Not everything about atrial fibrillation is a hot topic: drug-induced atrial fibrillation is an exception

Keywords::

• adverse effect
• atrial fibrillation
• cardiac safety
• drug-induced

Atrial fibrillation (AF) is the most common sustained tachyarrhythmia in the general population. It has become an “epidemic” since the beginning of this millennium mainly due to the continuous increase of the incidence and prevalence of AF worldwide [1]. Atrial fibrillation affects more than 2.5 million people in the United States and this number might reach 3.0 million by 2020 and 5.6 million by 2050 [2]. Individuals with AF have an increased incidence of mortality and stroke and a decreased cardiac function and quality of life. Even though 10 – 15% of AF cases occur in the absence of comorbidities, AF is often associated with both cardiovascular and non-cardiac abnormalities/conditions. The transition from a normal sinus rhythm to AF can be spontaneous or be induced by different physical and chemical factors such as hypothermia or medications. On the other hand, cardiac safety concern arising from the proarrhythmic effect is one of the leading causes for the recall of marketed drugs and the abandonment of drug development programs. Currently, the US Food and Drug Administration (FDA) requires both pre-clinical and clinical screenings of the proarrhythmic effect in the evaluation of most new drugs with a focus on the QT interval prolongation and torsades de pointes. However, QT prolongation is not the only mechanism through which a drug-related cardiac electrical disorder develops. Other types of non-QT-mediated arrhythmias have been reported in case series or clinical studies as cardiovascular side-effects induced by a drug—AF is one of them [3]. Dr Tamargo et al. performed a comprehensive, up-to-date review on the topic of “Drug-induced Atrial Fibrillation” which was published in recent issue of Expert Opinion on Drug Safety [4]. The authors should be commended for having provided a spectrum of almost all AF-inducing drugs and for explaining the possible mechanisms by which the drugs are associated with AF.

1. Difficulties in determining the relationship between AF episodes and certain drugs

To determine the proarrhythmic side-effect of a drug during clinical practice, it is expected that a certain type of arrhythmia reproducibly occurs in a portion of patients who are taking the same drug. The abnormal rhythm is supposed to convert to normal heart rhythm and not reoccur if the drug is discontinued. Other comorbidities and factors that may induce the arrhythmia should be excluded. Evidence from bench research linking the drug to the arrhythmia is also required. Cisapride and terfenadine are two examples matching all these criteria. Some patients on cisapride or terfenadine show QT interval prolongation on ECG and may experience syncope and/or die suddenly secondary to torsades de pointes. Subsequently, it has been demonstrated that these two drugs could alter the function of the rapid component of the delayed rectifier potassium current (Ikr) and then change the pattern of myocardial repolarization, which is responsible for the QT prolongation and torsades de pointes [5]. The evidence is strong enough to urge the FDA to announce a post-approval pharmaceutical withdrawal of cisapride and terfenadine. In contrast, the story of drug-induced AF is completely different. Atrial fibrillation has a high prevalence in the general population and can be noticed in > 10% of individuals over the age of 80. Atrial fibrillation is also associated with structural heart diseases for which, the patients may be taking “AF-inducing” drugs mentioned in the review by Dr Tamargo et al. Moreover, the mechanism of AF per se has not been well defined despite decades of research and debate. Bearing this in mind, we should be conservative when definitely and directly relating a certain drug to an AF episode.

2. Mechanism of drug-induced AF

Atrial fibrillation of any kind is not driven by a single mechanism. The initiation and the maintenance of this chaotic rhythm share different anatomical and pathological substrates in the atria while triggered activity plays a more important role in starting an AF episode. The “possible” mechanisms involved in drug-induced AF were discussed in the review by Dr Tamargo et al. However, these mechanisms, with few exceptions, need to be further demonstrated in the laboratory. In vitro changes of action potential duration and effective refractory period observed in myocytes or myocardial tissue may not apply to an intact heart in vivo when neurohumoral regulation and interactions with other factors are present. As proposed by the authors, genetic background serves not only as a determinant of the action of certain drugs, but also of their side-effects. While mutation is a very rare genetic variation, single nucleotide polymorphism (SNP) of a gene affects at least 1% of the general population. Polymorphic sequence variant (i.e., SNP) usually does not cause overt debilitating diseases but does contribute to disease susceptibility and influence drug responses. A genetic basis for drug response needs to be considered if the responses are heterogeneous in a cohort of patients with similar clinical characteristics. In the last few years, we have achieved notable progress toward SNP-guided personalized treatment; specifically the field of pharmacogenetics has expanded to study a broad range of cardiovascular drugs including antiplatetlet agents, warfarin, statins and diuretics [6]. One example is to titrate the therapeutic safety margin of warfarin based on the type of gene SNP associated with warfarin's pharmacokinetic (gene CYP2C9) and pharmacodynamic (genes VKORC1 and CYP4F2). Certain variations of these genes will alter the metabolism of warfarin, subsequently the dose requirement of warfarin [7,8]. Similarly, the SNP of cardiac ionic channel genes has been associated with the susceptibility to different arrhythmias including AF and the efficacy of antiarrhythmic drugs [9-11]. It is possible that patients carrying an AF-predisposing polymorphic genetic variation are more likely to develop AF compared to the “non-carriers” after taking a specific drug. Unfortunately, data on the relationship between specific SNP of cardiac ionic channel genes and the vulnerability to drug-induced AF are not available in the literature. Therefore, relevant studies are warranted, and adenosine-induced AF seems to be a good starting point.

3. Atrial fibrillation induced by antiarrhythmic drug

Interestingly, AF sometimes can be induced by antiarrhythmic drugs. Digxoin has been routinely prescribed in AF patients for more than 100 years, with a goal to control the ventricular response in AF. Occasionally, digxoin also helps in “converting” AF to normal sinus rhythm. Besides its positive inotropic effect, digxoin increases vagal activity, thereby slowing depolarization of pacemaker cells in the atrioventricular node. However, it is important to emphasize that we do have patients with vagal-mediated AF. If digxoin is prescribed in these patients, the vagal-mediated paroxysmal AF may progress to persistent AF due to continuous vagal nerve stimulation. Under this circumstance, it would be difficult to assess the pro-AF adverse effect of digxoin. In rare cases, digxoin-induced AF presents as digitalis intoxication when this medication is used for the treatment of heart failure [12-14]. As a strong sodium channel blocker, flecainide is a first-line drug in the management of AF. Specifically, it is recommended as the drug of choice when vagal-mediated AF is considered, possibly because flecainide has some vagolytic effect. However, flecainide also induces AF. This could reflect a “potent blockade of Na⁺ channels that slows the intra-atrial conduction velocity” [4]. Differently, previous and recent studies have shown that: i) flecainide reduces regional heterogeneity in effective refractory period and wavelength in the atria [15]; ii) flecainide leads to partial resynchronization of the electrical activity across the atrial wall; and iii) flecainide decreases transmural conduction of the atrial wall [16]. All the above effects improve the electrical homogeneity of the atrial myocardium and contribute to the antiarrhythmic/anti-AF potency of flecainide.

4. Drug-induced AF is frequently overlooked

I agree with Dr Tamargo et al. that the incidence of drug-induced AF is underestimated. This is because: i) “asymptomatic” AF is not reported; ii) even if the AF is detected, it could be attributed to the patient's underlying disease rather than the drug treating the disease; iii) drug-induced AF is not well recognized and understood by healthcare providers especially noncardiovascular professionals. In the literature, more than 80% of the existing studies on proarrhythmic side-effect discuss QT-related issues followed by drug-induced ventricular arrhythmia and heart block. This indicates that “drug-induced AF” is uncommon and poorly defined. Additionally, it reflects that drug-induced AF is overlooked or ignored. Since AF is not perceived as a life-threatening arrhythmia, it is likely that drug-induced AF is not considered as a “safety issue” by physicians or pharmacists. This situation is putting our patients at risk. Indeed, a thrombus could quickly form in the left atrium after drug-induced AF develops in a patient on chemotherapy and having a tumor-related hypercoagulable states. Similarly, a massive stroke secondary to an underdiagnosed/overlooked drug-induced AF by his/her antidepressants in an otherwise physically healthy individual would be a devastating complication. Even though it is probably too early to be certain of the incidence, pathophysiology, clinical features and prognosis of drug-induced AF, the data provided in Dr Tamargo et al.'s review are plenty to establish an “operative framework” to provide guidance in the daily practice. A physician should be aware of a list of “potential” AF-inducing drugs in order not to misdiagnose when a patient presents to him/her with drug-induced AF. Meanwhile, this information should be used to educate our patients when they are prescribed a drug with this potential side effect. In addition to self-monitoring AF-related symptoms, comparison of AF burden at different times, i.e., pre-administration and post-administration of a certain drug and after discontinuing of the drug if drug-induced AF is suspected, may help to diagnose drug-induced AF and distinguish it from spontaneous AF. However, criteria for the definition of new-onset drug-induced AF or for arrhythmia aggravation based on the underlying AF density needs to be further determined by clinical studies.

5. Expert opinion

Drug-induced AF is likely to be more common than what we have appreciated so far. This proarrhythmic effect should not be ignored solely because AF is not a life-threatening arrhythmia. There has been sufficient clinical data to suggest that many drugs are potential AF-inducing agents, however, under what situation and through which pathway theses drugs are associated with AF remain unclear. It is essential that healthcare providers are knowledgeable of these drugs and their possible pro-AF adverse effect. Further knowledge of these drugs will help in eliminating the underlying risks of underdiagnosed drug-induced AF. Basic research to explore the mechanisms of drug-induced AF and clinical studies incorporating established algorithms for the identification of drug-induced AF are critical to improve drug safety and future drug development.

Declaration of interest

The author states no conflict of interest and has received no payment in preparation of this manuscript. The author is supported from China by the Program for New Century Excellent Talents in University (NCET-09-0376); the National Natural Science Foundation (NSFC-30973601, NSFC-812101059) and the Scientific Research Foundation for the Returned Overseas Chinese Scholars (SFR ROCS 2008-101).