Systemic infusion of methacholine has been used to facilitate induction of atrial fibrillation. However, the dose-response relationship, reproducibility and effect of anesthetic agents on induction are not well understood. The use of methacholine to facilitate electrical induction of sustained (>10 minutes duration) atrial fibrillation or flutter was examined. In 25 dogs induction of atrial arrhythmias was attempted using a series of ten 50 Hz trains of 10 seconds duration delivered via an endocardial catheter in the baseline anaesthetized state and subsequently in the presence of graded doses of intravenous methacholine (maximum 5 microg/kg/min). Studies were repeated in 13 dogs to assess reproducibility. Twelve dogs (48%) had inducible sustained atrial flutter or fibrillation lasting greater than 10 minutes in the baseline state. During infusion of methacholine the remaining 13 (52%) dogs also had inducible sustained atrial flutter or fibrillation (mean infusion rate 1.6 +/- 1.9 microg/kg/min). Induction of sustained atrial flutter or fibrillation was reproducible in all but one dog. The type of anesthetic did not significantly affect inducibility. Induction of prolonged atrial fibrillation or flutter is possible in the baseline anaesthetized state in approximately half of dogs using high frequency programmed electrical stimulation. The yield of inducible sustained atrial fibrillation or flutter with programmed stimulation during intravenous infusion of methacholine was increased to 100%. Induction of sustained atrial fibrillation or flutter was highly reproducible.