Observations of torsades de pointes during therapy with terfenadine and astemizole has raised concern about the cardiac safety of non-sedating H1-antagonist agents. We compared cetirizine, another compound of that class, to D-sotalol and to astemizole in a model of acquired long QT syndrome. Open-chest surgery was performed in adult beagle dogs anaesthetized with halothane and thiopental. Bradycardia was produced with beta-adrenergic blockade and sinus node crush. Four left ventricular intramyocardial unipolar monophasic action potentials (MAP) were recorded during atrial pacing at basic cycle lengths (BCL) 400-1500 msec, before and during three successive 1-h drug infusions (0.14, 0.45 and 1.4 mg/kg/h for astemizole and cetirizine and 1.1, 2.2 and 4.5 mg/kg/h for D-sotalol). Dose- and bradycardia-dependent prolongations of MAP duration (MAPD) were produced by D-sotalol (P < 0.001) and astemizole (P < 0.001) but not by cetirizine. At BCL 1500 ms, the three infusions of astemizole prolonged endocardial MAPD from 323 +/- 8 msec (mean +/- SE) at baseline to 343 +/- 10, 379 +/- 13 and 468 +/- 26 msec, respectively (n = 9). Sotalol prolonged that MAPD from 339 +/- 6 msec to 377 +/- 7, 444 +/- 15 and 485 +/- 24 msec (n = 7). In contrast, cetirizine did not prolong MAPD: 341 +/- 8 msec at baseline Vs 330 +/- 8, 324 +/- 9 and 323 +/- 11 msec (n = 9). Drug-induced increase in transmural dispersion reached +79 +/- 19 msec after astemizole, +59 +/- 21 msec after D-sotalol and only +7 +/- 11 msec after cetirizine. Runs of ventricular tachycardias and torsades de pointes occurred during dose three of astemizole (5/9 dogs) and D-sotalol (4/7 dogs) but never during cetirizine. In the present model, astemizole and D-sotalol but not cetirizine prolonged MAPD and transmural dispersions of repolarization and produced torsades de pointes. These results suggest that the halothane-anaesthetized bradycardic dog could be a valuable model to discriminate drugs for their class III effects and proarrhythmic potencies.