HOHorseBmoderate
Around a quarter of colic cases seen by vets turned critical: which findings set them apart, which risk factors recur, and what the evidence shows.
Executive summary
Colic is not a single disease but a pain signal from the abdomen, with very different causes. In British surveys of the first veterinary examination, around a quarter of the classifiable cases were critical: needing hospital care or surgery, or fatal. Heart rate, pain severity, pulse quality, mucous membrane colour and absent gut sounds separated critical from other cases. The risk factors found most often were changes in feed and management, previous colic and crib-biting.
Key points
In the horse, colic describes pain in the abdomen. The word names a symptom, not a cause. Behind it can lie a build-up of gas, an impaction of the large colon, a displacement of sections of intestine, a strangulation that cuts off the blood supply, or inflammation. A 2023 review of the epidemiology of colic therefore states explicitly that colic is a syndrome made up of several disease processes and remains multifactorial in origin. That very variety explains why the same outward behaviour subsides without consequence on one occasion and turns life-threatening within a few hours on another.
The range can be put into figures. In a prospective British survey, 167 veterinary surgeons documented 1,016 primary examinations of horses showing signs of colic over thirteen months. Of the 822 cases that could be classified unambiguously, 76.4 per cent were resolved with simple medical treatment. The other 23.6 per cent were critical: they needed intensive medical treatment or surgery, died, or were put down.
In a strangulating lesion, the blood supply to a section of intestine is cut off and the tissue dies within hours. In an analysis of 896 Thoroughbred mares with surgically treated large colon volvulus, the duration of colic signs before admission was independently associated with survival to discharge. The median of that duration was two hours, and 88 per cent of the mares survived. That figure applies to a population presented very early and well looked after.
The signs reported are non-specific and overlap heavily between harmless and severe courses. In a British study of telephone triage, the practice teams surveyed named sweating and recumbency most often as pointers to a critical case. The same work also showed how inconsistent telephone advice turned out to be: movement and the safety of the owner or handler were the topics raised most often, and it was precisely on those that contradictory information was given.
Mild signs do not rule out a severe course. In a prospective survey of 120 large colon impactions seen at the primary examination, 67.5 per cent of the horses showed only mild signs of pain and 76 per cent had reduced gut sounds. Even so, 9.2 per cent of these cases ended critically, and 36.6 per cent needed more than one treatment. The distribution across the year was striking: 42.1 per cent occurred in winter, and 59 per cent of the horses had shortly beforehand experienced a change in management.
Only in part. In a survey of 1,564 horse owners, 45 per cent named the normal resting pulse correctly, 45 per cent the respiratory rate and 67 per cent the body temperature. In case scenarios put to them, 49 per cent confidently identified a surgical case, but only 9 per cent an impaction. There was no consensus on the behaviour at which the practice should be called.
The clearest picture so far comes from that same British survey of 1,016 primary examinations. Five features held up in the joint analysis as mutually independent pointers to a critical case: the level of the pain score, the heart rate, a capillary refill time over two and a half seconds, a weak pulse, and absent gut sounds in at least one of the four abdominal quadrants.
A second British analysis, of 941 colic cases seen out of hours, arrived at similar orders of magnitude on a different dataset: 23.9 per cent of the cases were critical, and 18 per cent of the horses were put down. In the joint analysis, a raised heart rate, the age of the horse and an altered mucous membrane colour remained associated with the critical outcome. The only feature the two analyses share is the heart rate; the remaining features differ between the two models. That two separate primary-care datasets both point to simple circulatory signs is among the more robust observations in this field.
| Finding | Observed association | Data basis |
|---|---|---|
| Raised heart rate | Independently associated with a critical outcome in both surveys | 1,016 primary examinations and 941 out-of-hours cases |
| Level of the pain score | The higher the score, the more often critical | 1,016 primary examinations |
| Capillary refill time over two and a half seconds | Associated with a critical outcome | 1,016 primary examinations |
| Weak pulse | Associated with a critical outcome | 1,016 primary examinations |
| Absent gut sounds in at least one quadrant | Associated with a critical outcome | 1,016 primary examinations |
| Altered mucous membrane colour | Associated with a critical outcome | 941 out-of-hours cases |
| Older age of the horse | Associated with a critical outcome | 941 out-of-hours cases |
The frequency figure quoted most often comes from a one-year prospective study on 31 horse premises with 1,427 horses in Virginia and Maryland. The crude incidence was 10.6 colic cases per 100 horse years. The median premises-level rate was 7 cases per 100 horse years, and the range between individual premises ran from 0 to 30. Three quarters of the episodes were mild, and 13 per cent of the horses had more than one episode in the same year.
The same study supplies the mortality figures. Per 100 horse years there were 0.7 colic deaths against 2.5 deaths from all causes combined. Put another way: 28 per cent of all deaths were due to colic, a higher share than for any other single cause. That is the real message of these numbers. Colic is mild most of the time, and it is at the same time the commonest single reason a horse dies.
Out of hours, colic dominates the caseload. In two British practices, 35 per cent of the 2,602 cases seen outside regular consulting hours were colic, ahead of wounds at 20 per cent and lameness at 11 per cent. Of all the out-of-hours cases, 13 per cent ended in euthanasia.
A context figure for the surgical route comes from the worldwide CEPEF4 survey of 47,396 general anaesthetics from 93 centres in 28 countries. The death rate within seven days was 4.2 per cent for colic procedures and 0.6 per cent for procedures without colic. Against the previous survey from 2002, the colic figure has fallen from 7.8 to 4.2 per cent. That is a figure for anaesthetic risk, not a prognosis for an individual horse.
The only systematic appraisal of the risk factors to date appeared in 2019. The working group screened 3,756 publications, retained 58 studies and assigned 22 categories of risk factor to three domains: horse, management, environment. For the narrower systematic part on management changes, 14 of 410 publications remained: one cohort study, eight case-control studies and five cross-sectional studies, carried out between 1990 and 2008. The largest blocks of evidence concerned feed changes and a recent change of housing.
The same work names the weakness of the field unusually plainly: 8 of the 14 included studies failed to meet the methodological criterion for handling confounders, and the methods used varied widely between them. The factors have therefore been found repeatedly, but rarely cleanly separated from other influences. Anyone reading them is reading associations, not demonstrated triggers.
It is the single factor found most often. In the prospective study on 31 premises, a change in concentrate feeding within a year went with an around three-and-a-half-fold higher odds of colic, and more than one hay change in the same year with roughly double. Quantity counted too: more than five kilograms of concentrate dry matter a day was associated with an around six-fold higher odds.
Time stabled and broken routine crop up just as repeatedly. A British case-control study of simple colonic obstruction and distension found crib-biting, increasing hours stabled, a recent change to the exercise programme, no worming with ivermectin or moxidectin in the preceding twelve months, and transport within the last 24 hours to be associated factors. The heaviest weights there fell on permanent stabling, crib-biting and recent transport.
For one particular and very severe form, the link with crib-biting is especially striking: entrapment of small intestine in the epiploic foramen, a narrow natural opening in the upper abdomen. In an international case-control study with 109 affected and 310 comparison horses, the odds for crib-biting horses were around 67 times higher in the point estimate; because of the small number of cases, however, that value remains very imprecise. A British parallel study with 77 cases and 216 comparison horses confirmed the direction and additionally found more time stabled in the preceding 28 days as well as a colic in the past year.
A previous colic is in fact the most stable single finding in the field. It raised the odds of a further episode by about three and a half times in the American cohort study and by about four times in the international case-control study on entrapment. Whether behind this lies a lasting anatomical or functional peculiarity of the individual horse, or simply a yard with a raised baseline risk, is an open question.
More recent work broadens the picture. An international case-control study from 2025 on strangulation by pedunculated lipomas compared 55 affected with 167 comparison horses. Alongside older age and certain breeds, two previously unknown factors stood out: a previous bout of laminitis and more time stabled in the four weeks beforehand. Weight management measures were associated with a lower risk. This is a single study with wide ranges of uncertainty.
The tapeworm Anoplocephala perfoliata sits at the junction between the small and the large intestine and has long been regarded as a colic factor. A Swedish case-control study with 67 horses in each group found an around sixteen-fold higher odds of colic where tapeworm eggs were detected in the droppings. The second part of the same result is notable: the same association could not be shown using the blood test.
A German case-control study on 620 hospital horses reached a different conclusion. Neither the antibody status against the tapeworm nor that against the large strongyle was related to colic status. Something else did stand out: horses wormed in the preceding week showed an around 2.4-fold higher odds of colic signs than horses whose treatment lay at least eight weeks back. Ponies were affected less often than warmbloods.
The contradiction can be explained, but it is not resolved. The studies measure in different ways: egg counts in the droppings, antibodies in blood, genetic detection. They examine different populations, and worming practice itself alters both sides of the equation. No cause follows from an association here, and no worming programme follows from any of these studies.
That parasites are not a historical topic is shown by a post-mortem series from Alberta. Among 510 horses examined between 2010 and 2022, 17.3 per cent had changes to the cranial mesenteric artery, 6.1 per cent active inflammation caused by migrating larvae of the large strongyle, and in 1.5 per cent of all cases this was the cause of death or euthanasia.
At a South African university hospital, 88 per cent of the 375 colic horses treated between 2019 and 2021 survived to discharge: 91 per cent of those treated medically and 77 per cent of those treated surgically. Lesions of the small intestine carried a worse prognosis than lesions of the large intestine.
For the period after discharge, an Italian follow-up over at least four years is available. Of 244 horses operated on, 203 were discharged and 176 followed up. Complications arose after discharge in 44.8 per cent, most often a further colic. The higher the complication score before discharge, the lower the long-term survival. Overall the authors describe long-term survival as favourable.
For one particular displacement, entrapment between the kidney and the spleen, a meta-analysis of 19 studies examined whether the choice of treatment alters survival. Neither the decision between a medical and a surgical approach, nor rolling versus trotting, nor the use of phenylephrine changed the probability of survival to a statistically meaningful degree. Since no randomised studies were available, this result should be read with caution; case selection remains decisive.
For small intestinal surgery, a 2024 review screened 3,024 publications in total. Of these, 104 met the inclusion criteria, but only 42 reported technique and outcome precisely enough to be analysed. Across all the anastomotic techniques described, short-term and long-term survival rates were consistently above 70 per cent. The authors explicitly rate the quality of this evidence as low and call for better-structured studies before techniques are compared with one another.
No proof follows from observational data, but a clear direction does. What has repeatedly gone together with colic are breaks in the routine: a change of concentrate, a change of hay, a sudden increase in time stabled, a journey, a change to the exercise programme. None of these associations has so far been tested as a preventive measure in a controlled trial. The conclusion that slower changes lower the colic rate is plausible and unproven at the same time.
On housing there is a pointer from Sweden. In a nine-month field study with 66 horses on the same premises, the frequency of colic was 0 per cent in the active open barn against 5 per cent in single boxes; lameness was also less frequent, at 18 against 26 per cent. In return, the horses in the active open barn more often had wounds from squabbles, without any more training days being lost. One yard, a convenience sample, no crossover between the systems: hypothesis-generating, not proof.
The second lever lies not with the horse but with people. A survey of 701 horse owners showed that most either had no intention of drawing up an emergency plan for colic or already considered their existing approach sufficient. Those who knew the British REACT awareness campaign put preventive recommendations into practice more often. A survey of 1,544 horse owners added to this: 58 per cent did not know how quickly irreversible damage to the intestine can develop, and simply providing evidence-based information barely changed their decisions.
Original analysis
| Factor | Best design found | Reported direction | Can be influenced day to day | What does not follow from it |
|---|---|---|---|---|
| Change of concentrate feed | Prospective cohort, largest block of evidence in the systematic review | Risk raised | Yes | That a slow change demonstrably lowers the risk |
| Change of hay | Prospective cohort | Risk raised | Yes | That the type of hay is itself the cause |
| Increasing time stabled, less turnout | Several case-control studies | Risk raised | Partly | That turnout alone protects |
| Crib-biting and windsucking | Several case-control studies | Risk strongly raised, especially for entrapment in the epiploic foramen | Barely | That stopping the behaviour lowers the risk |
| Colic in the past twelve months | Cohort and case-control studies | Risk raised | No | That the earlier colic causes the new one |
| Transport in the last 24 hours | Case-control study | Risk raised | Yes | That the length or type of the journey is what tips the balance |
| Tapeworm burden | Case-control studies with opposing results | Inconsistent | Yes | That worming on suspicion lowers the colic risk |
| Worming in the preceding week | Case-control study | Colic signs more frequent | Yes | That the treatment itself triggers the colic |
| Previous laminitis | Single case-control study from 2025 | Risk of strangulating lipoma raised | Partly | That preventing laminitis prevents colic |
| Dental disease | Cross-sectional studies in working horse populations | Risk raised | Yes | That the results transfer to European leisure horses |
| Restricted access to water | Single cross-sectional study without multivariable analysis | Risk raised | Yes | That the association is separated from other shortcomings in management |
Reliably, you cannot know without an examination, and that is the most important finding of the studies. In British practice data, heart rate, pain severity, pulse quality, mucous membrane colour and absent gut sounds separated critical from non-critical cases, all of them findings from the primary veterinary examination. Mild signs rule nothing out: among 120 large colon impactions, two thirds showed only mild signs of pain, and close to one case in ten still ended critically. In practice that means: observe, note what is changing, and ring the practice early if in doubt.
There is no evidence-based standard for this question. A British study of telephone triage found movement and safety to be the topics raised most often and at the same time found contradictory advice: owners were given different guidance depending on the practice. The study therefore led to the development of dedicated decision aids for practice teams. What follows from it is not a general rule but a recommendation on procedure: in a specific case this question belongs with the attending practice, ideally during the first phone call.
Yes, and that is the most stable finding in the whole field. In an American cohort study, a previous colic was associated with an around three-and-a-half-fold higher odds of another; in an international case-control study on small intestinal entrapment, with roughly four-fold. In the same cohort, 13 per cent of the affected horses had more than one episode in the year. After surgery, complications arose later in 44.8 per cent of the discharged horses, most often further colic. Cause and marker cannot be separated here.
Most episodes run a mild course. In the one-year farm study, three quarters of the cases were dealt with by no more than a single treatment. At the primary veterinary examination, 23.6 per cent were critical, and 23.9 per cent in the British out-of-hours setting, where 18 per cent of the horses were put down. Across all causes of death, 28 per cent of all deaths in the farm study were due to colic, more than to any other single cause. Colic is therefore usually harmless and still the commonest reason a horse dies.
The data are contradictory. A Swedish case-control study found an around sixteen-fold higher odds of colic where tapeworm eggs were present in the droppings; a German study on 620 hospital horses, by contrast, found no relationship between antibody status and colic status. In that same German study, horses wormed in the preceding week showed signs of colic more often than those whose treatment lay further back. A British study, conversely, found the absence of worming to be a risk factor. A worming programme therefore belongs with the practice, based on diagnostics rather than on the calendar.
A pointer, not a proof. In a Swedish field study with 66 horses over nine months, no colic occurred in the active open barn on the same premises, against 5 per cent of the horses in single boxes; lameness was also less frequent, at 18 against 26 per cent. In return there were more wounds from scuffles in the active open barn, without more lost training. A single yard, a convenience sample and no crossover between the systems: the result fits the case-control data on time stabled, but does not replace them.
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
Comparative Health
Radiocarbon data suggest the Achilles tendon core forms in growth and is barely renewed thereafter: what that means for healing, scar and the equine t
MUMultiple speciesBmoderateT2limited target data Open access
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
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.