Background

Idiopathic non-cirrhotic portal hypertension (INCPH) is a rare vascular liver disease characterized by presinusoidal portal hypertension (PHT) without cirrhosis [1]. It has been diagnosed in up to 84% of patients with common variable immune deficiency, hyper-IgM syndrome, and et al., although mechanisms linking immunological abnormalities to INCPH remain unknown [2]. Abdominal infections have also been implicated as one of the triggers for the development of INCPH through repeated obstruction of small to medium portal branches. However the evidence underlying this association is mainly from epidemiological studies [2]. The etiopathogenesis is so far poorly understood. Herein, we reported that a patient with INCPH was previously diagnosed with intestinal infection of T. Marneffei and hyper-IgE syndrome (HIES), a group of rare primary immunodeficiency disorders [3], which had never reported in INCPH before.

Case presentation

A 26-year-old man was hospitalized due to a 6-month history of fever, cough, expectoration and melena. In the primary hospital, labs reported decreased hemoglobin and increased serum IgE (Table 1). Anti-HIV antibody was not detected but immunodeficiency was found in decreased lymphocyte subsets counts, especially the natural killer T (NKT). Chest computed tomography (CT) showed a pneumatocele in the left lung and bilateral infectious lesions (Fig. 1A). The endoscopy found an ulcer at the ileocecal valve (Fig. 2A) and mild esophageal varices (Fig. 2C). The results of a culture of the sputum yielded T. Marneffei. Periodic acid-Schiff and Hexamine silver staining of the ileocecal biopsy tissue revealed T. Marneffei infection (Fig. 3A). Conclusively, intestinal infection of T. Marneffei and hyper-IgE syndrome (HIES) were included in the diagnosis, with the score = 28 according to NIH scoring system [4] (Table 2). The whole-exon gene sequencing identified compound heterozygous mutation in dedicator of cytokinesis-8 (DOCK8), the causing of HIES. He received intravenous amphoterincin B therapy (from small dose to 0.75 mg / kg/d) for 25 days and was discharged on an oral regimen of voriconazole (400 mg / d) for 15 months until the symptoms disappeared. Recheck of colonoscopy showed the ileocecal lesion was cured (Fig. 2B).

Table 1 Notable laboratory tests in the first and second hospitalization
Full size table
Fig. 1
figure 1

(A) Chest CT showed multiple infectious lesions in the upper lobe of the double lungs during the first hospitalization, which were absorbed obviously in the second hospitalization (B)

Full size image
Fig. 2
figure 2

The first colonoscopy showed an irregularly shaped ulcer at the ileocecal valve (A), which was found to be cured after the 15-month antifungal therapy (B). The esophageal varices was mild (C) when he had first esophagogastroduodenoscopy and was proved to aggravate to the severe gastroesophageal varices with the presence of red sign 29 months later (D)

Full size image
Fig. 3
figure 3

(A) Histopathologic examination for the ileocecal ulcer showed granulomatous inflammation with much round fungal structures (blue arrows; hexamine silver). (B) A fibrotic portal tract (black arrow) with obliteration of portal vein branches (H&E, magnifcation × 400). (C) Multiple thin walled vascular spaces in the portal tract (black arrow) with a periportal abnormal vessel (star). Compensatory enlargement of hepatic sinusoid (blue arrow) can be detected (H&E, magnifcation × 400)

Full size image
Table 2 NIH Scoring System of HIES with Clinical and Laboratory Tests
Full size table

However, the patient was readmitted to our hospital 14 months later because of recurrent melena. Laboratory tests showed decreased hemoglobin, decreased leukocyte, and normal liver function (Table 1). Endoscopy reveled the gastroesophageal varices significantly aggravated (Fig. 2D). Contrast-enhanced abdominal CT showed splenomegaly and gastroesophageal varices without obstruction of main portal vein or inferior vena cava (Fig. 4A). The liver biopsy was performed due to no evidence of chronic liver disease based on labs and previous history. The histopathological changes showed features of portal vein stenosis (Fig. 3B) and hyper-vascμlarized portal tracts (Fig. 3C) with out pseudolobule formation. Finally, the patient was diagnosed with INCPH.

Fig. 4
figure 4

A Contrast-enhanced abdominal CT angiography with three-dimensional reconstruction showed a smooth liver surface, splenomegaly and gastroesophageal varices (white arrow). B Hepatic vein-to-vein communication was detected by hepatic venography. C Portography showed irregularity and tortuosity of the peripheral portal branches with much abrupt interruptions (black arrow) and avascular area beneath the liver surface. Portovenous collaterals (white arrows) derived from gastric coronary vein were also displayed. D Post-TIPS portography revealed embolization of collaterals and clear stent flow from portal vein to systemic circulation

Full size image

Transjugular intrahepatic portosystemic shunt (TIPS) was performed to prevent variceal rebleeding. During the procedure, hepatic vein-to-vein communication (HVVC) was displayed via hepatic venography (Fig. 4B). Portography showed tortuosity of portal vein branches with much interruptions (Fig. 4C). The hepatic venous pressure gradient (HVPG) was detected as only 4 mmHg while the portal pressure gradient before and after stent placement were 18 mmHg and 6 mmHg, respectively. The collaterals was disappeared after TIPS (Fig. 4D) and there was no hepatic encephalopathy and recurrence of melena during the six-mounth follow-up.

Discussion and conclusion

The diagnosis of INCPH was based on radiological examination had excluded prehepatic and posthepatic PHT (such as extrahepatic portal venous obstruction and Budd Chiari syndrome), and liver biopsy displayed portal vein branches stenosis and hyper-vascularized portal tracts, rather than severe hepatic fibrosis or pseudolobule formation [5]. HVVC and normal HVPG detected by hepatic vein catheterisation as well as the portography features also conformed with the diagnosis [6]. The etiopathogenesis of INCPH is so far poorly understood. Immunological abnormalities, including primary antibody-deficiency and HIV infection as well as autoimmune disorders seem to be associated with INCPH [1]. However, autosomal recessive HIES caused by DOCK8 deficiency is a primary combined immunodeficiency which had never reported in INCPH before. Mutations in the DOCK8 gene encoding a guanine nucleotide exchange factor highly expressed in lymphocytes that regulates the actin cytoskeleton were identified as the most common cause for autosomal recessive HIES [3]. DOCK8 deficiency impairs immune cell migration, function and survival which characterized by allergic inflammation as well as susceptibility towards infections, autoimmunity and malignancy [7]. Previous studies have suggested the abnormalities of genes and specific adhesion molecules involved in lymphocyte activation in patients with INCPH [8, 9]. These suggest lymphocyte-endothelial cell cross-talk could play a role in the pathogenesis.

Other possible mechanisms of the portal vein branches damage in patients with immunological disorders includes abdominal infections, immunosuppressants, or anti-infection drugs. In our case, the anti-fungal drugs can not explain the disease because the signs of PHT appeared much earlier. It is noteworthy that the patient had an intestinal infection with Talaromyces marneffei. The fungus is gradually emerging in southeast Asia and southern China which HIV infection is regarded as the most underlying disease [10]. Recently T. Marneffei infections in non-HIV individuals are increasing whose clinical characteristics are atypical [11]. T. Marneffei is a primary lung pathogen that disseminates to other internal organs by lymphatic or hematogenous mechanisms [12]. In our case, there were definite evidences of the presence of T. Marneffei in the lung and intestine. Coupled with a close anatomical connection between the liver and the gut it can be concluded that the patient already had a disseminated disease of T. Marneffei. In other words, T. Marneffei may enter the liver through the superior mesenteric vein and cause endothelial and vascular damage. Additionally, T. Marneffei infection and INCPH exist in sequence, whereas the primary immunodeficiency was already present. Therefore, we hypothesized that the infection of T. marneffei played a major role in the pathogenesis of INCPH. As the DOCK8 deficiency affects the long-term retention of NKT-cell in the liver [13] and may therefore increase the susceptibility to infections, the intestinal involved of T. Marneffei as well as previous occult intestinal infections combined with immune disorders may be the cause of INCPH in this patient [14]. The association between T. Marneffei infections and INCPH has not been reported in previous studies. INCPH is more common in developing countries where bacterial infection is rampant [1, 15]. Also means that as long as it produces septic emboli which cause endothelial damage, microthrombosis, sclerosis, and obstruction of the small and medium-sized portal branches any intestinal infection leads to INCPH. This hypothesis is supported by an animal study that recurrent injection of heat killed Escherichia coli into portal vein can develop the rabbit model of INCPH [16].

As for treatment, the main purpose is to prevent PHT related complications for the lack of etiotropic therapy. A retrospective study including 41 patients with INCPH have demonstrated good effectiveness and safety of TIPS for PHT related complications [17]. In the current case, there is no episode of hepatic encephalopathy or abnormal liver function tests occurred after TIPS during the six-mounth follow-up. Despite the scarcity of evidence, TIPS may precede other therapies for INCPH due to preserved liver function and less post-TIPS complications.

This case demonstrated that, like other kinds of immunodeficiency, primary combined immunodeficiency can occur in INCPH. Intestinal infections may lead to further portal vein damage in such conditions. Thus, screening for enterogenic infection and immunological disorders in patients with unexplained PHT is necessary. Futher studies on mechanisms connecting intestinal infections and INCPH are needed.