The relation between propafenone dose, serum level, electrocardiographic parameters, antiarrhythmic drug efficacy and adverse effects was studied in 47 children with symptomatic supraventricular arrhythmias aged 1 day to 10.3 years (median 2.2 months) with a mean follow-up of 14.3 months. Propafenone trough serum levels were measured using gas chromatography. Oral propafenone (mean dose 353 mg/m2/day) was effective in 41 of the 47 patients (87.2%). Serum levels did not differ between patients responding and not responding to propafenone (0.45 +/- 0.40 vs 0.36 +/- 0.41 mg/liter). PR interval and QRS complex duration increased more significantly with propafenone dose increments (p < 0.001), than with propafenone serum levels (p < 0.05). At successful treatment PR interval and QRS complex were prolonged by a mean of 19.2 and 20.5% compared with pretreatment status. Five patients exhibited unexpected marked QRS complex prolongation (50 to 200%) despite low propafenone dosage (< 300 mg/m2/day) and level ranging from 0.05 to 1.33 mg/liter. Three patients (6.1%) were suspected of being "poor" metabolizers of propafenone. Mild chronic elevation of serum liver enzymes was observed in 5 patients treated with a larger dose (mean 448 mg/m2/day, p < 0.001). No proarrhythmia was noted on serial Holter monitors. One patient with Wolff-Parkinson-White syndrome and a normal heart had cardiac arrest after aspiration. Serial monitoring of PR interval and QRS complex duration was more useful for proper dosage adjustment than propafenone serum levels. Serum liver enzymes should be closely monitored when using higher propafenone doses. Malignant proarrhythmia could not be excluded in the 1 patient with cardiac arrest.