Table 5.

Association between TyG-BMI and reversion to normoglycemia in different sensitivity analyses.

Exposure	Model I (HR.,95%CI) p	Model I I(HR,95%CI) p	Model III(HR,95%CI) p
TyG-BMI (per 10 units)	0.970 (0.963, 0.977) <0.001	0.971 (0.965, 0.978) <0.001	0.970 (0.964, 0.976) <0.001
TyG-BMI quartiles
Q1	Ref	Ref	Ref
Q2	0.964 (0.911, 1.019) 0.195	0.952 (0.901, 1.005) 0.076	0.956 (0.908, 1.006) 0.086
Q3	0.860 (0.806, 0.917) <0.001	0.837 (0.787, 0.891) <0.001	0.858 (0.810, 0.907) <0.001
Q4	0.764 (0.710, 0.822) <0.001	0.789 (0.736, 0.846) <0.001	0.765 (0.718, 0.815) <0.001
P for ternd	<0.001	<0.001	<0.001

Model I involved a sensitivity analysis with participants (N=18,493) who had never smoked. Age, drinking status, HDL-c, AST, sex, DBP, Scr, ALT, family history of diabetes, LDL-c, SBP, and smoking status were all adjusted.

Model II was a sensitivity analysis conducted on participants (N=20,444) who had not consumed alcohol. Age, HDL-c, sex, AST, DBP, Scr, ALT, family history of diabetes, LDL-c, and SBP were all adjusted.

Model III involved a sensitivity analysis after excluding participants with a family history of diabetes (N=24,654). Age, drinking status, sex, HDL-c, AST, DBP, Scr, ALT, LDL-c, SBP, and smoking status were all adjusted.

CI stands for “confidence interval,” and “Ref” stands for “reference”.