Collagen is one of the few supplements where the headline claim holds up. Randomized trials of 2.5 to 10 grams of hydrolyzed peptides daily over 8 to 12 weeks show improved skin elasticity and hydration, with modest but reproducible effect sizes. Joint discomfort has weaker support in the same direction at 10 grams daily.

The claims that sell the most product — hair growth, gut lining repair, muscle building — have no adequate human trials behind them. This page separates the two groups, gives the doses the trials used, and covers the two things most buyers get wrong: the vitamin C cofactor, and counting collagen toward a daily protein target.

The Verdict

10 g/day of hydrolyzed collagen peptides covers both the skin and joint use cases; 2.5–5 g is enough for skin alone. For joint pain specifically, undenatured type II collagen at 40 mg/day is a different product with a different mechanism. Take it with vitamin C, since vitamin C is a required cofactor for collagen synthesis. Do not count collagen toward a protein target — it has no tryptophan and almost no leucine. Recurrent calcium oxalate stone formers should discuss it with a physician first.

What the evidence supports, claim by claim

ClaimDose used in trialsTime to effectStrength of evidence
Skin elasticity and hydration 2.5–10 g/day hydrolyzed peptides 8–12 weeks Good — multiple randomized trials and pooled analyses, consistent direction, modest effect size
Joint discomfort in athletes and osteoarthritis 10 g/day hydrolyzed peptides 3–6 months Moderate — several positive trials, some null results, effect smaller than an NSAID
Joint discomfort, undenatured type II route 40 mg/day UC-II 3–6 months Moderate — a different mechanism (oral tolerance), dosed 250× lower than peptides
Tendon and ligament collagen synthesis 15 g gelatin or peptides + ~50 mg vitamin C, 30–60 min before loading Weeks, with repeated exposure Emerging — small mechanistic studies; the timing and vitamin C are integral to the protocol
Bone mineral density in postmenopausal women 5 g/day specific peptides 12 months Limited — one notable trial, not widely replicated
Nail growth and brittleness 2.5 g/day 6 months Weak — small open-label studies without adequate controls
Hair growth, gut lining repair, muscle building — — Not supported — no adequate human trials for these claims

The mechanism behind the skin findings is better established than it once was. Collagen peptides are largely broken down during digestion, but specific di- and tri-peptides — prolyl-hydroxyproline in particular — survive intact, appear in the bloodstream, and have been shown to reach skin and stimulate fibroblast activity. That is a real pathway, and it explains why the effect exists and why it is small.

Forms and sources

  • Hydrolyzed collagen peptides. Enzymatically cut into short chains, dissolve in cold liquid, and the form used in nearly every positive skin and joint trial. This is the default.
  • Gelatin. The same protein, less broken down. Gels when cold and requires warm liquid. Works for the tendon protocol, where gelatin was the material used, and is inconvenient in a drink.
  • Undenatured type II collagen (UC-II). Not a peptide product. It works through oral tolerance — small intact fragments interacting with gut-associated immune tissue to reduce joint inflammation — which is why the dose is 40 mg rather than 10 g. Do not treat 40 mg of UC-II and 10 g of peptides as interchangeable.
  • Marine collagen. Fish skin and scales, mostly type I, slightly lower molecular weight. Suits pescatarians; contraindicated with fish allergy.
  • Bovine collagen. Hide-derived, types I and III, the cheapest and most common source.
  • Chicken and eggshell membrane collagen. Richer in type II; usually found inside multi-collagen blends.

Multi-collagen blends advertising types I, II, III, V, and X sound comprehensive and are worth a caution. Type II works at 40 mg only in its undenatured form; once hydrolyzed into a peptide blend, that mechanism is gone. A blend claiming type II coverage while listing hydrolyzed peptides is not delivering the joint mechanism the UC-II trials tested.

The five brands — listed alphabetically

Ordered alphabetically. No brand is ranked above another, because at the level of hydrolyzed bovine peptides these products are close to interchangeable and the price differences reflect packaging and flavoring more than material.

BrandWhat it offersSuited to
Ancient Nutrition Multi-source blends (bovine, chicken, fish, eggshell membrane) Buyers who want several collagen types in one product
Bulletproof Grass-fed bovine peptides, flavored and unflavored Buyers who want flavored options
Live Conscious Grass-fed bovine hydrolyzed peptides Lower price per serving
Sports Research Grass-fed bovine type I and III; NSF Certified for Sport on some SKUs Drug-tested athletes
Vital Proteins Grass-fed bovine peptides; broad retail availability, marine option Availability and a marine alternative

Judge on four things rather than on brand: grams of collagen per serving (many flavored products deliver 5–7 g per scoop, not 10), hydrolyzed peptides versus gelatin, published third-party testing, and cost per gram rather than cost per container.

Safety, contraindications, and the oxalate question

  • Calcium oxalate kidney stones. Collagen is roughly 10% hydroxyproline, and hydroxyproline is metabolized through glyoxylate to oxalate. Studies giving gelatin or hydroxyproline loads show measurable increases in urinary oxalate. For most people this is inconsequential; for a recurrent calcium oxalate stone former taking 15–20 g daily, it is worth raising with a physician before starting.
  • Allergy. Marine collagen is a genuine risk with fish or shellfish allergy. Bovine and chicken sources carry the corresponding risks. Source is printed on every label.
  • Heavy metals. Bovine hide and marine sources are plausible routes for cadmium and lead. Third-party testing is the only external check, since supplements are not reviewed for content before sale.
  • GI tolerance. Bloating and fullness at doses above 15 g are the most reported side effects, and are usually solved by splitting the dose.
  • Drug interactions. None of clinical significance are established. This is one of the few supplement categories where that is true.
  • Pregnancy. No specific contraindication, but no trial safety data either. Sourcing and third-party testing matter more here than usual.

What collagen does not do

  • Replace dietary protein. Incomplete amino acid profile; see the leucine note above.
  • Repair intestinal lining. The claim is mechanistically plausible and has not been tested adequately in humans. Glutamine has more direct support in this area, and even that is limited.
  • Reverse androgenic hair loss. A different mechanism entirely — dihydrotestosterone acting on follicles is not an amino acid supply problem.
  • Lower biological age. No epigenetic clock or mortality evidence exists for collagen supplementation.
  • Rebuild cartilage. Symptom improvement in joint trials is not the same as structural regeneration, and imaging endpoints have not shown the latter.

What to measure

  • Skin: photographs under fixed lighting and angle at baseline, 8 weeks, and 12 weeks. Subjective impressions at 12 weeks are unreliable without a reference image.
  • Joints: a numeric pain score for one specific activity, recorded weekly. “My knees feel better” is not measurable; “stairs went from 6/10 to 3/10” is.
  • Tendon work: load tolerance in the specific movement, tracked alongside the rehabilitation program. Collagen supports the loading protocol; it does not replace it.
  • Stop rule: if nothing has changed at 12 weeks on 10 g daily with adequate vitamin C, it is not working for you. This category has a low enough effect size that indefinite continuation without a measured change is the default failure mode.

Frequently Asked Questions

Does collagen actually work?

For skin, yes, modestly and reproducibly. Randomized trials and pooled analyses of 2.5–10 g daily over 8–12 weeks show improved skin elasticity and hydration, with effect sizes that are real rather than dramatic. For joint discomfort the evidence is weaker but positive on balance at 10 g daily over 3–6 months. For hair growth, gut lining repair, and muscle building, there are no adequate human trials, and those are the claims that sell most of the product.

How much collagen should I take and when?

For skin, 2.5–10 g daily of hydrolyzed peptides; timing does not appear to matter. For joints, 10 g daily of peptides, or 40 mg daily of undenatured type II collagen, which works by a different mechanism entirely. For tendon and ligament work, the protocol from the mechanistic studies is 15 g of gelatin or peptides plus about 50 mg of vitamin C, taken 30–60 minutes before loading the tissue — the timing exists because tendon blood flow is low and you want the amino acid peak to coincide with mechanical loading.

Why does collagen need vitamin C?

Because vitamin C is a required cofactor for prolyl and lysyl hydroxylase, the enzymes that hydroxylate proline and lysine residues during collagen synthesis. Without that hydroxylation, the triple helix does not stabilize properly — this is the biochemistry behind scurvy. Taking collagen peptides without adequate vitamin C supplies the building blocks while limiting the assembly step. Most people get enough from diet, but every tendon protocol in the literature includes vitamin C explicitly.

Can I count collagen toward my daily protein target?

No, and this is the most common mistake in the category. Collagen contains no tryptophan, which makes it an incomplete protein, and it is very low in leucine — roughly 0.4 g per 15 g serving against the 2.5 g needed to trigger muscle protein synthesis. It still counts as protein on a nutrition label, so someone taking 20 g of collagen daily and counting it toward a 130 g target is short by 20 g of usable protein. Count it separately from your protein target.

Marine or bovine — does it matter?

Rarely. Marine collagen has a slightly lower molecular weight and modestly higher absorption in some studies, and it suits pescatarians and anyone avoiding bovine products. Bovine hide is the cheaper source and dominates the market. Both deliver type I peptides with essentially the same glycine, proline, and hydroxyproline signature. Choose marine for dietary reasons or a bovine allergy, not for a performance difference. If you have a fish or shellfish allergy, marine collagen is a genuine risk.

Who should not take collagen?

Anyone with a history of calcium oxalate kidney stones should be cautious. Collagen is roughly 10% hydroxyproline, and hydroxyproline is metabolized to glyoxylate and then to oxalate, so large daily doses raise urinary oxalate excretion. In a recurrent stone former that is a meaningful consideration and worth discussing with a physician. Beyond that: fish allergy rules out marine collagen, and anyone with a bovine or chicken protein allergy should check the source. There are no significant drug interactions.

How do I choose a brand?

The five brands below are listed alphabetically and none is ranked above another, because at the peptide level the products are close to interchangeable. Judge on four things: a stated dose of at least 10 g per serving if joints are the goal, hydrolyzed peptides rather than unhydrolyzed gelatin if you want it to dissolve in cold liquid, third-party testing given that bovine hide is a plausible route for heavy metal contamination, and cost per gram rather than cost per tub. Flavored products often deliver less collagen per scoop than the tub size implies.

Related