Berberine, Glycine and the Science of Cautious Health Claims
A long-form podcast episode walking through berberine and colorectal cancer risk, a phase 1 trial on glycine and serine restriction, five nearly free blood and body measures, sunscreen as an anti-aging habit, a lithium re-analysis and vitamin K2 for arterial plaque. The recurring theme is how easily early or single-study findings get overstated online. Each topic ends with what the evidence supports — and where it clearly stops.
Overview
This is the semi-annual Physionic podcast, recorded outdoors, covering six topics in depth. It opens with a follow-up study on berberine, where 600 mg daily (split into two 300 mg doses) reduced precancerous growths in the colon, with the benefit still measurable six years after people stopped taking it. Two mechanisms are explored: berberine reduces glucose uptake by colorectal cancer cells in vitro, and it shifts the gut microbiome away from a bacterium linked to those cancers.
The episode then reviews five nearly free tests suggested by Dr Casey Means, and revises the list — swapping HDL cholesterol for non-HDL cholesterol, with three separate lines of evidence explaining why HDL turned out not to be a useful target. The glycine section covers a phase 1 trial where restricting glycine and serine alongside immunotherapy produced cancer regression, and preclinical tracing work suggesting serine, not glycine, is the actual driver via a carbon pathway feeding DNA building blocks.
Sunscreen is presented as a favourite anti-aging habit, affecting extrinsic aging only, and a critique of the UK Biobank sun exposure preprint argues red light — not UV — is the likely mediator of its observed benefits. Lithium is revisited after a meta-analysis suggested no benefit, with a detailed argument about which studies were included and excluded. Vitamin K2 MK-7 closes the analytical section: across roughly twenty studies most intervention trials showed no effect on arterial or calcified plaque, though long-term observational data still hint at a modest independent benefit. The episode ends on communication itself — how to navigate an information minefield when even careless voices now cite studies.
Key quotes
5It's not just that they don't take up as much glucose, but they also just can't survive — they undergo apoptosis, which means programmed cell death.
This does not mean that you should mega dose berberine. The only tested dose is 600 milligrams split into two doses.
The takeaway here is that glycine is probably not the issue.
Sunscreen specifically affects extrinsic aging. It has no real impact on intrinsic aging.
I don't care about being wrong as long as I end up right.
Key ideas
9The berberine dose that was actually tested
The trial used 600 mg of berberine per day, split into two 300 mg doses — not 600 mg twice. This is the only dose with clinical data behind it for this outcome.
A benefit that outlasted the supplement
After roughly two years of berberine in a randomised setting, participants were followed for six more years without it. Reduced rates of adenomas — precancerous growths in the colon — were still visible at follow-up.
Two mechanisms, direct and indirect
In lab work, berberine cut glucose uptake by colorectal cancer cells in a concentration-dependent way. In mice it also reshaped the microbiome, reducing a bacterium linked to faster cancer development.
Five measures you already get at a check-up
The original list was fasting glucose, triglycerides, HDL cholesterol, waist circumference and blood pressure. The appeal is that these come from a basic panel and a tape measure, with no extra cost.
Why HDL stopped being a target
Three lines of evidence converge: drugs that raised HDL showed no outcome benefit, adjusted observational data weakened the association sharply, and Mendelian randomisation found no lifetime protection from genetically high HDL.
The glycine and serine restriction trial
In a phase 1 trial, people with several cancer types followed a diet nearly free of glycine and serine alongside immunotherapy for three to nine weeks. Most saw some regression, some dramatic — but there was no control group.
Serine looks like the real lever
When preclinical work separated the two amino acids, restricting glycine alone did nothing while restricting serine alone reduced cancer growth. Tracing studies mapped the carbon pathway that feeds DNA building blocks.
Sunscreen and the two kinds of aging
Studies consistently point the same direction: daily sunscreen slows visible skin aging, in one trial to the point of appearing to halt it over five years. It acts on external damage, not on the passage of time.
Reading a meta-analysis by opening the studies
A lithium meta-analysis concluded no benefit, but several qualifying long-term trials showing an effect were missing, and very short trials were included against a disease that develops over decades.
Practical takeaways
6- 1
One good study is exciting, not settled 20:00
Even a well-run trial invites replication before it changes anything. Treat a single striking result as a reason to watch the field, not to rearrange your routine.
- 2
Look at non-HDL cholesterol instead of HDL 44:00
Total cholesterol minus HDL leaves the fractions that track most closely with cardiovascular risk. The figure discussed as a general reference point is under 130 mg/dL.
- 3
The simple numbers worth knowing 1:05:00
Fasting glucose under 100 mg/dL, blood pressure under 120/80, waist under 102 cm for men and 88 cm for women. Waist-to-hip ratio is a reasonable alternative to waist alone.
- 4
A cancer trial does not rewrite a supplement 1:20:00
The trial studied a specific clinical situation with two amino acids restricted at once. Outside that context there was no claim that everyday glycine intake drives cancer risk.
- 5
Daily sunscreen is an unglamorous, high-value habit 1:52:00
SPF 15 on ordinary days and SPF 30 in strong sun, reapplied a couple of times when outdoors, covering face, neck and arms. Sunscreen does not block the red and infrared parts of sunlight.
- 6
Ask which studies were left out 3:00:00
Confident supplement claims often rest on the one trial that found an effect while six others did not. Before acting on a claim, ask what the rest of the literature looks like.
Topics & chapters
15Welcome and the episode roadmap
A recording made outdoors while caring for family, laying out the six topics ahead. The format is semi-annual and deliberately long-form.
Berberine and colorectal cancer risk
A follow-up to a randomised trial: two years on berberine, then six years off, with reduced precancerous growths still detectable at the end.
How berberine appears to work
Reduced glucose uptake by cancer cells in vitro, plus a microbiome shift in mice away from a bacterium associated with colorectal cancer.
Why the lab dose is not a human dose
Concentration-dependent cell death in a dish does not translate into a case for higher intake. Comments about other cancer types are addressed here too.
Five nearly free tests for better health
Reviewing the list of fasting glucose, triglycerides, HDL, waist circumference and blood pressure, and checking each against current reference ranges.
The case against chasing HDL
Drug trials, adjusted observational data and Mendelian randomisation all point the same way. Non-HDL cholesterol replaces HDL on the revised list.
Glycine restriction and cancer regression
The phase 1 trial that produced striking tumour images, and the reasons it does not translate into a recommendation to stop glycine.
Serine, glycine and a carbon pathway
Preclinical tracing work separating the two amino acids and explaining how a carbon atom feeds nucleic acid production in dividing cells.
The favourite anti-aging habit: sunscreen
What daily use appears to do for visible skin aging and skin cancer risk, plus a look at the safety questions people raise about ingredients.
UV, red light and the UK Biobank preprint
A critique arguing that the observed benefits of sun exposure track red and infrared light rather than UV, using tanning bed data to separate the two.
Lithium revisited
A new meta-analysis found no benefit against Alzheimer's. Opening the included studies revealed missing long-term trials and very short trials that should not have counted.
Lithium salts versus orotate
Clinical trials used carbonate, sulfate and gluconate, and several showed an effect. Orotate's advantage rests on a single preclinical study.
Vitamin K2 and arterial plaque
Across roughly twenty studies, most intervention trials showed no effect on arterial or calcified plaque, despite widespread online confidence.
Answering the critiques on vitamin K
Working through comments about K1 versus K2, clotting risk, and why clinic anecdotes with five simultaneous interventions cannot isolate one ingredient.
Navigating an information minefield
Closing reflections on citations that no longer signal quality, an exchange with a listener, and the value of boring sources like textbooks.