Collagen Peptides: Doses Used in Clinical Trials

Collagen is the structural protein in skin, bone, tendon and cartilage, and intact collagen is a large insoluble triple helix the body cannot absorb whole. Manufacturers treat raw sources such as bovine hide, porcine skin, fish skin or chicken cartilage with heat and food-grade proteases, cutting the protein into fragments of roughly 2 to 10 kilodaltons. Hydrolyzed collagen, collagen hydrolysate, collagen peptides and hydrolysed gelatin all describe the same class of product.

The word "peptide" does a lot of work in marketing here. A collagen peptide product is not a single defined molecule the way semaglutide or BPC-157 is. It is a heterogeneous mixture of thousands of fragments whose profile varies by source animal, enzyme and process. Two 10 g scoops from different manufacturers are not chemically identical.

Partly, and as peptides rather than only as loose amino acids. A randomized, double-blind crossover study in six healthy volunteers gave a single 10 g dose dissolved in 200 mL of water, comparing fish, porcine and bovine sources. Prolyl-hydroxyproline (Pro-Hyp) reached maximum plasma concentrations within 60 to 120 minutes, and between 36% and 47% of circulating hydroxyproline remained peptide-bound depending on the hydrolysate tested. A meaningful fraction survives digestion as di- and tripeptides rather than being fully broken down.

One finding is worth holding onto when reading label claims: Gly-Pro was the most abundant free peptide in the products before ingestion, yet its plasma concentration afterwards was around 100 times lower than Pro-Hyp. What is abundant in the powder does not predict what ends up in blood. That study was funded by Rousselot, a collagen hydrolysate manufacturer, and several authors were employed there.

Absorption is not benefit. Showing a fragment in plasma establishes bioavailability, not that it rebuilds anything.

These are the amounts researchers administered in specific studies. They are not a recommendation, and no trial below was designed to find an optimal dose.

The clustering around 2.5 to 10 g is real, but it reflects what sponsors chose to test, not a dose-response curve anyone has mapped. Most commercial scoops land near 10 to 20 g, above what several of the positive trials used.

Skin elasticity. In the 2014 Proksch trial, both the 2.5 g and 5.0 g groups showed a statistically significant improvement in skin elasticity compared with placebo at 8 weeks, and elderly women still had significantly higher elasticity four weeks after the last dose. Skin moisture and transepidermal water loss trended positively in subgroup analysis but did not reach statistical significance. No side effects were noted.

Wrinkles. The companion 2014 trial gave 2.5 g of a branded bioactive collagen peptide to 114 women aged 45-65 for 8 weeks and reported a significant reduction in eye wrinkle volume of about 20% versus placebo, still present four weeks after the last dose. Suction blister biopsies showed 65% higher procollagen type I and 18% higher elastin; a 6% rise in fibrillin did not reach significance.

Joint discomfort. Zdzieblik 2017 gave 5 g daily to 139 physically active adults with functional knee pain for 12 weeks. Activity-related pain on a visual analogue scale improved more in the treated group (ΔVAS 19.5 ± 2.4) than placebo (13.9 ± 2.1, p = 0.046), confirmed by physician assessment (16.7 ± 1.8 versus 12.2 ± 1.8, p = 0.021). Pain at rest improved in both groups with no significant separation (p = 0.209). Note that this paper carries a published erratum.

Bone density. König 2018 is the longest of the four: 5 g daily for 12 months in postmenopausal women with age-related low bone density, 131 randomized and 102 completing. Lumbar spine BMD rose 3.0% on treatment and fell 1.3% on placebo (p = 0.030); femoral neck BMD rose 6.7% versus a 1.0% fall (p = 0.003). The formation marker P1NP rose significantly on treatment, while CTX-1 stayed flat on treatment and rose on placebo. The study was part-funded by GELITA, whose ingredient was used.

Pooled evidence. A 2021 systematic review and meta-analysis in the International Journal of Dermatology pooled 19 randomized double-blind controlled trials covering 1,125 participants aged 20 to 70, 95% of them women. It found favourable grouped results for skin hydration, elasticity and wrinkles versus placebo, and concluded that roughly 90 days of intake is where effects were demonstrated.

The limitations here are structural rather than incidental.

Most of the influential trials used branded ingredients and were funded by, or co-authored with researchers connected to, the companies selling those ingredients. That does not make the results wrong, but industry sponsorship concentrated in a small literature is a reason to weight it cautiously.

Nobody has run a proper dose-ranging study. That 2.5 g worked in one skin trial does not tell you whether 10 g works better, the same, or is simply wasted.

There is very little work comparing collagen peptides against an equal gram amount of ordinary dietary protein, so some of the observed effect could plausibly be a protein effect rather than a collagen-specific one. The published trials were generally not designed to separate those.

Effect sizes are mostly modest and measured with instruments, not mirrors. A significant change in a cutometer reading is not the same as a visible change you would notice unprompted.

If you came here after using a peptide dosing tool, the mental model does not transfer. Injectable research peptides arrive as a few milligrams of lyophilized powder, get reconstituted with a measured volume of bacteriostatic water, and are dosed in fractions of a millilitre read off insulin-syringe units. That workflow is what the reconstitution guide and the syringe measurement guide handle, and it is what every entry in the dosage calculator is built around.

Collagen peptides skip all of it. The product is a soluble food powder, measured by scoop or kitchen scale and stirred into liquid. There is no concentration to solve for and no dead space to account for. If the distinction between orally active food peptides and injectable research compounds is new to you, the beginners guide to peptides covers the categories and the peptide calculator guide explains what the calculator is and is not for.

For injectable compounds studied for skin and connective tissue, see the skin and hair category or the GHK-Cu dosage page — a different regulatory and practical class entirely.

In the United States, hydrolyzed collagen is handled as food, not as a drug. No FDA approval is required before a collagen peptide supplement goes on sale, and no collagen peptide product holds an FDA drug approval for treating any condition.

Some hydrolysates have gone through FDA's voluntary GRAS notification process. In GRAS Notice GRN 000713, Rousselot notified FDA of its view that hydrolyzed porcine trachea cartilage is generally recognized as safe in certain baked goods, beverages, milk products, fruit juices and candies at 1.2 to 2 g per serving. FDA's response concluded: "Based on the information that Rousselot provided, as well as other information available to FDA, we have no questions at this time regarding Rousselot's conclusion that HPTC is GRAS under its intended conditions of use." The same letter states it "is not an affirmation that HPTC is GRAS under 21 CFR 170.35."

That is narrower than most marketing implies. It concerns safety at specified food use levels and says nothing about efficacy for skin, joints or bone.

The trials above reported no notable side effects, and gastrointestinal complaints are the most commonly described issue in the wider literature. Because collagen is sourced from cattle, pigs, fish or poultry, source species matters for anyone with an allergy to that animal or observing dietary restrictions.

Appropriate protein load is a question for a clinician if you have reduced kidney function, and there is little trial data in pregnancy, in adolescents, or in people on multiple medications. Whether any supplement suits your situation, and at what amount, is a decision for a qualified healthcare professional who knows your history, not something to settle from a table of trial doses.