HOHorseCweak
Five case-control studies looked for a disturbed gut flora in free faecal water, four found none: what is established and what has never been measured.
Executive summary
Free faecal water is not a disease in its own right but an observation: droppings are passed in two phases, one liquid and one formed. No cause has been established. Five case-control studies covering 259 horses examined an altered gut flora, and four found no difference from stablemates without the sign. Feeding, faecal chemistry and fibre length have likewise failed to separate affected horses reliably from comparison horses.
Key points
Free faecal water is not a disease in its own right but an observation: the horse passes its droppings in two separate phases, one dark and liquid, the other formed. The liquid escapes before, during or after defecation, sticks to the tail and hindquarters and irritates the skin. Unlike diarrhoea, the balls of dung themselves stay formed. The American study of 2025 therefore excluded horses with persistently soft droppings.
Frequency has been surveyed once for Switzerland. In a telephone survey of 1,861 Swiss warmblood horses aged six to sixteen, 9 per cent of keepers reported free faecal water, against 34 per cent for lameness, 22 for colic and 13 for back problems. These figures come from the keepers' recollection, without veterinary confirmation and without a fixed case definition: an order of magnitude, not a measured prevalence.
In an online survey of 339 affected horses, breeds, ages, disciplines, coat colours, housing systems and feeding strategies were spread across every category recorded. The authors conclude that feeding may play a part but that its distribution does not on its own explain why the sign occurs. These data yield no risk profile that would place an individual horse.
The most common explanation holds that the gut flora of affected horses is out of balance. The idea has been tested five times, in Europe and North America, each time against unaffected horses on the same yard: 122 affected against 137 comparison horses. Four of these studies found no meaningful difference in the composition of the faecal flora. One found a difference, and its authors themselves call the effect sizes modest.
The Swiss study compared 16 affected horses with 15 kept in the same stable, sampled in spring and in autumn of the same year. Neither season nor health status altered the composition of the community at any taxonomic level, and diversity was the same. The role of the microbiome in how the sign arises, the authors write, remains to be clarified.
The largest study comes from Sweden and Norway: 50 affected horses, 50 comparison horses matched by yard, three sampling periods from autumn to late winter. Overall composition was similar between the groups; the sampling time, by contrast, weighed heavily. 14 genera differed in abundance in at least one period, Alloprevotella more often and Bacillus less often in the affected horses. All horses were culture negative for Clostridioides difficile and Clostridium perfringens.
Two further studies, ten against ten horses and 14 against 11, likewise found no difference. The second states explicitly that the earlier European work contradicted itself. The American study, 32 affected horses against 51 unaffected, is the first to see a different community structure and several differentially abundant groups, Alloprevotella again among them. The authors themselves call for larger numbers.
| Study | Affected against comparison | Sampling | Result |
|---|---|---|---|
| Switzerland, 2020 | 16 against 15 | Spring and autumn of the same year | No difference in composition or diversity |
| Sweden and Norway, 2021 | 50 against 50 | Three periods from autumn to late winter | Overall composition similar, marked effect of timing |
| Study with a transplantation arm, 2021 | 10 against 10 | Comparison horses from the same site | No difference in composition |
| Canada, 2024 | 14 against 11 | Once, in autumn | No difference; the authors note the contradiction in earlier work |
| United States, 2025 | 32 against 51 | Once, case definition confirmed by a veterinary surgeon | Different community structure, effect sizes modest according to the authors |
The second main hypothesis concerns the ration. A case-control study of 50 pairs matched by yard, in Sweden and Norway, compared feeding and management. The average amount of haylage and the management practices were the same in both groups. Affected horses, however, received twice as much concentrate per 100 kilograms of body mass per day, more starch and more water-soluble carbohydrates, but less digestible crude protein, less fibre, less straw and lucerne.
The authors themselves call these differences small: starting points for further study, not causes. Two reasons support that caution. The rations were reported by the keepers, not weighed. And the order of events remains open: anyone who has been wiping free faecal water for months changes the feeding, which may equally well have produced the observed difference.
In the survey of 339 horses, 58 per cent of keepers reported fewer signs after switching from haylage to hay and 46 per cent after a move to pasture. The third figure is the striking one: 17 per cent saw an improvement after nothing more than a change of batch of the same forage. All three are impressions without a comparison group.
A case series from 2022 describes a single horse whose free faecal water improved after a change of hay. The same authors state in the same text that the cause of the syndrome remains unknown. A case without a comparison remains a case without a comparison.
When liquid leaves the droppings, it is natural to examine the droppings. That is what two sub-studies did: 50 pairs in Scandinavia at three time points, 32 pairs in Germany once. Total short-chain fatty acids, pH and osmolality were similar between affected and unaffected horses. The authors read this as a sign that fermentation runs much the same way.
Differences appeared only in the Scandinavian sub-study, and only two: less lactic acid and a lower water-holding capacity of the droppings in affected horses. In the German sub-study of 32 pairs not one of the measures differed. Particle size distribution was the same in both sub-studies, which removes the basis for the widespread notion of fibre too coarse to bind water.
That a different fibre composition of the forage does shift how water is distributed was shown in six French trotters. On early-cut forage the droppings bound 12.6 grams of water per gram of dry matter, against 11.1 and 11.4 grams on the later cuts. Droppings carried 13.1 kilograms of water out of the body instead of 18.8 and 17.6, and urine 10.5 instead of 7.6 and 7.9. Intake and total excretion were unchanged.
A suspected cause can be tested by removing it and watching what follows. In free faecal water this has been tried several times, and each time the result runs against expectation.
A product containing montmorillonite and bentonite, whey and extracts of hops and wormwood was tested in eight horses in a complete crossover trial, each animal serving as its own control. Not one of the twelve faecal measures changed, from dung score to pH to the volatile fatty acids. The authors point out that these horses were healthy and call for trials in affected animals.
In ten affected horses, gut flora from a donor was transferred. Severity, scored by a veterinary surgeon, was lower on day 14 and stayed lower to day 168, without the recipients' faecal flora changing. There was no comparison group, no blinding, and a 0 to 4 scale built for this study alone. An improvement without a comparison group is compatible with the natural waxing and waning of the sign; the authors therefore call for controlled studies.
For gut flora transfer in horses generally, a systematic review from 2023 reaches the same conclusion: seven studies, small numbers throughout, mostly without a control group, no dependable information on donor selection, dose and procedure. In a randomised trial in 25 foals with diarrhoea, neither survival nor resolution of the diarrhoea differed between the groups. And foals with diarrhoea are not adult horses with free faecal water.
No controlled study has ever tested probiotics in free faecal water. A 2018 review states that a general benefit in horses is currently unsupported, that well-built studies are rare and contradictory, and that the labels of over-the-counter products often fail to reflect their contents. In a randomised field trial in 38 newborn foals, a multi-strain product did not even raise the proportion of lactobacilli in the droppings.
One last finding overturns a whole explanation. Improvement in dropping consistency under clioquinol was long ascribed to an action against protozoa. A study in eight healthy horses showed instead a broad antibacterial action: on day five a 90 per cent fall in species diversity, the spread of facultatively anaerobic families and salmonellae in the droppings of several animals. The authors urge restraint.
Why do five studies of the same question reach different answers? First, the case definition. In the Swedish work the keepers graded against a set template and an independent person then checked, with no meaningful deviation. In the American work veterinary supervision confirmed the two phases. The Canadian publication names no criterion at all.
Second, size. The largest microbiome study covers 100 horses, the smallest 20. If such a study finds no difference, that does not mean there is none, only that none became visible at that order of magnitude. Small differences stay below the limit of detection.
Third, the place of measurement. Only the droppings are examined, that is, the end product. What happens in the caecum and the front part of the colon, where fibre is fermented and large volumes of water are moved, remains unobserved in all the work published so far.
Fourth, a conceptual reason. A 2024 review states that variation in the faecal microbiome, within one horse as well as between horses, is so large that no clear definition of dysbiosis has yet been achieved for the horse, and that for no disease has an unambiguous causal link with dysbiosis been shown. Without a yardstick, no deviation can be established.
Demand for explanations rises in autumn, when the basic forage changes. No available study has measured whether free faecal water then occurs more often. Something else was measured: in the Swedish work the faecal flora of all horses changed markedly between October, the turn of the year and late winter, in affected and unaffected alike. The sampling time weighed more than health status.
In the Swiss work, by contrast, spring and autumn did not differ. Taken together the two results counsel caution towards any snapshot: sample affected and healthy horses at different times and what you end up measuring is the season.
Then there is the question of order. All the available work is case-control: it compares two groups at one point in time and cannot say whether a difference preceded the sign or followed it. For a sign that comes and goes and regularly prompts keepers to change ration and management, that is not a side question but the question.
It does not mean there is no cause, nor that nothing can be done. It means that no product and no feeding regime has shown an effect in affected horses. Persistent free faecal water, sore skin, weight loss or signs of colic belong in a veterinary examination, not in a trial plan from the internet.
As instructive as the contradictions is the list of explanations on which nothing at all can be found. For dental status, rank within the group, psychological load, worming programme and amount of exercise, this research found no study in horses with free faecal water. These explanations have not been refuted. They have not been examined.
With sand the case is subtler. The suspicion has been investigated, but in the wrong population: a 2024 study compared 74 colic horses with 74 comparison horses and found dry matter, particle size distribution and sand score similar between the groups. That says something about the sand test in colic. About free faecal water it says nothing, because the same question has never been put in affected horses.
And the most obvious thing is missing. No randomised, controlled trial of a change in feeding or management in affected horses. No follow-up of the same animals across a full year with uniform recording. That is the real finding of this field, and it explains why the advice diverges so widely: where nothing has been measured, experience fills the gap, and experience contradicts itself.
Original analysis
| Hypothesis | Strongest design found | Tested in affected horses | Result | What does not follow |
|---|---|---|---|---|
| Disturbed gut flora | Five case-control studies, 122 affected and 137 comparison horses | Yes | Four with no difference, one with a small difference in community structure | That dysbiosis is ruled out: the numbers are too small for that |
| Too much concentrate and starch | Case-control study of 50 pairs, rations reported by keepers | Yes | Twice the amount of concentrate per 100 kilograms in affected horses, called small by the authors themselves | That less concentrate reduces free faecal water |
| Haylage instead of hay | Online survey of 339 horses without a comparison group | Yes | 58 per cent of keepers report fewer signs after switching to hay | That the method of conservation is the cause: a mere change of batch also brought improvements |
| Fibre too coarse to bind water | Case-control study with particle measurement, 50 and 32 pairs | Yes | Particle size distribution similar in both sub-studies, water binding lower in only one | That the fibre length of the feed explains the sign |
| Derailed hindgut fermentation | Chemical analysis of faeces, 50 and 32 pairs | Yes | Short-chain fatty acids, pH and osmolality similar between the groups | That fermentation is unremarkable: measurement was on droppings, not in the hindgut |
| Clostridia as a trigger | Culture within the case-control study of 50 pairs | Yes | All horses negative for Clostridioides difficile and Clostridium perfringens | That pathogens play no part: only these two were looked for |
| Protozoa in the hindgut | Study of the effect of clioquinol in eight healthy horses | No | The substance acts broadly against bacteria rather than against protozoa | That protozoa play no part: the question was never put in affected horses |
| Sand in the hindgut | Case-control study of 74 colic and 74 comparison horses | No, in colic horses | Dry matter, particle size and sand score similar between the groups | Anything at all about free faecal water: the population examined was a different one |
| Clay minerals as a remedy | Complete crossover trial in eight healthy horses | No | No effect on a single one of the twelve faecal measures | That the product is ineffective in affected horses: they were not in the trial |
| Gut flora transfer | Series without a control group, ten affected horses | Yes | Severity lower to day 168, faecal flora unchanged | That the transfer works: without a comparison group the improvement cannot be assigned |
| Teeth, rank, stress, worming | No study in horses with free faecal water found in this research | No | No result, because no measurement | That these explanations have been refuted: they are untested |
There is no established answer to that so far. Two explanations above all have been tested. Five case-control studies compared the faecal flora of affected horses with that of stablemates without the sign: four found no meaningful difference, one found a difference with effect sizes its authors themselves call modest. A study of feeding and management in 50 pairs found small differences in concentrate feed, which its authors expressly decline to describe as causes. The chemical composition of the droppings and the fibre length likewise failed to separate affected from unaffected horses reliably. The honest position is this: several plausible explanations, none confirmed.
No product and no feeding regime has shown an effect in affected horses in a controlled study, because such studies do not exist. A clay mineral product changed not one of the twelve faecal measures in a crossover trial in eight healthy horses. After a transfer of gut flora into ten affected horses, severity fell up to day 168, but without a comparison group and without any change in the faecal flora, which also fits the natural waxing and waning of the sign. For probiotics there is no study in free faecal water at all. What makes sense in an individual case belongs to veterinary judgement.
The available data allow no clear statement. In the American study the affected horses were otherwise regarded as healthy on clinical examination, and horses with persistent diarrhoea were explicitly excluded. In the survey of 339 horses, however, 23 per cent had a history of colic, a proportion the authors rate as high compared with other published horse populations; whether a connection lies behind that or the self-selection of an online survey remains open. Irritation of the skin on the hindquarters and tail is described regularly. Weight loss, fever or signs of colic always belong in a veterinary examination.
Whether it occurs more often in winter has never been measured. Something else was measured, and it is instructive: in the Swedish and Norwegian study, with three sampling periods from autumn to late winter, the faecal flora of all horses changed markedly with the timing, in affected and unaffected alike. That effect of time was clearer than any difference between the groups. In the Swiss study, by contrast, spring and autumn did not differ. Anyone who notices more free faecal water in winter is also observing under other conditions: more time in the stable, more time spent with the horse, and conserved forage instead of pasture.
The available data cannot answer that question. In the survey of 339 horses, 58 per cent of keepers reported fewer signs after switching from haylage to hay and 46 per cent after a move to pasture. These are impressions, without a comparison group and without measurement, collected for a sign that fluctuates anyway. The third figure from the same survey is the striking one: 17 per cent saw an improvement after nothing more than a change of batch of the same type of forage. That sits poorly with a single cause in the method of conservation. Forage changes also affect the whole digestive tract and are worth discussing.
For free faecal water there is not a single study. A 2018 review states for horses in general that a benefit is currently unsupported, that well-built studies are rare and contradictory, and that the labels of over-the-counter products often fail to reflect the actual contents. In a randomised field trial in 38 newborn foals, a multi-strain product did not even raise the proportion of lactobacilli in the droppings. An older randomised study in 14 horses with acute enterocolitis showed a milder and shorter course under a yeast: a different clinical picture, and never repeated on this question.
Terms defined in this text
Equine Health
82 per cent of clinically healthy horses reacted to leptospires: what serology, genetics and the autoimmune trail really establish, and what they do n
HOHorseBmoderate Open access
Movement & Recovery
Of 750 horses at competitions, 44 per cent left no finger under the noseband: what the taper gauge, pressure sensors and a cadaver head show.
HOHorseBmoderate Open access
Equine Health
Only about one per cent of crib-biters began after contact with a crib-biting horse, although nearly half of owners believe it is copied.
HOHorseBmoderate Open access
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.