Integrating Psoas metrics and ascites into MELD-Na for survival prediction after TIPS: a temporal validation study
Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Background Accurate risk prediction following transjugular intrahepatic portosystemic shunt (TIPS) remains challenging. We evaluated the incremental prognostic value of incorporating continuous height-normalized transverse psoas muscle thickness (TPMT/H) and clinically significant ascites into the Model for End-Stage Liver Disease–sodium (MELD-Na), assessing temporal generalizability in a distinct later cohort. Methods This retrospective study included 328 patients with cirrhosis undergoing TIPS at the First Affiliated Hospital of the University of Science and Technology of China (USTC) between April 2015 and May 2022. Cox proportional hazards models adding TPMT/H, ascites, or both (MELD-NSA) to MELD-Na were derived in an earlier cohort ( n = 163) and evaluated without recalibration in a temporally distinct validation cohort ( n = 165). Model discrimination, prediction error, and calibration were evaluated at 6, 12, and 24 months. Results In the temporal validation cohort, MELD-NSA markedly improved prognostic discrimination, elevating the 24-month concordance index from 0.627 to 0.723 (paired difference: 0.096; 95% confidence interval 0.012–0.167). The C-indices of MELD-NSA at 6 and 12 months were 0.738 and 0.734, respectively. Furthermore, MELD-NSA demonstrated superior overall performance, yielding a lower 24-month Brier score (0.159 vs 0.177) and robust risk-group stratification across cohorts. Concurrently integrating continuous TPMT/H and ascites provided superior incremental prognostic value compared with individual additions. Conclusions Integrating quantitative muscle metrics and ascites burden with MELD-Na significantly enhances post-TIPS survival discrimination across calendar periods. This readily accessible, multidimensional risk profile provides a robust foundation for refining individualized patient counseling and post-procedural management.
Source: Research Square
Institution: University of Science and Technology of China
Language: English
Subfield: Hepatology
Field: Medicine
Domain: Health Sciences
Open Access status: green
Award: G8423720858
OpenAlex ID: W7225904005
