Type 2 diabetes individuals have a shorter QRS duration, prolonged QTc interval, flatter T wave, and reduced left ventricular cycle volumes compared to matched controls in the UK Biobank

Abstract Background Type 2 diabetes is associated with an increased risk of heart failure even in the absence of overt coronary artery disease, though underlying mechanisms are unclear. ECG and cardiac magnetic resonance (CMR) imaging provide windows into subclinical changes in myocardial structure and function and could be exploited to understand this. Purpose We investigate differences in ECG and CMR biomarkers in a large cohort of participants with type 2 diabetes, compared to a matched control cohort. We hypothesise that individuals with diabetes may exhibit concurrent differences in biomarkers, reflecting underlying cardiac remodelling. Methods We carried out a pair-matched cross-sectional study to examine the association between type 2 diabetes and multi-modal cardiac biomarkers. The exposure cohort consists of 1,781 UK Biobank participants with an imaging and ECG visit, prevalent type 2 diabetes and no known cardiovascular events at the time of visit. The control cohort is matched on age, sex and body mass index (BMI), and consists of 1,781 participants without diabetes. ECG and CMR biomarkers derived from UK Biobank data were used as outcome variables. We built four multiple multivariate linear regression models, incrementally adjusted for socio-demographic, lifestyle, and clinical covariates. Results Both cohorts were mostly male (63.6%), had a median age of 67 years [IQR: 72-85] and a median BMI of 27.8 kg/m2 [24.6-31.0]. Individuals with type 2 diabetes were more likely to be non-white (7.9% vs 2.7%, p<0.001), current or previous smokers (43.6% vs 40.6%, p=0.04), have lower diastolic blood pressure (79 [72-85] vs 81 [74-88] mmHg, p<0.001), and take cholesterol medication (72.3% vs 27.2%, p<0.001); anti-hypertensives (56.3% vs 27.3%, p<0.001); insulin (13.7% vs 0.1%, p<0.001); and metformin (53.1% vs 0.1%, p<0.001). They had a higher ventricular rate, shorter QRS duration, longer QTc interval, and flatter T wave amplitude, as well as lower left ventricular end-diastolic and end-systolic volumes, among other differences (Table 1). Covariates including cholesterol medication, metformin, and HbA1c were excluded as they were strongly correlated with diabetes, age and/or sex (Pearson correlation coefficient >0.3 or <-0.3). After full adjustment, we found that the exposure variable, diabetes, had a statistically significant association with ventricular rate, QRS duration, T wave offset, T wave amplitude in lead V3, left ventricular stroke volume and global average thickness (p<0.05) (Table 2). Conclusion Our results reveal that individuals with type 2 diabetes exhibit a shorter QRS, longer QTc, and lower end-diastolic and end-systolic volumes, compared to matched controls. This may be indicative of subclinical cardiac abnormalities, both electrophysiological and mechanical. Our association analysis suggests that other factors correlated with diabetes may be also responsible for certain biomarker differences and their underlying causes.