Skip to main content
. 2022 Jul 18;13:947157. doi: 10.3389/fendo.2022.947157

Table 6.

The result of the two-piecewise linear regression model.

Incident prediabetes HR, 95%CI P
Fitting model by standard Cox regression 1.111 (1.061, 1.164) <0.001
Fitting model by two-piecewise Cox regression
Inflection points of TG/HDL-c ratio 1.415
 ≤1.415 1.336 (1.134, 1.573) <0.001
 >1.415 1.055 (0.988, 1.126) 0.108
P for log-likelihood ratio test 0.021

We adjusted for age, sex, BMI, SBP, DBP, ALT, AST, sCr, smoking status, drinking status, and a family history of diabetes.