I’ve gotten multiple questions through the website and a few comments on this blog about the study published in the current issue of the Journal of the American Medical Association (JAMA) indicating that taking antioxidant supplements leads to increased mortality. I’ll lay out my thoughts on this paper, but before I forge ahead, I’ve got a bit of housekeeping to do.
I’m behind on commenting on the comments, but I’ll try to get caught up soon. One of the reasons I fall behind is that I’m overrun with spam. I get somewhere in the neighborhood of 200-300 spam hits per day, some of which are pages long. I have a spam filter on WordPress that is pretty vigilant, but the problem is, it is too vigilant. Spam almost never makes it through into the Comments Awaiting Moderation page of my blog, but, unfortunately, many comments get ensnared in the filter and end up in the Spam page. This means I have to go through all the spam one by one to root out the valid comments in there and transfer them. And this transfer isn’t easy: it requires four steps to complete. I don’t know what the spam filter keys on because I get comments from people who have commented multiple times that end up in the spam pile. I get comments from people who have never commented before that end up there. As far as I can see there is no rhyme or reason as to how this happens, but it does.
I hate going through the spam because it takes forever, so I have a tendency to let it stack up until there are hundreds before I finally tackle it. I just did and found a bunch of comments, some from several days ago. So, if you’re wondering what happened to your comment, it was probably languishing in the spam file until I just rescued it. As far as I can tell – with this program, at least – there is nothing you can do to ensure that your comments get to the right place except to rely upon my industry in plucking them from the spam. And since I hate to go through all the spam, I don’t do it as long as I can find any valid reason (to me) to put it off.
Speaking of comments, my commenting on them is becoming quite time consuming. I like the interchange of ideas, so I’ll keep on commenting as long as I reasonably can. What I am going to start doing, however, is simply putting the comments up that require no comment on my part. Many people write in simply commenting on something I’ve written or on a comment someone else has made. Often I don’t have anything to add beyond ‘Thanks for commenting.’ It will save me a bunch of time to simply post these kinds of comments without my pithy responses.
Now, as to the JAMA study…
First, without even reading it I don’t think a lot of it. Why? Because it’s a meta-analysis. In my opinion meta-analyses are more often than not used for one of two reasons: to make something statistically significant that really isn’t or to do the opposite. Let me explain.
Suppose you want to do a study to prove that eating fast food once a day for six months would lead people to gain weight. You recruit people into your study, weigh them, give them coupons for six months worth of fast food, then weight them at the end. If a statistically significant fraction of people in your study gained a significant amount of weight over the six months, you could conclude – based on the data generated in your study – that eating fast food daily for six months causes people to gain weight.
Let’s say, though, that when you crunched all the data in your study you found that there wasn’t a statistically significant difference in weight gain. When you look through all the data you notice that some people gained a lot of weight, some gained a little, some stayed the same, and some even lost weight eating fast food for six months. As you drill down deeper in your data you discover that almost all of the people who lost weight while eating the fast food were people who work in jobs requiring a lot of manual labor. If you remove their data from your analysis, you find that now the people remaining in the study do gain a significant amount of weight on the fast food diet.
You now write your paper and say: We recruited subjects from various white-collar occupations and found that those eating fast food daily for six months gained a significant amount of weight. You title your paper: Six month fast-food causes weight gain. The press picks up your study and writes: Cheeseburgers and fries pack on the pounds.
But the study isn’t valid because you can’t cherry pick your data after the fact then recreate your study to fit the data the way you want it to.
But that is exactly what happens in meta-analysis studies. Instead of a single study, researchers amass data from a number of studies and pass it off as one. Problem is, they already know the results of the studies they’re allowing into their meta-analysis. They can set the standards so that studies they don’t want to include don’t make the cut and don’t sully their ultimate data. Whenever any kind of meta-analysis is done that gets picked up on by the press as this one has, people crawl out of the woodwork claiming that this study or that one wasn’t included that would have changed the outcome. (Click here to see what I mean.)
Often meta-analysis studies are used to give statistical significance to some finding that hasn’t been shown to be statistically significant in many smaller studies. For example, if you flip a coin you expect it to land on heads 50 percent of the time and tails the other 50 percent. But would you bet a lot of money that if I flip a coin ten times it’s going to land on heads five times? I doubt it because you know that ten times isn’t enough to ensure a 50-50 outcome. If I were to flip the coin 100 times, it would probably come closer to hitting 50-50, but you still probably wouldn’t bet the farm on it. If I flipped in 1000 times it would come closer, and 10,000 times closer yet. The point is, the more tries, the closer the outcome to whatever the odds really are, which, in the case of flipping a coin, 50-50.
Let’s look at this a different way. Let’s say that you have a coin that you think is weighted in such a way that it is going to come up heads more than it’s going to come up tails whenever it’s flipped. You flip it ten times and it comes up heads seven times out of the ten. That wouldn’t prove that the coin was weighted because there weren’t enough flips to overcome the random chance that the seven heads wasn’t simply a fluke of random chance. You know you could flip a normal coin ten times and get seven heads from time to time. So now you flip it 100 times and you get 65 heads and only 35 tails. Now, it’s looking like maybe the coin is weighted. But, even a hundred flips could give you this outcome simply by chance, albeit less chance than with ten flips. You flip it 1000 times and get 680 heads. Now you’re on to something because 1000 flips is enough to overcome the little runs of heads and tails that will skew the outcome on ten flips or even a hundred. If you flip it 10,000 times and get 6,920 heads, you know the coin is weighted.
Let’s say I want to prove that eating fat causes heart disease. I recruit fifty people, determine the amount of fat in their diets, and wait. Let’s say that I divide them into quartiles of fat intake and I find that over 20 years the quartile of those with the greatest fat intake have 5 heart attacks and those in the lowest quartile have 4 heart attacks. No big deal. Doesn’t prove a thing because it isn’t statistically significant. Now let’s say that someone else does a study that is designed a little differently–let’s say this study looks at saturated fat and heart disease and finds the same thing. That the highest quartile of subjects has one more heart attack than the lowest quartile. Same thing. No big deal.
Then we have a study that shows just the opposite. The lowest quartile of fat intake has one heart attack more than the highest. And another, structured differently yet, that shows one more heart attack in the fat-eating group. None of these studies prove anything–none are statistically significant.
Now comes someone who wants to do a meta-analysis. He (or it could be a she, but let’s call him a he) already knows the outcomes of these studies. He then structures his meta-analysis to include data from all the studies showing a non-significant increase in heart disease in those eating the most fat and excludes those showing the opposite. How does he do this? By setting the standards by which any paper makes it into his meta-analysis. He then takes all the data from these little insignificant studies and gives it significance because of the bigger numbers.
It would be like trying to prove a coin was weighted by taking ten ‘studies’ of coin flips that went 6 heads/4 tails (which isn’t significant), eliminating ten studies that went 4 heads/6 tails, and adding the data up to saying that there were 60 heads out of 100 flips, which would be significant.
The same thing can be done in reverse to show that a particular finding is not significant by combining enough studies, the outcomes of which are already known, to get rid of any significant data. And meta-analysis don’t particularly pay attention to how well studies are done.
If a well-designed, well-executed study is performed showing that a low-carb diet significantly reduces blood pressure is combined in a meta-analysis with a crappy study showing that there is no change in blood pressure with a low-carb diet, then the overall significance of the excellent first study is diminished.
That’s why I don’t like meta-analysis studies and never give them much of a second look.
Everyone who wrote asked what I thought of the JAMA study. No one really asked what I thought of antioxidants. From the above it should be clear what I think of the merits of the study. I don’t have any faith in it one way or another. I would have to go through all the literature, see what studies weren’t included, and read all those that were to get any kind of handle on what I think is really going on, and I have neither the time nor the inclination to do that.
I do have my own opinions on the use of antioxidants in large doses, however, and, although no one asked, I’ll give them to you anyway.
Most people think of free radicals as these little electrons dashing around the body dinging the cells here and there and bringing about all the consequences of aging. If only we could quench (everyone always uses the word ‘quench’ when speaking of eliminating with free radicals) the little buggers, we could live forever.
That much is obviously false because a large number of researchers have given zillions of subjects huge quantities of various antioxidants without any real change in longevity. Antioxidants have been studied enough to show that they aren’t the magic bullet to significantly delay aging. Which seems strange, given what we know.
We know that free radicals cause damage, we know that the accumulation of free-radical damage is one of the major causes of aging, we know that in a test tube antioxidants neutralize free radicals, so why don’t we live longer when we take antioxidants?
First, when we take antioxidant supplements they go into our blood. Most of the free radicals and free radical damage isn’t in the blood. It’s deep within the mitochondria, the little sausage shaped organelles that are the power-generators within the cells. The supplements we take don’t make it into the mitochondria, so they’re not really effective in protecting them. If mitochondria get severely enough damaged, they die. If cells lose their mitochondria, they lose their power source, and they die. When enough cells die, we die.
Before we can understand how free radicals are created, we need to understand what happens to the food we eat. We know that food provides us with the energy we need to live, but most people don’t really understand how we use the food we eat. When we eat a steak, how do we use the energy contained in the steak to power ourselves? We use it to convert ADP into ATP. ATP (adenosine triphosphate) is the energy currency of the body. It is a molecule with high-energy phosphate bonds that when cleaved release the energy required to operate all of the body’s functions. ADP (adenosine diphosphate) is converted to ATP in the mitochondria. Energy is required for this process, and that energy comes from food.
Various metabolic pathways break down the food we eat and reduce it to high-energy electrons that end up in the mitochondria. These electrons are passed along from one complicated molecular structure to another along the inner mitochondrial membrane until they are finally handed off to oxygen, the ultimate electron receptor. (I’m really simplifying this process; entire books are written about it. I’m just giving you the most basic gist.) As these electrons are handed off from one complex to another, the energy they release during the transfer moves protons (hydrogen ions: H+) across this inner mitochondrial membrane. An electrochemical gradient is created when these hydrogen ions stack up on one side of the membrane. The electrochemical gradient is the force driving the production of ATP from ADP. Energy from food creates the electrochemical gradient, the electrochemical gradient drives the production of ATP, so, thusly, energy from food is converted into ATP.
As the high-energy electrons are passed along down the inner mitochondrial membrane they occasionally break free. When they break free, they become free radicals. These rogue free radicals can then attack other molecules and damage them. Because these free radicals are loosed within the mitochondria, the closest molecules for them to attack are the fats in the mitochondrial membranes. If enough of these fats are damaged, the membrane ceases to work properly. If enough of the membrane doesn’t work, the entire mitochodrium is compromised and ceases functioning. If enough mitochondria bite the dust, the cell doesn’t work and undergoes apoptosis, a kind of cellular suicide. This chronic damage and loss of cells is the basic definition of aging.
So, if free radicals cause this damage, why can’t we stop it with antioxidants? We do. But not the antioxidants that we take in supplement form–those don’t make their way into the interior of the mitochondria where the damage takes place. Nature has endowed us with our own antioxidant system located within the mitochondria where, so to speak, the rubber meets the road in terms of free radical damage. The antioxidants produced require sulfur, which comes from the sulfur-containing amino acids, i.e. methionine. There are certain substances contained in particular foods that stimulate the enzymatic machinery that increases the production of these intramitochondrial antioxidants. Sulforaphane, for instance, a substance found in broccoli sprouts greatly stimulates a particular enzymatic pathway within the mitochondria, resulting in an increased production of antioxidants where they need to be. Sulforaphane has been shown to prevent cancer, vascular damage, and a host of other disorders thought to result from excess free radical damage.
Our defense against free radicals, then, really comes in two forms. First, the production of antioxidants within the mitochondria, and, second, by making the fats in the mitochondrial membrane less prone to damage. How can we do that? By making them more saturated.
Saturated fats aren’t prone to free radical attack–only unsaturated fats can be damaged by free radicals. Fats that have double carbon-carbon bonds, i.e. unsaturated fats, are the only fats susceptible to free radical damage. If the fats in the mitochondrial membrane are more saturated, then the membrane is less prone to free radical damage.
Do we know this will work or are we guessing? We’re pretty sure this is the case for a couple of reasons. First, when animals are calorically restricted (so far the only sure-fire way to increase lifespan), their membranes become more saturated. It was first thought that caloric restriction would reduce the production of free radicals, but it turns out that it doesn’t. Calorically-restricted animals keep firing off free radicals at about the same rate as their non-calorically-restricted mates, but the fats in their membranes become more saturated, presumably providing protection against assault by free radicals, allowing the animals to live longer. Second, we can graph the degree of saturation of membranes against longevity, and when we do, we find that animals that live longer have more saturated membranes. Take a bat, for example, compared to a mouse. Both weigh about the same, but the bat lives for about 20 years, the mouse for three or four. The bat’s membranes are much more highly saturated than are a mouse’s.
How can we increase the saturation of our membranes? By eating more saturated fat. In papers I’ve read, authors have cautioned against this approach (not wanting to appear ‘nutritionally incorrect’ of course), then have gone ahead and written about how they created a group of study animals with greater membrane saturation by feeding them more saturated fat.
Another way we can increase the saturation of the fats in the membrane is by keeping insulin levels low. There are enzymes in the cells that both increase the length of fatty acid chains (called elongase enzymes) and those that desaturate (called desaturase enzymes) the fats. The desaturase enzymes can make fats less saturated. Insulin appears to activate these enzymes, so chronically elevated insulin levels would tend to keep the fats in the membranes less saturated and more susceptible to free radical attack. I would venture that this is one of the reasons that hyperinsulinemia shortens life. One of the constant findings in studies of centenarians is a low level of fasting insulin, which would make sense given the ability of excess insulin to make the membranes more prone to free radical damage.
Many people seem to think that the cellular membranes won’t function well if they contain more saturated fat. They believe that a more rigid membrane creates problems for the proper operation of all the receptors and other large protein structures that reside in the membrane. They are right in a way, since a certain degree of fluidity is necessary, but where I think they are wrong is in their belief that the degree of rigidity or fluidity of the membrane is determined by the degree of saturation of the fats in the membrane. It’s determined by methylation, as was discussed in the previous post.
When you put the whole puzzle together, it’s pretty easy to see why a whole-food low-carbohydrate diet works to maintain health and longevity.
It provides plenty of good quality saturated fat to help protect the cellular membranes from free radical attack. It provides plenty of methionine, which is both a source of sulfur for the antioxidants in the mitochondria and a source of methyl groups for methylation of the fats in the cellular membrane thereby keeping them more fluid while at the same time more saturated. And it keeps insulin levels low so that the fats are not desaturated more than necessary, once again keeping the membranes less prone to free radical damage.
(One other way that low-carb diets help with health and longevity is by keeping the cells de-junked. As we age junk proteins accumulate in the cells. Over time these junk proteins can compromise cellular function. The generation of ketone bodies, a common occurrence with low-carb diets, helps keeps the cells clean. See here for a previous post on the subject.)
I believe the first and most effective defense against free radical attack is a good diet. Second is moderate exercise. (The effects of exercise on free radicals could be another long post, but for now, take my word for it: exercise reduces the production of free radicals) Third is the addition of a few supplements. CoQ10 and lipoic acid both act as antioxidants, but more importantly, they serve to regenerate the bodies own antioxidants. And a good vitamin supplement without massive doses of specific antioxidants isn’t a bad idea.
I take krill oil, fish oil, and curcumin daily without fail. I also take a vitamin E daily to stabilize the fats in the fish and krill oil. I take CoQ10 and lipoic acid several times per week. I take a multivitamin every now and then. And I take vitamin D3 in large doses throughout the winter. From time to time I take this or that other supplement depending upon what’s going on with my health, i.e. do I feel like I’m getting a cold?
I’m not a big fan of large doses of specific antioxidants because we weren’t evolved to take them. Plants live in the sun and produce oxygen as their way of life. Both the sun and oxygen are harmful if not controlled. Plants have evolved a complicated antioxidant system to protect themselves from sun and oxygen damage. We consume these antioxidants when we consume plants. We get tiny amounts of a zillion different kinds of antioxidants, not massive amounts of single antioxidants. And we get all the raw materials for the production of our own antioxidants from meat. (This post has gone on long enough, so if you want to read more about my view on antioxidants, read Chapter 5 in the Protein Power LifePlan.)
In my view, this is how nature intended us to get our antioxidants, and, with the exceptions mentioned above, this is the way I intend to get mine.
Are antioxidants harmful?