Grain-based meals increase work and growth potential, but those of 5 lb or more incur an increasing risk of digestive disturbances involving rapid fermentation and metabolic disorders involving insulin resistance. They prompt a feeding-fasting cycle of metabolites and hormones that influences growth and skeletal development. Replacing grain with forage limits performance. Adding fat to an already balanced concentrate risks multiple incipient deficiencies. Including fat in a balanced formulation is safer but raises energy density, complicating feeding management. A family of fat-and-fiber feeds under development at Virginia Polytechnic Institute and State University is unusual because it has sufficient fat to be used like common concentrates and sufficient fiber to be fed as 100% of the ration versus 50% to 75% for traditional grain-based concentrates.
Grain is fed to provide more energy for growth and work. Improved performance is commonly observed up to a point until health becomes compromised by equine grain-associated disorders (EGAD)—a group of digestive disturbances involving rapid fermentation and a set of metabolic disorders involving insulin resistance (Table 1 and Table 2).1-18
This article highlights salient findings, critiques certain reports, and appraises alternatives to grain.
Forage and Grain
Horses evolved as grazing, hindgut-fermenting herbivores. Their staple (i.e., main feed energy source) was forage. Conserved seed was used for its logistic advantages by ancient armies and merchants. Grain feeding became more widespread about 350 years ago when improved pastures and NPK (nitrogen, phosphorus, potassium) fertilizers were introduced, and farm horses were expected to work harder. At about the same time, racing and other equine sports rose in popularity. Increasing demands for performance focused attention on the low energy in typical forages (e.g., pasture, hay), which provide about 1.8 to 2.2 Mcal (1000 kcal) of digestible energy (DE) per kg of dry matter (DM). This limitation was partially overcome by supplementing with grains, which contain 60% to 85% starch and about 3.2 to 3.9 Mcal/kg (hence the term concentrate).10
A bioenergetic model, which accumulates a series of typical efficiency factors for digestion and metabolism of each energy source in feeds,19 indicates that a ration of hay and oats (50:50), compared to hay alone (100:0), enables greater energy intake and has higher digestibility. An elite 1100-lb horse would need to consume 55 lb of hay, which is unrealistic, or 40 lb of hay and oats, which is possible with four meals of oats and access to hay most of the time, to yield 15 Mcal of net energy for competitive performance (Table 1). The predicted difference in bowel ballast is 60 lb (27 kg),19 a huge competitive advantage for using grain.
The National Research Council (NRC) recommended that concentrates should form 70% and 65% of the equine diet for growth and intense work, respectively.10 It cited experimental evidence associating high intakes of soluble carbohydrates with developmental orthopedic disease (DOD). Rapid growth was associated with DOD in one report, but not three others. The NRC also commented that muscle glycogen loading might improve performance, but with a greater tendency for exertional rhabdomyolysis (ER). The 1989 NRC10 stated that "[n]onstructural carbohydrates that escape prececal digestion…are subjected to anaerobic fermentation, largely in the cecum and colon." It gave no explicit warning, however, about rapid fermentation causing digestive disturbances (Table 1).
Fats and Oils
The 1989 NRC10 reviewed a dozen reports about fat and concluded that "fats and oils... can be added in limited amounts to equine diets. They will reduce dustiness and add significantly to the DE content." Subsequently, clearer evidence has emerged concerning beneficial effects of fat adaptation (i.e., a set of physiologic adaptations to feeding a higher fat diet during physical conditioning) on exercise performance.19
Bowel ballast, fecal weight and water, and water needed for heat elimination are reduced by replacing forage with grain (Figure 1) according to a bioenergetic model.19
These advantages for exercise are further enhanced by replacing 10% of hay and oats with vegetable oils (Figure 2)19
Predicted daily feed intakes are 55, 40, and 31 lb (25, 18, and 14 kg) DM for three diets of hay:oats:oil with proportions of 100:0:0, 50:50:0, and 45:45:10, respectively. According to the model,19 the decrease in bowel ballast from a 50:50:0 diet to 45:45:10 is about 30 lb (13.6 kg).
Metabolic studies have shown that fat adaptation increases oxidation of fatty acids and spares glucose and glycogen. Fat adaptation leads to less production of acid and heat during exercise.19 These metabolic advantages are further increased by combining dietary protein restriction with fat adaptation.20 Improvement in performance due to fat adaptation has been predicted biochemically only for long, slow work because fatty-acid transport into mitochondria is rate limiting.
Fat-adapted horses also turn on the high power source (anaerobic glycogen breakdown to lactate) when needed for sprinting. Faster times were recorded for fat-adapted horses in 1600 m on the track, mainly because of a faster first 200 m.21 Also, greater maximal accumulated oxygen deficits, longer run times to fatigue at 115% maximum oxygen consumption (Vo2max), and higher peak plasma lactate concentrations were observed in fat-adapted horses.22 The lactate threshold was increased by 12 weeks of conditioning and a high-fat diet.23 Increased plasma lactate during exercise is moderated during slow work but augmented during hard work in fat-adapted horses, which indicates improved metabolic regulation. This facility in fat-adapted horses has not been reported in other species.
Fat-adapted horses exhibited decreased spontaneous activity measured with pedometers and reactivity evaluated by startle tests.24 This calming effect may contribute to the effectiveness of high-fat diets in controlling some forms of ER.
In short, the fat-adapted equine athlete is calmer, lighter, more efficient, and better regulated metabolically; thus its performance should improve in most athletic events. Superior performance over long distances is predictable biochemically,19 but the advantages during intense work are surprising and have been demonstrated so far only in horses.20-23
Evidence that large grain-based meals, rapid starch fermentation, and an exaggerated feeding-fasting cycle of metabolites and hormones compromise health is based on epidemiologic studies, clinical trials, and physiologic experiments (Tables 1 and 2).
In the 1990s debate about health claims for foods (i.e., functional foods, medical foods, supplements), a differentiation was made between preventing or avoiding disease. Avoid corresponds to reducing or eliminating risk factors of a disease (associations found in epidemiologic studies). Good management, for example, can circumvent many risk factors of colic,1 thereby tending to avoid but not prevent the disease. In contrast, the claim of prevention is stronger and usually requires an empirical trial. For example, replacing about 25% of DE from starch with triglyceride interrupted the recurrence of recurrent ER in 16 of 19 horses after a 3- to 6-month adaptation.11 (These horses served as their own controls; thus the strength of this historic control depends on the consistently recurrent nature of these cases of ER.)
A nutritional health claim may also be based on anatomic and physiologic studies relating to pathogenesis. Feeding a grain-based meal (about 5 lb [2.3 kg] for an 1,100-lb [500-kg] horse) leads to profound circulatory, digestive, metabolic, and hormonal changes.2,25,26 Larger meals that exaggerate these changes may thereby contribute to digestive disturbances and metabolic disorders.
To determine the small intestine's capacity to digest starch, two horses were fed 1.4 kg (3 lb) of alfalfa hay and chopped corn to provide starch amounting to 0.20% to 0.55% of body weight (BW) per meal.25 Eleven meals providing up to 1,950 g of starch yielded 100 to 700 g of starch entering the cecum; however, five meals of 2,000 to 2,750 g of starch yielded 1,200 to 2,300 g of cecal starch. These investigators recommended limiting the starch in a meal to no more than 0.4% of BW (e.g., no more than 2,000 g of starch, or about 6 lb [2.7 kg] of grain, for an 1,100-lb [500-kg] horse). Other investigators recommended no more than 2 g of starch/kg (i.e., 0.2% starch, or about 3 lb [1.4 kg] of grain per meal) to avoid the risk of "dysbiosis" in the cecum.26
Cecal starch is rapidly fermented to lactic acid, which has a pK of 3.8 (compared with 4.7 for acetic acid); therefore, lactic acid is poorly absorbed and accumulates in the lumen.2,4 Lactate attracts water, setting the stage for osmotic diarrhea and cecal distention. Rapid fermentation also yields much gas, which may contribute to cecal distention.
Accumulating lactic acid lowers the pH below 6, lysing bacteria and releasing endotoxin, which may contribute to one form of laminitis.7 Acidic irritation of the mucosa may cause adhesion or entry of pathogenic bacteria and inflammation of the cecum and colon.
Feeding intermittent grain meals also generates more acidic conditions in the stomach than are found in a horse nibbling hay.5 A grain meal also raises serum concentrations of gastrin and insulin.6
The starch guides of 0.2% or 0.4% of BW have been directed mainly at avoiding exuberant fermentation.25,26 They are crude pointers for this purpose because the rate of starch digestion varies with type of grain, starch granule characteristics, and processing. The value of these starch guides is enhanced, however, by taking a wider view: More starch hydrolysis in the small intestine yields not only less starch escaping to the cecum but also more glucose entering the portal blood. The trade-off is between fermentative disturbances and the metabolic disorders of EGAD.
Deviations in insulin sensitivity are probably involved in the conditions listed in Table 2. A cluster of abnormalities in humans (i.e., syndrome X27) is associated with insulin resistance and a set of risk factors (e.g., obesity, dyslipidemia, hypertension) for non-insulin-dependent (type 2) diabetes mellitus and atheromatous cardiovascular disease. Genetic and other predispositions to insulin resistance are aggravated by a high-carbohydrate diet.28 The name changed to metabolic syndrome X, then simply metabolic syndrome,29 with progressively less emphasis on the low-fat, high-carbohydrate diet, which has been promoted as heart protective for about 40 years by several public health organizations.
The conditions listed in Table 2 have been called equine syndrome X.30 Peripheral Cushing's disease has been called the metabolic syndrome.9 The use of low-glycemic feeds has been recommended for managing these conditions.9,30
Feeding a meal of grain but not hay raises plasma concentrations of glucose and insulin for 4 to 6 hours in horses.31 The responses to grain are approached by a grain:hay (1:1) mixture,31 but they are greatly reduced by replacing most of the grain (starch) with fat and fiber32 (Figure 3).
Daily feeding of two grain-based meals sets up a feeding-fasting cycle of metabolites and hormones,33 which is alien to the nutritional heritage of horses. Elements in the feeding-fasting cycle—notably glucose, insulin, growth hormone, and insulin-like growth factor (IGF-1)—have been suggested to contribute to DOD.9,30,33
Developmental Orthopedic Disease
The collective term DOD includes clinical manifestations of dyschondroplasia, osteochondrosis, osteochondritis dissecans (OCD), and perhaps other pathologic entities. Predisposition to OCD is about one-third heritable and two-thirds environmental, with the best-identified factors being exercise (beneficial or traumatic) and nutrition. Nutritional factors include minerals (copper deficiency, calcium excess), vitamins (vitamin A excess), and energy sources (as argued here, starch and sugar).
Experimentally, OCD incidence was higher in horses fed DE at 130% the NRC recommendation than in those fed at 100%, with the extra DE provided as fat and starch.34 Other studies have focused attention on DE derived from glucose equivalents (mainly starch and sugar) in the etiology of OCD.14,35,36
Administration of glucose equivalents is the first step of two common tests of glucose metabolism—glucose tolerance and glycemic index. These tests share similar procedures but yield different information: The glucose tolerance test gives information on a test horse; the glycemic index test gives information on a test meal.37
In the glucose tolerance test, the magnitude of the response is usually measured as the time integral of concentration increments (i.e., observed minus baseline concentrations), which is approximated by the area under the curve (AUC):
AUC = mmol x L-1 x min-1
The glycemic index is calculated as follows37:
Glycemic index (expressed as a percentage) = Test AUC ÷ Standard AUC
In the OCD field,14,35,36 these tests have been oversimplified procedurally and confused conceptually as follows.
Plasma responses of glucose and insulin following a meal of sweet feed and hay were higher in four young standardbreds with radiographically diagnosed OCD than in 11 controls (Figure 4).14
Because these horses were in the same environment, the differences in responses were presumably more genetic than environmental. These and similar data served as a basis for patenting a diagnostic test for a predisposition to OCD.35
The diagnostic test compares an oral glucose tolerance test (OGTT) on an individual to tolerances of an OCD-free population.35 The "glucose challenge" according to the patent is any solid or liquid providing glucose sufficient to raise blood glucose concentration by 30% to 50%. The example is a 50:50 mixture of sweet feed and alfalfa hay. Instead of the AUC, the patent uses peak concentration or the rate of change in concentration as a measure of the response. Thus precision is less for both input and output in the OCD diagnostic test than in a routine glucose tolerance test.
A simpler modification of this diagnostic test became, in effect, a glycemic index test (GIT) of feeds on six thoroughbred farms.36 The incidence of OCD treated surgically was correlated with concentrations of glucose and insulin in a single sample of plasma taken 2 hours after a concentrate meal (standardized regarding nonstructural carbohydrates). The distribution of OCD incidences was uneven (i.e., 0%, 7%, 8%, 12%, 17%, 32%),37 which complicates regression analysis. The authors applied a linear regression, but reevaluation of the data suggests that a logarithmic or exponential relationship between OCD incidence and farm glucose response is more valid (Figure 4).36
The glycemic index was not diagnostic; it did not differentiate between affected and unaffected individuals on each farm. Consequently, the OCD predisposition was suggested to be environmental—an adaptation to the feed—rather than genetic. The authors reached the conclusion that young horses should be given a low-glycemic diet to avoid predisposition to OCD.
In the feeding-fasting cycle, changes in glucose and insulin are initial events. Subsequent changes in the somatotropic axis, predominantly growth hormone and IGF-1, exert more direct effects on chondrocyte proliferation and maturation according to many studies in other species.38 Thus we have proposed that exaggerated responses of growth hormone and IGF-1 are likely to be involved in the pathogenesis of dyschondroplasia, osteochondrosis, and OCD.30,33 Greater fluctuations in glucose and insulin are associated with higher daily levels of plasma IGF-1 concentration.33 These responses are attenuated by replacing starch and sugar with fiber and fat.30,33 These studies,14,30-33,36,37 when considered together, suggest that a genetic predisposition to some forms of DOD may be exacerbated by an environmental trigger (high-glycemic feed), which can be avoided by feeding less grain.
Current research is discriminating between sporadic ER and two forms of recurrent ER. One recurrent form of ER is found mainly in thoroughbreds. It involves neuromuscular conduction.13,39 Another recurrent form, polysaccharide storage myopathy, affects mainly cold- and warmbloods and quarter horses. It involves increased muscle glycogen synthesis and, perhaps, glucose uptake.12,40
The mode of inheritance is autosomal dominant with variable expression in recurrent ER39 and autosomal recessive in storage mypothy.40 Heritability (i.e., inherited variation divided by total variation) has not been estimated for any form of ER. Nevertheless, the causes of both forms appear to involve the interaction of genetic predispositions and environmental factors, and both are mitigated by replacing dietary starch and sugar with fiber and fat.11,13,41"43
Some animals affected by ER exhibit the syndrome every time or nearly every time they are exercised; thus they may serve as their own control. The first clinical report in which substituting fat for carbohydrate eliminated ER involved a team of racing sled dogs that was being carbohydrate loaded.41 The dogs were tying up on every run, but this stopped on the first run after a switch from a dry cereal diet to a canned meat product containing about 70% fat on an energy basis. In contrast, the timeline has been a matter of weeks and months, rather than the next day, in studies on ER in horses. Diets with digestible carbohydrates (i.e., starches and sugars) restricted to less than 15% DE and fat fortified to greater than 20% DE have led to elimination of recurrence of polysaccharide storage myopathy in 84 of 90 affected horses.42 In another group of 19 horses with undefined forms of recurrent ER, episodes were absent in 16 and mild in three after 6 months.11 In a switchback trial with 3-week intervals on five thoroughbreds with recurrent ER,43 mean postexercise serum creatine kinase activity was seven times higher while consuming a feed containing 40% and 5% DE from starch and fat, respectively, than when consuming a feed with 7% and 20% DE from starch and fat, respectively. The authors proposed five possible mechanisms, which were not mutually exclusive, for the beneficial effect. The most relevant possibility is the calming effect of a high-fat diet, which reduces plasma cortisol, a major anti-insulin hormone likely to contribute to insulin resistance.
Cereal grains are excellent horse feeds, but, as with human concentrates (e.g., butter, whiskey, sugar, salt), their use requires judicious care. A zone of increasing risk exists between the recommended safe meal of 5 lb25,26 (2.3 kg) and the 20-lb (9.1-kg) meal of grain that approximates the dose of starch and cellulose for inducing the well-known laminitis model.7 In one preliminary study,44 half the usual laminitis-inducing dose was given to healthy horses. Mild signs of pain were observed in some, but not all, horses in the hind region and were often difficult to localize with certainty in the abdomen, hindquarters, or feet. These variable, insidious, prodromal signs were awkward to summarize statistically but probably reflected inconsistent subclinical digestive disturbances.
While admitting that more precise data are needed on the effects of grain meals of 5 to 20 lb (2.3 to 9.1 kg), we recommend that the number, rather than the size, of meals should be increased to achieve desired high intakes of grain. Restricting grain-based meals to less than about 5 lb (2.3 kg) should almost certainly avoid digestive disturbances.2,25,26 Digestive upsets (Table 1) are usually overt, and the responses to grain restriction usually immediately evident, thereby reducing the incentive for costly controlled clinical trials.
Restricting grain and simply allowing more forage is inexpensive, safe, and effective but reduces energetic efficiency and performance potential.19 The usual upper limit of hay intake is about 2% BW,10 and such high loads result in accumulation of dead weight in the large bowel (Table 1) and, possibly, a round "hay belly."
Gastric ulceration in horses exhibiting little or no overt clinical signs appears to have both digestive and metabolic components. Dietary management has aimed to promote nibbling because gastric acidity is highest during intervals between meals.6 Antacids are claimed to reduce putative clinical manifestations but not ulcers per se.
Among the metabolic disorders, only ER and growth fluctuations that risk dysflexia16 have been demonstrated to respond to fat fortification and fat and fiber, respectively.11,15,41-43 For ER, the dose of vegetable oil is about 1 g/kg BW (about 1 lb/1,000 lb) or about 25% of DE.11,41 The benefits of fat adaptation become fully evident in 3 weeks to 6 months. The long-term danger of adding fat ("empty calories") to a previously balanced feed is the lowering of concentrations of essential nutrients on an energy basis, which risks incipient deficiencies.
Horses often reject dietary fat. Preference tests favor corn oil over many vegetable oils and animal tallows.45 Too rapid of an introduction can lead to shiny, greasy, and grayish stools or, eventually, loose steatorrhea. In our experience, smooth acceptance and formed stools are obtained when supplementary corn oil or a high-fat feed is introduced in 25% increments over 4 days.
Accommodation to increased dietary fat takes a few days to about 3 weeks, during which the lipolytic capacity of the small intestine appears to increase. In digestibility trials of various fats added up to 230 g fat per kg of a basal concentrate, apparent digestibility was 95%, true digestibility 100%, and endogenous fecal fat about 60 g/day.46 No adverse associative effect was found; undiminished fiber digestibility confirmed that no fat reached the large bowel.
Corn oil is usually measured by the cupful for horses. A standard measuring cup can hold 8 fl oz (250 ml) of water and 200 g (1.8 Mcal) DE of corn oil.10,47 An often-used routine with healthy horses is to start with one-quarter cupful per meal and to increase the oil in steps of 2 or 3 days to 1 or 2 cupfuls/day in 2 or 3 weeks, which is about 10% or 20% of the maintenance DE of an 1,100-lb (500-kg) horse.
For hypometabolic sick horses that have an energy expenditure about half-maintenance, the introduction is more gradual. For hypermetabolic sick horses with energy expenditures up to three times maintenance, oil introduction is more aggressive and may reach a total of 2 cupfuls, usually split into many doses, on the second or third day. A hypophagic, hypermetabolic sick horse metabolizes its own fatty and amino acids—nutrients that should comprise a large share of exogenous energy sources.
For prolonged use, oil or fat should be incorporated into a feed that is balanced on a DE basis for all essential nutrients. The NRC's Nutrient Requirements of Horses10 is like the corresponding volumes for dogs and cats: It provides mean minimum requirements rather than allowances like those for livestock and humans. If these requirements are followed exactly, feeds are formulated that are adequate only for horses whose requirements are average or below. The need for nutritionists with sufficient equine knowledge to devise their own set of nutrient allowances (with margins over requirements) is especially important when innovative or complex feeds are being formulated.
The fat content is 2% to 4% of DM in most unfortified concentrates, about 6% to 10% in most high-fat concentrates, and 15% to 25% in a few.42 Most high-fat feeds have less than 20% neutral detergent fiber, which is insufficient for use as complete feeds, and none has fiber in a long form (2 cm or more). Fat fortification raises the energy density by 10% to 50% (i.e., from the typical 3 Mcal/kg up to 4.5 Mcal/kg); thus a traditional concentrate should be replaced by commensurately less weight of a high-fat concentrate. More forage should be fed with a high-fat concentrate; in this way, more fiber is fed, as well as more fat, when a high-fat concentrate replaces a traditional grain-based concentrate.
Over the past decade, a series of studies at Virginia Polytechnic Institute and State University has developed a family of novel feeds featuring added fiber as well as fat so that the energy density is not raised above that typical of concentrates. These feeds contain sufficient fat to allow use as typical concentrates and enough fiber to enable exclusive use as complete feeds.15,20,23,24,30,32,33,47 All contain about 12% fat. They vary, however, in their carbohydrate profiles for specific purposes; examples are a feed for stall-fed athletes20 and a pasture supplement for reproduction and growth.32 This approach to energy sources has been described, including proteins and amino acids as well as carbohydrate profiles and fats.47 The primary experimental focus has been on performance, but the replacement of sugar and starch with fat and fiber also avoids molasses and grain.
Choosing to replace grains and molasses with fats and fibers will avoid EGAD but not necessarily prevent all presentations of disease complexes, such as diarrhea, colic, laminitis, DOD, and ER. Most of these syndromes are pleomorphic and multietiologic. The cause of each case may be viewed as the minimal set of events, conditions, and characteristics that are necessary and sufficient for the occurrence of the syndrome. Grain feeding is likely to be a sufficient causative factor in a few cases and a necessary factor in many, but not necessarily all. Controlling one causative factor for many common disorders is sound risk management—an avoidance maneuver, but not a prophylactic panacea.