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


Trial Selection

Trials were selected from all published RCTs involving study groups administered one or more kinds of statins, with the control intervention being either placebo or another conventional treatment (active control), in patients receiving treatment for either primary or secondary prevention of atrial fibrillation. Studies with protocols containing coadministration of other investigational agents were also included. Studies without original data related to the prevention of atrial fibrillation were excluded.

Search Strategy

A PubChem Compound search for all names of statins was conducted. The terms arrhythmia, atrial fibrillation, coronary heart disease, coronary artery disease, myocardial infarction, ischemic heart disease, hypertension, heart failure, and cardiac surgery were identified as health conditions. A literature search was performed of MEDLINE (through September 30, 2010), EMBASE (through 2010), and the Cochrane Controlled Trials Register (through 2010) by using the name of each drug as a key word combined with the health conditions to identify RCTs published in the English or Chinese language. In addition, abstracts from the scientific sessions of the American College of Cardiology, the American Heart Association, the European Society of Cardiology, and the North American Society of Pacing and Electrophysiology were manually searched (2001–2010), and the articles from the bibliographies of retrieved trials were scanned.

Trials were excluded if they were not RCTs, had no mention of statins as the primary intervention, the statin was used as a background intervention, and there was no mention of occurrence of atrial fibrillation. Two of the authors (Wang Z and Zhang Y) independently examined the titles and abstracts of all trials to eliminate irrelevant studies. Subsequently, the same authors examined the full texts of the remaining articles and the full text reports of all potentially relevant trials and assessed them independently for eligibility on the basis of the defined inclusion criteria. Data from included studies were extracted in duplicate. Authors of the published trial results, including abstracts, were contacted for required information when needed. Any discrepancies were resolved by discussion. To resolve disagreements, a final decision for trial eligibility and data extraction was made by the senior author (Hou Y).

Quality Assessment

The quality of the included studies and risk of bias were assessed in accordance with recent guidance,[29] in the following six domains: quality of random sequence generation and allocation concealment, blinding of participants and personnel, outcomes, description of incomplete outcome data, selective outcome reporting, and other sources of bias.

Statistical Analysis

Statistical analysis was performed based on the intent-to-treat principle, with participants not completing the study considered to be free of the event. Results of the atrial fibrillation outcome were expressed as odds ratios (ORs) with 95% confidence intervals (CIs) for each study. As the effect sizes have been sampled from a distribution of effect sizes, a pooled effect was calculated with a random-effects model.[30]

Heterogeneity was assessed by using the Q statistic (Qdf ), the p value (the result of a statistical test based on the Q statistic, in which p<0.10 represents statistical heterogeneity), and the I2 statistic (where I2>50% indicates a high degree of statistical heterogeneity).

Three subgroup analyses were performed: one to assess the effects of statins on primary and secondary prevention, the second to evaluate the effects of different statins, and a third to analyze the efficacy of different doses of atorvastatin on the prevention of atrial fibrillation.

A sensitivity analysis was performed by selectively excluding studies based on the quality assessment to check the consistency of the overall effect estimate. Funnel plots were created and Egger's regression method[31] was applied to determine the possible influence of publication bias.

All statistical analyses were performed by using Review Manager, version 5.0, software (RevMan; The Nordic Cochrane Centre, Copenhagen, Denmark) and SAS software (SAS Institute Inc., Cary, NC). Statistical significance was defined at a p value less than 0.05.


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.