Objective : Recently it has been recommended that all cirrhotic patients without previous variceal haemorrhage undergo endoscopic screening to detect varices and that those with large varices should be treated with B-Blockers ( American College of Gastro-enterology ). Using the Colour Doppler Velocity Profile (CDVP) , investigation of vascular haemodynamics and their relationship with esophageal varices ( EV ) presence and bleeding in patients with liver cirrhosis can be achieved . Methods: The haemodynamics of the portal vein ( PV) , [ PV diameter, mean velocity, flow value and congestion index ], splenic artery (pulsatility and resistive indices ) and hepatic vein were evaluated in 65 cirrhotic patients and 10 with no chronic liver disease as control volunteers. Esophagogastroduodenescopic evaluation was performed to all subjects. The haemodynamic features were compared between variceal patients without a history of esophageal variceal bleeding ( EVB), those with a history of EVB and those with no varices . A logistic regression model was employed to identify factors associated with EVB and the occurrence of varices. Results : Compared with the size of EV, patients with grade B and C Child-Pugh scores are at risk of having enlarged varices than grade A. The cirrhotic group has a significantly lower mean portal vein flow volume ( P<0.001 ), a significantly elevated mean portal vein congestion index ( P<0.001 ) and diameter ( P<0.05 ) compared with the controls. In the EV individuals, there was a higher incidence of low PV velocity ( P<0.05 ) and high PV congestion index than those without varices. Resistive index of the splenic artery was significantly higher in group with EVB than those without ( P<0.001 ). Logistic regression analysis revealed that, PV congestion index and splenic artery resistive index are independent esophageal variceal presence and bleeding ( respectively ) related factors. Conclusion : Increased resistive index of splenic artery may be a sensitive predictor of variceal bleeding. Portal vein congestion index and flow volume are also two statistically significant factors for EV prediction .