Association Between HDL Cholesterol and QTc Interval: A Population-Based Epidemiological Study

Previous experimental studies showed that increasing high-density lipoprotein cholesterol (HDL) cholesterol shortens cardiac ventricular repolarization and the QT interval corrected for heart rate (QTc). However, little is known about the epidemiological relationship between HDL and QTc. The potential antiarrhythmic effect of HDL cholesterol remains a speculative hypothesis. In this cross-sectional population based study in adults living in the Italian-speaking part of Switzerland, we aimed to explore the association between HDL cholesterol and the QTc interval in the general population. A total of 1202 subjects were screened. electrocardiogram (ECG) recordings, measurements of lipid parameters and other laboratory tests were performed. QTc was corrected using Bazett’s (QTc(Baz)) and Framingham (QTc(Fram)) formulas. HDL was categorized according to percentile distributions: <25th (HDL-1; ≤1.39 mmol/L); 25th–<50th (HDL-2; 1.40–1.69 mmol/L); 50th–<75th (HDL-3; 1.69–1.99 mmol/L); and ≥75th (HDL-4; ≥2.0 mmol/L). After exclusion procedures, data of 1085 subjects were analyzed. Compared with the HDL reference group (HDL-1), HDL-2 and HDL-3 were associated with a reduction of QTc(Baz) and QTc(Fram) duration in crude (HDL-2, QTc(Baz)/QTc(Fram): β-11.306/–10.186, SE 4.625/4.016; p = 0.016/0.012; HDL-3, β-12.347/–12.048, SE 4.875/4.233, p = 0.012/<0.001) and adjusted (HDL-2: β-11.697/–10.908, SE 4.333/4.151, p < 0.001/0.010; HDL-3 β-11.786/–11.002, SE 4.719/4.521, p = 0.014/0.016) linear regression models in women. In adjusted logistic regression models higher HDL, were also associated with lower risk of prolonged QTc(Baz)/QTc(Fram) (HDL-2: OR 0.16/0.17, CI 0.03–0.83/0.47–0.65; HDL-3: OR 0.10/0.14, CI 0.10–0.64/0.03–0.63) in women. Restricted cubic spline analysis confirmed a non linear association (p < 0.001). The present findings indicate an epidemiological association between HDL cholesterol and QTc duration. To draw firm conclusions, further investigations in other populations and with a prospective cohort design are needed.