Impact of Statins on Risk of Stroke: A Meta-Analysis

Nickole N Henyan, PharmD; Daniel M Riche, PharmD BCPS; Honey E East, MD; Pamela N Gann, PharmD


The Annals of Pharmacotherapy. 2007;41(12):1937-1945. 

In This Article


Data Sources

A systematic literature search of MEDLINE, EMBASE, Cumulative Index to Nursing & Allied Health Literature, and Web of Science from June 1975 through September 2006 was performed to identify controlled clinical trials of statin therapy that evaluated stroke as any endpoint (clinical trial acronyms defined in the Appendix). The search strategy is shown in Figure 1. A manual review of abstracts presented at meetings of the American College of Cardiology, the American College of Clinical Pharmacy, and the American Stroke Association from 2001 to 2006 was also conducted. References from review articles were reviewed to identify additional relevant studies.

Figure 1.

Search strategy diagram. Included studies may overlap among search strategies.

Inclusion Criteria

To be included in this meta-analysis, studies had to meet the following criteria: (1) controlled clinical trials versus placebo, (2) well-described protocol, and (3) data reported on incidence of all CVEs, ischemic stroke, or hemorrhagic stroke. Since major studies that included cerivastatin were terminated prior to completion, studies were excluded if cerivastatin was the active treatment (n = 2). Studies were also excluded from each analysis if there were no events in either group (n = 7) or if the control group included an active therapy or standard of care (n = 16). Abstracts not reporting stroke (n = 2) were also excluded. In cases in which there was more than one published report on the same population or group of patients (n = 4), the original article was selected for analysis. Two of the included trials reported transient ischemic attacks (TIAs) in combination with all CVEs. Considering the inadequacy of TIAs as a solid cerebrovascular endpoint, data on TIAs were not included in the analysis of all CVEs when reported separately.

Assessment of Literature Quality

The following methodological features most relevant to the control of bias were assessed: randomization, concealment, masking of treatment allocation, and withdrawals. All trials were independently reviewed by 3 investigators (NNH, DMR, PNG), with disagreement resolved by consensus.

Data Extraction

The following information was collected from each article: study design, sample size, follow-up duration, statin and dose evaluated, number and type of CVE, comorbid disease states, age, sex, race, concomitant medications, cholesterol levels, and smoking history.

Data Analysis

This meta-analysis was completed through the use of StatsDirect statistical software, version 2.4.5 ( Summary statistics were combined and weighted averages were calculated using a random effects (Der-Simonian and Laird methodology) model. Statistical heterogeneity was evaluated via the Q statistic (p < 0.1 was considered representative of significant statistical heterogeneity). Publication bias was assessed through visual inspection of funnel plots and the Egger weighted regression method, with a p value less than 0.05 considered representative of statistically significant publication bias. Data are reported as relative risks with 95% confidence intervals.


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.
Post as: