In a randomized controlled study examining the value of an intramuscular injection of lidocaine in the prehospitalization phase of suspected acute myocardial infarction, paramedics used an automatic injector to administer 400 mg of the drug into the patient's deltoid muscle before transport to the hospital. In a 33-month period, 7026 patients with acute chest pain were seen. Of the 6024 patients randomized (2987 to the lidocaine group and 3037 to the control group), 1935 (32 per cent) proved to have an acute myocardial infarction. In the entire 60-minute period of observation by continuous electrocardiography, primary ventricular fibrillation was observed in 8 treated and 17 control patients (P = 0.08). However, from 15 minutes after randomization onward, when plasma lidocaine levels were in the therapeutic range, only 2 cases of ventricular fibrillation occurred in the treated group, as compared with 12 in the control group (P less than 0.01). Ventricular tachycardia terminated a mean of 10 minutes after injection in six of nine lidocaine-treated patients with acute myocardial infarction but in none of five control patients with infarction (P less than 0.02). Mean plasma lidocaine levels were 3 micrograms per milliliter 11 to 20 minutes after injection in 369 consecutive patients. In 65 patients, levels were below 2 micrograms per milliliter, and in 15 patients, levels were above 6 micrograms per milliliter. Side effects were rare and did not contribute to mortality. We conclude that intramuscular lidocaine may be useful if given by a paramedic, another person, or the patient himself when acute myocardial infarction is suspected outside the hospital.