Studies on the efficacy of phenytoin administered directly into the cerebrospinal fluid in protecting against digoxin-induced arrhythmias were carried out in 20 anesthetized dogs. Phenytoin was administered in an average dose of 10 mg directly into the cisterna magna of 10 dogs. The other control dogs received only the vehicle for phenytoin intrathecally. Both groups of dogs subsequently received a toxic dose of digoxin (0.2 mg/kg) intravenously. The time after intravenous digoxin to the onset of ventricular premature beats, ventricular tachycardia and ventricular fibrillation (VF) was significantly shorter in control animals compared to the phenytoin-treated dogs. In the control group, one dog survived the 3-h observation period without developing ventricular fibrillation, whereas 5 of the 10 phenytoin-treated dogs survived this period without VF (P less than 0.05). Phenytoin had a similar protective effect against digoxin-induced arrhythmias in 10 other dogs that received phenytoin intravenously. In the animals that received phenytoin intrathecally, plasma concentrations of phenytoin were undetectable. Thus, in this experimental model, phenytoin exerts a protective effect against digoxin-induced ventricular arrhythmias which is mediated via the central nervous system.