ABSTRACT
Background: Many scoring systems have been widely applied in recent years for assessing and predicting the mortality across a broad spectrum of liver diseases. The two most commonly used prognostic models are the Child-Turcotte-Pugh score and the more recently described the model of end-stage liver disease [MELD] score and its derivatives
Aim of the work: The main objective of this work was to evaluate and compare the accuracy of the Child-Pugh score, the MELD score, and the new scores, MELD-Na, the integrated MELD [iMELD], and the MELD-to-sodium [MESO] index in the prediction of liver cirrhosis complications
Patients and methods: This study was conducted on 60 patients with chronic liver disease. They were divided into two groups: group I consisted of 17 [28.3%] patients with uncomplicated liver cirrhosis. They were 10 [58.8%] male and seven [41.1%] female patients. Their mean age was 58.7 +/- 10.7 years. Group II consisted of 43 [71.7%] patients with liver cirrhosis-related complications. They were 37 [86%] male and six [13.9%] female patients. Their mean age was 55.1 +/- 6.38 years. All were recruited from the Internal Medicine and Hepatology Department in Ain Shams University Hospitals. They were subjected to full medical history, clinical examination, lab investigations, upper gastrointestinal endoscopy, and imaging studies to assess the state of the liver disease and its complications. The MELD, its derivatives, and the Child-Turcotte-Pugh score were calculated for assessing their prognostic accuracy in relation to complications of chronic liver disease
Results: In this study, the most common complication was portal hypertension [70%], which is the leading cause for ascites and hematemesis due to esophageal varices. In this study, there was a highly significant statistical difference between group I and group II regarding albumin and International normalization ratio and a significant statistical difference between group I and group II regarding total bilirubin, platelets, and the sodium level [Na]. In this study, there was also a highly significant statistical difference between group I and group II regarding Child, MELD, MELD-Na, MESO index, and iMELD scores. Regarding the comparison between Child, MELD, MELD-Na, MESO index, and iMELD scores as predictors for liver cirrhosis complications, our study showed that the Child-Pugh score was the most sensitive score with the highest predictive accuracy. The Child score at a cutoff greater than 7.5 has a sensitivity of 97.7%, a specificity of 70.6%, and an accuracy of 0.95 in predicting the occurrence of complications. MELD at a cutoff greater than 12.5 has a sensitivity of 93%, a specificity of 70.6%, and an accuracy of 0.913 in predicting the occurrence of complications. MELD-Na at a cutoff greater than 16 has a sensitivity of 90.7%, a specificity of 70.6%, and an accuracy of 0.867 in predicting the occurrence of complications. The MESO index at a cutoff greater than 0.95 has a sensitivity of 90.7%, a specificity of 70.6%, and an accuracy of 0.919 in predicting the occurrence of complications. iMELD at a cutoff greater than 30.6 has a sensitivity of 95.3%, a specificity of 41.2%, and an accuracy of 0.841 in predicting the occurrence of complications. For the detection of the most common complication, that is, esophageal varices, the MELD score was found to be the most specific, with a specificity of 100% and a sensitivity of 43% and the cutoff point greater than 18. This was followed by the MESO index, with a specificity of 100% and a sensitivity of 39.2% and the cutoff point greater than 1.4, whereas other scores could not be significantly related to the prediction of such complications
Conclusion: This study concluded the superiority of the Child score over MELD, MELD-Na, MESO index, and iMELD scores in the prediction of liver cirrhosis-related complications, whereas the MELD score and the MESO index were the best for predicting the occurrence of esophageal varices