Back in 1972, Denham Harman predicted in a serious nutrition journal that an antioxidant supplement on top of a good normal diet might increase life expectancy by up to 7 years. Over the following decades the idea has been widely promoted that extra antioxidants in the form of supplements or antioxidant-rich foods would reduce our risk of chronic diseases like cancer and heart disease and help us to live longer healthier lives. Health gurus and those marketing foods and supplements have equated a designation of “rich in antioxidants” with “healthy”; merely demonstrating the presence of antioxidants in a food or supplement has been used as confirmation of its health-giving credentials. Ordinary foods have been designated as “superfoods” because of their high antioxidant content and new foods have been marketed to the British public as rich new sources of antioxidants. Efforts have even been made to breed new varieties of common plant foods with enhanced levels of antioxidants. The unquestioned message that has been promoted is “the more antioxidants the better”.
What is the basis of these extravagant claims and how strong is the supporting evidence?
What are antioxidants?
The name says it all; they are a variety of substances that block or quench the effects of harmful oxidising agents. During the normal metabolic processes of the cell, a range of highly reactive substances known collectively “free radicals” are produced and if not quenched these can cause oxidative damage to DNA, protein, fats and other components of the cell. Antioxidants should prevent or lessen this damage.
According to the oxidant theory of disease the cumulative damage caused by free radicals may be an important cause of cancer, heart disease and other degenerative age-related diseases and perhaps even a key factor in the aging process itself. For example DNA damage might lead to harmful mutations that cause cells to become cancerous. A number of harmful environmental stimuli are thought to produce at least some of their noxious effects by increasing the generation of free radicals e.g. radiation, UV light (sunlight), cigarette smoke, air pollutants and some harmful chemicals.
Not all of the effects of free radicals are seen as negative. When we have an infection and our white cells ingest invading bacteria then these are killed when the white cells produce a pulse of free radicals (the oxygen burst) that destroy the bacteria. Any infection or injury that leads to (chronic) inflammation will lead to increased free radical generation by white blood cells that infiltrate the damaged or diseased tissue. It is thus just possible that excess antioxidants might have negative consequences.
The body has evolved mechanisms to inactivate free radicals and repair the damage that they do. A number of essential vitamins and minerals are involved in our defences against free radicals. It is envisaged in the oxidant disease theory that some free radicals cause damage to cell components before they are quenched by the body’s antioxidant systems and this cumulative damage may eventually manifest as some chronic, life-shortening disease.
There are literally thousands of substances in food that have antioxidant activity. Some of these are essential nutrients but there are also many plant chemicals that are not essential nutrients but are nonetheless promoted as useful components of the diet because of their potential to act as antioxidants. Examples of antioxidants:
- Vitamins A, C and E the so-called ACE vitamins
- Essential minerals like selenium and zinc
- All 600 members of the carotenoid group of pigments; these include the vegetable form of vitamin A, beta-carotene, but most carotenoids have no vitamin A activity e.g. lycopene from tomatoes is a very powerful antioxidant but has no vitamin activity
- Many hundreds of flavonoids and other phenols and polyphenols found in foods like grapes, berries, other fruits, green tea, olive oil and red wine.
Most antioxidants are concentrated in foods from the fruit and vegetable group so a diet high in fruit and vegetables will tend to have large and varied antioxidant content.
The justification for taking extra antioxidants is that they minimise the damage done to our bodies by free radicals and thus slow down the development of chronic age-related diseases like cancer and heart disease. Smokers might be expected to get particular benefit from an increased intake of antioxidants because cigarette smoke increases free radical production. The belief is that high antioxidant intake should lead to reduced disease and delayed death and perhaps reduce the consequences of these harmful factors like smoke and radiation that increase free radical generation.
How strong is the evidence for the benefits of extra antioxidants?
There is a mass of so-called observational evidence which is generally consistent with the belief that extra antioxidants are good for us. If one measures the antioxidant intake of people or their blood levels of antioxidants then those with the antioxidant-rich diets tend to be healthier and live longer than those with lower antioxidant intakes. Fruits and vegetables are major sources of antioxidants and those who eat more fruits and vegetables tend to be healthier and live longer hence the advice to eat “five a day”. Observational studies cannot prove cause and effect but only show evidence of association i.e. that the two things tend to go together. High fruit and vegetable intake and thus high antioxidant intake is associated with increased life expectancy and reduced cancer and heart disease but this does not necessarily mean that more antioxidants cause these benefits or even that they are directly caused by fruit and vegetables. People who differ in their fruit and vegetable consumption also differ in many other respects particularly at the extremes of the range. Compared to people with the lowest antioxidant intake, people with the highest antioxidant intake and thus the highest fruit and vegetable consumption would also tend to:
- Have higher intakes of other essential nutrients
- Have higher intake of dietary fibre
- Have a lower (saturated) fat intake
- Consume less animal protein
- Have higher incomes and better education
- Have a higher level of physical activity
- Have a lower risk of being obese
- Be less likely to smoke or drink to excess
- Be generally more health conscious and adopt healthy lifestyle choices.
Epidemiologists who conduct observational studies cannot control these other variables in advance and so they try to statistically correct for them after the study has been completed. This process of correction is very imprecise and is liable to major errors and biases. Which variables do they choose to correct for and how do they carry out this correction process? They may have limited or unreliable information about some of these other so-called confounding variables. This correction process is an imprecise process and can be subject to considerable bias. For these reasons, it is accepted that observational studies cannot prove cause and effect.
Short term experiments
There are huge numbers of experimental studies using people and animals that have looked at the short term effect of giving subjects extra antioxidants in the form food or supplements. These experiments rely upon using short term proxy measures of an animal or person’s antioxidant status or some measure that is believed to indicate oxidant stress or oxidant damage. Most of these studies show that taking extra antioxidants improves antioxidant status or reduces the supposed indicators of oxidant stress or damage. This is almost inevitable given the way these measures of oxidant status or stress have been chosen and developed.
Positive results showing that a particular food or supplement improves antioxidant status or reduces some supposed measure of oxidant stress are often interpreted as providing evidence that the food or supplement may prevent cancer or heart disease. This is a great leap of faith and depends upon the assumption that these short term measures are good indicators of long term disease risk. Back in 2007, I came across a headline on the BBC News web-site with the headline Watercress “may cut cancer risk” describing the results of a study carried out in Northern Ireland. Healthy volunteers were persuaded to consume a whole bag (85g) of watercress every day for 8 weeks i.e. probably much more each day than most people consume in eight weeks! After 8 weeks these volunteers had higher levels of blood antioxidants, had reduced amounts of oxidative damage to the DNA in their white blood cells and their white blood cells appeared to be less damaged when exposed to an oxidising agent (hydrogen peroxide). Watercress, like most green leafy vegetables, is known to be a rich source of antioxidants and so one might well have predicted the general pattern of results at the outset and similar results would probably have been obtained with other green vegetables. Whilst I do not question the data, I question whether it provides significant evidence that eating normal amounts of watercress will reduce a person’s risk of cancer. This experiment did generate some very helpful headlines for the sponsors of this research, the Watercress Alliance!
Testing the effects of antioxidants in clinical trials
The evidence from observational studies and short term experiments eventually led to a number of long term clinical trials of antioxidant supplements being conducted. These are expensive and time-consuming studies where large matched groups of thousands of subjects are given supplements or placebos. The best of these studies have been conducted using a double blind design where neither the subjects nor the scientists know who is receiving real or placebo tablets until the data has been collected. Such randomised controlled trials (RCTs) are regarded as the “gold standard” of evidence in medical science because if well designed and executed they minimise bias and should give a proof level answer to the specific question being tested. As far as I am aware no RCTs have demonstrated any holistic benefit of antioxidants upon cancer, heart disease or risk of death. In some of them, especially those using smokers, the death rates were clearly higher in the supplemented groups when the supplement contained beta-carotene. This is ironic because smokers were chosen as subjects for these trials because they were considered more likely to benefit from taking extra antioxidants. Most of these RCTs have used varying combinations of the major antioxidant vitamins and minerals, including beta-carotene, vitamin E, vitamin C and selenium. These studies started being published in the early 1990s and the results have been quite consistent. Enough of these trials have now been conducted for them to be used in meta-analyses. Meta-analysis is a powerful tool where similar studies are combined together in a formal way to give a summative result from all the available studies; they have sometimes been dubbed the platinum standard of evidence. These meta-analyses confirm the lack of any indication that antioxidant supplements produce any reduction in cancer, heart disease or total mortality. They also confirm that large supplements of beta-carotene increase death rates at least in smokers. Some meta-analyses have even suggested very small negative effects of other antioxidant supplements or that large beta-carotene supplements may more generally harmful.
Although people who eat a diet rich in antioxidants from fruit and vegetables tend to be healthier and live longer than those who eat a low antioxidant diet, taking extra antioxidants does not prevent cancer or heart disease or increase life expectancy. There is clear evidence that some supplements increase death rates at least for smokers. Concentrated antioxidant supplements are more likely to do harm than good. A Department of Health report counselled against the use of concentrated beta-carotene supplements almost twenty years ago but supplements of this substance are still being marketed and sold.