Dr Finno's current address: University of California, Davis Population Health and Reproduction, Davis, California, USA. Dr McKenzie's current address: Oregon State University, Clinical Sciences, Corvallis, Oregon, USA. Email: email@example.com
Reasons for performing study: Recurrent exertional rhabdomyolysis (RER) occurs in fit, nervous Thoroughbreds fed high nonstructural carbohydrate (NSC) diets. Clinical signs are diminished by feeding low NSC, high fat diets; however, the mechanism is unclear.
Objective: To determine if the glucose, insulin and cortisol response to isocaloric diets varying in fat and NSC availability differ in fit vs. unfit Thoroughbreds with RER.
Materials and methods: Four fit (10 weeks treadmill training) RER Thoroughbred mares were exercised and fed 3 isocaloric (121 MJ/day) diets in a 5 day/diet block design. Two high NSC concentrates, sweet feed (SF) and a processed pelleted feed (PL) and a low starch high fat feed (FAT) were used. After 24 h of rest and a 12 h fast, horses ate half their daily concentrate. Blood sampled for [glucose], [insulin] and [cortisol] was obtained before, immediately after and at 30–60 min intervals for 420 min. After 3–6 months detraining period, the block design was repeated.
Results: Results for SF and PL were similar. Regardless of diet, cortisol was higher in fit vs. unfit horses. Fit horses on SF/PL had higher post prandial [insulin] and insulin:glucose ratio than unfit horses. FAT resulted in lower post prandial [glucose] and [insulin] vs. SF/PL. Higher [insulin] in fit vs. unfit horses was not seen on the FAT diet.
Conclusions: Increased post prandial [glucose], [insulin] and [cortisol] induced by high NSC, but not high fat, feeds are enhanced by fitness in RER horses. This combination may trigger rhabdomyolysis through increased excitability in RER Thoroughbreds.
Exertional rhabdomyolysis is a common cause of muscle pain and myodegeneration in horses (Lentz et al. 1999). An episode of exertional rhabdomyolysis can occur as an isolated event but, in some horses, it can be a chronic condition (Valberg et al. 1993). Several forms of chronic exertional rhabdomyolysis have been described, including type 1 polysaccharide storage myopathy (PSSM), affecting several breeds as well as recurrent exertional rhabdomyolysis (RER), which affects Thoroughbred horses (Valberg 1996). Horses with type 1 PSSM have a gain of function mutation in the glycogen synthase 1 gene (GYS1) resulting in high muscle glycogen and abnormal polysaccharide accumulation (McCue et al. 2008a). To minimise further stimulation of insulin by glycogen synthase, horses with PSSM are fed diets low in nonstructural carbohydrates (NSC) and high in fat (Valberg 1996; Valentine et al. 1997).
Recurrent exertional rhabdomyolysis occurs in approximately 5–10% of racing Thoroughbreds (MacLeay et al. 1999a; McGowan et al. 2002) and appears to be an inherited abnormality in the regulation of muscle contraction (MacLeay et al. 1999b). Intermittent episodes of muscle pain and high post exercise serum creatine kinase (CK) activity in RER horses are exacerbated by a nervous temperament, excitement and diet composition (Beech 1997; MacLeay et al. 1999a). RER differs from PSSM in that Thoroughbred horses with RER often have normal muscle glycogen concentrations, normal glucose tolerance tests (Valberg et al. 1999) and do not have a mutation in GYS1 (McCue et al. 2008b).
Thoroughbreds in race training require diets that contain 117 MJ of digestible energy (DE) per day traditionally provided as >6 kg/day of a high NSC concentrate (Southwood et al. 1993). Previous studies demonstrated that meeting daily energy requirements by feeding 88 MCal of DE in the form of high NSC (sweet feed) or high fat concentrates did not produce significant increases in 4 h post exercise serum CK activity in Thoroughbreds with RER (MacLeay et al. 1999c, 2000). However, when the amount of sweet feed fed to RER horses was doubled and exceeded daily requirements (120.6 MJ DE/day), a significant increase in post exercise serum CK activity occurred within 5 days. Horses on the high NSC diet developed hyperreactivity, a higher heart rate and a higher packed cell volume at rest. In contrast, when a 120.6 MJ/day diet, low in NSC and high in fat, was fed horses showed calmer behaviour, a lower heart rate, packed cell volume and serum CK activity (MacLeay et al. 1999c). It was hypothesised that one beneficial effect of replacing NSC with fat or fibre in the diet of RER horses is a decrease in nervousness/excitabilityassociated with lower blood glucose, insulin and cortisol concentrations.
Traditionally, it is recommended that RER horses are on a consistent exercise regimen because stall rest appears to increase the likelihood of an episode of exertional rhabdomyolysis (MacLeay et al. 1999a). Information is limited regarding the effect of fitness on the glycaemic responses to high NSC feeds in RER horses and on the effect of processing of high NSC feeds to increase availability of starch on the glycaemic response in RER horses. The purpose of this study was to determine if the glucose, insulin and cortisol responses to feeding isocaloric diets varying in fat and NSC availability would differ in fit vs. unfit horses with RER.
Materials and methods
Four Thoroughbred mares with RER (age 6, 9, 9 and 10 years, weight 480–585 kg) were studied. All horses were housed in an accredited facility and were cared for in accordance with principles outlined by the National Institute of Health for the care and use of laboratory animals.
Diagnosis of RER was on the basis of history (i.e. episodes of exercise-induced muscle cramping, signs of pain and stiffness) and increased serum CK activity. Histochemical analysis of muscle biopsy specimens from all horses with RER revealed increased numbers of central nuclei in type 2A and 2B fibres and lack of abnormal polysaccharide when specimens were stained with periodic acid Schiff (PAS) stain (Valberg et al. 1999). Additionally, in vitro contracture responses of external intercostal muscle bundles to stimulation by potassium, caffeine and halothane from the RER horses revealed decreased contracture thresholds when compared with muscle from healthy horses (Lentz et al. 1999).
Each horse was fed one of 3 diets in a replicated 3 × 3 randomised block design where each block lasted 5 days. During each block, horses were fed grass hay from the same source provided at 8.2 kg/500 kg horse/day divided in 2 feedings. All diets were fed to achieve 120.6 MJ/day in digestible energy, which surpassed the daily energy requirements of sedentary horses by 75%.
Diet 1 (SF) consisted of 4 kg of sweet feed1 combined with 0.4 kg of a ration balancer pellet, Diet 2 (PL) (Equine Pelleted Feed)1 consisted of 3.6 kg of pelleted sweet feed combined with 0.4 kg of a ration balancer pellet and Diet 3 (FAT) contained 4 kg of a commercial low starch high fat concentrate (Re-Leve)1. The ration balancer (Stamm 30)1 pellet was added to SF and PL to ensure these diets had similar protein content and that at least minimal daily requirements for vitamins and minerals were provided. It was not necessary to add the ration balancer to the FAT diet because it was already fortified. The composition of the concentrate rations and ration balancer is shown in Table 1 and the nutrient composition of each diet is shown in Table 2. A salt block was provided in each stall throughout the study. Horses were fed concentrate in 2 equal daily feedings, 10 h apart. The horses had unlimited access to water.
Table 1. List of ingredients
Formulation of the commercial products is proprietary. Ingredients are listed in descending order of inclusion.
The horses were trained 5 days a week (Monday–Friday) for 10 weeks on a high-speed treadmill while consuming on average 120.6 MJ/day of sweet feed and grass hay before the diet trial began (fit). Horses remained in their stalls and were not exercised on Saturdays and Sundays. Exercise consisted of repeated sets of 4 min walk (1.8 m/s), 2 min trot (4 m/s) and 2 min canter (7 m/s) at a 0° incline, until 20 min had elapsed. Each horse was then randomly assigned to one of the 3 diets (SF, PL or FAT). The horses consumed the diet for a 5 day period while continuing the same exercise protocol on the treadmill. Horses were observed for signs of tucked-up abdomen, muscle fasciculations in the flank, shifting lameness and sweating in which case exercise would have been terminated.
On the day prior to the test diet day, horses were weighed and not exercised. After aseptic preparation and local anaesthesia of the overlying skin, jugular venous catheters (14 gauge, 14 cm Angiocath)2 were placed the day prior to the test day (Day 5 of each diet period) and the horses were fasted for 12 h. On Day 6, at 08.00 h, baseline blood samples were taken 15 min before feeding and horses were then fed one-half of their daily concentrate rations as a test diet. Venous blood samples were obtained immediately after feeding the test meal (0 min) and then at 15, 30, 60, 120, 180, 240, 300, 360 and 420 min. The horses remained in their stalls throughout the sampling procedure. A glucometer (Precision QID)3 was used to measure blood glucose concentration. Plasma was obtained by centrifugation (1000 g for 20 min at 4°C) within 30 min of collection and frozen at -80°C until analysis. Plasma insulin and cortisol were measured in duplicate using solid-phase radioimmunoassay4, previously validated for specificity and accuracy in the horse (Reimers et al. 1981, 1982).
After 3–6 months of pasture rest where horses were fed grass hay and 2 kg of sweet feed per day (unfit), the 3 diets were again fed in a randomised block design for 5 days while horses rested in their stalls. The diet test procedure described above was repeated on Day 6 of each diet.
Data were analysed using NCSS Statistical and Power Analysis Software5. For each sample an insulin:glucose ratio (I:G) was calculated by dividing the plasma insulin by the glucose concentration. Baseline glucose, insulin and cortisol responses were analysed for the effect of diet and fitness using a 2-way ANOVA. The response of insulin, glucose and cortisol over time after consuming a test meal was analysed by repeated measures ANOVA blocked for diet in fit and in unfit horses. Peak glucose and insulin values were compared by 2-way ANOVA, blocked for diet and fitness. GLM ANOVA and Tukey Kramer post hoc tests with multiple comparisons was used to examine the effect of fitness and diet on glucose, insulin, insulin:glucose ratios and cortisol. Significance was set at P<0.05. Means ± s.e. are reported.
While exercising, all horses were able to maintain treadmill speeds without episodes of clinical rhabdomyolysis on all 3 diets. Bodyweights did not significantly differ between the 3 diets or change significantly between the fit and unfit states as measured on day 5 of each trial period.
On the day of sampling horses quickly and completely consumed their test meal within 35 min; range 11–35 min (mean 20 ± 1 min). There was no difference in the rate of consumption between the 3 diets.
Effect of diet
Form of grain: There was no significant difference in baseline or post prandial responses of glucose, insulin, I:G ratio or cortisol between the 2 high NSC diets (PL and SF) in fit or unfit RER horses.
Fit: Baseline glucose, insulin, I:G ratio and cortisol concentrations were not significantly different among diets in fit RER horses. Horses had a significantly greater glucose response after consuming the SF compared to the FAT diet. Mean glucose concentrations peaked after 60 min at 1.41 ± 0.165 g/l after consuming SF, peaked after 120 min at 1.53 ± 0.143 g/l min after consuming PL and after 60 min at 1.19 ± 0.08 g/l after feeding FAT. Peak glucose concentrations were not significantly different among the diets.
Insulin responses over time were greater (P<0.0033) after consuming the SF and PL diets compared to the FAT diet in fit horses. Mean insulin concentrations peaked after 120 min at 59.7 ± 15.8 μU/mL after consuming SF, after 120 min at 56.7 ± 13.9 μU/mL after consuming PL and after 60 min at 28.9 ± 15.9 μU/mL after feeding FAT. Peak plasma insulin concentrations were not significantly higher (P = 0.06) on SF and SP vs. FAT diets. The insulin:glucose ratios (I:G ratio) were significantly higher for the SF and PL diets compared to the FAT diet (Fig 1a).
Plasma cortisol responses and peak values were not significantly different among the diets. Cortisol increased after consuming the diets and peaked after 60 min at 44 ± 2 µg/l for the SF diet (Fig 2a), peaked after 60 min at 41 ± 6 µg/l with PL and after 30 min at 36 ± 7 µg/l with FAT (Fig 2b).
Unfit: Baseline glucose, insulin and cortisol concentrations were not significantly different among the 3 diets in unfit RER horses. There was no significant difference in glucose, insulin I:G ratio and cortisol responses or peak values among the diets in unfit horses. Mean glucose concentrations peaked after 60 min at 1.38 ± 0.67 g/l after consuming SF, after 60 min at 1.46 ± 0.82 g/l after consuming PL and after 60 min at 1.31 ± 0.27 g/l after consuming FAT. Mean insulin concentrations peaked after 60 min at 24.2 ± 4.5 μU/mL after consuming SF, after 60 min at 30.9 ± 5.4 μU/mL after consuming PL and after 60 min at 21.6 ± 5.0 μU/mL after consuming FAT. There was a trend (P = 0.09) toward higher insulin:glucose ratios (I:G ratio) for the SF and PL diets compared to the FAT diet (Fig 1b).
Plasma cortisol responses were not different among diets. Cortisol peaked immediately after consuming a SF meal at 35 ± 16 µg/l (Fig 2a), peaked immediately after consuming a PL meal at 35 ± 9 µg/l and peaked 30 min after consuming FAT at 39 ± 17 µg/l (Fig 2b). Peaks were not significantly different.
Effect of fitness
Baseline glucose concentrations were not affected by fitness whereas baseline insulin and cortisol were significantly higher in fit horses. There was no significant effect of fitness on the post prandial blood glucose response. In contrast, fit horses on the SF and PL diets had higher insulin responses and I:G ratios than unfit horses on the corresponding diets (P<0.05). Fit horses compared to unfit horses did not show a difference in insulin response or I:G ratio compared to the FAT diet (P>0.05) (Fig 1a,b).
Fit RER horses had significantly higher plasma cortisol concentrations on the SF diet compared to unfit horses on the corresponding diet (Fig 2a). Plasma cortisol concentrations were not significantly different after consuming the FAT diet in fit vs. unfit horses (Fig 2b).
The present study was conducted on a small number of horses susceptible to RER and did not include a control population of healthy Thoroughbreds. Therefore, any significant differences detected in the present study may only be applicable to RER Thoroughbreds and further studies of healthy horses would be necessary to determine if this is a unique or universal response. Furthermore, the small number of horses in the present study probably decreased our power to detect significant differences. For example, significant differences in insulin responses were not detected among diets in unfit horses where such responses have previously been shown to exist when larger groups of horses are studied (Williams et al. 2001). It is, therefore, quite remarkable that, in this study of 4 horses, a strong effect of fitness was detected on both the insulin and cortisol responses to diets varying in NSC.
An effect of diet and fitness on peripheral insulin sensitivity is well described in horses. In healthy fit Arabians, more efficient insulin-mediated glucose clearance was noted when horses were fed a high fat diet as compared to a high starch diet for 8 weeks (Treiber et al. 2006). A high NSC diet fed to 7 healthy Standardbred horses reduced peripheral insulin sensitivity as determined by euglycaemic hyperinsulinaemic clamp (Pratt et al. 2006). However, when these horses were fed a high NSC diet and trained, the reduced insulin sensitivity was mitigated, probably because exercise enhances insulin-induced glucose uptake in skeletal muscle (Pratt et al. 2006). An oral glucose tolerance test was not altered by fitness in the study by Pratt et al. (2006), as it was suggested that this test does not measure peripheral insulin sensitivity (Pratt et al. 2006). Rodiek et al. (1991) also demonstrated similar post prandial glucose and insulin values between fit and unfit horses when feeding a diet of either cracked corn or alfalfa, both providing 17.2 MJ/day. Based on these studies, we expected that the insulin response to the same high NSC meal would be unaffected or potentially lower in fit vs. unfit RER horse; however, we found the opposite to be true. The insulin response was greater and I:G ratios higher when fit RER horses were fed a high NSC load compared to unfit horses. This could be explained by either the need for more insulin to clear an equivalent glucose load in trained RER horses, an enhanced pancreatic beta cell response in fit RER horses or decreased clearance of insulin from the bloodstream in fit horses. The mechanism behind the heightened insulin response to high NSC diets in fit RER horses in the present study will require further elucidation.
Not only did fit RER horses consuming high NSC diets have a significantly higher baseline and post prandial insulin response they also had higher cortisol concentrations in response to feeding a high NSC diet. One of the difficulties in assessing plasma cortisol in horses is that plasma cortisol has a short half life of 1–1.5 h (Lassourd et al. 1996), can fluctuate with the pulsatile release of ACTH (Drake and Evans 1978) and has a circadian rhythm (Stull and Rodiek 1988). In the present study, samples were taken at the same time of day via a catheter placed the previous day and horses were not exercised on the day prior to the trial to minimise any acute effects of experimental stress on plasma cortisol concentrations (Stull and Rodiek 1988). The higher post prandial cortisol concentrations in RER horses fed a high NSC meal are in agreement with a previous study where foals fed a high NSC meal had a higher cortisol response compared to a lower NSC meal (Glade et al. 1984). In addition, higher baseline cortisol concentrations with training have also been reported in Thoroughbreds after 5 weeks of conditioning (Freestone et al. 1991). Other studies have shown no change (Vatistas et al. 1999) in serum cortisol after 9 weeks of conditioning. The higher cortisol in fit RER horses fed a high NSC diet suggest that fit RER horses are probably more easily stressed on a high NSC diet, which is consistent with the excitable behaviour noted in Thoroughbreds fed high grain diets (Davidson et al. 1998; Jansson et al. 2002; Lindberg et al. 2006), particularly those predisposed to RER (MacLeay et al. 1999a; McGowan et al. 2002).
It is interesting to note that previous studies have shown a close relationship between excitable behavior, high NSC diets, fitness and a predisposition to rhabdomyolysis in RER susceptible horses (MacLeay et al. 1999a; McGowan et al. 2002). The reason why this strong association exists is not known. The potential predisposing effect of a high NSC diet on RER may be through its ability to exacerbate nervous behaviour as a result of alterations in central nervous system (CNS) neurotransmitters. Insulin stimulates uptake of amino acids other than tryptophan into muscle tissue leaving relatively more circulating tryptophan to compete for uptake into the CNS. Increased uptake of tryptophan can enhance the synthesis of serotonin in the brain under certain conditions (Young 1996). Therefore, it is possible that the high insulin response in fit vs. unfit RER horses leaves relatively higher circulating serum tryptophan concentrations, which could alter CNS neurotransmitters over time and contribute to altered behaviour. A clear relationship between altered ratio of serum tryptophan to large neutral amino acids after feeding was not documented in a previous study (Wilson et al. 2007) comparing high NSC and high fat meals, although a trend toward an interaction between post prandial sampling time and diet was noted. The trend noted showed that foals fed a high fat/high fibre diet had a higher ratio of tryptophan to large neutral amino acids at the start and the end of feeding, while this ratio peaked in between these time points in foals fed a high starch/high sugar diet (Wilson et al. 2007). We hypothesise that Thoroughbreds with RER have an inherited abnormality (Dranchak et al. 2005) in which altered regulation of muscle contraction occurs in response to excitement and stress (Lentz et al. 1999). Higher insulin and cortisol levels in fit RER horses on high NSC diets could increase excitability and trigger rhabdomyolysis in a fashion somewhat similar to stress-induced malignant hyperthermia (Mickelson and Louis 1996).
A low starch, high fat diet fed for 3 weeks results in fewer episodes of rhabdomyolysis and lower serum CK activity in RER horses compared to an isocaloric high NSC diet without altering muscle glycogen or lactate metabolism (McKenzie et al. 2003). The results of the present study suggest that feeding a high fat diet to RER horses may mitigate the increase in baseline and post prandial cortisol concentration that occurs with fitness and this may represent a decrease in stress/excitability. Crandell et al. (1999) also found that feeding a diet high in fat for 5 weeks to fit horses lowers plasma cortisol concentrations. In the present study, horses were only fed the FAT diet for 5 days. It is possible that results could have been more or perhaps less pronounced if the adaptation period to high fat had been longer. A further effect of the high fat diet on behaviour may come from the lower baseline and post prandial serum insulin concentrations, which potentially could blunt the effect of insulin on altering tryptophan and CNS neurotransmitters.
In conclusion, this study showed that RER horses in the fit state compared to the unfit state had unexpectedly high insulin and cortisol responses to a high NSC meal but not a high fat meal. These results may provide additional evidence that a low NSC high fat diet may have a beneficial effect in RER horses by decreasing stress/excitable behaviour.
Conflicts of interest
1 Farmer's Mill, Inc, Hallway Feeds Lexington, Kentucky, USA.
2 Becton Dickinson, Franklin Lakes, New Jersey, USA.
3 Abbott Laboratories, Abbott Park, Illinois, USA.