Effect of statins in 'upstream therapy' of atrial fibrillation: better reliability with implantable cardiac monitors

The therapeutic potential to suppress atrial fibrillation (AF) is currently under evaluation for several non-antiarrhythmic drugs: angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers, aldosterone receptor antagonists, anti-inflammatory and antioxidative drugs, polyunsaturated fatty acids, and HMG-CoA reductase inhibitors, i.e. statins. Statins have shown multiple (‘pleiotropic’) beneficial effects beyond lowering of cholesterol: anti-inflammatory, antioxidant, and antifibrotic effects, modulation of matrix metalloproteinases, interaction with peroxisome receptors that regulate proliferation, and with endothelial nitric oxide synthase, protecting atrial myocardium during ischaemia.1 For these reasons, statins may be able to decelerate or even reverse structural remodelling in patients with AF. Additionally, pre-treatment with high-dose atorvastatin prevented electrical remodelling in a pericarditis dog model,2 and high-dose simvastatin attenuated downregulation of calcium channel subunits, preventing shortening of atrial refractoriness, a central element of electrical remodelling in AF.3 However, this sounds familiar: positive effects on intracellular calcium handling and atrial electrical remodelling have also been reported in animal studies with verapamil,4,5 but this drug disappointed in clinical trials, being unable to prevent atrial electrical remodelling and recurrence of AF.6,7 For statins and AF suppression, there are not only in vitro and animal studies available but also results from 16 clinical trials.1 Nine of these contained retrospective, registry or post hoc data; six were prospective. Results ranged from no effect to 77% relative risk reduction, i.e. basically from complete ineffectiveness to unique efficacy of statins in AF suppression according to these data. In part, differences can be explained by patient selection (acute myocardial infarction, heart failure, after cardiac surgery, post-cardioversion), and perhaps also by the use of different statins and dosages. However, one problem in assessing the efficacy of any treatment aimed at AF has not been sufficiently appreciated: the non-random distribution of AF recurrences and its implications for diagnostic monitoring.8 With the use of pacemaker memory data which provide continuous recordings of the atrial rhythm (the ‘100% picture’), the diagnostic yield of random, intermittent ECG and Holter recordings has been simulated, revealing a sensitivity for AF detection as low as 30%.9 This means that 70% of AF recurrences may go unnoticed and undetected by conventional diagnostic monitoring which covers <1% of the time. Symptoms are also unreliable to assess successful rhythm control.10–12 Therefore, the uncertainty of success rates of different approaches to treat and prevent AF mandates better diagnostic tools. This refers not only to antiarrhythmic and non-antiarrhythmic drugs but also to catheter ablation of AF. Future studies will have to provide implantable continuous monitors to evaluate the correct effect on AF recurrence.13 In this context, the study by Gillis et al. is highly welcome.14 With the use of a sophisticated dual-chamber pacemaker providing abundant memory for atrial electrograms and a dedicated algorithm for AF detection with a 96% specificity, the bias of non-random distribution of AF recurrences was eliminated. As a result, the use of statins was associated with a 22% absolute risk reduction of AF recurrence after 1 year. Short-lived AF episodes (≥5 min) as well as long AF episodes (≥7 days) were significantly less frequent in patients with a variety of statins at different dosages. Implantable devices offer the interesting parameter ‘AF burden’ which quantifies the average amount of AF over a pre-specified period of time as a percentage or in hours per day. This parameter also indicated in the present study that statin therapy may be associated with less AF. Unfortunately, AF burden was not used as primary end-point despite a publication by the same authors demonstrating its superiority over the end-point ‘time to first AF recurrence’ which is frequently not representative of treatment success.15 Of note, the effect of statins on AF recurrence was independent of and superior to that of ACE inhibitors. Despite bradycardia requiring pacing, the proportion of patients progressing to permanent AF was very low, with 7% in 3 years, potentially as a result of treatment with statins and ACE inhibitors. The present study is observational and non-randomized. Therefore, it leaves many questions open: patients receiving statins had a conventional indication for this treatment (hypercholesteraemia, coronary artery disease). Despite analysis with a multivariable logistic regression model, it is not clear if the effect on AF will also be present in the absence of these conditions, or if the effect of statins is in some way linked to them (e.g. beneficial action on disease progression with secondary effect on AF). The potential bias of non-randomized treatment is illustrated by the subgroup analysis of patients with and without class I or III drugs: patients with these drugs had an AF burden >3-fold as high as patients without class I or III drugs (statistically, however, non-significant), most probably due to the fact that these drugs were used in patients with more severe AF. Despite the justified enthusiasm about the pleiotropic effects of statins that may include atrial antiarrhythmic properties, the extent of the effect should not be ignored: in 63% of patients with ≥3 episodes of AF in the year preceding pacemaker implantation, statins were unable to prevent AF recurrence in the next year. This indicates that—despite the demonstrated preventive effects—statin therapy has only moderate ‘antiarrhythmic’ effects once chronic AF (paroxysmal or persistent) is established. In conclusion, the study by Gillis et al. demonstrates that statins can suppress AF recurrence to a certain extent. The results are particularly valuable because a continuously active AF monitor in the form of a pacemaker with a highly reliable AF detection algorithm was used. The results should encourage a randomized trial in patients with atrial-based devices as a continuous monitor. It also raises the question of whether the use of statins at an earlier stage (e.g. before documentation of the first AF episode) may be valuable in preventing AF, which would best be addressed, again, using implantable devices as monitors. Conflict of interest: none declared.