Assessment of MELD Scores as Predictors of Mortality in Patients with Decompensated Chronic Liver Disease with Variceal Hemorrhage at a third-level care center.

Introduction and Objectives

The variceal hemorrhage is the most common cause of decompensation in patients with chronic liver disease. Hemoglobin level has been used to classify severity; however, it is unreliable. This study aims to evaluate MELD scores as predictors of mortality in patients with variceal hemorrhage.

Materials and Patients

An observational, retrospective, comparative, and longitudinal study was conducted on patients hospitalized in the Gastroenterology department for one year, who were admitted with a diagnosis of variceal hemorrhage and met the criteria for applying the MELD score within the first 48 hours of hospitalization. Descriptive and inferential statistics were performed using ROC curves. Demographic variables, MELD, MELD Na, and MELD Lactate scores were evaluated. The initial and follow-up hemoglobin levels were assessed. Additionally, hospitalization days and discharge reasons were considered.

Results

A total of 96 patients were analyzed (60 women and 36 men) with an average age of 62 ± 8 years. Regarding the etiology of cirrhosis, Alcohol: 48, MASLD: 18, METALD: 8, Viral: 4, Unspecified: 6, Autoimmune: 16. In the analysis of ROC curves, it was found that there was a significant mortality prediction for the MELD, MELD Na, and MELD Lactate models. The MELD cutoff of 21.5 points presented an AUROC of 0.866 (95% CI: 0.71-1.00, p= <0.001), MELD Na of 20.5 had an AUROC of 0.848 (95% CI: 0.67-1.00, p= <0.001), and the MELD Lactate cutoff of 20.5 had an AUROC of 0.791 (95% CI: 0.644-0.939, p= 0.003)

Conclusions

In the analysis of ROC curves, the MELD, MELD Na, and MELD Lactate models demonstrated a significant predictive capacity for mortality. The AUROC values were 0.866, 0.848, and 0.791 respectively, confirming their utility in clinical practice for patients with chronic liver disease admitted in the context of decompensation due to variceal bleeding.

Ethical statement: Data confidentiality and participant protection are guaranteed.

Declaration of interests: None.

Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.