Introduction

Peptic ulcer bleeding (PUB) is a major cause of gastrointestinal-related morbidity and mortality. Various clinical scoring systems such as AIMS65, Glasgow-Blatchford Score (GBS), and ABC score have been used for mortality prediction. However, their predictive performance varies, and improved tools may enhance clinical decision-making. This study aimed to compare the predictive accuracy of traditional scoring systems with machine learning models for 30-day mortality in PUB patients.

Aims & Methods

We retrospectively analyzed data from patients with endoscopically confirmed PUB. Independent variables included clinical parameters, laboratory results, and endoscopic findings. Five conventional scoring systems (GBS, AIMS65, ABC score, Pre- and Post-Rockall scores) were evaluated. A machine learning model based on logistic regression was developed using relevant clinical features. Predictive performance was assessed using area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity. Optimal cut-off values for each score were determined using Youden’s Index.

Results

Among conventional scores, the ABC score achieved the highest AUROC (0.892), followed by AIMS65 (0.864). The machine learning model showed superior performance with an AUROC of 0.989, sensitivity of 66.7%, and specificity of 96.7%. Compared to traditional scoring systems, the machine learning model achieved comparable or higher sensitivity and significantly improved specificity.

Conclusion

Machine learning, particularly logistic regression, provided more accurate prediction of 30-day mortality in PUB compared to conventional scoring systems. These findings support the integration of machine learning models into risk assessment strategies for PUB patients.

References

Laursen, S. B., Oakland, K., Laine, L., Bieber, V., Marmo, R., et al. (2021).
ABC score: a new risk score that accurately predicts mortality in acute upper and lower gastrointestinal bleeding: an international multicentre study.
Gut, 70(4), 707–716.
https://gut.bmj.com/content/70/4/707.abstract

Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) is the best choice to treat choledocholithiasis and obstructive jaundice in pancreatic or biliary neoplasms. Endoscopic ultrasound (EUS) has shown the best diagnostic accuracy for detecting common bile duct (CBD) stones, also showing an essential role for correctly staging pancreatic and biliary neoplasms. EUS and ERCP have been considered historically as two separate procedures; however, during the last decade, a same-session execution has gained popularity. Towards this process, bilio-pancreatic endoscopy was born.

Aims & Methods

We aimed to evaluate the efficacy and safety of same-session EUS-ERCP (SSEE) in treating choledocolithiasis and distal malignant biliary strictures compared to two-session procedures (TSEE). We retrospectively studied all consecutive patients enrolled in a 2-year period (January 2021 and December 2023) undergoing EUS and ERCP to treat choledocolithiasis or obstructive jaundice for malignant distal CBD strictures at an Italian tertiary referral Centre. Patients were divided into two groups: “same session EUS and ERCP group” (SSEE) and “two sessions EUS and ERCP group” (TSEE). Procedural ERCP success, the onset of major adverse events (moderate-to-severe-post-ERCP acute pancreatitis [ASGE guidelines], moderate-to-severe early and/or delayed bleeding [ASGE guidelines], duodenal and/or biliary perforation [Stapfer classification]), and all variables conventionally associated with the procedure were compared between the two groups of therapeutic strategy.

Results

At the end of the study, 200 patients were included (M/F 103/97; mean age: 71 years). The main features of the patients are reported in Table 1. Hundred (50%) patients were treated by the SSEE approach, while the remaining 100 (50%) by the TSEE modality. No statistically significant difference between the two groups was recorded in terms of ERCP overall success (95% vs 99%; p=n.s.). When considering major adverse events, no significant difference was seen between the two groups (0% vs 1%; p=n.s.). Regarding procedural features, the total procedure time (EUS + ERCP) did not significantly differ between the two groups (71.40 min vs 75.05 min; p=n.s.). We did not record major complications (acute pancreatitis, acute cholangitis, acute cholecystitis) during the waiting time between EUS and ERCP (average of 4.03 days) in the TSEE group. By contrast, a mean 3-day gain in terms of hospitalization length was observed in the SSEE group.

Conclusion

Our single-centre experience shows that both SSEE and TSEE approaches are safe and effective in patients undergoing EUS plus ERCP. Delaying the execution of ERCP from EUS, with reduced and appropriate inter-procedural waiting times, proves to be safe and effective, especially in cases involving specific patient conditions (altered anatomy, need for anaesthesiologic assistance, major comorbidities etc.) or the pathology to be treated. Finally, the SSEE approach seems to be more cost-effective with a gain of 3-mean-day in terms of hospitalization length.

Introduction

EUS-guided choledochoduodenostomy using a lumen-apposing metal stent (EUS-CDSL) has been shown in randomized controlled trials (RCT) to be effective compared to endoscopic retrograde cholangiopancreatography (ERCP) in patients with advanced malignant distal biliary obstruction (MDBO). We aimed to further ascertain the safety of EUS-CDSL, which has been limited by inadequate trial sample sizes.

Aims & Methods

We undertook aggregate and individual patient data (IPD) meta-analyses following the PRISMA-IPD statement. A literature search was performed from January 2013 to November 2024 using OVID MEDLINE, EMBASE, Cochrane Library, and ISI Web of Science. Additional searches were performed for abstracts and the grey literature. RCTs comparing EUS-CDSL to ERCP were included. The primary outcome was the rate of procedure-related 14-day severe or fatal adverse events (AE). Secondary outcomes included technical success, 30-day AE, clinical success, stent dysfunction, and procedure time.

Results

A total of 2679 citations were screened with three RCTs included (519 patients). There was no difference in severe or fatal AEs between EUS-CDSL and ERCP (OR 0.64, 95% CI 0.22- 1.83 for IPD data and RR 0.76, 95% CI 0.22, 1.77 for aggregate data). Technical success was greater for EUS-CDSL (OR 4.22, 95% CI 2.35,7.61 and RR 1.14, 95% CI 1.04, 1.25) with shorter procedural time (mean difference -10.50 minutes (95% CI -15.28, -5.73). No significant differences were noted in 30-day AEs, clinical success, and stent dysfunction.

Conclusion

Our aggregate and IPD meta-analyses demonstrate the safety of EUS-CDSL as a first-line alternative to ERCP in advanced MDBO. In addition, EUS-CDSL is technically superior in terms of success rate and procedure time.

Introduction

A biliary stricture is a narrowing of intrahepatic and/or extrahepatic biliary system with upstream dilatation¹; 70-80% of biliary strictures are malignant¹ therefore histological diagnosis is necessary to distinguish benign from malignant causes and guide management. Strictures without confirmed histology are referred to unexplained biliary strictures². ESGE (European Society of Gastrointestinal Endoscopy) guidelines recommend an integrated diagnostic approach that combines multiple techniques to improve diagnostic accuracy. Endoscopic ultrasound(EUS) has demonstrated high sensitivity for evaluating biliary strictures and unexplained common bile duct (CBD) dilatation, due to its ability to provide high-resolution images of CBD walls and surrounding structures, so is the preferred diagnostic modality for both imaging and tissue acquisition(TA), even in absence of abnormal laboratory test.¹ The introduction of ancillary techniques, as DFI (Detective Flow Imaging), which enables the detection of low-velocity blood flow without the use of contrast media, represents a potential useful addition in the diagnostic algorythm³.

Aims & Methods

The aim of this single-centre retrospective study is to validate the role of EUS-DFI in characterizing unexplained biliary strictures and CBD dilatation. We included consecutive patients (pts) with unexplained biliary strictures/CBD dilatation who underwent biliopancreatic EUS between January 2024 and April 2025 at the Digestive Diseases Unit of Sant’Andrea University Hospital, Rome. During EUS, biliary stricture (lesions or thickening of CBD wall) were described as hypoechoic or isoechoic. DFI was performed to evaluate strictures: if microvascularization was observed, vascular pattern was classified as regular or irregular⁴. Results were compared with the final diagnosis, defined by histopathology and/or ≥6 months clinical follow-up as gold standard. TA was performed with EUS-FNB (fine needle biopsy) and/or ERCP(Endoscopic Retrograde Cholangiopancreatography) intraductal biopsies/brushing.

Results

Of 13 patients included, 8(61.5%) were male, median age 72 (range 56-86). All underwent Magnetic Resonance Imaging/Contrast-Enhanced Computed Tomography as first line imaging: 3(23%) had a perihilar stricture, 9(69%) a medio-distal, 1(8%) had “in toto” CBD thickening. Among the 13 strictures, EUS identified 10(76.9%) lesions, with a mean diameter of 17 mm (median 16 mm; range, 5–35 mm), and 3(23.1%) CBD wall thickening. Twelve (92%) strictures were hypoechoic, 1(8%) was isoechoic. EUS-TA was performed in all pts, revealing 7(54%) cases of malignancies, of whome 5 cholangiocarcinoma and 2 pancreatic ductal adenocarcinoma (PDAC), 3(23%) high-grade dysplasia, and 3(23%) fibroinflammatory disease on histopatological evaluation. All pts underwent DFI-EUS. Microvascularization was detected in 12 cases (92%), with an irregular pattern observed in 9 pts. Among these, 5 (55.6%) were diagnosed with cholangiocarcinoma, 1(11.1%) with PDAC infiltrating CBD, and 3(33.3%) with high-grade dysplasia. The 3 benign strictures demonstrated a homogeneous pattern. One case (8%), corresponding to a PDAC, appeared hypovascular on DFI.

Conclusion

In a small cohort of patients with unexplained or indeterminate biliary strictures, the novel DFI-EUS technique proved to be feasible for assessing microvascularization in CBD stenosis and revealed characteristic irregular patterns or hypovascularization in malignant lesions. Further prospective studies with a larger patient population are needed to confirm these preliminary findings.

References

1. Facciorusso, A. et al. Diagnostic work-up of bile duct strictures: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 57, 166–185 (2025).
2. Zhang, W., Chai, N., Wu, Q. & Linghu, E. The new criteria for differential diagnosis of indeterminate biliary stricture under super minimally invasive peroral cholangioscopy. Chin Med J (Engl) 137, 255–256 (2024).
3. Yamashita, Y. et al. A Novel Endoscopic Ultrasonography Imaging Technique for Depicting Microcirculation in Pancreatobiliary Lesions without the Need for Contrast-Enhancement: A Prospective Exploratory Study. Diagnostics 11, 2018 (2021).
4. Yamashita, Y. et al. Novel Technique of Endoscopic Ultrasonography for the Differential Diagnosis of Gallbladder Lesions and Intraductal Papillary Mucinous Neoplasms: A Single-Center Prospective Study. Diagnostics 13, 2132 (2023).

Introduction

Avoidant Restrictive Food Intake Disorder (ARFID) is increasingly recognized within the paediatric population and specialist clinics. However, there is sparse data on the prevalence and burden of ARFID within the adult general population. We addressed this issue.

Aims & Methods

A population-based Internet survey with pre-defined demographic quotas was conducted across the UK and USA in 2023. The survey collected data on demographics, body mass index (BMI), ARFID (using the nine-item ARFID screening tool, NIAS), disorders of gut-brain interaction (using the Rome IV diagnostic questionnaire), PHQ-15 somatic symptoms, PHQ-4 anxiety and depression, PROMIS-10 quality of life, and healthcare utilisation.

Results

In total, 4002 adults (mean age 47 [range 18-91] years and 50% females) completed the survey. Overall, 26.0% (n=1035) screened positive for ARFID, with similar results between the UK (23.0%) and USA (29.0%). The prevalence of positive ARFID screens by NIAS subscale were 15.3% for poor appetite, 13.9% for picky eating, and 6.2% for fear of aversive consequences. In all, 18.2% had a single ARFID domain, 5.7% had two domains and 1.9% had all three domains.
ARFID was more common in females than males (29.6% vs. 22.1%, OR 1.5), and highest amongst people aged 18-39 years (31.6%), followed by those aged 40-64 years (25.0%), and lowest in those 65 years and over (16.1%); p<0.001.
Individuals with ARFID - compared to those without ARFID - had a significantly higher prevalence of disorders of gut-brain interaction (51% vs. 37.5%, OR 2.2), including irritable bowel syndrome (11.8% vs. 4.0%, OR 3.2) and functional dyspepsia (25.8% vs. 7.0%, OR 4.6). They were significantly more likely to be underweight (BMI<18.5; 7.0% vs. 1.5%), experience greater somatic symptoms and mood disturbances, reduced mental and physical quality of life, and increased healthcare utilization (i.e. medication use and doctor visits). An increase in the number of ARFID domains was associated with greater general health impairment; p<0.001.

Conclusion

Elevated ARFID symptoms are common within the adult general population, affecting 1-in-4 people, and associated with substantial health burden. Increased awareness of ARFID will aid clinical service provision and future research on how to best manage this patient group.

Disclosure

Nil

Introduction

Small intestinal bacterial overgrowth (SIBO) is characterized by the presence of excessive gram-negative aerobic and anaerobic colonic bacteria in the small intestine^1,2, leading to increased carbohydrate fermentation and gas production. Symptomatology frequently overlaps with other gastrointestinal disorders³. This study aimed to identify factors associated with test positivity and improve patient selection for diagnostic evaluation.

Aims & Methods

Retrospective, unicentric study. A cohort of 34 patients underwent glucose hydrogen and methane breath testing for SIBO. Demographic, clinical, laboratory data and test results were analyzed.

Results

The cohort included 21 females (61.8%) and 13 males (38.2%), with a mean age of 46.5 ± 15.9 years. Chronic diarrhea was the leading indication for testing (82.4%), followed by abdominal pain (8.8%). Tests results were positive in 4 cases (11.8%). Positive patients had a higher mean age (56.0 ± 10.8 vs. 46.1 ± 15.8 years, p = 0.236) and a comparable gender distribution. Among positive patients, diarrhea was reported in 100% of cases, followed by abdominal distension (75%) and abdominal pain (50%). Folic acid levels were higher in the positive group (11.65 vs. 6.54 ng/mL, p = 0.074), while fecal calprotectin levels were significantly elevated (480 vs. 44.26 µg/g, p = 0.027). Altered small intestine anatomy due to prior surgical interventions was strongly associated with test positivity (p = 0.012), while systemic disorders affecting gastrointestinal motility showed a non-significant trend (p=0.136). Proton pump inhibitor use was not associated with test positivity (p = 1.0). Rifaximin treatment led to symptom improvement in all positive cases, but with no significant difference compared to negative patients (p = 0.444).

Conclusion

Elevated serum folate and fecal calprotectin levels may serve as indicators of underlying SIBO in the appropriate setting. Altered small intestine anatomy was strongly associated with test positivity, and could indicate SIBO under adequate intestinal transit time. These findings highlight the need for larger studies to refine patient selection and improve diagnostic strategies.

References

1. UpToDate. Small intestinal bacterial overgrowth: Etiology and pathogenesis. Available at: https://www.uptodate.com/contents/small-intestinal-bacterial-overgrowth-etiology-and-pathogenesis. Accessed January 19, 2025.
2. Pimentel M, Saad RJ, Long MD, Rao SSC. ACG Clinical Guideline: Small Intestinal Bacterial Overgrowth. Am J Gastroenterol. 2020;115(2):165-178. doi:10.14309/ajg.0000000000000501.
3. Sachdev AH, Pimentel M. Antibiotics for irritable bowel syndrome: Rationale and current evidence. Curr Gastroenterol Rep. 2012;14:439–45.

Introduction

Functional dyspepsia (FD) is a chronic gastrointestinal disorder characterised by meal-related symptoms, which a subset of patients attribute to the consumption of gluten-containing foods [1]. Recent evidence suggests that subtle duodenal immune activation is a feature of FD, as evidenced by elevated eosinophils [2] and impaired epithelial barrier function [3]. FD patients also exhibit an altered duodenal mucosa-associated microbiota (d-MAM) when compared to controls [4]. Given the microbiome’s involvement in the digestion of gluten, we hypothesised that the d-MAM determines the capacity for gluten to disrupt the epithelium and drive immune responses in FD patients.

Aims & Methods

Duodenal biopsies were collected from 10 FD patients and 7 asymptomatic controls. The d-MAM was cultured from biopsies in a habitat simulating media [5] containing 1% wheat-gluten, for up to 48hrs. Microbial cell pellets were frozen for future microbiome sequencing and the cell-free supernatants, containing d-MAM-digested gluten, were collected. Caco2 cells were challenged with the d-MAM-digested gluten, in a Transwell model of the intestinal epithelium (n=3, 2 technical replicates). Transepithelial electrical resistance (TEER) and protein expression of markers of epithelial integrity were measured in Caco2 cells and compared between treatments. Peripheral blood mononuclear cells (PBMCs) from FD patients (n=20) and controls (n=20) were stimulated with pooled samples of FD or control d-MAM-digested gluten, or left untreated, for 44 hours. For comparison to host enzymatic breakdown of gluten, matched PBMCs were also treated with α-chymotrypsin digested, deaminated gluten (CT gluten). The T cell phenotype and presence of activation-induced markers was assessed in PBMCs by flow cytometry.

Results

Challenge of Caco2 cells with gluten digested by the FD d-MAM resulted in a significant increase in Zonula occludens-1 (p=0.007), and reduction in Myd88 (p=0.03), compared to Caco2 cells challenged with control d-MAM-digested gluten. However, TEER readings were not altered between treatment groups in Caco2 experiments. In PBMC experiments, in response to stimulation with FD or control d-MAM-digested gluten, both FD and control PBMCs exhibited heightened expression of activation markers (CD4⁺OX40⁺CD25⁺), when compared to their respective unstimulated cells (p<0.05). When assessing the T-cell phenotype we observed that FD d-MAM-digested gluten promoted an increase in T-helper (Th) 2 cells in both FD (p=0.007) and control (p=0.0001) PBMCs. Meanwhile, control d-MAM-digested gluten stimulation resulted in elevated Th17.1 cells in FD (p<0.0001) and control (p=0.02) PBMCs, when compared to unstimulated cells. Alternatively, CT gluten stimulation did not alter any of the assessed lymphocyte populations in FD or control PBMCs when compared to unstimulated cells, suggesting host digestion of gluten isn’t sufficient to induce immune responses in FD.

Conclusion

Our findings suggest that the d-MAM of FD patients digest gluten in a way that promotes epithelial barrier remodelling and altered immune responses compared to the d-MAM from asymptomatic individuals. The promotion of a Th2 T-cell phenotype by the FD d-MAM is consistent with reports of increased Th2-associated cells, such as eosinophils, in FD patients [6]. These results highlight the likely role of microbiome-gluten interactions in FD, indicating the potential for novel dietary therapies and microbial approaches to management.

References

1. Potter, M., et al., Incidence and prevalence of self-reported non-coeliac wheat sensitivity and gluten avoidance in Australia. Med J Aust, 2020. 212(3): p. 126-131.
2. Shah, A., et al., Duodenal Eosinophils and Mast Cells in Functional Dyspepsia: A Systematic Review and Meta-Analysis of Case-Control Studies. Clinical Gastroenterology and Hepatology, 2022. 20(10): p. 2229-2242.e29.
3. Puthanmadhom Narayanan, S., et al., Duodenal Mucosal Barrier in Functional Dyspepsia. Clin Gastroenterol Hepatol, 2022. 20(5): p. 1019-1028.e3.
4. Shanahan, E.R., et al., Alterations to the duodenal microbiota are linked to gastric emptying and symptoms in functional dyspepsia. Gut, 2023. 72(5): p. 929-938.
5. Schooth, L., Defining the Role of the Mucosa-Associated Microbiota in Digestive Health and Disease Using Novel ex-vivo Combinations of Microbe Culture with Metagenomics. 2024 [Doctoral thesis, The University of Queensland], UQ eSpace.
6. Burns, G.L., et al., Type 2 and type 17 effector cells are increased in the duodenal mucosa but not peripheral blood of patients with functional dyspepsia. Front Immunol, 2022. 13: p. 1051632.

Disclosure

JP, EH, CN, SS, SR, MS, KH, GB, and KD have no disclosures to report.
NT: Financial support from: BluMaiden (microbiome advisory board) (2021), Comvita Mānuka Honey (2021) (digestive health), Biocodex (functional dyspepsia tool), Microba (microbiome advisory board) outside the submitted work. In addition, Dr. Talley has a patent Nepean Dyspepsia Index (NDI) 1998, Biomarkers of IBS licensed, a patent Licensing Questionnaires Talley Bowel Disease Questionnaire licensed to Mayo/Talley, a patent issued, “Diagnostic marker for functional gastrointestinal disorders” Australian Patent Application WO2022256861A1via the University of Newcastle and UniQuest (University of Queensland) and ”Methods and compositions for treating age-related neurodegenerative disease associated with dysbiosis” US Application No. 63/537,725. Dr. Talley is supported by funding from the National Health and Medical Research Council (NHMRC) to the Centre for Research Excellence in Digestive Health and he holds an NHMRC Investigator grant.
SK: Grants from National Health and Medical Research Council (Ideas Grant and Centre for Research Excellence), grants from Viscera Labs (Research contract), grants from Microba Life Science (Research contract), personal fees from Gossamer Bio, personal fees from Anatara Lifescience, personal fees from Immuron, personal fees from Microba Life Science.