Prediction of anesthetic torsadogenicity using a human ventricular cell model.

This simulation study was designed to predict the torsadogenicity of sevoflurane and propofol in healthy control, as well as type 1 and type 2 long QT syndrome (LQT1 and LQT2, respectively), using the O'Hara-Rudy dynamic model. LQT1 and LQT2 models were simulated by decreasing the conductances of slowly and rapidly activating delayed rectifier K currents (I and I, respectively) by 50%, respectively. Action potential duration at 50% repolarization level (APD) and diastolic intracellular Ca concentration were measured in epicardial cell during administration of sevoflurane (1 ~ 5%) and propofol (1 ~ 10 μM). Torsadogenicity can be predicted from the relationship between APD and diastolic intracellular Ca concentration, which is classified by the decision boundary. Whereas the relationships in control and LQT1 models were distributed on nontorsadogenic side in the presence of sevoflurane at all tested concentrations, those in LQT2 models were shifted to torsadogenic side by concentrations of ≥ 2%. In all three models, propofol shifted the relationships in a direction away from the decision boundary on nontorsadogenic side. Our findings suggest that sevoflurane, but not propofol, exerts torsadogenicity in patients with reduced I, such as LQT2 patients. Caution should be paid to the occurrence of arrhythmia during sevoflurane anesthesia in patients with reduced I.