BACKGROUND: Large esophageal varices (LOV) were diagnosed by endoscopy in patients with cirrhosis. Noninvasive method would be valuable. AIMS: To evaluate the diagnostic performance of Fibroscan for LOV prediction and to investigate the prognostic value of liver stiffness (LS) in cirrhosis. PATIENTS AND METHODS: One hundred and eighty-three patients with cirrhosis (103 alcohol, 58 viral, and 22 others) underwent an endoscopy and a Fibroscan. Of those patients, 41 (22.4%) had LOV. RESULTS: Median LS was 33.66 kPa (range: 12-75), higher in patients with LOV than those without (51.24 +/- 1.61 vs. 29.81 +/- 1.82 kPa, p < 0.0001), and in alcoholic than nonalcoholic (40.39 +/- 1.75 vs. 25.73 +/- 1.82, p < 0.0001). In whole population, a LS > or =48 kPa predicted LOV with sensitivity, specificity, positive, negative predictive values (PPV, NPV) of 73.2, 73.2, 44.1, and 90.4%, respectively, and an area under ROC curve (AUROC) of 0.75 (CI 95%: 0.69-0.82). For alcoholic cirrhosis, LS was > or =47.2 kPa with sensitivity, specificity, PPV, NPV of 84.6, 63.6, 44, and 92.5%, respectively, AUROC 0.77 (0.68-0.85). For viral cirrhosis, a LS > or =19.8 kPa generated diagnostic values of 88.9, 55.1, 26.7, and 96.4% and 0.73 (0.60-0.84). Sixteen (8.75%) patients died at 1 year. In multivariate analysis, LS was not predictive of mortality. CONCLUSIONS: Etiology of cirrhosis has strong impact on LS cutoff for diagnosis of LOV. Studies should be performed with homogenous cirrhosis etiology.
BACKGROUND: Large esophageal varices (LOV) were diagnosed by endoscopy in patients with cirrhosis. Noninvasive method would be valuable. AIMS: To evaluate the diagnostic performance of Fibroscan for LOV prediction and to investigate the prognostic value of liver stiffness (LS) in cirrhosis. PATIENTS AND METHODS: One hundred and eighty-three patients with cirrhosis (103 alcohol, 58 viral, and 22 others) underwent an endoscopy and a Fibroscan. Of those patients, 41 (22.4%) had LOV. RESULTS: Median LS was 33.66 kPa (range: 12-75), higher in patients with LOV than those without (51.24 +/- 1.61 vs. 29.81 +/- 1.82 kPa, p < 0.0001), and in alcoholic than nonalcoholic (40.39 +/- 1.75 vs. 25.73 +/- 1.82, p < 0.0001). In whole population, a LS > or =48 kPa predicted LOV with sensitivity, specificity, positive, negative predictive values (PPV, NPV) of 73.2, 73.2, 44.1, and 90.4%, respectively, and an area under ROC curve (AUROC) of 0.75 (CI 95%: 0.69-0.82). For alcoholic cirrhosis, LS was > or =47.2 kPa with sensitivity, specificity, PPV, NPV of 84.6, 63.6, 44, and 92.5%, respectively, AUROC 0.77 (0.68-0.85). For viral cirrhosis, a LS > or =19.8 kPa generated diagnostic values of 88.9, 55.1, 26.7, and 96.4% and 0.73 (0.60-0.84). Sixteen (8.75%) patients died at 1 year. In multivariate analysis, LS was not predictive of mortality. CONCLUSIONS: Etiology of cirrhosis has strong impact on LS cutoff for diagnosis of LOV. Studies should be performed with homogenous cirrhosis etiology.
Authors: Leire Pérez-Latorre; Antonio Rivero-Juárez; Víctor Hontañón; Cristina Díez; Francisca Cuenca; Maria Luz Martín-Carbonero; María L Montes; José M Bellón; Teresa Aldámiz-Echevarría; Ana Carrero; Antonio Rivero; Juan González-García; Juan Berenguer Journal: Open Forum Infect Dis Date: 2016-10-03 Impact factor: 3.835