Does Collagen Actually Work? The Evidence, Graded Honestly

Collagen supplements work for some things and not others, and the difference is well documented. For skin hydration and elasticity the evidence is strong: a 2021 meta-analysis of 19 randomised controlled trials in 1,125 people found consistent improvements. For joint pain, tendon health, muscle with training and bone density the evidence is moderate and dose-dependent. For hair growth, gut healing and weight loss it is weak or missing. Anyone who tells you collagen does everything, or that it does nothing, has not read the trials.

This is the page we wish existed when we started KUDU. It grades each claimed benefit by the quality of evidence behind it, gives the doses the trials used, answers the sceptic's argument that collagen is just digested into amino acids, and ends with a simple way to test whether it is working for you.

How to read the evidence

Three things decide whether a collagen claim holds up: whether there are placebo-controlled human trials, whether they used a dose you can actually buy, and whether the effect showed up in a measurement rather than a questionnaire. Almost every positive trial used hydrolysed collagen peptides at 2.5g to 20g a day for 8 to 24 weeks. Almost every negative or inconclusive result involved tiny doses, short durations or outcomes collagen was never likely to change.

Skin: strong

The de Miranda meta-analysis (International Journal of Dermatology, 2021) pooled 19 randomised, placebo-controlled trials and found significant improvements in skin hydration, elasticity and wrinkle depth with hydrolysed collagen over 8 to 12 weeks. Individual trials such as Proksch 2014 measured the changes with instruments, not opinions, at 2.5g to 5g a day. This is the best-supported claim in the category and the one most sceptics concede.

Joints: moderate

In active adults with knee discomfort, 5g daily for 12 weeks reduced activity-related pain versus placebo (Zdzieblik 2017). In 147 athletes, 10g daily for 24 weeks reduced joint pain (Clark 2008). In knee osteoarthritis, 10g for 6 months improved pain scores (Benito-Ruiz 2009). The effects are real but modest, need 12 weeks, and a 1.2g trial showed how quickly benefit disappears at low doses. Full breakdown: does collagen help joint pain.

Tendons and ligaments: moderate, with exercise

15g of collagen or gelatin with vitamin C taken an hour before exercise roughly doubled a blood marker of collagen synthesis (Shaw 2017). 5g daily with calf training increased Achilles tendon size and stiffness over 14 weeks (Jerger 2022). 2.5g with rehab improved Achilles tendinopathy over 6 months (Praet 2019). 5g for 6 months improved chronic ankle instability (Dressler 2018). Every one of these paired collagen with loading; none show benefit from collagen alone. Protocol: collagen for tendonitis.

Muscle and body composition: moderate, with training

Three 12-week trials in older men, young men and women gave 15g daily with resistance training and found greater gains in fat-free mass and strength than training alone (Zdzieblik 2015, Kirmse 2019, Jendricke 2019). Collagen is low in leucine and is not a substitute for complete protein; the effect seems to come from connective tissue and extra total protein. See collagen vs whey.

Bone: moderate

131 postmenopausal women taking 5g daily for 12 months increased bone mineral density at the spine and hip while placebo lost bone (König 2018), and a 4-year follow-up in the same group reported continued gains. One high-quality trial plus a follow-up is promising but not yet a body of evidence.

Recovery and soreness: moderate

20g daily for 9 days reduced muscle soreness and sped recovery of jump performance after damaging exercise (Clifford 2019). One trial, but a clean one.

Nails: moderate

2.5g daily for 24 weeks increased nail growth rate by about 12% and cut broken nails by 42% (Hexsel 2017). Open-label, so weaker than the skin trials, but consistent with them.

Hair: weak

No randomised trial shows collagen makes hair grow faster or thicker in healthy adults, and it does not affect male pattern hair loss. It supplies the amino acids keratin is built from, which is a plausible mechanism and nothing more. See collagen for hair and nails.

Gut: weak

Collagen is rich in glycine and glutamine and type III collagen is part of the gut wall, but human trials on gut symptoms are small, short and mostly uncontrolled. Plausible, unproven.

Weight loss: weak

Collagen is a satiating, low-calorie protein (85 calories per 20g) and the training trials showed more fat loss with collagen than placebo, but there is no evidence collagen causes weight loss on its own.

The sceptic's argument, answered

The standard objection: "collagen is just protein, it gets digested into amino acids like any other, so it cannot target your skin or joints". Two findings answer it. First, hydrolysed collagen is not fully broken down to single amino acids: specific di- and tri-peptides, particularly prolyl-hydroxyproline, are absorbed intact and detected in blood within an hour (Iwai 2005). Second, those peptides act as signals, stimulating fibroblasts and chondrocytes to produce more collagen and matrix in laboratory studies. That is why collagen outperforms an equal dose of a generic protein in the skin and tendon trials, which would not happen if the amino acids were all that mattered.

Why some people say it did nothing

• Dose too low: 1g to 2g gummies and 2.5g beauty shots are below every joint, tendon and muscle trial.
• Stopped too early: skin needs 8 weeks, joints 12, tendons 3 to 6 months.
• Wrong outcome: expecting hair regrowth or weight loss, where the evidence is weak.
• No loading: tendon and muscle benefits need exercise alongside.

How to test it on yourself in 12 weeks

1. Pick one measurable outcome: skin photos in the same light, a nail growth mark, a joint pain score, next-day soreness.
2. Take 10g to 20g of hydrolysed collagen every day. Before training if joints or tendons are the goal.
3. Check at 4, 8 and 12 weeks. Skin first, joints last.

KUDU liquid collagen delivers 20g of hydrolysed bovine collagen per ready-to-drink sachet, 0g sugar, 85 calories, Informed Sport certified, with a 60-day money-back guarantee that covers the trial period above.

Run the 12-week test with KUDU →

Does collagen work: FAQs

Is there scientific proof that collagen works?

Yes for skin: a meta-analysis of 19 placebo-controlled trials found improved hydration, elasticity and wrinkles. Joint, tendon, muscle and bone benefits have moderate evidence at 5g to 20g a day.

Is collagen just broken down into amino acids?

Not entirely. Specific collagen di- and tri-peptides are absorbed intact and act as signals to collagen-producing cells, which is why collagen outperforms generic protein in the trials.

How long does collagen take to work?

Skin 8 weeks, joints 12 weeks, tendons 3 to 6 months, bone 12 months. See our full timeline.

What does collagen not do?

Regrow hair, raise testosterone, cause weight loss on its own, or fix a tendon without exercise.

Which collagen has the best evidence?

Hydrolysed collagen peptides at 2.5g to 20g a day. Bovine (type I and III) is the source in most of the sports and joint trials.

Related reading

• How long does collagen take to work?
• How much collagen per day
• How collagen is absorbed
• Hydrolysed collagen vs collagen peptides

Sources

• de Miranda RB et al. Effects of hydrolyzed collagen supplementation on skin aging: a systematic review and meta-analysis. Int J Dermatol 2021.
• Proksch E et al. Skin Pharmacol Physiol 2014. Hexsel D et al. J Cosmet Dermatol 2017.
• Zdzieblik D et al. Appl Physiol Nutr Metab 2017. Clark KL et al. Curr Med Res Opin 2008. Benito-Ruiz P et al. Int J Food Sci Nutr 2009.
• Shaw G et al. Am J Clin Nutr 2017. Jerger S et al. Scand J Med Sci Sports 2022. Praet SFE et al. Nutrients 2019. Dressler P et al. J Sports Sci Med 2018.
• Zdzieblik 2015 (Br J Nutr), Kirmse 2019, Jendricke 2019 (Nutrients). Clifford T et al. Amino Acids 2019.
• König D et al. Nutrients 2018; Zdzieblik D et al. 4-year follow-up, Nutrients 2021.
• Iwai K et al. J Agric Food Chem 2005.