Risk factors for decompensated liver cirrhosis and the construction of a nomogram prediction model
- VernacularTitle:肝硬化失代偿的危险因素及列线图模型构建
- Author:
Yuanyuan LIANG
1
;
Huifang QU
2
;
Xiyue WANG
3
;
Yan WANG
4
;
Hezhao ZHANG
4
;
Jun XU
4
Author Information
- Publication Type:Journal Article
- Keywords: Liver Cirrhosis; Risk Factors; Model for End-Stage Liver Disease; Nomogram
- From: Journal of Clinical Hepatology 2026;42(6):1327-1334
- CountryChina
- Language:Chinese
- Abstract: ObjectiveTo investigate the independent risk factors for decompensation in patients with liver cirrhosis， to construct a nomogram-based risk assessment model， and to assess its risk assessment performance and clinical value by comparing it with Model for End-Stage Liver Disease （MELD）， MELD combined with serum sodium concentration （MELD-Na）， and MELD 3.0 scoring system. MethodsA retrospective analysis was performed for 514 patients with liver cirrhosis who attended The First Hospital of Shanxi Medical University from January 2020 to May 2025， and related data were collected， including demographic data and laboratory markers. According to the presence or absence of decompensation， the patients were divided into compensation group with 275 patients and decompensation group with 239 patients. The Mann-Whitney U test was used for comparison of continuous data between two groups， and the chi-square test was used for comparison of categorical data between two groups. The least absolute shrinkage and selection operator regression analysis was used for screening of variables， and the multivariate logistic regression analysis was used to identify independent risk factors for decompensated liver cirrhosis and construct a nomogram model. The receiver operating characteristic （ROC） curve， calibration curve， and decision curve analysis were used to assess the discriminatory ability， calibration， and net clinical benefit of the model. ResultsThe multivariate logistic regression analysis showed that low hemoglobin （odds ratio ［OR］=0.984， 95% confidence interval ［CI］： 0.969 — 0.999， P<0.05）， low lymphocytes （OR=0.564， 95%CI： 0.383 — 0.830， P<0.05）， high international normalized ratio （OR=3.131， 95%CI： 1.242 — 7.891， P<0.05）， and low serum sodium （OR=0.922， 95%CI： 0.872 — 0.975， P<0.05） were independent risk factors for decompensation events in patients with liver cirrhosis. Construct the prediction model equation： Logit（P）=intercept value－0.016×hemoglobin－0.573×lymphocytes+1.141×INR－0.081×serum sodium. The nomogram constructed based on these factors had a good discriminatory ability， with an area under the ROC curve （AUC） of 0.787， a specificity of 77.8%， and a sensitivity of 67.4%， and it had significantly better predictive performance than MELD score （AUC=0.718）， MELD-Na score （AUC=0.719）， and MELD 3.0 score （AUC=0.725）. The calibration curve showed good consistency between the probability of risk assessed by the model and the observed probability. The decision curve analysis showed that compared with the MELD-based scores， this model provided higher net clinical benefit across a wide range of risk thresholds. ConclusionThis study successfully constructed and validated a nomogram risk assessment model incorporating hemoglobin， international normalized ratio， lymphocytes， and serum sodium. This model has good clinical practicability and can help to achieve early identification of high-risk patients with decompensated liver cirrhosis and optimize intervention strategies in clinical practice.
