Soaking hay: why no soaking time guarantees a result on sugar

What soaking actually takes out of the hay

Soaking hay exploits a simple property of the stem: part of its carbohydrate is water-soluble and leaves with the water. The question is therefore never whether soaking removes something. It is how much, how reliably, and what it carries off besides the sugar. In a laboratory series on British hays, the starting content of water-soluble carbohydrates lay between 123 and 230 grams per kilogram of dry matter. After a bath of up to sixteen hours in water at around eight degrees, a good quarter of that was missing on average, but in the individual batch anywhere between 6 and 54 per cent.

What does ‘sugar in hay’ actually mean?

Three terms sit side by side. Water-soluble carbohydrates cover the simple sugars and the fructans. Ethanol-soluble carbohydrates essentially capture the simple sugars alone. Non-structural carbohydrates add starch on top. A laboratory can therefore report three different figures for the same hay without any one of them being wrong.

The distinction is not academic, because the fractions behave differently in the water bath. In that same laboratory series, 56 per cent of the glucose, 45 per cent of the sucrose and 41 per cent of the fructose disappeared on average, but only 24 per cent of the fructans. Crude protein content did not change meaningfully. Soaking therefore acts selectively: it washes out the small, readily soluble molecules and leaves the long-chain fructans behind out of all proportion.

Why the figures in circulation differ by a factor of two

Anyone looking for a soaking time finds figures from a quarter of an hour to sixteen hours and reduction rates of around 20 to 50 per cent. Those figures do not contradict one another; they come from different trials under different conditions. A European expert panel gives a fall in nutrients including water-soluble carbohydrates of 24 to 43 per cent for seven to sixteen hours at room temperature. A laboratory series on three hays measured 18 per cent after an hour and a half, 38 per cent after nine hours and 42 per cent after sixteen.

Four things separate these trials: the hay batch, the water temperature, the agitation in the bath and the fraction measured. A fifth point comes on top, which hardly any discussion mentions: the laboratory method. A comparison of methods on pasture grasses showed that a cheap colorimetric procedure systematically underestimates fructan content. Across the same samples, the mean from liquid chromatography came to 154 grams per kilogram of dry matter, that of the colorimetric procedure to 82. The work concerns fresh pasture grass and not conserved hay, but it explains why two analyses of the same batch can diverge.

Time, temperature, agitation: what really drives the result

The table carries the central message of that work: at the same duration of sixteen hours, the removal of water-soluble carbohydrates ranged from 28 to 49 per cent, purely because the water was warmer or was agitated. The authors conclude that temperature and agitation determine how much is removed, and not duration alone. The last columns are remarkable: sucrose and glucose pass almost entirely into solution, while two thirds of the fructans stay in the stem. Anyone whose target is precisely the fructans is holding the weakest tool.

Does soaking for longer achieve more?

On the available trials, hardly. In a trial on four meadow hays at 20 degrees, fifteen minutes already lowered almost every nutrient measured to a marked degree, and extending to thirty minutes, one hour or twelve hours removed no more. An older stable measurement of dust exposure reached the same conclusion: no advantage from wetting for longer.

That turns the usual logic on its head. The idea that a night in the water is more thorough than a quarter of an hour finds no support in these data as far as sugar removal goes. The collateral costs, by contrast, fall due within that first quarter of an hour: in the same trial, metabolisable energy dropped by 5 to 15 per cent, and the crude protein digestible before the large intestine, amino acids included, fell by 35 per cent. Crude fibre and lignin rose arithmetically, because the soluble fractions were missing and what remained was concentrated.

The spread is the real result

The most striking observation in this literature is not a number but its blurriness. In a laboratory series on three hays with three bales each, the mean losses came to 18 per cent after an hour and a half, 38 per cent after nine hours and 42 per cent after sixteen. The spread around those means was, at nine and at sixteen hours, as large as the means themselves. The authors therefore call the course of leaching expressly inconsistent and unpredictable, and they found no relationship between the extent of leaching and bacterial growth.

In practice this means that a percentage cannot be used to work back to how much sugar is still in the hay after soaking. That is exactly what the laboratory series mentioned at the outset showed on its own material. Despite an average loss of 27 per cent, seven of the hays examined stayed above the limit of 100 grams of water-soluble carbohydrates per kilogram of dry matter that the same authors name as an upper bound for animals at risk of laminitis. That limit is a proposal from the specialist literature and not a threshold tested in a clinical study.

What stays in the water besides the sugar

Water is not a selective solvent: it takes with it whatever is soluble. The group that tested four meadow hays at 20 degrees therefore concludes that soaking can degrade the feeding value of the hay and make ration calculation on a dry matter basis uncertain, precisely in horses with a metabolic disorder or with asthma.

At first sight that contradicts the English laboratory series, in which no soaking procedure changed the crude protein content. The contradiction is more a misunderstanding about the quantity measured. Crude protein is the total amount. The share digestible before the large intestine is a fraction calculated from that total, and it says how much of it can be taken up ahead of the large intestine. In the work cited, this fraction was calculated from nitrogen fractions and not measured in the animal; not a single horse was fed there.

How far the loss of dry matter shifts a ration is shown by a feeding trial in twelve overweight horses and ponies over six weeks. A ration of 1.25 per cent of body weight as dry matter was planned. Because soaking washed out nutrients and dry matter, only around 1 per cent actually arrived, corresponding to 64 per cent of the maintenance requirement for digestible energy. The restriction turned out 23.5 per cent sharper than intended, and the animals lost on average just under 1 per cent of body weight per week, the most sensitive of them close to 2 per cent.

Another group measured, in the soaked hay of its own trial, 38 per cent less non-structural carbohydrate, 6.78 per cent less digestible energy and leached water-soluble minerals. An analysis of 33 studies on sodium, potassium and chloride metabolism draws attention to the same point: ponies later found to have pituitary dysfunction digested potassium less well than healthy animals, which deserves attention with soaked hay. That analysis did not investigate soaking itself.

The other side: the microbial count rises

Soaking does not improve the hygienic quality of hay, it worsens it. In a laboratory series with five hays and five treatments, the bacterial count rose from around 60,000 to around 350,000 colony-forming units per gram of dry matter, roughly a sixfold increase; the mould count fell from 1,148 to 692. Steaming alone, by contrast, brought bacteria down to around 1,000 and moulds to two, while leaving the sugar content practically unchanged. The authors record that soaking meaningfully worsened hygienic quality.

A second work sequenced the viable bacteria of four hays under three modes of preparation. Soaking raised the share of Gram-negative bacteria and reduced the diversity of the community, whereas high-temperature steam lowered the number of live organisms without reducing diversity. The authors therefore regard steaming as the more suitable procedure. That conclusion rests on a microbiological finding and not on a health outcome measured in the horse.

Then there is keeping quality. In a trial with a fifteen-minute soak, bacteria, moulds and yeasts increased in the soaked hay after six, twelve and twenty-four hours of storage, at 10 degrees as much as at 25. Six warmblood mares took up both soaked and steamed hay more slowly in the first hour than untreated hay; the authors rate the soaked hay as the less well accepted overall. Soaked hay is therefore a fresh product.

What has been measured in the horse itself

The literature named so far measures the hay, not the horse. The only randomised trial that compares soaked and dry hay directly tested three forages in twelve ponies after a meal of 0.25 per cent of body weight. The insulin response over five hours was lowest with the soaked hay, almost three times as high with dry hay and around six times as high with haylage. The glucose response barely differed between the three feeds. Six of the twelve ponies were regarded as insulin dysregulated.

A second work found nothing. In nine healthy Standardbred racehorses, neither soaking nor steaming changed the acute glucose response to a meal, even though soaking had measurably lowered non-structural carbohydrates, soluble protein and potassium. The horses ate less of the soaked hay and spent less time on it. What was measured here, however, was blood glucose and not insulin, the value that matters for endocrine laminitis.

The evidence is clearer for dust. A wearable particulate monitor in the breathing zone of ten horses recorded for eight hours after feeding. Exposure fell by more than half with the soaked hay, in the first twenty minutes from 160 to 53 micrograms per cubic metre for the finest particles and from 2,829 to 970 for the coarser ones. The authors themselves record that no health outcome was collected.

A trial in ten horses with severe asthma, five per group, tested hay soaked for 45 minutes over six weeks, with dried-out hay removed between meals. Airway resistance fell in both groups, on soaked hay from a median of 1.89 to 0.61 and on lucerne pellets from 2.47 to 1.59. The authors add that this strict routine is likely to strike owners and keepers as too laborious. The comparison was moreover against pellets and not against dry hay.

What follows from this, and what does not

A European consensus statement on equine metabolic syndrome puts not sugar but insulin dysregulation at the centre: it is the consistent feature of endocrine laminitis, and high insulin values trigger laminitis. For the feed, the document names a target of under 10 per cent non-structural carbohydrate and describes management as a long-term feeding and exercise strategy that demands consistency.

That puts the widespread fixation on fructans into perspective, of all fractions the one that resists water best. In a randomised trial in twelve ponies, five of them with previous laminitis, glucose and fructose produced marked glucose and insulin peaks, whereas inulin as a fructan produced only minimal changes. These were pure sugars given as a supplement and not the sugars of a hay, and the response was stronger in autumn than in spring.

Steaming is the obvious alternative, but it does not solve the sugar problem. In a trial on two hays, ninety minutes of steaming lowered water-soluble carbohydrates by 13 per cent only in the mouldier hay, and not at all in the barely mouldy one. Moulds fell by at least 91 per cent in both hays, and intake of the barely mouldy hay rose over two hours from 0.64 to 2.02 kilograms. Whether steaming lowers sugar therefore hangs on the batch.

A review of environmental management in equine asthma sums up the trade-off: soaking and steaming improve the hygienic quality of the feed but lower its nutritional value, so that compensation becomes necessary. The most critical point of environmental control, however, is neither procedure but the fidelity with which owners and keepers carry it out. This is a narrative review without a systematic search.

Where the contradictory soaking figures come from: condition, fraction measured and open question for each series of measurements

Original analysis

Limitations and uncertainty

• Almost all the figures come from laboratory series on very few hay batches, mostly three to five per piece of work, and from Britain, Germany or Canada. No Swiss series of measurements on Swiss hay could be found in this search, even though the batch is the largest source of spread.
• The editorial requirement to base at least half the sources at the level of systematic review, meta-analysis or randomised trial cannot be met here: only four of the twenty works used reach that level, and the single meta-analysis investigates not soaking but electrolyte balance. No systematic review of soaking exists.
• Only one randomised trial compares soaked and dry hay directly: twelve ponies, one meal after a fasting period, five hours of observation. No clinical outcome such as laminitis was collected in any of the works cited here.
• The frequently quoted loss of protein digestible before the large intestine was calculated from nitrogen fractions and not measured in the animal; no horse was fed in that work. The contradiction with the unchanged crude protein of other laboratory series therefore remains arithmetical in nature.
• The microbial counts differ considerably between works, and in one case even between the abstract and the full text of the same publication. What holds is the order of magnitude and the ranking of the procedures, not the individual figure.
• The upper bound of 100 grams of water-soluble carbohydrates per kilogram of dry matter and the target of under 10 per cent non-structural carbohydrate are proposals from the specialist literature and from consensus, not thresholds tested in a trial.

Open questions

• Does soaking tailored to the measured batch lower the risk of laminitis, or does it only alter the insulin curve after the meal?
• What does the raised microbial count of soaked hay mean for the horse? No health outcome has so far been collected after soaking.
• Why does crude protein stay unchanged while the arithmetically determined fraction digestible before the large intestine falls by a third: measurement artefact or real loss?
• Can the actual removal from a batch be estimated in the stable, through a fibre correction for instance, instead of being borrowed from averages established elsewhere?

Frequently asked questions

How long do I have to soak hay for the sugar to come down?

No generally valid duration can be derived from the studies, and that is the most important finding in this field. In a trial on four meadow hays, most of the soluble fractions were already washed out after fifteen minutes, and thirty minutes, one hour or twelve hours removed no more. In another laboratory series, removal ranged from 28 to 49 per cent at an identical sixteen hours, purely because of water temperature and agitation. And even with the percentage known, the final content remains open, because it depends on the starting content of the batch. Anyone who needs a figure needs an analysis of their own hay.

Is soaking overnight better than half an hour?

For sugar removal, the longer time brings no demonstrated additional benefit on the available data; for the nutrient balance and the microbial count it is unfavourable. A quarter of an hour already lowered metabolisable energy by 5 to 15 per cent in one trial, and the protein calculated as digestible before the large intestine by 35 per cent. A laboratory group warns expressly about the multiplication of unwanted micro-organisms during very long soaking in warm water, and a European expert panel recommends no more than one to two hours in warm weather. For dust exposure too, an older stable measurement found no advantage in wetting for longer.

My horse has had laminitis. Is soaked hay safe then?

No, soaking does not automatically make any hay suitable. In one laboratory series, seven hays stayed above the upper bound of 100 grams of water-soluble carbohydrates per kilogram of dry matter that the same authors name for animals at risk, despite an average loss of 27 per cent. A European consensus statement moreover puts not sugar but insulin dysregulation at the centre of endocrine laminitis, and names a feed target of under 10 per cent non-structural carbohydrate. Whether and how soaked hay comes into question for a particular horse belongs with the attending veterinary practice, on the basis of an analysis of the batch.

Does hay also lose protein and minerals when it is soaked?

Minerals yes; with protein it depends on the quantity measured. Several works found water-soluble minerals leached out, one of them 6.78 per cent less digestible energy as well. Total crude protein stayed unchanged in the English laboratory series, while a German work saw the protein fraction calculated as digestible before the large intestine fall by 35 per cent. That is not a genuine contradiction: the one figure describes the total amount, the other the share that can be taken up ahead of the large intestine. That fraction, however, was measured in the forage and not in the horse.

Can I soak the hay in the morning and feed it in the evening?

The data argue against it. In a trial with a fifteen-minute soak, bacteria, moulds and yeasts increased in the soaked hay after six, twelve and twenty-four hours of storage, at 10 degrees as much as at 25. Soaked hay therefore behaves like a fresh product and not like dry hay. The soaking itself had already raised the bacterial count roughly sixfold in another laboratory series. How long a given portion can reasonably stand in a given stable cannot be derived from these trials.

Is steaming better than soaking?

For hygiene yes, for sugar no. In one laboratory series, steaming brought bacteria down to around a thousand and moulds to two colony-forming units per gram, while leaving the sugar content practically unchanged; soaking lowered the sugar and let the bacterial count rise. Only the sequence of soaking and then steaming lowered both. In a further trial, steaming lowered water-soluble carbohydrates in only one of two hays. Anyone with dust and mould in mind chooses differently from someone who has to bring the sugar down.

Sources

1. Longland AC, Barfoot C, Harris PA. Effects of soaking on the water-soluble carbohydrate and crude protein content of hay. The Veterinary Record, 2011 (Laboratory study | Horse)DOI 10.1136/vr.d157
   Soaking removed on average 27 per cent of the water-soluble carbohydrates with a range of 6 to 54 per cent, yet seven hays still lay above the limit of 100 grams per kilogram of dry matter proposed as an upper bound, while crude protein did not change meaningfully.
2. Longland AC, Barfoot C, Harris PA. Effect of period, water temperature and agitation on loss of water-soluble carbohydrates and protein from grass hay: implications for equine feeding management. The Veterinary Record, 2014 (Laboratory study | Horse)DOI 10.1136/vr.101820
   Water temperature and agitation, and not duration alone, determined how much was removed: after sixteen hours, 28 per cent of the water-soluble carbohydrates were missing at 8 degrees against 49 per cent at 16 degrees with agitation and rinsing, with fructans resisting far more strongly than sucrose and glucose.
3. Bochnia M, Pietsch C, Wensch-Dorendorf M, Greef M, Zeyner A. Effect of Hay Soaking Duration on Metabolizable Energy, Total and Prececal Digestible Crude Protein and Amino Acids, Non-Starch Carbohydrates, Macronutrients and Trace Elements. Journal of Equine Veterinary Science, 2021 (Laboratory study | Horse)DOI 10.1016/j.jevs.2021.103452
   Fifteen minutes already lowered almost every nutrient measured to a marked degree and longer soaking removed no more, yet those same fifteen minutes cost 5 to 15 per cent of the metabolisable energy and 35 per cent of the protein calculated as digestible before the large intestine, amino acids included.
4. Moore-Colyer M, Longland A, Harris P, Zeef L, Crosthwaite S. Mapping the bacterial ecology on the phyllosphere of dry and post soaked grass hay for horses. PLoS One, 2020 (Laboratory study | Horse)DOI 10.1371/journal.pone.0227151
   Leaching and bacterial growth ran inconsistently and unpredictably and were unrelated to one another: the mean losses came to 18, 38 and 42 per cent after an hour and a half, nine hours and sixteen hours, with the spread at the longer times as large as the mean itself.
5. Moore-Colyer MJ, Lumbis K, Longland A, Harris P. The effect of five different wetting treatments on the nutrient content and microbial concentration in hay for horses. PLoS One, 2014 (Laboratory study | Horse)DOI 10.1371/journal.pone.0114079
   Nine hours of soaking lowered the water-soluble carbohydrates from 126 to between 79 and 83 grams per kilogram of dry matter but multiplied the bacterial count roughly sixfold, while steaming lowered the microbial load and left the sugar untouched; only the sequence of soaking and then steaming lowered both.
6. Daniels S, Hepworth J, Moore-Colyer M. The haybiome: Characterising the viable bacterial community profile of four different hays for horses following different pre-feeding regimens. PLoS One, 2020 (Laboratory study | Horse)DOI 10.1371/journal.pone.0242373
   The mode of preparation shaped the bacterial profile more strongly than the hay itself: soaking raised the share of Gram-negative bacteria and reduced diversity, high-temperature steam lowered the count of viable organisms without loss of diversity, the conclusion resting on a microbiological and not on a clinical outcome.
7. Glatter M, Bochnia M, Wensch-Dorendorf M, Greef JM, Zeyner A. Feed Intake Parameters of Horses Fed Soaked or Steamed Hay and Hygienic Quality of Hay Stored following Treatment. Animals, 2021 (Controlled trial | Horse)DOI 10.3390/ani11092729
   Soaked hay proved not to be storage stable, as bacteria, moulds and yeasts increased after six, twelve and twenty-four hours at 10 and at 25 degrees, and it was less well accepted by six warmblood mares than untreated or steamed hay.
8. Argo CM, Dugdale AH, McGowan CM. Considerations for the use of restricted, soaked grass hay diets to promote weight loss in the management of equine metabolic syndrome and obesity. The Veterinary Journal, 2015 (Controlled trial | Horse)DOI 10.1016/j.tvjl.2015.07.027
   Soaking washed out so much dry matter and nutrient that instead of the planned 1.25 per cent of body weight only around 1 per cent arrived, corresponding to 64 per cent of the maintenance requirement and to a restriction 23.5 per cent sharper than intended.
9. Carslake HB, Argo CM, Pinchbeck GL, Dugdale AHA, McGowan CM. Insulinaemic and glycaemic responses to three forages in ponies. The Veterinary Journal, 2018 (Randomised trial | Horse)DOI 10.1016/j.tvjl.2018.03.008
   At equal dry matter, soaked hay triggered the lowest insulin response, dry hay almost three times as high and haylage around six times as high, while the glucose response barely differed; this is the only randomised trial that compares soaked and dry hay directly.
10. McGowan CM, Dugdale AH, Pinchbeck GL, Argo CM. Dietary restriction in combination with a nutraceutical supplement for the management of equine metabolic syndrome in horses. The Veterinary Journal, 2013 (Randomised trial | Horse)DOI 10.1016/j.tvjl.2012.10.007
    A supplement and not the soaking was randomised, both groups receiving soaked hay; what remains usable is the measurement in the hay, with 38 per cent less non-structural carbohydrate and 6.78 per cent less digestible energy, together with the finding that the improvement in insulin sensitivity depended on the starting state and on weight loss.
11. Owens TG, Barnes M, Gargano VM, Julien L, Shoveller AK. Nutrient content changes from steaming or soaking timothy-alfalfa hay: effects on feed preferences and acute glycemic response in Standardbred racehorses. Journal of Animal Science, 2019 (Controlled trial | Horse)DOI 10.1093/jas/skz252
    Thirty minutes of soaking lowered soluble protein, non-structural carbohydrates and potassium, yet the acute glucose response of the healthy Standardbred racehorses changed neither through soaking nor through steaming, and the horses ate less of the soaked hay.
12. Ivester KM, Ni JQ, Couetil LL, Peters TM, Park JH. A wearable real-time particulate monitor demonstrates that soaking hay reduces dust exposure. Equine Veterinary Journal, 2025 (Controlled trial | Horse)DOI 10.1111/evj.14425
    In the breathing zone of ten horses, soaked hay lowered particulate exposure by more than half, in the first twenty minutes from 160 to 53 micrograms per cubic metre for the finest particles, with the authors themselves recording that no health outcome was collected.
13. Westerfeld R, Payette F, Dubuc V, Manguin E, Leclere M. Effects of soaked hay on lung function and airway inflammation in horses with severe asthma. Journal of Veterinary Internal Medicine, 2024 (Controlled trial | Horse)DOI 10.1111/jvim.16919
    Hay soaked for forty-five minutes, with dried-out hay removed between meals, kept airway obstruction under control in ten horses with severe asthma, but the authors themselves regard this strict routine as too laborious for many owners and keepers.
14. Clements JM, Pirie RS. Respirable dust concentrations in equine stables. Part 2: the benefits of soaking hay and optimising the environment in a neighbouring stable. Research in Veterinary Science, 2007 (Other | Horse)DOI 10.1016/j.rvsc.2006.12.003
    Soaking lowered the respirable dust concentration in the breathing zone to a marked degree without longer wetting bringing any further advantage, and the conditions in the neighbouring stable altered exposure as well, which puts the effect of a single measure into perspective.
15. Maier I, Kienzle E. A Meta-Analysis on Quantitative Sodium, Potassium and Chloride Metabolism in Horses and Ponies. Animals, 2025 (Meta-analysis | Horse)DOI 10.3390/ani15020191
    This analysis of 33 studies concerns electrolyte balance and not soaking; the only usable finding is that ponies later found to have pituitary dysfunction digested potassium less well, a point the authors ask readers to bear in mind with soaked hay.
16. Durham AE, Frank N, McGowan CM, Menzies-Gow NJ. ECEIM consensus statement on equine metabolic syndrome. Journal of Veterinary Internal Medicine, 2019 (Professional guideline | Horse)DOI 10.1111/jvim.15423
    The consensus statement puts insulin dysregulation and not sugar at the centre of endocrine laminitis, puts the nutrient loss at 24 to 43 per cent for seven to sixteen hours of soaking, names under 10 per cent non-structural carbohydrate for the feed and advises no more than one to two hours in warm weather.
17. Diez de Castro E, Fernandez-Molina JM. Environmental Management of Equine Asthma. Animals, 2024 (Other | Horse)DOI 10.3390/ani14030446
    This narrative review holds that soaking and steaming improve hygienic quality but lower nutritional value and therefore call for compensation, and that the most critical point of environmental control is the fidelity with which owners and keepers carry it out.
18. Earing JE, Hathaway MR, Sheaffer CC, Hetchler BP, Martinson KL. Effect of hay steaming on forage nutritive values and dry matter intake by horses. Journal of Animal Science, 2013 (Controlled trial | Horse)DOI 10.2527/jas.2013-6333
    Steaming lowered the water-soluble carbohydrates by 13 per cent only in the mouldier of the two hays and not at all in the barely mouldy one, but it reduced moulds by at least 91 per cent in both and tripled intake of the barely mouldy hay.
19. Longland AC, Dhanoa MS, Harris PA. Comparison of a colorimetric and a high-performance liquid chromatography method for the determination of fructan in pasture grasses for horses. Journal of the Science of Food and Agriculture, 2012 (Laboratory study | Horse)DOI 10.1002/jsfa.5555
    The cheap colorimetric procedure systematically underestimated the fructan content of pasture grasses compared with liquid chromatography, on average 82 against 154 grams per kilogram of dry matter, and does not serve as a substitute according to the authors: the fructan value of an analysis depends on the laboratory method.
20. Borer KE, Bailey SR, Menzies-Gow NJ, Harris PA, Elliott J. Effect of feeding glucose, fructose, and inulin on blood glucose and insulin concentrations in normal ponies and those predisposed to laminitis. Journal of Animal Science, 2012 (Randomised trial | Horse)DOI 10.2527/jas.2011-4236
    The three sugars acted differently: glucose separated ponies with previous laminitis most sharply from healthy animals, fructose produced weaker excursions and inulin as a fructan only minimal changes in glucose and insulin, with the response stronger in autumn.