Can Transient Elastography Replace Liver Histology for Determination of Advanced Fibrosis in Alcoholic Patients A Real-life Study Filip Janssens, MD,* Nicolas de Suray, MD,* Hubert Piessevaux, MD, PhD,* Yves Horsmans, MD, PhD,* Philippe de Timary, MD, PhD,w and Peter Sta ¨rkel, MD, PhD* Background/Goals: Consensus is lacking whether cut-off values for different fibrosis stages using transient elastography (TE, FibroScan) are universally applicable to all liver diseases. We evaluated the performance of TE in predicting severe fibrosis ( ZF3) in alcoholic patients using cut-off values validated for chronic hepatitis C. Study: Patients admitted for alcohol withdrawal were prospectively evaluated by TE and biochemistry for aspartate aminotransferase to platelet ratio index (APRI) and Forns score calculations. If TE revealed severe fibrosis ( ZF3), hepatic venous pressure gradient measurements and transjugular liver biopsy were proposed results of which were correlated and compared with TE measurements or APRI and Forns scores. Results: Among 239 patients, 72 had liver TE scores ZF3 and 23 declined liver biopsy leaving a final study population of 49 patients. Compared with biopsy, 32 patients were correctly classified by TE, whereas 16 patients differed by 2 fibrosis stages yielding a positive predictive value of 65% for liver fibrosis ZF3 at TE. Specificity and sensitivity of TE improved beyond 75% and 70%, respectively, with modified cut-offs of 17 (F3) and 21.1 kPa (F4). Areas under the receiver operating characteristic curves were 0.766 and 0.864 for severe fibrosis ( ZF3) and cirrhosis, respectively. APRI and Forns scores performed less well than TE regarding sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating characteristic values. A significant correla- tion was found between hepatic venous pressure gradients and liver stiffness values at TE. Conclusions: TE with modified cut-offs has the potential to predict advanced fibrosis and significant portal hypertension in alcoholic patients. APRI and Forns scores are of limited usefulness in alcoholics. Key Words: alcoholic liver disease, liver stiffness, APRI, Forns, portal hypertension (J Clin Gastroenterol 2010;44:575–582) A lcohol-induced liver fibrosis is characterized by exces- sive deposition of extracellular matrix components, especially collagen, because of increased matrix production and/or decreased matrix degradation. Progression of fibro- sis to cirrhosis in alcoholic patients who continue drinking is a major cause of morbidity and mortality. The only means to date that might reliably assess liver fibrosis stage in these patients is performance of a liver biopsy with subsequent histologic assessment. Difficulties associated with liver biopsy include a poor patient acceptance, relatively high costs, and its highly invasive nature carrying a substantial risk of complications with morbidity between 0.3% and 0.6% and mortality of 0.05%. 1 Over the last years, many efforts have been devoted to the development of noninvasive markers and tests that might reliably predict fibrosis stages in chronic liver diseases. One of the newer developments consists of measuring liver stiffness measurements (LSMs) by transient elastography (TE, Fibroscan) for noninvasive liver fibrosis assessment. 2 TE has been widely validated in hepatitis C virus (HCV) patients and seems to reasonably well predict histologic fibrosis stages in these patients. 3,4 TE has not yet been extensively studied in alcoholic liver disease (ALD) and there is ongoing debate as to whether cut-off levels for assessment of fibrosis stages in chronic HCV patients can be applied to other chronic liver diseases such as ALD. Therefore, we studied the performance of TE in assessing advanced fibrosis ( ZF3) focusing on the popula- tion of actively drinking alcoholic patients that constitutes more than 80% of alcoholic patients seen by gastroenter- ologists or alcohol addiction specialists in their daily practice. In this particular population, we evaluated the TE cut-offs validated for HCV patients with respect to histology considered as the gold standard. We then examined the performance of 2 easily available, noninvasive biochemical markers, the aspartate aminotransferase (AST) to platelet ratio index (APRI) and the Forns score in the same study population. PATIENTS AND METHODS Study Population Patients admitted between January 1, 2006 and February 29, 2008 to our unit for alcohol detoxification and rehabilitation were included in the study. A complete standard program of alcohol withdrawal was composed of 2 weeks of hospitalization separated by 1 week of Copyright r 2010 by Lippincott Williams & Wilkins Received for publication July 28, 2009; accepted November 15, 2009. From the Departments of *Gastroenterology; and wAdult Psycho- pathology, St Luc University Hospital, Universite´ Catholique de Louvain, Brussels, Belgium. The authors have nothing to declare. In particular, there is no conflict of interest or financial support to be declared for any of the authors. Reprints: Peter Starkel, MD, PhD, Department of Gastroenterology, St Luc University Hospital, Universite´ Catholique de Louvain, Avenue Hippocrate 10, 1200 Brussels, Belgium (e-mail: Peter.Starkel@ uclouvain.be). ORIGINAL ARTICLE J Clin Gastroenterol  Volume 44, Number 8, September 2010 www.jcge.com | 575