Equine laminitis: Induced by 48 h hyperinsulinaemia in Standardbred horses

Authors

  • M. A. De LAAT,

    1. Australian Equine Laminitis Research Unit, School of Veterinary Science, The University of Queensland, Brisbane, Queensland 4072, Australia; Equine Division, Department of Clinical Veterinary Science, University of Liverpool, Neston, CH64 7TE, UK; and
    Search for more papers by this author
  • C. M. McGOWAN,

    1. Australian Equine Laminitis Research Unit, School of Veterinary Science, The University of Queensland, Brisbane, Queensland 4072, Australia; Equine Division, Department of Clinical Veterinary Science, University of Liverpool, Neston, CH64 7TE, UK; and
    Search for more papers by this author
  • M. N. SILLENCE,

    1. Faculty of Science and Technology, Queensland University of Technology, Brisbane, Queensland 4001, Australia.
    Search for more papers by this author
  • C. C. POLLITT

    1. Australian Equine Laminitis Research Unit, School of Veterinary Science, The University of Queensland, Brisbane, Queensland 4072, Australia; Equine Division, Department of Clinical Veterinary Science, University of Liverpool, Neston, CH64 7TE, UK; and
    Search for more papers by this author

Summary

Reasons for performing study: Hyperinsulinaemia is known to induce laminitis experimentally in healthy ponies with no history of the condition. Horses are more insulin sensitive than ponies and whether prolonged hyperinsulinaemia and euglycaemia would have a similar laminitogenic effect requires study.

Objectives: To determine if laminitis results when the prolonged euglycaemic hyperinsulinaemic clamp technique (p-EHC) is applied to clinically normal Standardbred horses, and to monitor hoof wall temperature seeking an association between vascular activity and laminitis development.

Methods: Eight young, clinically normal Standardbred horses were assigned into 4 pairs and within each pair, one was assigned randomly to either treatment (n = 4) or control (n = 4) groups. Treated horses received continuous infusions of insulin and glucose until clinical signs of laminitis developed, at which point the horses were subjected to euthanasia. Control horses received an equivalent volume of a balanced electrolyte infusion for the same period. Hoof wall surface temperature (HWST) was monitored continuously throughout the experimental period.

Results: All horses in the treatment group were calculated to have normal insulin sensitivity. All treated horses, and none in the control group, developed laminitis (P = 0.01). Pronounced digital pulses were a feature of the treatment group, while insignificant digital pulses occurred in control horses. HWST was higher and less variable in treated horses once hyperinsulinaemia was established.

Conclusions: Healthy Standardbred horses subjected to prolonged hyperinsulinaemia develop laminitis within 48 h, demonstrating that laminitis in horses can be triggered by insulin.

Potential relevance: Insulin resistance and the associated hyperinsulinaemia place horses and ponies at risk of developing laminitis. This study demonstrates a need for prompt management of the persistent hyperinsulinaemia seen in some endocrinopathies.

Introduction

Laminitis has been defined as failure of the hoof lamellar-distal phalangeal attachment apparatus (Pollitt 2004), and despite extensive research, its pathogenesis is not explained. Previous investigations have focused on alimentary and inflammatory induction models (Garner et al. 1975; Galey et al. 1991). However, many cases of laminitis are associated with endocrine dysfunction (Anon 2000), suggesting a need for research focused on endocrinopathic laminitis. Conditions associated with endocrinopathic laminitis include equine Cushing's syndrome (Walsh et al. 2009), equine metabolic syndrome or pasture-associated laminitis (Treiber et al. 2006) and iatrogenic corticosteroid administration (Ryu et al. 2004). Common to all these conditions is the development of insulin resistance (IR), manifest as hyperinsulinaemia (Frank 2006; McGowan 2008) and many studies have linked IR to the development of laminitis (McGowan et al. 2004; Bailey et al. 2007; Walsh et al. 2009). Recently, induced hyperinsulinaemia was shown to trigger the development of clinical laminitis in healthy ponies (Asplin et al. 2007a), independent of changes in blood glucose concentration. However, ponies have been shown to be more insulin resistant than horses (Jeffcott et al. 1986; Rijnen and van der Kolk 2003) and insulin resistance and pasture-associated laminitis are frequently documented in ponies (Treiber et al. 2006; Bailey et al. 2008). Whether a similar response to induced hyperinsulinaemia would occur in insulin sensitive horses was uncertain.

The aim of the current study was to determine if laminitis would result after prolonged euglycaemic hyperinsulinaemia in Standardbred horses in the absence of pre-existing hormonal aberrations and insulin resistance. The hypothesis was that persistently high serum insulin concentrations would result in lamellar dysfunction in clinically normal, Standardbred horses. In addition, hyperinsulinaemia causes profound changes in the peripheral microcirculation (Baron 2002), and a vascular pathogenesis for laminitis has long been sought (Bailey et al. 2004). Hoof wall surface temperature (HWST) reflects the state of dermal vessels beneath the hoof wall and as such is a noninvasive method of assessing underlying perfusion (Hood et al. 2001). Thus a secondary aim of this study was to measure HWST continuously in hyperinsulinaemic and control horses seeking an association between vascular changes and laminitis.

Materials and methods

Subjects

Eight healthy Standardbred racehorses obtained from different properties in south-east Queensland were purchased for use in this study. All horses (7 geldings, 1 filly) were recently retired from racing (<6 weeks) and as a result were fit, in moderate-light body condition (body condition score 2.5–4.5/9) (Henneke et al. 1983), mean ± age 5.4 ± 1.95 years. There was no sign of regional adiposity in any individual and all horses scored 0–1/5 on the cresty-neck scoring system (Carter et al. 2009). All of the horses were unremarkable on physical examination, and haematology, biochemistry and urinalysis test results were within the accepted laboratory ranges for equids (Olson et al. 1993). Lateral radiographs of both forelimb feet, taken before the study, showed normal anatomy in all cases, and clinical inspection of the hooves did not detect any abnormalities consistent with previous episodes of laminitis. All horses were videotaped while walking and trotting for evaluation of soundness and any unsound horse was eliminated from the experiment. Dietary intake in the 72 h preceding the experiment was limited to hay and lucerne chaff. Unrestricted access to lucerne chaff, hay and fresh water was provided during the study.

Prolonged euglycaemic-hyperinsulinaemic clamp (p-EHC) technique

The horses were paired and assigned randomly within each pair to either the control or treatment group. Each pair of horses was studied separately so the experiment consisted of 4 replicates. The study was completed in 4 weeks in sub-tropical, mid-winter in a fully enclosed, air-conditioned building with internally located stables (mean ± s.e. ambient temperature of stables: 15.9 ± 0.4°C). Ambient temperature was recorded separately for each horse.

Horses were allowed 24 h acclimatisation to the stables during which time a clinical examination was performed. Prior to the experiment, the skin over both jugular veins of each horse was clipped, surgically prepared and infused with local anaesthetic (2% Lignocaine)1. Flexible, extended use, 14 gauge i.v. catheters (Milacath)2 were inserted into both veins and sutured in place. The left catheter was used to administer the infusions while the right was capped (3-way stopcock)2 for blood sampling. Blood (20 ml) was taken from all horses to establish resting serum insulin and blood glucose concentrations, including 3 samples drawn 10 min apart for insulin sensitivity (SI) calculations in treated horses. Blood samples were also obtained for routine haematology and biochemistry. Free-catch urine samples were assessed for glycosuria by the glucose-oxidase method using a dipstick (Multistix-10SG)3.

Horses in the treatment group received a combined infusion of insulin and glucose via the prolonged-EHC clamp technique which was modelled on the EHC technique described by DeFronzo (DeFronzo et al. 1979) and modified to last for up to 72 h (Asplin et al. 2007a). A constant infusion of insulin was used to produce hyperinsulinaemia, while the rate of glucose infusion was adjusted throughout the experiment to maintain euglycaemia. The amount of glucose administered provides a proxy measure of the SI of muscle and adipose tissue (Firshman and Valberg 2007). Horses in the treatment group received an initial bolus (45 miu/kg bwt) of insulin (Humulin-R)4 diluted in 50 ml of 0.9% saline and administered over 60 s. Immediately following bolus administration, an insulin infusion (6 miu/kg bwt/min) was commenced concurrently with an infusion of 50% glucose (Baxter)5 solution (10 µmol/kg bwt/min). Blood samples for serum insulin concentration determination (Fig 1) included 3 samples obtained at 10 min intervals during the steady state period. The steady state period was defined as a 30 min period when euglycaemia (5 ± 1 mmol/l) had been achieved without the need to alter the glucose infusion rate, and occurred between 1.5 and 3.5 h after bolus administration in all treated horses. Blood glucose was measured using a handheld glucometer (Accu-check)6, which had been calibrated against the hexokinase method using equine blood (data not shown). Samples were taken every 5 min during the first 3.5 h. Once steady state was achieved, blood glucose was analysed every 30 min until euthanasia. The glucose infusion rate was adjusted to maintain euglycaemia.

Figure 1.

Mean serum insulin concentrations (µiu/ml) of treated horses (●) infused with insulin (6 miu/kg bwt/min) compared to control horses (○) infused with balanced electrolyte solution (0.57 ml/kg bwt/h) at various time-points over the experimental period. Serum insulin concentration peaked at 1309 µiu/ml at 4 h post bolus infusion in the treated horses. The time points represent the blood sampling intervals for the treated horses throughout the infusion period. The control horses were sampled at 5 h intervals throughout the infusion period as shown.

Samples were drawn for comparative blood glucose and serum insulin levels in all horses for the duration of the infusion period (Fig 1). Control horses received an infusion of balanced, electrolyte solution (Hartmanns)5 at a rate of 0.57 ml/kg bwt/h for the same period as their paired horse. An EHC was not performed to determine insulin sensitivity in control horses, so an assumption of insulin sensitivity was made based on basal insulin concentrations. Blood samples were placed immediately in plain plastic vacutainers (Vacuette)7, left to clot at room temperature for 30 min and then centrifuged for 10 min at 3000 g. Serum was pipetted into 1 ml aliquots and stored at -80°C until analysed. Serum insulin concentrations were measured by a radioimmunoassay kit (Coat-a-count)8 previously validated for use in horses (McGowan et al. 2008). Where necessary, sample dilutions were prepared using equine serum that contained undetectable levels of insulin, as this kit can suffer intolerance when assay buffer is used as the diluent (Tinworth et al. 2009). Assays were performed in duplicate.

Glucose metabolism (M) was used as a measure of tissue sensitivity to exogenous insulin (DeFronzo et al. 1979) and was calculated during the steady state period of the p-EHC clamp for each treated horse. The amount of glucose metabolised per unit of endogenous insulin (M-to-I) was also calculated for the treated horses as a secondary measure of insulin sensitivity. Calculations were made following standard protocols and included space corrections (Rijnen and van der Kolk 2003).

Hoof wall surface temperature (HWST)

All horses were fitted with noninvasive, surface thermistor probes (Tinytag)9 placed on the dorsal midline of both forelimb hooves 25 mm below the hairline. The probes were connected to dataloggers that recorded the temperature of the hoof wall every minute throughout the experimental period.

Development of clinical laminitis and lamellar histopathology

The infusions continued until the development of lameness in the treated horses. The horses were monitored continuously throughout the experimental period for changes in appetite, body temperature, heart and respiratory rates. Immediately at the onset of Obel grade 2 lameness the horses were subjected to euthanasia and necropsy. All hooves of each horse were disarticulated at the metacarpo-phalangeal joint and dissected to obtain blocks of lamellar tissue that were placed in 4% paraformaldehyde for 24 h, embedded in paraffin, sectioned at 5 µm and stained with haematoxylin and eosin (H&E) and periodic acid-Schiff (PAS). Lamellar tissues were examined by light microscopy and judged positive or negative for laminitis histopathology by one of the authors (C.C.P.) who was blinded to the group of origin.

Radiographic assessment

Plain lateral radiographs of all forelimb hooves were taken before the study and just before euthanasia. Standard measurements to assess changes in hoof and distal phalangeal anatomy were performed on all radiographs, by a blinded investigator experienced in performing these measurements.

Statistical analysis

Comparison of clinical development of laminitis between control and treatment groups was performed using a Fisher's Exact Probability test. Radiographic measurements were compared between groups using a Welch t test. The Wilcoxon rank sum test was used to compare blood glucose and serum insulin values between groups and these values are presented as median (interquartile range). A repeated measures analysis of variance (ANOVA) was used to analyse HWST with ambient temperature as a covariate in the analysis. Histopathological classification of lamellae was compared using a Chi-squared test. Statistical significance was set at P<0.05. All other results are presented as mean ± s.e.

Ethical clearance

The experimental protocol was approved by the animal ethics committee of The University of Queensland, which ensures compliance with the Animal Welfare Act of Queensland (2001) and the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (7th edition 2004). All horses were monitored continuously throughout the experimental period by a registered veterinarian.

Results

Clinical outcomes

All horses in the treatment group developed laminitis compared to none in the control group (P = 0.01). The early clinical signs of laminitis included palpably increased digital pulses, restlessness and episodic shifting of the feet and began at 31.5 ± 4.65 h, with a progression to more consistent shifting of weight and turning to look at the hind feet at 40.5 ± 3.87 h. The mean time taken for the development of Obel grade 2 laminitis; stiff gait at walk, reluctance to trot, difficulty turning (Obel 1948) and hence the time of euthanasia was 46 ± 2.31 h from the start of the insulin infusion.

All horses maintained body temperature, pulse (HR) and respiratory rates (RR) within reference ranges prior to and during the study except for one control horse which was briefly pyretic (39.9°C) following a mild bout of diarrhoea. Average HR and RR were 36 ± 3.26 beats/min and 18 ± 1.15 breaths/min, respectively, in the treatment group prior to the experiment and did not differ significantly at the onset of Obel grade 2 laminitis. Body temperature did not change significantly in the treatment group for the duration of the study. No significant changes in appetite, body temperature, HR or RR were seen amongst the control group. Haematology, serum biochemistry and urinalysis remained within the reference ranges for all horses during the study. Glycosuria was not detected.

Prolonged euglycaemic-hyperinsulinaemic (p-EHC) clamp technique

Basal mean (range) serum insulin and blood glucose concentrations in the treated horses were 13.1 (9.0–35.6) µiu/ml and 6.25 (5.35–7.30) mmol/l, respectively and did not differ from those recorded in control horses (11.6 [9.70–17.0]µiu/ml and 5.0 [4.6–5.7] mmol/l). One treated horse (TH1) did have elevated basal serum insulin concentrations in relation to the other horses; however, subsequent SI calculations showed this horse to be insulin sensitive (Table 1). Another treated horse (TH4) was mildly hyperglycaemic prior to the experiment commencing; however, this horse was of a nervous disposition and the result was attributed to mild stress hyperglycaemia. This horse also proved to have normal SI (Table 1). Blood glucose concentrations also did not differ during the experiment: 4.6 (4.3–5.6) mmol/l for the treated horses and 5.8 (5.4–6.4) mmol/l for the control group. In treated horses serum insulin concentrations increased rapidly from basal levels over the first 4 h of the infusion period and then stabilised for the remainder of the infusion (Fig 1). Peak serum insulin concentrations reached 1309 µiu/ml at 4 h post bolus, and the mean serum insulin concentration for the treatment group was 1036 ± 129 µiu/ml compared to 13.59 ± 2.92 µiu/ml for the control group (P<0.01).

Table 1. Comparison of glucose metabolism and insulin sensitivity measurements for horses in the treatment group (n = 4) obtained during a prolonged-euglycaemic hyperinsulinaemic clamp
VariableTH1TH2TH3TH4Mean ± s.e.
  1. TH1-TH4: Treated horse 1–4, I basal: basal serum insulin, I steady state: serum insulin concentration (I) at steady state, M: amount of glucose metabolised and a measure of sensitivity of tissues to exogenous insulin, M-to-I ratio: glucose metabolised per unit of endogenous insulin.

I basal (µiu/ml)55.67.1915.610.722.3 ± 11.3
I steady state (µiu/ml)10041108105112931114 ± 63
M (mmol/kg bwt/min)0.0250.0220.0210.0260.024 ± 0.001
M-to-I ratio (10−6)3.572.862.922.913.07 ± 0.17

Glucose metabolism (M) values calculated for the treatment group showed all treated horses to be sensitive to exogenous insulin (Table 1). Similarly, the M-to-I ratios indicated that none of the horses in the treatment group were insulin-resistant (Table 1).

Hoof wall surface and ambient temperature

Hoof wall surface temperature varied markedly within and between control horses throughout the infusion period with a coefficient of variation of 42% at the end of the experiment (Fig 2). There was also marked variation in HWST in treated horses initially, but once high insulin concentrations were reached, this variation reduced substantially (P<0.05) so that by the end of the experiment the coefficient of variation was only 4%. In addition, the HWST was higher (P<0.05) in treated horses (Fig 2).

Figure 2.

Mean hoof wall surface temperature (HWST) ± s.e., was recorded for horses infused with 6 miu/kg bwt/min of insulin (inline image) and control horses infused with 0.57 ml/kg bwt/h of balanced electrolyte solution (●) over a 48 h period. The mean ambient temperature (± s.e.) has been subtracted from the mean HWST for both treatment (n=4) and control (n=4) groups. The error bars of control horse HWST, when compared to treated horses, show greater variability.

Lamellar histopathology

All hoof lamellae of the control horses were classified as normal and laminitis negative (Pollitt 1996), whereas 3 out of 4 horses from the treatment group were laminitis positive in all 4 hooves.

The lightest treated horse (328 kg) had histologically normal lamellae in the right hind hoof, but lamellar changes consistent with laminitis in the remaining 3 hooves. The histopathology of the affected hooves included marked lamellar lengthening and attenuation, basement membrane separation at the tips of the secondary epidermal lamellae (SEL) and rounding of the basal cell nuclei (Fig 3). Comparison of the microscopic appearance between the 2 groups was statistically significant (P<0.001).

Figure 3.

Photomicrographs of hoof lamellar histology from control (a, b) and treated (c, d) horses in transverse section. The secondary epidermal lamellae (SELs) of control horses are uniform in length, show symmetrical angulation to the primary epidermal lamellar (PEL) axis, and have rounded tips. The tips of secondary dermal lamellae (SDLs) extend deeply between adjacent SELs and are close to the PEL. SEL basal cell nuclei (solid white arrow in b) are mostly ovoid and apical, and the lamellar basement membrane is closely apposed to the SEL perimeter (white arrowheads). In treated horses, SELs are lengthened, attenuated with pointed tips (black arrowheads in c and d), and the angulation is asymmetrical. Basal cell nuclei are rounded, and many are pyknotic. SDLs on either side of the PEL are obliterated and many basal cells are confluent. Wavy strands of basement membrane, no longer apposed to basal cells, are present at SEL tips (white arrowheads in d). Stain = haematoxylin and eosin.

Radiographic assessment

No radiographic evidence of laminitis was found in any of the pretreatment radiographs and there was no difference in the hoof wall to distal phalanx distance either between groups, or between pre- and post treatment radiographs. Distal phalanx location as determined by the vertical distance from the tip of the distal phalanx to the ground was not different between groups and showed no evidence of rotation or sinking relative to the dorsal hoof wall either before or after the study in any horse.

Discussion

The results are consistent with the hypothesis that prolonged, elevated serum insulin concentrations cause laminitis in the absence of pre-existing insulin-resistance. Despite having normal insulin sensitivity and no evidence of endocrine disease, all the treated horses developed laminitis in 48 h or less. Although the serum insulin concentrations achieved in this study were higher than those commonly associated with equine Cushing's disease (McGowan et al. 2004), or those reported with some cases of pasture-associated, endocrinopathic laminitis (Treiber et al. 2005), concentrations >1000 µiu/ml have occasionally been reported in the field (Reeves et al. 2001). Interestingly, very high serum insulin concentrations have been documented in recurrently laminitic ponies in the absence of clinical signs of laminitis (Bailey et al. 2008). This could suggest that very high serum insulin concentrations alone are not necessarily laminitogenic. Alternatively, it is possible that the high value observed in the Bailey et al. (2008) study did not persist long enough to trigger the condition, as under natural conditions insulin concentrations can fluctuate markedly over time. The minimum period and degree of hyperinsulinaemia required to induce laminitis is yet to be determined, and it is possible that the threshold for the onset of insulin-induced laminitis differs on an individual basis. It is also possible that the degree of tissue resistance to insulin exhibited by ponies affected by pasture-associated laminitis confers them some protection against the toxic effects of insulin. In the present study the subjects had a lower BCS than commonly reported in ponies at risk of pasture-associated laminitis (Treiber et al. 2006) and were sensitive to the effects of insulin and it is possible that this insulin-sensitive state resulted in a more rapid and pronounced response to very high serum insulin concentrations.

Mean M value in this study is higher than previously calculated for ponies and similar to those calculated for horses (Rijnen and van der Kolk 2003; Asplin et al. 2007a) indicating that ponies in previous studies were less sensitive to exogenous insulin than our horses and is consistent with ponies being intrinsically less insulin sensitive than horses. The M-to-I ratio calculated in this study (3.07 ± 0.17) is similar to those recorded previously in ponies (3.5 ± 1.1) (Asplin et al. 2007a) and horses (3.8 ± 2.0) (Rijnen and van der Kolk 2003), a finding inconsistent with reports that show ponies to be inherently more insulin resistant than horses (Jeffcott et al. 1986). However, at very high serum insulin concentrations the M-to-I ratio may underestimate the degree of insulin sensitivity of subjects undergoing a p-EHC following suppression of hepatic glucose production and clearance by saturated insulin receptors (DeFronzo et al. 1979). Therefore, at such high serum insulin concentrations, the p-EHC may more accurately report maximal glucose metabolism as opposed to actual insulin sensitivity.

The ability to prevent insulin-induced laminitis requires a better understanding of its pathogenesis. Speculative theories on the pathogenesis of endocrinopathic laminitis are sometimes extrapolated from the human literature and include alterations in blood flow to the hoof, increased expression of lamellar proinflammatory cytokines that activate inflammation and immunopathology and altered glucose metabolism within the hoof lamellae (Bailey et al. 2004; Johnson et al. 2004).

Vasodilation is a well defined vascular action of insulin (Baron 1994). Increased blood flow and vasodilation in human (Baron and Clark 1997) and rat (Zhang et al. 2004) limb skeletal muscle has been reported following insulin administration. Vascular endothelial cells contain insulin receptors and signaling pathways that activate NO synthase (Rattigan et al. 2007), therefore upregulating NO production in the presence of insulin. NO is a potent vasodilator and increased perfusion of the hooves of treated horses may have resulted from vasodilation secondary to hyperinsulinaemia via this NO pathway. The control horses exhibited more variability in HWST, which may reflect periods of physiological vasodilation and vasoconstriction and their hoof lamellae were lesion free. The observation that HWST did not fall between 25 and 45 h in treated horses, whereas HWST in control horses did, suggests that elevated hoof temperature may be an important feature in the prodromal stage of insulin-induced laminitis. The potential of using elevated HWST as a diagnostic tool or prognostic indicator in the development of insulin-induced or naturally-occurring pasture-associated laminitis is worth exploring.

Polymerase chain reaction studies of glucose transport proteins within hoof lamellae have demonstrated the abundance of GLUT-1 receptors in lamellar tissues indicating that hoof lamellar tissue is capable of up taking glucose independently of insulin (Asplin et al. 2007b). Notwithstanding the high requirement for glucose of hoof lamellae (Wattle and Pollitt 2004), this suggests that failure of glucose uptake is an unlikely cause of lamellar pathology. Conversely, an increase in glucose uptake may contribute to the pathogenesis of endocrinopathic laminitis. It has been shown experimentally that hyperinsulinaemia is able to mediate blood flow-induced glucose uptake in normal muscle (Clark et al. 2003). Studies on human limb blood flow have found that the rate of insulin-mediated glucose metabolism is closely coupled with vasodilation (Mather et al. 2000) further supporting this theory. Despite the fact that in the current study blood glucose concentrations were maintained in the normal range, it is possible that local fluctuations in plasma glucose concentration or an increased uptake of glucose by flow-saturated GLUT-1 transport proteins as a result of increased perfusion resulted in glucotoxic lamellar pathology. In human diabetics glucotoxic endotheliopathy produces advanced glycation endproducts (AGEs) (Uemura et al. 2001), the release of reactive oxygen species (ROS) leading to impaired antioxidant defence (Giugliano et al. 1996) and microvascular dysfunction (Yuan et al. 2007), and the presence of these in lamellar tissue is the subject of further study in our laboratory. Glucose-induced oxidative stress in man has been implicated in the pathogenesis of atherosclerosis and up-regulated MMP-9 activity (Uemura et al. 2001). Digital lamellae may be particularly sensitive to damage by ROS (Loftus et al. 2007), and the presence of MMP-9 in the epidermal lamellae has been documented (Mungall et al. 1998). Thus, it is feasible that laminitis occurring secondary to hyperinsulinaemia may be due, at least in part, to the pathological consequences of lamellar glucotoxicity and oxidative stress. A possible relationship between insulin-induced vasodilation and an increased rate of glucose loading in lamellar tissue deserves further investigation.

Alternatively, vasodilation and elevated perfusion pressure upstream of the microvasculature could result in plasma extravasation and local oedema as occurs in man (Yuan et al. 2007), while dilation of arteriovenous anastomoses diverting blood away from the microvascular bed may promote lamellar hypoxia and coagulopathies (Hood et al. 1993). Digital hypoperfusion, secondary to a decrease in endothelium dependent vasodilation, could be responsible for an ischaemia-reperfusion type injury in the lamellae.

Whether by disturbances in vascular function, glucose metabolism, or from a direct toxic effect on the lamellar tissues, it is clear from the outcome of this study that excess circulating insulin has a rapid, detrimental effect on the lamellar epidermis. The lamellar histopathology showed some similarities to carbohydrate-induced laminitis including basement membrane disruption, lamellar attenuation and rounding of basal cell nuclei (Pollitt 1996), but differed to insulin-induced lamellar lesions in laminitic ponies (subject to a p-EHC) where only occasional, small areas of basement membrane separation at the tips of the SEL were identified (Nourian et al. 2009). There may be a relationship between bodyweight and the severity of the laminitis, with heavier horses suffering more severe clinical disease and an increased degree of basement membrane separation at the tips of the SEL when compared with the lighter horses and ponies (Nourian et al. 2009). Histopathological correlation between the two research models and also between horses and ponies with insulin-induced laminitis is yet to be established. Currently, there are few available data on the histopathological lesions seen in naturally-occurring cases of endocrinopathic laminitis. Syndromes of glucocorticoid excess leading to laminitis have been shown to lead to weakening and separation of the dermo-epidermal interface and lengthening and attenuation of the primary and secondary dermal lamellae (Johnson et al. 2002), findings not dissimilar to those in the current study. Further histological and ultrastructural study of lamellar samples from horses with hyperinsulinaemic laminitis is in progress.

There may also be a relationship between bodyweight and time to onset of laminitis. In this study, the time taken for the development of Obel grade 2 laminitis (Obel 1948) was 46 ± 2.3 h in treated horses (427 ± 68 kg) with a mean serum insulin concentration of 1036 ± 129 µiu/ml. In a previous study using the same technique (serum insulin concentration of 1036 ± 55 µiu/ml) in ponies (258 ± 25 kg), the mean time taken for the development of Obel grade 2 laminitis (Obel 1948) was 55 ± 5.5 h.

Marked hyperinsulinaemia induced laminitis within 48 h in healthy Standardbred horses with normal SI and euglycaemia. The induction of laminitis in the absence of endocrine disease or insulin resistance in these horses emphasises the pathogenicity of hyperinsulinaemia and importance of early detection and management of hyperinsulinaemia in equids, particularly those at risk of developing pasture-associated, endocrinopathic laminitis. Future research into endocrinopathic laminitis is required to improve our understanding of the mechanisms involved in the pathogenesis of the disease and to further elucidate individual risk factors and thresholds for the onset of naturally occurring disease in the equine population.

Acknowledgements

This study was funded by the Rural Industries Research and Development Corporation, Australia. The authors thank Dr Jon Curlewis for his assistance with the insulin assays and Brian Hampson for measuring the radiographs. Technical assistance from Michelle Visser and Alireza Nourian was much appreciated.

Manufacturers' addresses

1 Troy Laboratories Pty, Ltd. Smithfield, New South Wales, Australia.

2 Mila International Inc., Erlanger, Kentucky, USA.

3 Bayer Diagnostics, Pymble, New South Wales, Australia.

4 Eli-Lilly Australia Pty Ltd, West Ryde, New South Wales, Australia.

5 Roche Diagnostics, Mannheim, Germany.

6 Baxter Healthcare Pty Ltd, New South Wales, Australia.

7 Greiner Bio-One, Kremsmünster, Austria.

8 Siemens Diagnostics Ltd, Macquarie Park, New South Wales, Australia.

9 Gemini Data Loggers (UK) Ltd, Chichester, West Sussex, England.

Author contributions All authors contributed to the initiation, conception, planning and writing of this study. Its execution was by M.A.D., C.M.M. and C.C.P., with pathology by M.A.D. and C.C.P., and statistics by M.A.D., C.M.M. and M.N.S.

Ancillary