The post feeding glycaemic and insulin response to copra meal in horses Academic research paper on "Animal and dairy science"

Accepted Manuscript

Title: The post feeding glycaemic and insulin response to copra meal in horses

Author: N. Richards T.J. Kempton

PII: S0377-8401(15)30013-4

DOI: http://dx.doi.org/doi:10.1016/j.anifeedsci.2015.09.003

Reference: ANIFEE 13366

To appear in: Animal Feed Science and Technology

Received date: 12-2-2015

Revised date: 1-9-2015

Accepted date: 5-9-2015

Please cite this article as: Richards, N., Kempton, T.J.,The post feeding glycaemic and insulin response to copra meal in horses, Animal Feed Science and Technology (2015), http://dx.doi.org/10.1016/j.anifeedsci.2015.09.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The post feeding glycaemic and insulin response to copra meal in horses

N Richards3 and TJ Kemptonb

a Equilize Horse Nutrition Pty Ltd PO Box 11034 Tamworth NSW 2340 Australia b Stance Global Pty Ltd, PO Box 764 Kenmore QLD 4069 Australia tim@stanceglobal.com (corresponding author)

Abstract

Knowing the effect a feed ingredient has on post-feeding glycaemic and insulin responses is important when managing conditions like equine metabolic syndrome, laminitis and the polysaccharide storage myopathy form of tying up in horses. Feeds that cause minimal post-feeding disturbances to plasma glucose and insulin are desirable. This study was conducted to determine the post feeding glycaemic and insulin response in horses to copra meal (Copra Meal; 11% nonstructural carbohydrate; NSC) and to compare this to the responses observed to low NSC pasture (Pasture; 7 % NSC) and higher NSC extruded pellet (Pellet; 25.3 % NSC) and sweetfeed (Sweetfeed; 33.7 % NSC) rations.

Copra Meal did not increase plasma glucose levels above those observed in horses grazing the low NSC Pasture while the Pellet and Sweetfeed rations caused significant post-feeding rises in plasma glucose. The Pellet and Sweetfeed rations also raised post feeding plasma insulin levels significantly compared to Pasture and Copra Meal. Copra Meal raised plasma insulin levels significantly higher than that observed in horses grazing Pasture from 15 to 60 minutes post feeding, after which time there was no significant difference. The results of this study demonstrated that the NSC content of a

feed ingredient will influence the post-feeding glycaemic and insulin response to that feed, with low NSC feeds giving significantly lower responses than higher NSC feeds. The insulin response to Copra Meal observed in this study was small both in magnitude and time in comparison to levels reported in the literature needed to induce laminitis and as such are unlikely to be physiologically significant. Copra Meal is a feed ingredient that could be considered for use in diets for horses needing to remain on a low NSC diet that will cause minimal disturbance to their post feeding glucose and insulin levels.

Keywords

Horse, plasma glucose, plasma insulin, non-structural carbohydrate, copra meal, insulin resistance

Highlights

• Four feeds with varying non-structural carbohydrate (NSC) contents were studied.

• Two were low in NSC (Pasture; 7 % NSC and Copra Meal; 11 % NSC).

• Two were higher in NSC (Pellets; 25.3 % NSC and Sweetfeed; 33.7 % NSC).

• Low NSC feeds gave rise to lower post feeding plasma glucose and insulin responses.

• Copra Meal appears suitable for horses who need a low NSC diet.

Introduction

Copra or coconut meal is a co-product of the coconut oil industry and consists of the coconut 'meat' or flesh after a majority of the oil has been extracted. While the quality and nutritional characteristics of copra meal can be variable depending on whether it is sun or

'kiln' dried, how it is stored, its moisture content (Thanaraj etal. 2007) and whether it is mechanically or solvent oil extracted, copra meal is considered a valuable livestock feed (Guarte et al. 1996).

When considering its use as a feed for equines, of particular interest is its low non-structural carbohydrate content (NSC; defined as the sum of starch and water soluble carbohydrates (Frank et al. 2010). Horses and ponies with health conditions including laminitis, insulin resistance, equine metabolic syndrome, Cushing's disease and the polysaccharide storage myopathy form of tying up (PSSM) are required to be maintained on a diet low in NSC that will not cause large fluctuations in post feeding blood glucose and insulin levels (Borgia 2009; Geor & Harris 2009; Frank et al. 2010; Geor 2010; Harris 2012). Current recommendations suggest horses prone to or with these conditions should be maintained on diets containing less than 12% NSC (Geor & Harris 2013).

Mechanically expelled copra meal (8 - 10 % Oil, 15 MJ/kg digestible energy) is reported to contain 10.3% NSC on an as fed basis (Richards 2008). Based on this NSC content and the post feeding glycaemic and insulin responses reported by Gordon et al. (2007) to a feed containing 9.43% NSC, copra meal could be considered an appropriate feed ingredient for horses that are required to be maintained on a low NSC, yet high energy diet. However, there is currently scant information available in the literature to describe the likely postfeeding glycaemic and insulin response to feed ingredients with this level of NSC.

This study aims to describe the changes in post-feeding plasma glucose and insulin when horses grazing low NSC pasture are fed either copra meal or moderate to high NSC commercially prepared feeds.

Materials and methods

The study was conducted at Newcastle, NSW Australia during the autumn period from late April until the end of May (34 days). The experimental protocol was approved by the Animal Ethics Committee at the University of New England, Armidale NSW Australia.

Animals

Four horses, mean age 16.3 ± 2.6 years, weighing 455 ± 26 kg and mean body condition score (Henneke et al. 1983) 4.8 ± 1.2 were used in the study (Table 1). Horses were weighed at the commencement and the completion of the experimental period using digital Ruddweigh livestock scales. Condition score of each horse was assessed using the 9 point Henneke Body Condition Scoring System (Henneke et al. 1983) on the first and last days of the experimental period. Temperament of the horses was assessed during the 14 day adaptation period to ensure suitability.

Four diets were used during the study. These were Pasture only (Pasture), consisting of 90% kikuyu (Pennisetum clandestinum) and 10% white clover (Trifolium repens) with 7% DM NSC. This was used as the control for the study. Treatment diets were Pasture supplemented with either Stance Equine CoolStance copra meal (Copra Meal) with 11% DM NSC; Mitavite Economix extruded and pelleted feed (Pellet) with 25.3% DM NSC; or Barastoc Cool Command sweetfeed (Sweetfeed) with 33.7% DM NSC. Details of NSC analysis are following.

Feedstuffs analysis and in vitro digestibility

Because of the variable nature of commercially prepared feeds that are made using least-cost ration formulation methods the amount of NSC can vary between bags of the same branded product. To avoid this potential bag to bag variation in NSC content impacting the glycaemic and insulin responses observed, single bags of the Pellets and Sweetfeed were kept aside and used only to

provide the horse's feed rations on the morning when blood collection was carried out. Likewise a single bag of copra meal was used to provide all Copra Meal rations on blood sampling days. Samples were collected from each of these bags and sent to a commercial feed testing laboratory (Dairy One, Ithaca NY USA) for analysis of crude protein (AOAC 990.03), crude fat (AOAC 2003.05), water soluble carbohydrates, ether soluble carbohydrates and starch (Table 2). Water soluble carbohydrates and ether soluble carbohydrates were determined using the method of Hall et al. (1999). Starch was determined using the YSI 2700 SELECT Biochemistry Analyser. Briefly, samples are pre-extracted for sugar by incubation in a 40 °C water bath and filtration on Whatman 41 filter paper. Residues are thermally solubilized using an autoclave and then incubated with glucoamylase enzyme to hydrolyse starch and produce dextrose. Prepared samples are then injected into the sample chamber of the YSI Analyser where dextrose diffuses into a membrane containing glucose oxidase. The dextrose is immediately oxidised to hydrogen peroxide and D-glucono-4-lactone. The hydrogen peroxide is detected amperometrically at the platinum electrode surface. The current flow at the electrode is directly proportional to the hydrogen peroxide concentration, and hence to the dextrose concentration. Starch is determined by multiplying dextrose by 0.9. In vitro digestible NSC was determined (Richards 2003) to estimate the amount of NSC that was expected to be digested in the small intestine and absorbed as glucose (Table 2). In brief, samples were ground to 0.5 mm before 100 mg of sample was incubated at 38°C for 15 minutes in the presence of excess a-amylase and amyloglucosidase. The amount of glucose produced was then measured.

Morning and afternoon pasture samples were collected from the immediate area the horse on this diet was grazing between 0900 and 1000 hours and between 1300 and 1500 hours on each of the blood collection days. Multiple areas of the pasture were sampled, pooled, mixed and sub-sampled to give as representative sample of what the horses were consuming as possible. Pasture samples were immediately dried using a microwave oven and packaged in a labelled, airtight ziplock bag before being sent to Dairy One, Ithaca NY USA for analysis of crude protein, crude fat, starch, water

soluble carbohydrates and ether soluble carbohydrates using the analytical methods previously described (Table 2).

Adaptation period

All horses were wormed and had their teeth floated by an equine dentist during the 14 day adaptation period of the study. During this initial period, horses were fed an equal mixture of the Meal, Pellets and Sweetfeed (dampened down with water) at a combined rate of 0.5 % bodyweight (approximately 2.3 kg), divided into 2 meals fed at 0800 and 1800 per day. During this period, horses were allowed access to pasture from 0800 till 1800 hours. They were then yarded overnight without access to hay or pasture. Each horse was exercised at the trot for 30 minutes each day between 0600 and 0800 (except on blood collection days).

Experimental period

During the experimental period, horses were fed the treatment diets from individual nose bags to prevent horses sharing or dropping feed. The horses on these diets were allowed access to pasture from 0800 hours till 1800 hours before being yarded overnight. This feeding regime was used to maintain consistency between this study and others designed to measure post-feeding glycaemic and insulin responses in horses (Pagan 1999; Richards 2003; Borgia 2009; Vervuert et al. 2009).

Treatment diets were fed in a randomised 4 x 4 latin square, with blood collection occurring on the morning of the fifth day of each latin square period. Horses on the Pasture diet were given access to pasture 24 hours of the day to simulate the natural feeding patterns of a horse. When on the Pasture diet, horses were supplemented with 0.2 g/kg bodyweight of a commercial vitamin and

mineral supplement once per day to cover their vitamin and mineral requirements. All horses had ad-libitum access to fresh water and a salt lick.

The Copra Meal, Pellets and Sweetfeed were fed at a rate of 0.5% bodyweight per day divided into two meals fed at 0800 hours and 1800 hours. The evening feed following blood collection consisted of a 50% mix of the horse's old diet and new diet in the latin square rotation. The following mornings feed was a 25% to 75% mix of the old and new diets while the evening feed on day one of each rotation was 100% of the horse's new diet. Horses changing onto the pasture diet were given 24 hour access to pasture from the afternoon of the blood testing days.

Feed refusals were recorded as any feed left after the horse had one hour's access to the feed. If no refusal was recorded, feed intake was taken to be the amount fed. Pasture intake was not measured.

Blood collection

On days 5, 15, 20 and 25 of the experimental period venous blood samples were collected via an indwelling catheter (inserted at 6 am) for measurement of plasma glucose and insulin.

Once the catheters were inserted, a baseline blood sample was taken and catheters were flushed with a heparin (50 i.u./ml) saline (0.09%) solution. On completion of catheter insertion the horses were fed the diet they had been adapted to in the previous 4 days. Horses on the Meal, Pellet and Sweetfeed diets were allowed to eat their morning meal before being tethered for the remainder of the collection period. The horse on the Pasture diet was allowed to graze for the duration of the collection period.

Blood samples were collected over a 6 hour period with time zero corresponding to when horses received their treatment diet meal. For horses on the pasture diet time zero was started at a time chosen close to when the treatment diets were fed. Samples were taken every 15 minutes for the first hour, every 30 minutes for the following 3 hours and then every hour for the remaining 2 hours

for a total of 13 samples per horse, including the baseline sample. The catheters were flushed with heparin saline after each blood sample was taken. All blood samples consisted of 10 mL of venous blood and were collected into lithium heparinised blood collection tubes that were immediately centrifuged at 2000 g for 15 min. The plasma was then transferred to labelled plastic vials. Plasma samples were stored at -16°C until analysed for glucose and insulin concentrations.

Blood analysis

Prior to glucose and insulin analysis, plasma samples were thawed and re-centrifuged at 2000 g for 10 min to remove plasma proteins that had precipitated from solution on thawing. The plasma glucose concentration was determined using the Dimension® clinical chemistry system on a DADE XL clinical auto-analyser (Dade Behring Inc, Newark, DE 19714, USA). The GLU Flex™ reagent cartridge (Cat No DF39A) was used as the in vitro reagent.

Plasma insulin concentrations were determined using an in-house radioimmunoassay (RIA)(Albano et al. 1972) which has been previously validated for equine insulin(Freestone et al. 1991; Richards 2003; Borer-Weir et al. 2012). This was carried out at the University of New England, Armidale NSW.

The polyclonal antiserum (Sigma-Aldrich. St Louis, Missouri, USA) was raised in guinea pigs against purified bovine insulin. Bovine insulin was used as both the standard and tracer. The tracer was iodinated using the iodogen method (Pierce. Rockford, Illinois, USA). Separation of bound insulin from free insulin was achieved using polyethylene glycol 6000.

Statistical analysis

For horse measured outcomes, including plasma insulin and glucose parameters, a restricted maximum likelihood (REML) procedure was used to assess the effects of various factors, specifically diet, period, and time, as well as interactions between these terms. Significance of terms was assessed using Wald chi-square and F tests. Non-significant (P > 0.05) interactions were removed

from the model. In addition, a random term was included for horse, to take account of there being multiple observations on each horse and significance of horse effects were assessed using REML likelihood ratio tests by fitting a reduced model without horse in the model. For a number of the outcomes measured, data had to be log-transformed to ensure a normal distribution of residuals, as well as constant variance.

Based on the fitted model, tables of model-based means were calculated, and where necessary, back-transformed to the original scale to facilitate a meaningful interpretation. When significant differences were obtained from the overall Wald tests, pairwise comparisons of means were undertaken using approximate least significant differences (LSDs) by examining the means on the transformed scale.

Because of the limited number of observations, a simple analysis of variance (ANOVA) was used to assess pasture differences, and period differences. For testing of before versus after bodyweights, a paired t-test was used.

Results

Bodyweights

There was no significant difference in average bodyweight over the 25 days from the commencement of the experimental period (455 ± 26.0 kg) to the end of the experimental period (456 ± 24.7 kg; P = 0.83).

Feed and pasture non-structural carbohydrates

The pasture had the lowest total NSC content (7.0 % DM), while the Copra Meal (11.0% DM), Pellets (25.3% DM) and Sweetfeed (33.7% DM) had increasing concentrations of NSC (Table 2).

Digestible NSC varied between diets, with the Pellets having the highest digestibility and Sweetfeed the lowest. Using these results to determine the amount of NSC digested in the small intestine, the

Pellets had the highest level of digestible NSC (215 g/kg) and provided the highest digestible NSC daily intake for the 3 diets (Table 3).

Pasture starch, water soluble carbohydrates (WSC), ether soluble carbohydrates (ESC) and total NSC levels were, on average, higher in the pasture samples collected in the afternoons of blood testing days, compared to samples collected in the mornings, with total NSC being significantly higher in the afternoons (7.5% DM) compared to the mornings (6.5% DM; Table 4).

Pasture starch, water soluble carbohydrates (WSC), ether soluble carbohydrates (ESC) and total NSC levels did not differ significantly between Latin square periods.

Feed intakes and refusals

Horse 4 recorded a feed refusal of 440 g and 526 g of the Copra Meal diet on days 4 and 5 of the latin square period during which this horse was on this diet.

Glycaemic response

The Pasture and Copra Meal diets gave the lowest post-prandial plasma glucose responses which were not significantly different from each other (P>0.05). Pellets gave the highest post-prandial glycaemic response, with the plasma glucose levels for horses on this diet being significantly higher than plasma glucose levels measured on the Pasture or Copra Meal diets from 30 minutes post feeding until 240 minutes post feeding (P<0.05; Figure 1). Likewise the horses on the Pellet diet had significantly higher plasma glucose concentrations than horses on the Sweetfeed diet from 45 minutes until 240 minutes post feeding. Plasma glucose levels were significantly higher in horses fed the Sweetfeed compared to horses fed the Copra Meal or Pasture diets from 30 minutes until 180 minutes post feeding (Figure 1).

The Pellets gave rise to the highest post-feeding peak plasma glucose level, which was significantly higher (P<0.05) than the peak plasma glucose concentration observed when horses were consuming

the other diets. There was no significant difference between the Copra Meal and Pasture diets for peak plasma glucose (Table 5). There was no significant difference between diets for time to peak plasma glucose (Table 5).

Diet had a significant effect (P=0.007) on the area under the plasma glucose response curves (Table 5). The Pellet diet gave a significantly higher area under the glucose response curve than the Sweetfeed, Copra Meal or Pasture diets, while the Sweetfeed gave a significantly higher area under the curve when compared to the Copra Meal and Pasture diets. There was no significant difference between the Copra Meal and Pasture diets for area under the glycaemic response curve (Table 5).

Insulin response

The plasma insulin levels reported in this study are comparable to the reference range serum insulin levels reported when measured using a similar RIA by Freestone et al. (1991), allowing for expected difference between plasma and serum insulin levels (Manley et al. 2007).

The Pasture diet gave the lowest post-feeding plasma insulin response which was significantly lower than the Copra Meal diet from time 15 to 60 minutes. After 60 minutes there was no significant difference between these two diets. The post-prandial plasma insulin levels observed in horses on the Pasture diet were significantly lower than the levels seen in horses on the Pellet diet from time 30 to 240 minutes and significantly lower than the Sweetfeed diet from time 60 to 180 minutes (Figure 2).

The Pellets gave the highest post feeding plasma insulin response which was significantly higher than the Copra Meal diet from time 45 to 240 minutes. The Sweetfeed gave the second highest post feeding insulin response, however it was only significantly higher than the Copra Meal diet at time 180 minutes (Figure 2). The Copra Meal diet displayed a rapid post feeding insulin response and was significantly higher than all diets at time 15 minutes. It then only remained significantly higher than the Pasture diet until time 60 minutes (Figure 2).

The Pellets gave rise to a significantly higher (P<0.05) post-feeding peak insulin level compared to other diets, while the Copra Meal and Sweetfeed diets had significantly higher peak insulin levels than the Pasture diet. There was no significant difference in peak insulin between the Copra Meal and Sweetfeed diets (Table 5). There was no significant difference between diets for time to peak plasma insulin (Table 5).

The Pellet diet gave rise to a significantly higher (P=0.001) area under the insulin response curve than all other diets (Table 5). Pasture gave the lowest area under the insulin response curve, followed by the Copra Meal and Sweetfeed diets, however these diets were not significantly different from one another (Table 5).

Discussion

Results of this study showed that lower NSC feed ingredients gave rise to lower post feeding glycaemic responses. Post-feeding plasma glucose was influenced by feed type, with the low NSC Pasture and Copra Meal inducing significantly lower glycaemic responses than seen when horses consumed higher NSC Pellet and Sweetfeed rations. Of note, Copra Meal did not increase plasma glucose levels above those seen when horses were grazing low (7 %) NSC pasture, suggesting that Copra Meal could be incorporated into rations where low post-feeding glucose responses are a priority (for example horses with the PSSM form of tying up).

Post feeding insulin responses were not as clearly related to a feed ingredients NSC content. This apparent disconnect between post-feeding glucose and insulin responses has been previously reported (Gordon et al. 2007; Borgia 2009). There was no significant difference in peak insulin observed between the Copra Meal (11 % NSC) and Sweetfeed (33.7 % NSC) diets, despite the fact the Copra Meal caused a significantly lower post-feeding plasma glucose response when compared to the Sweetfeed. While peak insulin did not differ between the Copra Meal and Sweetfeed diets, the Sweetfeed did cause post feeding insulin levels to remain elevated for longer, with insulin

following the Sweetfeed meal significantly higher at 180 minutes post feeding compared to the Copra Meal.

The Copra Meal diet raised plasma insulin levels significantly higher than the low NSC pasture for a short period of time (1 hour) post-feeding, after which time there was no significant difference between these diets. A prolonged elevation of insulin to supraphysiological levels has been shown to cause laminitis in otherwise normal ponies and standardbred horses under experimental conditions (Asplin et al. 2007; De Laat et al. 2010; de Laat et al. 2012). Practically however, it is currently unknown how long plasma insulin needs to remain elevated, what level it needs to be elevated to and whether it needs to remain elevated continuously or if serial elevations over several days or weeks will have the same effect in causing laminitis. De Laat et al (2012) do however indicate that the combination of the magnitude of hyperinsulinaemia and the length of time it is maintained may be what determines the severity of lamellar damage in the hoof.

Differences between study design, analysis methods and use of both plasma and serum to measure insulin makes absolute comparison of insulin concentrations between research papers difficult. However, with Bailey et al. (2008) reporting no clinical signs of laminitis in ponies with a mean insulin level of 69.5 ± 19.8 uIU/ml while grazing summer pastures in the UK and de Laat et al. (2012) needing to maintain serum insulin levels at an average of 200 uIU/ml for 48 hours to cause histopathological evidence of laminitis in at least one hoof of insulin sensitive standardbred horses, it would be reasonable to conclude that neither the magnitude of the rise in insulin observed in this study following consumption of the copra meal (average peak of 9.6 uIU/ml), nor the length of time it was maintained is going to be physiologically significant from a laminitis or insulin resistance perspective. It should be noted that the horses used in this study, perhaps with the exception of Horse 4, all showed evidence of good insulin sensitivity. The glucose and insulin responses to the feed ingredients described here may be different in horses with compromised insulin sensitivity and this is something that requires further investigation.

Copra Meals higher than expected insulin response may be due to a number of factors. Copra meal has a relatively unique amino acid profile, with high levels of leucine and phenylalanine (Son et al. 2014). Leucine and phenylalanine are known to cause a rise in plasma insulin levels following consumption in humans (Floyd et al. 1966; van Loon et al. 2000; Yang et al. 2010) and this may be what caused the rise in plasma insulin following consumption of the Copra Meal ration. A further confounding factor that must be considered is the effect of the overnight fast imposed on horses being fed the Copra Meal, Sweetfeed and Pellet rations but not the Pasture. While the protocol used allows this study to be compared to others using similar protocols, the effect fasting for 12 hours prior to being fed a single meal before being fasted again for 6 hours during sample collection has on post-feeding glycaemic and insulin responses is unknown at this point. Further research looking at the post-feeding glycaemic and insulin response to Copra Meal whilst having continuous access to a low NSC forage would be of practical value.

Finally, this study demonstrated that digestibility of a feeds NSC content (Table 3) appears to affect the observed post-feeding glycaemic and insulin responses as previously reported by Richards (2003), with the lower NSC but higher 'digestible NSC' Pellet giving rise to higher glycaemic and insulin responses when compared to the higher NSC but lower 'digestible NSC' Sweetfeed. As such the digestibility of a source of NSC should be considered alongside the total NSC content of a feed ingredient when selecting feeds for horses requiring a diet that will not cause significant disturbance to their post feeding glycaemic and insulin responses.

Conclusion

Copra meal was shown in this study to cause no significant changes in post feeding plasma glucose levels when compared to a low NSC (7 % DM) pasture and resulted in significantly lower postfeeding glucose levels compared to commercial sweetfeed and extruded pellet rations. Copra Meal resulted in post-feeding insulin levels that were significantly higher than those observed for horses grazing the low NSC pasture for the first 60 minutes post feeding. However, the level of insulin

measured and the short period of time it actually remained elevated mean that the observed increase in plasma insulin is unlikely to be physiologically significant when considering its use for horses that need to remain on low NSC diets that will not cause significant disturbance to their postfeeding insulin responses.

AUTHOR DECLARATION

We wish to draw the attention of the Editor to the following facts which may be considered as potential conflicts of interest and to significant financial contributions to this work. 'Stance Equine' who funded this study also supply copra meal for commercial gain to the horse industry. Dr Timothy J Kempton, a co-author of this paper is the owner and managing director of Stance Equine. While Dr Kempton co-authored the paper, Stance Equine had no influence on the design of the reported study nor were they involved in any aspect of the collection or analysis of plasma or pasture/feed samples. All statistical analyses were conducted by Statistical Advisory & Training Service Pty Ltd and again Stance Equine had no input into the analyses. We are confident that Stance Equine's Involvement has not influenced the outcome of this study.

We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.

We confirm that we have given due consideration to the protection of intellectual property associated with this work and that there are no impediments to publication, including the timing of publication, with respect to intellectual property. In so doing we confirm that we have followed the regulations of our institutions concerning intellectual property.

We further confirm that any aspect of the work covered in this manuscript that has involved experimental animals has been conducted with the ethical approval of all relevant bodies and that such approvals are acknowledged within the manuscript.

We understand that the Corresponding Author is the sole contact for the Editorial process (including Editorial Manager and direct communications with the office). He is responsible for communicating with the other authors about progress, submissions of revisions and final approval of proofs. We confirm that we have provided a current, correct email address which is accessible by the Corresponding Author and which has been configured to accept email from the Animal Feed Science and Technology Journal.

Signed by all authors as follows:

Nerida Richards

6th February 2015

Timothy J Kempton

6th February 2015

Role of the funding source

'Stance Equine' who funded this study also supply copra meal for commercial gain to the horse industry. TJ Kempton, a co-author of this paper is the owner and managing director of Stance Equine. While Dr Kempton co-authored the paper, Stance Equine had no influence on the design of the reported study nor was involved in any aspect of the collection or analysis of any samples (plasma or pasture/feed). The Copra Meal used in the study was purchased from a feed store without their knowledge of where it would be purchased from. In addition, all statistical analyses were conducted by Statistical Advisory & Training Service Pty Ltd and again Stance Equine had no input into the analyses. We are confident that Stance Equine's Involvement has not influenced the outcome of this study.

Acknowledgements

The authors wish to thank Stance Equine for funding this study and the capable assistance of Dr Britta Walner DVM, Associate Professor Jim McFarlane and Tarsha Macklinshaw in handling the horses, collecting and analysing samples. We would also like to thank Statistical Advisory & Training Service Pty Ltd for the statistical analysis of the study.

References

Albano J.D.M., Ekins R.P., Maritz G. & Turner R.C. (1972) A sensitive, precise radioimmunoassay of serum insulin relying on charcoal separation of bound and free hormone moieties. Endocrinologica 70, 487-509.

Official Methods of Analysis of AOAC INTERNATIONAL (2012) 19th Ed., AOAC INTERNATIONAL, Gaithersburg, MD, USA, Official Methods 990.03 and 2003.05

Asplin K.E., Sillence M.N., Pollitt C.C. & McGowan C.M. (2007) Induction of laminitis by prolonged hyperinsulinaemia in clinically normal ponies. Vet J 174, 530-5.

Bailey S.R., Habershon-Butcher J.L., Ransom K.J., Elliott J. & Menzies-Gow N.J. (2008) Hypertension and insulin resistance in a mixed-breed population of ponies predisposed to laminitis. American Journal of Veterinary Research 69, 122-9.

Borer-Weir K.E., Bailey S.R., Menzies-Gow N.J., Harris P.A. & Elliott J. (2012) Evaluation of a

commercially available radioimmunoassay and species-specific ELISAs for measurement of high concentrations of insulin in equine serum. American Journal of Veterinary Research 73, 1596-602.

Borgia L.V., S. Watts, K. and Pagan, J (2009) Glycemic/insulinemic response to feeding hay with different water soluble carbohydrate content in healthy and polysaccharide storage myopathy-affected horses. Journal of Equine Veterinary Science 29, 355 - 7.

De Laat M.A., McGowan C.M., Sillence M.N. & Pollitt C.C. (2010) Equine laminitis: Induced by 48 h hyperinsulinaemia in Standardbred horses. Equine Veterinary Journal 42, 129-35.

de Laat M.A., Sillence M.N., McGowan C.M. & Pollitt C.C. (2012) Continuous intravenous infusion of glucose induces endogenous hyperinsulinaemia and lamellar histopathology in Standardbred horses. Vet J 191, 317-22.

Floyd J.C.J., Fajans S.S., Conn J.W., Knopf R.F. & Rull J. (1966) Stimulation of Insulin Secretion by Amino Acids. Journal of Clinical Investigation 45, 1487-502.

Frank N., Geor R.J., Bailey S.R., Durham A.E. & Johnson P.J. (2010) Equine Metabolic Syndrome. Journal of Veterinary Internal Medicine 24, 467-75.

Freestone J.F., Wolfsheimer K.J., Kamerling S.G., Church G., Hamra J. & Bagwell C. (1991) Exercise induced hormonal and metabolic changes in Thoroughbred horses: effects of conditioning and acepromazine. Equine Veterinary Journal 23, 219-23.

Geor R.J. (2010) Nutrition and Exercise in the Management of Horses and Ponies at High Risk for Laminitis. Journal of Equine Veterinary Science 30, 463-70.

Geor R.J. & Harris P. (2009) Dietary management of obesity and insulin resistance: countering risk for laminitis. Vet Clin North Am Equine Pract 25, 51-65, vi.

Geor R.J. & Harris P. (2013) Laminitis. In: Equine Clinical and Applied nutrition (eds. by Geor RJ, Harris P & Coenen M), p. 696. Elsevier.

Gordon M.E., Jerina M.L., King S.L., Davison K.E., Young J.K. & Raub R.H. (2007) The Effects of

Nonstructural Carbohydrate Content and Feeding Rate on Glucose and Insulin Response to Meal Feeding in Equine. Journal of Equine Veterinary Science 27, 489-93.

Guarte R., Muhlbauer W. & Kellert M. (1996) Drying characteristics of copra and quality of copra and coconut oil. Postharvest Biology and Technology 9.

Hall M.B., Hoover W.H., Jennings J.P. & Webster T.K.M. (1999) A method for partitioning neutral

detergent-soluble carbohydrates. Journal of the Science of Food and Agriculture 79, 2079-86.

Harris P. (2012) Update on nutrition and pasture associated laminitis. In: 51st British Equine Veterinary Association Congress BEVA, p. 176, Birmingham, United Kingdom.

Henneke D.R., Potter G.D., Kreider J.L. & Yeates B.F. (1983) Relationship between condition score,

physical measurements and body fat percentage in mares. Equine Veterinary Journal 15, 371 - 2.

Manley S.E., Stratton I.M., Clark P.M. & Luzio S.D. (2007) Comparison of 11 human insulin assays: implications for clinical investigation and research. Clin Chem 53, 922-32.

Pagan J.D. (1999) Recent Developments in Equine Nutrition. Cornell Nutrition Conference for Feed Manufacturers 61, 160 -7.

Richards N. (2003) Enhancing starch digestion in the equine small intestine. In: School of Rural Science. University of New England, Armidale, NSW Australia.

Richards N. (2008) The non-structural carbohydrate content of some commercially available horse feeds in Australia. In: Australian Equine Science Symposium (ed. by Bryden WL), Gold Coast, Australia.

Son A.R., Hyun Y., Htoo J.K. & Kim B.G. (2014) Amino acid digestibility in copra expellers and palm kernel expellers by growing pigs. Animal Feed Science and Technology 187, 91-7.

Thanaraj T., Dharmasena N.D.A. & Samarajeewa U. (2007) Comparison of quality and yield of copra processed in CRI improved kiln drying and sun drying. Journal of Food Engineering 78, 144651.

van Loon L.J., Saris W.H., Verhagen H. & Wagenmakers A.J. (2000) Plasma insulin responses after

ingestion of different amino acid or protein mixtures with carbohydrate. Ameriacn Journal of Clinical Nutrition 72, 96-105.

Vervuert I., Voigt K., Hollands T., Cuddeford D. & Coenen M. (2009) Effect of feeding increasing

quantities of starch on glycaemic and insulinaemic responses in healthy horses. Vet J 182, 67-72.

Yang J., Chi Y., Burkhardt B.R., Guan Y. & Wolf B.A. (2010) Leucine metabolism in regulation of insulin secretion from pancreatic beta cells. Nutr Rev 68, 270-9.

Table 1: The sex, age, breed, bodyweight, condition score and temperament of the horses used in the study.

Horse Sex Age Breed Bodyweight Condition Temperament

ID (kg) Score (0-9)

1 Gelding 15 Thoroughbred 491 4 Quiet

2 Gelding 16 Stockhorse 454 4.5 Nervous

3 Gelding 20 Quarter Horse 430 4 Quiet

4 Mare 14 Stockhorse 445 6.5 Quiet

476 Table 2: The digestible energy, crude protein, starch, water soluble carbohydrate (WSC), ether

477 soluble carbohydrate (ESC), total non-structural carbohydrate (NSC; Starch + WSC) and mineral

478 characteristics of the 4 diets used in the study. All values are presented on a dry matter basis.

Pasture3 Copra meal Pellets Sweetfeed

Digestible Energy (MJ/kg) 9.7 15.0 12.0 13.0

Crude Protein (g/kg) 173 239 151 162

Crude Fat (g/kg) 55 88 54 88

Starch (%) 0.4* 0.1 18.0 26.3

WSC (%)b 6.7* 10.9 7.3 7.4

ESC (%)c 6.8* 6.9 3.3 4.8

Total NSC (%)d 7.0* 11.0 25.3 33.7

Calcium (g/kg)e 3.5 2.3 12.0 10.0

Phosphorous (g/kg) e 3.1 5.0 7.0 5.0

Magnesium (g/kg) e 1.9 3.6 4.9 4.1

Potassium (g/kg) e 26.4 20.0 14.2 8.2

Sodium (g/kg) e 1.0 0.4 6.8 4.8

Chloride (g/kg) e 14.3 7.7 13.7 9.7

Copper (mg/kg) e 6 33 37 66

Zinc (mg/kg) e 28 47 110 180

Manganese (mg/kg) e 42 46 191 150

Iron (mg/kg) e 190 171 194 240

479 a Average of 8 individual samples collected and analysed

480 b Water Soluble Carbohydrate; c Ether soluble Carbohydrate; d Total Non-Structural Carbohydrate

481 e Mineral levels are shown 'as analysed' by Dairy One (Ithaca NY) for the Pasture or as specified by the

482 manufacturing companies for the Copra Meal, Pellets and Sweetfeed.

484 Table 3: The daily non-structural carbohydrate (NSC) intake, NSC digestibility, total digestible NSC

485 and total average digestible NSC intake for the Copra Meal, Pellet and Sweetfeed diets.

Copra Meal Pellets Sweetfeed

Dry Matter (%) 95.3 90.1 90.0

NSC Intake (g/day) 250 575 767

NSC Digestibility (% DM) 69 85 62

Total Digestible NSC (g/kg DM) 76 215 209

Total digestible NSC intake (g/day) 190 489 475

486 DM - Dry Matter

488 Table 4: The average starch, water soluble carbohydrate (WSC), ether soluble carbohydrate (ESC),

489 and total non-structural carbohydrate (NSC) for pasture samples collected on the morning and

490 afternoons of blood sampling days. All results presented on a dry matter basis.

Pasture Average Morning Collections Pasture Average Afternoon Collections P Value

Starch (%) 0.2 ± 0.15 0.5 ± 0.49 0.444

WSC (%)a 6.3 ± 0.42 7.0 ±0.74 0.103

ESC (%)b 6.3 ± 1.24 7.3 ± 0.68 0.059

Total NSC (%)c 6.5 ± 0.39 7.5 ± 0.34 0.011

491 Water Soluble Carbohydrate; Ether soluble Carbohydrate; Total Non-Structural Carbohydrate

493 Table 5: The peak plasma glucose, time to peak plasma glucose, area under the plasma glucose

494 curve, peak plasma insulin, time to peak plasma insulin and area under the plasma insulin curve

495 values for the Pasture, Copra Meal, Pellet and Sweetfeed diets.

Pasture Copra Meal Pellets Sweetfeed

Peak Glucose (nM/l) 4.2b ± 0.39 4.4b ± 0.39 7.4a ± 0.97 5.6b ± 0.84

Time to Peak Glucose (mins) 143 ± 37.8 116 ± 55.3 98 ± 15 90 ± 0

Area Under Glucose Curve* 23.7a ± 1.61 23.9a ± 1.73 28.0b ± 5.30 32.0c ± 5.21

Peak Insulin (uIU/ml) 1.8c ± 1.87 9.6b ± 14.96 16.0a ± 19.06 10.6b ± 16.1

Time to Peak Insulin (mins) 120 ± 64.8 45 ± 17.3 68 ± 28.7 68 ± 56.8

Area Under Insulin Curve* 8.8 ± 11.02 18.3 ± 28.46 33.3 ± 54.93 49.9 ± 61.72

496 a b c Different superscripts in the same row indicate significant differences (P<0.05).

497 * Area Under Glucose Curve (P<0.007); #Area Under Insulin Curve (P<0.001)

Time (minutes)

499 Figure 1: Average post feeding glycaemic response curves for the Pellet, Sweetfeed, Meal and

500 Pasture diets. Error bars show standard error of the mean.

Pellets Sweetfeed Copra Meal ■ Pasture

/ A ^ i,i \

Hi* t......i.....t.

0 30 60 90 120 150 180 210 240 270 300 330 360

Time (minutes)

502 Figure 2: Average post feeding insulin response curves for the Pellet, Sweet Feed, Copra Meal and

503 Pasture diets. Error bars show standard error of the mean.