ABSTRACT
Electrocardiographic [ECG] abnormalities are independently associated with poor outcomes in the general population. Their prevalence and determinants were assessed in the understudied African American population with type 2 diabetes. Standard 12-lead ECGs were digitally recorded in 635 unrelated African American-Diabetes Heart Study [AA-DHS] participants, automatically processed at a central lab, read, and coded using standard Minnesota ECG Classification. Age- and sex-specific prevalence rates of ECG abnormalities were calculated and logistic regression models were fitted to examine cross-sectional associations between participant characteristics and ECG abnormalities. Participants were 56% women with a mean age of 56 years; 60% had at least one minor or major ECG abnormality [23% >/= 1 major [or major plus minor], and 37% >/= 1 minor [with no major]]. Men had a higher prevalence of >/= 1 minor or major ECG abnormality [66.1% men vs. 55.6% women, p = 0.0089]. In univariate analysis, age, past history of cardiovascular disease, diabetes duration, systolic blood pressure, sex and statin use were associated with the presence of any [major or minor] ECG abnormalities. In a multivariate model including variables, female sex [OR [95% CI] 0.79 [0.67, 0.93]], statin use [0.79 [0.67, 0.93]] and diabetes duration [1.03 [1.0, 1.05]] remained statistically significant. Nearly three out of five African Americans with diabetes had at least one ECG abnormality. Female sex and statin use were significantly associated with lower odds of any ECG abnormality and diabetes duration was significantly associated with higher odds of any ECG abnormality in the multivariable model