To investigate the timing of gastroscopy for gastrointestinal bleeding induced by portal hypertension and the application value of antithrombin Ⅲ in predicting active bleeding.

Clinical data of 152 cirrhosis-induced portal hypertension patients experiencing gastrointestinal bleeding who underwent emergency gastroscopy in Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine from August 2018 to March 2025 were analyzed retrospectively. The informed consents of all patients were obtained and the local ethical committee approval was received. Among them, 104 patients were male and 48 female, aged from 32 to 74 years, with a median age of 57 years. All patients were divided into two groups according to whether active bleeding was found. Emergency detection rate of gastrointestinal bleeding, implementation rate of anesthesia, re-treatment rate, rebleeding rate, hemostasis success rate and incidence of complications were analyzed. All patients were divided into two groups according to whether they were anesthetized or not. The impact of anesthesia on emergency gastroscopy for gastrointestinal bleeding was assessed in portal hypertension patients. The relationship between the main indexes of coagulation function and active bleeding was analyzed.

The detection rate of gastrointestinal bleeding in 152 patients was 48.0%(73/152). The implementation rate of anesthesia in the active bleeding group was 30.1%(22/73), significantly lower than 60.8%(48/79) in the non-active bleeding group (χ²=14.321, P<0.05). The re-treatment rate in the active bleeding group was 84.9%(62/73), significantly higher than 40.5%(32/79) in the non-active bleeding group (χ²=31.732, P<0.05). The rebleeding rate was 23.3%(17/73), significantly higher than 10.1%(8/79) in the non-active bleeding group (χ²=4.782, P<0.05). In the anesthesia group, the re-treatment rate was 42.9%(30/70), significantly lower than 70.0%(64/82) in the non-anesthesia group (χ²=19.819, P<0.05). The rebleeding rate was 8.6%(6/70), significantly lower than 23.2%(19/82) in the non-anesthesia group (χ²=5.857, P<0.05). In the active bleeding group, the antithrombin Ⅲ was 44.3(29.1, 48.4)%, significantly lower than 75.8 (55.9, 88.3)% in the non-active bleeding group, and the difference was statistically significant (Z=-6.095, P<0.05). ROC curve analysis showed that antithrombin Ⅲ had certain predictive and diagnostic value for active bleeding, with an AUC of 0.786, a sensitivity of 0.899, a specificity of 0.784 and an accuracy of 84.2%, respectively. The optimal cut-off value of antithrombin Ⅲ diagnosis was 50.9%.

For patients with gastrointestinal bleeding induced by portal hypertension, significant clinical benefits can be obtained from emergency gastroscopy after full preoperative preparation and anesthesia evaluation when vital signs are stable. Antithrombin Ⅲ has certain predictive value for patients with active bleeding, which can be used as a reference marker for the timing of gastroscopy.