MUMultiple speciesBmoderateT0not transferable
At gallop a foreleg carries 14.0 newtons per kilogram: why in the horse it is the supporting limb, not the fracture, that decides the outcome.
Executive summary
It is not the bone that decides, but the setting in which it would have to heal. A horse weighs several hundred kilograms, stands day and night on four limbs and cannot lie down for weeks. At gallop, peak forces of 14.0 newtons per kilogram of body weight were measured on a single foreleg. If it offloads one limb permanently, the hoof of the opposite limb is put at risk. That chain limits the outlook, not bone healing itself.
Key points
The question usually comes up after an accident on the racecourse or in competition, and the common answer is that horse bone simply does not heal. That is not so. A horse's bone heals. What decides the outcome is the setting in which it would have to heal: an animal of several hundred kilograms standing on four columns, one that cannot lie down for weeks and that, after a general anaesthetic, wants to get back on its feet under its own power before it can judge the situation.
Four conditions interlock here. Each of them has been studied separately, and none of them concerns fracture healing itself.
This article sets out the evidence on these four points and holds it against the human situation. It says nothing about any individual horse and passes no judgement on a decision taken on the spot by a vet together with the owner or keeper.
The load at gallop has been measured directly. In a study of seven ridden Thoroughbreds crossing force plates, the peak vertical force reached 14.0 newtons per kilogram of body weight on the non-leading foreleg, 13.6 on the leading foreleg, and 13.6 and 12.3 on the hind limbs. A second result of the same work deserves attention: the load was split roughly evenly between the forehand and the hindquarters and shifted with speed, not in the often quoted sixty-to-forty ratio. The authors suggest that an even distribution across four limbs limits fatigue.
For the question of a broken leg, though, it is standing that counts most. In an experiment with thirteen Standardbreds, the undisturbed animals carried a median of 27.3 per cent of their body weight on the foreleg under observation. In the horses in which a plate shoe on the opposite limb forced an offloading posture, the figure was 38.7 per cent. Mean body weight in that group was around 458 kilograms.
Only for a short time. Offloading one limb shifts the load onto the one opposite, which then stands under raised pressure day and night. Horses do lie down, but only in short spells. Bed rest over weeks, of the kind people know, does not exist in a stable.
The complication with the gravest consequences strikes not the broken limb but the healthy one beside it. Inside the hoof a fine layer of tissue, the lamellar apparatus, interlocks the pedal bone with the hoof capsule. When that attachment fails, the pedal bone sinks or rotates: that is laminitis. When it appears in the limb opposite a painful injury, it is called supporting-limb laminitis.
A review from the group that supplies most of the experimental data describes the process in two steps. First the constant alternation between loading and unloading that normally goes with moving about is missing, and perfusion of the lamellae suffers. Then the extra load tips the already weakened tissue into acute laminitis.
Several measurements exist for the first step. In an experiment with thirteen Standardbreds, a marker of oxygen shortage reached a value of 196 in the lamellar tissue of the more heavily loaded limb, against 42 in the comparison horses; outside the loaded hooves nothing of the sort was measurable. In a randomised experiment with nine Standardbreds, measures of lamellar perfusion improved as soon as the load changed frequently: thirty minutes of walking was enough. An hour of one-sided standing, by contrast, changed nothing measurable, and an hour says nothing about several days.
Imaging published in 2026 in eight healthy horses found, during quiet standing, areas of the inner sole in all eight animals and areas of the lamellae in seven of eight that did not take up a marker of perfusion. Those areas were absent once the horses had been allowed to move beforehand. The authors state expressly that this does not establish a link with supporting-limb laminitis in affected horses.
Probably both, and standing still weighs more heavily than long assumed. In an experiment from 2025, restricted freedom of movement without any extra load was enough to multiply dying cells in the lamellar tissue: 45 instead of four per lamella. The group concludes from this that movement has to be preserved, not merely load avoided.
Two figures from the same world appear to contradict each other. In the caseload of a British practice with 65,327 registered horses, eleven cases of supporting-limb laminitis were found over nine years, that is 0.02 per cent. They appeared between four and a hundred days after the injury, at a median of fourteen and a half days, and close to three quarters of those affected were Thoroughbreds. The authors also record that the complication was not confined to horses that had stopped bearing any weight at all on the injured limb.
Look instead only at horses that really are sparing a limb, and the figure becomes large. In an analysis of 113 horses in casts, fourteen developed supporting-limb laminitis, that is twelve per cent. Two features went with it: body weight and the duration of the cast. Whether a fracture was present made no difference. Any painful one-sided lameness can lead to it.
The second point at which fracture treatment fails lies not in the stable but in the recovery box. After a general anaesthetic a horse gets up under its own power, often before it can judge the situation, and the freshly screwed limb immediately takes the full mass.
How much that step weighs is shown by a worldwide survey of 47,396 general anaesthetics from 93 centres in 28 countries, in which the Vetsuisse Faculty in Zurich also took part. Mortality within seven days was 1.2 per cent overall and 0.6 per cent for procedures other than colic, in both cases lower than in the preceding survey of 2002. Among the 227 deaths outside colic, fractures came first at 35.7 per cent, ahead of abdominal complications at 18.1 per cent. The authors record that these fractures arose predominantly during the recovery phase, and that healthy horses were among them.
That figure describes the anaesthetic risk of a mixed population, not the outlook for a horse with a freshly operated fracture. Centres reported voluntarily, which over-represents well equipped clinics and, if anything, understates true mortality.
The claim that a broken leg means the end for a horse is, in that generality, out of date. What decides is not the name of the bone but the form of the break: whether it stays closed, whether it splits into few or many pieces, and whether it can be screwed together stably.
Standing surgery shows this most clearly. In an analysis of 245 racehorses operated on standing under sedation between 2007 and 2021 for a longitudinal fissure of the pastern or the cannon bone, 98 per cent left the clinic alive, and around three quarters raced again, at a median of 241 days after the procedure. Earnings per start, the proportion of winners and the proportion of placed horses did not differ meaningfully before and after surgery. These are two selected fracture forms at a specialist centre.
For pastern joint arthrodesis, a meta-analysis of 21 studies covering 458 horses sums up: 90 per cent survived, 65 per cent returned to their intended use, 12 per cent developed a wound infection. That separation is the real point. Surviving and being usable again are two different outcomes, and around twenty-five percentage points lie between them here.
The same operation on the same bone can succeed almost every time or fail almost every time depending on the patient. In a case series on complete diaphyseal fractures of the fore and hind cannon bone, three of ten adult horses survived, but ten of eleven foals. In open fractures the gap was sharper still: one of eight adults against six of seven foals. The authors name body weight, age and infection as the decisive quantities and mark a weight above 320 kilograms in their series as a risk factor.
Infection is no marginal problem here. In an analysis of 155 horses with internal fixation, 14.2 per cent developed a surgical site infection. The affected animals left the clinic alive twelve times less often than the rest. Local antibiotic delivery could not be shown to be protective in this dataset.
History counts too. An analysis of 115 post-mortem reports from Californian racehorses with a complete tibial fracture found a pre-existing stress fracture in most of them; 68 per cent of the breaks happened in training rather than in a race, and 73 per cent in two- and three-year-old horses. A meta-analysis of risk factors in flat racing puts together a frequency of around 1.2 catastrophic musculoskeletal injuries per 1,000 starts and names age, race class, going, previous injury and findings at the pre-race examination as repeatedly associated factors.
In people, comparable bone tissue heals under entirely different conditions. A Cochrane review from 2024 brings together 53 studies with 4,489 adults after surgically treated ankle fracture. Early weight bearing within three weeks probably improved function slightly over the first six months, without raising the frequency of further procedures. The difference remained small, however, and did not consistently reach a clinically meaningful threshold.
What is taken for granted in that work is impossible in the horse. A person chooses the timing and the dose of their loading, walks on crutches, sits, lies down and takes a splint off. The horse can do none of this. The question of whether to load early or late therefore does not transfer between the species: here it is not a treatment question but a question of species.
Then the figures move closer together. A meta-analysis of eighteen studies on non-operated hip fracture in frail older people found a mortality of 36 per cent at thirty days and 60 per cent at one year; at six months just under ten per cent were mobile again.
This comparison has a limit that has to be named. The non-operated people in that analysis were the frailest patients of all, so their mortality mixes the consequences of the missing operation with their starting condition. The direction still reads clearly: where load-bearing function cannot be restored, a fracture becomes a life-threatening situation in people too. The horse is no special case in this. It only reaches that point earlier and more often.
No rule for an individual horse follows from this evidence. It explains why the weighing-up after a broken leg looks different from the one in human medicine, and it explains just as much why that weighing-up has shifted: selected fracture forms can now be treated with a high probability of success, others still cannot.
What the studies do not supply is a yardstick by which an outsider could judge whether a particular decision was right. Fracture form, age, weight, temperament, level of pain and the possibility of weeks of aftercare all enter into it together, and none of the publications cited here weighs them against one another.
Original analysis
| Situation | Population | Reported outcome | What does not follow | Design |
|---|---|---|---|---|
| Standing osteosynthesis, longitudinal fissure of the pastern or cannon bone | 245 racehorses, one centre, 2007 to 2021 | 98 per cent left the clinic alive, 75 per cent raced again, at a median of 241 days | Nothing about open or shattered fractures: the fracture forms were pre-selected | Retrospective cohort |
| Pastern joint arthrodesis | 458 horses from 21 studies | 90 per cent survived, 65 per cent returned to their intended use | That survival and usability are the same thing: around 25 percentage points lie between them | Meta-analysis of retrospective series |
| Complete diaphyseal fracture of the fore or hind cannon bone | 10 adult horses and 11 foals | 3 of 10 adults and 10 of 11 foals survived | That the technique fails: body mass, age and infection separate the groups | Case series |
| Disruption of the suspensory apparatus, double arthrodesis | 26 racehorses, three clinics | 13 of 26 were usable at pasture without painkillers beyond six months | That the rest stayed free of complications: 77 per cent had short-term complications | Case series |
| Infection after internal fixation | 155 horses, one centre, 2008 to 2016 | 14.2 per cent developed a surgical site infection | That infection merely delays healing: those affected left the clinic alive twelve times less often | Retrospective cohort |
| Supporting-limb laminitis after a cast | 113 casted horses at a referral clinic | 12 per cent developed laminitis in the supporting limb | That a fracture determines the risk: body weight and cast duration did | Retrospective cohort |
| Supporting-limb laminitis in a practice caseload | 65,327 horses in a British practice, 2005 to 2013 | 11 cases, equal to 0.02 per cent | That the complication is equally rare when a limb is already injured: the denominator is the whole caseload | Cross-sectional |
| Deaths around general anaesthesia | 47,396 anaesthetics, 93 centres, 28 countries | 0.6 per cent mortality within seven days excluding colic, of which 35.7 per cent from fractures | That the same risk applies to a horse with a freshly operated fracture: the population is mixed | Prospective cohort |
Yes, and for certain fracture forms it is now the rule. Of 245 racehorses operated on standing under sedation for a longitudinal fissure of the pastern or the cannon bone, 98 per cent left the clinic alive and around three quarters raced again. For a complete diaphyseal fracture of a large bone shattered into many pieces the picture reverses: in one case series three of ten adult horses survived, but ten of eleven foals. The outcome hangs on the form of the break, on body weight, on age, and on whether the repair holds stably.
Because the weighing-up rarely concerns the break alone. A horse cannot lie down for weeks, it shifts the load onto the opposite limb and puts that hoof at risk. After a general anaesthetic it has to get up under its own power, and in a worldwide survey fractures were the commonest cause of death outside colic at around 36 per cent. Add a fracture form that cannot be screwed together stably, and what remains is weeks of painful treatment with a poor outlook. That weighing-up is done by the attending practice together with the owner or keeper, not by a statistic.
A disease of the healthy limb, not the broken one. When one limb carries more weight for days to weeks because the other one hurts, the attachment between the pedal bone and the hoof capsule suffers inside the hoof. In experiments with healthy horses a marker of oxygen shortage rose markedly in the more heavily loaded limb, and restricted freedom of movement on its own already raised the number of dying cells in the lamellar tissue. In the caseload of a British practice the complication was rare in absolute terms at 0.02 per cent; after a cast, by contrast, it occurred in twelve of every hundred horses.
No, and that is the commonest misconception around this question. The bone itself heals. What is missing are the conditions: a person can offload a fracture, walk on crutches, sit and lie down, and choose when to bear weight. A Cochrane review of surgically treated ankle fractures compares exactly such loading plans, and the horse has no such choice. Then there is the mass: at gallop, peak forces of 14.0 newtons per kilogram of body weight were measured on one foreleg, and even in quiet standing a single foreleg carries about a quarter of the body weight.
Because the cast itself carries a risk. In an analysis of 113 casted horses, twelve per cent developed supporting-limb laminitis, and two quantities went with it: body weight and the duration of the cast. Full-limb casts and those with transfixation pins loaded the limb more than half-limb casts. A cast also immobilises exactly the movement that, according to the available experiments, keeps perfusion going inside the hoof. For some fractures it is the right choice, but it is not a harmless interim measure.
In the care, yes; in the outcome, not always. As long as surgery is possible, the human outlook is incomparably better. Where load-bearing function is not restored, that changes: in a meta-analysis of eighteen studies on non-operated hip fractures in frail older people, mortality at one year was 60 per cent, and at six months just under ten per cent were mobile again. Those people were, however, the frailest patients of all, which is why the share attributable to the missing operation and the share attributable to their starting condition cannot be separated.
Terms defined in this text
Comparative Health
Ten of sixteen clinically unremarkable horses stayed under 30 minutes of REM sleep: measured lying times and what sleep loss actually shows.
MUMultiple speciesBmoderateT1hypothesis only Open access
Comparative Health
Of 107 untreated sarcoids, 1.9 per cent regressed: what links bovine and human papillomaviruses, and where the comparison ends.
MUMultiple speciesBmoderateT1hypothesis only Open access
Equine Health
In 392 Quarter Horses, 57 per cent of the healthy horses carried a P variant: what the two assessments of P2, P3 and P4 actually found.
HOHorseBmoderate Open access
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.