Primary biliary cholangitis (PBC) is a relatively common autoimmune liver disease.^1-3 Histologically, PBC manifests as chronic cholestasis which is characterized by non-suppurative progressive injury to the small intrahepatic bile ducts. It is generally divided into four phases: phase I, cholangitis; phase II, peri-congestive cholangitis; phase III, progressive fibrosis; and phase IV, cirrhosis.^3,4 Portal hypertension (PH) refers to is an elevated pressure in the portal vein due to various causes, and it is associated with clinical manifestations such as splenomegaly, hypersplenism, esophagogastric fundal varices and ascites, etc.⁵ PH can be divided into pre-hepatic, hepatic, and post-hepatic PH, with hepatic PH is further classified divided into presinusoidal, sinusoidal, and post-sinusoidal types.⁵ As PBC progresses, it can eventually lead to biliary cirrhosis, resulting in a variety of complications, including PH. However, there exists a subset of patients with PBC who have not yet developed cirrhosis histologically but with PH and may experience with severe complications, including splenomegaly, hypersplenism, and even gastrointestinal hemorrhage.⁶ Related studies have shown that 5% to 10% of patients with precirrhotic PBC develop esophageal varices under endoscopy. Their results suggest that low platelets (PLT), low albumin (ALB), and high total bilirubin (TBIL) are independent predictors for the development of esophageal varices in early PBC.^7,8 Furthermore studies have shown that PH in early PBC is considered to be presinusoidal PH, which may be associated with nodular regenerative hyperplasia.^6,9,10 A study of hepatic venous pressure gradient (HVPG) measurements in PBC patients shows that about 34% of precirrhotic PBC patients have high-risk PH, which is associated with portal vein and hepatic sinusoid lesions.¹¹

However, in clinical practice, the risk of liver biopsy is higher when there are cirrhosis-related complications, leading to less frequent utilization of liver histological examination. The existing studies have primarily focused on histological staging, and the histopathological mechanisms underlying precirrhotic PBC with PH remain unclear. Furthermore, these studies did not exclude other autoimmune liver diseases such as autoimmune hepatitis. Our study is the first to comprehensively investigate the details of bile duct injury or loss in precirrhotic PBC patients with PH. Most of the existing studies have relied on gastroscopy as the diagnosis method of PH, but this method may overlook the diagnosis of patients with periesophageal varices. In addition, gastroscopy and other screening tests for esophagogastric varices are usually recommended only for patients with cirrhosis and are not routinely used for patients with precirrhotic PBC.⁷ Consequently, there are limited reports on PH in precirrhotic PBC patients. Imaging examination is a noninvasive examination, and more convenient than gastroscopy, ensuring the diagnosis of patients with periesophageal varices. However, there is a lack of studies on using images as a screening tool for PBC combined with PH.

This study aimed to elucidate the clinical significance and histologic pathogenesis of PH in patients with precirrhotic PBC by examining the prevalence, clinical characteristics, risk factors, histologic characteristics, and response to ursodeoxycholic acid (UDCA) in precirrhotic with PH.

1. Patients

This was a retrospective study on patients who were diagnosed with precirrhotic PBC after liver tissue puncture examination from January 2017 to December 2022 in the Second Hospital of Nanjing. We excluded patients according to the following exclusion criteria: (1) combination of chronic liver diseases other than PBC; (2) no imaging examination within 3 months before or after the liver puncture procedure; (3) those with incomplete clinical data; or (4) cirrhosis. According to the inclusion and exclusion criteria, we included a total of 412 patients diagnosed with PBC based on liver histology and excluded 205 patients with other liver diseases (198 autoimmune hepatitis, 5 viral hepatitis, and 2 drug-induced hepatitis), 20 patients with no imaging data, four patients with incomplete clinical data, and 18 patients with cirrhosis. Finally, 165 precirrhotic PBC patients were enrolled in the study. In this retrospective study, the data were anonymous, and the requirement for informed consent was waived by the Medical Ethics Committee of the Second Hospital of Nanjing (No. 2022-LY-kt101). This study was approved by the examination of the Second Hospital of Nanjing. The study protocol conformed to the ethical guidelines of the Declaration of Helsinki.

1) Diagnosis of PBC

PBC can be defined by meeting at least two of the following criteria: 1. serum alkaline phosphatase (ALP) ≥2×upper limit of normal (ULN) or serum glutamyl transpeptidase ≥5×ULN; 2. serum anti-mitochondrial antibody positive or anti-mitochondrial antibody-M2 positive; 3. bile duct injury in the portal area on liver histology.¹²

2) Diagnosis of PH

PH can be defined by excluding other causes of PH and meeting at least one of the following specific manifestations: (1) specific manifestations (1. gastric, esophageal, or ectopic varices, 2. portal hypertensive hemorrhage, 3. imaging suggestive of portal collateral circulation); (2) nonspecific manifestations (1. ascites, 2. PLT <150×10⁹/L, 3. spleen length diameter ≥13 cm).¹³

3) UDCA Response Criteria

For all patients treated with UDCA, the administered dose was 13–15 mg/kg/day. UDCA response was defined as a 40% reduction or normalization of the ALP level (meeting the Barcelona criteria).¹⁴

2. Histological grading staging

1) Nakanuma staging system

The Nakanuma staging system was based on fibrosis, bile duct loss, and lichen red-positive particle deposition.¹⁵ For fibrosis, a score of 0 indicates almost no fibrosis or fibrosis confined to the portal area; a score of 1 indicates fibrosis extending beyond the portal area with occasional incomplete fibrous septa; a score of 2 indicates bridging fibrosis with lobular disorganization; a score of 3 indicates cirrhosis, characterized by extensive fibrosis and regenerative nodules. For the absence of ducts, a score of 0 signifies the presence of interlobular bile ducts in all the portal areas of the specimen; a score of 1 and 2 indicate the loss of bile ducts is evident in <1/3 and in 1/3-2/3 of the portal areas, respectively; a score of 3 means that bile ducts were absent in >2/3 of portal tracts. Lichen red-positive particles are copper-binding proteins in the lysosomes of hepatocytes, and their deposition reflects the degree of chronic bile salt sludge. A score of 0 indicates no deposition of hepatocytes around the portal area, scores of 1 signifies deposition in 2/3 of the portal area and a score of 2 falls between the previous two manifestations. After scoring each item, a total score was calculated to determine the stage of progression. Stage I (no progression) corresponds to a total score of 0, Stage II (mild progression) to a total score of 1-3, Stage III (moderate progression) to a total score of 4-6, and Stage IV (severe progression) to a total score of 7-9. In cases where lichen planus staining is not feasible, the sum of the fibrosis and bile duct defect scores can be used to determine the stage as follows: Stage I (no progression) with a total score of 0, Stage II (mild progression) with a total score of 1-2, Stage III (moderate progression) with a total score of 3-4, and Stage IV (severe progression) with a total score of 5-6.

2) Fibrosis and inflammation staging

Fibrosis was classified as stages 0-4: stage 0, no fibrosis; stage 1, fibrosis confined to the portal area; stage 2, fibrosis around the portal area or septum fibrosis between the portal area and portal area, but the relationship between blood vessels are complete; stage 3, fibrosis with distorted liver structure without obvious cirrhosis; stage 4, possible or definite cirrhosis of liver. Inflammation was classified as stages 0-4: stage 0, no inflammation; stage 1, inflammatory but no liver damage; stage 2, focal necrosis or acidophil bodies; stage 3, severe focal cell damage; and stage 4, fusion necrosis.¹⁶

3) Active cholangitis grading

The active cholangitis was divided into 0-3 grades. Grade 0 means no cholangitis; grade 1 means that one damaged bile duct with obvious chronic cholangitis; grade 2 means that there are more than two bile ducts with obvious chronic cholangitis; grade 3 means that at least one damaged bile duct with chronic non-suppurative cholangitis (granulomatous cholangitis).¹⁵

4) Grading on the degree of cytokeratin 7 (CK7) hepatocyte positivity

The CK7 hepatocyte positivity was divided into 0-3 grades. Grade 0 means absent; grade 1 means that positive cells in less than 1/3 of the portal area, grade 2 means that positive cells in 1/3-2/3 of the portal area; grade 3 means that positive cells in more than 2/3 of the portal area.¹⁷

3. Statistical analysis

Continuous variables are expressed as median and interquartile range and the Mann-Whitney U test was used to compare the differences between the two groups. Categorical variables were expressed as counts and percentages and the chi-square test was used to compare the differences between groups. Multifactorial logistic regression was used to analyze the risk factors for PBC with PH. The recovery probability of liver function was determined for these groups using the Kaplan-Meier method, and results were compared using the log-rank test. p<0.05 was considered significant. Statistical analysis was performed using SPSS statistical software 25.0 (IBM Corp., Armonk, NY, USA).

1. Basic characteristics of the study population

A total of 165 PBC precirrhotic patients were enrolled, including the non-PH group (n=125) and PH group (n=40) (Fig. 1). The demographic and baseline characteristics of the enrolled patients were summarized in Table 1. Among the study population, 137 (83%) were female and the median age was 52 years. The characteristics of the two groups of patients are demonstrated in Table 1. The age of patients in the PH group was slightly higher than that in the non-PH group (54 years vs 52 years, p=0.043). Compared to the non-PH group, patients with PH had lower PLT (90×10⁹/L vs 189×10⁹/L, p<0.001) and white blood cell (WBC; 3.63×10⁹/L vs 5.08×10⁹/L, p<0.001) counts, lower ALB levels (37.9 g/L vs 41.9 g/L, p<0.001), higher levels of ALP (1.79 ULN vs 1.13 ULN, p<0.001), TBIL (18.35 μmol/L vs 13.25 μmol/L, p<0.001), liver stiffness measurement value (9.8 kPa vs 6.2 kPa, p<0.001) and a longer prothrombin time (PT) (11.55 seconds vs 10.90 seconds, p=0.001), which were the manifestations of advanced liver disease. Compared to the patients without PH, PBC patients with PH had higher scores of aspartate aminotransferase to PLT ratio index (1.41 vs 0.54, p<0.001), fibrosis-4 (FIB-4; 4.68 vs 1.81, p<0.001), and Mayo score (1.33 vs 0.52, p<0.001). It was also suggested that the combination of PH meant further disease progression. Since PLT was part of the Mayo score and PT, ALB, and TBIL were part of the FIB-4 index, they were not included in the multifactorial analysis. That liver stiffness measurement value is significantly correlated with FIB-4 index, Mayo score, and aspartate aminotransferase to PLT ratio index, it is not included in the multivariate analysis. The result of the multifactorial analysis is shown in Table 2. Low WBC count, high ALP level, and high Mayo score, and FIB-4 index were the risk factors for precirrhotic PBC combined with PH (p=0.015, p=0.002, p=0.040, p=0.012, respectively).

Figure 1. Flowchart of study. PBC, primary biliary cholangitis.

Table 1 . Baseline Characteristics of Precirrhotic Primary Biliary Cholangitis Patients.

Characteristic	Whole cohort (n=165)	PH group (n=40)	Non-PH group (n=125)	p-value
Age, yr	52 (48–58)	54 (50–57)	52 (46–59)†	0.043
Female sex	137 (83.0)	33 (82.5)	104 (83.2)	0.988
Antibody positive	144 (87.2)	35 (87.5)	109 (87.2)	0.960
AMA positive	110 (66.7)	23 (57.5)	87 (69.6)	0.158
AMA-M2 positive	78 (47.2)	22 (55.0)	56 (44.8)	0.261
Anti-gp210 positive	47 (28.4)	13 (32.5)	34 (27.2)	0.518
Anti-sp100 positive	20 (12.1)	7 (17.5)	13 (10.4)	0.231
ANA positivity	126 (76.4)	30 (75.0)	96 (76.8)	0.816
IgG, g/L	13.8 (11.4–16.6)	14.8 (10.5–17.4)	13.6 (11.2–16.5)	0.261
IgM, g/L	3.10 (1.76–4.12)	2.65 (1.59–4.55)	2.62 (1.64–4.08)	0.780
PT, sec	10.9 (10.4–11.6)	11.6 (10.7–12.2)	10.9 (10.4–11.6)†	0.001
FIB, g/L	2.60 (2.17–2.89)	2.45 (2.05–2.79)	2.58 (2.16–2.89)	0.092
PLT, ×109/L	180 (118–235)	90 (64–145)	189 (145–244)†	<0.001
WBC, ×109/L	4.73 (3.82–5.64)	3.63 (2.84–4.20)	5.08 (4.12–6.03)†	<0.001
ALP, ULN	1.73 (1.07–2.77)	1.79 (1.31–2.72)	1.13 (0.76–1.78)†	<0.001
GGT, ULN	3.77 (1.39–6.72)	4.07 (1.56–6.79)	2.91 (1.13–6.33)	0.207
ALT, ULN	1.03 (0.59–2.02)	1.20 (0.58–1.90)	0.99 (0.55–1.93)	0.962
AST, ULN	1.05 (0.72–1.65)	1.25 (0.85–1.99)	1.00 (0.67–1.56)	0.053
ALB, g/L	41.2 (38.3–44.5)	37.9 (35.6–43.3)	41.9 (39.0–44.7) †	<0.001
GLO, g/L	30.0 (27.1–33.5)	31.7 (27.7–34.9)	29.4 (26.9–33.0)	0.179
TBIL, μmol/L	13.7 (9.9–19.5)	18.4 (13.7–37.1)	13.3 (9.2–17.3)†	<0.001
APRI	0.68 (0.38–1.09)	1.41 (0.80–2.03)	0.54 (0.35–0.97)†	<0.001
FIB-4	1.96 (1.29–3.22)	4.68 (3.13–6.67)	1.81 (1.17–2.45)†	<0.001
Mayo score	0.55 (0.25–1.02)	1.33 (0.72–1.90)	0.52 (0.18–0.92)†	<0.001
LSM, kPa*	6.7 (5.5–9.5)	9.8 (8.5–14.0)	6.2 (5.3–8.0)†	<0.001
Clinical manifestation
Pruritus	15 (9.0)	6 (15.0)	9 (7.2)	0.135
Feeble	58 (35.1)	11 (27.5)	47 (37.6)	0.244
Jaundice	18 (10.9)	11 (27.5)	7 (5.6)†	<0.001
Ascites	9 (5.4)	7 (17.5)	2 (1.6)†	0.001
Poor appetite	26 (15.7)	4 (10.0)	22 (17.6)	0.323
Sjogren's syndrome	14 (8.4)	5 (12.5)	9 (7.2)	0.295
Epigastric discomfort	11 (6.6)	6 (15.0)	5 (4.0)†	0.015
Asymptomatic	67 (40.6)	10 (25.0)	57 (45.6)†	0.021

Data are presented as median (interquartile range) or number (%)..

PH, portal hypertension; AMA, anti-mitochondrial antibody; ANA, anti-nuclear antibodies; IgG, immunoglobulin G; IgM, immunoglobulin M; PT, prothrombin time; FIB, fibrinogen; PLT, platelet; WBC, white blood cells; ALP, alkaline phosphatase; ULN, upper limit of normal; GGT, gamma-glutamyl transpeptidase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ALB, albumin; GLO, globulin; TBIL, total bilirubin; APRI, AST to PLT ratio index; FIB-4, fibrosis-4; LSM, liver stiffness measurement..

*A total of 110 patients completed transient elastography; ^†There was significant difference between PH and Non-PH groups..

Table 2 . Multifactorial Analysis of Factors Associated with Precirrhotic Primary Biliary Cholangitis Combined with Portal Hypertension.

Variable	β	p-value	OR (95% CI)
Age	–0.034	0.319	0.966 (0.903–1.034)
White blood cell	–0.577	0.015	0.562 (0.352–0.895)
Alkaline phosphatase	0.755	0.002	2.127 (1.309–3.459)
Mayo score	0.697	0.040	2.007 (1.034–3.899)
APRI	–1.272	0.060	0.280 (0.074–1.056)
FIB-4 index	0.573	0.012	1.774 (1.133–2.778)

APRI, aspartate aminotransferase to platelet ratio index; FIB-4, fibrosis-4; OR, odd ratio; CI confidence interval..

2. Clinical manifestations of precirrhotic PBC with PH

Compared to the non-PH group, patients in the PH group showed a higher prevalence of more clinical manifestations such as jaundice (27.5% vs 5.6%, p<0.001), ascites (17.5% vs 1.6%, p=0.001), and epigastric discomfort (15% vs 4%, p=0.015), indicative of advanced liver disease (Table 1). Additionally, the results demonstrated that a larger proportion of patients in the non-PH group were without any clinical symptoms (45.6% vs 25.0%, p=0.021).

3. Imaging manifestations of precirrhotic PBC patients in different histologic stages

The imaging manifestations of PH in precirrhotic PBC patients with different histologic stages are presented in Table 3. Among the 19 patients with stage 1 histology, one patient had specific manifestations (varicose veins). Among the 98 patients with histological stage 2, 17 patients had specific manifestations (varicose veins) and two of them also had portosystemic collaterals. Among the 48 patients with histological stage 3, 22 patients had specific manifestations (varicose veins), including two patients with portosystemic collaterals and one patient with PH bleeding.

Table 3 . Imaging Manifestations of Precirrhotic Primary Biliary Cholangitis Patients.

Characteristic	Stage I (n=19)	Stage II (n=98)	Stage III (n=48)	p-value
Specific	1 (5.3)	17 (17.3)	22 (45.8)	<0.001
Gastric, esophageal, or ectopic varices	1 (5.3)	17 (17.3)	22 (45.8)	<0.001
Portal hypertensive bleeding	0	0	1 (2.1)	0.406
Portosystemic collaterals at imaging	0	2 (2.0)	2 (4.2)	0.755
Not specific	3 (15.8)	40 (40.8)	29 (60.4)	0.002
Spleen size ≥13 cm in the largest axis	2 (10.5)	23 (23.5)	23 (47.9)	0.002
Ascites	0	4 (4.1)	5 (10.4)	0.211
Platelet count <150×109/L	2 (10.5)	28 (28.6)	15 (31.3)	0.201

Data are presented as number (%)..

4. Histologic comparison of precirrhotic PBC patients with and without PH

There were no significant differences in epithelioid granulomas, active cholangitis, bile duct reaction, interfacial inflammation, and inflammation staging between the PH group and non-PH group in Table 4. The loss ratio of bile duct (0.33 vs 0.18, p=0.002), the degree of cytokeratin 7 positivity (p=0.008), and the degree of fibrosis (p=0.002) in PH group were significantly higher than those in the non-PH group, while the ratio of lymphoid follicles (0.09 vs 0.16, p=0.018) in PH group was significantly lower than that in the non-PH group. Compared to the non-PH group, patients in the PH group showed a higher prevalence of obliterative portal venopathy (40.0% vs 11.2%, p<0.001), incomplete septal fibrosis (12.5% vs 0.8%, p=0.003), portal tract abnormalities (72.5% vs 33.6%, p<0.001) and non-zonal sinusoidal dilatation (32.5% vs 14.4%, p=0.018) (Table 4, Fig. 2).

Figure 2. Histological characteristics associated with portal sinus vascular disorders in some precirrhotic primary biliary cholangitis patients with portal hypertension. (A) Nodular regenerative hyperplasia (hematoxylin and eosin [H&E] staining, ×100). (B) Obliterative portal venopathy (H&E staining, ×200) (C) Portal tract abnormalities (multiplication) (H&E staining, ×100). (D) Portal tract abnormalities (aberrant vessels: herniation of the portal vein) (H&E staining, ×100). (E) Non-zonal sinusoidal dilatation (H&E staining, ×100). (F) Incomplete septal fibrosis (H&E staining, ×100).

Table 4 . Histological Comparison of Precirrhotic Primary Biliary Cholangitis Patients with or without PH.

Characteristic	PH group (n=40)	Non-PH group (n=125)	p-value
Lymphoid follicle	24 (60.0)	97 (77.3)	0.061
Lymphoid follicle ratio*	0.09 (0.00–0.19)	0.16 (0.06–0.31)	0.018
Epithelioid granuloma	29 (72.5)	92 (73.6)	0.922
Epithelioid granuloma ratio*	0.13 (0.05–0.25)	0.20 (0.07–0.37)	0.225
Bile duct loss ratio*	0.33 (0.19–0.50)	0.18 (0.00–0.37)	0.002
Active cholangitis			0.161
0	0	2 (1.6)
1	14 (35.0)	27 (21.6)
2	13 (32.5)	38 (30.4)
3	13 (32.5)	58 (46.4)
CK7			0.008
0	9 (22.5)	55 (44.0)
1	9 (22.5)	25 (20.0)
2	1 (2.5)	14 (11.2)
3	21 (52.5)	31 (24.8)
Bile duct reaction	27 (67.5)	81 (64.8)	0.779
Interfacial inflammation	33 (82.5)	94 (75.2)	0.297
Fiber staging			0.002
0	2 (5.0)	10 (8.0)
1	1 (2.5)	28 (22.4)
2	31 (77.5)	80 (64.0)
3	6 (15.0)	7 (5.6)
Inflammation staging			0.344
1	7 (17.5)	12 (9.6)
2	16 (40.0)	78 (62.4)
3	15 (37.5)	35 (28.0)
4	2 (5.0)	0
Nakanuma staging			<0.001
I	1 (2.5)	18 (14.4)
II	17 (42.5)	81 (64.8)
III	22 (55.0)	26 (20.8)
Characteristics associated with PSVD
Obliterative portal venopathy	16 (40.0)	14 (11.2)	<0.001
Nodular regenerative hyperplasia	2 (5.0)	5 (4.0)	0.677
Incomplete septal fibrosis	5 (12.5)	1 (0.8)	0.003
Portal tract abnormalities†	29 (72.5)	42 (33.6)	<0.001
Non-zonal sinusoidal dilatation	13 (32.5)	18 (14.4)	0.018

Data are presented as number (%) or median (interquartile range)..

PH, portal hypertension; CK7, cytokeratin 7; PSVD, porto-sinusoidal vascular disorder..

*The ratio is the number of bile duct defects, lymphoid follicles, and epithelioid granuloma confluent areas to the total number of confluent areas per patient; ^†Portal tract abnormalities include multiplication, dilatation of arteries, periportal vascular channels and aberrant vessels..

5. The subsequent clinical outcomes and UDCA response between the two groups after UDCA treatment

We identified patients with patients who did not meet the UDCA response criteria after UDCA treatment as non-responders. There was a significant difference in the un-recovery probability of liver function within 1 year between the PH group and the non-PH group (p=0.002) (Fig. 3). In addition, there was a significant difference in the probability of decompensation events between the two groups after UDCA treatment (p<0.001) (Fig. 3).

Figure 3. The subsequent clinical outcomes of primary biliary cholangitis (PBC) patients in the portal hypertension (PH) group and the non-PH group after ursodeoxycholic acid treatment. Liver function unrecovery probability and compensation event probability of PBC patients were calculated by the Kaplan-Meier method. There was significant difference between the two subgroups (p<0.05).

The risk of developing PH in patients with liver cirrhosis is very high. However, the present study found that manifestations of PH, such as esophageal varices, collateral circulation formation, and occasionally life-threatening variceal bleeding may occur in PBC at precirrhotic state. It has been reported that the development of esophageal varices in the early stage of the disease is a unique disease feature of PBC patients, which rarely occurs in the non-cirrhotic period of other liver diseases such as viral hepatitis.^18-20 However, there is an increased risk of liver biopsy when PH-related complications occur in clinical work, so liver histological examination is rarely performed. In PBC patients combined with PH, there may be an additional presinusoidal component of PH that cannot be evaluated by HVPG.^9,21 Therefore, among these patients, HVPG may underestimate the prevalence and severity of PH. In addition, gastroscopy and other methods are recommended to screen esophageal-gastric varices only in patients with histological stage IV, and this method is not routinely recommended in early-stage patients. Furthermore, this method may overlook some patients with extraesophageal varices and cannot fully represent the patients with PH.^7,10 Related studies have proved that varicose veins can be evaluated by imaging examination, and it has high sensitivity and specificity, which can be used in the diagnosis of PH.^22-27 By means of imaging examination, the method of defining PH is more comprehensive according to the diagnostic criteria of specificity and non-specificity. This is a retrospective study analyzed a large number of patients with PBC and is the first study to focus on combined PH in patients with simple precirrhotic stage PBC from multiple perspectives.

Our study revealed that 24.2% of patients with precirrhotic PBC may experience PH. According to Nakanuma's pathologic staging, there were 5.2% (1/19) of patients with PH in the histologic stage I of our study, 17.3% (17/98) of the patients in stage II had PH, and 45.8% (22/48) of the patients in stage III had comorbid PH. The occurrence of combined PH increases as the histology progresses (p<0.001). Some studies have proved that the cause of PH in patients with histological stage IV PBC is the same as most other causes of liver cirrhosis, which is related to the pseudolobular nodule of cirrhosis. The resistance to blood flow in the blood sinusoids within the pseudolobular nodules increases, thereby leading to PH. As for PH occurring in precirrhotic PBC, studies have demonstrated that some histological changes in the course of PBC are linked to esophageal varices, including inflammation confined to the portal vein and periportal area, portal vein lesions, and nodular regenerative hyperplasia. All liver biopsies in our study were of adequate length and quality and were reviewed by an experienced pathologist. The results showed that the percentage of lymphoid follicles in patients with PBC with PH in the precirrhotic was significantly lower than that in patients without PH. Compared with the group without PH, the percentage of bile ducts missing, the degree of cytokeratin 7 positivity, and the stage of fibrosis were significantly higher than those in patients with PH. The results suggest a potential association between PBC with PH and cholestasis. A study on cholestasis has proposed that the development of PH is associated with reduced production of nitric oxide (NO).¹⁸ NO is produced by endothelial cell NO synthase. The function of endothelial cell NO synthase in patients with cholestasis may be changed, resulting in a reduction of NO released by endothelial cells in the liver, which impair normal vascular tension normally, resulting in PH.

Related studies reported that male sex, low ALB, elevated TBIL, and/or prolonged PT were predictors of the development of early esophageal varices in PBC,^7,8 which is consistent with some of the results of our univariate analysis, as shown in Table 2. In addition, our multifactorial analysis showed that high ALP, low WBC, high Mayo score, and high FIB-4 index were risk factors for PBC with PH. A study of patients with early PBC with esophageal varices showed that a high ALP ratio at diagnosis indicated the presence of presinusoidal PH and may be significantly associated with the presence or development of esophageal varices in patients with precirrhotic PBC.^8,28 Sakisaka et al.²⁹ have also reported other possibilities that the high ALP ratio may reflect changes in the tight junction between biliary epithelial cells and hepatocytes. The relationship between the changes of tight junction and the development of esophageal varices should be further investigated. There is a significant correlation between low WBC and precirrhotic PBC in patients with PH,⁸ which may be related to hypersplenism. Hypersplenism is a common clinical manifestation in patients with PH. Hypersplenism leads to overactive phagocytosis, which increases the destruction of WBC and PLT, resulting in a significant decrease in WBC and PLT. TBIL, ALB, and PT are the main biochemical indexes reflecting the anabolism, detoxification, and reserve function of the liver. The Mayo score, as a comprehensive assessment of TBIL, ALB, and PT, evaluates the status of hepatic function. In PBC patients with PH, the liver anabolism function decreased, bilirubin metabolism was blocked, and the synthesis ability of ALB and coagulation factors decreased. FIB-4 index includes alanine aminotransferase, aspartate aminotransferase, PLT, and the patient's age, which can be used as an index to evaluate the degree of liver fibrosis. Patients with a higher degree of fibrosis are more likely to have PH according to the histological results, resulting in a higher FIB-4 index.

In terms of clinical manifestations, patients with PH were more likely to show jaundice, ascites, and epigastric discomfort.^2,20 Due to the reduced anabolic function of the liver in combined PH, ALB synthesis is reduced, plasma colloid osmotic pressure is reduced, and portal vein pressure is increased, which leads to the formation of ascites.³⁰ When the intrahepatic bile duct is injured, the ability to convert indirect bilirubin into direct bilirubin decreases, resulting in the abnormal excretion of bilirubin, which leads to jaundice. The study showed that jaundice indicates a considerable degree of destruction of the intrahepatic bile duct, which may cause a strong inflammatory response to hepatic regenerative nodules and portal vein,⁹ leading to PH.

We also compared the subsequent clinical outcomes and the response to UDCA between the two groups after UDCA treatment. It has been previously reported that UDCA can delay the development of esophageal varices in patients with PBC.^17,31 Our findings support the idea that UDCA may help prevent PH or at least may not aggravate the development of PH in PBC patients, suggesting a relationship between cholestasis and the occurrence and progression of PH.

However, our study had some limitations. This study is a retrospective clinical analysis and PBC patients with histologically advanced (cirrhotic stage) were not included in the study to compare the difference between cirrhotic PH and non-cirrhotic PH in PBC. In addition, we did not measure the HVPG of the patients. In the future, we plan to continue expanding the sample size, collect advanced histological cases, and further study the pathogenesis of PBC with PH.

In summary, the percentage of precirrhotic PBC patients with PH at the time of diagnosis was 24.2% (40/165). High ALP decreased PLT count, high Mayo score, and high FIB-4 index at the time of diagnosis were effective predictors of the presence of PH in precirrhotic PBC patients, and tests related to PH should be reviewed regularly in such patients. The presence of combined PH in precirrhotic of PBC is associated with cholestasis. Cholestasis due to bile duct defects may affect the function of endothelial cells in portal vein vessels, which in turn may lead to venous hypertension.

No potential conflict of interest relevant to this article was reported.

Study concept and design: Y.F.H., Y.F.Y. Data acquisition; Data analysis and interpretation: Y.F.H., S.X.L., H.L.L., Z.X.D., S.S.W., M.Y.C., L.W., Q.F.X., Y.D.Z., D.X.L. Drafting of the manuscript: Y.F.H. Critical revision of the manuscript for important intellectual content: S.X.L., Y.F.Y. Approval of final manuscript: all authors.