High-dose proton pump inhibitor (PPI) therapy has been recommended to prevent rebleeding of high-risk peptic ulcer (PU) after hemostasis. Vonoprazan has been proven to be noninferior to PPIs in various acid-related diseases. This study aimed to compare the efficacy of vonoprazan vs PPI for preventing high-risk PU rebleeding after hemostasis. A multicenter, randomized, noninferiority study was conducted in 6 centers. Pre-endoscopic and endoscopic therapy were performed according to standard protocol. After successful hemostasis, patients with high-risk PU bleeding (Forrest class Ia/Ib, IIa/IIb) were randomized into 1:1 to receive vonoprazan (20 mg twice a day for 3 days, then 20 mg once a day for 28 days) or high-dose PPI (pantoprazole intravenous infusion 8 mg/h for 3 days, then omeprazole 20 mg twice a day for 28 days). The primary outcome was a 30-day rebleeding rate. Secondary outcomes included 3- and 7-day rebleeding rate, all-cause and bleeding-related mortality, rate of rescue therapy, blood transfusion, length of hospital stay, and safety. Of 194 patients, baseline characteristics, severity of bleeding, and stage of ulcers were comparable between the 2 groups. The 30-day rebleeding rates in vonoprazan and PPI groups were 7.1% (7 of 98) and 10.4% (10 of 96), respectively; noninferiority (within 10% margin) of vonoprazan to PPI was confirmed (%risk difference, -3.3; 95% confidence interval, -11.2 to 4.7; P < .001). The 3-day and 7-day rebleeding rates in the vonoprazan group remained noninferior to PPI (P < .001 by Farrington and Manning test). All secondary outcomes were also comparable between the 2 groups. In patients with high-risk PU bleeding, the efficacy of vonoprazan in preventing 30-day rebleeding was noninferior to intravenous PPI. (ClinicalTrials.gov, Number: NCT05005910).

Background and study aims Endoscopic submucosal dissection (ESD) enables en bloc removal of colorectal neoplasms regardless of size. Submucosal fibrosis is a significant factor for technical difficulty and poor outcomes. We assessed the predictive factors for severe submucosal fibrosis and the ESD outcomes. Patients and methods Patients undergoing ESD from January 2006 to September 2017 were retrospectively reviewed. The degree of submucosal fibrosis was classified into three types: no fibrosis (F0), mild fibrosis (F1), and severe fibrosis (F2). F0 and F1 cases were grouped as non-severe fibrosis for comparison with the severe fibrosis group. Predictors of severe submucosal fibrosis and ESD outcomes were evaluated. Results ESD was performed in 524 lesions (60 % male; mean age, 67.8 years). Eighty lesions with severe fibrosis (15.3 %) were observed. The overall en bloc resection rate and curative resection rate were 94.3 % and 77.7 %, respectively. Rates of en bloc resection (91.2 % vs. 94.8 %, P = 0.2) and perforation (7.5 % vs. 5.6 %, P = 0.45) were no different between severe fibrosis and non-severe fibrosis groups. However, incidences of non-curative resection and low resection speed were significantly higher in the severe fibrosis group. Among protruding lesions, tumor height and volume were significantly greater in the severe counterparts. A diameter ≥ 40 mm, endoscopic finding of the tumor beyond fold, and fold convergence were independent risk factors for severe fibrosis. Conclusions Severe submucosal fibrosis is a significant risk factor for non-curative resection and a long procedural time. Tumor size and morphology might help to predict the severity of fibrosis.

Objective: Age and alarm features are commonly used as indicators for endoscopy in dyspeptic patients; however, the age cut-off and the predictive value of these parameters for identifying upper gastrointestinal (UGI) malignancies are uncertain. Design: Cross-sectional study. Setting: Data were extracted from the Gastrointestinal Endoscopy Centre of Siriraj Hospital, Thailand, during 2005–2011. Participants: Consecutive patients underwent a first-time upper endoscopy for dyspepsia. Patients with previous surgery, suspected UGI malignancy by imaging, or indefinite biopsy results on prior examination were excluded. Main outcome measures: Alarm features included dysphagia, unintentional weight loss, GI bleeding/anaemia, and persistent vomiting. The diagnostic performance of each alarm feature and different age cut-off values were evaluated. Results: A total of 4664 patients (mean age: 52.0±14.4 years, 66% female) were included. Alarm symptoms were presented in 21.6%. The prevalence of active Helicobacter pylori infection was 26.3%. Fifty-eight (1.2%) patients had UGI malignancy. The prevalence of malignancy significantly increased with increasing age (0.6% in patients aged 60 years (p<0.001)). Cancer was found in two patients aged <50 years who did not have alarm features. Patients with alarm features had a higher prevalence of malignancy (OR 22.3, 95% CI 10.5 to 47.4; p10, while dysphagia and GI bleeding/anaemia had a PLR >10 in patients <50 years old. Conclusion: Despite the overall limited value of age and alarm features, persistent vomiting, dysphagia, and GI bleeding/anaemia are strong predictors for malignancy in patients aged <50 years. Without these symptoms, cancer prevalence is negligible; thus, they are worthy guidance for endoscopic evaluation in this age group.

Background and Aim: Treatment of acute lower gastrointestinal bleeding (LGIB) remains problematic, and clinical data is limited compared to that of upper GIB. This study aimed to describe the clinical outcomes and predictors of rebleeding and validate the performance of proposed scoring systems in patients with acute overt LGIB.Methods: Patients with LGIB who underwent colonoscopies between 2013 and 2018 were retrospectively reviewed. Overt LGIB patients who presented within 72 h after bleeding onset were included. Demographics, comorbidities, initial management, endoscopic finding, and treatment outcomes were collected. Factors associated with rebleeding were explored, and the performance of Oakland, NOBLAD, and Strate scores regarding mortality and rebleeding were validated.Results: A total of 537 patients from 3402 (age 72 years, 63–80) were included. Of this, 53% took antithrombotic agents and 59% required red cell transfusion, with a median of 4 red cell units. The most common diagnoses were diverticular bleeding (31.3%) and colorectal polyp/cancer (28.9%). The median time to colonoscopy was 2.3 days, and 80.3% of patients did not receive any hemostatic intervention. The 30‐day mortality and rebleeding were 2.6% and 18.3%, respectively. Patients with radiation proctitis, angioectasia, diverticulosis and using dual antiplatelet drugs were associated with recurrent bleeding. The risk scores showed low performance in predicting recurrent bleeding and mortality.Conclusion: Acute, overt LGIB was common among elders with comorbidities. The rebleeding risk was mostly linked to underlying lesions and the use of antiplatelet drugs. The performance of current risk stratification scores remains unsatisfactory and requires further development.

Background: Acute upper gastrointestinal bleeding (UGIB) is common in clinical practice and has a wide range of severity. Along with medical therapy, endoscopic intervention is the mainstay treatment for hemostasis in high-risk rebleeding lesions. Predicting the need for endoscopic intervention would be beneficial in resource-limited areas for selective referral to an endoscopic center. The proposed risk stratification scores had limited accuracy. We developed a machine learning model to predict the need for endoscopic intervention in patients with acute UGIB. Methods: A prospectively collected database of UGIB patients from 2011 to 2020 was retrospectively reviewed. Patients older than 18 years diagnosed with UGIB who underwent endoscopy were included. Data comprised demographic characteristics, clinical presentation, and laboratory parameters. The cleaned data was used for model development and validation in Python. We conducted 80%–20% split sample training and test sets. The training set was used for supervised learning of 15 models using a stratified 5-fold cross-validation process. The model with the highest AUROC was then internally validated with the test set to evaluate performance. Results: Of 1389 patients, 615 (44.3%) of the cohorts received the endoscopic intervention (293 variceal- and 336 nonvariceal-bleeding interventions). Eighteen features, including demographic characteristics, clinical presentation, and laboratory parameters, were selected as input for 15 machine learning models. The result revealed that the linear discriminant analysis model could achieve the highest AUROC of 0.74 to predict endoscopic intervention. The model was validated with the test set, in which the AUROC was increased from 0.74 to 0.81. Finally, the model was deployed as a web application by Streamlit. Conclusions: Our machine learning model can identify patients with acute UGIB who need endoscopic intervention with good performance. This may help primary care physicians prioritize patients who need referrals and optimize resource allocation in resource-limited areas. Further development and identification of more specific features might improve prediction performance. Trial Registration: None (Retrospective cohort study) Patient & Public Involvement: None