OUP user menu

Usefulness of statins in preventing atrial fibrillation in patients with permanent pacemaker: a systematic review

Pasquale Santangeli, Giuseppe Ferrante, Gemma Pelargonio, Antonio Dello Russo, Michela Casella, Stefano Bartoletti, Luigi Di Biase, Filippo Crea, Andrea Natale
DOI: http://dx.doi.org/10.1093/europace/euq044 649-654 First published online: 26 February 2010


Patients with permanent pacemakers (PM) are at high risk of developing atrial fibrillation (AF). Minimal ventricular pacing modalities have been demonstrated to reduce AF in such patients, although they are not suitable for patients with advanced atrioventricular conduction disease. Recent evidences suggest that 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (i.e. statins) may represent a new strategy to prevent AF in patients at risk. In this article, we sought to review data regarding the effectiveness of statin therapy in preventing AF patients with a PM. We reviewed all available studies that assessed the effect of statin therapy on the occurrence of AF in patients with PM, implanted due to sinus node dysfunction or atrioventricular conduction disease. Moreover, a random effect inverse variance-weighted meta-analysis was performed, by entering directly the logarithm of the hazard ratio (HR) of AF provided in the multiple Cox regression analyses from each study. Three studies were identified, including 552 patients, of whom 159 received statins. Follow-up ranged from 1 to 2.77 years. Two studies (one observational and one prospective randomized) included predominantly patients with sinus node dysfunction (70% and 91% of patient population, respectively) and, consistently, showed a beneficial effect of statins on the occurrence of AF. On the other hand, the study including predominantly patients with atrioventricular block (60% of patient population) failed to show a beneficial effect of statins on AF occurrence. The HR for AF occurrence for the cumulative data was found to be 0.43 (95% confidence interval: 0.28–0.67, P < 0.001). Statistical heterogeneity between included studies was not detected (χ2 = 1.68, P = 0.43, I2 = 0%), although significant clinical differences were found in terms of study design, patient populations, statins use and dosage and AF-monitoring capabilities. Statins may represent a novel treatment strategy to prevent the occurrence of AF in patients with PM, especially for those who had a PM implanted due to sinus node dysfunction. Basing on our findings, a randomized clinical trial with a proper design to evaluate the utility of statins in preventing AF in these patients is warranted.

  • Atrial fibrillation
  • Statins
  • Permanent pacemaker
  • Sinus node dysfunction
  • Atrioventricular
  • Block


Atrial fibrillation (AF) is the most frequent tachyarrhythmia in patients with a permanent pacemaker (PM), occurring in up to 50% of patients implanted due to atrioventricular block1 and up to 65% of those implanted due to sinus node dysfunction,2,3 and represents the main cause of morbidity in these patients after correction of symptomatic bradycardia with PMs.4 Most of the AF episodes occur asymptomatically and can be detected only by means of modern PMs with advanced rhythm-monitoring capabilities.1,3,5 The accuracy of device measurement of frequency and duration of AF episodes has been validated by a number of studies comparing the accuracy of device atrial arrhythmia detection to Holter monitoring.6,7 Moreover, these brief and asymptomatic device-recorded AF episodes carry a significant prognostic relevance. In the mode selection trial, the presence of any atrial high-rate episodes was an independent predictor of total mortality [hazard ratio (HR): 2.48; 95% confidence interval (CI): 1.25–4.91; P = 0.0092] and stroke (HR: 2.79; 95% CI: 1.51–5.15; P = 0.0011),3 and these observations have been extended to the increased risk of any arterial embolism in patients with device-detected AF episodes longer than 1 day.8 A large ongoing clinical trial is further assessing the value of asymptomatic atrial high-rate episodes for the prediction of stroke and other vascular events.9

The only therapeutic strategy to date that has been demonstrated to be effective in reducing the occurrence of AF in these patients is minimizing ventricular pacing by dedicated algorithms in order to prevent the ventricular desynchronization caused by pacing the right ventricular apex,10,11 a condition consistently associated with an increased risk of AF.4,12,13 However, minimal ventricular pacing algorithms are not yet widely diffused and are not suitable for patients with advanced atrioventricular block.10

Recently, several clinical and experimental studies have reported a protective effect of the 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (i.e. statins) against the occurrence of AF.14

In particular, statin use has been associated with a reduced prevalence of AF in retrospective studies,15 with a reduced incidence of perioperative AF following thoracic or coronary bypass surgery,16 and also with a lower recurrence of AF following electrical cardioversion.17 To date, studies evaluating the effect of statins in preventing AF in patients with PMs are few and with controversial results.1820 Due to the abovementioned significant impact on prognosis, together with the limited strategies available to prevent AF in such patients, the clinical relevance of evaluating the possible antifibrillatory effects of statins specifically in PM patients is even more important than in other AF patient populations.

Thus, this review assesses the potential value of statin therapy in preventing AF in patients with a PM. To this aim, we performed a literature search using the Pubmed, EMBASE, and CINHAIL databases, as well as the scientific session abstracts in Circulation, Journal of the American College of Cardiology, European Heart Journal, and Heart Rhythm from January 2000 to November 2009, including all available studies that assessed the role of statin therapy in the prevention of AF in patients with a PM. A detailed description of the search strategy is available in the appendix.

Published clinical studies

Of 64 citations retrieved, 26 studies assessing the effect of statin therapy on the incidence or recurrence of AF were identified. Nine studies were excluded because the study population did not include patients with PM and 14 studies because patients with PM were mixed to patients without. Finally, three studies were selected and included in this review: one randomized and two observational (Figure 1).1820

Figure 1

Selection process of studies included in the systematic review. AF, atrial fibrillation; PM, pacemaker.

Amit et al.18 retrospectively assessed the association between statin therapy and AF incidence or recurrence in a cohort of 264 patients with a PM, predominantly (60% of study patients) implanted due to atrioventricular conduction disease (Table 1). Atrial fibrillation developed in 70 (26%) patients after a median follow-up of 2 years. At multivariable analysis, which took into account all significant baseline differences between the statin and no-statin groups (Table 1), the use of statins showed a trend towards reduced risk of AF occurrence, although this reduction was not statistically significant (HR: 0.59; 95% CI: 0.31–1.12).

View this table:
Table 1

Characteristics of the studies included in the meta-analysis

Amit et al.18Tsai et al.20Gillis et al.19
Number of patients264106185
DesignRetrospectiveProspective randomizedProspective cohort
PopulationSND (40%) or AVB (60%)SND (70%) or AVB (30%)SND (91%) or AVB (9%)
Follow-up2 years (1–2.73 years)1 year2.77 ± 0.77 years
Mode of AF diagnosisECG, Holter ECG, PM interrogationPM interrogationPM interrogation
EndpointIncidence of AFIncidence of AFRecurrence of AF
Degree of adjustment of hazard ratioAge, gender, HTN, CAD, past AF, DM, LVEF, antiarrhythmic use, PM type, PM indicationAge, gender, PM indication, PM type, LA volume, LVEFAge, gender, HTN, CAD, DM, LVEF, LA diameter, antiarrhythmic use, beta-blocker use, percent of VP
  • AF, atrial fibrillation; SND, sinus node dysfunction; AVB, atrioventricular block, Std., standard; HTN, hypertension; CAD, coronary artery disease; DM, diabetes mellitus; LVEF, left ventricular ejection fraction; LA, left atrium; PM, pacemaker; VP, ventricular pacing.

At variance with the study of Amit et al.,18 another observational study by Gillis et al.19 recently assessed the effect of statin therapy on AF recurrence in a cohort of 185 patients with previous history of paroxysmal AF and a PM implanted, mainly (70%) due to sinus node dysfunction (Table 1). At 1-year follow-up, 69% of patients experienced AF recurrence, defined as any episode lasting >5 min on PM interrogation. Statin therapy was significantly associated with a lower rate of AF recurrence at multiple logistic regression analysis [odds ratio (OR): 0.33; 95% CI: 0.14–0.74; P = 0.007], after adjustment for all the confounding variables reported in Table 1. Moreover, patients receiving statins had a significantly lower AF burden during follow-up (0.10 h per day vs. 0.39 h per day, P = 0.0059).

In keeping with these positive results are the findings of Tsai et al.20 in a population of patients with no previous history of AF. The authors tested in an open-label prospective randomized trial the efficacy of statin therapy in preventing the occurrence of atrial high-rate episodes detected at PM interrogation. At variance with the studies by Amit et al.18 and Gillis et al.,19 in which patients treated with statins received different statin types and doses, in this study patients randomized to statins received a fixed dose (i.e. 20 mg) of atorvastatin.

A total of 106 patients undergoing PM implantation (91% due to sinus node dysfunction) were randomized. At 1-year follow-up, treatment with atorvastatin was associated with a reduced occurrence of long atrial high-rate episodes (HR: 0.33; 95% CI: 0.14–0.79; P = 0.015), defined as episodes lasting ≥10 min. Furthermore, patients randomized to atorvastatin showed a favourable left atrial remodelling, as assessed by echocardiography, with a significant reduction of left atrial volumes (from 42.4 ± 12.1 mL to 39.7 ± 10.9 mL; P < 0.05) at the end of follow-up. On the other hand, statin therapy failed to prevent short atrial high-rate episodes (i.e. episodes lasting ≥1 min but <10 min).

Taken together, these clinical studies suggest a possible beneficial effect of statin therapy in preventing either incidence or recurrence of AF in patients with a PM, especially for those who had sinus node dysfunction as the indication to PM implantation.19,20

The retrospective series reported by Amit et al. was the only one that did not show a clear reduction in AF with statin use.18 Although this study included the largest cohort of patients and had the longest follow-up, several factors may have contributed to this negative result. Indeed, patients treated with statins were at higher risk of developing AF, with a significantly higher incidence of hypertension (88% vs. 71%, P = 0.01), coronary artery disease (69% vs. 29%, P < 0.01), and diabetes mellitus (42% vs. 26%, P = 0.02), and an imprecise number of patients in the control group was actually treated intermittently with statins. Moreover, patients in the statin group had also a greater prevalence of previous history of AF (24% vs. 14%, P = 0.08) as compared with those not receiving statins, therefore consistently showing a higher risk of developing AF during follow-up.

As the authors acknowledged, all these confounding factors may have not been fully adjusted by the use of a multiple regression model. Furthermore, at variance with the studies by Gillis et al.19 and Tsai et al.,20 no specific PM programming was necessary in this study,18 possibly affecting the reliability of AF detection at PM interrogation. Indeed, it is essential that the devices are programmed appropriately to avoid far-field R-wave oversensing and atrial undersensing to ensure a high sensitivity and specificity for AF detection.2,3,5,7,9

Moreover, the fact that statin-treated patients of the studies by Amit et al.18 and Gillis et al.19 were actually treated with different types and doses of statins, and the lack of adjustment for statin types and doses may affect both the entity and the direction of the statin treatment effect. Indeed, there is growing evidence that pleiotropic effects of statins, including their effects on AF, are at least dose-related.21

Finally, significant differences in clinical indication to PM implantation between the studies of Amit et al.18 and those by Gillis et al. and Tsai et al.19,20 may account for the reported different results. Indeed, the majority of patients included in the study by Amit et al. had a PM implanted due to AV block, whereas the studies by Gillis and Tsai had sinus node dysfunction as the main reason for PM implantation (91% and 70% of patients, respectively). Basing on these data, it is tempting to hypothesize that the beneficial effect of statins in preventing AF in patients with a PM is limited to patients with sinus node dysfunction (see also Pathophysiological mechanisms).

Pathophysiological mechanisms

From a pathophysiological standpoint, basic and clinical evidence suggest that statins may exert antiarrhythmic effect by improving endothelial nitric oxide availability22 and reducing inflammation23 and oxidative stress.24

These effects, also called pleiotropic properties of statins, largely exert through inhibition of isoprenoid intermediates of the cholesterol synthesis pathway. Isoprenoid intermediates,25 such as farnesylpyrophosphate and geranylgeranylpyrophosphate, are important for the post-translational modification of GTP-binding proteins, such as Ras, RhoA, or Rac1, which are implicated in the regulation of inflammation and oxidative stress.26 Moreover, statins may directly up-regulate the endothelial nitric oxide sinthase.27 Accordingly, clinical studies evaluating the prevention of AF by statins have mostly focused on clinical settings in which increased inflammation, oxidative stress, and endothelial dysfunction have been clearly linked to increased risk of AF, such as after cardiac or thoracic surgery.1517,2830 The pathophysiological rationale for evaluating the effect of statins on AF in patients with a PM is less straightforward. Indeed, the development of AF in such patients (especially in those with sinus node dysfunction) appears to be linked to the occurrence of structural changes in the atria such as fibrosis and dilatation31,32 and is accelerated by increasing the amount of right ventricular pacing.10,33,34

Interestingly, basic science data have linked endothelial dysfunction, inflammation, and oxidative stress with both structural changes in the atria leading to atrial fibrosis and dilatation, as well as with biochemical changes in the individual atrial myocytes, such as hypertrophy or changes in ion channel density or distribution.14 Furthermore, right ventricular pacing has been recently demonstrated to cause systemic endothelial dysfunction, thus suggesting a novel mechanistic explanation for its striking association with the occurrence of AF.35

Thus, the beneficial effect of statins in preventing AF in patients with a PM may exert both slowing the progression of atrial structural and electrophysiological remodelling through their anti-inflammatory and anti-oxidative properties, and also improving the endothelial dysfunction caused by ventricular pacing. However, the exact mechanism(s) by which statins are antiarrhythmic in patients with a PM requires further investigation.


To provide an overall estimate of the effect of statin therapy in preventing AF in patients with a PM, we performed a meta-analysis including the publications selected above. From each article we extracted the following data: the total number of patients in the treatment group and control group, the number of events in both groups, the length of follow-up, and a relative risk measure of AF with its 95% CI from each study.

Statistical analysis

Data from each study were entered as logarithm HR with its standard error and combined with a fixed effect inverse variance-weighted method to obtain the summary estimate of the endpoint, expressed as HR with 95% CI. DerSimonian and Laird36 random effect method was also used as sensitivity analysis, even in the absence of statistically detectable heterogeneity. To provide an estimate of the treatment effect that may be minimally affected by baseline differences in the treatment and control group characteristics, the value of HR, reported in each study, adjusted for the largest number of confounding factors and/or provided in the model of multiple Cox regression analysis adjusted for the maximum number of covariates (Table 1), was used. The presence of heterogeneity among studies was evaluated with Cochrane Q χ2 test, and inconsistency was assessed with I2 test that describes the percentage of the variability in effect estimates that is due to heterogeneity: values of 25%, 50%, and 75% correspond to low, moderate, and high I2.37 Publication bias was assessed using the asymmetry linear regression of Egger's test38 and displayed as funnel plot of precision (standard error of log HR) against the treatment effect (displayed as HR). Statistical level of significance for the summary treatment effect estimate was two-tailed P<0.05. Heterogeneity and publication bias were considered statistically significant at a two-tailed P < 0.1. Analyses were performed using the software package STATA 9.0 (Stata Corporation, College Station, TX, USA).


Overall, 552 patients with PM were included, of whom 159 (28%) received statins. Table 2 summarizes the baseline clinical characteristics of study patients.

View this table:
Table 2

Baseline characteristics of study patients

Amit et al.18 (n = 264)Tsai et al.20 (n = 106)Gillis et al.19 (n = 185)
Statins (n = 51)No statins (n = 213)Statins (n = 52)No statins (n = 54)Statins (n = 57)No statins (n = 128)
Age, years, (mean ± SD; median/range)72 ± 871 ± 1370 ± 1372 ± 1370 (65–76)71 (65–78)
Male sex, n (%)28 (55)106 (50)23 (43)25 (47)35 (61)61 (48)
Hypertension, n (%)44 (88)150 (71)*48 (93)53 (98)35 (61)65 (51)
Diabetes, n (%)21 (42)54 (26)*6 (12)4 (7)
CAD, n (%)35 (69)61 (29)*29 (51)38 (30)*
LVEF (%)60.4 ± 12.362.5 ± 16.363 (56–69)70 (52–70)
PM type (%)DDD (59), VDD (35), AAI (6)DDD (51), VDD (36), AAI (13)DDD (76), AAI (24)DDD (72), AAI (28)DDD (100)DDD (100)
Percent of VP, (mean ± SD; median/range)70.9 ± 40.869.8 ± 41.881 (60–98)83 (63–95)
Beta-blockers, n (%)33 (66)52 (41)*
ACE-inhibitors, n (%)36 (63)68 (53)
Antiarrhythmics, n (%)11 (22)26 (12)31 (54)88 (69)
AF occurrence, n (%)10 (20)47 (22)3 (6)10 (19)*36 (63)110 (86)*
  • SD, standard deviation; CAD, coronary artery disease; LVEF, left ventricular ejection fraction; PM, pacemaker; VP, ventricular pacing; FU, follow-up.

  • *P < 0.05 for comparisons between statins and no-statins groups.

Adjusted HR was available in two studies from Cox proportional regression18,20 and adjusted OR from multiple logistic regression analysis in one study.19 Factors and covariates of adjusted HR from each study are reported in Table 1. The length of follow-up varied from 118,20 to 2.77 years.19

Therapy with statins was associated with a significant reduction of the risk of developing AF at follow-up (HR: 0.43; 95% CI: 0.28–0.67; P < 0.001) (Figure 2). No statistical heterogeneity was detected (χ2 1.68, P = 0.43, I2=0), and a random effect model used for estimating the effect of treatment yielded the same result (HR: 0.43; 95% CI: 0.28–0.67; P < 0.001). On visual estimation, publication bias was not detected (Figure 3), although Egger's test found evidence of small publication bias [bias −5.45, 95% CI (−13.53, 2.61), P = 0.07].

Figure 2

Forest plot showing the adjusted hazard ratio of atrial fibrillation associated with statin use in each study and the overall adjusted HR. Square boxes denote HR; horizontal lines represent 95% confidence interval.

Figure 3

Funnel plot of the standard error of logarithm hazard ratio against log HR, displayed as HR on logarithmic scale. There is no presence of publication bias on visual estimation.

Implications for further studies

The results of our meta-analysis have to be interpreted as only hypothesis-generating, rather than conclusion-drawing. Indeed, although the pooled treatment effect is quite impressive, with a reduction of more than 50% of the risk of developing AF, the small number of studies (of which only one was randomized), with different clinical indications to PM implantation, inconsistent outcomes of interest (i.e. new-onset AF and AF recurrence), and different lengths of follow-up, make our results not definite, but worthy of further investigation.

Interestingly, a positive treatment effect came from the two studies that had sinus node dysfunction as the main indication to PM implantation,19,20 and the more pronounced benefit was reported in the study by Gillis et al.,19 which included patients with history of paroxysmal AF and reported a significantly greater number of AF events at follow-up. Thus, our results suggest that an appropriately designed randomized double-blind clinical trial testing the effectiveness of statin therapy in preventing AF recurrence in a homogeneous population of patients with sinus node dysfunction undergoing PM implantation would be warranted to clarify the possible benefit of statins in these patients.


Statin therapy may represent a novel treatment strategy to prevent AF in patients with PMs, especially for those who had a PM implanted due to sinus node dysfunction.

The results of our meta-analysis suggest a striking reduction of AF occurrence associated with statin use in these patients and generate an appealing hypothesis to be tested in a large double-blinded randomized placebo trial.

Conflict of interest: A.N. has received compensation for belonging to the speakers' bureau for St. Jude Medical, Boston Scientific, Medtronic, and Biosense Webster and has received a research grant from St. Jude Medical. The other authors declare no conflicts of interest.


Search strategy

The investigators independently searched Pubmed, EMBASE, and CINHAIL using highly sensitive and specific strategies. Search keywords included: ‘atrial fibrillation’, ‘prevention’, ‘treatment’, ‘statins’, ‘HMG-CoA reductase’, ‘pacemaker’, ‘sinus node disease’, ‘sinus node dysfunction’, ‘sick sinus syndrome’, ‘atrioventricular block’, ‘atrioventricular conduction disease’. Searches were updated to November 2009. No language restriction was used. Proceedings from the annual American Heart Association, American College of Cardiology, European Society of Cardiology, Heart Rhythm, and Europace meetings for the past 9 years were also manually searched.


View Abstract