How to understand the evidence from clinical trials of nutrients

Over the last 50 years, thousands of clinical trials have been performed to test the efficacy of various nutrients for managing medical problems. At times, the results are all over the place. This makes some people conclude that it is better to stay off the nutrient supplements till conclusive results appear.

In this article, I will give you my views on how to correctly interpret such trial results and draw meaningful conclusions from them.

Problems with Nutrient Trials

Many researchers, who are from the world of pharmaceutical clinical trials, find nutrient trial results very difficult to digest. A chemical entity either works, or does not. A nutrient should also either ‘work or not work’. But that is not correct.

The reason is nutrients are different in nature from pharmaceuticals. Let us understand the main points of differentiation.

Nutrients Come from Varied Sources

Depending on the source, the exact same nutrient might be completely different.

In general, a pharmaceutical-grade chemical is identical the world over. Atorvastatin is a pharmaceutical medicine, from a class known as statins that are used for lowering cholesterol.

Atorvastatin is atorvastatin; and atorvastatin 10 mg will be atorvastatin 10 mg anywhere you buy it in the world (with slight, permissible variation in the exact weight). It will behave exactly the same way independent of where you bought it from.

There are 633 brands of atorvastatin available in my country (India), and unless some company is doing something unethical, all of them contain exactly the same thing—atorvastatin—and nothing else, or at least, nothing else that matters for cholesterol-lowering.

But omega-3 fish oil is different in every single trial.

First of all, fish oils come from fish, which is a source of variation. Depending on whether you are using salmon, krill, anchovy, herring, mackerel, or sardine, the contents of your fish oil will change. Here is a statement from the National Institutes of Health about this:

Understood? Neither did I.🤣 I quoted it to point out that fish oils themselves have so much variation in them, even if they are standardised to what is called a ‘180 mg DHA—120 mg EPA’ dose.

And, yes, the spelling is right. It is the National Institutes of Health, the plural of the word institute. It is the U.S.A. government’s apex body involved in biomedical and public health research. It comprises of 27 separate institutes and hence the plural form is used. If a statement comes from them (it?), consider it a fairly official statement from the U.S.A. government itself.

Nutrients Work Together

Many food items work together with other nutrients. Unlike most pharmaceutical products, their efficacy changes based on the presence or the absence of other nutrients. Here is a comprehensive article on this website about the nutrients that work together.

On the other hand, some nutrients don’t work well in the presence of other nutrients. Let’s say that we plan to investigate whether omega–3 fish oils are beneficial for asthma and allergies. Asthma and allergy symptoms are, primarily, inflammatory responses of the body. Now, omega–3 oils are known to be anti-inflammatory (they reduce inflammation) and we logically think that they should help.

Alas, omega–3 fish oils also contain omega–6 oils. Typically, omega–6 in fish oils is about 15% of the omega–3 present in it. And, omega–6 are ‘generally’ pro-inflammatory (increase inflammation). The reason I said ‘generally’, is because, in the last few years, some components of the omega–6 oils are also found to be anti-inflammatory.

To complicate matters even more, different fatty fish, whose oils are used for studies about omega–3, contain drastically different levels of omega–3 and omega–6 oils.

I don’t have the numbers for various fish oil brands, but I have given below omega–3 and omega–6 contents of edible versions of fatty fish themselves. It is quite likely that the preparation of fish oil used for a study would have a completely different profile of omega–6 oils in it, altering its effect on the inflammation in asthma and allergies.

As you can see, the ratio varies from 8% to 3108%, a variation of a factor of 388 times. Just to add a visual perspective, here is the bar graph of the oil contents in the same fish below.

Good luck if you are going to draw a conclusion, which says something like this: “Our study showed omega–3 fish oil did …”.

Starting Levels of the Nutrient

At the start of a trial, participants may have varying levels of the nutrient in their bodies.

Using our earlier example, when you start a clinical trial for atorvastatin, every single participant has the same starting level in his blood: zero.

Atorvastatin is not a part of the human food chain, nor is it a native nutrient in our body. You measure the amount of atorvastatin given to a participant, and you know exactly the maximum amount that could be present in the person’s body.

However, when you start a trial for omega–3 fish oils, different individuals will have different blood plasma levels of omega–3 oils, at the start of the trial. The amount of omega–3 fish oil you give to the person is not the same as the amount at the disposal of his body. And the results of the trial will depend on the amount his body has, and not the amount you give him in the trial.

So unless you measure the blood plasma levels of various omega–3 oils, before, during, and after the trial, how can you comment on whether giving fish oil helps? Your conclusions should not mean anything.

In fact, if you take a closer look at the data in many nutrient trials, you will find that people who were deficient in the nutrient at the start, benefit from taking it. However, those who were already replete with adequate amounts of that nutrient, do not benefit much.

When the data are taken together, it averages out, and hides, the extra benefit seen in deficient people.

Now, it is not a good idea to use only the really deficient patients for the trial, without giving them additional supportive treatment. How can you knowingly allow seriously deficient patients to potentially suffer, during the treatment?

Of course, there are certain types of trials, such as cross–over trials, to take care of this. But, they are done rarely.

Triage Theory for Nutrients

If having high starting values of a nutrient may wash out any possibility of conclusive results, very low starting values run a risk of falling afoul of the Triage Theory. Dr Bruce Ames proposed his Triage theory in 2007. Read: What is Triage Theory?

The logic here is simple: If the body is severely deficient in any nutrient, it starts to ration it out. It is first used for immediate and essential body functions. Long–term requirements, such as cell repair and buildup, are postponed until the nutrient is available in adequate quantity.

Of course, you also would do that if your house faced a water–shortage. Out goes a long bath in a bathtub. You will conserve the water for drinking and cooking needs. Your body does exactly that, theoretically. And now, new research proves that the theory is correct from this study on Vitamin K.

So what will happen if we involve participants with severe deficiency of omega–3 fatty acids in our trials? The omega–3 oil supplement that we give them will be first diverted for the most essential tasks, such as building the brain cell membranes and not towards reducing asthma inflammation.

Here are 17 science-based benefits of omega–3 fish oils, but not all of them are equally vital for survival. So the participants’ bodies are going to choose where to use the newly acquired omega–3 fish oil, and most likely, that is not going to be to reduce asthma and allergy symptoms. How do you expect to see a positive result from such a trial?

And, the last time I checked, atorvastatin was not needed for any of the vital, or even non-vital, body functions. So when you give atorvastatin to a participant, his body is unlikely to divert it to fix his brain or eyesight. It will be used for lowering cholesterol, or whatever else it will be tested for.

In my view, the triage theory is the main reason why nutrient trials will never give as conclusive results as pharmaceutical trials will. Your body will always override your need to target a specific use of a nutrient, triaging it for the most important need.

Sample Size and Trial Duration

Many pharmaceutical trials are halted before they reach their endpoints because the participants face worse side effects, or develop some medicinal toxicity.

On the other hand, since nutrients are a part of our lives, their trials usually cause no extra harm, whether you see a benefit or not. So, in most nutrient trials, you will never notice any conclusion saying the situation worsened, or the patients were harmed.

Usually, the outcome in nutrient trials is that there was no result—the conclusion that the benefits observed were within the statistical range of the randomness.

In other words, the conclusions say that the benefits were so less pronounced that they could have happened by chance, without any genuine, underlying effect.

Remember this when you hear a conclusion that says “A showed no relation with B”. It actually means “the relation of A with B was so statistically insignificant that, we conclude that, the effect was just an artifact of pure luck—there was no real effect”.

As we discussed above, people start with non-zero levels of nutrients in their bodies. So the difference noticed is usually not that drastic. For example, if we take two patients, in a pharmaceutical clinical trial, you may have starting points as 0 and 0, and endpoints as 18 and 23. In a similar situation with a nutrient clinical trial, you may have starting points as 8 and 11, and the endpoints may be 18 and 23.

Basic statistics tell that in the former (pharmaceutical) case, the change is statistically more significant. On the other hand, in the latter (nutrient) case, the change might be outside of what is called the 95% confidence interval. That makes its statistical significance low. Hence, the nutrient trial might conclude that “there was no evidence to prove that…”.

I know that I am making this sound so trivial. But you get the point.

Type II Error and Resolving Power

In statistics, the example above is called type II error, which is when you conclude that there is no effect when there is actually an effect. The statistical resolving power of a trial is the likelihood that an effect is detected by the trial, when there is, in fact, an effect.

In other words, for nutrient-based clinical trials, you will need a higher resolving power of the trial, than for an equivalent pharmaceutical clinical trial.

To rephrase, if you apply similar statistical rigour to both the cases, the nutrient trial may not show any conclusive results, while the pharmaceutical trial may.

The ways to increase the resolving power of a trial are to enhance the effect (get sicker patients), add more participants, or extend the timeline of the trial.

Involving patients with more severe conditions may not work well, since in some extremely sick patients, the relevant nutrient or the medicine may not work properly (as clarified earlier with the example of severe asthma patients).

Increasing the number of patients involved in the trial makes the result less likely to occur by chance. That may help the conclusions to fall within acceptable statistical limits.

Extending the timeline (duration) of the trial makes the differences between treated (‘case‘) and untreated (‘control‘) people more prominent. Nutrients often need longer time to work their benefits than pharmaceuticals. Many nutrient trials are done only for 3 to 6 months, which might be just too short a duration to show benefits.

The problem is both the latter effects are non-linear in nature. To get double the confidence in your conclusions (statistical significance), you may need four times the number of participants. To get double the confidence in your conclusions, you may need to increase the trial timeline by many times, depending on what you are studying.

All these have cost and time implications. As a result, a clinical trial of nutrients, in general, may throw a non-result when the same quality and quantity of testing for a pharmaceutical molecule may give a positive result.

Note that a non-result is not a negative result. If you did not notice any black swans or ghosts, it does not prove that they don’t exist anywhere else in the world.

Fake Medicines are Rare, Fake Nutrient Supplements Abound

How do we know that the nutrient tablet used for studies actually contains the ingredient claimed and tested for? If fake products abound on our store shelves, can we trust the study results, especially if they don’t show any benefit of the nutrient?

Here is a detailed article on this website with dozens of links showing how fake nutrient supplements abound in many countries, including on the shelves of well–known retail chains.

The article talks about many herbal supplements that don’t contain the herbs their labels claim or are laced with dangerous or unapproved chemicals. And this happens right under the nose of the U.S. Food and Drug Administration (US FDA).

The nutrients industry just has to follow an Honour Code, with self-regulation. The pharmaceutical industry has to follow a much stricter regulatory control. Broadly, in the nutrient world, it is the job of the regulator (FDA) to prove that the tablet is not safe. In the pharmaceutical world, it is the job of the manufacturing company to prove that the medicine is safe.

Thus, the chances of a drug tablet not containing the exact amount of its active ingredient are quite less. But that possibility exists in the nutrients world. One more reason to be careful about trusting negative results from a supplements trial!

Side Benefits of Nutrients

In a pharmaceutical clinical trial, a chemical is tested for a single purpose or end-point. Atorvastatin is given to a patient to lower his cholesterol.

Usually, such medicine will not have secondary benefits, such as improving eye health or sharpening mental acuity. Thus, if the trial shows no benefit for cholesterol-lowering, atorvastatin should have no reason to be taken.

This is not the case with nutrients. For example, fish oils are known for their numerous benefits in other body systems. Read a comprehensive article on this website: Omega–3 oils: A complete guide.

Even if omega–3 fish oils were to show no benefit for asthma, they are still worth taking given their benefits for heart, brain, eye, skin, liver, and joint health. After all, the fish oil soft gel does not know that it has to fix your asthma and not your heart.

Bioavailability

Depending on the budget, researchers choose different qualities of nutrients. For example, minerals from organic sources such as food ingredients have two to ten times more bioavailability (absorption) in the intestines than those that are just inorganic salts.

If you are technically minded, organic sources bind minerals with a covalent bond which is much harder to break with stomach acid. On the other hand, inorganic salts have ionic bonds, which are easily broken giving rise to mineral ions in the stomach. Such ions can easily bind to phytates and oxalates in plant food, forming compounds that don’t dissociate in the digestive tract. Such bound minerals get eliminated through stools.

Simply by choosing an inexpensive, inorganic salt of a mineral instead of a plant-based and expensive source of it, you reduce its absorption in the body by five times. How will you get the same result, even if the underlying nutrient such as chromium is the same? That is why we see so much divergence in multiple research trials, though every single published paper talks about ‘chromium does this‘ or ‘chromium does not do this‘. Maybe, one should say ‘plant-sourced chromium does this‘ or ‘Inorganic chromium salt does not do this‘?

Financing

Nutrients are found in nature. So they cannot be patented. That means, even if someone does extensive research on a nutrient, such as turmeric, one cannot have exclusivity in marketing products based on it after conclusively proving its benefits. So there is very little incentive for a company to sponsor such trials.

Keep in mind that nutrient trials need a long time to results. Without adequate funding, especially for buying expensive, organically-sourced (as explained above) nutrients, many trials are kept short and involve cheap sources. No wonder they don’t show great results.

On a personal level, I have noticed that when people who take cheap, run-of-the-mill nutrient supplements switch to high-quality, organic, and usually more expensive supplements, they start seeing much better results. Of course, there is no place for personal observations in science. So take my observations as just that—a single data point.

Selection Bias

I have noticed that people usually have their opinions and choose whatever test reports that support their views as the proof that they are right. It is called Selection Bias—you select the data that suits your view. I hope you don’t fall prey to that weakness.

Clickbait Article Writing

Many people do not read the original research papers written in arcane, medical language. They rely on the summary written by some news source.

Often, such writers use what is called clickbait. They will write a provocative caption that may induce you to click on the website link. After all, ‘Trial fails to show benefit of vitamin D in breast cancer‘ sounds too drab to attract your attention. So how about ‘Consumers are fooled to spend billions of dollars on useless vitamin D supplements‘?

I read an article claiming that ‘there is no benefit of omega–3 fish oil for asthma in adults‘. Actually, the underlying trial on the basis of which this statement was made involved only obese adolescents.

Keep in mind that obesity is an inflammatory condition, and it can reduce omega–3 fish oil’s benefit for asthma. The findings for obese youngsters cannot be extended blindly to all adults. So, this was a clickbait title.

Shallow Reading

Some sites are not intentionally misleading. They may caption the same article as ‘Vitamin D has no benefit in breast cancer‘. But this is also an incorrect caption. The right caption should be ‘A recent trial failed to show benefit of vitamin D in breast cancer‘. After all, there could have been multiple other trials earlier that showed the benefit.

So it is imperative that a reader reads a full article before taking life-changing health decisions. It is not a good idea to just read the title of the article and carry it as a take-home message.

Now, I must confess that even I use simplistic captions. How can you have a long, complex title for a simple, non–medical website article such as this?

References and Additional Resources

A good website should also give links to the original research paper or article. This is to help the readers differentiate between the sensational and the scientific. For example, the present website always gives links to the original articles for any of the major claims made.

For those who are seriously interested, it is always a good idea to read the commentary by the original authors in the discussions section of a paper. Often, the authors mention the limitations of their trial.

Here is an article on this website about the benefits of omega–3 fish oils in asthma and allergies. It analyses the last 30 years of clinical trials on that subject, using all the points discussed above.

My Views

Well, I have given a lot of my views in the article above. But in short, the nutrient research is more likely to progress by consensus observations, and not by clinical trials.

First published on: 13th February 2017
Image credit: Image by Gerd Altmann from Pixabay
Last edited on: 17th July 2022