Simple surrogate equations to predict controlled attenuation parameter values for screening non-alcoholic fatty liver disease in a Chinese population

Hanying Liu; Xiao Li; Xiaodong Han; Yan Zhang; Yanting Gu; Lianjie Sun; Junfeng Han; Yinfang Tu; Yuqian Bao; Wenkun Bai; Haoyong Yu

doi:10.3389/fmed.2022.894895

Simple surrogate equations to predict controlled attenuation parameter values for screening non-alcoholic fatty liver disease in a Chinese population

Front Med (Lausanne). 2022 Jul 22:9:894895. doi: 10.3389/fmed.2022.894895. eCollection 2022.

Authors

Hanying Liu¹, Xiao Li², Xiaodong Han³, Yan Zhang², Yanting Gu², Lianjie Sun², Junfeng Han¹, Yinfang Tu¹, Yuqian Bao¹, Wenkun Bai², Haoyong Yu¹

Affiliations

¹ Department of Endocrinology and Metabolism, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai Diabetes Institute, Shanghai Clinical Center of Diabetes, Shanghai Key Laboratory of Diabetes Mellitus, Shanghai Key Clinical Center for Metabolic Disease, Shanghai, China.
² Department of Ultrasound in Medicine, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai Institute of Ultrasound in Medicine, Shanghai, China.
³ Department of General Surgery, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai, China.

Abstract

Objective: Non-alcoholic fatty liver disease (NAFLD) is one of the leading causes of chronic liver disease. The controlled attenuation parameter (CAP) obtained by FibroScan reflects the level of liver steatosis in patients with obesity. Our study aimed to construct a simple equation to predict the CAP, to facilitate the screening and monitoring of patients at high risk for NAFLD.

Methods: A total of 272 subjects were randomly divided into derivation and validation cohorts at a ratio of 1:2. The derivation set was used for constructing a multiple linear regression model; the validation set was used to verify the validity of the model.

Results: Several variables strongly correlated with the CAP were used to construct the following equation for predicting CAP values:CAP1 = 2.4 × BMI + 10.5 × TG+ 3.6 × NC + 10.3 × CP +31.0, where BMI is body mass index, TG is triglyceride, NC is neck circumference and CP is C-peptide. The CAP1 model had an R ² of 0.764 and adjusted R ² of 0.753. It was then simplified to derive CAP2 included only simple anthropometric parameters: CAP2 = 3.5 × BMI + 4.2 × NC + 20.3 (R ² = 0.696, adjusted R ² = 0.689). The data were well fitted by both models. In the verification group, the predicted (CAP1 and CAP2) values were compared to the actual CAP values. For the CAP1 equation, R ² = 0.653, adjusted R ² = 0.651. For the CAP2 equation, R ² = 0.625, adjusted R ² = 0.623. The intra-class correlation coefficient (ICC) values were 0.781 for CAP1 and 0.716 for CAP2 (p < 0.001). The actual CAP and the predicted CAP also showed good agreement in Bland-Altman plot.

Conclusion: The equations for predicting the CAP value comprise easily accessible variables, and showed good stability and predictive power. Thus, they can be used as simple surrogate tools for early screening and follow-up of NAFLD in the Chinese population.

Keywords: FibroScan; controlled attenuation parameter; non-alcoholic fatty liver disease; prediction equation; transient elastography.