Ask how much collagen to take and the answer from brand pages is usually one number — often "10–20 g" — with no trial attached. The trials themselves did not agree on one number, because they were not asking one question. The studies that measured muscle used one amount; the ones that measured knees, tendons, bones or skin used others.
This page lines up the trial doses by what was measured, then sets them against what is actually sold: every on-market collagen-peptide powder label in the US government's label database that states an amount. It reports doses; it does not recommend one, per our editorial standards.
The trial doses, by what was measured
| Outcome | Trial | Who | Daily amount | How long |
|---|---|---|---|---|
| Muscle and strength (with training) | Zdzieblik 2015, Br J Nutr | 53 men with sarcopenia, average age 72 | 15 g | 12 weeks |
| Muscle and strength (with training) | Jendricke 2019, Nutrients | 77 premenopausal women | 15 g | 12 weeks |
| Activity-related joint pain | Clark 2008, Curr Med Res Opin | 147 varsity and club athletes (97 analysed) | 10 g | 24 weeks |
| Knee pain in sport | Zdzieblik 2017, Appl Physiol Nutr Metab | 139 athletes | 5 g | 12 weeks |
| Achilles tendon size (with training) | Jerger 2022, Scand J Med Sci Sports | 40 young men | 5 g | 14 weeks |
| Bone mineral density | König 2018, Nutrients | 131 postmenopausal women | 5 g | 12 months |
| Collagen-synthesis marker | Shaw 2017, Am J Clin Nutr | 8 men, crossover | 5 or 15 g gelatin + vitamin C | 3 days |
| Skin elasticity | Proksch 2014, Skin Pharmacol Physiol | 69 women aged 35–55 | 2.5 or 5 g | 8 weeks |
Read down the amount column and the pattern is plain: the muscle outcome was tested at 15 g; everything else at 10 g or less. Neither muscle paper tested a lower amount, so the 15 g is the dose that was studied rather than a dose shown to be needed. Collagen is also low in leucine, the amino acid most tied to muscle protein synthesis, and whether collagen builds muscle at all is disputed; the argument is set out on collagen peptides for muscle growth.
What the table does not contain is a head-to-head dose comparison for muscle. No trial found for this page gave one group 10 g and another 15 g and measured strength or lean mass. The only within-study dose contrast is Shaw 2017, where 15 g of vitamin C-enriched gelatin raised a blood marker of collagen synthesis and 5 g did not do so significantly — in eight men, over three days, on a marker rather than on a tendon.
Skin is outside this site's lane; Peptide Glow Journal's collagen page covers that evidence.
What a scoop actually contains: 212 labels
To see how the trial amounts compare with what is sold, this page queried the NIH Office of Dietary Supplements' Dietary Supplement Label Database on 2026-09-27 for on-market powders whose product name includes "collagen peptides", then read the collagen amount from each label's supplement-facts panel.
212 labels from 94 brands state a collagen amount per serving.
| Collagen per serving | Labels |
|---|---|
| About 20 g | 50 |
| 15 g or more (all of them the 20 g group) | 50 |
| 10 g to under 15 g | 109 |
| 5 g to under 10 g | 47 |
| Under 5 g | 6 |
| Median | 11 g |
The three most common amounts are about 20 g (50 labels), 11 g (43) and 10 g (40). Three findings follow:
- 162 of the 212 labels (76%) give less than the 15 g the muscle trials used in a single serving. A buyer reading "muscle" on the tub and the trial on a blog is often looking at two different amounts.
- Almost every label (206 of 212) gives at least the 5 g used in the knee, tendon and bone trials.
- The 11 g cluster is a scoop size, not a research dose. Read label by label, most of those entries are an 11 g scoop of powder that declares 10 g of protein — collagen powder is not pure protein, so the collagen weight and the protein weight on one panel differ by about a gram. No trial found for this page used 11 g.
A label amount is what the maker declares, not what a laboratory found in the tub. The database records labels, not unique products — one brand's flavours and sizes can appear as several entries — and it does not test anything.
Who ran the trials
The dose record is thinner than eight rows make it look. Five of the eight trials come from one University of Freiburg group (König, Zdzieblik, Jendricke, Jerger and colleagues). Four of the five carry a co-author from the Collagen Research Institute in Kiel, as does the skin trial, and four name "specific" or "bioactive" collagen peptides — branded ingredients — in their titles.
In 2025 the British Journal of Nutrition published a corrigendum to the 2015 sarcopenia trial (PMID 40859798). The original acknowledgements said part of the costs were paid by Gelita AG and that there were no conflicts of interest. The corrected text states that the research received no specific grant and that Gelita AG provided the study samples and paid an amount covering participant compensation and sample analysis. The results were not changed. What changed is what a reader knows about who supplied and paid for the most-cited collagen muscle trial. This site's collagen and recovery page found the same pattern in the recovery trials.
Timing and form
Timing. The Freiburg muscle trials gave the supplement after training sessions; Shaw's gelatin study gave it an hour before exercise. No trial found for this page compared times of day on a muscle, tendon or joint outcome, so "best time" is a practice question without a trial behind it.
Form. "Collagen peptides", "hydrolysed collagen" and "collagen hydrolysate" describe the same thing: collagen broken into short chains that dissolve in water. Gelatin is less broken down. The trials above used hydrolysates except Shaw 2017, which used gelatin.
Protein arithmetic. Collagen counts toward protein intake, but it contains no tryptophan and little leucine, so it is an incomplete protein. The comparison with whey is on collagen vs whey protein, and the pairing with creatine on creatine and collagen peptides.
The honest summary
- Muscle trials: 15 g a day, 12 weeks, with resistance training (two trials, 130 people).
- Joints: 5–10 g; tendon and bone: 5 g; skin: 2.5–5 g.
- No trial compared doses for muscle. The only dose contrast (5 vs 15 g gelatin) used a blood marker in eight men.
- 212 US labels: median 11 g a serving; 162 below 15 g; 50 at about 20 g.
- Most dose trials come from one group testing one supplier's peptides, and a 2025 corrigendum added that the supplier provided the product and paid participant and analysis costs in the key muscle trial.
Limits of this page
The label census covers products whose name includes "collagen peptides" and whose panel states a collagen amount; blends, capsules and products named otherwise are outside it, and the database lags the market. Trial populations were specific — older sarcopenic men, premenopausal women, athletes with pain — and results in them may not carry to others. Search volumes are DataForSEO's US clickstream estimates for 2026-09-21.
Sources and dates
- Zdzieblik D, Oesser S, Baumstark MW, Gollhofer A, König D. Br J Nutr 2015;114:1237–45 — PMID 26353786; corrigendum Br J Nutr 2025;134:440 — PMID 40859798 (full text via Europe PMC, PMC12580965), both read 2026-09-27.
- Jendricke P, Centner C, Zdzieblik D, Gollhofer A, König D. Nutrients 2019;11:892 — PMID 31010031.
- Clark KL, Sebastianelli W, Flechsenhar KR, et al. Curr Med Res Opin 2008;24:1485–96 — PMID 18416885.
- Zdzieblik D, Oesser S, Gollhofer A, König D. Appl Physiol Nutr Metab 2017;42:588–95 — PMID 28177710.
- Jerger S, Centner C, Lauber B, et al. Scand J Med Sci Sports 2022;32:1131–41 — PMID 35403756.
- König D, Oesser S, Scharla S, Zdzieblik D, Gollhofer A. Nutrients 2018;10:97 — PMID 29337906.
- Shaw G, Lee-Barthel A, Ross ML, Wang B, Baar K. Am J Clin Nutr 2017;105:136–43 — PMID 27852613.
- Proksch E, Segger D, Degwert J, et al. Skin Pharmacol Physiol 2014;27:47–55 — PMID 23949208.
- NIH Office of Dietary Supplements, Dietary Supplement Label Database API v9 — search "collagen peptides", on-market powders, 212 labels with a stated collagen amount, retrieved 2026-09-27.