increased erythrocyte sedimentation rate of 34 mm/h (Fig. 1E). Transferrin and ceruloplasmin levels were within normal ranges. Specific antibodies, including anti-SARS-CoV-2 spike antibody and AT1R antibodies, were utilized for further mechanistic investigation (details below) (Fig. 1F).

Hemogram (A), kidney (B), liver (C), pancreas (D), inflammation (E), and antibodies (anti-SARS-CoV-2 Spike antibody & angiotensin II receptor type 1 (AT1R) antibody) (F). Day 0 is the date of admission and doses and duration of medications are at top. Improvement was noted after starting steroids and plasma exchange (D7). Hb drop was noted after liver biopsy (D5). Reference of positivity: anti-SARS-CoV-2 Spike antibody (≥0.8 U/mL, quantitative Roche Elecsys® Anti-SARS-CoV-2 S assay), angiotensin II receptor type 1 (AT1R) antibody (>17 U/mL, enzyme immuno-assay (EIA), CellTrend GmbH). Normal ranges of laboratory testes other than antibodies were presented in Supplementary Table S1. ALT alanine transaminase, AST aspartate transaminase, BUN blood urea nitrogen, Cre creatinine, CRP C-reactive protein, D-bil direct bilirubin, Hb hemoglobin, PLT platelet, T-bil total bilirubin, WBC white blood cell.

Percutaneous liver biopsies conducted on day 5 after admission revealed centrilobular cholestasis and bile duct damage, characterized by cholangiocyte vacuolation. The bile ducts were lost in some of the portal areas (Fig. 2A). These findings were consistent with obstructive cholestasis resulting from the inflammatory destruction of the intrahepatic bile ducts, a condition known as vanishing bile duct syndrome (Fig. 2A). Importantly, this pattern did not suggest an infectious etiology. Additional immunostaining unveiled positive C4d deposition in endothelial cells, as demonstrated in Fig. 2B. This finding strongly indicated antibody-mediated liver injury, resembling the pattern of antibody-mediated rejection (AMR) seen in solid organ transplantation.

Histopathological (A, B) and immunohistopathological (C) examinations of liver biopsy. A Marked centrilobular cholestasis. B Mild lymphocytic (square) and neutrophilic (circle) infiltration in the portal areas with bile duct damage and cholangiocyte vacuolation (black arrows). C Positive C4d deposition in portal venous and capillary endothelial cells (black arrows). 200X.

Initial management included supportive medications for symptom relief and short-term empirical antibiotics (doxycycline and levofloxacin, ended immediately after liver biopsy) for suspected atypical infection with multiple organ involvement in inflammation (liver, kidney, and pancreas). We initiated oral prednisolone at a dosage of 10 mg twice daily on day 6, immediately following the receipt of the biopsy results.

On day 7, we initiated plasma exchange as a bridge to liver transplantation. The patient underwent two courses of plasma exchange, with the first course (four times) performed on days 7, 10, 12, and 14. The laboratory data on day 8 after the first plasma exchange showed a decrease in serum bilirubin levels and an improvement in renal function (as indicated in Fig. 1B, C). The serum bilirubin levels showed a declining trend following a series of plasma exchange (as shown in Fig. 1C), dropping from 18.45 mg/dL on day 13 to 14.35 mg/dL on day 16. However, they later increased to 19.49 mg/dL on day 18 when plasma exchange was not continued. A similar pattern was observed for the direct bilirubin levels. Therefore, we performed additional six times of plasma exchange and switched the steroids from oral prednisolone back to intravenous methylprednisolone 8 mg every 6 h since day 20. Intractable pruritus, especially at night and unresponsive to antihistamines and cholestyramine, seemed subjectively responsive to an increased dose of steroids. We tapered off the steroids from day 22 to day 25 and observed a transient increase in serum bilirubin levels, along with increased levels of amylase, lipase, and a decline in renal function. On days 22 and 42, serum bile acid levels were 122 µM and 40 µM, respectively. The normal range is <10 µM, while levels exceeding 50 µM are considered elevated.

On day 6, the patient experienced a transient episode of mild epigastric pain. Subsequently, amylase and lipase levels reached their peak at 1717 U/L and 5784 U/L on day 8, respectively (compared to initial data 123 U/L and 290 U/L on day 4) (see Fig. 1D). Abdominal CT on day 5 and follow-up ultrasound scans on day 8 and 19 revealed the absence of gall stones, common bile duct stones or dilatation, and the pancreas parenchyma remained intact without necrosis or swelling. Under the impression of mild acute pancreatitis, temporary parenteral nutritional support was initiated, and oral food intake was not resumed until amylase levels dropped to 108 U/L and lipase levels to 307 U/L on day 16.

Recent literature suggests that immune-mediated liver injury following COVID-19 vaccination and the potential mechanisms and antibody candidates involved warrant further investigation19,20,27. In this case, we observed a decreasing trend in the anti-SARS-CoV-2 spike antibody level (peak, 12,371 BAU/mL on day 32; positive, ≥0.8 BAU/mL), the vaccine-induced antibodies against the viral spike protein, measured using a quantitative Roche Elecsys® Anti-SARS-CoV-2 S assay (Fig. 1F). This decline in antibody levels could possibly be attributed to the effects of therapeutic plasma exchange (major) and a gradual ‘natural’ fading over time post-vaccination (minor)28,29. The peak level was considered high compared to a COVID-19-naïve population receiving a similar course of COVID-19 vaccination with a geometric mean concentration of anti-SARS-CoV-2 spike IgG level of 3898 BAU/mL (range: 3303–4600 BAU/mL, 28 days post boost dose), indicating a stronger immune response28. The clinical course in this case suggested adverse events associated with COVID-19 vaccination, supported by clinical findings and chronological correlation. COVID-19 mRNA vaccines are known to potentially induce systemic inflammation and multiorgan injury by provoking proinflammatory responses and eliciting autoimmune reactions9,26,27. Three main hypothetical mechanisms through which COVID-19 vaccines could trigger autoimmunity are: molecular mimicry (immune cross-reactivity due to similarities between certain vaccine components and specific human proteins), the production of autoantibodies (e.g., anti-platelet factor 4 antibodies), and the specific vaccine adjuvants (by triggering innate inflammatory responses)9,30. The biopsy results and the clinical manifestations in this case, along with the insights derived from the non-HLA antibody-mediated graft injury, particularly the presence of anti-angiotensin II type 1 receptor (AT1R) antibodies, prompted us to further investigate the potential involvement of these antibodies in the pathogenesis. Ideally, the vaccine simulates an actual infection by mimicking the virus and the immune system reacts similarly to an actual infection. However, increased AT1R antibody levels have been observed in cases of multiorgan tissue damage and acute respiratory distress syndrome in COVID-19 patients31,32. We found that the AT1R antibody level in this patient was elevated (peak, 166.61 U/mL; positive, >17 U/mL) (Fig. 1F). AT1R antibody levels were determined in the serum using a capillary electrophoresis–marked enzyme immunoassay developed at CellTrend GmbH according to the manufacturer’s instructions.

Rheumatologists recommend pharmacological combinations of hydroxychloroquine and cilostazol to treat thrombotic antibody-mediated inflammation. Although ferritin levels decreased during the clinical course, they remained highly elevated compared to the upper limit of normal (Fig. 1E). The CRP level exhibited a similar trend but returned to normal after day 28 (Fig. 1E).

As jaundice gradually improved, steroids were tapered from intravenous methylprednisolone to oral prednisolone at a dose of 10 mg twice daily starting on day 47. Additionally, colchicine was introduced, and the oral prednisolone dose was titrated to 12.5 mg twice daily in response to fluctuations in liver enzyme levels (Fig. 1C), even as serum bilirubin levels continued to decrease (peak AST, 64 U/L on day 113; ALT, 149 U/L on day 64). The patient was discharged on day 57 and continued with regular outpatient visits. Throughout the course of treatment, the patient’s mental status was within normal limit, and he was able to ambulate and articulate freely. Details of the follow-up laboratory results and medication usage are presented in Fig. 1. On day 171, abdominal magnetic resonance imaging and cholangiopancreatography revealed unremarkable findings in the liver, gall bladder, kidneys, and pancreas.