February 13th, 2012
Meta-Analysis Raises More Questions About Routine Use of Aspirin for Primary Prevention
Larry Husten, PHD
Although aspirin can reduce the risk for cardiovascular (CV) events, the associated increase in bleeding suggests that it should not be used routinely in people without prior CV disease, say the authors of a meta-analysis published in the Archives of Internal Medicine.
Sreenivasa Rao Kondapally Seshasai and colleagues combined data from nine clinical studies including more than 100,000 participants who were followed for a mean of 6 years. They found a significant reduction in CV events, but not CV mortality, and an increased risk for important bleeding events:
- CV Events: OR 0.90, CI 0.85-0.96, number needed to treat (NNT): 120
- Nonfatal MI: OR 0.80, CI 0.67-0.96, NNT: 162
- CV Death: OR 0.99, CI 0.85-1.15
- Nontrivial bleeding events: OR 1.31, CI, 1.14-1.50, number needed to harm, 73
The authors conclude that in the absence of further studies identifying patients likely to benefit from aspirin therapy, “a reappraisal of current guidelines appears to be warranted, particularly in countries where a large number of otherwise healthy adults are prescribed aspirin, since a significant proportion of them may develop bleeding complications.”
In an accompanying comment, Samia Mora agrees with the authors that routine use of aspirin for primary prevention should not be recommended. But, she writes, “it is reasonable to consider using aspirin for primary prevention in higher-risk individuals without known CVD (above 1% CVD event rate per year) if they are deemed to have a greater benefit to risk ratio and after taking into account patient preferences.”
If you read articles in the magazine “Chance”, a statistical journal published by the department of mathematics at Dartmouth, or check either of my two blogs that analyze the statistical errors inherent in meta-analysis (at http://www.ghthomas.blogspot.com), it becomes obvious that the results of a meta-analysis should be viewed as a hypothesis to then be tested is a proper double-blind fashion. Peto’s articles in Lancet were mathematically flawed. I can offhand recall two meta-analyses that gave diametrically opposing results to the question of anti-coagulation in a certain clinical situation.