Vitamin e and omega-3: what to believe: observational studies or randomized controlled trials?

The practice of conventional medicine has markedly changed since the introduction of the concept of the evidence-based medicine. Randomized controlled study design and large sample size were the only justifications for level A or B evidence at the summit of what is called the evidence pyramid. A lot of medical interventions that were based on a plethora of basic researches and multiple large real world or observational studies in humans became questioned now by the results of even a single large sized randomized controlled trial (RCT). The conflicting evidences for the value of vitamin E and Omega-3 fatty acids in cardiovascular diseases are famous examples for such perplexity. This article discusses this problem on the basis of scientific, ethical, and statistical critical appraisal. To conclude, in this era of overwhelming flow of data, it should be emphasized in short, fast-to-read articles that it is important to consider not only the level of evidence "as dictated by the study design and sample size" but also the relevance of evidence. Studies tell us about populations but we treat individuals. The type of the studied individuals, the enrollment criteria, the methodology, the dose of the studied drug and all the combined medications in the study should be clearly considered whenever the reported results are to be generalized beyond the specific situation studied. Comparing the effect of an active drug against placebo by giving either one of them to a group already treated with other multiple drugs (optimum medical therapy) could be a misleading indicator for the pure efficacy of the active drug. Many confounding variables such as known "or unknown" drug-drug interactions, sharing mechanisms of action or unexpected adverse drug reactions can afflict only the group randomized to take the active drug. These variables will not affect the control group simply because they add to their optimized multiple drug therapy an inert placebo.