The role of antioxidants in health and aging has been in the spotlight this year. The February 2013 issue of Scientific American published a provocatively titled article “The Myth of Antioxidants.”1 The month prior, Nobel laureate James Watson, the co-discoverer of the double helix structure of DNA, suggested that antioxidant supplementation “more likely causes than prevents cancer.”2 In this editorial, I speak to this controversial topic and examine the evidence regarding antioxidant vitamin supplementation for the benefit of our health-conscious readership.
Analyses of the Efficacy of Antioxidant Therapies in Chronic Diseases
The strongest evidence against antioxidant supplementation in preventing cardiovascular disease is a meta-analysis reported in Lancet in 2003, involving 81,788 subjects on vitamin E and 138,113 on beta carotene. Supplementation did not prevent cardiovascular disease, and beta carotene supplementation was associated with a small but significant increase in mortality.3
Another meta-analysis in 2006 reported no benefit from vitamin C or vitamin E supplementation in preventing cancer.4 Some experts disagree with these conclusions, questioning the dosage, type of antioxidant used, timing of therapy, and the appropriateness of study subjects, suggesting that antioxidants work only in a subset of study participants.5
I reject these post hoc arguments because, among other reasons, it is not difficult to prove an effective intervention works when there is a clear mechanism of action and therapeutic pathway. The Nobel laureate and pathologist Sir Howard Florey needed only 8 white mice, four treated and four controls, to prove that penicillin works.6 Further, adjusting a theory, such as the role played by oxidative stress in cardiovascular disease or cancer, after a hypothesis has proved wrong is both inelegant and should damage the prestige of the theory. Even worse is making the theory so adjustable that no experimental result is excluded.7 The excuses for the failure of antioxidant trials offered above are examples of post hoc modifications which exclude no experimental result.
Clinical experts may also continue to believe in the benefit of antioxidant supplementation because of faith in laboratory data showing that reactive oxygen species and free radicals are harmful. Of course, in vitro data must be confirmed in vivo. Unfortunately, this basic tenet of the scientific method is ignored too often in the 21st century.
In vitro, oxidized LDL causes inflammation and cell death--these mechanisms are extrapolated to suggest a role for oxidized LDL in atherogenic activity in vivo.8 It showed in humans that oxidized LDL is present in veins, which do not develop atherosclerosis. In arteries, oxidized LDL accumulation did not correlate with the severity or location of atherosclerosis. This is because oxidized LDL normally circulates in blood and lymph, and binds nonspecifically to a number of widely-distributed molecules in the human body, including collagen and glycosaminoglycans. In the case of oxidized LDL, in vitro data are not applicable in vivo.8
Historical Perspective on Oxidative Stress
It is instructive to examine how today’s widespread interest in oxidative stress began. In 1981, molecular physiologist D. Neil Granger et al. showed that capillary damage following reperfusion of acutely ischemic ileum was attenuated by administration of superoxide dismutase, a free radical scavenger, administered intra-arterially to the affected intestine 15 minutes prior to reperfusion.9 These data show successful intervention in an acute injury. This is a markedly different scenario than what presumably causes chronic diseases such as atherosclerosis and cancer.
It is well-accepted that the progression from normal through dysplasia to cancer involves the accumulation of multiple mutations over years. The mainstream “modified response to injury” hypothesis of atherogenesis also postulates a prolonged preclinical phase beginning with the development of fatty streaks in youth.
What is the evidence for chronic oxidative stress? Defining an abnormality requires determining what is normal. For example, hypertension is defined as increased blood pressure over normal, defined in many populations over time. Similarly, diabetes is defined by multiple elevations of blood glucose over time. The community of researchers studying aortic stiffness have taken the step to determine normal values for pulse wave velocity,10 an increasingly utilized marker of aortic stiffness. I could not find a normal value for the redox potential of human blood after a Google search.
Searching Educus, my preferred biomedical search engine, using the phrase “normal human redox potential” yielded seven hits, only one of which appeared promising: “Experimental contribution to redox potential. VII. Normal human blood redox potential” in the journal Acta Argentina de Fisiologia y Fisiopatologia, published in 1952. Unfortunately, the citation did not include an abstract. Using Educus and searching the phrase “human oxidative stress” yielded 1910 hits. Using PubMed and searching the same terms yields 845 and 116,077 hits, respectively. Clearly, the extensive literature about “oxidative stress” is based on a minimal foundation of hard data. That is why I call “oxidative stress” the alchemy of the 21st century. The emperor is not wearing any clothes.
The idea that antioxidants play no role in aging is supported by data showing that changing expression of most antioxidant enzymes in mice has no effect on lifespan, as discussed in the February 2013 Scientific American article mentioned above. I believe biomechanical material failure is more important in determining longevity (see my previous commentaries on this site entitled “Materials Science in Medicine: Not Just for NASA Engineers” and “Can Exercise Damage Your Arteries?”).
In the late 1990’s, my graduate student at Louisiana State University Health Sciences Center in New Orleans, Graham Hemelt, found that vitamin E supplementation had no consistent effect on blood viscosity at low shear rates in humans (unpublished data). Another study showed that plasma vitamin C levels but not dietary vitamin C intake were inversely associated with calculated blood viscosity at an unspecified shear domain.11
I have yet to be convinced of the value of antioxidant supplementation. I do not believe there is convincing evidence that oxidative stress plays a role in any disease except reperfusion injury. I see oxidative stress as a house of cards built on a foundation of quicksand. More rigorous application of the scientific method could cause this theory to collapse.
3. Vivekananthan DP, Penn MS, Sapp SK, Hsu A and Topol EJ. Use of antioxidant vitamins for the prevention of cardiovascular disease: meta-analysis of randomised trials. The Lancet. 2003; 361: 2017-23.
10. The Reference Values for Arterial Siffness’ Collaboration. Determinants of pulse wave velocity in healthy people and in the presence of cardiovascular risk factors: ‘establishing normal and reference values’. Eur Heart J. 2010;31:2338–2350.
For Further Reading: