Noninvasive indices associated with esophageal varices in children with portal hypertension: a single-center exploratory diagnostic accuracy study

Abstract Background Esophageal varices (EV) are a major complication of pediatric portal hypertension. Data on the diagnostic accuracy of noninvasive indices for identifying EV in Egyptian children remain scarce. This exploratory study aimed to evaluate the association between readily available clinical, laboratory, and sonographic parameters and the presence of EV in children with portal hypertension at a tertiary referral center. Patients and methods In this cross-sectional diagnostic accuracy study, 41 children with chronic liver disease, extrahepatic portal vein obstruction (EHPVO), or cirrhosis were consecutively enrolled at Assiut University Children Hospital between January 2022 and January 2023. All participants underwent clinical evaluation, laboratory testing, abdominal and Doppler ultrasonography, and calculation of noninvasive indices including platelet count, spleen-size z -score, platelet/spleen ratio, and Aspartate Aminotransferase-to-Platelet Ratio Index (APRI). Esophagogastroduodenoscopy (EGD) served as the reference standard for diagnosis and grading of varices. Results Esophageal varices were detected in 31 children (75.6%). Compared with those without varices, affected children had significantly lower hemoglobin (9.4 ± 1.8 vs. 11.9 ± 1.3 g/dL, p = 0.027), platelet counts (139.7 ± 87.6 vs. 245.7 ± 61.0 × 10⁹/L, p = 0.001), and albumin (2.9 ± 0.6 vs. 3.9 ± 0.3 g/dL, p < 0.001), with prolonged prothrombin time (14.9 ± 2.9 vs. 12.6 ± 1.1 s, p = 0.021). Sonographic findings showed smaller liver size (9.8 ± 1.4 vs. 12.0 ± 1.9 cm, p = 0.002), larger spleen size (128.2 ± 3.7 vs. 88.1 ± 1.6 mm, p = 0.002), and higher APRI scores (1.69 ± 1.65 vs. 0.28 ± 0.16, p = 0.011). On multivariate analysis, splenomegaly (aOR 8.45, 95% CI 3.21–22.28), elevated APRI (aOR 5.10, 95% CI 2.30–11.30), and thrombocytopenia (aOR 4.82, 95% CI 1.95–11.92) were associated with EV in the exploratory multivariate model, though confidence intervals were wide, reflecting the small sample size. Spleen length ≥ 120 mm demonstrated the highest area under the receiver operating characteristic curve (AUC 0.88; bootstrap-corrected 0.86). In an exploratory etiology-stratified analysis, APRI was significantly associated with EV in the cirrhosis subgroup ( n = 25) but not in the EHPVO subgroup ( n = 16). Conclusion In this single-center exploratory study, splenomegaly, thrombocytopenia, and elevated APRI showed significant associations with the presence of EV in children with portal hypertension. These findings suggest that readily available clinical and laboratory parameters may have potential utility for risk stratification; however, the small sample size, high disease prevalence, and absence of external validation preclude definitive conclusions. Prospective multicenter studies with larger, etiology-stratified cohorts are needed to confirm these observations and establish reliable diagnostic thresholds.