The Supplement Dose Gap Report 2026: Trial Doses vs. US Label Doses

Last updated: September 6, 2026 · Published by Eternal Elixir

Published: September 6, 2026 · By Eternal Elixir Editorial · Original research · 12 compounds · 39 products · 35 cited trials

Ask the internet how much NMN, berberine or TUDCA to take and you get a number. Ask where that number came from and the trail usually ends at another supplement brand. So we went and checked: for twelve of the most-searched longevity compounds sold in the United States, what daily dose did published human trials actually use — and what dose does the label on the bottle actually deliver?

We built the dataset from scratch. Every trial dose in this report was read verbatim out of a PubMed abstract and is attached to the PMID it came from; nothing is inferred from a review, a brand page or a round number in circulation. Every label dose was read live from a US storefront on September 6, 2026. The two sets are then placed side by side.

The result is a measurable gap. Only 4 of the 12 compounds have a typical US serving that falls inside the range of doses used in the human trials we retrieved. Three sit below that range, four sit above the highest dose ever tested, and one has no published human dose range to be measured against at all.

The full dataset: trial dose vs. label dose, 12 compounds

Each row pairs the range of daily doses we retrieved from published human trials with the label serving doses we read across 4 US retailers. “Oral RCTs” counts PubMed records tagged as randomized controlled trials in humans that also mention oral supplementation, queried live on September 6, 2026. Rows are ordered by the size of the gap.

Tongkat ali returns no search-volume figure for its head dosage term in the dataset we queried, so its demand cell is left empty rather than estimated. Where only one trial dose was retrieved, the “range” is that single value and the comparison should be read accordingly.

How many US longevity supplements are dosed inside the human-trial range?

Four of twelve: NMN, spermidine, glutathione and Panax ginseng. Each has a median panel serving that sits inside the band of doses used in the trials we retrieved.

The other eight split three ways. TUDCA, nicotinamide riboside and berberine come in below the trial doses at a typical single serving. Tongkat ali, shilajit, pterostilbene and resveratrol come in above the highest dose tested, though for the first two extract-ratio labelling makes the comparison much less clean. Calcium alpha-ketoglutarate has no published human dose range to be measured against.

Which compound has the biggest dose gap?

TUDCA. It is the clearest case in the dataset because the arithmetic is simple: the only published randomized human trial we could retrieve administered 2 g (2,000 mg) per day for 16 weeks, in people with progressive multiple sclerosis. US labels in our panel deliver between 250 mg and 600 mg per serving. The panel median of 500 mg is a quarter of the trial dose, and reaching 2,000 mg from a 500 mg capsule takes four capsules a day.

TUDCA also carries the highest demand-to-evidence ratio in the set: about 600 US searches a month asking how much to take, against just 3 indexed oral-supplementation randomized trials. More people are asking the question each month than there are trials to answer it, by two orders of magnitude.

Is the most-searched compound dosed correctly?

No, and this is the finding we found most striking. Berberine draws 5,800 US dosage searches a month across its two main phrasings — more than twice any other compound in this report, and more than NMN, TUDCA, spermidine, resveratrol and nicotinamide riboside combined. It also has a substantial literature by this category’s standards: 109 randomized human trials.

Those trials used 600 mg/day at the low end (0.3 g twice daily) and up to 2,000 mg/day. The median US berberine capsule in our panel contains 500 mg. A single capsule therefore sits below every dose we retrieved; a two-capsule serving clears the bottom of the range comfortably. Products sold as multi-capsule servings, such as the 1,200 mg Nutricost formulation, land inside it as sold.

Are spermidine supplements underdosed?

Less clearly than we expected, and the honest answer is that nobody yet knows. This is the compound where the literature itself is unsettled rather than the market.

The four human trials we retrieved used 0.9 mg/day, 6 mg/day, 15 mg/day and 40 mg/day — a 44-fold spread. A 2023 report administering 15 mg/day to healthy volunteers concluded that supplements below 15 mg/day are unlikely to exert short-term effects. A 2026 trial in healthy older adults reported effects on immune cell senescence and vaccine responses at 6 mg/day. A 2024 study found that even 40 mg/day had minimal effect on circulating polyamines. These are not easy to reconcile.

Every spermidine product in our panel — three products from two US brands, including our own — declares 10 mg of spermidine trihydrochloride per serving. That is inside the retrieved range, but it sits in exactly the zone the literature disagrees about. Anyone quoting a confident spermidine dose is choosing a side in an open question.

Which compound has the weakest human dose evidence?

Calcium alpha-ketoglutarate. This is worth stating plainly because Ca-AKG is sold widely as a longevity ingredient, including by us. PubMed returns 34 randomized human trials for alpha-ketoglutarate, which sounds reassuring until you read them. The bulk are clinical-nutrition studies of ornithine alpha-ketoglutarate given at gram-scale doses to surgical, burn or critically ill patients, or arginine alpha-ketoglutarate in sports and orthopaedic contexts. Those are different molecules answering different questions at a dose scale far above the 500 mg calcium salt on the shelf.

The 1,000 mg/day figure that circulates for Ca-AKG traces to a non-randomized human study of biological-age markers, not to a randomized controlled trial. A 2022 review in Trends in Endocrinology & Metabolism surveying alpha-ketoglutarate as a human dietary supplement reaches the same place: the human dosing question is open. Anyone quoting a confident Ca-AKG dose is quoting a convention, not a trial.

Are any longevity supplements dosed above what trials tested?

Four of the twelve, and pterostilbene is the cleanest example. The published human trials we retrieved used 10 mg/day and 100 mg/day in a 2025 pilot, and 200 mg/day (100 mg twice daily) in a short-term safety evaluation of a 90% pterostilbene extract. Single capsules sold in the US reach 500 mg — two and a half times the highest dose in that published set, including in our own 500 mg product. Resveratrol is a milder version of the same pattern: trials clustered at 80–500 mg/day, US labels at 600–700 mg.

Shilajit and tongkat ali also sit above their trial figures on paper, but here the comparison is genuinely muddier and we would not lean on it. Trials tend to report a standardized extract dose; labels often report a concentrated extract ratio such as 20:1 or 200:1. A 2,000 mg 200:1 tongkat ali serving is not directly comparable to the 600 mg used in a trial, because the two numbers are not measuring the same thing.

Being dosed above the tested range is not the same as being unsafe. It means the dose has not been formally studied at that level, which is a gap in knowledge rather than a finding of harm.

Where does public demand most outrun published evidence?

We divided monthly US dosage-intent search volume by the number of indexed oral-supplementation randomized trials to get a demand-to-evidence ratio. A high number means a lot of people are asking a question the literature has barely addressed.

TUDCA (200 searches per trial) and berberine (171) top the list, and both are compounds our data flags as under-dosed on the shelf. Shilajit (80) and NMN (74) follow. At the other end, Panax ginseng and resveratrol have far more trials than questions — decades of literature and almost no one asking.

How has US interest in these compounds changed since 2021?

We pulled five years of US Google Trends interest for four of these compounds, indexed on a shared scale so they can be compared with each other. The 2026 figures cover January through early September only.

Berberine has grown roughly nine-fold on this index since 2021 and is by a wide margin the most-searched compound in the group — which makes the gap between its label dose and its trial dose the most consequential finding here. NMN has grown about seven-fold, with most of that jump landing in 2026. Spermidine has grown from a rounding error to a small but real signal. Calcium alpha-ketoglutarate was queried alongside these four but stayed below the index threshold for the entire five-year window: it is a compound with far more shelf presence than search presence.

How many servings would it take to reach the trial dose?

This is the practical version of the same question. Using the lowest daily dose retrieved from a published human trial, and the label servings in our panel, here is the multiplier.

For 7 of the eleven compounds with a retrieved trial dose, one serving of every product in our panel already clears the lowest dose used in a trial. The exceptions are TUDCA, NR, Berberine and Glutathione. TUDCA is the starkest: reaching 2,000 mg from a 500 mg capsule takes four capsules a day. Nicotinamide riboside needs two servings to reach 1,000 mg. Berberine and glutathione need one serving of most products in the panel but more of the smallest ones — three servings of a 100 mg liquid glutathione, for instance. These are the compounds where a label that looks adequate quietly is not.

Compound-by-compound: the trials behind each dose

Every dose figure above traces to a specific published study. Here they are, with PubMed identifiers so you can read them yourself. Where we retrieved only one or two trial doses for a compound, that is stated rather than smoothed over.

TUDCA

Retrieved trial dose: 2,000 mg/day (one retrieved trial)  ·  PubMed human trials: 27 (21 randomized, 3 oral supplementation)  ·  Panel: 4 products from 3 brands, median 500 mg  ·  Verdict: Below (0.25x lowest trial dose)

• 2,000 mg/day — PMID 39447576: Med 2025 – 2 g/day for 16 weeks in progressive multiple sclerosis

Panel products read September 6, 2026: Eternal Elixir TUDCA (500 mg) · Nutricost Tudca Capsules (250 mg) · Nutricost Tudca Powder (500 mg) · Codeage Liposomal TUDCA+ (600 mg).

See Eternal Elixir TUDCA 500mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

NR

Retrieved trial dose: 1,000–2,000 mg/day  ·  PubMed human trials: 39 (31 randomized, 25 oral supplementation)  ·  Panel: 3 products from 3 brands, median 500 mg  ·  Verdict: Below (0.50x lowest trial dose)

• 1,000 mg/day — PMID 32320006: 2020 RCT – 1,000 mg/day for 6 weeks
• 1,000 mg/day — PMID 37159264: 2023 RCT – 1,000 mg/day
• 2,000 mg/day — PMID 38016950: GeroScience 2024 – doses above 2,000 mg/day have not been tested in humans

Panel products read September 6, 2026: Eternal Elixir Nicotinamide Riboside Capsules (500 mg) · Nutricost Nicotinamide Riboside + Resveratrol (500 mg) · Codeage Liposomal Nicotinamide Riboside+ (500 mg).

See Eternal Elixir Nicotinamide Riboside 500mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Berberine

Retrieved trial dose: 600–2,000 mg/day  ·  PubMed human trials: 133 (109 randomized, 34 oral supplementation)  ·  Panel: 5 products from 4 brands, median 500 mg  ·  Verdict: Below (0.83x lowest trial dose)

• 600 mg/day — PMID 31926918: Lancet Gastroenterol Hepatol 2020 – 0.3 g twice daily
• 1,000 mg/day — PMID 18397984: 2008 RCT – 1.0 g daily for 3 months
• 1,000 mg/day — PMID 40029660: 2025 RCT – 500 mg twice daily
• 2,000 mg/day — PMID 40029660: 2025 RCT – 1,000 mg twice daily

Panel products read September 6, 2026: Eternal Elixir Berberine Capsules (500 mg) · Nutricost Berberine Capsules (600 mg) · Nutricost Berberine HCl + Ceylon Cinnamon (1,200 mg) · Double Wood Berberine with Ceylon Cinnamon (500 mg) · Codeage Berberine Phytosome+ (500 mg).

See Eternal Elixir Berberine HCl 500mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Ca-AKG

Retrieved trial dose: not established in the randomized human literature  ·  PubMed human trials: 51 (34 randomized, 13 oral supplementation)  ·  Panel: 1 product from 1 brand, median 500 mg  ·  Verdict: No established range

• PMID 34952764 — Trends Endocrinol Metab 2022 – review of alpha-ketoglutarate as a human dietary supplement
• PMID 23043451 — Curr Med Res Opin 2012 – arginine alpha-ketoglutarate, a different salt and context

Panel products read September 6, 2026: Eternal Elixir Ca-AKG Capsules (500 mg).

See Eternal Elixir Ca-AKG 500mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Tongkat Ali

Retrieved trial dose: 600 mg/day (one retrieved trial)  ·  PubMed human trials: 6 (6 randomized, 3 oral supplementation)  ·  Panel: 2 products from 2 brands, median 1,500 mg  ·  Verdict: Above (2.50x highest trial dose)

• 600 mg/day — PMID 33559971: 2021 – 600 mg/day for two weeks

Panel products read September 6, 2026: Eternal Elixir Tongkat Ali Capsules (2,000 mg) · Nutricost Tongkat Ali Capsules (1,000 mg).

See Eternal Elixir Tongkat Ali 200:1 — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Shilajit

Retrieved trial dose: 250–500 mg/day  ·  PubMed human trials: 9 (7 randomized, 6 oral supplementation)  ·  Panel: 3 products from 3 brands, median 1,000 mg  ·  Verdict: Above (2.00x highest trial dose)

• 250 mg/day — PMID 35933897: Phytomedicine 2022 – 125 mg twice daily
• 500 mg/day — PMID 35933897: Phytomedicine 2022 – 250 mg twice daily

Panel products read September 6, 2026: Eternal Elixir Shilajit Capsules (1,000 mg) · Nutricost Shilajit Capsules (1,000 mg) · Codeage Liposomal Himalayan Shilajit+ (500 mg).

See Eternal Elixir Shilajit 20:1 Extract — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Pterostilbene

Retrieved trial dose: 10–200 mg/day  ·  PubMed human trials: 7 (7 randomized, 4 oral supplementation)  ·  Panel: 2 products from 2 brands, median 300 mg  ·  Verdict: Above (1.50x highest trial dose)

• 10 mg/day — PMID 40024751: 2025 pilot – 10 mg/day for 12 weeks
• 100 mg/day — PMID 40024751: 2025 pilot – 100 mg/day for 12 weeks
• 200 mg/day — PMID 37671486: 2023 safety evaluation – 200 mg/day (100 mg twice daily), 90% pterostilbene extract

Panel products read September 6, 2026: Eternal Elixir Trans-Pterostilbene Capsules (500 mg) · Nutricost Pterostilbene Capsules (100 mg).

See Eternal Elixir Trans-Pterostilbene 500mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Resveratrol

Retrieved trial dose: 80–500 mg/day  ·  PubMed human trials: 305 (262 randomized, 146 oral supplementation)  ·  Panel: 3 products from 2 brands, median 600 mg  ·  Verdict: Above (1.20x highest trial dose)

• 80 mg/day — PMID 27194304: 2016 RCT – 80 mg/day in combination with EGCG
• 150 mg/day — PMID 31996311: 2020 RCT – 150 mg/day for 4 weeks
• 150 mg/day — PMID 32564438: J Bone Miner Res 2020 – 75 mg twice daily
• 400 mg/day — PMID 37843843: RCT – 400 mg/day for 6 weeks, endothelial function
• 500 mg/day — PMID 27664491: RCT – 500 mg/day
• 500 mg/day — PMID 39565038: RCT – 500 mg/day for 8 weeks

Panel products read September 6, 2026: Eternal Elixir Resveratrol Capsules (600 mg) · Nutricost Resveratrol Capsules (700 mg) · Nutricost Organic Resveratrol (600 mg).

See Eternal Elixir Trans-Resveratrol 600mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

NMN

Retrieved trial dose: 250–1,200 mg/day  ·  PubMed human trials: 23 (17 randomized, 14 oral supplementation)  ·  Panel: 6 products from 3 brands, median 500 mg  ·  Verdict: Within range

• 250 mg/day — PMID 36797393: Sci Rep 2023 – long-term 250 mg/day, well tolerated
• 250 mg/day — PMID 42082179: 2026 RCT – 250 mg/day for 24 weeks
• 300 mg/day — PMID 34238308: JISSN 2021 – dose-ranging arm, 300 mg/day
• 600 mg/day — PMID 36482258: GeroScience 2023 – NAD and physical performance peak at 600 mg/day
• 1,200 mg/day — PMID 34238308: JISSN 2021 – dose-ranging arm, 1,200 mg/day
• 1,200 mg/day — PMID 41705654: 2025 crossover – 1,200 mg/day for 7 days

Panel products read September 6, 2026: Eternal Elixir NMN Capsules (500 mg) · Eternal Elixir NMN 1200 Advanced Longevity Complex (500 mg) · Nutricost NMN (500 mg) · Codeage Liposomal NMN 1000 Capsules (1,000 mg) · Codeage Amen NMN (500 mg) · Codeage Liposomal NMN Platinum (500 mg).

See Eternal Elixir NMN 500mg Capsules — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Spermidine

Retrieved trial dose: 0.9–40 mg/day  ·  PubMed human trials: 65 (41 randomized, 18 oral supplementation)  ·  Panel: 3 products from 2 brands, median 10 mg  ·  Verdict: Within range

• 0.9 mg/day — PMID 35616942: SmartAge RCT – 0.9 mg spermidine/day from wheat germ extract
• 6 mg/day — PMID 42169618: Aging Cell 2026 – 6 mg/day for 13 weeks, well tolerated
• 15 mg/day — PMID 37111071: 2023 – 15 mg/day in 12 healthy volunteers; authors state doses below 15 mg/day are unlikely to exert short-term effects
• 40 mg/day — PMID 39405978: 2024 – 40 mg/day purified spermidine, minimal effect on circulating polyamines

Panel products read September 6, 2026: Eternal Elixir Spermidine Capsules (10 mg) · Codeage Amen Spermidine (10 mg) · Codeage Liposomal Spermidine NAD+ (10 mg).

See Eternal Elixir Spermidine 10mg — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Glutathione

Retrieved trial dose: 250–1,000 mg/day  ·  PubMed human trials: 2,529 (1,683 randomized, 706 oral supplementation)  ·  Panel: 4 products from 3 brands, median 750 mg  ·  Verdict: Within range

• 250 mg/day — PMID 24791752: Eur J Nutr 2015 – 250 mg/day oral GSH
• 500 mg/day — PMID 20524875: J Dermatolog Treat 2012 – 500 mg/day in two divided doses
• 1,000 mg/day — PMID 24791752: Eur J Nutr 2015 – 1,000 mg/day oral GSH

Panel products read September 6, 2026: Eternal Elixir L-Glutathione Pills (2,000 mg) · Nutricost Glutathione Capsules (500 mg) · Codeage Liposomal Glutathione+ (1,000 mg) · Codeage Nanofood Liquid Liposomal Glutathione (100 mg).

See Eternal Elixir L-Glutathione — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Panax ginseng

Retrieved trial dose: 100–2,000 mg/day  ·  PubMed human trials: 109 (91 randomized, 31 oral supplementation)  ·  Panel: 3 products from 2 brands, median 1,000 mg  ·  Verdict: Within range

• 100 mg/day — PMID 23567596: 2013 RCT – 100 mg/day
• 200 mg/day — PMID 40289951: RCT – 200 mg/day including 24 mg ginsenosides
• 250 mg/day — PMID 41838185: 2026 RCT – 250 mg/day for 4 weeks
• 1,000 mg/day — PMID 37218691: RCT – 1 g/day with standard chemotherapy
• 2,000 mg/day — PMID 22909282: RCT – 2 g/day for 4 weeks

Panel products read September 6, 2026: Eternal Elixir Korean Red Panax Ginseng (1,600 mg) · Nutricost Panax Ginseng Capsules (1,000 mg) · Nutricost Korean Red Ginseng (1,000 mg).

See Eternal Elixir Korean Red Panax Ginseng — our own product for this compound, included in the panel above and measured on the same basis as everything else.

Methodology

Compound selection. We covered twelve compounds that are sold as single-active oral supplements in the United States and are searched for by dose. The list was fixed before any data was collected.

Trial doses. For each compound we queried PubMed through the NCBI E-utilities API on September 6, 2026, filtering to human records tagged as clinical trials or randomized controlled trials, then to those additionally mentioning oral supplementation. We retrieved the abstracts of the highest-relevance records and extracted every stated daily dose verbatim. A dose appears in this report only if we read it in a retrieved abstract, and each one is listed above against the PMID it came from. 35 PubMed records are cited in total.

This means the ranges here describe the doses in the trials we retrieved, not an exhaustive census of every trial ever run. Where a compound has a large literature, such as resveratrol, other trials will have used doses outside our range. Where the retrieved set is one or two studies, such as TUDCA and tongkat ali, the “range” is thin by construction and we have flagged it in the tables. We preferred a narrow, fully-sourced set to a wider one padded with numbers we could not point at.

Label doses. We read 39 products from 4 US retailers — Eternal Elixir, Nutricost, Codeage and Double Wood — from their own live public product data on September 6, 2026, recording the milligrams of the named compound per labeled serving. Where a brand publishes a serving of more than one capsule, we used the declared serving, not the per-capsule figure. Products whose public data did not state a milligram figure for the named compound were excluded rather than estimated; this removed BulkSupplements and partiQlar products from the dose calculations, though both sell into this category.

Search demand. US monthly search volumes for dosage-intent queries were taken from a commercial Google Ads-derived keyword dataset, queried for the United States in English in September 2026. Where a compound has more than one common dosage phrasing, the volumes are summed and the constituent terms are named in the tables. Five-year interest figures are Google Trends indices for the United States, normalized across the terms queried together, averaged by calendar year.

Derived metrics. “Panel median” is the median milligrams per labeled serving across the products we read for that compound. “Verdict” compares that median to the retrieved trial range. “Searches per trial” divides monthly dosage search volume by the count of indexed oral-supplementation randomized trials.

Limitations

• The trial ranges are a retrieved sample, not a census. This is the most important caveat. A wider literature search would widen several ranges and could move a verdict from “above” or “below” to “within”.
• Some panels are very small. Ca-AKG rests on a single product, which is our own; pterostilbene and tongkat ali on two each. A median of two products is an average, not a market picture — the pterostilbene median of 300 mg is a figure no product on the shelf actually carries.
• Extract ratios are not milligram-equivalents. For shilajit, tongkat ali and Panax ginseng, a label figure describes a concentrated extract while trials often used a differently standardized preparation. Those above-range verdicts are the weakest in the set and should not be read as evidence of overdosing.
• Salt forms differ. Spermidine trihydrochloride, berberine HCl and calcium alpha-ketoglutarate each contain less of the active moiety than the total salt weight. Where labels declare the salt, our figure is the salt.
• The product panel is a sample too. Thirty-nine products across four retailers shows the direction and size of the gaps; it is not the whole US market, and it over-represents single-active products.
• Trial dose does not mean optimal dose. A dose used in a trial is a dose someone chose to test in a specific population for a specific outcome, sometimes in patients rather than healthy adults. Falling inside the range is not proof a product works, and falling below it is not proof one does not.
• Glutathione counts are inflated. Glutathione appears in thousands of trials as a measured biomarker rather than an administered supplement, so its trial counts are far higher than its oral-supplementation literature warrants. Its dose figures come only from trials that administered it orally.

Reusing this data

The chart at the top of this page and every table below it are free to republish, quote or cite, in whole or in part, with attribution to Eternal Elixir and a link to https://eternalelixir.com/supplement-dose-gap-report/. The chart is available directly at this address. If you are writing about supplement dosing and want the underlying per-product rows or the exact PubMed queries used, the methodology above describes both precisely enough to reproduce, and we are happy to be contacted for the raw table.

Suggested citation: Eternal Elixir. The Supplement Dose Gap Report 2026: Trial Doses vs. US Label Doses. September 6, 2026. https://eternalelixir.com/supplement-dose-gap-report/

Frequently Asked Questions

How much NMN should you take according to human trials?

The human trials we retrieved used 250, 300, 600 and 1,200 mg per day. A 2023 randomized trial in healthy middle-aged adults reported that blood NAD and physical performance responses peaked at 600 mg daily, and several trials used 250 mg/day for 12 to 24 weeks. Typical US products deliver 500 mg per serving, which falls inside that range.

Which longevity supplement has the biggest gap between its label and the research?

TUDCA. The only published randomized human trial we could retrieve used 2,000 mg per day for 16 weeks. US products in our panel deliver 250 to 600 mg per serving, so the median label serving is about a quarter of the trial dose and reaching it takes roughly four servings.

Are spermidine supplements underdosed?

The evidence is genuinely unsettled rather than clearly against the labels. Retrieved human trials used 0.9, 6, 15 and 40 mg per day. One 2023 report concluded that doses below 15 mg per day are unlikely to exert short-term effects, while a 2026 trial reported effects at 6 mg per day. US products sit at 10 mg per serving, inside that spread but in its most contested part.

How much berberine did trials use?

The randomized trials we retrieved used 600 mg per day at the low end, given as 0.3 g twice daily, and up to 2,000 mg per day. Many US berberine capsules contain 500 to 600 mg, so a single capsule sits at or below the lowest trial dose while a two-capsule or 1,200 mg serving reaches into the range.

Is there a proven human dose for calcium alpha-ketoglutarate?

No. The randomized human literature on alpha-ketoglutarate is dominated by ornithine alpha-ketoglutarate given at gram-scale doses in clinical nutrition settings, and by arginine alpha-ketoglutarate in sports contexts. Neither establishes a daily dose for the 500 mg calcium salt sold as a longevity supplement. The commonly quoted 1,000 mg per day figure comes from a non-randomized study, not a controlled trial.

Can a supplement be dosed too high compared to trials?

It can be dosed higher than anything that has been formally tested, which is different from being unsafe. Pterostilbene is the clearest case in our data: retrieved human trials used 10 to 200 mg per day, while single capsules on the US market reach 500 mg. That means the dose has not been studied at that level, not that a specific harm has been shown.

Which longevity compounds are dosed in line with the research?

Four of the twelve we examined have a median US label serving inside the retrieved human trial range: NMN, spermidine, glutathione and Panax ginseng.

Where does this dose data come from?

Trial doses were read verbatim from PubMed-indexed human study abstracts retrieved through the NCBI E-utilities API in September 2026, and each is cited by PMID. Label doses were read from the live public product data of four US retailers on the same date. Search figures come from a Google Ads-derived US keyword dataset and Google Trends. The full method, including what was excluded, is set out in the methodology section.

What this means if you are choosing a supplement

The useful takeaway is not that the industry is broken. Four of twelve compounds line up with the research, and for several of the rest the mismatch is a labelling-convention problem rather than a real one. The useful takeaway is narrower: for three specific compounds — TUDCA, nicotinamide riboside and berberine — the number on the front of the bottle is smaller than the number in the trials, and the fix is reading the serving size rather than assuming one capsule is a dose.

For a fourth, calcium alpha-ketoglutarate, there is no trial number to compare against at all, and any brand quoting one with confidence — ours included — is quoting a convention. Our own pterostilbene capsule sits above every dose we could find in a published trial, and our own spermidine capsule sits in the middle of an argument the literature has not settled. We would rather publish that than not.

If this report changes one habit, we would like it to be checking the milligrams per serving against the studies before checking the price.

Browse the full Eternal Elixir range, or read our companion guides to the best NMN supplements, spermidine supplements and TUDCA supplements in the US.

This report describes doses used in published research and doses printed on product labels. It is not medical advice and not a recommendation to take any particular amount. Talk to a qualified healthcare professional before starting or changing a supplement, particularly if you take prescription medication or have an existing health condition.