Table 4 Threshold effect analysis of ALT/HDL-C on the prevalence of NAFLD in different g genders, ages, and BMI using the two-piecewise linear regression model.

From: Utility of the serum alanine aminotransferase to high density lipoprotein cholesterol ratio in evaluating nonalcoholic fatty liver disease and liver fibrosis

Prevalence of NAFLD

Adjusted OR (95% CI), P value

All participants

 

Fitting by the standard linear model

2.915 (2.290, 3.709), < 0.001

Fitting by the two-piecewise linear model

Inflection point

1.054

ALT/HDL-C < 1.054

11.099 (7.734, 15.929), < 0.001

ALT/HDL-C > 1.054

0.821 (0.624, 1.080), 0.158

Log likelihood ratio

< 0.001

Men

 

Fitting by the standard linear model

2.345 (1.747, 3.148), < 0.001

Fitting by the two-piecewise linear model

Inflection point

1.240

ALT/HDL-C < 1.240

5.951 (3.889, 9.107), < 0.001

ALT/HDL-C > 1.240

0.902 (0.677, 1.202), 0.483

Log likelihood ratio

< 0.001

Women

 

Fitting by the standard linear model

3.665 (2.369, 5.670), < 0.001

Fitting by the two-piecewise linear model

Inflection point

0.528

ALT/HDL-C < 0.528

125.411 (49.341, 318.755), < 0.001

ALT/HDL-C > 0.528

0.837 (0.530, 1.323), 0.483

Log likelihood ratio

< 0.001

Age 40–59 years

 

Fitting by the standard linear model

3.950 (2.319, 6.727), < 0.001

Fitting by the two-piecewise linear model

Inflection point

1.029

ALT/HDL-C < 1.029

11.743 (5.655, 24.384), < 0.001

ALT/HDL-C > 1.029

0.426 (0.164, 1.106), 0.079

Log likelihood ratio

< 0.001

BMI > 30 Kg/m2

 

Fitting by the standard linear model

1.912 (1.381, 2.646), < 0.001

Fitting by the two-piecewise linear model

Inflection point

1.042

ALT/HDL-C < 1.042

9.400 (5.136, 17.202), < 0.001

ALT/HDL-C > 1.042

0.924 (0.440, 1.940), 0.150

Log likelihood ratio

< 0.001

  1. Age, gender, race, hypertension, BMI, T2DM, smoke, physical activity level, WC, LSM, DBP, SBP, CRP, fast glucose, fast insulin, HbA1c, TBIL, ALP, GGT, AST, TG, and SUA were adjusted. In the subgroup analysis for gender, the model was not adjusted for gender; in the subgroup analysis for age, the model was not adjusted for age; and in the subgroup analysis for BMI, the model was not adjusted for BMI.