Effect of feeding glucose , fructose , and inulin on blood glucose and insulin concentrations in normal ponies and those predisposed to laminitis 1

ABSTRACT: Identifi cation of ponies (Equus caballus) at increased risk of pasture-associated laminitis would aid in the prevention of the disease. Insulin resistance has been associated with laminitis and could be used to identify susceptible individuals. Insulin resistance may be diagnosed by feeding supplementary water-soluble carbohydrate (WSC) and measuring blood glucose and insulin concentrations. The aim of this study was to assess the glycemic and insulinemic responses of 7 normal (NP) and 5 previously laminitic (PLP), mixed breed, native UK ponies fed glucose, fructose, and inulin [1 g/(kg.d) for 3 d] or no supplementary WSC (control) in spring and fall after a 7-d adaptation to a pasture or hay diet. Blood samples were taken for 12 h after feeding on each day, and baseline and peak concentrations and area under the curve (AUC) for glucose and insulin were recorded. Linear mixed models were used for statistical analysis. Differences between PLP and NP groups were most marked after glucose feeding with differences in peak glucose (P = 0.02) and peak insulin (P = 0.016) concentrations. Season and diet adaptation also affected results. Peak concentrations of glucose and insulin occurred 2 to 4 h after WSC feeding. Peak insulin concentration was greater and more variable in fall, particularly in PLP adapted to fall pasture. Baseline glucose and insulin concentrations varied between individuals and with season and diet adaptation but were not greater in PLP than NP. Insulin AUC was greater in PLP than NP after feeding both glucose and fructose (P = 0.017), but there were no differences between PLP and NP in glucose AUC. Glycemic and insulinemic changes were less (P ≤ 0.05) after feeding fructose than glucose, although differences between PLP and NP were still evident. Minimal changes in glucose and insulin concentrations occurred after inulin feeding. Measurement of peak insulin 2 h after feeding of a single dose of glucose (1 g/kg) may be a simple and practical way to aid identifi cation of laminitis-prone ponies before the onset of clinical disease, particularly when ponies are adapted to eating fall pasture.


INTRODUCTION
Laminitis is a common, painful, and debilitating condition, which primarily affects ponies at pasture in the United Kingdom (Menzies-Gow et al., 2010).One of the major inciting causes may be excessive consumption of grass high in water soluble carbohydrate (WSC, sugars and fructans; Bailey et al., 2004a).Some individuals appear predisposed to laminitis, whereas others, often grazing the same pasture, are never affected.Accurate identifi cation of at-risk individuals would aid in instituting preventive countermeasures.Some, but not all, recurrently laminitic ponies are insulin resistant, which could prove useful in identifi cation (Treiber et al., 2006).However, not all insulin-resistant ponies are prone to laminitis (Jeffcott et al., 1986).
Differences in insulin concentration between normal and laminitis-prone ponies were exacerbated when inulin, a fructan, was included in the diet for 48 h at 3 g/(kg .d) divided into 3 daily feeds (Bailey et al., 2007).
Previously laminitic ponies had an exaggerated response, increasing their serum insulin concentration by 5.5-fold vs. a 2-fold increase in normal ponies.This indicates that insulin resistance and possibly laminitis susceptibility could be diagnosed by feeding supplementary fructan WSC.This could be used as a dynamic test to predict laminitis susceptibility by measuring the change in serum insulin concentration.The optimum time to measure the insulinemic response has not been determined, although this may vary with different situations, and the effects of season or basal diet on the response are unknown.The suitability of using other WSC (glucose or fructose) to stimulate insulin release in ponies has not been investigated.
The aim of the current study was to assess the metabolic responses to feeding supplementary WSC (glucose, fructose, or inulin) in a group of normal or previously laminitic ponies.The effect of season (spring or fall) and previous diet (pasture or hay) on the responses were also assessed.

MATERIALS AND METHODS
The protocol was approved by the Royal Veterinary College Ethics and Welfare Committee (Hatfi eld, Hertfordshire, UK) and was carried out under UK Home Offi ce license.

Animals
Seven normal (NP) and 5 previously laminitic (PLP) ponies were used.The ponies were all mixed UK native breeds (mainly New Forest, Welsh, and cross breeds) and were not related.The ponies were maintained within a closed research herd at the Royal Veterinary College.The NP had no history of laminitis within a 5-to 10-yr period.The PLP had recurrent bouts of pasture-associated laminitis within the previous 5 yr, which was diagnosed by experienced equine veterinary surgeons based on clinical signs and digital radiography where required.None of the ponies displayed clinical signs of active laminitis during the study.None of the ponies had clinical signs of pituitary pars intermedia dysfunction, and all showed normal cortisol suppression in response to a dexamethasone suppression test.All NP were mares and the PLP included 4 mares and 1 gelding.The mean age and BW for the NP were 19 ± 4 yr and 341 ± 71 kg, respectively, and for the PLP were 16 ± 3 yr and 289 ± 60 kg, respectively.The mean BCS on a scale of 1 to 9 (Henneke et al., 1983) of the NP was 5.4 ± 1.1 and PLP was 5.0 ± 1.0 and did not vary over the course of the study.

Study Design
The study had a randomized, 16-period crossover design.Feed trials were performed under 4 different seasonal and dietary adaptations: spring pasture, spring hay, fall pasture, and fall hay.Pasture-adaptation occurred after the ponies had been out on pasture for at least 7 d and hay-adaptation occurred after they had been group housed in a barn bedded with shavings and eating hay for at least 7 d.The study was performed in late spring (May to June) and late fall (October to November).Each individual pony was fed a control diet (no supplementary WSC) and each of 3 supplementary WSC (glucose, fructose, and inulin) under each seasonal and dietary adaptation.
The day before each feed trial began, the ponies were housed if they had been previously out on pasture, and an indwelling venous catheter was placed under mepivacaine local anesthesia (Intra-Epicaine, Dechra, Shrewsbury, UK).The ponies had free access to water and the same batch of soaked Timothy (Phleum pratense L.) hay at all times when housed.They were not fasted before supplementary WSC feeding, but had no access to concentrate feed before the fi rst (baseline) blood sample was obtained.Each pony was fed 1 g/kg of glucose (Dextrose Monohydrate, BHB, Lincoln, UK), fructose (Tate & Lyle Fruit Sugar, London, UK), or inulin (Orafti Group, Tienen, Belgium) mixed with 500 g commercially available chaff-based feed (Happy Hoof, Spillers Effem Equine, Ltd., Mars HorseCare UK, Ltd., Milton Keynes, UK; Table 1) once daily at 0800 h for 3 consecutive days (designated as a single feed trial).Control studies were performed where no supplementary WSC was fed and the ponies received only 500 g of chaffbased feed once daily at 0800 h for 3 d.After each 3 d feeding trial, there was a washout period of 7 d when the ponies ate only grass or hay.
Blood samples (20 mL at each sampling point, and total volume of 440 mL over 3 d) were obtained from the catheter at baseline (before feeding) and for 12 h after feeding on each day of the feeding trial (d 1 at 0.5, Table 1.Nutritional analysis of commercial chaff-based feed used during the study (as-fed) 1  1, 1.5, 2, 2.5, 3, 4, 6, 8, 10, and 12 h; d 2 and 3 at 2, 4, 8, and 12 h).Blood was sampled over a 12-h period to obtain maximal information on the time course of insulinemia and glycemia after feeding, as it is possible that insulin-resistant animals may have a delayed return to normal blood glucose and insulin concentrations.Blood for measurement of serum insulin concentrations was collected into plain vacutainers and allowed to clot at 37°C for at least 20 min.Fluoride-oxalate vacutainers were used to obtain plasma for glucose measurement and were kept on ice once collected.All samples were centrifuged (3,000 × g) for 10 min at 4°C and serum or plasma samples were then stored at -80°C until analysis.

Sample Analysis
Glucose was analyzed using the glucose-oxidase method by a laboratory (Clinical Pathology Laboratory, Royal Veterinary College).Insulin was analyzed using a commercially available RIA kit designed for use in humans (Coat-A-Count, Siemens, Surrey, UK) previously validated for use in horses; Reimers et al., 1982) and also validated in our laboratory.Intra-and inter-assay repeatability were 6.5 ± 5.1 and 7.4 ± 3.4%, respectively.All samples were assayed undiluted and dilutions were used to bring insulin concentrations into the working range of the RIA (up to 389 μIU/mL) where appropriate using insulin-depleted equine serum (IDS) prepared in our laboratory.Dilutional parallelism was observed when IDS was used to dilute serum samples (recovery on dilution, 96 ± 27%) but not when the zero standard supplied by the manufacturers was used for dilution (recovery on dilution, 72 ± 15%; Borer et al., 2012).

Measurements and Statistics
The baseline and peak concentration of glucose and insulin reached during each feeding trial were recorded.The area under curve (AUC) for glucose and insulin for the 3 d in each feeding trial was calculated based on the trapezoid rule.Normality of the distribution of the data was assessed using visual inspection of histograms and the Kolmogorov-Smirnov test (GraphPad Prism, GraphPad Software, San Diego, CA).Results were log transformed before statistical analysis if they were not normally distributed or if the plot of the residuals vs. predicted values was not evenly distributed.All results except baseline glucose concentrations were log transformed before statistical analysis.Linear mixed models were used to analyze the data (PASW Statistics, SPSS, Chicago, IL) using individual pony as the subject and a compound symmetry repeated covariance format with season, diet adaptation (hay or pasture), and WSC (control, glucose, fructose, or inulin) as repeated measures.For baseline glucose and insulin concentrations, a repeated factor for the different feed trials (1 to 4) was used in the mixed model to account for the fact that no WSC had yet been fed and that 4 measurements were obtained for each pony for each season and diet adaptation.Group (NP or PLP), season, dietary adaptation, WSC (or the repeated factor), and the interaction between these factors were included as fi xed effects.Bonferroni post-hoc tests were used where differences were found.Signifi cance was set at P ≤ 0.05.Results are presented as median values (interquartile range) of the actual values before log transformation.

Baseline Glucose Concentration
Overall, baseline glucose was affected by season (P = 0.02) but not by group, diet adaptation, or the repeated factor for trial number (Table 2).There were interactions between season and diet adaptation (P < 0.001), season, and the repeated factor for trial number (P < 0.001), and group, season, and diet adaptation (P = 0.046).In NP, there was no difference in baseline glucose between different diet adaptations within season.In PLP in spring, baseline glucose after adapting to hay was greater than after pasture, whereas in fall, baseline glucose after adapting  1 Results are presented as median (interquartile range) of 4 baseline measurements from each individual for each seasonal and dietary adaptation (for NP, n = 28; for PLP, n = 20).
2 Baseline insulin was log transformed before statistical analysis using mixed models; results are reported as untransformed values. 3Effect of season (P = 0.02 for glucose and P < 0.001 for insulin).Differences between individual measurements not shown in table as pooled values are reported.
to pasture was greater than after hay.Baseline glucose concentrations varied between the different feeding trials, although median concentrations were within the laboratory reference range (4 to 6 mmol/L) at all times.

Baseline Insulin Concentration
Overall, log baseline insulin was affected by season (P < 0.001) but not group, diet adaptation, or the repeated factor for trial number (Table 2).There was an interaction between season and the repeated factor (P = 0.028).Baseline insulin concentrations varied between the different feeding trials and were more variable in PLP than NP.Baseline insulin concentrations were greater in fall than spring in 3 separate feed trials in PLP and 2 trials in NP.

Peak Glucose Concentration
Group (P = 0.02), diet adaptation (P = 0.045), and WSC (P < 0.001), but not season, affected log peak glucose; however, there were also interactions between group and season (P = 0.035) and group and WSC (P < 0.001; Table 3).Peak glucose concentrations were observed 2 to 4 h after feeding each WSC.
Peak glucose concentrations were greater in PLP than NP in the spring hay-adapted glucose feeding trial and in both the fall pasture and hay-adapted glucose feeding trials (P ≤ 0.05).Peak glucose concentrations were greater in both groups after feeding glucose compared with control (no supplementary WSC) feeding in both seasons and diet adaptations (P ≤ 0.05).Peak glucose concentrations in both NP and PLP were lower after feeding fructose and inulin compared with glucose; however, there were differences between PLP and NP after feeding fructose in fall after both diet adaptations (P ≤ 0.05).

Peak Insulin Concentration
Log peak insulin was affected by group (P = 0.016), season (P < 0.001), diet adaptation (P = 0.002), and WSC (P < 0.001), and there was an interaction between group and season (P = 0.045; Table 3).Peak insulin concentrations were observed 2 to 4 h after feeding each WSC.The greatest median (interquartile range) peak insulin concentration in NP was 452.5 (242.4)μIU/ mL in the fall hay-adapted glucose feeding trial.In PLP, the greatest median (interquartile range) peak insulin concentration was 989.0 (441.3)μIU/mL in the fall pasture-adapted glucose feeding trial.In the fall pastureadapted glucose feeding trial, all NP had peak insulin concentrations < 350 μIU/mL and all PLP had peak insulin concentrations > 500 μIU/mL.
Peak insulin concentrations were greater in PLP than NP in both the spring and fall pasture-adapted glucose feeding trials and the fall pasture and hay-adapted control feeding trials (P ≤ 0.05).Peak insulin concentrations Table 3. Peak glucose and insulin concentrations in 7 normal (NP) and 5 previously laminitic ponies (PLP) after control feeding and feeding 1 g/kg glucose, fructose, or inulin once daily for 3 d in spring and fall after adaptation to a basal pasture or hay diet  Each water-soluble carbohydrate (WSC) was fed mixed with 500 g of a commercial chaff-based diet (Happy Hoof, Spillers Effem Equine, Ltd., Mars HorseCare UK, Ltd., Milton Keynes, UK).Control feeding was just 500 g of commercial chaff-based diet.Blood samples were taken for the next 12 h on each day of feeding.All results are presented as median (interquartile range).All results were log transformed before statistical analysis using mixed models; results are reported as untransformed values. 2 Greater than Control (no WSC feeding) for same pony group, season, and basal diet (P ≤ 0.05). 3Greater in PLP than NP for same season, basal diet, and WSC feeding (P ≤ 0.05). 4Greater after feeding basal hay diet than pasture diet for same pony group, season, and WSC feeding (P ≤ 0.05). 5Greater in fall than spring for same pony group, basal diet and WSC feeding (P ≤ 0.05).
were greater in both groups after glucose compared with control feeding in both seasons and diet adaptations (P ≤ 0.05).Peak insulin concentrations in PLP were greater in both control feeding trials in fall compared with spring (P ≤ 0.05).There were also differences in peak insulin concentration between groups in the spring pastureadapted fructose, fall pasture-adapted fructose, and fall hay-adapted inulin feeding trials, although concentrations varied widely and the range of concentrations in each group overlapped considerably (P ≤ 0.05).

Area Under Curve for Glucose
Log glucose AUC was affected by diet adaptation (P = 0.036) and WSC (P < 0.001) but not group or season (Table 4).There were interactions between group and season (P = 0.029) and group and WSC (P = 0.007).Glucose AUC was greater in both NP and PLP after glucose compared with control feeding in both seasons and after both diet adaptations except for the NP fall pasture-adapted trial (P ≤ 0.05).The glucose AUC was also greater after fructose compared with control feeding in both trials in PLP in fall (P ≤ 0.05).

Area Under Curve for Insulin
Log insulin AUC was affected by group (P = 0.017), season (P < 0.001), diet adaptation (P = 0.022), and WSC (P < 0.001), and there were interactions between group and season (P = 0.029) and group and WSC (P = 0.037; Table 4).Insulin AUC was greater in PLP compared with NP in the spring and fall pasture-adapted glucose feeding trials and the fall pasture-and hayadapted control feeding trials (P ≤ 0.05).Insulin AUC was also greater in PLP than NP in the fructose feeding trials in both seasons and after both diet adaptations (P ≤ 0.05), although there was wide variation in values, particularly in PLP.Insulin AUC was greater in both NP and PLP after glucose compared with control feeding in both seasons and after both diet adaptations (P ≤ 0.05).Insulin AUC values were greater in fall than spring in PLP in the pasture-adapted glucose and pasture-and hay-adapted control feeding trials (P ≤ 0.05).
None of the ponies showed clinical signs of laminitis at any point during the different feeding trials.Mild diarrhea (softened feces) was observed in 1 PLP pony during the fructose feeding trials but no clinical signs of abdominal pain were noted.Table 4. Area under the glucose and insulin concentration-time curves in 7 normal (NP) and 5 previously laminitic ponies (PLP) after control feeding and feeding 1 g/kg glucose, fructose, or inulin once daily for 3 d in spring and fall after adaptation to a basal pasture or hay diet 1

Control
Glucose Fructose Inulin

DISCUSSION
One of the key goals of laminitis research is to identify susceptible ponies before the onset of clinical signs to instigate preventive countermeasures (Harris et al., 2006).The association between insulin resistance and laminitis susceptibility may provide a means of taking such a step by identifying insulin-resistant individuals, which are thought to be at increased risk of laminitis.Insulin resistance is defi ned as a sub-normal response of peripheral tissues to a normal serum insulin concentration (Treiber et al., 2005).In ponies, tissue insulin resistance is usually compensated for by an increased pancreatic secretion of insulin, resulting in hyperinsulinemia (Jeffcott et al., 1986).However, not all laminitis-prone ponies are insulin-resistant and not all insulin-resistant individuals are prone to laminitis (Jeffcott et al., 1986).
Measurement of serum insulin concentrations 2 h after feeding glucose when the ponies were adapted to eating pasture seemed to be the most consistent way to distinguish between NP and PLP in the current study.Differences between NP and PLP were more marked when adapted to fall rather than spring pasture.All NP had peak insulin concentrations of <350 μIU/mL in the fall pasture-adapted glucose feeding trial, whereas all PLP had peak insulin concentrations of >500 μIU/ mL.Measurement of peak insulin concentration has the advantage of requiring fewer blood samples, as opposed to calculation of insulin AUC, which can also distinguish between the groups of ponies but requires multiple blood samples over a prolonged period.However, the present result was based on the responses of a small number of ponies, which may not be fully representative of the equine population as a whole.The usefulness of peak insulin concentration measured 2 h after glucose feeding requires further testing in a larger group of individuals of unknown laminitis susceptibility to assess its predictive value, specifi city, and sensitivity.
In the United Kingdom, laminitis has a seasonal incidence with cases predominantly occurring in spring.In ponies at a rescue center in southern England, the greatest prevalence (2.6%) and incidence (16 cases/1,000 animals) of laminitis were in May, associated with a greater number of hours of sunshine (Menzies-Gow et al., 2010).Increased basal insulin concentrations have also been reported in grazing horses in the United States in April (55 ± 10 μIU/mL; McIntosh et al., 2007;Frank et al., 2010a), which may correspond to pasture WSC content.Conversely, in November in the United Kingdom, the prevalence and incidence of laminitis were 0.6% and 2 cases/1,000 animals, respectively (Menzies-Gow et al., 2010).Thus, this study was conducted to coincide with the times of year when laminitis prevalence is predicted to be at its greatest and lowest.
Interestingly, in the current study, changes in serum insulin concentration in response to feeding supplementary WSC were greater in fall compared with spring, particularly in PLP.This was an unexpected result and contradicts previous work, in which PLP had increased serum insulin concentrations compared with NP during the summer when grazing pasture, but not during the winter when fed hay (Bailey et al., 2008).There was an increased variability in insulin concentrations in the fall in the current study, which, when combined with the small numbers of animals studied, may have contributed to the difference in results between the current study and the previous work (Bailey et al., 2008).The pasture WSC may have been unusually high during the fall in the current study.Grass stems are greater in WSC than leaves (Watts, 2005), thus, relative overgrazing of short grass in fall could have resulted in ingestion of greater concentrations of WSC in the ponies.Unfortunately, one of the limitations of the current study is that WSC analysis of pasture was not performed and the WSC concentrations of pasture in spring and fall are unknown.
Laminitis has been induced in normal ponies by infusing insulin for 72 h using a prolonged euglycemichyperinsulinemic clamp (pEHC) technique (Asplin et al., 2007).Serum insulin concentrations achieved in that study were 1,036 ± 55 μIU/mL and the authors hypothesized that insulin was directly responsible for causing laminitis (Asplin et al., 2007).Serum insulin concentrations above 1,000 μIU/mL were recorded in PLP in the current study on many occasions without occurrence of any clinical signs of laminitis.However, in the current study, hyperinsulinemia was pulsatile in response to feeding, rather than sustained for 72 h, and this may be the reason why ponies in the current study did not succumb to clinical disease.Alternatively, in the pEHC technique, infusion of large amounts of glucose is required to maintain euglycemia in the face of profound hyperinsulinemia and it is possible that glucose toxicity or high cortisol may have been responsible for inducing laminitis in that study (Asplin et al., 2007).
There was a great deal of variation in both baseline glucose and insulin concentrations both between individuals and in different seasons and dietary adaptations.This confi rms previous reports of the variability of these measurements and limitations of using baseline concentrations to assess insulin resistance in ponies (Firshman and Valberg, 2007;Pratt et al., 2009).Baseline insulin concentrations were relatively high in both groups of ponies, particularly in fall.Median concentrations in both groups were above 20 μIU/mL, which has been suggested as a cut-off value to defi ne hyperinsulinemia in horses and ponies (Frank et al., 2010b).This may be because the ponies in the current study were not fasted before baseline measurements were obtained.However, relatively high baseline measurements (27 ± 8 μIU/mL) have also been recorded in a group of Thoroughbred (TB) mares adapted to pasture in spring, which was hypothesized to be due to increased consumption of WSC from pasture (Staniar et al., 2007).The baseline insulin concentrations measured in the ponies in the current study indicates that ponies, regardless of their laminitis susceptibility, are relatively insulin resistant in comparison with horses, agreeing with previous studies (Jeffcott et al., 1986;Pratt et al., 2009).
In human epidemiological studies, it has been reported that overconsumption of fructose, rather than glucose, is responsible for the increasing incidence of human metabolic syndrome (HMS), insulin resistance, and associated cardiovascular complications in human subjects (Johnson et al., 2007).In previous equine studies in Standardbred and Arabian horses, fructose supplementation resulted in a lower glycemic and insulinemic response compared with glucose supplementation (Bullimore et al., 2000;Vervuert et al., 2004), which agrees with the results of the current study.The association between fructose consumption and HMS is thought to be due to the production of uric acid.As fructose is metabolized to fructose-1-phosphate, ADP is released, which is further metabolized to uric acid (Nakagawa et al., 2006;Johnson et al., 2007).Uric acid can cause endothelial dysfunction by inhibiting nitric oxide synthase, a key component of HMS and a proposed pathophysiological mechanism for laminitis (Johnson, 2002;Bailey et al., 2004a;Khosla et al., 2005;Frank et al., 2010b).However, most mammals possess uricase, an enzyme responsible for metabolism of uric acid to allantoin, but it is non-functional in humans and nonhuman primates (Varela-Echavarria et al., 1988).Thus, the role of fructose and uric acid may be less important in inducing endothelial dysfunction and metabolic syndrome and increasing the risk of vascular complications, such as laminitis in equines than humans.However, feeding fructose [3 g/(kg .d) divided into 3 feeds for 2 wk], but not glucose or inulin, to obese, insulin-resistant TB horses caused increases in resting insulin concentrations and decreases in insulin sensitivity measured by minimal model analysis (Geor et al., 2010).Interestingly, no changes in insulin sensitivity occurred in obese, noninsulin-resistant TB horses after fructose feeding in the same study (Geor et al., 2010).Peak insulin concentrations were not reported in that study.Differences between the study of Geor et al. (2010) and the current study may be related to the use of obese horses rather than non-obese ponies, a longer period of WSC feeding, and feeding a larger dose of WSC.Differences may also have occurred because of the different outcome measures investigated in the 2 studies, a short term insulinemic response to feeding WSC compared with a longer-term change in insulin sensitivity as a result of WSC feeding.
In the current study, minimal differences in insulin concentration between PLP and NP were noted after feeding inulin.In contrast, there was a previous study, in which inulin [3 g/(kg .d), divided into 3 feeds for a 48-h period], caused a 5.5-fold increase in insulin concentrations in PLP and a 2-fold increase in NP (Bailey et al., 2007).A lower dose of inulin [1 g/(kg .d)] was fed in the current study and it is possible that differences between NP and PLP would have been enhanced if larger doses had been fed.Inulin was selected as a fructo-oligosaccharide to resemble the fructan WSC commonly found in pasture grass, and thus to mimic conditions encountered by grazing ponies, which may gorge on lush pasture.Fructans are storage compounds in grass, which accumulate during conditions that favor photosynthesis over growth, such as days with bright sunshine and cool nights (Longland and Byrd, 2006).Overconsumption of fructan WSC in particular has been associated with an increased risk of laminitis (Bailey et al., 2004a).Feeding inulin at 3 g/(kg .d) to NP and PLP caused decreases in fecal pH and increases in fecal concentrations of vasoactive amines such as tryptamine and tyramine (Crawford et al., 2007).In the circulation, these amines cause digital vasoconstriction and they could be a trigger factor for laminitis, providing a possible link between overconsumption of grass and the vascular events associated with laminitis (Elliott et al., 2003;Bailey et al., 2004b).However, there are structural differences between inulin, which is a linear molecule linked by β-(2-1) bonds, and the fructans found in most pasture grasses, which tend to be branched molecules with both β-(2-1) and β-(2-6) bonds (Roberfroid, 2005;Longland, 2007).Thus, inulin may not be the best model for the overconsumption of fructan WSC commonly found in pasture.
The supplementary WSC dose of 1 g/(kg .d) was selected due to concerns over feeding larger amounts of WSC in a single feed to laminitis-prone ponies and the risk of inducing laminitis in these individuals.Additionally, large doses of fructose cause gastrointestinal distress and diarrhea in humans because of a limited capacity for fructose absorption in the small intestine (Ravich et al., 1983).Previous studies of feeding fructose to horses have used 0.7 g/kg (Bullimore et al., 2000;Vervuert et al., 2004), and there was concern that a dose of fructose > 1 g/kg could cause gastrointestinal problems in the ponies.However, it must be acknowledged that larger doses of WSC would have been likely to cause more marked changes in glucose and insulin concentration and differences between NP and PLP may have been more obvious.
In the current study, the mean age of NP and PLP was similar but both groups of ponies were aged.Peak insulin concentration and AUC for insulin in response to feeding hay and a standard starch and sugar feed have recently been shown to be greater in normal horses aged over 19 yr in comparison with adult horses aged 5 to 12 yr (Rapson et al., 2011).It is possible that different responses to those in the current study may have been obtained in young or immature ponies.
In conclusion, the most noticeable changes in glucose and insulin concentration were observed after feeding glucose, particularly when ponies were adapted to eating fall pasture.Measurement of insulin concentration 2 h after feeding glucose (1 g/kg) distinguished PLP and NP ponies and could be a useful basis for identifi cation of individuals at increased risk of laminitis.However, further work is necessary in a larger group of ponies of unknown laminitis susceptibility to determine the predictive power of this measurement and to determine specifi c cut-off values that may identify laminitis-susceptible individuals.

Table 2 .
Baseline glucose and insulin concentrations in 7 normal (NP) and 5 previously laminitic ponies (PLP) in spring and fall after adaptation to a basal pasture or hay diet1