The effect of anatomical factors on ECG amplitudes – a cardiac magnetic resonance study

Background Electrocardiographic criteria for left ventricular hypertrophy (ECG-LVH) show poor diagnostic performance. Distance factors, such as the chest wall – left ventricle (CWLV) distance and subcutaneous adipose tissue (SAT) have been proposed as alternative modulators of ECG amplitudes both on a theoretical and a clinical basis. We aimed to investigate the independent associations of CWLV, SAT, and LVM with ECG amplitudes. Methods In this cross-sectional study using cardiac magnetic resonance (CMR) imaging in 49 healthy volunteers (22 female, mean age 37.1 years, median BMI 22.8 kg/m 2 ), we analysed the CWLV distance, SAT and LVM in relation to QRS amplitudes from a resting 12‑lead ECG. Results In a multivariable regression model, CWLV distance, measured with excellent reliability (inter-observer ICC = 0.995, p < 0.001), was the strongest predictor of Sokolow-Lyon voltage (SV1 + RV5), followed by LVM (std. β = −0.48 and 0.38 respectively, both p < 0.002, adj. R 2 = 0.53). SAT correlated with most ECG amplitudes to a similar degree as did LVM, but in the opposite direction. In females, CWLV distance but not LVM correlated with Sokolow-Lyon voltage. Conclusions We found that a non-cardiac variable (CWLV) was a stronger predictor of ECG amplitudes than was LVM, and that SAT was negatively correlated with ECG amplitudes, even in a sample with a normal median BMI. These findings suggest that distance factors may influence the ECG of non-obese persons and contribute to the limited accuracy of ECG-LVH criteria. Our findings also question the validity of Sokolow-Lyon voltage for assessment of LVM in females.