Which type of collagen is best to take?

Which type is best for you depends on your goal. For joints, two types have been studied: native type II collagen (also known as UC-II) and hydrolysed collagen peptides. They work in different ways. UC-II is unprocessed and appears to suppress inflammation in the joint via the immune system. Hydrolysed collagen is broken down into small peptides that can reach joint tissue and thereby protect cartilage.

For knee complaints caused by osteoarthritis, UC-II (40 mg per day) performed well in an RCT with 191 participants: after six months, pain, stiffness and function improved significantly more than with placebo and better than glucosamine with chondroitin. That is a concrete positive signal. At the same time, an independently funded study in 68 participants found that a combination of UC-II and hydrolysed peptides did not work better than a placebo after 12 weeks. The evidence is therefore not clear-cut, and the study that rated UC-II positively was conducted in a setting that was not entirely free from industry interests.

For bones, specific research has been carried out into hydrolysed collagen peptides (5 grams per day). In an RCT with 131 postmenopausal women, bone density of the spine and hip increased measurably after 12 months compared to placebo. This is an interesting result, but it comes from a single study in a specific group.

For skin, low-molecular-weight collagen peptides (1,000 mg per day) appear to improve hydration, wrinkles and elasticity after 12 weeks, as shown in an RCT with 64 participants. Gelatine with vitamin C before exercise can also stimulate the production of new collagen, but this has only been demonstrated in a very small study with eight men. Vitamin C is in any case necessary for collagen production in your own body.

A caveat that applies to this entire field of research: multiple studies have been (co-)funded by manufacturers of collagen ingredients, including a review article written entirely by employees of a collagen producer. This increases the risk that positive results are presented as somewhat larger than they actually are. The only fully independent RCT we have here found no effect at all. That makes caution appropriate.