Collagen Before and After: What Really Changes in Skin, Joints, Nails and Hair, and When, According to the Trials

Written by Dimitrun Irene. Updated October 26, 2026.

A real collagen before and after looks like this: at eight to twelve weeks, skin that is measurably more hydrated and a few percent more elastic, with crow's feet that are slightly shallower in a 3D scan and not obviously different in a bathroom mirror; at twelve to twenty-four weeks, knees that ache less after exercise; at six months, nails that break less and grow a little faster. That is what the controlled trials show when they photograph, scan and measure people rather than ask them. It is a smaller, slower and more consistent story than the side-by-side photos that dominate search results, most of which involve different lighting, a tan, and a brand to sell.

This article lays out what changes and when, with the numbers from the studies, what a realistic after looks like for skin, joints, nails, hair and body composition, why most before-and-after photos cannot be trusted, and how to run an honest version on yourself.

The classic pair of photos has four problems that have nothing to do with collagen. Lighting: an overhead bathroom light in the before, a ring light in the after, and every wrinkle on a face changes depth. Hydration and time of day: skin is plumper after a glass of water and in the morning. Makeup, filters and a tan, which can be absent in one shot and present in the other. And selection: a brand publishes the dozen customers whose photos look best and nobody sees the rest. The honest way to measure skin is with a corneometer for hydration, a cutometer for elasticity and a 3D profilometer for wrinkle depth, which is what the trials below used. The honest way to measure your own result is described near the end of this article.

The table combines the main randomized, placebo-controlled trials and the largest meta-analysis. Doses, products and populations differ, so treat the numbers as the size of effect you can reasonably expect rather than a promise.

The sizes are modest and the direction is consistent, which is the profile of a real but small effect. For a sense of why the timing is what it is, see how long collagen peptides take to work.

The best-known skin trials, by Proksch and colleagues in 2014, gave women aged 35 to 55 either 2.5 g or 5 g of a specific collagen peptide or placebo for eight weeks. Skin elasticity, measured by a suction probe, rose about 7% more in the collagen groups than placebo, with the biggest difference in women over 50. A companion trial with 2.5 g measured eye wrinkles with a 3D scanner: wrinkle volume fell by around 20% relative to placebo at eight weeks, and skin procollagen content rose. The 2021 meta-analysis pooling 19 trials found consistent improvements in hydration, elasticity and wrinkles after about 90 days.

Translate that to a mirror and it means skin that feels less tight and papery, makes up more smoothly, and in a side-by-side scan shows slightly shallower fine lines. It does not mean fewer deep wrinkles, a lifted jaw or a different face. People who see dramatic photo changes at twelve weeks have usually also started drinking more water, wearing sunscreen or sleeping better, which is fine, but it is not the collagen. Collagen peptides work by signalling the skin's fibroblasts to make more of their own collagen, and that process needs vitamin C; the detail is in collagen and vitamin C.

Joint benefits take longer and are measured differently: by pain scores during and after activity. In a 24-week trial of 147 college athletes, 10 g of collagen hydrolysate a day reduced joint pain at rest, when walking, standing and carrying, compared with placebo, with the clearest effect in those with knee pain. A 12-week trial in young active adults with activity-related knee pain found the same at 5 g. In people with mild to moderate knee osteoarthritis, trials over 3 to 6 months have shown small reductions in pain and stiffness scores. The mechanism appears to be stimulation of cartilage cells to produce matrix, which is slow; expect nothing for the first month and a noticeable difference, if any, between weeks 12 and 24. Training and recovery specifics are in collagen peptides for joints and exercise.

Nails are the cleanest result in the collagen literature because they can be measured with a ruler. In a 24-week open-label trial, 2.5 g of collagen peptides daily increased nail growth rate by 12% and cut the frequency of broken nails by 42%, with most participants reporting less brittleness and peeling; the benefit persisted four weeks after stopping. The limitation is that the trial had no placebo group. Hair is a different matter. There is no randomized trial showing collagen peptides increase hair density, thickness or growth. Hair is made of keratin, not collagen, and the amino acids in collagen are not special for keratin synthesis. People who notice thicker hair are usually also eating more protein overall or correcting a deficiency. The evidence for each tissue is weighed in collagen peptides for skin, hair and nails.

The before-and-after that gets the most attention online, and deserves the most scepticism, is the body transformation. Collagen peptides are protein, about 9 g per 10 g scoop, but they are low in leucine and lack tryptophan, so they are a poor muscle-building protein compared with whey or a balanced diet. In the one well-known trial, 53 older men with sarcopenia who did resistance training three times a week for 12 weeks gained more fat-free mass and lost more fat with 15 g of collagen a day than with placebo. Without the training, there is no evidence collagen changes body composition at all. If building muscle is the goal, the comparison in collagen peptides vs protein powder is the one to read.

If you want to know whether collagen is doing anything for you, set up the test before the first scoop. Photograph your face in the same spot, same time of morning, no makeup, same phone, same lighting, with the window behind the camera; mark the floor where you stand. Pinch-test the back of your hand and count how many seconds it takes to flatten. Note how long your nails take to grow past a mark on the nail, and how many break in a month. Rate any joint pain from 0 to 10 after your usual activity. Then take 2.5 to 10 g daily, as covered in how to take collagen peptides, and repeat the measurements at 8, 12 and 24 weeks without looking at the earlier ones first. Change nothing else for the first twelve weeks, or you will not know what did what. If nothing has moved by week 24, collagen is not doing anything measurable for you, and the money is better spent on sunscreen, protein and sleep.

Collagen peptides do not build a permanent reserve; they signal cells to make more collagen while they are present. In the Proksch trial, the elasticity gain was still detectable four weeks after stopping and then faded. Nails revert over a growth cycle. Joint benefits fade over months. There is no evidence that cycling on and off is necessary or helpful, and no evidence of harm from continuous use; the question is covered in do you need to cycle collagen peptides.

If hair is your main goal, read what the evidence says about collagen for hair growth before judging your before-and-after photos.

A trustworthy collagen before and after is measured, not photographed: skin a few percent more elastic and hydrated at 8 to 12 weeks, eye wrinkles slightly shallower on a scanner, knees less sore by 12 to 24 weeks, nails stronger by six months, and no proven change to hair or, without training, to body composition. The effect sizes are small, consistent across many trials, and reversible when you stop. If a photo shows a transformation bigger than that, something other than collagen changed. If you want to know what it does for you, measure yourself the same way the trials did and give it the twelve weeks the data say it needs. Choosing a product that matches the trial doses is covered in how to choose collagen peptides.

Collagen peptides are a dietary supplement, not a medicine. This article summarizes published research for general information.