Article Text


Effects of soaking on the water-soluble carbohydrate and crude protein content of hay
  1. A. C. Longland, BSc, PhD, DIC1,
  2. C. Barfoot, BSc, RNutr2 and
  3. P. A. Harris, MA, PhD, DipECVCN, VetMB, MRCVS3
  1. Equine and Livestock Nutrition Services, Pantafallen fach, Tregaron, Ceredigion, Wales, SY25 6NG
  2. MARS Horse Care UK, Old Wolverton, Buckinghamshire, MK12 5PZ
  3. WALTHAM Centre for Pet Nutrition, Waltham-on-the-Wolds, Leicestershire, LE14 4RT
  1. E-mail for correspondence elns.longland{at}

The aim of this study was to determine the amounts of water-soluble carbohydrate (WSC), WSC constituents and protein leached from a range of UK hays soaked according to common practice. Initial hay WSC content ranged from 123 to 230 g/kg dry matter (DM). Soaking the hays for up to 16 hours in water at a mean temperature of 8°C resulted in a mean loss of 27 per cent (range 6 to 54 per cent) of hay WSC. The mean percentage losses of WSC constituents were 24 per cent (range 14 to 31) for fructan, 41 per cent (range 21 to 70) for fructose, 45 per cent (range 28 to 100) for sucrose and 56 per cent (range 29 to 100) for glucose. The mean crude protein content of the initial hays was 58.7 g/kg DM (range 30 to 86 g/kg DM) and this value was not affected significantly by soaking. Despite a mean WSC loss of 27 per cent, the WSC contents of seven of the hays remained above the suggested upper limit for laminitic animals of 100 g/kg DM.

Statistics from

LAMINITIS is a painful and debilitating condition, which may necessitate euthanasia (Hinckley and Henderson 1996, Pollitt and others 2003). Pasture grass species commonly found in temperate regions, such as the UK, may accumulate high levels (>250 g/kg dry matter [DM]) of water-soluble carbohydrates (WSCs). Hays from such pastures may also contain substantial levels of WSC, often in excess of 200 g WSC/kg DM (Halling and others 2005). The WSCs largely comprise fructans (poly and oligosucrosyl fructose) and the simple sugars, glucose, fructose and sucrose. Pasture-associated laminitis in temperate regions is currently thought to be a sequel to over-consumption of WSC fractions, potentially associated with the development of chronic metabolic disorders such as insulin resistance, especially in overweight animals (Carter and others 2009, Geor and Harris 2009), or acute digestive disturbances associated with the rapid fermentation of fructans (Pollitt and others 2003). Indeed, laminitis can be induced clinically in healthy horses by administration as a bolus, after a short adaptation period, of inulin-type fructans at 5 g/kg bodyweight (Kalck and others 2009) or 7.5 g/kg bodyweight (Pollit and others 2003). Furthermore, Geor and others (2009) showed that ingestion of fructose (the main constituent of fructan) at 3 g/kg bodyweight per day resulted in reduced sensitivity to insulin in thoroughbred mares exhibiting insulin resistance.

In an attempt to reduce the intake of WSCs, owners of equids predisposed to laminitis often soak the hay intended to be fed to such animals in water. It has been shown that nutrients, including soluble carbohydrates, may be leached from hay during the soaking process (Warr and Petch 1992, Blackman and Moore-Colyer 1998, Watts and Sirois 2003, Cottrell and others 2005). However, published studies evaluating the effect of soaking hay on nutrient contents have tended to use very large volumes of water (1:20 to 1:50 w/v), prolonged periods of soaking or chopped hay, all of which may promote WSC loss. Furthermore, it is not known whether soaking hay straight from the bale in compacted ‘flakes’, compared with hay that has been shaken loose from the ‘flake’, affects the rate of nutrient leaching. A small pilot survey revealed that in practice, at least in parts of the UK, long-stemmed hay is often soaked in just enough water to submerge it completely (equivalent to a hay to water ratio of 1:10 to 1:15 w/v), for either 20 to 40 minutes, two to four hours or overnight. Therefore, the aim of this study was to emulate common practice and determine the amounts of total WSC, WSC constituents and crude protein (CP) leached from a range of different UK grass hays, either shaken free of the flake or left compacted in the flake when soaked in just sufficient water to allow total submersion (1:12 w/v) for various time periods.

Materials and methods

Nine different long-stemmed hays, produced and offered for sale to the equine industry in different regions of England and Wales in the summer of 2008 were obtained in December 2008. The hays (Table 1) were mixed species meadow hays or hays that contained predominantly ryegrass (Lolium species) or timothy (Phleum pratense). The study was carried out in mid-December 2008, a time of year when many horses are kept stabled for all or part of the day. Hay bales were weighed and measured and their density (ρ) was calculated (ρ=mass/volume). Two kilogram flakes of each hay were either shaken loose or kept in their compacted flakes, and duplicates were submerged separately in 65 litre plastic tubs containing 24 litres of tap water at ambient temperature (mean 8°C, range 5 to 10°C). Hays were sampled (approximately 100 to 150 g) at 20 minutes, 40 minutes, three hours and 16 hours after submergence. The original hays and soaked hay samples were dried in a forced draught oven at 60°C until a constant weight was obtained, milled to pass a 1 mm steel mesh and stored in 50 ml ‘Sterilin’ tubes at ambient temperature before chemical analysis.

Table 1

Hay description, origin, dry matter (DM), water-soluble carbohydrate (WSC) and crude protein (CP) contents and bale density

Chemical analyses

The total WSC content of dried and milled hays (approximately 150 mg) was determined in filtrates resulting from extraction in 25 ml deionised water for one hour with constant stirring, by the phenol-sulphuric acid method (Dubois and others 1956). Separation and quantification of WSC constituent fractions (fructan, sucrose, glucose and fructose) in seven of the loose hays was achieved by ion-exchange high-performance liquid chromatography (Cairns and Ashton 1993) and total nitrogen concentrations were determined in all the ‘loose’ hays using a Leco FP 428 nitrogen analyser (Leco Corporation). They were then expressed as CP (N×6.25).

Statistical analysis

Significant differences in nutrient contents of hays after soaking were determined by two-way analysis of variance (GenStat). Comparisons between treatments were made by least significant difference equations. The effects of shaking the hays free from the flake (loose) or leaving the hay densely compacted in the flake (dense) on hay DM and WSC contents averaged across all hays and soak times were compared. The effects of duration of soaking on the DM and WSC contents averaged across all hays (loose and dense) were also compared. Comparisons of the effect of duration of soaking on the mean fructan, sucrose, glucose and fructose contents averaged across all hays tested were also made. Relationships between WSC content and amounts of WSC leached were subjected to regression analysis.


The mean DM of the original hays was 827 g/kg (range 763 to 871), and the mean density of the bales was 69.5 kg DM/m3 (range 52.0 to 85.5 kg DM/m3) (Table 1). There were no overall significant differences in the DM or WSC contents of either loose or dense hays when submerged in water for up to 16 hours (data not tabulated). The mean DMs of all hays, both loose and dense, were 243, 192, 176 and 157 g/kg, after being soaked for 20 minutes, 40 minutes, three hours and 16 hours, respectively. The DM of hays was significantly reduced after 20 minutes of soaking, with further significant reductions in DM up to 16 hours after submergence (Table 2).

Table 2

Mean dry matter (DM) (g/kg) and water-soluble carbohydrate (WSC) content (g/kg DM) of all nine hays at each soak time

The original hays contained between 123 and 230 g WSC/kg DM (mean 171 g WSC/kg DM) (Tables 1, 2). Reductions in WSC content across all nine hays were significant (P<0.05) when soaked for three or 16 hours (Table 2). After soaking for 20 minutes, 40 minutes and three hours, hays had lost a mean of five, nine and 16 per cent of their total WSC content, respectively. After being submerged for 16 hours, the mean loss of WSC from hays was 27 per cent (Table 2) (range 6 to 54 per cent [Table 3]). There was no relationship between the WSC content of the original material and the degree of loss at any time period.

Table 3

Water-soluble carbohydrate (WSC) content (g/kg dry matter [DM])* of the original hays and after different periods of soaking, when shaken free of the flake (loose) or when left undisturbed within the flake (dense)

Due to the lack of effect of soaking on water retention or nutrient leaching from loose compared with dense hays, further analyses on the effects of soaking hay on individual WSC fractions and total CP content were confined to the hays shaken loose from the flake.

The mean losses of WSC fractions, fructan, sucrose, glucose and fructose from all hays tested are shown in Table 4. There was incomplete separation of some of the different WSC fractions from samples of hays three and seven, rendering it difficult to unequivocally quantify each WSC constituent. Therefore, the data for these two hays were not included. Fructan contents of the original hays ranged from 71 to 131 g/kg DM (mean 106 g/kg DM), those of sucrose ranged from 8.8 to 27.1 g/kg DM (mean 16 g/kg DM), glucose contents ranged from 14.8 to 31.5 g/kg DM (mean 22 g/kg DM) and fructose contents ranged from 22.5 to 60.5 g/kg DM (mean 39 g/kg DM) (Tables 4, 5).

Table 4

Effects of soaking for up to 16 hours on the mean water-soluble carbohydrate (WSC) constituents (g/kg dry matter [DM])* of seven different hays shaken loose from the flake

Table 5

Water-soluble content (WSC) constituent contents (g/kg dry matter [DM])* of seven of the original hays shaken loose from the flake and after submergence in tap water for up to 16 hours

After 40 minutes of submergence, hay fructan contents were significantly lower (P<0.05) than from the unsoaked hay and the majority of the diffusible fructan appeared to have leached into the liquor after three hours of submergence (Table 4). The sucrose and fructose contents of hays were significantly reduced (P<0.05) after 16 hours of soaking, whereas hay glucose contents were significantly reduced (P<0.05) after a three-hour soak (Table 4). The mean percentage losses of fructan at 20 minutes, 40 minutes, three hours and 16 hours after submergence were 9, 11, 21 and 23 per cent. Corresponding percentage losses for sucrose were 6, 25, 31 and 45 per cent, for fructose they were eight, 15, 26 and 41 per cent, and for glucose they were 10, 27, 46 and 56 per cent.

The losses of WSC constituents from individual hays are shown in Table 5. Fructan losses varied markedly between individual hays, ranging from 5 to 20 per cent at 20 minutes and from 14 to 31 per cent after 16 hours. Likewise, there was a large range of losses of sucrose and glucose from different hays. Thus, all of the sucrose and glucose had leached from hay 9 by three hours, whereas only 21 and 29 per cent of these sugars, respectively, had leached from hay 1 by three hours (Table 5). With the exceptions of hays 1 and 9, a mean of 60 per cent of the fructose remained in the hay after 16 hours of soaking. In hays 1 and 9, however, only 30 and 40 per cent, respectively, of the fructose remained in the hays after being soaked for 16 hours.

CP contents of the original hays ranged from 30 to 86 g/kg DM (Table 1). The overall effects of soaking hay for up to 16 hours on hay CP content were variable and not significant. Thus, slight increases and decreases were apparent in the CP content of hays after soaking for up to 16 hours, such that the CP contents of hays 3, 4, 6 and 8 had increased by 2, 18, 16 and 20 per cent, respectively, whereas those of hays 1, 2, 5 and 7 had decreased by 33, 8, 4 and 14 per cent, respectively (data not shown).


This study serves to illustrate that although soaking hay results in reductions of WSC content, the extent of the losses was highly variable between hays, and not related to their initial WSC content.

The hays absorbed water rapidly; within 20 minutes of submersion, they were a mean of approximately 4.4 times their original weight and after 16 hours, they were approximately five times their original weight. Thus, when fully submerged, a feed of 6 kg air-dry hay would become 26 kg after 20 minutes and 30 kg after 16 hours of soaking. Hay soaked even for short periods, therefore becomes very heavy for handling and subsequent feeding to equids. It is not known whether the high water content of soaked hay affects the subsequent voluntary DM intake by equids.

Although the hay bales varied in density, with the least dense bale being approximately 60 per cent of the density of the most dense bale, there was no significant relationship between bale density and hay water uptake.

There was no overall significant difference in either the DM or WSC contents of loose or dense hays after soaking. This suggests that water is capable of penetrating the hay quickly when it has been baled at the densities reported here. The noted lack of effect of the density of the hay on total WSC egress might indicate that hays should be left in the flake when soaked, to reduce the release of respirable particles into the airspace, such as spores from moulds and yeasts, which can pose a potential health hazard to both horses (Clarke 1987) and those handling the hay (Gregory and Lacey 1963).

The total WSC contents of the original hays determined in this study were within the range reported for grass hay within the UK (Ministry of Agriculture, Fisheries and Food [MAFF] 1992). The amount of WSC leached from the hays was not related to the treatment (loose or dense hay) or to the total WSC content of the original hay. The percentage losses reported in the present study are consistent with the findings of other authors who have reported mean losses of sugar or WSC, after soaking hays for 30 minutes, of 6 per cent (Blackman and Moore-Colyer 1998), 12 per cent (Watts and Sirois 2003) and 17 per cent (Warr and Petch 1992). After 12 hours of submergence, Warr and Petch (1992) reported losses of 30 to 48 per cent. The tendency towards higher values reported by Warr and Petch (1992) and Watts and Sirois (2003) may reflect the higher soak water volumes used by these authors. In addition, the water was warmer (28°C) in the study by Watts and Sirois (2003) than in the present study (8°C) and the hay was chopped, which may have aided the egress of sugars from the hays. The amounts of WSC lost from the different hays in the present study varied greatly, which has been a common feature in other studies. For example, Watts and Sirois (2003) reported sugar losses of 12 to 45 per cent from six grass hays after 60 minutes of soaking, and in the study by Warr and Petch (1992), 30 to 48 per cent of the WSC had leached into the liquor from three hays after 12 hours of soaking.

Of the WSC constituents of the original hays, fructan was the most abundant, accounting for a mean of 57 per cent of the total WSC fraction. Corresponding values for the simple sugars were: fructose 21 per cent, glucose 12 per cent and sucrose 8 per cent. The predominance of fructan in the WSC fraction of grass has been noted elsewhere (Ince and others 2003). Sucrose is frequently the next most abundant WSC constituent in fresh grass. However, during haymaking, only grass at the top of the swath receives much light, and thus the sucrose in grass lower down the swath may be depolymerised to glucose and fructose, as under aerobic conditions in the dark, grass sucrose is degraded rapidly (Cairns and others 2008).

Fructan can be used to clinically induce laminitis in healthy horses by administration of a bolus of oligofructose. In the present study, soaking hays reduced fructan contents by a mean of 21 per cent after three hours with little further loss by 16 hours. Three of the hays contained in excess of 100 g fructan/kg DM after prolonged soaking. If such soaked hays were fed at 2 to 2.5 per cent of bodyweight per day (DM basis), animals would ingest approximately half of the amount of fructan used to clinically induce laminitis in healthy horses. However, there is increasing evidence that voluntary DM intake in horses can be in excess of 5 per cent of their bodyweight per day (Argo and others 2002, Smith and others 2007), and therefore it would be unwise to allow ad libitum access to such soaked hay for equids predisposed to laminitis.

The total CP content of the hays was generally low, ranging from 30 to 86 g/kg DM, the lower value being more akin to the CP content found in straw (MAFF 1992, Dairy One 2003). Hay 8, which had the lowest CP content, was very mature, stemmy hay, with little apparent leaf content.

According to current recommendations, a horse with a bodyweight of 500 kg requires 630 g CP/day to meet its maintenance protein requirements (National Research Council [NRC] 2007). Only four of the nine hays examined in the present study would have met or exceeded the maintenance protein requirements of such a horse being fed a hay-only diet at 2 per cent of its bodyweight per day. Indeed, a horse being fed hay 8 as its sole feed would have received less than half of its maintenance protein requirement. If such low protein hays were to form the majority of the diet for horses, then a high-quality protein supplement should be fed to redress deficiencies.

The effect of soaking hays on CP-content determination was variable, with apparent losses and gains in CP from the different hays virtually cancelling each other out. Such equivocal results were reported by Moore-Colyer (1996), where CP contents of soaked hays apparently increased with soaking. Likewise, summation of the amounts of CP present in both soaked hays and the resultant liquors gave total CP values that were 16 to 23 per cent greater than that in the original hay (Warr and Petch 1992). This may indicate that soaking in some way allows greater extractability of the total nitrogen, leading to relatively greater values than the unsoaked material.

The effect of prolonged soaking on the vitamin and mineral status of the hays was not examined in the present study. However, Blackman and Moore-Colyer (1998) reported losses of potassium of approximately 40 per cent from hays after soaking for 30 minutes. Further studies are required to determine whether prolonged soaking results in continued losses of potassium and to investigate the effects of prolonged soaking on the content of other minerals and nutrients from hays.

The concurrent presence of insulin resistance and pasture-associated laminitis in equids is increasingly being recognised. There is also mounting evidence of the adverse effects of an elevated intake of fructose on human health and the development of insulin resistance (Basciano and others 2005) and elevated fructose intake has also been linked to insulin activity in rats (Zavaroni and others 1980). It was recently demonstrated that administering fructose, but not glucose, at 3 g/kg bodyweight per day significantly reduced insulin sensitivity in mares with insulin resistance, but not in healthy mares (Geor and others 2009). The same study showed a statistical trend for a similar effect with inulin-type fructan. This may suggest differential and heightened sensitivity to inulin and fructose in animals predisposed to metabolic dysfunction compared with their healthy counterparts. In the present study, fructose was present in the original hays at up to 60 g/kg DM, and although substantial amounts were lost from several of the hays after prolonged soaking, fructose was generally the least readily leached of the simple sugars. In hay 2, it still constituted up to 43 g/kg DM after having being soaked for 16 hours. However, for animals fed at 2 to 2.5 per cent bodyweight per day, this level of fructose would be approximately one-third of that used to elicit decreased insulin sensitivity in insulin-resistant mares. It should be noted that it appears from indirect studies that some inulin-type fructans can undergo at least partial degradation in the equine foregut (Coenen and others 2006), the resultant hydrolysates potentially adding to the free fructose pool from ingested herbage. If this is the case, then it is possible that elevated fructose levels could arise within the gut, conceivably aiding the development of insulin resistance in susceptible equids.

A minimum ratio of 1:12 (w/v) hay:water was required to entirely submerge the hays in the 65 litre tubs. Thus 144 litres of liquor per day would result from soaking 12 kg of hay for a 500 kg horse fed at 2.5 per cent bodyweight per day. The resultant liquor from hay 9 in the present study would provide approximately 960 g WSC, of which 420 g would be fructan and approximately 180 g each of sucrose, glucose and fructose. This would provide a readily available carbon source for microorganisms, and would constitute a potent pollutant. Indeed, the biological oxygen demand of the liquor produced from soaking hay for 12 hours in the study by Warr and Petch (1992) was substantial and many times greater than raw sewage. It is clear that hay-soak liquor should be disposed of responsibly, and on no account should it be drained into watercourses or static ponds.

In conclusion, soaking hays for 16 hours at a mean temperature of 8°C resulted in significant losses of total WSC and all of the constituent WSC fractions. However, there was great variation between individual hays in the amounts of WSC leached. Franks (2009) recommended a maximum threshold of 120 g non-structural carbohydrates/kg forage DM for obese, laminitic and insulin-resistant horses, which is equivalent to the 100 g WSC/kg forage DM commonly recommended as the upper threshold for WSC contents of diets for equids predisposed to laminitis (Genrick 2008). Soaking for 16 hours resulted in only two of the hays containing less than 100 g WSC/kg DM, and two of the hays still contained in excess of 160 g WSC/kg DM.

When formulating diets that include soaked hays for equids predisposed to laminitis, hays with an initially low WSC content should be chosen and reanalysed after soaking. As there was no apparent benefit of shaking hay loose from the flake on the egress of WSC, the authors recommend that hays should be soaked in compacted flakes, to reduce the potential negative impact of increased airborne respirable particles on human and equid respiratory health.

Further work on the methods of improving the leaching of WSC from hays may aid in the provision of suitable forage for lamimitis-prone equids. Additionally, studies on the effects of different grass species, hay-making processes, maturities of the crop, time of day and weather conditions at harvest on subsequent WSC losses may contribute to the understanding of factors affecting leaching of WSC from soaked hays. The effects of prolonged soaking on the leaching of other nutrients from soaked hays is also worthy of investigation, to allow appropriate supplementation of any depleted but essential nutrients.


View Abstract


  • Provenance not commissioned; externally peer reviewed

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.