MUMultiple speciesBmoderateT2limited target data
Radiocarbon data suggest the Achilles tendon core forms in growth and is barely renewed thereafter: what that means for healing, scar and the equine tendon.
Executive summary
Tendons consist largely of type I collagen. Radiocarbon measurements suggest that the core of the human Achilles tendon is essentially laid down during longitudinal growth and is barely renewed thereafter. The superficial digital flexor tendon of the horse counts as a functionally comparable energy-storing tendon. Both heal through repair tissue rather than genuine rebuilding: equine tendon scar was found to contain more type III collagen. That fits the high re-injury rates seen in both species.
Key points
Tendons are connective tissue built for tension. Most of their dry weight is type I collagen, which bundles up from fibrils into fibres and then into fascicles, the largest subunits of a tendon. Between those fascicles sits the interfascicular matrix: a softer layer that is rich in elastin in the equine tendons studied. Mechanical testing on excised equine tendons showed that it is not the fascicles themselves but precisely this intervening layer that decides how far a tendon can stretch and how completely it springs back.
A positional tendon transmits force and sets a joint. An energy-storing tendon does the same, and on top of that takes up elastic energy at foot strike and returns it at push-off. That lowers the energy cost of locomotion, but it demands strains close to the failure load. The human Achilles tendon and the superficial digital flexor tendon of the horse both belong to the second group: same job, same narrow safety margin.
In a direct comparison of two equine tendons with different jobs, the energy-storing tendon was the one with the lower values for collagen building instructions and for degradation markers, despite its higher cell density. The purely positional tendon from the same limb turned its matrix over noticeably faster. The authors consider it possible that this very slowdown in the energy-storing tendon favours the accumulation of damage: the tissue under the greatest strain is the one that renews itself least.
How fast tendon tissue turns over was an open question for a long time. Published estimates of the half-life of tendon protein ranged from two months to two hundred years, a difference of more than a thousandfold. A Copenhagen research group settled the question by a detour: the atmospheric nuclear weapons tests between 1955 and 1963 sharply raised the carbon-14 content of the air. Tissue laid down at that time carries the signature of its year of manufacture.
In 28 forensic samples of human Achilles tendon, the carbon-14 content measured matched atmospheric values from several decades before the samples were taken. On that calculation the tendon core is laid down in roughly the first 17 years of life, that is during longitudinal growth, and is essentially not replaced afterwards. Four skeletal muscle samples from the same bodies, by contrast, showed ongoing turnover. In the group's own assessment, the finding probably explains why tendons regenerate so poorly.
No. A synthesis of 27 intervention studies with 264 healthy adults in total found clear increases in tendon stiffness after at least eight weeks of raised loading, and smaller increases in cross-sectional area. What mattered was the magnitude of the load, not the type of muscle contraction. A broader analysis of 61 papers with 763 people reached the same result and showed in addition that the rise in stiffness comes mainly from altered material properties and only in small part from a thicker tendon. Core remodelling and measurable adaptation are therefore two different processes: the tendon becomes stiffer and somewhat thicker without its load-bearing scaffold being exchanged.
The same group later compared diseased and healthy tendons. In 25 tendinopathic Achilles tendons a substantial share of the collagen had been renewed in adulthood; in 10 healthy tendons it had not. The modelling suggested that roughly half of the collagen in the diseased tendon had already been slowly replaced years before the first symptoms. Whether this raised turnover is a very early phase of disease or a pre-existing risk factor cannot be decided from these data.
In comparative research, injury to the superficial digital flexor tendon in the sport horse counts as one of the best-supported animal models of a human disorder, namely overstrain injury of the Achilles tendon. A review from 2014 notes that no equivalent rodent or rabbit model exists for this function. The reason is simple: few species move on a tendon that works close to its material limit at the gallop.
The frequency is striking. In an analysis of 10,262 Japanese Thoroughbreds from a single year group, forelimb tendonitis affected 11.1 per cent of the horses, while the same injury in the hindlimb, with six cases, barely occurred at all. A British ultrasound series in 263 National Hunt horses found tendon changes in 24 per cent of the animals over two seasons, with differences between training yards of 10 to 40 per cent that the authors did not, however, class as established.
Because function, load and age coincide. The energy-storing tendon of the forelimb carries most of the load at the gallop, its remodelling is slow, and damage adds up over the years. In the Japanese analysis, horses aged three and over were far more affected than two-year-olds. What did not work is just as telling: routine ultrasound at three-month intervals did not predict the acute rupture in the British series.
Pull fascicles from an energy-storing tendon and from a positional tendon in the same equine limb, and the fascicles themselves barely differ. The difference lies in between. The interfascicular matrix of the energy-storing tendon allows more sliding between neighbouring fascicles and springs back more elastically, so it recovers better after repeated loading. With advancing age this layer became stiffer in the energy-storing tendon and tolerated loading cycles less well, while the fascicles remained largely unchanged.
A key reason for that is elastin. Biochemical and immunohistological work localised elastin specifically to the interfascicular matrix of the energy-storing tendon; with age its quantity fell and its arrangement became more irregular. In a further experiment on excised equine tendons, targeted enzymatic degradation of elastin lowered the recovery capacity and the fatigue resistance of exactly that layer, without measurably changing the mechanics of the fascicles.
As early as 1980, a study of partially ruptured equine tendons showed the decisive difference: normal tendon tissue consists almost entirely of type I collagen, whereas the scar tissue additionally contained 20 to 30 per cent type III collagen. Type III forms thinner fibrils that bear less load. The repair cells also resembled myofibroblasts more than tendon cells, which suggests that they do not come from the tendon itself but have migrated in from the surrounding tissue.
More recent work supports this picture. In an experiment on isolated tendon fascicles from rat tails, a single strain above two per cent was already enough to degrade the material properties, while structurally visible collagen damage appeared only from six per cent upwards. Over several days in culture, the tendon core repaired that damage barely at all. The authors conclude that the resident cells mainly break tissue down and send signals outwards instead of rebuilding it themselves.
Because the repair tissue matches the original neither structurally nor mechanically. A review from 2022 describes the typical failures: inflammation that does not settle in time, excessive ingrowth of blood vessels, and immature tissue that never reaches the mechanical competence required. Without appropriate mechanical stimuli the fibre alignment stays disordered, and the cross-linking that gives a tendon its stiffness forms only incompletely.
In the horse, implantation of the animal's own bone marrow cells into the injured tendon has been followed for years. In a follow-up of 113 racehorses over at least two years after their return to full work, the re-injury rate was 27.4 per cent, and 25.7 per cent among the National Hunt horses. That was lower than in previously published comparison groups, though without a concurrent control group. A small randomised trial in 12 horses with career-ending injury found a better-ordered fibre structure in the treated tendons after six months.
A systematic analysis from 2024 pooled 21 studies of cell and plasma treatments in the horse. Cell treatments did not raise the rate of return to performance, but they did lower the risk of re-injury. The authors explicitly stress the weak quality of the studies. A comparative review of 123 papers on platelet-rich plasma in horse and human found positive effects mostly in studies at high risk of bias, and absent effects more often in the methodologically better ones.
In humans the evidence base is larger, but no more satisfying. A synthesis of 33 studies with just under 36,000 people found that after acute Achilles tendon rupture, surgery brought fewer re-ruptures and more frequent return to sport, but also more infections and more nerve injuries. An analysis of 14 randomised trials with 1,628 participants showed a similar pattern: fewer re-ruptures after surgery, but more complications after the open procedure. Platelet-rich plasma changed the outcomes in neither treatment pathway.
For the painful, unruptured tendon, loading itself does the work. A randomised trial in 58 people with chronic Achilles tendon complaints compared eccentric training with heavy slow resistance training: both clearly improved function and pain, the effect held at 52 weeks, and no difference between the approaches emerged. An analysis of 21 randomised trials with 994 people supports this picture: against other forms of training, eccentric training was equivalent. For shockwave therapy, by contrast, a review at the Achilles tendon found only small and inconsistent effects, and a network analysis of non-surgical approaches could not name a best option.
The mechanism transfers. Both tendons store elastic energy, both have a core that is remodelled very slowly, both heal through scar tissue, in both the risk rises with age, and in both part of the specialisation sits in the matrix between the fascicles. This pattern is consistent across independent measurement methods and across several species, which makes a shared biology far more likely than coincidence.
The clinical picture as a whole does not transfer. The review from 2014 states explicitly that the equine tendon is a model for the acute rupture in athletes, not for the full spectrum of human tendon complaints and in particular not for chronic tendon pain. The loading profiles also differ fundamentally: a racehorse carries half a tonne on four limbs at the gallop, a human runs upright on two.
Three observations hold up to scrutiny. First, in the adaptation of healthy tendons the magnitude of loading appears to be the more effective lever than the chosen form of exercise: the synthesis of intervention studies found an association with intensity, not with contraction type, and the head-to-head comparison of two training forms in symptomatic tendons produced no difference. Second, prevention weighs more heavily than repair, because on current knowledge the tendon core is not built anew.
Third, supplements are no substitute for loading. A synthesis of 19 studies with 768 adults found favourable effects on tendon morphology under collagen peptides combined with regular training, but rated its own certainty in that statement as very low. Without a training stimulus, nothing comparable was shown in this body of work.
Original analysis
| Method | Quantity measured | Time horizon | Blind spot | Species in the cited studies |
|---|---|---|---|---|
| Carbon-14 bomb pulse | Year of formation of the collagen in the tendon core | Lifelong, decades | Says nothing about the current synthesis rate | Human |
| Messenger RNA for collagen I and III | Current building instructions of the cells | Hours to days | No evidence that the protein is actually incorporated | Horse |
| Degradation markers of type I collagen | Ongoing breakdown of the existing matrix | Days to weeks | Does not separate breakdown from new formation | Horse |
| Tissue autofluorescence | Non-enzymatic cross-linking, indirectly matrix age | Years | Non-specific as to the cause | Horse |
| Collagen hybridising peptide | Mechanically unfolded collagen triple helix | Immediately after loading | Does not capture functional damage below the detection threshold | Rat, outside the body |
| Ultrasound cross-sectional area | Size and echo pattern of the tendon | Weeks to months | Did not predict the acute rupture in a two-year series | Horse |
| Stiffness measurement in the living | Mechanical response of the tendon to force | Weeks to months | Separates material change from shape change only indirectly | Human |
On current knowledge, not in the sense of a genuine new build. The load-bearing core of the human Achilles tendon is laid down during longitudinal growth and is essentially not replaced afterwards. After an injury, repair tissue forms; in equine tendons it contains 20 to 30 per cent type III collagen, whereas healthy tendon tissue consists almost entirely of type I. This scar tissue bears less load. Loading can make a tendon stiffer and somewhat thicker, but it does not exchange the scaffold.
Because the two tissues handle their matrix in fundamentally different ways. In the same bodies in which the tendon core held collagen decades old, skeletal muscle showed ongoing turnover. Tendon tissue is also poor in cells and poorly supplied with blood. In a laboratory experiment on fascicles from rat tail tendons, mechanical damage went essentially unrepaired over several days: the resident cells tended to break tissue down and send signals outwards instead of rebuilding it themselves.
The evidence is weak and the conclusion uncertain. A synthesis of 19 studies with 768 healthy adults found favourable effects on tendon morphology under collagen peptides combined with regular training. The authors themselves, however, rated their certainty in exactly that statement as very low. Without accompanying training, nothing comparable was shown. This text deliberately names no amounts and gives no recommendation.
Functionally very similar, clinically not identical. Both are energy stores that take up energy at foot strike and give it back at push-off, both work close to their material limit, both have a slowly remodelled core and heal through scar. A review from 2014 states explicitly, however, that the equine tendon is a model for the acute rupture in athletes, not for the whole spectrum of human tendon complaints and in particular not for chronic tendon pain.
Because damage accumulates and the matrix between the fascicles ages. In an analysis of 10,262 Japanese Thoroughbreds, horses aged three and over had a markedly higher risk than two-year-olds. At tissue level, the interfascicular matrix of the energy-storing tendon became stiffer and more prone to fatigue with age, and its elastin decreased and arranged itself more irregularly. That explains the link plausibly, but it is a laboratory finding and not an individual risk marker.
That question is decided by the treating doctor, not by an article. The evidence shows a trade-off: a synthesis of 33 studies with just under 36,000 people found fewer re-ruptures and more frequent return to sport after surgery, but also more infections and more nerve injuries. An analysis of 14 randomised trials with 1,628 participants likewise showed fewer re-ruptures after surgery, but more complications after the open procedure. Platelet-rich plasma improved the outcomes in neither group.
Comparative Health
Serum amyloid A rose about 227-fold in an equine experiment, haptoglobin only 1.1-fold. How inflammation starts, how it ends, what CRP and SAA measure
MUMultiple speciesBmoderateT2limited target data Open access
Equine Health
What counts is not visible dust but the respirable fraction in the horse's breathing zone: figures from trials on hay, steaming and haylage.
MUMultiple speciesBmoderateT1hypothesis only Open access
Evidence Lab
Evidence levels grade confidence, not truth: what randomised, observational and animal studies support, how GRADE rates them and where the pyramid bre
MUMultiple speciesBmoderateT0not transferable Open access
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.