Collagen: for muscle or for joints?

Collagen is sold for everything at once: skin, joints, tendons, and now muscle. Yet these are two different questions, with two literatures that bear no resemblance to each other. As a protein for building muscle, it loses the moment you pit it against whey. As a product for tendons and knees, trials find an effect — smaller than what the label says, but measured. Here are the two, separately.

As a protein: whey wins when they are compared head-to-head

The most direct trial gave 30 g of whey or 30 g of collagen peptides, twice a day for six days, to 22 women with an average age of 69. Whey increased muscle protein synthesis both at rest and after the session, markedly more than collagen. The latter only stimulated synthesis in the trained leg, in the short term, and nothing over time.

And the gap holds even when leucine is equalized. In a ten-week resistance training trial in 22 young untrained adults, both groups received 35 g of protein with the same amount of leucine (3 g). Whey still thickened the thigh more (+8.4% vs. +5.6%) and the biceps (+10.1% vs. +6.0%). Strength and power, on the other hand, progressed equally in both groups.

The same observation comes up in endurance athletes. In eleven athletes taking 20 g after exercise and 40 g at bedtime, myofibrillar protein synthesis was 13% higher with alpha-lactalbumin (a milk protein) than with collagen. And in 25 young men during a heavy training week, 2 × 15 g of collagen per day successfully raised glycine, proline, and hydroxyproline in the blood — but without increasing either muscle protein synthesis, nor even that of its connective tissue, which is what it is sold for.

“But trials show it makes you gain muscle?”

Yes, and they are real. In 53 men aged 72 with sarcopenia, 15 g per day for twelve weeks of strength training resulted in +4.2 kg of lean mass versus +2.9 kg with placebo, and +16.5 Nm of quadriceps strength versus +7.3 Nm. In 77 non-menopausal women, same dose, same duration: more lean mass, more grip and leg strength, and more fat mass lost than with placebo. In 97 untrained middle-aged men: +3.42 kg of lean mass versus +1.83 kg with placebo.

The point that changes everything: these three trials compare collagen to a placebo, not to another protein. They say that adding it is better than nothing. The useful question for you, who must choose what to put in your shaker, is "collagen or whey?" — and there, the equal-leucine trial gives the advantage to whey. The only one to have had all three groups, the one in middle-aged men, finds no difference between collagen and whey, but in an analysis that its authors themselves describe as exploratory.

Tendons: the place where it does something

It is this same review that brings the best news about collagen. Out of the eight randomized trials, all combined with training, supplementation (15 to 30 g per day, with at least 50 mg of vitamin C) increases tendon cross-sectional area and stiffness, with a high level of evidence.

Two trials in female soccer players show the magnitude. In 17 players, 30 g of hydrolyzed collagen after each session for ten weeks (with 500 mg of vitamin C in both groups): patellar tendon stiffness +18.0% vs. +5.1% with placebo. In 11 professional female players in pre-season: +15.4% vs. +4.6%. In the first trial, maximum strength and muscle thickness did not budge: collagen acted on the tendon, not on the muscle.

Sample sizes of 11 and 17 people: that is small, and it must be read as such. But the result goes in the same direction as the review, and it is the most consistent use with what collagen is.

Knees that hurt during exertion

In 139 athletes who had knee pain during activity, 5 g of collagen peptides per day for twelve weeks reduced pain slightly more than placebo (drop of 19.5 points vs. 13.9), and they had less need for other treatments. A second trial in 180 active young adults (18 to 30 years old) confirmed: −21.9 mm vs. −15.6 mm on the pain scale.

Look at the two figures together: the placebo already does almost as much. The gap attributable to collagen hovers around 6 points. It is statistically significant, confirmed by the doctors' examination in both trials — and modest.

Knee osteoarthritis

A 2019 meta-analysis finds a decrease in the total WOMAC score (−8.00) and the visual analog scale for pain (−16.57) with collagen. But upon closer inspection, only stiffness improves significantly; pain and functional impairment, taken individually, do not. A 2023 meta-analysis of four trials and 507 patients finds pain relief (standardized mean difference −0.58), with no more adverse effects than placebo — and itself warns that the included trials are prone to bias and better-designed ones are needed.

Undenatured type II collagen is a distinct product, taken at a very small dose. In a trial of 191 people, 40 mg per day for 180 days outperformed the placebo and the glucosamine-chondroitin combination on the WOMAC score, with good tolerance. Finally, a pilot study in 30 people showed favorable knee cartilage changes on MRI after 24 weeks of hydrolyzed collagen. Pilot, thirty people: a clue, not proof.

Every figure in this article comes from a study you can open: the underlined mentions lead to the summary of the work it is drawn from, written based on the abstract alone. Several trials involve small sample sizes. Informative content, which does not replace the advice of a healthcare professional: joint pain that lasts deserves to be examined.