To improve the clinical value of ambulatory Holter electrocardiographic (ECG) monitoring as a tool of antiarrhythmic therapy control, a new statistical model was developed. In a patient group at increased risk of sudden cardiac death, the spontaneous variability of ventricular arrhythmias was assessed, with simultaneous consideration of single ventricular premature complexes, couplets and salvos. The study included 100 patients who suffered from coronary heart disease or idiopathic dilated cardiomyopathy and for whom greater than 30 ventricular premature complexes/h and couplets had been demonstrated on the last Holter ECG before the study. Between 3 and 12 Holter recordings were made for each patient in a drug-free state; the mean follow-up period was 260 days (maximum 1,403). The mean hourly values of the ectopic events (EE) were assessed separately for ventricular premature complexes, couplets and salvos. The spontaneous variability (SV) was calculated for single ventricular premature complexes, couplets and salvos as SV = log (EEday 2 + 0.01/EEday 1 + 0.01) and linked in one, two and three dimensions. Compared with the consideration of only one type of arrhythmia (one-dimensional model), the simultaneous use of two or three types of arrhythmia (two- or three-dimensional model) resulted in considerably lower reduction and aggravation rates as sufficient proof of drug effects. With control intervals up to 1 week, the one-dimensional model yielded reduction rates for ventricular premature complexes, couplets and salvos of -63%, -90% and -95%, respectively. In contrast, with the three-dimensional model, the rates were -28%, -72% and -88%. The corresponding aggravation values were +370, +1,114% and +2,189% versus +38%, +256% and +747%.(ABSTRACT TRUNCATED AT 250 WORDS)