Table of Contents
Mechanisms and Prediction of Drug-induced Cardiotoxicity
For more mechanistic detail and clinical evaluation see also Cardiotoxic Drugs
The question of ventricular fibrillation and sudden death induced by the therapeutic use of quinidine has been discussed in the literature since at least the 1920s. It was 1964 when the link between quinidine syncope and ventricular fibrillation was first made and 1966 when what has become the classic form of quinidine induced ventricular arrhythmia “ torsades de pointes” was first described. Many cases of polymorphic ventricular tachycardia (in which the QRS shape varies widely without the classical periodicity of torsades) are probably also manifestations of this phenomenon. Most cases of quinidine induced arrhythmia occur at therapeutic drug concentrations and are not due to drug overdose.
Similar behaviour was reported with disopyramide and subsequently for many other antiarrhythmic agents. Significant worsening of ventricular arrhythmias occurs in 5-15% of patients on an antiarrhythmic drug and if further evidence of the risk was needed, it was provided by the findings of the cardiac arrhythmia suppression trial (CAST) which demonstrated unequivocally the lethal potential of Class I antiarrhythmic agents in patients with chronic ventricular arrhythmias.
MECHANISM OF DRUG INDUCED ARRHYTHMOGENESIS
Arrhythmias are generally considered to arise either from disorders of conduction, in which reentry is the commonest and best known, or from abnormal or inappropriate automaticity. The Vaughan Williams classification of antiarrhythmic drugs is shown in the table.
Mechanisms for induction of reentrant arrhythmias
Drug effects on conduction
A reentrant circuit depends for its continuing existence on a fine balance between conduction time around the circuit and the refractory periods of the various components of the reentrant pathway. If the conduction time ever falls below the refractory period of part of the circuit then the advancing wave front will meet only refractory tissue and the arrhythmia will terminate. Theoretically, therefore an ideal antiarrhythmic agent would tend to accelerate conduction and prolong refractoriness within the substrate for reentry. Many of the drugs available to clinicians prolong refractory periods in myocardium, but none has been convincingly shown to accelerate conduction in therapeutic use. Almost invariably, in fact, conduction tends to be slowed. This combination of conduction slowing and prolongation of refractory period can be either proarrhythmic or antiarrhythmic. With currently available techniques, we really have no way of predicting which of these outcomes is likely for a given drug in a given patient.
Drug effects on refractoriness
While most clinically useful agents prolong refractoriness, lignocaine, mexiletine and tocainide tend to shorten it, particularly in low concentrations. This latter mechanism may explain some cases of drug associated arrhythmogenesis in patients with reentrant tachycardias. Stable reentrant arrhythmias rarely if ever occur in a normal ventricle, presumably because conduction is fast and uniform enough to prevent it. It seems likely that most reentrant ventricular tachyarrhythmias depend upon slow conduction through myocardium that has been partially depressed by ischaemia or some other change (including drug effects) but not depressed to the point where conduction is completely blocked and reentry impossible. Flecainide and other Class Ic drugs have relatively poor selectivity for depolarised tissue and flecainide has recently been shown to render even normal myocardium capable of sustaining ventricular tachycardia both in vitro and in vivo.
Automatic arrhythmias due to antiarrhythmic drugs
Early after depolarisations
Most forms of automaticity known to cause tachyarrhythmia are suppressed by antiarrhythmic drugs. The major exception to this rule is the form of triggered automaticity due to so called “early after depolarisations”. Early after depolarisations (EADs) represent a marked slowing of repolarisation probably due to the reduction of the normal repolarising outward potassium current. If this prolongation is long enough and the voltage conditions appropriate a series of automatic action potentials may be “triggered”. This process can occur as a single event or as an oscillating series of action potentials depending on the prevailing conditions of voltage, calcium concentrations etc.
This phenomenon can be induced in vitro by a number of antiarrhythmic drugs exhibiting Class I properties such as the tricyclic antidepressants and the phenothiazines. This induction of EADs is now thought to be the basis of the drug induced long QT syndromes and their associated arrhythmias including “ torsade de pointes”. According to this theory, the slowing of repolarisation leads directly to the QT prolongation. The EADs are present and they are manifest on the ECG as prominent, bizarre TU waves and, if triggered activity occurs, ventricular tachyarrhythmias result.
Much evidence to support the EAD concept now exists. A number of factors known to be associated with enhanced risk of “torsade de pointes”, such as bradycardia and hypokalaemia, render EAD induction more likely in in vitro studies. Similarly magnesium, lignocaine, and overdrive pacing, which are often of therapeutic value in torsade, reduce or abolish EADs in vitro.
The Vaughan Williams' Class Ia antiarrhythmic agents quinidine, disopyramide and procainamide are all capable of producing EADs and have all been reported to cause torsade de pointes. This is also true of the Class III drugs amiodarone and sotalol, although the incidence appears somewhat less. The Class Ib agents, lignocaine, mexiletine and tocainide do not produce EADs (indeed they tend to reverse them) and have not been convincingly reported to cause torsade de pointes. Arrhythmogenesis due to these agents probably relates to effects on conduction or refractoriness discussed above. The newer Class Ic compounds such as flecainide and encainide do not usually cause significant slowing of repolarisation and have not been definitely shown to cause torsade, although they can certainly lead to proarrhythmic responses. Most of these probably relate to their marked depression of conduction; although it should be noted that there are isolated case reports of marked QT prolongation (not due to QRS widening) for both drugs at therapeutic blood concentrations. Finally, it should be remembered that a number of psychoactive agents including the tricyclic antidepressants have been implicated in causing torsade de pointes. The mechanism is probably the same as for the Class Ia drugs which they closely resemble.
It is now impossible to deny the fact that many, if not all, of the agents that have been used for many years to treat arrhythmias are also quite capable of causing them. There appear to be at least two important cellular mechanisms for proarrhythmic effects. The first of these leads to facilitation of reentry based arrhythmias and is due to critical changes in conduction velocity and refractory period, which render potential reentry pathways more likely to sustain a tachyarrhythmia. This mechanism probably underlies most cases of proarrhythmia produced by traditional Class I antiarrhythmic agents. Since it derives from the same drug effects, which are antiarrhythmic in other contexts, it may well prove to be an insoluble problem. The second mechanism results from excessive prolongation of action potential duration leading to the production of early after depolarisations during the repolarising phase of the action potential. These in turn are thought to underlie the bizarre form of ventricular tachycardia commonly known as torsade de pointes. This problem is particularly seen with Class Ia antiarrhythmic agents and the new Class III agents.