Low psoas muscle index associates with long-term mortality in cirrhosis: construction of a nomogram

Abstract

Background: To develop a nomogram incorporating indicator of muscle waste to prognosticate long-term mortality in liver cirrhosis (LC), and identify the prognostic impact of psoas muscle index (PMI). Methods: A total of 251 LC patients who underwent computed tomography were included in this study. Multiple Cox regression was performed, and sex-specific nomogram models incorporating PMI were developed. The utility of the proposed models were evaluated by Harrell’s concordance index (C-index), calibration curve and decision curve analysis. X-tile was used to determine optimal cutpoint for stratifying subjects with distinct outcomes. Subgroup analysis was implemented in terms of age and MELD score. The correlation between PMI and gait speed was also evaluated. Results: On multiple analysis, independent predictors for 3-year all-cause mortality were age, BMI, PMI and MELD for males, and age, PMI and MELD for females. Both nomogram models gave rise to moderately strong discrimination, with a C-index of 0.792 (95% CI: 0.723–0.861) in males and 0.715 (95% CI: 0.637–0.793) in females, respectively. The calibration curve implied predicted survival corresponding optimally with the actual outcomes. The proposed models were feasible in clinical settings based on decision curve analysis. On subgroup analysis, PMI might confer valid predictive value on LC patients with MELD <15. Moreover, a definitely positive correlation between PMI and gait speed was revealed. Conclusions: Our proposed nomogram embedding PMI rendered an individualized predictive tool for long-term mortality in LC. The diminishing value of PMI is likely indicative of muscle dysfunction.