The analysis of linear/nonlinear coupling between heart rate and QT intervals

Two groups of subjects (21 healthy, 21 hypertensive patients) were tested during two measurement protocols (exercise, tilt). Two models of QT-RR coupling were tested - linear (LM) and nonlinear (NLM). QTc reproducibility was used as evidence of model validity. Results: LM was significantly better (lower QTc irreproducibility) than NLM (P<;0.01) during the tilt-table test. The differences were not significant during the exercise test; NLM was slightly better in hypertensive patients, LM in healthy subjects. Maximal irreproducibility of QTc over subjects was found in both tests with NLM. Conclusion: Evidence of QT-RR nonlinearity cannot be based on minimal error between QT and model QT. QTc reproducibility must also be tested and the excitation specificity of QT-RR coupling must be considered.