BACKGROUND Sedation limits the clinical utility of classical H1 antihistamines, while newer antihistamines such as loratadine and terfenadine are non-sedating. However, clinical use of the terfenadine has been associated with rare but severe cardiac arrhythmias, in particular torsades de pointes. OBJECTIVE To establish a quantitative experimental model for assessing the sedating and cardiotoxicity potential of non-sedating and sedating antihistamines. METHODS Drugs were administered intravenously and the integrated amplitude of the cortical electroencephalogram (EEG) signal was recorded. The threshold dose that depressed EEG activity was compared with the dose required to inhibit by 50% the peripheral bronchospasm elicited by 10 micrograms/kg i.v., of histamine. In separate studies, the electrocardiogram (ECG) and cardiovascular effects of loratadine (30 and 100 mg/kg, i.v.), terfenadine (10 mg/kg, i.v.), promethazine (5 mg/kg, i.v.) and diphenhydramine (20 mg/kg, i.v.) were evaluated. RESULTS The sedating antihistamines, diphenhydramine and promethazine, depressed the integrated EEG at doses between 0.6 and 2.0 times their peripheral antihistamine doses. Loratadine had no EEG depressant activity at 100 mg/kg, i.v., a dose more than 170 times its ED50 (0.58 mg/kg, i.v.) against histamine bronchospasm. We were unable to evaluate the EEG effects of terfenadine, because it produced cardiovascular collapse at 10 mg/kg, i.v. Loratadine and promethazine did not produce adverse cardiovascular effects, nor did they alter normal ECG activity. Diphenhydramine produced bradycardia followed by a transient hypertensive phase without affecting the QTc interval. In contrast, terfenadine elicited hypotension, bradycardia and significant arrhythmogenic activity, causing a prolongation of the QTc interval and a torsades de pointes--like ventricular arrhythmia. Pharmacokinetic studies after i.v. administration of loratadine (30 and 100 mg/kg) demonstrated plasma levels of loratadine and its major metabolite descarboethoxyloratadine to be several orders of magnitude greater than levels found in humans at the clinical dose of 10 mg. CONCLUSIONS The CNS depressant effects of H1 antihistamines are promethazine approximately diphenhydramine >> loratadine = placebo. Of the non-sedating antihistamines, loratadine was devoid of adverse cardiovascular effects whereas terfenadine caused a pronounced disruption of the normal ECG, characterized by a torsades de pointes-like effect.