Non Endoscopic Predictors of Esophageal Varices in Cirrhotics

Abstract

Objective: To identify non-invasive predictors of esophageal varices in patients of Liver Cirrhosis. 
Place and duration of study: First Medical unit, Nishtar Hospital Multan from January 2003 to December 2005. 
Patients and methods: Sixty five patients diagnosed as Cirrhosis of liver and without any history of hematemesis 
and/or melena were included in this study. These patients underwent complete clinical, biochemical and 
ultrasonographic evaluation. Four variables selected to predict the presence of esophageal varices were platelet 
count, spleen diameter, portal vein diameter and platelet count/spleen diameter ratio. Upper GI endoscopy was done 
in all these patients to see esophageal varices. 
Results: Thirty three patients had esophageal varices while varices were not found in 32 patients. Best cut off values 
of spleen diameter (160 mm, p-value <0.001) and portal vein diameter (13.5 mm, p-value of <0.001) were 
statistically significant for prediction of presence of varices. Positive and negative predictive values for spleen 
diameter were 90% and 81% respectively. Positive and negative predictive values for portal vein diameter were 74% 
and 77% respectively. Best cut off values of platelet count (145 X 103/µL, p-value-0.486), and platelet count/spleen 
diameter ratio (1200, p-value-0.153) were statistically not significant for prediction of presence of varices. 
Conclusion: Spleen diameter (cut off value 160 mm) and portal vein diameter (cut off value 13.5 mm) have very 
good predictive values (positive and negative) and can be used as predictors for presence of varices in patients of 
cirrhosis with no past history of bleeding.