Liver stiffness and the prediction of clinically significant portal hypertension and portal hypertensive complications. (April 2014)

• Record Type: Journal Article
• Authors: Kitson, Matthew T.; Roberts, Stuart K.; Colman, John C.; Paul, Eldho; Button, Peter; Kemp, William
• Is Part Of: Scandinavian journal of gastroenterology. Volume 50:Number 4(2015)
• Page Start: 462
• Page End: 469
• Publication Date: 2014-04
• Subjects: decompensation -- liver disease -- liver stiffness -- portal hypertension -- transient elastography; Gastroenterology -- Periodicals; Digestive organs -- Diseases -- Periodicals; 616.33
• DOI: 10.3109/00365521.2014.964758
• Languages: English
• ISSNs: 0036-5521

Abstract

Abstract: Objective. Clinically significant portal hypertension (CSPH) is associated with increased risk of liver disease complications, but its identification requires invasive methods. Liver stiffness (LS) measurement via transient elastography correlates with the presence of CSPH. We, therefore, evaluated LS as a noninvasive tool in the prediction of CSPH and portal hypertensive complications. Material and methods. Ninety-five consecutive patients successfully underwent measurement of hepatic venous pressure gradient (HVPG) and LS on the same day. Recent laboratory tests were correlated. Patients were followed up for development of portal hypertensive complications. Predictors of CSPH and complications were identified. Results. Seventy-six (80%) were male and mean age was 56.8 ± 9.3 years. Ninety-three percent and 72% of patients had cirrhosis and esophageal varices, respectively. Only LS (r 2 = 0.38; p < 0.0001) and international normalized ratio (r 2 = 0.21; p = 0.02) were independently associated with HVPG. An LS >29.0 kilopascal (kPa) predicted CSPH with 71.9% sensitivity, 100% specificity, 100% positive predictive value (PPV), and 56.0% negative predictive value (NPV). An LS <25.0 kPa in those with platelet count >150 × 10 9 /L excluded CSPH with 91.7% sensitivity, 100% specificity, 100% PPV, and 90% NPV. Ninety patients were followed up for a median duration of 15.1 months. CSPH and LS >34.5 kPa predicted portal hypertensive complications with 100% and 75.0% sensitivity, 40.3% and 69.4% specificity, 43.1% and 52.5% PPV, and 100% and 86.2% NPV, respectively. Conclusion. LS shows promise as a noninvasive marker of CSPH and portal hypertensive complications. Combining LS with platelet count improves diagnostic accuracy in the exclusion of CSPH.