INTRODUCTION
Azathioprine was introduced in the 1960s as an immunosuppressive agent, initially for the prevention of transplant rejection and later for its ability to facilitate corticosteroid withdrawal in the treatment of various autoimmune diseases. Thiopurines are widely used in the management of inflammatory bowel disease (IBD). Azathioprine (AZA) and mercaptopurine are effective in maintaining remission.1 The main factor limiting their use is the potential occurrence of adverse effects (AEs) in some patients.2 Azathioprine is a rare cause of portal hypertension (PH) due to its induction of a liver small vessel disorder known as portosinusoidal vascular disease (PSVD).
Recently, cases of AZA-related PH have been reported, though its occurrence is considered to be rare.3 We describe a report of PSVD in a Crohn's disease (CD) patient associated with AZA therapy, and the clinical course after a 4-year follow-up.
CASE REPORT
A 48-year-old man, followed for 15 years for ileocolonic CD initially presenting with chronic diarrhea and complicated by a surgically treated intra-abdominal abscess. There were no anoperineal lesions or extraintestinal manifestations. Deep and sustained remission was achieved with long-term azathioprine (125 mg [2.5 mg/kg/d]).
Fifteen years after beginning of AZA, and during routine follow-up, bicytopenia with leukopenia and thrombocytopenia (without anemia) was detected (WBC = 3,600/mm3, platelets = 91.000/mm3, Hb = 13 g/dL), persisting despite dose reduction (from 125 mg to 50 mg) over 4 months and subsequent discontinuation of AZA. Liver function tests remained persistently normal (prothrombin time: 82%, ALT: 24 IU/L, AST: 32 IU/L, total bilirubin: 5 mg/L, alkaline phosphatase (ALP: 123 IU/L, GGT: 54 IU/L). Esophagogastroduodenoscopy was performed which showed several Grade II-III esophageal varices without red signs.
Hepatic Doppler ultrasound showed a dysmorphic liver, no intrahepatic or extrahepatic biliary duct dilation, and signs of portal hypertension (PH) due to intrahepatic block. There was homogeneous splenomegaly (22 cm in length) and extensive portal thrombosis in the right portal vein, portal bifurcation, and origin of the left portal vein, without morphological signs of portal cavernoma. The hepatic veins and retrohepatic inferior vena cava were patent with normal flow (Figure 1).
The results of an evaluation for chronic liver disease were negative for hepatitis B surface antigen, hepatitis C virus antibody, antinuclear antibody, anti-mitochondrial antibodies, anti-smooth muscle antibody, and anti-liver kidney microsomal type 1 antibody. Magnetic resonance cholangiography to rule out sclerosing cholangitis associated with CD was normal. The stiffness measured with FibroScan test was 9.4 kPa (IQR 1.5; IQR/median = 16%; 10 valid measurements out of 10 total), which was below the range of liver cirrhosis, and controlled attenuation parameter was 156 dB/m.
In summary, this case involves a 48-year-old man with CD who developed presinusoidal intrahepatic portal hypertension during long-term azathioprine therapy, suggesting small-vessel liver disease. Therefore, we concluded that PSVD was potentially related to azathioprine. A liver biopsy was performed.
The specimen was of good quality (2 cm core with 14 portal tracts). Histological examination by an expert pathologist confirmed PSVD of the liver with marked sinusoidal dilatation; obliterative portal venopathy, and nodular regenerative hyperplasia (NRH), and neither fibrosis nor evidence of hepatic cirrhosis, were observed (Figure 2). The biopsy met adequacy criteria for the exclusion of cirrhosis (≥15 mm length, with at least 1 fragment ≥10 mm) according to the 2026 PSVD/noncirrhotic portal fibrosis (NCPF) consensus,4 and the presence of NRH satisfied a major histologic criterion for PSVD/NCPF.
Applying the newly proposed diagnostic score,4 this patient scored 9 points (signs of portal hypertension: +3; histologic lesions: +5; associated condition: +1; concomitant cause of cirrhosis: 0), corresponding to definite PSVD/NCPF (≥5 points).
A comprehensive etiologic workup was performed to identify potential associated conditions of PSVD. The complete work-up, including autoimmune, celiac, hematologic, and thrombophilia screening, was unremarkable and is summarized in Table 1. The patient had no history of exposure to hepatotoxic drugs, herbal preparations, or chemotherapeutic agents such as thioguanine, didanosine, or oxaliplatin.
As all other causes of liver damage were excluded, AZA therapy was believed to be the cause of liver injury. The patient had no epidemiological exposure to a schistosomiasis-endemic area and no clinical or biological findings suggestive of schistosomal infection; schistosomiasis was therefore not considered a relevant differential diagnosis. After discussion with the patient, azathioprine treatment was withdrawn, and biologic therapy was scheduled with meticulous monitoring.
Intestinal status: The patient was asymptomatic, with CD in remission. Our management consisted of discontinuation of azathioprine (AZA) and initiation of carvedilol (12.5 mg/d) to prevent variceal bleeding due to portal hypertension combined with anticoagulants (enoxaparin later switched to Sintrom), and initiation of anti-TNFα therapy for CD. The patient was monitored every 6 months with abdomen ultrasound.
Complete recanalization of the portal vein was achieved at 6 months. After 4 years of follow up, CD remains in clinical and endoscopic remission. Platelet count gradually increase from 91.000 to 120.000/mm3.
Although portal hypertension remained stable, there was a mild progression of liver stiffness to 10.9 kPa (Table 2). This patient has been continuing outpatient follow-up without any complications in 4 years thereafter. The patient is currently treated with anti-TNF and regularly attends follow-up visits at our center.
DISCUSSION
Azathioprine (AZA) and 6-mercaptopurine (6-MP) are associated with frequent AEs, typically occurring within 4–6 weeks of initiation. Up to 25% of patients experience AEs, with an annual risk of 7% per patient-year of treatment.5,6 Approximately 20% of IBD patients treated with thiopurines develop severe AEs requiring treatment discontinuation.7 Common idiosyncratic AEs include skin rashes, fever, acute pancreatitis, and arthralgias. Major dose-dependent AEs are myelosuppression and dyspepsia.
Thiopurine-related liver injury can be both idiosyncratic and dose-dependent. Hepatotoxicity may develop months to years after treatment initiation and can often be managed with dose reduction in nearly half of cases.8 In our patient, liver injury became clinically apparent after 15 years of AZA therapy without prior liver test abnormalities. A systematic review of 34 studies (3,485 patients) estimated a mean prevalence of AZA/6-MP-induced liver injury at 3.4%, with no significant difference between the 2 drugs.9 AZA and 6-MP accounted for 1%–2% of drug-induced liver injury cases in the DILIN network.10
PSVD, previously termed NRH, idiopathic portal hypertension, or noncirrhotic intrahepatic portal hypertension, frequently manifests as portal hypertension (PH) and its complications.11,12 According to the recent multisociety consensus statement, the terms PSVD and NCPF are considered equivalent and may be used interchangeably under a unified diagnostic framework.4 There are more than 60 conditions associated with PSVD and these include, in addition to drug exposures, genetic, hematologic, prothrombotic, immune, inflammatory, and infectious diseases; a comprehensive listing is available.4 AZA-associated PSVD with PH accounts for 0.5–1.25% of cases.13 The mechanism linking AZA to PSVD remains unclear. Musumba et al propose that glutathione depletion caused by AZA leads to endothelial cell injury in sinusoids and hepatic veins, resulting in nonthrombotic vascular occlusion.14 This triggers hepatic hypoperfusion, parenchymal atrophy, and subsequent fibrosis/PH,15 characterized by alternating zones of atrophy and adaptive hepatocellular hyperplasia.16 The pathogenesis of PSVD in IBD is multifactorial. The largest series (37 cases from 11 French tertiary centers, Groupe d’Étude Thérapeutique des Affections Inflammatoires Digestives group) reported a cumulative NRH risk of 0.5% at 5 years and 1.25% at 10 years.
Diagnosis occurred at a mean of 48 months post-AZA initiation [6–187 months], with 38% developing PH. Risk factors include male sex and stenosing phenotype.13 Another study identified high-risk groups as men with ≥50 cm small bowel resection before or after AZA initiation.17 Notably, PSVD is incidentally found in 6% of thiopurine-naïve IBD patient's postintestinal resection,18 possibly due to surgery-induced vitamin B12, B6 and folate malabsorption and subsequent hyperhomocysteinemia.14 PSVD may remain asymptomatic for years with normal liver tests,19 as seen in our patient. Diagnosis relies on liver biopsy, though interobserver variability is high, even among experienced pathologists.20 PSVD should be suspected in 2 clinical scenarios: unexplained liver test abnormalities and/or noncirrhotic intrahepatic portal hypertension (PH).21 In a French study of 3,600 liver biopsies performed for unexplained liver enzyme abnormalities, PSVD was identified in 159 cases (4%).22 In a recent international study (27 centers) of 587 patients with PSVD, most patients had preserved liver function, but many developed PH-related complications over time.
At the time of diagnosis, about one-third already had symptoms such as variceal bleeding or ascites, while the remainders were diagnosed incidentally. AZA exposure was identified in about 8% of cases, mainly in patients with IBD.23 PSVD should be considered after ruling out recognized chronic liver diseases (eg, alcohol misuse, metabolic syndrome, hepatitis B/C, autoimmune hepatitis, hepatic vein obstruction). Liver Test Abnormalities in PSVD: Transaminase elevation (moderate in most cases) in two-thirds of patients, ALP elevation in 50% of cases, occasionally exceedingly twice the upper limit of normal and Gamma-glutamyl transferase elevation.24,25 PH may present with normal/near-normal liver function tests and liver stiffness measurement below 10 kPa.26 Diagnostic criteria of PSVD include the absence of cirrhosis in a liver biopsy and the detection of specific or nonspecific histological findings with or without portal hypertension.11 Management focuses on discontinuing the causative medication (eg, thiopurines) and addressing Portal hypertension complications.
Portal vein thrombosis, a potential consequence of PSVD due to intrahepatic flow obstruction, may develop during follow-up.24
The natural history of patients with PSVD is derived from small studies. To date, there is scarce information regarding its natural history and prognostic factors. The prognosis is strongly determined by the severity of the associated underlying conditions.
During a median follow-up of 68 months. Magaz et al, found that about 15% developed a first variceal bleeding, 18% developed new or worsening ascites and 16% of developing portal vein thrombosis within 5 years. Despite these complications, overall survival was good, with transplant-free survival rates of 97% at 1 year, 83% at 5 years, and 72% at 10 years.
The study did not report major differences in the clinical course or prognosis of AZA related PSVD compared with other associated conditions. The main prognostic factors identified were age, presence of ascites, serum bilirubin, albumin, and creatinine levels, as well as the severity of the underlying associated condition. Patients with severe systemic diseases, had a worse prognosis.
In contrast, those without major comorbidities tended to have a stable course and good long-term outcomes.23 Few data are available regarding the natural history of PSVD after discontinuation of AZA and remains uncertain. In our case, after AZA cessation. there was an increase in platelets counts and portal hypertension remained stable. However, liver stiffness progression suggests ongoing microvascular damage.
Ferlitsch et al, reported an improvement of portal pressure in thiopurine-exposed NRH patients, after the cessation of thiopurines.27 According to Seiderer et al, discontinuation of azathioprine led to an improvement in portal hypertension symptoms.28 Whereas Pola et al observed that NRH and hepatoportal sclerosis induced by 6-mercaptopurine were irreversible.29 Given the lack of standardized treatment protocols for PSVD, withdrawal of suspected causative agents like azathioprine is recommended. Management of portal hypertension should follow the Baveno VII consensus for variceal bleeding,30 while liver transplantation may be indicated in patients with refractory ascites or end-stage hepatic dysfunction.31
Azathioprine is a drug widely used in medical practice. It is generally effective and well tolerated. The occurrence of PSVD is a rare hepatic AE that must be systematically screened for.
Thiopurines should be permanently discontinued following the onset of PSVD. This case highlights the necessity of a sustained follow-up in patients undergoing long-term thiopurine therapy.
DISCLOSURES
Author contributions: F. Azizi: Contributed to patient management, data collection, literature review, and drafting of the manuscript. R.
Ouldgougam, Y. Aouameur, SES Bourezak and BM El Amin: Contributed to patient management, data acquisition, and critical revision of the manuscript. M.
Gasri contributed to the radiological evaluation and interpretation of imaging findings. ZC Amir performed the histopathological analysis and contributed to interpretation of pathological findings. All authors reviewed and approved the final version of the manuscript and agree to be accountable for all aspects of the work.
F. Azizi is the article guarantor.
Financial disclosure: None to report.
Informed consent was obtained for this case report.