Résumé
We aimed to evaluate the correlation between serumascites albumin concentration gradient [SAAG] with esophageal varices [EV] presence and grading, and to assess the relationship between SAAG measurements and the occurrence of gastrointestinal hemorrhage in cirrhotic patients with ascites. Our study included 45 nonalcoholic cirrhotic cases with ascites. They had routine clinical, ultrasonographic and laboratory investigations including ascitic fluid analysis. They had measurement of SAAG computed. An upper gastrointestinal endoscopy was done in all cases to assess the presence and size of EV. 36 of our patients [80%] had EV. The mean SAAG level was 1.46 +/- 0.27 gm/dL for all cases. No correlation was found between SAAG and any of the studied clinical or biochemical parameters. By using the ROC Curve, a SAAG value at a level of [>1.55gm/dL], was a good predictor of the presence of EV with 100% sensitivity and 71.4% specificity. The presence of EV was positively correlated with serum bilirubin, prothrombin time [PT], and spleen size. Meanwhile, it was negatively correlated with serum albumin, serum total protein, platelet count and total protein in ascetic fluid. On univariate analysis of variants associated with the presence of large esophageal varices, only the presence of splenomegaly could predict high grade varices. On comparing patients with and without bleeding varices, the EV grade, portal vein diameter [PVD], spleen size and creatinine level were significantly higher in the group of bleeding varices [p values were 0.002, 0.006, 0.01 and 0.012 respectively] A SAAG score [>/=1.55 gm/dL] is a useful predictor of the presence of EV in cirrhotic patients with ascites. This finding can assist clinicians in determining the urgency of care and referral for upper gastrointestinal endoscopy in cases with ascites. Meanwhile, SAAG was not valuable in screening and predicting complications, such as bleeding from esophageal varices