Spleen and Liver Stiffness to Detect Esophageal Varices in Children with Biliary Atresia.

OBJECTIVES To determine the accuracy of non-invasive parameters such as liver (LS) and spleen stiffness (SS) to detect esophageal varices (EV) in children with biliary atresia (BA). METHODS Children with BA between 2000 and 2015 were recruited. All underwent esophagogastroduodenoscopy and transient elastography. Demographic data, laboratory investigations, alanine transferase-to-platelet ratio index (APRI), and Varices Prediction Rule (VPR) score were collected. RESULTS A total of 51 children (mean age 10.63 years, standard deviation [SD]=6.08 years; 53% males) were enrolled. There were differences in onset and outcome of portoenterostomy, spleen palpablility, platelet count, albumin, LS, SS, and VPR between the varice and varice-free groups (P < 0.05). In the varice group, the median LS was 18.12 [interquartile ratio, IQR 13.15-19.12] and the median SS was 46.85 [IQR 25.95-54.55] kPa. In the varice-free group, the median LS was 7.85 [IQR 5.88-16.75] and the median SS was 16.54 [IQR 11.75-21.75] kPa. Both LS and SS were higher in the varice than the varice-free group (P < 0001). The area under the receiver operating characteristic curve of LS, SS, spleen palpability, platelet count, APRI and VPR were 0.734, 0.870, 0.817, 0.810, 0.751 and 0.794, respectively. Using a cut-off value of 12.5 kPa for LS, the sensitivity and specificity were 80% and 70% respectively. Using a cut-off value of 28.9 kPa for SS, the sensitivity and specificity were 75% and 87% respectively. Combination of LS and SS to diagnose varices increased the specificity to 93%. CONCLUSIONS SS as a single marker had the best diagnostic value to predict esophageal varices in children with BA. The combination of SS and LS furthermore increased the diagnostic yield.