Electrophysiologic studies have shown that intravenous magnesium sulfate prolongs atrioventricular (AV) nodal conduction and refractoriness and thus could play a role in the management of patients with paroxysmal AV reentrant supraventricular tachycardia (SVT). The present study evaluates the clinical and electrophysiologic effects of intravenous magnesium sulfate in patients with SVT and compares them with those of adenosine triphosphate (ATP), one of the most potent drugs in the treatment of this arrhythmia. Patients with inducible sustained SVT were treated with ATP (10 or 20 mg) and magnesium sulfate (2 g over 15 seconds) during electrophysiologic study. If the tachycardia failed to terminate by the sixth minute, an additional 2 g dose of magnesium was given. ATP (10 or 20 mg) was significantly better than magnesium for terminating induced tachycardias (14 of 14 vs 6 of 14, p < 0.0001). Arrhythmia termination with ATP was due to anterograde AV nodal blockade in all but 1 patient who developed retrograde block over an accessory pathway with decremental conduction. Arrhythmia termination by magnesium was due to retrograde block over an accessory pathway in 3 patients (including the patient with accessory pathway exhibiting decremental conduction), anterograde AV nodal conduction Mock in 2 patients and premature ventricular complexes in 1 patient. During induced tachycardias, only AH intervals were prolonged by ATP, whereas magnesium significantly prolonged AH and QRS intervals. Short-lasting side effects (chest pain, flushing, nausea) occurred after both drugs were administered but were more severe after magnesium. In conclusion, there seems to be little rote for the clinical use of intravenous magnesium sulfate for terminating SVT. The electrophysiologic effects of magnesium sulfate are more complex than previously believed. Magnesium sulfate may block retrograde conduction in accessory pathways.