Hepatic venous pressure gradient after balloon-occluded retrograde transvenous obliteration and liver stiffness measurement predict the prognosis of patients with gastric varices

Figure 1 shows the flow diagram of the patients enrolled in the study. Eighty-one consecutive cirrhosis patients with portal hypertension who therapeutically or prophylactically underwent BRTO for GVs at Hiroshima University Hospital between January 2011 and December 2021 were initially included. Of the 81 patients, 12 and 29 patients were excluded due to a lack of measurement of HVPG before (pre-HVPG) and after BRTO (post-HVPG), respectively. Four patients were excluded due to loss to follow-up. As a result, 36 consecutive patients with both pre- and post-HVPG measurements were enrolled in this study. All patients underwent contrast-enhanced computed tomography (CT) scanning before BRTO to clarify the hemodynamics around the GVs, including the collateral vein and porto-systemic shunt [8]. Similarly, gastrointestinal endoscopic examinations were performed in all patients to confirm the indication of BRTO as GVs bleeding and/or GVs with rapid growth or appearance of a red color (RC) sign, or GVs classified as F3 based on the classification of the Japan Society for Portal Hypertension [24]. Patients in whom the drainage route from the GVs was not via the gastro-renal vein did not undergo BRTO [8].

This study was conducted in accordance with the ethical principles of the Declaration of Helsinki and approved by the Institutional Review Board of Hiroshima University. Written informed consent was obtained from each patient after providing them a detailed explanation of the study.

BRTO was performed using the established procedure, as previously reported [14, 17]. Briefly, all patients underwent selective angiography of the celiac and superior mesenteric arteries, along with CT during arterial portography via the superior mesenteric arteries and/or splenic arteries using an angio-CT system (Aquilion LB, Canon Medical Systems, Ohtawara, Japan) to evaluate porto-systemic collaterals around the GVs via the femoral artery before BRTO. To perform the BRTO procedure, a 5-Fr balloon catheter (Selecon MP catheter; Terumo Clinical Supply, Gifu, Japan) was inserted into the gastro-renal shunt through the inferior vena cava via the right femoral or right jugular vein under local anesthesia. Balloon-occluded retrograde venography was performed to determine the hemodynamics of the GVs and collateral veins after blockade of outflow vessels from the gastro-renal shunt by balloon occlusion, to categorize the GVs according to Hirota’s classification [11]. After the GVs were visualized by retrograde venography under fluoroscopy, the BRTO procedure was performed using sclerosing agents, commonly 10% ethanolamine oleate (Oldamin; Takeda Pharmaceutical, Osaka, Japan) mixed with the same volume of non-ionic contrast medium iopamidol (Iopamiron 300; Bayer Healthcare, Osaka, Japan), with 5% ethanolamine oleate mixed with iopamidol (EOI) under balloon occlusion. When necessary, minor collateral shunt vessels were embolized using a 50% glucose solution or microcoils prior to the 5% EOI injection [25]. To prevent renal dysfunction related to hemolysis, which occurs as an adverse effect of 5% EOI, 2,000 or 4,000 units of haptoglobin was administered intravenously to all patients before BRTO. To avoid incomplete therapeutic efficacy and pulmonary infarction due to an unstable thrombus, the balloon catheter was retained in the draining vein with balloon inflation overnight, and removed after confirmation of complete obliteration of the varices by retrograde venography [25]. When obliteration of the shunt was insufficient on retrograde venography, an additional BRTO was subsequently performed until the inflow vessels disappeared. All patients underwent contrast-enhanced CT scanning and gastrointestinal endoscopy one week after BRTO to clarify the technical success, which was estimated by observation of low attenuation, disappearance of enhancement, and reduction in size of the GVs.

Prior to the BRTO procedure, baseline HVPG was measured using the 5-Fr balloon catheter, which was inserted through the superior or inferior vena cava into the right hepatic venous branch via the right femoral or right jugular vein [23, 26]. After setting the zero-reference point, free hepatic venous pressure and wedged hepatic venous pressure were measured using a nanometer (Polygraph MSC-7000; Fukuda Denshi, Tokyo, Japan) by taking enough time to stabilize the mean pressure. Pre-HVPG was calculated as the difference between the above-measured wedged and free venous pressures. On the day after BRTO, the retained balloon catheter in the draining shunt vein with balloon inflation overnight was removed after confirmation of complete obliteration by retrograde venography. Free hepatic venous pressure, wedged hepatic venous pressure, and HVPG were measured again on the day after BRTO (post-HVPG) before removing the balloon catheter system.

Liver stiffness measurements were made by transient elastography based on ultrasound using a FibroScan-502® Touch system (Echosens, Paris, France) with M-probe and XL-probe. Patients were placed in the supine position with the right hand at maximal abduction for right lobe liver scanning. Measurements were considered valid when there were at least 10 measurements with LSM values of ≥ 60% and an interquartile range of < 30%, and the median value of these measurements was used for analysis.

Endoscopic findings of the EVs and GVs were evaluated according to the classification system of the Japanese Society for Portal Hypertension and Esophageal Varices [24]. The form (F) of EVs was classified as small straight (F1), enlarged tortuous (F2), or large coil-shaped (F3). The RC sign was also classified based on the criteria of the Japanese Society for Portal Hypertension and Esophageal Varices [24]. Endoscopic examinations were performed every 6 months or 1 year after the BRTO to follow-up the EVs and GVs. Worsening of the form and RC sign compared to baseline on follow-up endoscopy was defined as aggravation of EVs. The images were evaluated by two expert endoscopists. Aggravated EVs were treated by endoscopic injection sclerotherapy (EIS) or endoscopic variceal ligation (EVL) when the attending physicians determined that treatment for the EVs was needed.

Laboratory and physical data were collected before BRTO, and survival and EVs aggravation were examined after BRTO, according to our previous report [17]. Quantitative variables were compared using the Wilcoxon signed-rank test. Optimal cut-off points for quantitative variables were determined by receiver operating characteristic curve analysis, if necessary. Cumulative survival and EVs aggravation rates were determined using the Kaplan–Meier method, and cumulative curves between subgroups were compared using the log-rank test. The Cox proportional hazard model was used to estimate the significance of the independent variables. Statistical significance was set at P < 0.05. Statistical analyses were performed using IBM SPSS Statistics for Windows (version 22.0; IBM Corp., Armonk, NY, USA).

Baseline characteristics of the 36 patients who underwent BRTO are shown in Table 1. Of the 36 patients, seven patients (19%) underwent BRTO for GVs bleeding. The remaining 29 patients (81%) were prophylactically treated by BRTO because the GVs were diagnosed to have a high risk for bleeding based on endoscopic findings, such as presence of the RC sign or F3. The seven patients who experienced GVs bleeding underwent endoscopic or balloon tamponade treatment and subsequent BRTO. Clinical success of BRTO was achieved in all patients, as estimated by performing contrast-enhanced CT scans during the one week period following BRTO. None of the patients experienced severe complications related to BRTO treatment.

Changes in HVPG in each patient are shown in Fig. 2. Before BRTO, the measured HVPG was higher than the normal limit (< 5 mmHg) [3], reflecting the presence of portal hypertension, in 31 out of 36 patients. After BRTO, HVPG increased in more than half of the patients. It increased in 21 patients (58.3%), was maintained in six patients (16.7%), and decreased in nine patients (25.0%). Post-HVPG was significantly elevated compared to the value of pre-HVPG (P < 0.01), and was higher than the normal limit in 35 out of 36 patients.

After BRTO, patients were observed for a median duration of 24.5 (3–140) months. Twenty-three patients (63.9%) had co-existence of EVs before treatment, and EVs were exacerbated in 19 patients (52.8%) during the observation period. Among these patients, the exacerbated EVs were endoscopically treated in 12 patients: two patients by EIS and 10 patients by EVL. Cumulative EVs exacerbation rates 1, 3 and 5 years after BRTO were 27, 67 and 73%, respectively (Fig. 3a). Median time to the exacerbation of EVs was 10.5 months. Univariate analysis of the factors associated with EVs exacerbation after BRTO showed that aspartate aminotransferase (AST) and liver stiffness measurement (LSM) correlated significantly with EVs exacerbation (Table 2). Notably, post-HVPG also correlated with EVs exacerbation, although there was no correlation between pre-HVPG and EVs exacerbation. Multivariate analysis of the above factors identified AST ≥ 37 IU/L (hazard ratio [HR] 3.09, 95% CI 1.01–9.47, P = 0.04) and elevated post-HVPG of ≥ 13 mmHg (HR, 3.04, 95% CI 1.25–7.38, P = 0.01) as independent risk factors for EVs exacerbation after BRTO. Evaluation of optimal cut-off points of post-HVPG for predicting EVs exacerbation, as determined by receiver operating characteristic (ROC) curve analysis (Additional file 1: Fig. 1a), indicated an area under the curve (AUC) of 0.79 at the post-HVPG cut-off value ≥ 13 mmHg, with a sensitivity of 0.74 and specificity of 0.71. When the patients were classified by their post-HVPG values, patients with a post-HVPG value of ≥ 13 mmHg were more likely to develop EVs exacerbation compared to patients with post-HVPG values < 13 mmHg (Fig. 3b).

During the observation period, 14 patients (38.9%) died due to various causes: six due to the progression of hepatocellular carcinoma (HCC), two due to infectious diseases, three due to gastroesophageal variceal bleeding, two due to liver failure, and one due to colon cancer. Cumulative survival rates 1, 3 and 5 years after BRTO were 88, 67 and 55%, respectively (Fig. 4a). In terms of factors associated with prognosis, univariate analysis showed that AST, γ-glutamyl transpeptidase and LSM had a significant correlation with poor prognosis after BRTO (Table 3). Multivariate analysis identified an LSM value of ≥ 25.1 kPa as an independent risk factor of poor prognosis after BRTO (HR, 8.27, 95% CI 1.04–65.52, P = 0.04). In addition, ROC curves, which were used to determine the cut-off value of LSM yielding the highest combined sensitivity and specificity with respect to prognosis after BRTO (Additional file 1: Fig. 1b), indicated a value of 25.1 kPa. ROC curve analysis at this cut-off LSM value showed an AUC of 0.78, sensitivity of 0.91 and specificity of 0.65. Interestingly, no patient with LSM < 25.1 kPa died during the observation period (Fig. 4b). In contrast, patients with LSM ≥ 25.1 kPa showed a significantly poor survival rate compared to patients with LSM < 25.1 kPa.