Validity of fatty liver prediction scores for diagnosis of fatty liver by Fibroscan
Seogsong Jeong a , b , Sun Jae Park a , Seong Kyun Na c , Sang Min Park a , d , Byung-Cheol Song e , Yun Hwan Oh f , g , ∗
a Department of Biomedical Sciences, Seoul National University Graduate School, Seoul 03080, Korea
b Department of Biomedical Informatics, CHA University School of Medicine, CHA University, Seongnam 13488, Korea
c Department of Internal Medicine, Inje University College of Medicine, Seoul 50834, Korea
d Department of Family Medicine, Seoul National University Hospital, Seoul National University College of Medicine, Seoul 03080, Korea
e Department of Internal Medicine, Jeju National University Hospital, Jeju National University College of Medicine, Jeju 63241, Korea
f Department of Family Medicine, Jeju National University Hospital, Jeju National University College of Medicine, Jeju 63241, Korea
g Department of Family Medicine, Chung-Ang University Gwangmyeong Hospital, Chung-Ang University College of Medicine, Gwangmyeong 14353, Korea
∗Corresponding author at: Department of Family Medicine, Chung-Ang University Gwangmyeong Hospital, Chung-Ang University College of Medicine, Gwangmyeong 14353, Korea.
E-mail address: swimayo@gmail.com (Y.H. Oh).
Abstract：Background: The Korea National Health and Nutrition Examination Survey nonalcoholic fatty liver disease (K-NAFLD) score was recently developed with the intent to operationally define nonalcoholic fatty liver disease (NAFLD). However, there remained an external validation that confirmed its diagnostic performance, especially in patients with alcohol consumption or hepatitis virus infection.
Methods: Diagnostic accuracy of the K-NAFLD score was evaluated in a hospital-based cohort consisting of 1388 participants who received Fibroscan®. Multivariate-adjusted logistic regression models and the contrast estimation of receiver operating characteristic curves were used for validation of the K-NAFLD score, fatty liver index (FLI), and hepatic steatosis index (HSI).
Results: K-NAFLD-moderate [adjusted odds ratio (aOR) = 2.53, 95% confidence interval (CI): 1.13-5.65] and K-NAFLD-high (aOR = 4.14, 95% CI: 1.69-10.13) groups showed higher risks of fatty liver compared to the K-NAFLD-low group after adjustments for demographic and clinical characteristics, and FLI-moderate and FLI-high groups revealed aORs of 2.05 (95% CI: 1.22-3.43) and 1.51 (95% CI: 0.78-2.90), respectively. In addition, the HSI was less predictive for Fibroscan®-defined fatty liver. Both K-NAFLD and FLI also demonstrated high accuracy in the prediction of fatty liver in patients with alcohol consumption and chronic hepatitis virus infection, and the adjusted area under curve values were comparable between K-NAFLD and FLI.
Conclusions: Externally validation of the K-NAFLD and FLI showed that these scores may be a useful, noninvasive, and non-imaging modality for the identification of fatty liver. In addition, these scores also predicted fatty liver in patients with alcohol consumption and chronic hepatitis virus infection.
2024, Vol. 23 
