Natural progression and prediction markers in non‐clinically significant oesophageal varices in children

Abstract Objectives Limited literature exists on non‐clinically significant varices (nCSV) and progression in children with portal hypertension (PHT). This study investigates trends and associations in this cohort. Methods This retrospective cohort study analysed 70 children with nCSV undergoing surveillance endoscopy between January 2012 and 2024. Laboratory parameters, prediction scores and fibroscan results were collected. Statistical analysis include Mann–Whitney U test, chi‐squared test and receiver operating characteristic. Results Ten children (14.3%) presented with portal vein thrombosis (PVT), 26 (37.1%) with non‐biliary atresia chronic liver disease (CLD) and 34 (48.6%) with biliary atresia (BA). Twenty‐five children (35.7%) had variceal progression, with median years until progression of 3 years recorded in PVT and CLD (PVT: 1–6 years, CLD: 2–8 years), and 2 years (1–10 years) in BA. Haemoglobin count (Hb) (area under the curve [AUC] = 0.943), risk score (AUC = 0.748), and spleen stiffness by fibroscan (SSM) (AUC = 1.00) revealed optimal accuracy in predicting progression in PVT, with similar findings in CLD (von Willebrand Factor score [vWFAg score]: AUC = 1.00, risk score: AUC = 0.767, SSM: AUC = 0.882). Suboptimal accuracy was seen in BA biomarkers. Conclusions Risk score is a reliable marker to monitor variceal progression in CLD and PVT. Interim noninvasive scores could be trialled along with surveillance OGD to validate results. Caution is advised extending endoscopy period for children with BA. Due to small subgroup sizes, larger cohort studies are needed to validate SSM and vWFAg score in children with nCSV.