Background/Aims: We investigated the factors related to spleen length and the diagnostic accuracy of a model using spleen length corrected by related factors, for the prediction of varices needing treatment (VNT).
Methods:Various prediction models for VNT including spleen length were analyzed in the cohort of compensated advanced chronic liver disease (cACLD), defined as liver stiffness (LS) ≥10 kPa in a recent study. The associated factors for spleen length were identified in healthy subjects to improve the prediction of VNT.Results: Among 1,218 cACLD patients, VNT was noted in 249 patients (20.4%). On multivariate analysis, longer spleen length, lower platelet count, and higher LS value were independent predictors for VNT (all p<0.001). In multivariate analysis of 1,041 healthy subjects, age (β=-0.027), sex (β=0.762), and body mass index (β=0.097) were found to be significant factors for spleen length (all p<0.001). Using the β values, the estimated spleen length was calculated. To improve the prediction of VNT, the ratio of measured and estimated spleen length was calculated. Based on binary regression analysis results, the LS value-spleen ratio to platelet score (LSRPS) was calculated as follows: 0.027×LS value (kPa)+2.690×measured/estimated spleen ratio-0.011×platelet count (cells×10 9 /L)-4.215. The area under the receiver operating characteristic of the LSRPS for VNT was 0.820, which was significantly higher than 0.797 of LS value-spleen diameter to platelet ratio score (LSPS) (p=0.006).Conclusions: Spleen length is influenced by age, sex, and body mass index in the Asian population. The LSRPS using the measured/estimated spleen ratio had higher diagnostic accuracy than LSPS in predicting VNT in patients with cACLD.