BACKGROUND AND STUDY AIMS: Non-invasive predictors of medium/large oesophageal varices (LOVs) could reduce the number of screening endoscopies. As portal hypertension is a consequence of liver fibrosis, serum fibrosis markers were evaluated together with other variables as possible non-invasive predictors of medium OV/LOV. PATIENTS AND METHODS: A total of 154 cirrhotic patients with splenomegaly and 30 healthy control subjects were recruited in a prospective study in two gastroenterology centres in Upper Egypt. Clinical parameters assessed included Child-Pugh class, liver size and ascites. Laboratory parameters included complete blood count, liver function tests, and aspartate aminotransferase (AST)/platelet ratio. Transforming growth factor-β(1) (TGF-β(1)), alpha(2) macroglobulin (A(2)M) and hyaluronic acid (HA) were assayed. Ultrasonographic examination was done for assessment of liver span, portal vein diameter and detection of minimal ascites. Oesophageal varices were diagnosed and graded by oesophagogastroduodenoscopy. RESULTS: Fifty-four patients (35%) had no or small varices and 100 (65%) patients had medium OV/LOV by endoscopy. On multivariate analysis, the independent predictors of medium OV/LOV were the presence of ascites (β=0.258, p=0.047) and serum HA (β=0.449, p=0.009). The receiver operating characteristic curve for HA showed the area under the curve to be 0.916. The sensitivity, specificity, positive and negative predictive values and diagnostic accuracy of HA at a cut-off value of 207μgl(-1) were 94%, 77.8%, 88.7%, 87.5% and 88.3%, respectively. CONCLUSIONS: The presence of ascites and serum HA level higher than 207μgl(-1) can predict the presence of medium OV/LOV in cirrhotic patients. This would help physicians to identify patients who would most likely benefit from screening endoscopy and thus, reduce costs and discomfort from unnecessary endoscopic procedures.
BACKGROUND AND STUDY AIMS: Non-invasive predictors of medium/large oesophageal varices (LOVs) could reduce the number of screening endoscopies. As portal hypertension is a consequence of liver fibrosis, serum fibrosis markers were evaluated together with other variables as possible non-invasive predictors of medium OV/LOV. PATIENTS AND METHODS: A total of 154 cirrhotic patients with splenomegaly and 30 healthy control subjects were recruited in a prospective study in two gastroenterology centres in Upper Egypt. Clinical parameters assessed included Child-Pugh class, liver size and ascites. Laboratory parameters included complete blood count, liver function tests, and aspartate aminotransferase (AST)/platelet ratio. Transforming growth factor-β(1) (TGF-β(1)), alpha(2) macroglobulin (A(2)M) and hyaluronic acid (HA) were assayed. Ultrasonographic examination was done for assessment of liver span, portal vein diameter and detection of minimal ascites. Oesophageal varices were diagnosed and graded by oesophagogastroduodenoscopy. RESULTS: Fifty-four patients (35%) had no or small varices and 100 (65%) patients had medium OV/LOV by endoscopy. On multivariate analysis, the independent predictors of medium OV/LOV were the presence of ascites (β=0.258, p=0.047) and serum HA (β=0.449, p=0.009). The receiver operating characteristic curve for HA showed the area under the curve to be 0.916. The sensitivity, specificity, positive and negative predictive values and diagnostic accuracy of HA at a cut-off value of 207μgl(-1) were 94%, 77.8%, 88.7%, 87.5% and 88.3%, respectively. CONCLUSIONS: The presence of ascites and serum HA level higher than 207μgl(-1) can predict the presence of medium OV/LOV in cirrhotic patients. This would help physicians to identify patients who would most likely benefit from screening endoscopy and thus, reduce costs and discomfort from unnecessary endoscopic procedures.