A Retrospective Analysis of 581 Cases of Esophagogastric Variceal Bleeding Complicated by Portal Vein Thrombosis at a Single Center.

Abstract

Background: Esophagogastric variceal bleeding (EGVB) and portal vein thrombosis (PVT) are common complications of liver cirrhosis and frequently coexist, with reported prevalence rate up to 32.5%. Their coexistence complicates clinical management and is associated with adverse outcomes. Early identification of PVT may improve patient management and prognosis.

Objective: To investigate the prevalence, clinical characteristics, and risk factors of PVT in patients with EGVB and to develop a nomogram for the early identification of PVT.

Methods: This retrospective case-control study included 581 consecutive cirrhotic patients with EGVB treated at a single center. Patients were classified into PVT and non-PVT groups according to imaging findings. The demographic characteristics, clinical manifestations, laboratory parameters, and treatment histories were compared between the two groups. Risk factors associated with PVT were identified using logistic regression analysis, and a predictive nomogram was subsequently developed to estimate the probability of PVT occurrence.

Results: Among the 581 patients, 201 patients (34.6%) had concomitant PVT. The median age was 55 years (interquartile range, 48.0-64.0 years), and 384 (66.1%) were male. Compared with the non-PVT group, patients with PVT more frequently presented with abdominal pain (15.9% vs. 10.3%, p = 0.048) and hepatic hydrothorax (18.9% vs. 12.4%, p = 0.034) than the non-PVT group. Multivariable analysis identified previous splenectomy (OR = 4.615, 95% CI: 2.418-8.806; p < 0.001) and endoscopic treatment (OR = 2.414; 95% CI 1.644-3.545; p < 0.001) were as independent risk factors for PVT. Patients with Yerdel Grade IV had a significantly higher prevalence of splenectomy than those with Yerdel Grade I (59.1% vs. 23.4%). Similarly, the prevalence of PVT was significantly higher among patients with a history of endoscopic treatment than among those without (46.9% vs. 26.6%, p < 0.001). D-dimer and fibrin degradation product (FDP) levels were significantly elevated in the PVT group. Receiver operating characteristic analysis identified optimal cutoff values of 0.995 μg/mL for D-dimer and 2.665 μg/mL for FDP. A nomogram incorporating significant predictors demonstrated moderate discrimination for identifying PVT (AUC = 0.751).

Conclusion: PVT was present in approximately one-third of cirrhotic with EGVB. Previous splenectomy and endoscopic intervention were independently associated with PVT, while abdominal pain and hepatic hydrothorax were more common clinical manifestations. The nomogram showed moderate diagnostic performance and may facilitate early identification of PVT; however, external validation is required before routine clinical application.