Collagen helps your joints

In people with osteoarthritis, collagen supplements have been linked in trials to a small reduction in joint pain and a small improvement in function. A 2024 review of 35 randomised trials, covering 3,165 people, found that pain fell by about 8.5 points on a 100-point scale more than with the comparison treatments. The review's authors treat 9 points as the smallest change a patient would notice. Function improved by a similar small amount, and side effects were no more common than with the comparisons.

Most of those trials lasted under six months, and the review recorded 20 of the 35 as not independent of industry funding, 10 as unclear and 5 as independent. Searches found about ten more osteoarthritis trials published since the review closed its search; seven of the eight that used a placebo reported a benefit and one did not. The largest, in 212 people with knee osteoarthritis, tested a capsule combining 40 mg of undenatured type II collagen with boswellia extract, methylsulfonylmethane and vitamins C and D. Symptom scores improved more with the capsule than with placebo over six months, and use of rescue painkillers did not differ. The capsule's maker is listed as the trial's sponsor, and a five-ingredient capsule cannot show how much of the effect came from collagen. A 2025 trial run without manufacturer funding, in 68 people with knee osteoarthritis, found no difference from placebo in pain, function or use of rescue painkillers over 12 weeks; both groups also received muscle-strengthening exercises and rescue painkillers as part of standard care. A 2022 review of 19 studies in rheumatoid arthritis and osteoarthritis rated most of them poor quality and noted that several reported high rates of adverse effects and lower effectiveness than routine treatments. Every rheumatoid arthritis trial in it used intact type II collagen, and the most recent dated from 2011.

For people without a joint diagnosis, searches found about 18 trials, most of them small. At least five name a company that makes or sells collagen as a funder, and most reported improvement over placebo on at least one measure of joint pain or movement. In 2011 a European Food Safety Authority panel assessed one of them, a 24-week trial of 10 g a day in 147 student athletes with joint discomfort, submitted by a collagen maker in support of a joint health claim. It found no difference from placebo on any of the trial's 15 pain and discomfort scores once the number of comparisons was taken into account, and concluded that a benefit for joint health in active people had not been established. Two of the largest point in opposite directions. In 180 physically active adults aged 18 to 30 with knee pain during exercise, 5 g a day of collagen peptides for 12 weeks reduced activity-related pain by 21.9 points on a 100-point scale, against 15.6 points with placebo; pain at rest and after squats did not differ. That trial was partly funded by GELITA, which also supplied the peptides. In 167 active adults with a median age of 63 and self-reported knee pain, 10 g a day for 12 weeks reduced pain by the same amount as placebo, and markers of inflammation and cartilage turnover did not change.

Collagen supplements come in different forms, tested at very different doses. Most trials of hydrolysed collagen, also sold as collagen peptides, used about 1 to 10 g a day, while undenatured type II collagen (UC-II) is taken in milligrams. In a trial sponsored by its manufacturer, in which two of the three authors were company employees, 40 mg a day of UC-II for six months improved overall knee osteoarthritis scores in 191 people more than placebo and more than glucosamine with chondroitin. The 2025 trial that found no effect combined UC-II with under 1 g a day of hydrolysed collagen. The 2024 review found no clear difference between forms but described those comparisons as underpowered, and its trials also included gelatin, combination products and several collagen injections given directly into the joint.

Across the 2024 review, adverse events and withdrawals occurred at similar rates with collagen and with the comparisons, and the 2025 trial reported no adverse effects in either group. In the 212-person trial, mild adverse events were reported by about one in five people in both groups, and one person taking the capsule left the trial after an allergic reaction. The undenatured type II collagen in that capsule came from chicken, and the trial did not enrol people with a known intolerance to chicken protein. None of the sources on this claim studied interactions with medicines, and none followed people for longer than about a year.