Collagen for Tendons and Joints: What the Evidence Actually Shows

Somewhere along the way, collagen stopped being a niche supplement and became a category. Powders, gummies, coffee creamers, protein bars. The pitch usually goes like this: your tendons and joints are made of collagen, so eating collagen rebuilds them.

That logic is a bit like saying your car is made of steel, so you should eat steel. Your digestive system breaks collagen down into amino acids like any other protein, and your body then uses those amino acids wherever it needs them. It does not route them to your Achilles tendon because you asked nicely.

But here is the interesting part. The research is not empty. It is just much narrower, and much stranger, than the marketing. There is a real mechanism, real trials, and real limits, and knowing which is which changes how you should think about spending money on it.

TL;DR: Collagen is a protein, and your body breaks it into amino acids before using it, so it is not a targeted repair kit. The most interesting evidence is about timing: 15 g of collagen or gelatin taken roughly an hour before loading exercise appears to increase collagen synthesis markers, and a 10-week trial in female soccer players found greater patellar tendon stiffness with 30 g collagen plus vitamin C after training. The effects are modest, the studies are small, and a 2026 systematic review rated the evidence for tendon and ligament tears as low certainty. If you use it, take 15 to 30 g with vitamin C about an hour before your loading session, and treat progressive loading, sleep, and enough total protein as the things that actually build tendon.

YouTube: Related video

What collagen actually is

Collagen is the most abundant protein in your body. It is the structural scaffolding in skin, bone, cartilage, and tendon. A tendon is roughly two thirds collagen by dry weight, mostly type I, arranged in long parallel fibres that transmit force from muscle to bone.

Tendons are famously slow to adapt. Compared with muscle, which can visibly change in weeks, a tendon remodels over months. Blood supply is low, metabolic rate is low, and the tissue responds to one specific signal: mechanical load. That is why tendinopathy rehab is built around progressive loading and not much else.

Collagen supplements come in a few forms. Hydrolysed collagen peptides are collagen broken into short chains, which dissolve in liquid. Gelatin is the same thing in a less processed form, and it sets into a jelly. Both deliver the amino acids glycine, proline, and hydroxyproline in unusual proportions compared with normal dietary protein.

The mechanism that got researchers interested is not "collagen goes to your tendon." It is this: a dose of those specific amino acids shows up in your blood about an hour later, and if you load the tendon at that moment, the building blocks are available while the tissue is signalling for repair. Timing, not targeting.

The study everyone cites

The paper that launched a thousand supplement brands came from Keith Baar's group and was published in the American Journal of Clinical Nutrition in 2017. Shaw and colleagues ran a randomised, double-blind, crossover trial in eight healthy young men. Each participant completed three conditions: 0 g, 5 g, or 15 g of gelatin in a drink containing 48 mg of vitamin C, taken one hour before six minutes of rope skipping. They repeated the skipping three times a day for three days.

The results were striking on paper. The 15 g dose produced roughly double the amino-terminal propeptide of collagen I (PINP) in blood at four hours compared with placebo, and the effect persisted across the 72 hours of the protocol. The engineered ligament experiments in the same paper showed increased collagen content when treated with serum from the gelatin group.

Two caveats matter, and they rarely make it into the marketing.

First, the sample was eight men. That is a mechanistic pilot, not a population study.

Second, and more importantly, PINP is primarily a marker of bone collagen turnover, not tendon collagen synthesis. The authors acknowledged this themselves, and a commentary published alongside the paper made the same point: the blood marker was likely reflecting bone, which turns over faster than tendon. The engineered ligament data were suggestive, but they came from a single donor's cells.

So the honest summary of that study is: a plausible mechanism with a small signal, not proof that your Achilles gets stronger. It was always a starting gun, not a finish line.

What the tendon trials show

Since then, researchers have done the harder work of measuring actual tendon properties.

The most useful trial for this question was published in Frontiers in Physiology in 2023. Researchers pair-matched 17 female under-21 soccer players from a Women's Super League academy and randomised them to 30 g of collagen hydrolysate plus 500 mg vitamin C, or an energy-matched placebo, taken after each training session. Both groups trained three days a week for ten weeks.

The results were modest but real. Patellar tendon stiffness increased 18.0 percent in the collagen group versus 5.1 percent in the placebo group. Young's modulus, a measure of the tissue's material quality, rose 17.3 percent versus 4.8 percent. Tendon cross-sectional area did not change in either group, and neither did knee extension strength or muscle thickness.

That is a genuinely encouraging finding, and it is also a small one. Seventeen teenage athletes, one tendon, ten weeks. The effect size is meaningful for a tissue that normally takes months to change, but it is a single trial in a specific population.

Other studies point in the same general direction with varying quality. A 2026 systematic review in the Journal of Functional Morphology and Kinesiology looked at collagen supplementation and tendon-related structural and performance outcomes and found inconsistent results across trials, which is exactly what you would expect for a small literature with different doses, forms, and loading programmes.

Where the evidence gets thin

If you have a tendon or ligament tear, or you are recovering from surgery, the picture is much less clear.

A 2026 systematic review in Frontiers in Medicine by Latini and colleagues searched for clinical evidence on hydrolysed collagen peptides in tendon and ligament injuries and post-repair settings. It found six eligible studies. Three provided direct evidence in tear or post-surgical populations, and three more provided indirect evidence in Achilles tendinopathy, chronic ankle instability, and rotator cuff tendinopathy.

The signal was cautiously positive: improvements in pain, function, analgesic use, or imaging outcomes. But the authors were blunt about the quality. Most oral interventions were multicomponent, meaning the collagen was bundled with other ingredients, so you cannot attribute the effect to collagen. Only one direct tear study used a clearly defined low-molecular-weight peptide formulation. Their conclusion was that the evidence is sparse, heterogeneous, and of low certainty, and that properly blinded placebo-controlled trials with well-characterised formulations are still needed.

Translated: there is a hint of benefit, and not enough evidence to build a treatment plan on.

The joint pain question

Joints are a different tissue again. Cartilage has almost no blood supply and an extremely slow turnover, and the outcome measures in joint supplement trials are notoriously soft: subjective pain scores, sometimes with small and inconsistent functional changes.

You will find meta-analyses reporting modest improvements in knee osteoarthritis pain with collagen hydrolysate. You will also find that the better-controlled the trial, the smaller the effect tends to look. If joint pain is your main concern, the evidence is weaker than it is for tendon loading, and the non-supplement interventions (strength training, weight management, activity modification) have far stronger support.

Dose, timing, and the vitamin C question

If you are going to try it, the research points to a specific protocol rather than a scoop in your morning coffee.

The timing logic is the part most worth keeping. The point of taking it before loading is to have the amino acids in circulation when the tendon is being asked to remodel. Taking it and then sitting at a desk for six hours is unlikely to do much.

One thing to be clear about: the vitamin C dose is not established. The 2017 study used a small amount because it came bundled in a drink. The 2023 trial used 500 mg. Vitamin C is a cofactor in collagen synthesis, so the rationale is sound, but nobody has shown that more is better.

What collagen will not do

A short list, because the marketing is long:

• It will not fix a tendon without loading it. Tendons respond to mechanical stress. A supplement is at best an additive to that signal, and the trials that show anything all pair collagen with a loading programme.
• It will not replace protein. Collagen is an incomplete protein, low in leucine and missing tryptophan. If you are using it, it should sit on top of your normal protein intake, not replace it. Our guide to how much protein per day covers the baseline.
• It will not make you recover faster from a hard week. There is no good evidence that collagen improves HRV, resting heart rate, or subjective recovery. Those track sleep, load, alcohol, illness, and stress far more closely.
• It will not rebuild cartilage in a meaningful way. Cartilage turnover is so slow, and the trials so soft, that any claim here is speculative.
• It will not hurt you at normal doses. Collagen peptides are generally well tolerated. That is a low bar, but it is a real one.

What actually builds tendon

If you strip everything else away, tendon adaptation comes down to four things, and collagen is at best a distant fifth.

Progressive loading. Tendons adapt to load, and they need enough of it to signal change without so much that they break down. For a tendinopathy, that means isometric or heavy slow resistance work, progressed over months. For a healthy athlete, it means strength training that actually loads the tendon, like calf raises, Nordic hamstring curls, and heavy squats.

Sleep. Tissue repair happens largely while you sleep, and short sleep is associated with worse recovery across the board. If you are choosing between a collagen powder and an extra hour of sleep, the sleep wins. Our breakdown of recovery tools and what the evidence says ranks these interventions honestly.

Total protein and energy availability. You cannot build connective tissue in a deficit if the deficit is large enough to shut down repair signalling. Adequate protein and enough calories are the boring foundation.

Time. Tendons remodel over months. A ten-week trial showing a stiffness change is genuinely fast for tendon tissue. Any programme that promises a fixed tendon in three weeks is selling something else.

A practical eight-week approach

If you want to try collagen with a decent chance of knowing whether it did anything, treat it as a structured block rather than a permanent habit.

1. Pick one tendon and one load. An Achilles with a history of niggles, a patellar tendon after a heavy squat block, or a hamstring origin. Choose a loading exercise you can progress and measure.
2. Set a baseline. Record your current tolerance: how many pain-free single-leg calf raises, what load on the exercise, and how the tendon feels the morning after training.
3. Take 15 to 30 g collagen with vitamin C, 45 to 60 minutes before the loading session. Two to three sessions a week is enough, and it matches the trial protocols.
4. Progress the load slowly. Add reps or weight every week or two. The supplement is not the intervention. The loading is.
5. Re-test at eight weeks. Same test, same conditions. If tendon tolerance has improved and your training load has also gone up, you have your answer, though you will never know for certain how much the powder contributed.
6. Decide. If nothing changed, drop it. It is not a foundation supplement like protein or vitamin D, and there is no reason to keep paying for it out of habit.

How to know if it is working

This is where wearable data helps, as long as you use it for the right question.

Collagen will not move your HRV. But the training block you pair it with will show up clearly. A loading programme that is working looks like this over four to six weeks: training load rising in steps, resting heart rate stable or drifting down slightly, HRV holding around its baseline, and sleep duration steady. A loading programme that is too aggressive looks like load rising while resting heart rate climbs and HRV drops below baseline, which is the pattern that precedes most overuse problems.

That distinction is the useful one, because tendon trouble usually arrives when load outpaces recovery, not when a supplement is missing. Century AI pulls your health score, recovery score, and sleep trends into one view so you can see whether a new loading block is being absorbed or whether you are accumulating fatigue. It is a much better use of your watch than trying to detect a supplement's effect.

The short version

Collagen is a protein, not a repair kit. The interesting research is about giving your body the raw materials at the moment you load the tissue, and the best trials suggest a modest improvement in tendon stiffness when collagen is combined with training. The evidence in tears and post-surgical recovery is low certainty, and the joint pain evidence is softer still.

If you want stronger tendons, load them progressively, sleep enough, eat enough protein, and be patient. If you want to add collagen on top, take 15 to 30 g with vitamin C about an hour before training, give it eight to twelve weeks, and measure something real rather than trusting how you feel about the powder.

Century AI helps you understand your body with a daily health score, recovery score, and sleep insights, using the watch you already wear.