Antioxidant supplements - in practice

In the last post on this topic, we saw that there are tentative theoretical reasons for why it might be worthwhile augmenting our natural stores of antioxidants with supplements (pharmacological or natural, like fruits). As I emphasised there, by itself, this is not enough. We have to go and do the experiments and find out the truth of the matter.

Of course, there are experiments and then there are experiments. One all the varieties of study methodologies, the one most prone to misinterpretation by the press and public are the observational studies. Let's say we want to see if antioxidants have any effect on cancer prevention. Since vegetables are a good source of antioxidants, we can go and collect two large groups of people - those who eat lots of vegetables and those who almost never eat vegetables. We then collect data on the incidences of various cancers from each group and compare them.

Now let's say that the veggie-guzzling group showed a statistically significant reduction in the number of cancers that developed. Hooray for antioxidants, we may be tempted to say - they seem to reduce cancer!

But this would be a gross misreading of the experiment, although a very common one. The problem with observational studies is that the two groups already exist, and are simply picked by the researcher. Seeing as they already exist, they are not randomly generated, and thus prone to hidden statistical bias. For instance, the above result is certainly compatible with the theory that antioxidants reduce cancer, but it is also compatible with the theory that:

• Some other substance(s) in vegetables, and not antioxidants, reduces cancer.
• The people who tended to eat more vegetables also tended to smoke less, and so developed less lung cancer.
• The people who tended to eat more vegetables also tended to exercise more, and thus showed reduced cancer rates.
• The people who tended to eat more vegetables also tended to be wealthier, and could thus afford regular cancer surveillance screenings, thus lowering their incidence of cancer.

Apart from the first option, you can see that the observational studies are perpetually vulnerable to perfectly valid objections of the type "The people who tended to eat more vegetables also tended to [X], and it was actually [X] that reduced the cancer rates, not the antioxidants in vegetables." In other words, it is almost impossible to separate a correlation (which observational studies show very nicely) from a cause (which is really what we need to know about).

Discussing the pitfalls of observational studies is necessary, because there are many examples of such observational studies that show that antioxidant intake (usually in the form of vegetables and fruits, but sometimes in the form of artificial supplements) correlates with healthier outcomes (e.g. lower cancer rates, lower rates of macular degeneration, lower rates of strokes and heart attacks, etc.). Whilst these studies did show the need for further research, by themselves they are not convincing.

For the convincing study types, we must turn to randomised control trials (RCT), the gold standard of studies. In this setup, two groups are selected from a population at random. In this way, they should have the same number of fruit eaters, the same number of smokers, the same number of doctors, the same number of gamblers, etc., etc.. Only once we have the two groups do we introduce the variable that we wish to test for; in other words, antioxidants are then given to one group, and another substance, usually a placebo, is given to the other.

For good measure, the RCTs are usually double-blinded if possible - neither the patients nor the people administering the antioxidants know whether a particular person is getting the antioxidant or the placebo. This is to ensure that no subtle psychological factors play themselves out here. (For instance, a patient who knows he's getting the placebo may opt to eat more healthily to compensate.)

Although RCTs do have their down side, it should be clear that they are much more accurate in determining whether antioxidants cause better health outcomes, rather than are simply correlated with them.

And so, I can now reveal the results to you... (drum roll, please)...

With very few exceptions, properly designed randomised control trials have failed to show a benefit from antioxidant supplementation. In other words, antioxidants don't appear to be of benefit at present.

To give you some idea of the research, I looked at the four most prestigious medical journals - the Lancet, the Journal of the American Medical Association, the British Journal of Medicine, and the New England Journal of Medicine for recent research on this topic. Since these journals usually only accept the best designed trials, they are the most reliable. Here is a sampling:

• This study showed that "Vitamin E had no benefit in patients with mild cognitive impairment".
• This study concluded that "There is insufficient evidence to support the role of dietary antioxidants, including the use of dietary antioxidant supplements, for the primary prevention of early AMD [age-related macular degeneration]".
• This study "provides no evidence to support the use of antioxidant or folinic acid supplements in children with Down’s syndrome."
• This study found that "in patients at high risk for cardiovascular events, treatment with vitamin E for a mean of 4.5 years has no apparent effect on cardiovascular outcomes."
• This study's authors believe that the use of antioxidant vitamins in the setting of patients with coronary disease and low HDL levels "must be questioned".
• The results of this study "[did] not support the hypothesis that depletion of vitamin E, selenium, cysteine, or riboflavin has a role in the development of kwashiorkor."
• This recent study ended off by stating that, "In this large, long-term trial of male physicians, neither vitamin E nor vitamin C supplementation reduced the risk of major cardiovascular events. These data provide no support for the use of these supplements for the prevention of cardiovascular disease in middle-aged and older men."
• This metanalysis concluded, on the basis of several well-designed randomised control trials that vitamin E had no beneficial effect cardiovascular disease, and that beta carotene [a vitamin A derivative] led to a slightly increased risk of cardiovascular death.
• This metanalysis looked at the effects of common antioxidant supplements on mortality in several diseases, and showed that "treatment with beta carotene, vitamin A, and vitamin E may increase mortality." [bold type added]

Of course, there are dissenting results in the midst of this, but they are at present hugely out-numbered by the RCTs that demonstrate either no benefit (usually) or increased risk (rarely). The latter is particularly disturbing, as it suggests that even the old assumption on antioxidants ("they may not work, but at least they aren't harmful") may be wrong.

I must stress that the above is not some strict and comprehensive review of the literature - it is meant for interested laymen. The realities are more complex and subtle. However, I think that the research may accurately be summarised thus:

Although there are both some theoretical reasons and several observational studies that indicate that antioxidants may be protective against several diseases, the randomised control trials have not, on the whole, backed this up. Most of them have failed to show an improvement in mortality rates. At present - and this may change as new evidence comes in! - there is no strong evidence to suggest that supplemented antioxidants are routinely beneficial.

Furthermore, on present evidence, we can conclusively say that popular "complementary" health adverts that claim extravagant benefits for antioxidants are simply wrong. And attempting to back up your claims by alluding to one or two observational studies from 20 years ago, whilst ignoring the succeeding RCTs on the same topic, is frankly dishonest. At best you can say that the evidence is equivocal - you certainly can't pretend that the research unanimously shows that, say, antioxidants prevent cancers and skin aging.

This rather startling conclusion generates a paradox: if supplemented antioxidants are unhelpful, why would the body be apparently going to such lengths to have its own precious antioxidant systems? I'm not sure of the answer, but there are a few options:

• Option 1: The body is already at its optimal free radical/antioxidant balance (remember: free radicals are beneficial in selected instances), and any additional antioxidants don't result in additional benefits.
• Option 2: The body thinks it is at its optimal free radical/antioxidant balance, based on what was best for most of our evolutionary past, though not what is best for us today - and thus it fails to absorb or utilise any other supplemented antioxidants we may feed it.
• Option 3: The administered antioxidants are too blunt an instrument for the body to derive benefit from. For instance, it would be nice to have elevated antioxidant levels in the brain (where it may protect against Alzheimer's disease) but it might not be a good idea to have these elevated levels in neutrophils (which require free radicals to destroy the bacteria). Thus consuming extra antioxidants results is both positive and negative effects, which usually cancel each other out.

There may be other plausible explanations too - feel free to add them in the comments section if you wish!