Comparison of transient elastography, serum markers and clinical signs for the diagnosis of compensated cirrhosis

BACKGROUND AND AIMS: Non-invasive diagnosis of compensated cirrhosis is important. We therefore compared liver stiffness by transient elastography, APRI score, AST/ALT ratio, hyaluronic acid and clinical signs to determine which modality performed best at identifying compensated cirrhosis. METHODS: Patients undergoing evaluation at a single center were recruited and had clinical, serological, endoscopy, radiological imaging, liver stiffness measurement and liver biopsy. Patients were stratified into cirrhotic and non-cirrhotic. RESULTS: In 404 patients (124 cirrhosis), transient elastography was diagnostically superior to the other modalities yielding an AUC 0.9 +/- 0.04 compared with hyaluronic acid (AUC 0.81 +/- 0.04: P < 0.05), clinical signs (AUC 0.74 +/- 0.04: P < 0.05), APRI score (AUC 0.71 +/- 0.03: P < 0.05) and AST/ALT ratio (AUC 0.66 +/- 0.03: P < 0.05). The optimum cut-off for transient elastography was 12 kPa giving a sensitivity of 89% and specificity of 87% for cirrhosis. In 238 hepatitis C patients (87 cirrhosis), transient elastography yielded an AUC 0.899 +/- 0.02 for cirrhosis and in 166 non-HCV patients (37 cirrhosis) the results were similar with an AUC 0.928 +/- 0.03; with transient elastography being superior to HA, APRI, AST/ALT and clinical signs for all etiologies of cirrhosis (P < 0.05 for all). Importantly, transient elastography was statistically superior at identifying cirrhosis in 38 biopsy proven Childs Pugh A cirrhotics with no clinical, biochemical or radiological features of cirrhosis or portal hypertension (AUC 0.87 +/- 0.04). CONCLUSION: Transient elastography accurately identified compensated cirrhosis; a liver stiffness of >12 kPa represents an important clinical measurement for the diagnosis of cirrhosis.