How Twelve People Made a Very Big Claim

LONGEVITY LATEST ISSUE 20 COMPANION · 22 JULY 2026

LONGEVITY LATEST · DEEP DIVE

How Twelve People Made a Very Big Claim

An 83% drop in inflammation from twelve participants isn't stronger evidence than a 5% drop from five thousand — and the reason why will change how you read every supplement headline.

By Christian Thomsen · Companion to Issue 20 · 22 July 2026 · ~6-minute read

You've read the newsletter, so you know where GlyNAC landed: C+ for the hallmarks claim, C for the deficiency case, and a Hype Check on what a patent adds to the price of two commodity amino acids. What the newsletter didn't have room for is the part that matters most, and it isn't really about GlyNAC.

It's this. When a very small trial reports a very large effect, the size of that effect is a reason to distrust the number — not to be impressed by it. Almost nobody reading supplement coverage knows this, including some of the people writing it. Let me show you the machinery.

The filter that guarantees inflation

Statistical significance isn't a measure of how true something is. It's a threshold a result must clear to get called a finding — and the threshold effectively moves with sample size. With five thousand people, a genuine 5% effect clears it comfortably. With twelve people per arm, noise is enormous and that same effect will almost never clear it. It comes out as a non-result and goes in a drawer.

So ask the question backwards. Given that a twelve-person study did report a significant result, what must have happened? Either the true effect is gigantic, or that sample landed on a favourable roll of the dice. And since gigantic true effects are rare in nutrition — nutrients aren't drugs, and they're being added to bodies that already contain them — the odds tilt hard toward the dice.

This is the significance filter, and it doesn't merely add uncertainty. It biases the reported number upward in a predictable direction. Andrew Gelman and John Carlin formalised it in 2014 as a "Type M error" — an error of magnitude. In an underpowered study, any result that survives to publication is systematically exaggerated, often twofold or more, even when the underlying effect is perfectly real.

One precision, because it matters: this says nothing about whether an effect exists, only about how big it is. Direction, plausibly right. Magnitude, almost certainly inflated. And the inflation isn't fraud or sloppiness — it's arithmetic, and it happens to careful researchers running careful experiments.

Now put GlyNAC's numbers beside that. IL-6 down 83%. Insulin resistance down 68%. Gait speed up 19%. Striking for a potent pharmaceutical — and these are two amino acids you can buy by the kilogram, in twelve people, reported as within-group changes.

So the trial may be immaculate, the direction may be right, and the magnitudes still almost certainly won't survive. Which raises the obvious question: has anybody tried? Somebody has. It just isn't the study you've heard about.

Why the second lab matters more than the first

First, though, notice where all the good news comes from. Nearly every human GlyNAC result traces to one research group — Rajagopal Sekhar's team at Baylor College of Medicine. That group has been strikingly consistent, which sounds reassuring and isn't, quite. A single lab shares its assay methods, its participant pool, its measurement quirks and its expectations. Consistency within one lab tells you the lab is consistent, not that the effect exists outside it.

Hence the rule: one lab is a hypothesis, three labs is a finding.

It's why the National Institute on Aging built the Interventions Testing Program the way it did — three independent sites, shared protocol, genetically varied mice — after the field accumulated a graveyard of compounds that worked beautifully in one lab and nowhere else. And on incentives, without being conspiratorial: Baylor holds the patent and licensed it to Nestlé Health Science. That doesn't make the research dishonest, and I've seen nothing suggesting it is. It simply leaves one question open, and that question is doing more work here than any p-value: has anyone without a stake found this?

As it happens, somebody with a very large stake went looking.

The replication nobody quotes

Nestlé Health Science licensed the patent. Nestlé Research then ran the biggest test of it anyone has run — and here is the part I didn't expect when I started reading: they published the disappointing answer themselves.

In 2022, in Frontiers in Aging: 114 healthy older volunteers, randomised, placebo-controlled, three doses of glycine and NAC — 2.4, 4.8 and 7.2 g/day. Primary endpoint, total glutathione. Result: placebo 903.5 mg/L, top dose 959.6 mg/L, p = 0.278. A miss.

Three things deserve more attention than they get.

It was short. Two weeks, against sixteen in Sekhar's trial. If GlyNAC's benefits accrue slowly, a fortnight isn't a fair test — a real limitation, and I don't want to lean on this trial harder than its design allows.

It found something anyway, and the something is weaker than it looks. In participants who began with low glutathione and high oxidative stress, the higher doses did raise glutathione. But that analysis was post-hoc: the subgroup was identified after the results were in, the two top doses pooled, the change measured within those participants rather than against placebo. Each choice multiplies the ways a null can be made to look like a signal. It's suggestive. It isn't a finding — and it's why the newsletter graded that claim C rather than B.

It came from an interested party. The awkward one — and it sits in tension with the first, so let me name that rather than paper over it. Commercially motivated research bends toward positive findings, so a null from the patent-holder's own company is harder to dismiss than a null from a neutral one. But that argument only carries as far as the design does, and a two-week study isn't a decisive test of a sixteen-week claim. Both are true. What the trial establishes isn't that GlyNAC fails; it's that the effect isn't large or fast enough to surface in the biggest sample anyone has assembled, in a fortnight, in ordinary healthy older adults.

It remains the largest human GlyNAC trial published to date. It is also absent from every piece of marketing I could find — which tells you something about the gap between what a company's scientists will publish and what its marketing department will quote.

What would actually settle this

Here's what would move GlyNAC from C+ to B in this newsletter, roughly in order of weight:

An independent group replicating any of it. Not the same lab with more people — a different lab, different assays, ideally no patent exposure. Even half the reported effect would be real news.

A sample in the hundreds, run for months not weeks. Sixteen weeks was a reasonable duration; twenty-four participants was not a reasonable number. Both need to be true at once, and no trial has managed that.

An outcome that isn't a blood marker. Gait speed is the most interesting thing in the 2023 trial precisely because it's functional. Scale that up — falls, fractures, admissions — and you'd have something no marker panel can give you.

A comparison against the ingredients bought separately. Nobody has tested branded GlyNAC against plain glycine plus plain NAC. Until someone does, the patented pairing is a commercial claim wearing a scientific one's clothes.

If two of those land, I'll re-grade in public and say so. If none do within a couple of years, that silence is its own answer — the field is littered with promising small results nobody could be bothered to replicate, because replication doesn't sell anything.

What this means for you

The practical advice is unchanged and deliberately dull: eat the glycine, skip the subscription, and if you want to experiment, buy the two ingredients separately and measure something.

The newsletter leaves you with three rules — big study small effect, primary endpoint first, one lab is a hypothesis. Here's the uncomfortable thing about all three: they're easy to apply to a supplement you never wanted, and very hard to apply to one you already bought. That asymmetry is where most otherwise-sceptical people come unstuck, and it's worth more vigilance than any individual study.

Applied consistently, those rules would have saved a lot of people a lot of money on resveratrol, on antioxidant megadoses, and on most of what's been sold as an ageing breakthrough this century.

One thing before you go

The newsletter poll asks you to name your private threshold for proof — twenty-four people, a second lab, a hard outcome, or nothing at all. Most people have never said it out loud, which is exactly why it's so easy to quietly lower it when a headline flatters something you already wanted to buy.

So pick your line before the next study lands, not after. Then reply and tell me whether reading this moved it. I read every one — and the answers decide how hard Issue 21 comes down on the muscle-supplement shelf.

This is the Issue 20 Deep Dive. Longevity Latest runs one every week — the long version of the argument the newsletter only has room to start. Next week we leave the supplement shelf for muscle: what protein dose actually does in older adults, how resistance training bends mortality curves, and why the tissue nobody sells you is the one that predicts the most.

Sources and further reading

1. Kumar P, Liu C, Suliburk J, et al. Supplementing glycine and N-acetylcysteine (GlyNAC) in older adults improves glutathione deficiency, oxidative stress, mitochondrial dysfunction, inflammation, physical function, and aging hallmarks: a randomized clinical trial. J Gerontol A Biol Sci Med Sci. 2023;78(1):75–89. PMID: 35975308.

2. Kumar P, Liu C, Hsu JW, Chacko S, Minard C, Jahoor F, Sekhar RV. Glycine and N-acetylcysteine (GlyNAC) supplementation in older adults improves glutathione deficiency, oxidative stress, mitochondrial dysfunction, inflammation, insulin resistance, endothelial dysfunction, genotoxicity, muscle strength, and cognition: results of a pilot clinical trial. Clin Transl Med. 2021;11(3):e372. PMID: 33783984.

3. Lizzo G, Migliavacca E, Lamers D, et al. A randomized controlled clinical trial in healthy older adults to determine efficacy of glycine and N-acetylcysteine supplementation on glutathione redox status and oxidative damage. Front Aging. 2022;3:852569. PMID: 35821844 — largest human GlyNAC trial published to date; primary endpoint not met; subgroup finding post-hoc.

4. Kumar P, Osahon OW, Sekhar RV. GlyNAC supplementation in mice increases length of life by correcting glutathione deficiency, oxidative stress, mitochondrial dysfunction, abnormalities in mitophagy and nutrient sensing, and genomic damage. Nutrients. 2022;14(5):1114. PMID: 35268089.

5. Miller RA, Harrison DE, Astle CM, et al. Glycine supplementation extends lifespan of male and female mice. Aging Cell. 2019;18(3):e12953. PMID: 30916479 — NIA Interventions Testing Program, three independent sites.

6. Gelman A, Carlin J. Beyond power calculations: assessing Type S (sign) and Type M (magnitude) errors. Perspect Psychol Sci. 2014;9(6):641–651 — the formal account of effect-size inflation in underpowered studies.

7. Button KS, Ioannidis JPA, Mokrysz C, et al. Power failure: why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci. 2013;14(5):365–376. PMID: 23571845.

8. Ioannidis JPA. Why most published research findings are false. PLoS Med. 2005;2(8):e124. PMID: 16060722.

9. Paulsen G, Cumming KT, Holden G, et al. Vitamin C and E supplementation hampers cellular adaptation to endurance training in humans: a double-blind, randomised, controlled trial. J Physiol. 2014;592(8):1887–1901. PMID: 24492839 — markers of mitochondrial biogenesis blunted; performance outcomes unchanged.

10. Sayin VI, Ibrahim MX, Larsson E, et al. Antioxidants accelerate lung cancer progression in mice. Sci Transl Med. 2014;6(221):221ra15. PMID: 24477002.

11. Le Gal K, Ibrahim MX, Wiel C, et al. Antioxidants can increase melanoma metastasis in mice. Sci Transl Med. 2015;7(308):308re8. PMID: 26446958.

© 2026 FrontWave Media Ltd · Longevity Latest

This article provides general educational information and is not medical advice. N-acetylcysteine can interact with medications. If you are pregnant or breastfeeding, taking prescribed medication, have asthma, have a history of cancer, or have a diagnosed condition, consult your physician before use and do not stop prescribed treatment in favour of a supplement.

© 2026 FrontWave Media Ltd · Longevity Latest 1