Open AccessTimely diagnosis and staging of non‐alcoholic fatty liver disease using transient elastography and clinical parameters
Author(s) -
Shieh Christine,
HalegouaDe Marzio Dina L,
Hung Matthew L,
Fenkel Jonathan M,
Herrine Steven K
Publication year - 2020
Publication title -
jgh open
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.546
H-Index - 8
ISSN - 2397-9070
DOI - 10.1002/jgh3.12385
Subject(s) - transient elastography , fatty liver , medicine , steatohepatitis , gastroenterology , fibrosis , univariate analysis , alanine transaminase , hepatology , aspartate transaminase , diabetes mellitus , liver biopsy , multivariate analysis , disease , biopsy , liver fibrosis , endocrinology , biology , biochemistry , alkaline phosphatase , enzyme
Abstract Background and Aim There is no standardized guideline to screen, image, or refer patients with non‐alcoholic fatty liver disease (NAFLD) to a specialist. In this study, we used transient elastography (TE) to examine the fibrosis stages at which patients are first diagnosed with NAFLD. Subsequently, we analyzed metabolic markers to establish cut‐offs beyond which noninvasive imaging should be considered to confirm NAFLD/non‐alcoholic steatohepatitis fibrosis in patients. Methods Charts spanning July 2015–April 2018 for 116 NAFLD patients who had TE performed were reviewed. Univariate and multivariate analysis of metabolic markers was conducted. Results At the first hepatology visit, TE showed 73% F0–F2 and 27% F3–F4. Univariate analysis showed that high‐density lipoproteins (HDL), hemoglobin A1c (A1c), aspartate transaminase (AST), and alanine transaminase (ALT) were significantly different between the F0–F2 and F3–F4 groups. Multivariate analysis showed that AST ( P = 0.01) and A1c ( P = 0.05) were significantly different. Optimal cut‐offs for these markers to detect liver fibrosis on TE were AST >43 U/L and A1c >6.6%. The logistic regression function combining these two variables to reflect the probability ( P ) of the patient having advanced fibrosis (F3–F4) on TE yielded the formula: P = e R /(1 + e R ), where R = −8.56 + 0.052 * AST + 0.89 * A1c. Conclusions Our study suggested that >25% of patients presenting to a specialist for NAFLD may have advanced fibrosis (F3–F4). Diabetes (A1c >6.6%) and AST >43 U/L were the most predictive in identifying NAFLD patients with advanced fibrosis on imaging. We proposed a formula that may be used to prioritize NAFLD patients at higher risk of having advanced fibrosis for specialist referral and imaging follow‐up.