Publication:

Comparison of Model for End Stage Liver Disease-Na score and Model for End Stage Liver Disease score in predicting in-hospital mortality in patients with end stage liver disease: an observational study

Background Model for End-stage Liver Disease (MELD) score was originally developed to predict mortality after trans jugular intrahepatic portosystemic shunt. Hyponatremia is the most common electrolyte abnormality in End Stage Liver Disease (ESLD). Incorporating serum sodium into MELD score increases its predictive accuracy.

Methods This is an observational study conducted on 50 patients of ESLD admitted from October 2012 to September 2014. Study population was divided into survivor and non-survivor groups. MELD score and Model for End Stage Liver Disease-Na (MELD-Na) score was calculated and compared between the groups.

Results Out of 50 participants, 20 (40%) died in the hospital due to cirrhosis related complications.  The average age was 44.7±12.040 years in the survivor group and 54.1±9.910 years in the non-survivor group. The mean MELD score and MELD-Na score was found to be higher in non-survivors group (28.5 and 30.5) compared to survivors group (22.03 and 25.67) which was statistically very significant. Majority of the patients in survivor group had MELD score between 10-19 (43.3%) and 30-39 (36.7%). In the non-survivor group majority of patients had score of more than 20 (80%). MELD-Na score has better sensitivity (90%) compared to MELD score (80%) at a cut off value above 22. However, MELD score has better specificity (60%) compared to MELD-Na score (43.3%) at the same cut off value.

Conclusions MELD-Na score was higher in non-survivor group with good predictability for in-hospital mortality and there was good correlation between both the scores in terms of degree of agreement and MELD-Na score was more sensitive compared to MELD score.