Print Digg

Article

Re-Analysis of Meta-Analysis

Friday, September 10, 2010


Different results were found in a recent re-analysis of the meta-analysis published in the 2007 Journal of the American Medical Association (JAMA), which linked antioxidant supplements to all-cause mortality.

The Copenhagen University Hospital meta-analysis published in JAMA in 2007 looked at data from 68 randomized clinical trials with 232,606 participants. They reported a 16% increase in mortality risk from supplement intake of popular antioxidants such as vitamins A, C, E, and beta-carotene.

A recent University of Honheim (Stutthart, Germany) re-analysis of the Copenhagen meta-analysis published in the August 2010 Nutrients found quite different results:

  • • 36% of the trials showed beneficial outcomes to antioxidant supplementation
  • • 60% showed no differences
  • • Only 4% showed negative outcomes


The re-analysis of the original meta-analysis found benefits correlated to vitamins C and E, selenium, beta-carotene and zinc. The re-analysis investigators suggested that supplementation of these nutrients appeared to be most effective in treating and preventing chronic health conditions when the user had inadequate levels of these micronutrients from diet alone.

One wonders why the re-analysis investigators did not look at exactly the same micronutrients looked at in the meta-analysis.  

Since new data being released from Health & Human Services Healthy People 2010 is suggesting fewer than 11% of the U.S. population actually consumes even 5 servings of fruits and vegetables a day, it seems reasonable to suggest that 90% of the U.S. population may be consuming inadequate levels of micronutrients from their diet alone, therefore increasing their risk of chronic degenerative disease. .

The new Health & Human Services recommendation for fruits and vegetables is 9-13 servings a day to reach micronutrient recommended daily intake (RDA), also referred to as dietary reference intake (DRI).

Meta-analyses too often wreak havoc, particularly when the lay press turns frequently unreliable science into headlines. The 2005 meta-analysis that linked vitamin E to all-cause mortality was almost immediately disputed by reputable re-analysis. Unfortunately too many doctors still tell patients to avoid supplemental vitamin E, based on a completely flawed vitamin E meta-analysis, published in Annals of Internal Medicine.

Ellen Troyer, MT MA
Biosyntrx CEO / Chief Research Officer

PEARL

Andrew Shao, PhD, Senior VP of Scientific and Regulatory Affairs for the Council of Responsible Nutrition said it best when he commented on the antioxidant mortality risk meta-analysis and re analysis:

"The Copenhagen antioxidant meta-analysis simply examined numbers to see if the risk for total mortality increased in those who took antioxidant supplements. The German re-analysis investigators examined the same studies to see if there were statistically significant benefits demonstrated in the study group, and if the effects were biologically plausible. Not surprisingly, they reached different conclusions."

"The obvious question: Could a large part of flawed nutrient meta-analysis be eliminated by requiring that the investigators look for and report both benefits and risks, if they are to be considered valid for publication."

Crestpoint Management, LTD instrument announcement:
otaka dilator 6-182-2

References

Mortality in Randomized Trials of Antioxidant Supplements for Primary and Secondary Prevention, Bjelakovic G, Nikolova D, et al. JAMA 2007;297:842-857 [full article]

Re-analysis: Reexanination of a Meta-Analysis of the Effect of Antioxidant Supplementation on Mortality and Health in Randomized Trials. Biesalski H, Grune T, Tinz J, Zooller I, and Jeffrey Blumberg. Nutrients August 2010 Volume 2, Issue 9 [abstract]

High-dosage vitamin E supplementation may increase all-cause mortality. Miller ER, Pastor-Barriuso R, et al. Ann Intern Med. 2005 Jan 4; 142(1):37-46 [abstract]

Re-analysis: Risk of Mortality with Vitamin E Supplements: the Cache County study. Hayden KM, et al. Am J Med  2007 Feb;120(2):180-4 [abstract]