Does Collagen Work? What the Evidence Shows

The short answer is: it depends entirely on what you mean by "work". The evidence varies considerably by outcome, and a major 2025 analysis has cast real doubt on some of the most widely marketed claims. Here is what the research actually shows.

The 2025 American Journal of Medicine meta-analysis

A meta-analysis published in the American Journal of Medicine in 2025 (Myung SK, Park Y. Am J Med. 2025;138(9):1264–1277) pooled 23 randomised controlled trials covering 1,474 participants, examining collagen supplementation against skin-ageing markers such as hydration, elasticity and wrinkles. Across all 23 trials combined, collagen did show significant improvement in all three measures. The finding that matters comes from splitting those trials up: in trials that did not receive pharmaceutical industry funding, and separately in the higher-quality trials, the effect disappeared.

That split is the whole story, and it is a more uncomfortable one than a flat null result would be. It means the positive findings are concentrated in the studies with a commercial interest in the outcome and in those with weaker methodology. It does not prove collagen has no effect on skin — absence of evidence is not evidence of absence — but anyone citing the pooled figure as support is quoting the number that the authors themselves went on to qualify. The honest position is that the picture is unsettled.

It is also worth noting that UK regulations do not permit any health claim linking collagen to skin, hair, nails, joints or bones, regardless of what individual studies suggest. That constraint exists for a reason: the regulatory threshold for a permitted claim is high, and the current evidence does not meet it.

Where research interest has been stronger: tendons and connective tissue

A different area of research has attracted attention. Shaw and colleagues (Shaw G, Lee-Barthel A, Ross ML, Wang B, Baar K. Am J Clin Nutr. 2017;105(1):136–143) investigated whether vitamin C-enriched gelatin taken before exercise was associated with changes in connective tissue markers. The protocol examined in that study was 15 g of vitamin C-enriched gelatin consumed approximately 60 minutes before a bout of exercise designed to load the relevant tissue. Note that the study used gelatin, not hydrolysed collagen powder; these are related but distinct products, and whether findings transfer between them has not been established.

The study measured markers in the blood and in engineered tissue models rather than clinical outcomes in living tendons, so it is important not to overstate what was shown. It is separately established that vitamin C contributes to normal collagen formation in the body — this is an authorised claim for vitamin C, not for collagen supplements.

Subsequent work has explored similar protocols in athletes returning from soft tissue injuries, with some positive findings, though sample sizes remain small and study designs vary. This is an area of ongoing research; it is not conclusive. Anyone managing a soft tissue injury or recovery from one should speak to a physiotherapist or their GP rather than rely on supplementation alone.

Why the outcomes differ

It is worth understanding why the tendon and connective tissue research may show more signal than the skin research, even if neither is definitive.

• Tendons and ligaments have relatively poor blood supply and limited capacity for repair. They are also under direct mechanical load during exercise, which appears relevant to when synthesis occurs.
• Skin ageing involves many overlapping processes — UV exposure, hormonal changes, hydration, genetics — making it harder to isolate any single dietary variable.
• The amino acid profile of hydrolysed collagen (rich in glycine, proline and hydroxyproline) is specific and distinct from whey or plant proteins, which may matter more in some tissues than others.

None of this means the skin evidence will never improve. It means that, right now, the claims made for collagen and skin are ahead of what the research supports.

What collagen is not

Collagen powder is not a complete protein source for general muscle support. It is low in essential amino acids, particularly leucine, which plays a key role in muscle protein synthesis. If your primary goal is muscle maintenance or growth, protein contributes to a growth in muscle mass and to the maintenance of muscle mass — but for that purpose, whey, faba bean, or another protein with a complete essential amino acid profile is likely to serve you better than collagen. Collagen and a general protein supplement are not the same thing and are not interchangeable.

Who has been studied

The research to date has primarily examined people doing regular exercise that loads tendons and connective tissue — running, jumping, resistance training. The Shaw protocol involved vitamin C-enriched gelatin taken before the relevant exercise session. This research is ongoing and inconclusive; it describes what has been studied, not a recommended course of action.

If you are considering collagen for a specific recovery concern or have an existing injury, speak to your GP or a physiotherapist. Supplementation is not a substitute for appropriate clinical advice.

The honest summary

• The 2025 meta-analysis found a significant skin-ageing effect across all pooled trials, but no effect in the independently funded and higher-quality subsets.
• The tendon and connective tissue research is more encouraging but still limited in scale and largely conducted using gelatin rather than hydrolysed collagen.
• The timing and dose used in the Shaw protocol (15 g vitamin C-enriched gelatin, 60 minutes pre-exercise) is specific to what was studied — it is not the same as taking collagen at any time in any amount, and the product used was gelatin.
• Collagen is not a general protein supplement and should not be treated as one.
• UK regulations do not permit health claims for collagen in any of the areas it is most commonly marketed for.

The evidence for collagen is neither as strong as some brands imply nor entirely absent. The honest position is that it depends on what outcome you are interested in, and for most of the popular claims, the science is still catching up. If you want the practical questions rather than the evidence base, we cover how much collagen per day and choosing between bovine and marine separately.