Dataset Information

Individual mortality risk predictive system of patients with acute-on-chronic liver failure based on a random survival forest model.

ABSTRACT:

Background

The basis of individualized treatment should be individualized mortality risk predictive information. The present study aimed to develop an online individual mortality risk predictive tool for acute-on-chronic liver failure (ACLF) patients based on a random survival forest (RSF) algorithm.

Methods

The current study retrospectively enrolled ACLF patients from the Department of Infectious Diseases of The First People's Hospital of Foshan, Shunde Hospital of Southern Medical University, and Jiangmen Central Hospital. Two hundred seventy-six consecutive ACLF patients were included in the present study as a model cohort (n = 276). Then the current study constructed a validation cohort by drawing patients from the model dataset based on the resampling method (n = 276). The RSF algorithm was used to develop an individual prognostic model for ACLF patients. The Brier score was used to evaluate the diagnostic accuracy of prognostic models. The weighted mean rank estimation method was used to compare the differences between the areas under the time-dependent ROC curves (AUROCs) of prognostic models.

Results

Multivariate Cox regression identified hepatic encephalopathy (HE), age, serum sodium level, acute kidney injury (AKI), red cell distribution width (RDW), and international normalization index (INR) as independent risk factors for ACLF patients. A simplified RSF model was developed based on these previous risk factors. The AUROCs for predicting 3-, 6-, and 12-month mortality were 0.916, 0.916, and 0.905 for the RSF model and 0.872, 0.866, and 0.848 for the Cox model in the model cohort, respectively. The Brier scores were 0.119, 0.119, and 0.128 for the RSF model and 0.138, 0.146, and 0.156 for the Cox model, respectively. The nonparametric comparison suggested that the RSF model was superior to the Cox model for predicting the prognosis of ACLF patients.

Conclusions

The current study developed a novel online individual mortality risk predictive tool that could predict individual mortality risk predictive curves for individual patients. Additionally, the current online individual mortality risk predictive tool could further provide predicted mortality percentages and 95% confidence intervals at user-defined time points.

SUBMITTER: Zhang ZQ

PROVIDER: S-EPMC8318661 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:To validate prognostic scores for acute decompensation of cirrhosis and acute-on-chronic liver failure in Brazilian patients.This is a prospective cohort study designed to assess the prognostic performance of the chronic liver failure-consortium (CLIF-C) acute decompensation score (CLIF-C AD) and CLIF-C acute-on-chronic liver failure score (CLIF-C ACLF), regarding 28-d and 90-d mortality, as well as to compare them to other prognostic models, such as Model for End-Stage Liver Disease (MELD), MELD Sodium (MELD-Na), Child-Pugh (CP) score, and the CLIF-C Organ Failure score (CLIF-C OF). All participants were adults with acute decompensation of cirrhosis admitted to the Emergency Department of a tertiary hospital in southern Brazil. Prognostic performances were evaluated by means of the receiver operating characteristic (ROC) curves, area under the curves (AUC) and 95%CI.One hundred and thirteen cirrhotic patients were included. At admission, 18 patients had acute-on-chronic liver failure (ACLF) and 95 individuals had acute decompensation (AD) without ACLF, of which 24 eventually developed ACLF during the course of hospitalization (AD evolving to ACLF group). The AD group had significantly lower 28-d (9.0%) and 90-d (18.3%) mortality as compared to the AD evolving to ACLF group and to the ACLF group (both P < 0.001). On the other hand, 28-d and 90-d mortalities were not significantly different between AD evolving to ACLF group and ACLF group (P = 0.542 and P = 0.708, respectively). Among patients with ACLF, at 28 d from the diagnosis, CLIF-C ACLF was the only score able to predict mortality significantly better than the reference line, with an AUC (95%CI) of 0.71 (95%CI: 0.54-0.88, P = 0.021). Among patients with AD, all prognostic scores performed significantly better than the reference line regarding 28-d mortality, presenting with similar AUCs: CLIF-C AD score 0.75 (95%CI: 0.63-0.88), CP score 0.72 (95%CI: 0.59-0.85), MELD score 0.75 (95%CI: 0.61-0.90), MELD-Na score 0.76 (95%CI: 0.61-0.90), and CLIF-C OF score 0.74 (95%CI: 0.60-0.88). The same occurred concerning AUCs for 90-d mortality: CLIF-C AD score 0.70 (95%CI: 0.57-0.82), CP score 0.73 (95%CI: 0.62-0.84), MELD score 0.71 (95%CI: 0.59-0.83), MELD-Na score 0.73 (95%CI: 0.62-0.84), and CLIF-C OF score 0.65 (95%CI: 0.52-0.78).This study demonstrated that CLIF-C ACLF is the best available score for the prediction of 28-d mortality among patients with ACLF. CLIF-C AD score is also useful for the prediction of mortality among cirrhotic patients with AD not fulfilling diagnostic criteria for ACLF, but it was not superior to other well-established prognostic scores.