Table 4.

Multivariable Adjusted Associations of Selected Risk Factors With Occurrence of New ECG Abnormalities

Variablea	New Major ECG Abnormality		Any New ECG Abnormality
	HR (95% CI)	P Value	HR (95% CI)	P Value
Age, yr (EDIC year 1)	1.04 (1.02–1.06)	0.001	1.02 (1.01–1.03)	<0.001
Sex (male vs female)	0.92 (0.67–1.26)	0.60	1.04 (0.90–1.19)	0.61
Current smokingb (yes vs no)	1.75 (1.22–2.53)	0.002	1.12 (0.93–1.33)	0.23
Weighted mean systolic blood pressureb, mm Hg	1.03 (1.01–1.05)	0.003	1.01 (1.00–1.02)	0.045
Weighted mean HbA1cb, c (per 10% increase)	1.16 (1.04–1.29)	0.008	1.02 (0.98–1.07)	0.33

ECG indicates electrocardiogram; EDIC, Epidemiology of Diabetes Interventions and Complications (EDIC) Study; HbA1c, glycosylated hemoglobin; HR, hazard ratio; HDL, high‐density lipoprotein.

^aIn addition to age and sex, the most significant variables from each category in Table 3 (smoking, blood pressure, lipid, microalbuminuria, and HbA1c) were selected to be included in building separate multivariate models. In the new major ECG abnormality model, the lipid variable (ie, non‐HDL at EDIC year 1 or weighted mean preceding event/censoring) and microalbuminuria (ever) were no longer significant and were deleted after adjusting for smoking and weighted mean systolic blood pressure. In the any new ECG abnormality model, no other risk factors were nominally significant. The proportional hazard assumption was met for all variables in the models.

^bIncluded in the models as time dependent covariate.

^cThe log HbA1c value was used so that the hazard ratio per c‐fold change in risk are c^1.51197 and c^0.23687 where 1.51197 and 0.23687 are the estimated regression coefficient for major and any abnormality, respectively; a value of c that equals 1.1 corresponds to a 10% increase in the HbA1c value.