The Trouble with Nutrients
Ask any ten people you know about the benefits of vitamin supplements, and you are bound to get ten different answers. Search the web, and the opinions are nearly infinite. How can we possibly know whom to trust? How is any logical conclusion possible?
One approach to this problem is to consider how we know what we know. Philosophers would call this epistemology, and it is a useful starting point for any complex question. For example, in medicine, knowledge is acquired through various types of “research”, and the “gold standard” of research is: the randomized double-blind placebo-controlled trial.
Let’s break that phrase down a little.
a “controlled trial” is simply an experiment designed to test a certain treatment using different groups of subjects: a “test” group (which gets the treatment) and a “control” group (which doesn't )
a “placebo controlled” trial is when the control group gets a sugar pill instead of simply nothing
“randomized” means that the subjects are assigned to either the control group or the test group at random
a trial is “double-blind” when neither the testers nor the subjects know who is getting the placebo and who is getting the real treatment
So basically, we take a group of similar people and split them randomly into two groups. We hire other people to distribute real pills to one group and sugar pills to the other, where neither the distributors nor the subjects know which is which. And then we measure the outcomes and compare the two groups.
This allows us to tell whether a treatment is actually doing anything, and if it is, whether it is actually because of the treatment or if people are just responding to taking a pill everyday or getting attention from a doctor or the phase of the moon or some other unrelated factor. This is a good way to eliminate bias and objectively test if a drug is effective.
However, it is an absolutely absurd way to test a nutrient — for multiple reasons.
To illustrate why this is, let’s use a simple example and say we want to compare Prozac and vitamin D as treatments for depression. For our hypothetical study, we’ll randomly assign depressed patients to a Prozac treatment group, a vitamin D treatment group, and a placebo group.
The Challenges
Defining Disease
Classically, physicians believed that diseases were caused by problems. Believe it or not, that is not actually the case now. Most modern physicians commonly treat “diseases” whose cause is unknown, eg, high blood pressure, arthritis, cancer, diabetes. In fact, in psychiatry, most conditions are defined as not being attributable to “the physiologic effects of … another medical condition.” When it comes to the case of depression, if we actually know the cause, we can't have the “disease.”
So when doctors prescribe antidepressants, they are not treating a physiological problem but are instead “treating” a list of symptoms that a committee has decided to call “Major Depressive Disorder.”
How did that happen?
Well, if we take groups of people with the same symptoms and keep giving them different psychoactive drugs, eventually we'll find some that work better than others (at least for some of the people, some of the time) and voila! An antidepressant is born (and, incidentally, so is a small fortune). Not only do we get a new drug, it “confirms” that “Major Depressive Disorder” must be a “real” disease.
Prozac, by definition, treats this type of “depression:” a cluster of symptoms that may have numerous different causes (as long as you don’t know what they are). If it didn't treat that particular definition of that “disease,” it wouldn't be available. No one suffers from a Prozac deficiency.
Vitamin D, on the other hand, treats one thing: vitamin D deficiency. No committee decided that our bodies need vitamin D to function properly; deficiency is a fundamental and undeniable physiologic problem in itself. It may cause a wide array of different symptoms, however, because, like all vital nutrients, vitamin D has many different roles in the body.
In a growing child with severe Vitamin D deficiency, you may get bone deformities that leave you crippled, while an adult with a mild deficiency may have only a spell of insomnia — ie, one cause but many symptoms. Some people (quite a few in my experience) will have symptoms that overlap with “depression,” such as fatigue, sleep changes, reduced concentration, etc. If no one checks their vitamin D levels, they may very well be diagnosed with depression.
But back to our hypothetical study and our two treatment groups. Some of the depressed people in the Prozac group will probably respond positively — about two-thirds, usually (based on past studies). Some of those in the vitamin D group will probably also respond, but these numbers will be all over the map — literally, as it turns out. If it's a group of Floridians in August (who probably already have good levels of vitamin D), very few will respond. If it's a group of Canadians in February (many of whom have low vitamin D), however, more will respond.
Regardless, it will almost always be fewer than the Prozac group, because although they all have “depression” (which Prozac treats), they don't all have vitamin D deficiency (which vitamin D treats). In terms of response rates, the Prozac group would score highest, then the vitamin D group (maybe), then the placebo group. We would conclude that Prozac is a better treatment for depression than vitamin D.
By defining our "diseases" as syndromes without specific causes, we have greatly predisposed in favor of pharmaceuticals.
Imagine how different our conclusions would be, however, if we defined the illness as vitamin D deficiency. How effective would Prozac be at strengthening your bones, for example? And what if your depression were caused by vitamin D deficiency — would Prozac do better than vitamin D? With fewer side effects?
But we can’t do that study, because we have already defined depression as something without a known cause. Ironically, no one with known vitamin D deficiency can have “Major Depressive Disorder” (although many people with vitamin D deficiency are, in fact, depressed).
Of course, your average depressed person doesn’t know he or she is deficient in vitamin D. There is a simple test for that, but why would your doctor perform it, if the “studies show” that vitamin D is not effective for depression anyway? The circular logic here is head-smackingly maddening.
On Your Mark…
We all start out with a zero body level of Prozac: everyone's at the same starting line, so to speak. So in our hypothetical study, the Prozac group goes from zero Prozac in their blood to some Prozac in their blood (probably within a certain range, with some individual differences).
However, we all start out with some level of all vital nutrients, including vitamin D. Everyone is starting the race at a different place. Some people may be 100 yards behind from the very beginning, and some may be 100 yards ahead. In the vitamin D group, patients go from some vitamin D to more vitamin D — which may be good for some people, but may actually push others into a toxic range. It’s possible that you may hurt as many people as you help, which makes the net effect less positive.
But in order to “randomize” our sample, we have to split our subjects up into “test” and “control” groups irrespective of their starting points. If we treat or divide them differently, it wouldn't be “random.” In all the dozens of studies l've read on vitamin D (and other nutrients, for that matter), almost no one checks initial levels, probably because their experiment would no longer be considered “randomized.” Studies that aren’t “randomized” are considered “poorer” quality, and poor quality studies don’t get published.
By declaring that all “good” studies be “randomized” this way, we again greatly predispose to positive results for the drug and negative outcomes for the nutrient.
The Un-control Group
To review: in our hypothetical study two of our groups get a treatment, and the remainder get a “placebo” (sugar pill). For our drug group, this makes sense: Prozac vs. no Prozac.
But what about our vitamin D group? How do you keep the control group from getting vitamin D? Do you keep everyone indoors the entire time? And what about vitamin D in food — do you restrict food intake? Is it even ethical to deprive people of a vital nutrient which you know they need?
Except under very tightly regulated situations, you can never have a true “control” group for a nutrient study. Instead of Drug vs. Placebo, it will always be Nutrient vs. Placebo + Nutrient.
The point of the control group is to eliminate variables, but it doesn’t work for nutrient studies. All control groups do is dilute the effect of the treatment. When it comes to nutrient studies there are other parameters that are better indicators of validity than just whether it is “placebo-controlled.” Once again, requiring that the study be constructed in this way places the nutrient at a disadvantage.
Dosage (Size Matters)
“It is only the dose which makes a thing poison.”
This is true for both drugs and nutrients, of course. However, as we mentioned earlier, when it comes to starting points, we are all the same for drugs and all different for nutrients. A small dose of Prozac, therefore, is likely to have a small effect on most people. However, a small dose of vitamin D is likely to have a negligible effect on someone who is severely deficient and yet may be toxic to someone with an already high vitamin D level.
Since we can't test levels ahead of time (that would be un-random) and since we can't assign different doses anyway (that would be un-blinded), this poses a conundrum for the researcher. Since most researchers do not want to cause toxicity in otherwise healthy people, they select a very small dose of a nutrient out of caution. While that may be admirable in terms of the safety of your subjects, it definitely shortchanges the impact of the nutrient. In the case of vitamin D, a dose of 400 IU is common in many studies.
(In my practice, I commonly test a blood level, then recommend a supplement of vitamin D, and then re-test about a month later. Many people on a dose of 5000 IU per day show only a negligible increase in their blood levels! I personally had to take 10,000 IU daily for many months just to bump my numbers barely over into the "sufficient" range.)
In fact, there is quite a range of doses people need to get to the same blood level. One reason for this is that we all already have a natural method for getting vitamin D and — spoiler alert — taking a pill is not it. Everyone’s body is looking for vitamin D *from the sun.* Our bodies react differently to oral vitamin D supplements, and pills may just not work for some people. Some people may just need to be in the sun.
Prozac, however, is equally unnatural for all of us:.
no one's body is “looking” for Prozac somewhere else — in fact, no one's body is looking for it at all (no one has a Prozac deficiency, remember?)
the body doesn’t make special pathways that build it or construct special proteins to carry it around in the blood or have special control mechanisms to direct where it goes
it just shows up and does what it does
The drug has been designed for oral administration as a pill. Before being tested in humans, drugs go through animal testing so the manufacturer knows what range of doses will cause what range of blood levels. The dose-response curve is pretty well established for most drugs. If it is erratic or irregular, the developers scrap it and look for something more reliable.
Drug manufacturers are also required to figure out what the lethal dose is (in lab animals), so they can steer well clear of the toxic limit when they design studies in people.
So, again, it is much easier to do a study with a “therapeutic” dose of Prozac which is within a pretty small and predictably non-toxic range, than with a “therapeutic” dose of vitamin D which may vary widely and which researchers will usually under-dose to avoid toxicity.
Alone, Together
Because they need to be double-blinded and placebo controlled, nutrient studies generally use nutrients in pill form. What exactly is the double-blind placebo control for broccoli, by contrast? But, like with vitamin D, pills are not the way we were designed to get nutrients. Broccoli, for example, contains folate, vitamin C, sulfur, and iron: what a coincidence that those nutrients all work together! Nutrients naturally come in synergistic combination packs (otherwise known as food), which affects how they get absorbed and how they work. They don't work alone, any more than a single spark plug does.
You could have great spark plugs, and an entire fully functional car, for that matter, but be deficient in one thing — say, gas — and you won't get very far. Would it make any sense to take 100 different cars, replace one spark plug in each of them and then compare a random parameter like 0–60 acceleration — without any attention to the make or model, the rest of the engine, the transmission, the gas in the tank, the air in the tires, or any of the other thousands of parts that make up a car? And yet that is *exactly* how much sense it makes to study nutrients in isolation.
On the other hand, most modern drugs like Prozac block something in the body. You can do that pretty easily with a single agent: throw the same wrench into a hundred different gear boxes, and you can effectively disable the car — most cars, in fact. No complicated process there.
The difference between nutrients and drugs is like the difference between building a functional automobile from scratch and breaking one.
This is not to say that drugs are never helpful; they do sometimes work. I do actually prescribe them. But drugs and nutrients are completely separate entities, and it is absurd to study them the same way and then draw any even half-way logical conclusions. Even though the randomized. double-blind, placebo-controlled trial may be grade "A" or “Level 1” evidence for a drug study, it is pretty useless when evaluating nutrients. It is misleading [at best] to grade evidence on a Level 1–5 scale and apply that standard to all fields of study indiscriminately.
Why on earth would anyone think that biological systems could fit so neatly into this single, tightly circumscribed method of study? Life is just more complex than that.
Examining the epistemology therefore explains why there is so much conflicting information about nutrients and natural herbs. When people (including doctors) solely rely on randomized, double-blind, placebo-controlled trials to assess nutrient therapy, chaos ensues. Rather than throw up our hands in despair, though, if we are patient and a bit more discerning, we actually can extract meaningful information from the evidence available. Like most things that require some study and effort, the results are well worth our time.
A Different Approach
In the coming posts we will be examining the evidence about nutrients, herbs, and natural supplements, but be forewarned that I will not be relying on the traditional, overly reductionistic “1–5” evidence scale that is usually used to stratify medical studies. Since the vast majority of medical literature strives to adhere to this scale, it means we will be looking elsewhere most of the time.
I will be relying more on the observational and descriptive work of researchers who have painstakingly traced biochemical and physiological pathways in both healthy and disease states, as well as the clinical experience of myself and my colleagues. One advantage we have is that we can test for substances which should be in the body — like vitamin D, which we’ve discussed. There is very little down side in correcting a deficiency; even if it doesn’t resolve the issues you were targeting, there will be some other benefit.
In most cases, I recommend that you find appropriate testing before taking extra nutrients, especially in massive doses, although there are a few exceptions to this. There are direct-to-consumer laboratories now and fee-for-service wellness centers focused on optimal nutrition where you can access such testing, even if your regular doctor is not willing or able to order tests for you.
In the next few weeks, we will be looking at some natural substances that are particularly helpful to the immune system (including vitamin D, as it turns out), but in a contextual way that allows you to understand what might apply to you or not. For example, we will look at what natural substances do, what co-nutrients are necessary, what deficiency looks like, what toxicity looks like, what levels you can test, and various other information. We’ll look at nutrients first and then at various herbs, which have their own fascinating plant chemistry.
Until then,
Stay thinking, my friends.
