09:46
creatine
Reacting to Dr. Rhonda Patrick's podcast interview with creatine researcher Dr. Darren Candow, Physionic argues that while creatine likely does boost cognition in the metabolically/cognitively stressed and the aging brain, the specific meta-analysis Candow cites is methodologically flawed (it double-counts studies), and that reaching brain-relevant creatine levels probably requires up to 20 grams a day rather than the standard 5-gram muscle dose.
The brain consumes about 20% of daily energy, and creatine's clearest evidence lies in populations who are metabolically or cognitively stressed: sleep deprivation, hypoxia, jetlag, or aging.
Mechanistically, creatine taken up into brain cells like neurons gets phosphorylated into phosphocreatine, a fast-acting reserve energy system that can rapidly regenerate spent ATP when neurons fire and ATP stores drop within seconds.
Dr. Candow's claim (as relayed via Dr. Patrick's podcast): creatine's brain effects are not generally detected in people who are not cognitively stressed, but are detected in people doing cognitively demanding tasks or in older adults (generally over 60).
Physionic pushes back on using the referenced meta-analysis as evidence, calling it 'not a good one' despite otherwise agreeing with the underlying conclusion.
The critique's core: the meta-analysis includes the same underlying studies (e.g., the Alvis study, the McMorris studies) multiple times by treating separate measurements from the same trial as distinct entries, which is an improper inclusion since it's the same outcome measured repeatedly.
This effectively counts the same participant group up to seven times, massively inflating the apparent number of participants and manufacturing an artificial effect in the pooled memory outcome.
In the same flawed analysis, creatine shows no effect on other cognitive outcomes like processing speed, reinforcing suspicion that the memory 'effect' is an artifact of the inflated, duplicated data rather than a genuine finding.
Physionic states this doesn't mean creatine has no cognitive effect: he says he has covered other, corrected analyses and independent randomized controlled trials that still find effects consistent with what Dr. Candow describes.
On dosing, muscle protocols don't require a loading phase, but the best brain studies reportedly use around 20 grams a day for at least a week — creating an apparent conundrum when comparing muscle, bone, and brain dosing needs.
The proposed reason 5 grams is insufficient for the brain: the blood-brain barrier, walled off by astrocytes (glial cells with 'feet' that extend to form the barrier alongside endothelial cells), reportedly has limited or no ability to express the creatine transporter, according to a review Physionic cites — so low blood creatine levels from a 5g dose don't cross into the brain effectively.
The endothelial cells lining the blood side of the barrier do express the creatine transporter and can take up creatine, but the astrocyte layer beyond them remains an obstacle whose bypass mechanism at higher doses is not well understood.
Physionic's own prior analysis, published roughly two years earlier, reached a similar conclusion about needing more than 5 grams for brain effects, though without strong proof at the time.
Final position: there is some evidence, though not strong evidence in Physionic's assessment, that 20 grams of creatine may be more consistently effective than 5 grams for brain-specific benefits; he says he'd like to see more research on this specific question.
A specific, reusable meta-analysis red flag is named: counting the same trial multiple times because it reported multiple measurements of the same outcome inflates apparent sample size and can manufacture statistical significance that vanishes once duplicates are removed.
Physionic notes an unexplained inconsistency in the forest plot itself — he's unsure why the plot appears to move in favor of the control group on the graph even though the paper's text states creatine is the one experiencing the benefit, and speculates (without confirming) that it might be a time-based test where a longer time is the 'worse' outcome.
The same molecule needs fundamentally different dosing protocols depending on target tissue: no loading phase is needed for muscle, but the brain reportedly needs a loading-like ~20g/day for at least a week, because of a tissue-specific bottleneck (the astrocyte-guarded blood-brain barrier) that doesn't apply to muscle uptake.
Physionic explicitly separates 'is the effect real' from 'is this piece of evidence good' — he maintains the cognitive effect likely exists (citing other corrected analyses and RCTs) even while discarding this specific meta-analysis as poor evidence for it, modeling a distinction between source-critique and claim-rejection.
The mechanism by which higher blood creatine concentrations manage to cross the astrocyte barrier at all is described as an open question — Physionic says he looked in the literature and speculates the barrier may not be fully impermeable at high concentrations, but states plainly 'we don't have a complete understanding yet.'
«Now, our brain is small from stature, but it uses 20% of our daily energy.»
— 00:34
«That's where creatine seems to have the best lines of evidence.»
— 00:51
«the phosphocreatine is a reserve energy system inside your cells.»
— 01:59
«Dr. Dr. Kandow makes an excellent point that creatine's effectiveness in the brain is not generally detected in people who are not cognitively stressed, but it is detected in people who are doing cognitively stressing tests or they're older.»
— 02:21
«This meta analysis is not a good one.»
— 02:45
«they made a massive mistake by including the same study over and over again.»
— 03:27
«The best studies for the brain look at 20 grams a day for at least a week.»
— 06:20
«there's some evidence, although I still wouldn't call it strong evidence, that 20 gram of creatine might be better than five grams of creatine.»
— 09:15
Reception
Engaged but cautiously skeptical audience; genuine interest in creatine's benefits undermined by persistent concerns about hair loss, kidney safety, and ineffectiveness for some users.
The video functions as a secondary-source check on a podcast claim, carefully separating the underlying physiological effect (which it treats as plausible) from the quality of the specific meta-analysis used to support it, backing that separation with a concrete citation-duplication critique rather than generic skepticism. Its headline dosing conclusion — that 20g may outperform 5g for brain effects — is explicitly flagged by the source itself as suggestive rather than strong evidence.
09:46