Measurement of drug levels is becoming increasingly popular to optimise the dosage of various drugs. In the case of antiarrhythmic drugs, the narrow therapeutic margin of most of these agents and a direct relationship between their pharmacological effects and plasma concentrations would justify more widespread use of monitoring. Optimum plasma concentration ranges have been described for lignocaine (lidocaine), procainamide, quinidine and, more recently, also for disopyramide, mexiletine, tocainide and other new antiarrhythmics. A critical analysis of the original data shows, however, that therapeutic and toxic levels are not so well defined as often assumed: small numbers of patients, marked interindividual variability, sometimes inadequate documentation of arrhythmias and lack of standardised blood sampling characterise many of these studies. Uncertainty about the reliability of concentration-effect relationships also arises when active drug metabolites are identified or there are marked concentration-dependent changes of drug protein-binding. In addition, abolition of various types of arrhythmias might require different drug concentrations. Nevertheless, therapeutic monitoring can be of practical value in patients with life-threatening ventricular arrhythmias and can also greatly facilitate dosage adjustment in cases with renal hepatic or severe cardiac failure. For a correct interpretation of drug levels, the time of blood sampling, dosage regimen, duration of treatment, pharmacokinetic principles, and the clinical condition of the patient must be taken into account. Further studies are needed to define the optimum therapeutic range for several drugs and to evaluate the usefulness of plasma concentration measurements in routine antiarrhythmic treatment.