High-normal serum carcinoembryonic antigen levels and increased risk of diabetic peripheral neuropathy in type 2 diabetes

Table 2 Impacts of serum CEA levels on outcomes of nerve conduction indices by multivariable linear regression analyses

Models	B (95% CI)	β	t	p	Adjusted R²
Composite Z-score of latency
Model 0	0.308 (0.200 to 0.417)	0.270	5.608	< 0.001	0.073
Model 1	0.201 (0.097 to 0.305)	0.176	3.803	0.001	0.239
Model 2	0.141 (0.022 to 0.260)	0.119	2.326	0.021	0.342
Model 3	0.147 (0.020 to 0.274)	0.128	2.286	0.023	0.393
Model 4	0.152 (0.023 to 0.280)	0.132	2.330	0.021	0.416
Composite Z-score of amplitude
Model 0	− 0.324 (− 0.441 to − 0.207)	− 0.263	− 5.461	< 0.001	0.069
Model 1	− 0.197 (− 0.310 to − 0.085)	− 0.160	− 3.462	0.001	0.239
Model 2	− 0.173 (− 0.303 to − 0.042)	− 0.133	− 2.608	0.010	0.341
Model 3	− 0.193 (− 0.338 to − 0.048)	− 0.151	− 2.626	0.009	0.360
Model 4	− 0.209 (− 0.354 to − 0.064)	− 0.164	− 2.838	0.005	0.398
Composite Z-score of NCV
Model 0	− 0.452 (− 0.581 to − 0.324)	− 0.328	− 6.935	< 0.001	0.107
Model 1	− 0.359 (− 0.487 to − 0.230)	− 0.260	− 5.495	< 0.001	0.204
Model 2	− 0.291 (− 0.441 to − 0.140)	− 0.200	− 3.794	< 0.001	0.299
Model 3	− 0.294 (− 0.449 to − 0.127)	− 0.207	− 3.621	< 0.001	0.369
Model 4	− 0.297 (− 0.453 to − 0.129)	− 0.210	− 3.662	< 0.001	0.406

CEA was natural logarithmically transformed for the regression analyses
Model 0: unadjusted
Model 1: adjusted for age, sex, diabetic duration, BMI, SBP, DBP, hypertension, alcohol consumption, and statins uses
Model 2: additionally adjusted for ALT, TBI, albumin, lipid profiles, UA and eGFR
Model 3: additionally adjusted for HbA1c, fasting C-peptide and glucagon
Model 4: additionally adjusted for antidiabetic treatments

ISSN: 1758-5996