• Wah-Kheong Chan, Sombat Treeprasertsuk, George Boon-Bee Goh, Jian-Gao Fan, Myeong Jun Song, Phunchai Charatcharoenwitthaya, Ajay Duseja, Yock-Young Dan, Kento Imajo, Atsushi Nakajima, Khek-Yu Ho, Khean-Lee Goh, and Vincent Wai-Sun Wong.
• Gastroenterology and Hepatology Unit, Department of Medicine, University of Malaya, Kuala Lumpur, Malaysia. Electronic address: wahkheong2003@hotmail.com.
• Clin. Gastroenterol. Hepatol. 2019 Nov 1; 17 (12): 2570-2580.e37.

Background & AimsMeasuring liver stiffness only in patients with indeterminate or high nonalcoholic fatty liver disease (NAFLD) fibrosis scores (called a 2-step approach) was reported to reduce indeterminate or discordant results while maintaining the accuracy to identify patients with advanced fibrosis. We aimed to validate this approach using data collected from the Gut and Obesity in Asia Workgroup.MethodsWe performed a retrospective analysis of data from 759 patients with biopsy-proven NAFLD (24% with advanced fibrosis), seen at 10 centers in 9 countries in Asia, from 2006 through 2018. By using liver biopsies as the reference standard, we calculated percentages of misclassifications and indeterminate or discordant results from assessments made based on fibrosis scores (NAFLD fibrosis score [NFS] or Fibrosis-4 score) and liver stiffness measurements (LSMs), alone or in combination. The analysis was repeated using randomly selected subgroups with a different prevalence of advanced fibrosis (histologic fibrosis stage ≥F3).ResultsIn groups in which 3.7% and 10% of patients had advanced fibrosis, a 2-step approach (using the NFS followed by LSM only for patients with indeterminate or high NFS) and using a gray zone of 10 to 15 kPa for LSM, produced indeterminate or discordant results for 6.9% of patients and misclassified 2.7% of patients; only 25.6% of patients required LSM. In the group in which 10% of patients had advanced fibrosis, the same approach produced indeterminate or discordant results for 7.9% of patients and misclassified 6.6% of patients; only 27.4% of patients required LSM. In groups in which 24% and 50% of patients had advanced fibrosis, using LSM ≥10 kPa alone for the diagnosis of advanced fibrosis had the highest accuracy and misclassified 18.1% and 18.3% of patients, respectively. These results were similar when the Fibrosis-4 score was used in place of NFS.ConclusionsIn a retrospective analysis, we found that a 2-step approach using fibrosis scores followed by LSM most accurately detects advanced fibrosis in populations with a low prevalence of advanced fibrosis. However, LSM ≥10 kPa identifies patients with advanced fibrosis with the highest level of accuracy in populations with a high prevalence of advanced fibrosis.Copyright © 2019 AGA Institute. Published by Elsevier Inc. All rights reserved.

18 keywords...

hide…