In his book Bad Pharma, Ben Goldacre talks about how placebos often do not make good controls (he also talks about this briefly in one of his TED Talks starting at 7:50). The reason is that placebos are equivalent to no treatment, which isn’t helpful when the condition we’re exploring already has an established treatment. What we want to know, he argues, is not whether an intervention does anything at all, but rather whether it’s better than what we already do.
The same problem of improper controls exists in the nutrition industry, and is responsible for a plethora of bad science. Do branched chain amino acids really provide any benefits to an athlete, or does taking the same amount of a whole protein do the same thing? Are specialized carbohydrate supplements like Vitargo actually more effective, or does any high-molecular weight glucose-based carbohydrate provide the same results? We don’t really know (though we can suspect the answer) because trials on these products use no controls, placebos for controls, or improper controls that are not truly comparable.
On this site, the most frequently encountered culprit of improper controlling is dietary collagen protein, such as collagen hydrolysate; in every trial I’ve seen, a placebo is used for the control rather than a complete protein supplement. This is a problem because collagen is a protein—a poor-quality protein, yes, but a protein nonetheless. And as a protein, we know that dietary collagen will have at least some effect on our whole-body nitrogen kinetics, including our tendons and ligaments. Thus, when trials on collagen hydrolysate appear to show benefits, it’s not surprising.
The problem is that we already know of another nutrient that affects our whole-body nitrogen kinetics, and does so more effectively for cheaper: high-quality dietary protein. So the question isn’t “Does collagen hydrolysate positively influence tendons and ligaments [or whatever]“, it’s “Does collagen hydrolysate positively influence tendons and ligaments [or whatever] to a greater extent than a comparable amount of whey protein [or an equivalent]“. The answer to the latter question is far less clear, but biochemically there’s little reason to suspect it would be.
If I were to design a trial on dietary collagen, I would want to see a few things:
- High-Quality Protein as a Control: Any trial that attempts to demonstrate an effect of collagen hydrolysate should compare that effect to a whole protein. We know that whole proteins of high-quality are effective for collagen growth and repair, or at least that inadequate protein impairs collagen growth and repair.
- Controlled Protein Intake in General: Protein supplements of any kind will only be effective relative to the amount of protein in a diet; a person who consumes 60 grams of protein per day on average will get much more benefit from an extra 20 grams of protein than someone who consumes 120 grams of protein per day. All participants included in a study on protein should have similar daily protein intakes to mitigate this.
- A Third Group Using an Amino Acid-Fortified Whole Protein (Optional): Beyond the collagen and control group, it would be useful to see how a third group that used a whole protein fortified with collagen-specific amino acids fared. The specific goal here would be to buoy any amino acid the whole protein might be slightly low in relative to collagen and boost it slightly. This would give some insight into the relative importance of certain amino acids in the creation or limitation of collagen synthesis.
Obviously, this makes a trial considerably harder to run than just comparing dietary collagen to a placebo—but it would also give us much better insight. As it stands now, I can see no plausible reason for why dietary collagen should be more effective than another whole protein. It doesn’t mean there isn’t a reason it is, but without good data to back the hypothesis, it doesn’t hold water, and I’ll continue to recommend people avoid supplemental dietary collagen and stick with high-quality proteins.