Baseline characteristics of patients
Between February 2017 and March 2021, 192 patients with NASH cirrhosis or HBV cirrhosis with acute variceal bleeding were enrolled. A total of 38 patients were excluded due to incomplete information, 10 due to portal vein thrombosis, 10 due to previous TIPS treatment, 8 due to hepatocellular carcinoma, 4 due to obviously impaired liver function with a Child–Pugh score >13, and 2 due to ectopic variceal bleeding. Among the 120 included patients, 38 (31.7%) had NASH cirrhosis, and 82 (68.3%) had HBV cirrhosis.
The baseline characteristics of all patients with NASH cirrhosis and HBV cirrhosis are shown in Table 1. The mean age of NASH cirrhosis was 56.7 (interquartile range [IQR], 50-65) years, the mean follow-up time was 27.6 months, and that of HBV cirrhosis was 49.2 years (IQR 44-56) and 24.7 months, respectively. Among patients with NASH cirrhosis, the proportion of females was 55.3%, which was significantly higher than that of patients with HBV cirrhosis (17%, P<0.001). Nineteen patients (50%) with NASH cirrhosis had metabolic syndrome, 18 patients (47.4%) were overweight or obese, 28 patients (73.7%) had diabetes mellitus, 10 patients (26.3%) had hypertriglyceridemia, and 8 patients (21%) had hypertension.
Biochemical analysis of liver function showed that NASH cirrhosis patients had better liver function results and significantly lower Model for End-Stage Liver Disease (MELD) scores (P = 0.034). Alanine aminotransferase (P = 0.031) and the international normalized ratio (P = 0.004) were significantly higher in the HBV group, while the level of gamma-glutamyltransferase (GGT) was significantly higher in the NASH group (P = 0.034). In addition, the platelet count was higher in NASH cirrhosis (P = 0.003).
HVPG/PPG measurement
Patients with NASH cirrhosis had a lower portal pressure (26.3 ± 6.1 vs. 30.1 ± 4.7; P < 0.001), lower WHVP (24.1 ± 5.3 vs. 27.6 ± 5.5; P = 0.001), lower HVPG (15.3 ± 3.8 vs. 18.0 ± 4.8; P = 0.003) and lower PPG (18.0 ± 3.7 vs. 20.0 ± 3.4; P = 0.007) than those with HBV cirrhosis (Table 2). Remarkable correlation between the HVPG and PPG was determined by a Pearson correlation coefficient of 0.78. The R-squared value was 0.608. High HVPG levels were more frequently found in HBV cirrhosis. The HVPG level in 3 (7.9%) NASH patients versus 25 (30.5%) HBV patients was greater than or equal to 20 mmHg (P = 0.006). Low HVPG (<10 mmHg) levels were observed in 3 (7.9%) NASH patients and 3 (3.7%) HBV patients. After successful TIPS treatment, the PPG significantly decreased from 18.0 ± 3.7 mmHg vs. 20.0 ± 3.4 mmHg to 7.6 ± 4.1 mmHg vs. 9.2 ± 3.8 mmHg (NASH cirrhosis vs. HBV cirrhosis,). The PPG level after TIPS treatment of NASH cirrhosis was significantly lower than that after TIPS treatment of HBV cirrhosis (P = 0.04). Compared to the baseline, an PPG value, a mean reduction of 10.4 mmHg was observed in NASH cirrhosis and 10.8 mmHg in HBV cirrhosis. No significant difference was found between them. After TIPS treatment, the PPG effectively decreased to a level of <12 mmHg in 104 (86.7%) patients, but all patients achived a sufficient reduction of PPG of more than 20%. The 1-year shunt dysfunction rate was 94.1% vs. 91.9% and 2-year shunt dysfunction rate was 82.0% vs. 75.4% (NASH group vs. HBV group).
Rebleeding
During the follow-up, a total of 38 patients (11 NASH patients and 27 HBV patients) had at least one variceal rebleeding. Analyzed by competing risk analysis, there were no significant differences in the cumulative probability of rebleeding at 6 months (5.3% vs. 12.2%), 1 year (14.6% vs. 18.6%) or 2 years (22.7% vs. 32.9%) between the two groups. Furthermore, Kaplan–Meier survival curves indicated no significant difference in the overall rebleeding rate between NASH and HBV cirrhosis. Variceal rebleeding patients had higher baseline HVPG and PPG levels than nonrebleeding patients in both groups. Patients with PPG greater than 12 mmHg after TIPS placement were at higher risk for rebleeding than those without (71.2% vs. 50.0%, P = 0.059). Patients with a higher HVPG level of ≥20 mmHg had a significantly higher variceal rebleeding rate than those with an HVPG of <20 mmHg (64.3% vs. 21.7%, P < 0.001). According to competing risk model, the observed cumulative probability of variceal rebleeding was significantly higher in those with an HVPG ≥20 mmHg than in those with an HVPG <20 mmHg at the 6-month (25.0% vs. 5.4%), 1-year (36.2% vs. 11.4%) and 2-year (49.9% vs. 23.0%) follow-ups. In logistic regression, an HVPG ≥20 mmHg was associated with an increased risk of variceal rebleeding (HR 6.48; 95% CI 2.59-16.23; P < 0.001) compared with an HVPG <20 mmHg. The effect with the PPG after TIPS did not reach significance. The c-statistic for baseline HVPG and PPG for predicting variceal rebleeding were 0.82 (95% CI 0.66-0.97; P = 0.002) and 0.72 (95% CI 0.53-0.92; P = 0.027) in NASH patients, and the optimal threshold for baseline HVPG and PPG were ≥17.0 mmHg (specificity 85.2%, sensitivity 72.7%) and ≥20.9 mmHg (specificity 96.3%, sensitivity 54.5% [Fig. 1A]). In the HBV group, the c-statistic for baseline HVPG and PPG for predicting variceal rebleeding were 0.75 (95% CI 0.36-0.86; P < 0.001) and 0.71 (95% CI 0.59-0.83; P = 0.002), and the optimal threshold for baseline HVPG and PPG were ≥21.6 mmHg (specificity 48.1%, sensitivity 92.7%) and ≥20.2 mmHg (specificity 74.1%, sensitivity 63.6% [Fig. 1B]). The ROC cruves of PPG after TIPS placement were not statistically different. Elevation of the baseline HVPG level per 1 mmHg increased the rebleeding risk by 1.50 in NASH cirrhosis (95% CI 1.11-2.03; P = 0.008) and 1.23 in HBV cirrhosis (95% CI 1.09-1.40; P = 0.001). The survival curves of variceal rebleeding in the NASH and HBV groups according to an HVPG are depicted in Fig. 2. The incidence of rebleeding was significantly higher in patients with an HVPG ≥17 mmHg in the NASH group (HR 7.06; 95% CI 1.88-26.56; P = 0.001). Multivariate analysis showed that an HVPG ≥17 mmHg (HR 9.40; 95% CI 1.85-47.70; P = 0.007), lower albumin (HR 1.25; 95% CI 1.06-1.48; P = 0.007), and higher GGT (HR 1.02; 95% CI 1.01-1.03; P = 0.002) were independent predictors of variceal rebleeding in the NASH cirrhosis group (Table 3). The PPG before and after TIPS had no significance in the Cox model. The median survival time was 22 months for NASH patients with an HVPG ≥17 mmHg and 26 months for HBV patients with an HVPG ≥17 mmHg, although no significant difference was found between the two groups.
Other complications of cirrhosis
At enrollment, 19 (50.0%) NASH patients had ascites, 3 (7.9%) had encephalopathy, and 2 (5.3%) had acute-on-chronic liver failure. During a mean follow-up of 27.6 months, 13 (34.2%) patients with NASH cirrhosis developed cirrhotic complications other than variceal bleeding, including encephalopathy (7 patients), HCC (1 patient), acute-on-chronic liver failure (7 patients), and ascites (4 patients). Fifty (61.0%) HBV cirrhosis patients had ascites, 10 (12.2%) had encephalopathy, 7 (8.5%) had liver failure, and 4 (4.9%) had HCC at the time of the first visit of the study. Compared with patients with NASH cirrhosis, 42 (51.2%) patients with HBV cirrhosis developed cirrhotic complications other than variceal bleeding (22 patients with encephalopathy, 10 patients with HCC, 7 patients with liver failure, 8 patients with ascites) during 2 years of follow-up. None of patient received liver transplants. Although the NASH group showed a lower total incidence of cirrhotic complication outcomes, the occurrence rates of hepatic encephalopathy (26.3% vs. 39%), acute-on-chronic liver failure (23.7% vs. 34.1%), and ascites (60.5% vs. 70.7%) were similar between the two groups. According to the Kaplan–Meier analysis, the incidence of HCC was significantly higher in HBV cirrhosis (17.1% vs. 2.6%, P = 0.008). In the NASH group, the HVPG level was significantly higher in patients with cirrhotic complications than in those without complications (16.2 ± 3.9 vs. 13.4 ± 2.8; P = 0.026). The prevalence of cirrhotic complications increased with the HVPG level. Each 1 mmHg elevation in an HVPG was associated with a 27.8% increase in the risk of clinical events (P = 0.035). The c-statistic of an HVPG for the predictive value of cirrhotic complications was 0.75 in NASH patients (95% CI 0.59-0.90; P = 0.014).
During follow-up, 1 patient with NASH died due to liver failure, and 2 patients with HBV died due to HCC and lethal variceal bleeding. Furthermore, there were no differences in survival between the groups.