Statin Therapy for the Prevention of Atrial Fibrillation

A Meta-analysis of Randomized Controlled Trials

Zhongsu Wang; Yong Zhang; Mei Gao; Jiangrong Wang; Qing Wang; Xiaojun Wang; Lequn Su; Yinglong Hou


Pharmacotherapy. 2011;31(11):1051-1062. 

In This Article


Benefits of Statin Therapy on Atrial Fibrillation

In addition to their lipid-lowering effects, statins have been shown to have antiinflammatory and antioxidant activity, and to treat endothelial dysfunction, increase nitric oxide bioavailability, stabilize atherosclerotic plaques, and modulate the autonomic nerve system.[32] All these effects may influence the structural and electrical remodeling of the atria that contributes to the development and perpetuation of atrial fibrillation.

Compared with a previous meta-analysis,[27] we relaxed the inclusion criteria to include a greater diversity of patient groups in order to examine the benefits of statins on primary and secondary prevention of atrial fibrillation, and to assess the influence of drug type and dosage. Similar to the conclusion from the previous meta-analysis, we concluded from the overall effect size that the use of statins significantly decreases the risk of atrial fibrillation (OR 0.59, 95% CI 0.45–0.76).

According to the subgroup analysis, the benefits of statins were marked for both secondary prevention and new-onset atrial fibrillation (Table 2), and no significant differences were noted between these patient groups. Several studies have shown that inflammatory mediators or markers, such as interleukin-6 and high-sensitivity C-reactive protein levels, are increased in patients with atrial fibrillation,[33] whereas statins are more beneficial in patients with higher inflammatory marker levels.[34] From these results, greater benefit could be predicted for secondary prevention. In our meta-analysis, however, the difference between primary and secondary prevention was not statistically significant. We noticed that patients undergoing cardiac surgery were included in the primary prevention subgroup. As these patients also had significantly elevated C-reactive protein levels (1.63–23 mg/dl), the differences in inflammatory marker levels between the two subgroups became insignificant, which may be one of the reasons leading to the insignificant difference between the two patient groups.

Another subgroup analysis was performed to assess the activity of the individual drugs. These analyses revealed that atorvastatin displayed a statistically significant benefit compared with the other three statins (Table 2), but the number of rials involving pravastatin, rosuvastatin, and simvastatin was too small; thus, more RCTs are necessary to further support this conclusion.

In the atorvastatin subgroup, we found that the dose of drug (10–80 mg/day) may have contributed to the heterogeneity. Subgroup analysis evaluating the benefits of different doses of atorvastatin on atrial fibrillation indicated that marked benefit was observed in the 10–40-mg/day subgroup, whereas intensive atorvastatin therapy (80 mg/day) showed no beneficial effect (Table 2). Statins show several benefits, including lipid lowering and antiinflammatory effects, and atorvastatin's preventive effect on atrial fibrillation may be more significant at a lower dose (10–40 mg/day), whereas the antifibrillatory effect became smaller or even disappeared when the dose was increased to 80 mg/day.

Atrial fibrillation was a commonly observed adverse event in a study of atorvastatin as lipidlowering therapy.[35] Over 12 months, 2 of 234 patients in the atorvastatin 80-mg/day group and 1 of 233 patients in the atorvastatin 10-mg/day group developed atrial fibrillation (p=0.57). In the Aggrastat to Zocor (A to Z) trial that included 4497 patients, different simvastatin dose regimens were compared, and the onset of atrial fibrillation was identified by adverse-event reporting.36 During the 2-year follow-up, the rate of atrial fibrillation was higher in the high-dose simvastatin therapy group (p=0.10). Although there was no statistically significant evidence (p>0.05, and atrial fibrillation was only reported as an adverse event), the study also suggests that the rate of atrial fibrillation may increase as the dose of drug increases, similar to the results found in our meta-analysis.

However, in the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction (PROVE IT–TIMI) 22 study, 4162 patients were enrolled and randomly assigned to intensive therapy (atorvastatin 80 mg/day) or moderate therapy (pravastatin 40 mg/day).[36] As 2.9% of patients receiving the high dose and 3.3% of patients receiving the moderate dose developed atrial fibrillation over the next 2 years (no significant difference between groups), this study showed no increased atrial fibrillation risk with higher-dose statins. Furthermore, another group conducted an RCT that included 104 consecutive patients undergoing cardiac surgery and found that, although the rate of postoperative atrial fibrillation was lower in the atorvastatin 80-mg/day group compared with the 10-mg/day group, the high-dose regimen for 7 days before cardiac surgery conferred a nonsignificant reduction in postoperative atrial fibrillation (p=0.43).[37] Overall, the underlying mechanisms of these dose-dependent antifibrillatory effects of statins remain unclear, and further investigations are urgently needed.

Sensitivity Analysis

Sensitivity analysis was performed in the overall study as well as the subgroup analyses. After excluding the seven trials[11,16,19–23] in which atrial fibrillation was not a prespecified end point, no obvious change was observed (Table 2). We inferred from the sensitivity analyses that the trials in which atrial fibrillation was not a prespecified end point were not outliers with a dominating effect in this meta-analysis. The results provided a rationale for including these studies in the subgroup analyses.

Publication Bias

The funnel plots were visually asymmetrical, and the Egger's regression also indicated obvious publication bias in the atorvastatin subgroup, but we did not find substantial publication bias in the all-inclusive meta-analysis. Some procedures (e.g., funnel plot, regression, failsafe N) are based entirely on a model of publication bias and would have no meaning for a "small-study effect" (as previously described[38,39]) or are only sensitive to the relationship between sample size and effect size. However, if the populations or protocols are different among the trials, legitimate heterogeneity will be reflected in the effect size, as observed in our study. According to the result of the subgroup analysis of the atorvastatin group, the effect size was related to dose, and this was also be observed in the funnel plot (Figure 5). In the funnel plot, the OR values increased as the dose of drug increased, implying that the asymmetry of the funnel plot is caused by the difference in dose, but does not reflect publication bias. Moreover, to avoid publication bias, the literature search was performed through MEDLINE, EMBASE, and the Cochrane Controlled Trials Register, and we made efforts to locate unpublished studies.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.