Diagnostic accuracy of non-invasive tools to detect esophageal varices keeping esophagogastroduodenoscopy as gold standard in patients with compensated chronic liver disease

Abstract

ABSTRACT Objectives: To determine the diagnostic accuracy of Aspartate Aminotransferase to Platelet Ratio Index (APRI) and Platelet Count to Spleen Diameter Ratio (PC/SD) having upper esophagogastroduodenoscopy findings as the reference standard for predicting esophageal varices in patients with compensated chronic liver disease. Methodology: This prospective cross-sectional validation study was carried out at the Department of Medicine, Khyber Teaching Hospital (KTH) Peshawar from August 2025 to January 2026. A total of 161 consecutive patients aged 20 to 70 years with compensated chronic liver disease (CLD) who fulfilled the eligibility criteria were enrolled using non probability consecutive sampling technique. Patients with peptic ulcer disease, contraindications to upper gastrointestinal endoscopy, hepatocellular carcinoma or altered upper gastrointestinal anatomy were excluded. APRI and PC/SD ratio were calculated for all participants and upper gastrointestinal endoscopy was performed as the reference standard. Diagnostic performance was evaluated by calculating sensitivity, specificity, positive predictive value, negative predictive value and overall diagnostic accuracy. Results: A total of 161 patients with CLD were included, of whom 101 (62.7%) had esophageal varices (EVs) on upper gastrointestinal endoscopy. The APRI score demonstrated 100% sensitivity, 65.0% specificity and an overall diagnostic accuracy of 87.0%, whereas the PC/SD ratio showed 51.5% sensitivity, 100% specificity, and 69.6% accuracy. Both markers demonstrated excellent discriminatory performance, with AUCs of 0.943 for APRI and 0.939 for the PC/SD ratio (p = 0.315). APRI was significantly more sensitive, while the PC/SD ratio was significantly more specific for detecting EVs (both p < 0.001). Increasing APRI values and decreasing PC/SD ratio values were significantly associated with higher grades of EVs (p < 0.001). Conclusion: Both the APRI and PC/SD ratio demonstrated acceptable diagnostic performance for identifying EVs in patients with CLD. These inexpensive and readily available biomarkers may assist in risk stratification and prioritizing patients for endoscopic evaluation particularly in resource limited settings. However, upper gastrointestinal endoscopy remains the reference standard for the diagnosis of EVs.