Read on Twitter
Since @ElysiumHQ e-mailed me this morning about their new product (Vitamin B megadoses), it's worth noting that in the best studies, and over and over again, long term micronutrient megadosing *fails* to improve health outcomes, and sometimes makes them worse. But why? (...)
In general, hopes of finding a chemical fountain of youth are often based on association studies (people with good outcomes have X in their blood), along with a plausible mechanism for why X is causal—most commonly, that X has "anti-oxidant" or "protective" effects.
Having established this, supplement companies then prepare megadoses of X far in excess of what can be achieved by diet alone—e.g., in Elysium's case, 200x the recommended dietary allowance of B12.
When RCTs finally get around to determining the effects of these treatments—many years after the money has been made—the results are less than disappointing. Vitamin A, D, and E megadoses are associated with significant *increases* in cancer incidence.
Primum non nocere alone should suggest that attempting to alter basic blood chemistry should be done with caution if at all but, legally, gurus are off the hook and can market as they please. (I rather like Elysium's design—it's spare, like a reformation-era protestant church.)
In any case, I can see three reasons why megadoses are associated with nil or negative outcomes.
1. association, not cause.
2. synergistic effects.
3. inhibited hormesis.
1. association, not cause.
2. synergistic effects.
3. inhibited hormesis.
#1 is the easiest. Did you ever notice that micronutrients that improve health outcomes are associated with foods that poor people (in the early 21st Century) *don't* eat?
I mean, how did it just so happen that the cure for aging accidentally ended up in red wine?
I mean, how did it just so happen that the cure for aging accidentally ended up in red wine?
Other things associated with living longer: living near, and eating out of, Whole Foods.
This may seem like a cheap shot, but when SES has such a strong effect, it's hard to control for without an RCT. Usually the way people get around this is looking for a mechanism...
This may seem like a cheap shot, but when SES has such a strong effect, it's hard to control for without an RCT. Usually the way people get around this is looking for a mechanism...
The idea is that if you can find a good reason why the absence of the molecule is bad, then getting someone's blood chemistry to 100x the normal levels will produce less bad, and more good.
Here's the human metabolic network. Not the most complicated one I could find on the web.
Here's the human metabolic network. Not the most complicated one I could find on the web.
The takeaway from this image is that biochemistry is non-linear. There are loops and feedback processes that maintain systems in homeostasis.
Imagine you notice schools that give kids math homework have better outcomes. On the basis of this, you increase math homework 100x.
Imagine you notice schools that give kids math homework have better outcomes. On the basis of this, you increase math homework 100x.
Excellent reasons to believe that math homework is causally relevant to doing well in math. But with homework megadosing either...
1. everyone ignores the extra assignments, and carries on as before. Homeostasis.
2. other subjects are neglected, etc.—unindented side-effects.
1. everyone ignores the extra assignments, and carries on as before. Homeostasis.
2. other subjects are neglected, etc.—unindented side-effects.
There's a more general principle in play here: evolution is an amazing engineer, and we are highly optimized for survival already.
In the immortal words of Regina Spector: I've got a perfect body cause my eyelashes catch my sweat. *Particularly* at the biochemical level.
In the immortal words of Regina Spector: I've got a perfect body cause my eyelashes catch my sweat. *Particularly* at the biochemical level.
Indeed, the only places we should expect to see gains are where we are placing people in (biochemical) conditions that they did *not* evolve to thrive under.
Which brings me to the one pair of interventions that reliably does have effects visible in RCTs...
Which brings me to the one pair of interventions that reliably does have effects visible in RCTs...
Calorie restriction / fasting, and exercise. (It's debated whether the effects of fasting were originally mistaken for the effects of CR.)
In both cases, these interventions appear to work because they *increase* stress on the body temporarily (hormesis).
In both cases, these interventions appear to work because they *increase* stress on the body temporarily (hormesis).
A simple example of this is how the benefits of fasting appear to be counteracted by taking some of the miracle nutrients marketed by Elysium et al.—in this case, Vitamin C and E, which interfered with the cancer-fighting properties of... Oreo cookies. https://www.sciencealert.com/doing-the-5-2-diet-avoid-antioxidants-new-research-suggests
A similar thing is known by those who work: NSAIDs (the anti-inflammatories a million times more powerful than some weird megadose of turmeric extract) are counterproductive for the beneficial effects of exercise.
All of this, by the way, should be distinguished from the therapeutic use of micronutrients in the presence of disease. Elysium's most recent ($480/year) supplement bases its entire pitch around a study of older adults with "mild cognitive impairment".
This is like suggesting that we should chronically dose ourselves with anti-histamines to help breathe better. At best it's unproven; at worst, it's actively harmful (and yes, the dosages of B vitamins that Elysium puts in the bottle are associated with elevated risk of cancer.)
I think this is a great idea, although I've never seen it used as a guiding principle for research in this domain. Most medical research is (naturally) interested in curing diseases, rather than attempting to improve outcomes for the "healthy". https://twitter.com/KateandPie/status/1295073856378605569?s=20
Best bet for a 21st Century miracle drug: a chemical that increases stress, is commonly consumed by the urban poor, and increases the risks of communicable diseases.
I think you're falling for the fallacies above. It's often possible to identify a mechanism. But it's hard to explain why a megadose will have the positive effect unless it's to counteract a "modern" problem (in which case, solve the modern problem). https://twitter.com/ultimape/status/1295074716097576960?s=20
Kind of loving the shade thrown on the "proprietary" blend by Wikipedia (perhaps @gwern had a role in that). https://twitter.com/againstutopia/status/1295073684487581707?s=20
As far as I can tell, a very slight alteration of Michael Pollan's recommendation would overwhelm the effects of any supplement on the market:
1. eat food
2. mostly plants
3. not too much
(the modification to 3: skip meals or fast once in a while.)
1. eat food
2. mostly plants
3. not too much
(the modification to 3: skip meals or fast once in a while.)
