SEARCH

SEARCH BY CITATION

Keywords:

  • horse;
  • laminitis;
  • insulin;
  • proxies;
  • season

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

Reasons for performing study: Insulin resistance may be a risk factor for pasture-associated laminitis. Diagnosis of insulin resistance could help identify individuals at increased risk of laminitis.

Objective: To calculate proxy measurements of insulin sensitivity (reciprocal of the square root of insulin: RISQI and quantitative insulin sensitivity check index: QUICKI) and insulin secretory response (modified insulin-to-glucose ratio: MIRG) based on basal glucose and insulin concentrations in normal (NP) and previously laminitic (PLP) ponies.

Methods: Proxies were calculated in 7 NP and 5 PLP from 20 separate measurements of insulin and glucose taken in spring, summer and winter when ponies were adapted to eating either pasture or hay. Proxies were RISQI: Insulin-0.5, QUICKI: 1/(log[fasting Insulin]+ log[fasting Glucose]) and MIRG: (800−0.3×[Insulin-50]2)/[Glucose-30]. A modified insulin-to-glucose ratio for ponies (MIGRP) was investigated using: (3000−0.012 ×[Insulin-500]2)/[Glucose-30]. Statistical analysis used linear mixed models.

Results: Diet did not significantly affect measurements, so values were pooled for further analysis. RISQI (mean ± s.d.) was lower in PLP (0.26 ± 0.15 [mu/l]-0.5) than NP (0.29 ± 0.12 [mu/l]-0.5; P = 0.05). QUICKI was lower in PLP (0.31 ± 0.05) than NP (0.33 ± 0.04; P = 0.047). There was no difference in MIRG between NP and PLP. MIGRP (median [interquartile range]) was greater in PLP (4.0 [7.9][muins]2/10·l·mggluc) than NP (2.6 [3.2][muins]2/10·l·mggluc; P = 0.022). In spring, NP had higher RISQI and QUICKI and lower MIGRP than PLP (P<0.001). In PLP, RISQI and QUICKI were higher in summer than spring (P<0.02) and MIGRP was lower in summer than other seasons (P<0.01). In NP, RISQI, QUICKI and MIGRP were each different between seasons (P<0.017). MIRG did not vary with season.

Conclusions: RISQI, QUICKI and MIGRP, but not MIRG, differentiated between NP and PLP. None of the proxies accurately identified individual PLP. Seasonal changes in insulin sensitivity and insulin secretory response were apparent.

Potential relevance: Current proxy measurements cannot determine an individual's laminitis susceptibility. MIGRP may be useful in hyperinsulinaemic animals.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

There is currently a great deal of interest in the links between pasture-associated laminitis, insulin resistance and metabolic syndrome [1]. Diagnosis of insulin resistance may help identify laminitis-prone animals and allow the application of preventive countermeasures. However, not all laminitis-prone animals are insulin resistant and not all insulin-resistant animals are prone to laminitis. Diagnosis of insulin resistance in equines is difficult, with no general agreement over the best method to use. Quantitative measurements of insulin sensitivity using a euglycaemic hyperinsulinaemic clamp (EHC) or computerised minimal model analysis [2] are not practical or cost effective for use in the field. Basal insulin concentration is affected by individual, diurnal and seasonal variation and is unsuitable to estimate insulin sensitivity [3–5].

In human studies, proxies based on basal insulin and glucose concentrations correlate well with quantitative measurements of insulin sensitivity [6,7].

Proxies have also been developed and used in 46 healthy horses [8]. The reciprocal of the square root of insulin (RISQI) estimates insulin sensitivity and the modified insulin-to-glucose ratio (MIRG) estimates pancreatic β-cell responsiveness and insulin secretory response [8]. RISQI and MIRG in healthy horses correlated reasonably well (correlation coefficients of 0.774 and 0.754, respectively) with quantitative measurements of insulin sensitivity obtained from minimal model analysis, and use of both proxies allows differentiation between compensated and uncompensated insulin resistance [8]. Reference ranges for RISQI and MIRG were created and the results divided into quintiles based on differing degrees of insulin resistance [8]. Quantitative insulin sensitivity check index (QUICKI) may be the best proxy to use in insulin-resistant human patients [7] and correlated well (r = 0.738) with insulin sensitivity in healthy horses [8].

Previous authors have encountered problems with interpretation of MIRG in hyperinsulinaemic animals [4,9,10]. If insulin concentrations are >50 µiu/ml, despite normal glucose concentrations, MIRG calculations are erroneously low or negative, incorrectly indicating failed pancreatic compensation. Previously, negative MIRG values have been eliminated from analyses [4,10] but this causes the loss of information from an important subset of animals and in hyperinsulinaemic equines an adaptation of the MIRG formula may be required.

The objective of the current study was to calculate proxies based on glucose and insulin measurements in ponies which were well characterised according to their predisposition to laminitis. The effect of season and a pasture or hay diet on measurements was assessed. Secondly, the proxies were used in an attempt to determine an individual's predisposition to laminitis. Additionally, a modified insulin-to-glucose ratio adapted from the MIRG proxy was developed to eliminate negative MIRG values in hyperinsulinaemic equines.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

The work was carried out under a project licence granted under the Animals Scientific Procedures Act (1986) and approved by the institutional ethics and welfare committee. Twelve ponies were studied, 7 of which had no recorded history of laminitis (normal ponies: NP; all intact females) and 5 of which suffered recurrent bouts of pasture-associated laminitis, including at least one episode in the previous 3 years, diagnosed by experienced equine veterinary surgeons based on clinical signs (previously laminitic ponies: PLP; 4 intact females and 1 gelding). None of the ponies had clinical signs of laminitis when blood samples were taken. None of the ponies were obese or showed any clinical signs of pituitary pars intermedia dysfunction (PPID). All ponies had normal cortisol suppression during an overnight dexamethasone suppression test on more than one occasion during the course of the study. The mean ± s.d. age and weight of the NP was 19 ± 4.3 years and 314 ± 71 kg and PLP 16.3 ± 2.7 years and 289 ± 60 kg, respectively. Bodyweight did not change significantly over the course of the study. The median (and interquartile ranges) for body condition score (BCS, on a scale of 1–9) [11] of the NP was 6 (2) and for PLP were 5 (2), and did not vary over the study. All ponies were mixed, native British pony breeds (predominantly New Forest and Welsh ponies and crossbreeds) and were unrelated.

The current study used baseline blood samples obtained from the ponies during 20 separate studies in which glucose and insulin were measured (total: 240 blood samples). Two of these studies were performed in the spring, the first when the ponies were at pasture (pasture adapted) and the second after the ponies had been kept outside in a dirt paddock without access to grass and fed only hay for 3 weeks (hay adapted). Nine studies were performed in the summer, 5 when the ponies were pasture adapted and 4 when the ponies were housed and hay adapted. A further 9 studies were performed in the winter, 5 with the ponies pasture adapted and 4 when housed and hay adapted.

None of the ponies were fasted before blood sampling. They had free access to hay or pasture and water, but no concentrate feed, before sampling. Blood for insulin analysis was collected into serum vacutainers and allowed to clot at 37°C for 20 min. Fluoride oxalate vacutainers were used to collect blood for glucose analysis and kept on ice until centrifuged. All samples were centrifuged (3000 g) at 4°C for 10 min. Serum or plasma was harvested and stored at -80°C. Glucose was analysed by a commercial laboratory using the glucose oxidase methoda. Serum insulin was analysed using a commercially available radioimmunoassay (RIA) kitb, validated for use in equines in the authors' laboratory. When this kit is used to measure high concentrations of insulin in equine samples, dilutional parallelism does not occur if the zero calibrator supplied in the kit is used as a diluent as recommended by the manufacturers [12]. However, dilutional parallelism is seen when insulin-depleted equine serum (IDS) is used as a diluent [12]. All samples were assayed undiluted and samples expected to contain insulin concentrations above the working range of the RIA (>350 µiu/ml) were diluted using IDS prepared in the authors' laboratory by continuously mixing 50 mg activated charcoalc per 1 ml serum for 20 h at 20°C. Serum was then centrifuged twice (3000 g) at 4°C for 20 min, and the supernatant decanted after each centrifugation. Supernatant was filtered through a 0.2 µm syringe filter (Anachem Supatop)d and stored at -80°C. A sample of IDS was analysed in each RIA to confirm no insulin was detected.

Proxies were calculated using the following formulae [8]:

  • image
  • image
  • image

Additionally, a modified insulin-to-glucose ratio adapted from the MIRG proxy (MIGR for ponies; MIGRP) was investigated using the following formula, which was developed based on initial analysis of the results of the current study, and confirmed by correlations with previous insulin resistance data:

  • image

Glucose concentrations in SI units (mmol/l) were converted into mg/dl before insertion into the formulae. Insulin concentration in µiu/ml was used for calculation of formulae. Results for RISQI and MIRG were categorised into reference quintiles based on previously published work. Quintiles for RISQI were: 1st: 0.152–0.295; 2nd: 0.296–0.335; 3rd: 0.336–0.393; 4th: 0.394–0.470; 5th: 0.471–0.953 [mu/l]-0.5. Quintiles for MIRG were: 1st: 1.20–2.12; 2nd: 2.13–3.48; 3rd: 3.49–4.54; 4th: 4.55–5.27; 5th: 5.27–10.67 [muins]2/10·l·mggluc[8].

Data analysis

Linear mixed models (SPSS, PASW Statistics, Version 18)e were used to analyse the data using individual pony as the subject and baseline insulin, RISQI, QUICKI, MIRG or MIGRP as the dependent variables. Season and diet adaptation were included as repeated measures with a compound symmetry repeated covariance format. Group (NP or PLP), season, diet adaptation and the interactions between these 3 factors were included as fixed effects. Bonferroni post hoc tests were used where significant differences were found. Baseline insulin and MIGRP were not normally distributed based on visual inspection of histograms and the Kolmogorov-Smirnov test, and therefore these variables were log transformed before statistical analysis. Differences between individuals were analysed using a Kruskal-Wallis test, with Dunn's post hoc test where applicable (GraphPad Prism Version 5)f.

Spearman correlations were calculated between MIGRP and MIRG, both before and after exclusion of negative MIRG values. Additionally, Spearman correlations were calculated between MIGRP, MIRG and quantitative measures of insulin sensitivity calculated from computerised minimal model analysis of a frequently sampled i.v. glucose tolerance test previously performed on the same individuals. The quantitative measures used were SI (insulin sensitivity) and AIRg (acute insulin response to glucose).

Results are presented as mean ± s.d. unless stated otherwise, and significance was set at P≤0.05.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

Adaptation to hay or pasture did not significantly affect any of the measurements and there was no interaction between diet adaptation and group and/or season (P≥0.09). Thus, the values from different studies on hay and pasture were pooled for further analysis.

Log insulin (median [interquartile range] was significantly (P = 0.024) greater in PLP (19.3 [43.6]µiu/ml) than in NP (12.7 [16.5]µiu/ml). There was a significant effect of season (P = 0.001) on baseline insulin and a significant interaction between group and season (P<0.001; Table 1). Log insulin was significantly higher in PLP than in NP in spring (P<0.001). In NP, there were significant differences between seasons, with log insulin in spring significantly lower than in summer and winter, and log insulin in summer significantly lower than in winter (P<0.013). In PLP, log insulin was significantly lower in summer than in spring and winter (P<0.02).

Table 1. Baseline insulin concentrations and proxy measurements in 7 normal (NP) and 5 previously laminitic (PLP) ponies calculated from blood samples obtained in spring (2 occasions), summer (9 occasions) and winter (9 occasions)
SeasonInsulin (µiu/ml)RISQI [mu/l]-0.5)QUICKIMIRG [muins]2/10·l·mgglucMIGRP [muins]2/10·l·mggluc
NPPLPNPPLPNPPLPNPPLPNPPLP
  1. RISQI = Reciprocal of the square root of insulin, calculated as Insulin-0.5; QUICKI = Quantitative insulin sensitivity check index, calculated as 1/(log[Insulin])+log[Glucose]); MIRG = Modified insulin-to-glucose ratio, calculated as (800−0.3×[Insulin-50]2)/(Glucose-30). Values are reported after exclusion of 1 negative MIRG value in NP and 9 negative MIRG values in PLP; MIGRP = Modified insulin-to-glucose ratio for ponies, calculated as (3000−0.012×[Insulin-500]2)/(Glucose-30); RISQI, QUICKI and MIRG are reported as mean±s.d.; Insulin and MIGRP were not normally distributed, therefore values are reported as median (range). These values were log transformed before statistical analysis; * Significantly different from PLP at the same season for the same proxy measurement (P<0.001); a–cWithin a column, values with different superscript letters are significantly (P<0.02) different from each other.

Spring6.4 (2.7–31.3)*a25.9 (5.1–241.4)a0.41±0.14*a0.19±0.11a0.37±0.04*a0.29±0.04a4.8±2.79.6±3.01.3(0.6–5.9)*a5.1(1.1–28.8)a
Summer10.8 (2.9–172.4)b13.4 (3.3–131.0)b0.31±0.12b0.29±0.13b0.33±0.04b0.33±0.05b7.8±12.37.4±3.92.2(0.5–63.5)b2.7(0.7–21.9)b
Winter19.5 (4.4–80.0)c26.3 (1.6–370.0)a0.24±0.08c0.24±0.16a,b0.31±0.03c0.31±0.06a,b8.6±3.18.3±3.63.8(1.0–13.4)c5.1(0.6–41.6)a

Baseline insulin concentrations in NP ranged from 2.7 to 172.4 µiu/ml and in PLP from 1.6 to 370.0 µiu/ml. Baseline glucose concentrations in NP ranged from 2.1 to 6.6 mmol/l (median 5.0 mmol/l) and in PLP from 3.5 to 6.8 mmol/l (median 4.9 mmol/l).

RISQI was significantly (P = 0.05) lower in PLP (0.26 ± 0.15 [mu/l]-0.5) than in NP (0.29 ± 0.12 [mu/l]-0.5). Season significantly (P = 0.001) affected RISQI; however, there was also a significant (P<0.001) interaction between group and season (Table 1). The reciprocal of the square root of insulin was significantly higher in spring in NP than in PLP (P<0.001). In NP, there were significant differences between seasons, with RISQI in spring significantly higher than in summer and winter and RISQI in summer significantly higher than in winter (P<0.002). In PLP, RISQI was significantly higher in summer than in spring (P = 0.014).

RISQI ranged from 0.08 to 0.61 [mu/l]-0.5 in NP and from 0.05 to 0.80 [mu/l]-0.5 in PLP. There were significant differences in RISQI values between different individuals (P<0.0001) and the 5 PLP did not have the lowest RISQI values. Nine of the ponies (5 NP and 4 PLP) had mean RISQI values that fitted into the 1st (lowest) quintile (0.152–0.295 [mu/l]-0.5) [8]. Three ponies (2 NP and 1 PLP) had mean RISQI values in the 3rd quintile (0.336–0.393 [mu/l]-0.5).

QUICKI was significantly (P = 0.047) lower in PLP (0.31 ± 0.05) than in NP (0.33 ± 0.04). There was a significant effect of season (P = 0.004) and a significant interaction between group and season (P<0.001; Table 1). QUICKI was significantly higher in spring in NP than in PLP (P<0.001). In NP, there were significant differences between seasons, with QUICKI in spring significantly higher than in summer and winter and QUICKI in summer significantly higher than in winter (P<0.006). In PLP, QUICKI was significantly higher in summer than in spring (P = 0.007).

QUICKI ranged from 0.24 to 0.42 in NP and from 0.22 to 0.50 in PLP. There were significant differences in QUICKI values between individuals (P<0.0001) and the 5 PLP did not have the lowest QUICKI values.

MIRG in NP was 7.3 ± 10.6 [muins]2/10·l·mggluc and in PLP -0.3 ± 48.6 [muins]2/10·l·mggluc (Table 1). One negative MIRG value was calculated in the NP, owing to one high insulin concentration. After exclusion of this value, MIRG in NP was 7.9 ± 8.6 [muins]2/10·l·mggluc. Nine negative MIRG values were calculated in 4 of the PLP. After exclusion of these values, MIRG in PLP was 8.0 ± 3.7 [muins]2/10·l·mggluc. There was no significant effect of group (P = 0.087) or season (P = 0.391) on MIRG after exclusion of the negative values.

MIRG ranged from -67.5 to 100.4 [muins]2/10·l·mggluc in NP and from -444.5 to 18.9 [muins]2/10·l·mggluc in PLP. There were significant differences in MIRG between individuals (P<0.0001) and the 5 PLP did not have the highest MIRG values. After exclusion of negative values, 10 ponies (5 NP and 5 PLP) had MIRG values that fitted into the 5th (highest) quintile (5.27–10.67 [muins]2/10·l·mggluc) [8]. One NP had an MIRG above the 5th quintile (13.0 ± 20.9 [muins]2/10·l·mggluc). This was due to one outlying MIRG of 100.4 [muins]2/10·l·mggluc owing to a low glucose concentration (2.1 mmol/l) at one measurement point. After exclusion of this outlier, the MIRG for this individual was 8.4 ± 4.1 [muins]2/10·l·mggluc, placing it into the 5th quintile. One NP had a MIRG in the 4th quintile (4.55–5.27 [muins]2/10·l·mggluc).

Log MIGRP (median [interquartile range] was significantly (P = 0.022) greater in PLP (4.0 [7.9][muins]2/10·l·mggluc) than in NP (2.6 [3.2][muins]2/10·l·mggluc). Season significantly (P<0.001) affected log MIGRP; however, there was also a significant interaction between group and season (P = 0.001; Table 1). Log MIGRP was significantly higher in PLP than in NP in spring (P<0.001). In NP, there were significant differences between seasons, with log MIGRP in spring significantly lower than in summer and winter and log MIGRP in summer significantly lower than in winter (P<0.017). In PLP, log MIGRP was significantly lower in summer compared to spring and winter (P<0.01).

MIGRP ranged from 0.5 to 63.5 [muins]2/10·l·mggluc in NP and from 0.6 to 41.6 [muins]2/10·l·mggluc in PLP. There were significant differences in MIGRP between individuals (P<0.0001) and the 5 PLP did not have the highest MIGRP values.

Spearman correlations between MIGRP, MIRG and quantitative measures of insulin sensitivity (SI and AIRg) are shown in Table 2. MIGRP correlated significantly with MIRG after exclusion of negative MIRG values. Both MIGRP and MIRG after exclusion of negative values correlated significantly with AIRg. Before exclusion of negative values, MIRG did not correlate with MIGRP, SI or AIRg.

Table 2. Spearman correlations between MIGRP and MIRG and between MIGRP, MIRG and quantitative measures of insulin sensitivity measured by minimal model analysis
 MIGRPMIRG (after exclusion of negative values)MIRG (before exclusion of negative values)
Spearman rP valueSpearman rP valueSpearman rP value
  1. MIGRP = Modified insulin-to-glucose ratio for ponies; MIRG = Modified insulin-to-glucose ratio. Values are reported before and after exclusion of 1 negative MIRG value in NP and 9 negative MIRG values in PLP; AIRg = Acute insulin response to glucose measured by minimal model analysis; SI = Insulin sensitivity measured by minimal model analysis; Significant correlations are highlighted in bold.

MIGRP 0.85 0.0004 -0.030.91
AIRg 0.60 0.04 0.65 0.02 0.380.23
SI-0.550.07-0.400.200.090.78

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

In the current study, RISQI, QUICKI and MIGRP could separate groups of ponies based on their predisposition to laminitis; however, differences were small and there was a large overlap between groups. Additionally, they could not accurately determine an individual's susceptibility to laminitis. MIRG was unable to distinguish between NP and PLP.

Proxies were developed to allow prediction of insulin sensitivity and insulin secretory response, which were previously measured by more quantitative methods [8]. These latter methods are expensive, time-consuming and technically difficult, and thus unsuitable for use in the field or in large-scale trials [8]. Insulin sensitivity measured by minimal model analysis (SI) correlated well with RISQI (r = 0.774) and insulin secretory response (acute insulin response to glucose: AIRg) correlated well with MIRG (r = 0.754) in healthy horses [8]. Similar correlations between minimal model parameters and proxies have been reported in human studies [7]. The total predictive power of the proxies in terms of assessing insulin sensitivity in healthy horses was 78–80%, and both proxies had a high specificity (85% for SI and 88% for AIRg) but a low sensitivity (45% for SI and 50% for AIRg) [8].

Reference intervals for RISQI (0.159–0.917 [mu/l]-0.5) and MIRG (1.24–10.26 [muins]2/10·l·mggluc) have been published based on measurements in 46 healthy horses [8]. The horses were Thoroughbreads and Arabians, of which 12 were pregnant mares and 12 weanlings. The division of RISQI and MIRG into reference quintiles aimed to characterise different degrees of insulin resistance. Only 3 of the ponies in the current study fitted into the 3rd quintile for RISQI (2 NP and 1 PLP) and all the others fitted into the 1st (lowest) RISQI quintile, indicating the lowest insulin sensitivity. Eleven of the ponies in the current study fitted into the 5th (highest) quintile for MIRG and only 1 NP fitted into the 4th quintile, indicating a high level of pancreatic β-cell responsiveness, insulin secretory response and pancreatic compensation. Thus, based on RISQI and MIRG values in the current study most ponies, regardless of their predisposition to laminitis, are relatively insulin resistant. Other investigators have also reported that ponies are more insulin resistant than horses [4,13].

The proxy measurements used in man are based on fasting glucose and insulin concentrations allowing maintenance of a steady state of insulinaemia [7]. The ponies in the current study were not fasted before blood sampling and had free access to grass or hay, but not concentrate feed, before sampling. Variation in feed intake before sampling may have accounted for some of the variation in measurements seen in the current study. However, previous studies evaluating RISQI and MIRG in equines have not fasted animals before sampling, allowing comparison of the results of the current study with previous investigations [4,8–10]. Rodents do not have a physiological fasting state, and the imposition of a period of fasting before blood sampling causes depletion of hepatic glycogen content and increases in insulin sensitivity [7]. Equines have evolved to graze virtually continuously, and it is unclear whether fasting would significantly alter measures of insulin sensitivity and secretion, evaluated using either proxy estimates or more quantitative measurements.

Proxies based on baseline glucose and insulin concentrations (in the fasted state in man) reflect mainly the effect of insulin on hepatic glucose production and thus hepatic insulin sensitivity [7]. In contrast, dynamic tests based on responses to the administration of glucose and/or insulin reflect both hepatic and peripheral insulin sensitivity [7,14]. In horses, insulin resistance is thought to arise predominantly as a result of a subnormal response of peripheral tissues (particularly skeletal muscle) to insulin, rather than hepatic insulin resistance. The relative contributions of hepatic and peripheral tissues to insulin resistance have not been investigated in horses, and this may further limit the usefulness of steady-state proxies in horses. In man, hepatic and skeletal muscle insulin resistances are usually proportional to each other [7], but whether this is true for horses is unknown. In man, hepatic insulin resistance is characterised by high fasting glucose concentrations [15], a feature which is not usually apparent in horses and was not present in the ponies investigated in the current study.

One of the main sources of variation in measurements in the current study is likely to be the inherent biological fluctuation in serum insulin concentrations, as demonstrated by the large range of baseline insulin concentrations in both groups. This has also proved a problem in evaluating proxies in man, particularly in insulin-resistant individuals, who have an even greater variation in insulin concentration [6]. Seasonal variation in insulin concentration related to the carbohydrate content of pasture has also been previously demonstrated in horses [5]. Interpretation of insulin concentrations is further hampered by problems associated with validation of the assay in horses [7,12,16].

One of the problems encountered in the current study and in previous studies which have measured MIRG in equines is that individuals which are hyperinsulinaemic (>50 µiu/ml) have erroneously low or negative MIRG values [4,9,10]. MIRG represents the pancreatic β-cell response, with higher MIRG values indicating greater pancreatic compensation and lower values suggesting glucose tolerance [4]. A negative MIRG would thus indicate no secretion of insulin in response to glucose and failed pancreatic compensation. None of the ponies in the current study were hyperglycaemic, so failure of pancreatic insulin secretion is unlikely. Previous investigators have eliminated negative MIRG values from their analyses [4,10] and this was also done in the current study to avoid skewing the results, but this eliminates information from an important subset of hyperinsulinaemic individuals. Another problem was encountered in one individual which had a low baseline glucose and thus a very high MIRG.

In an attempt to increase the utility of MIRG for hyperinsulinaemic equines, an amendment to the previous formula (MIGRP) was calculated to reduce the likelihood of negative MIRG values in hyperinsulinaemic animals. The aims of the MIGRP were to provide an index value (of the insulin response to glucose) that did not start to decrease as soon as serum insulin exceeded 50 µiu/ml, giving the misleading impression of failure of pancreatic compensation. The new index more accurately reflects the situation in ponies, which are often in a state of compensated insulin resistance with basal insulin concentrations approaching 500 µiu/ml. Therefore, MIGRP continues to increase up to this point; however, when glucose rises (indicating failure of pancreatic compensation) MIGRP still declines, as was the original intention for this index. MIGRP was significantly correlated with MIRG after exclusion of negative MIRG values, indicating that the new equation is still suitable for estimation of pancreatic insulin response. Additionally, MIGRP was significantly correlated with AIRg (the quantitative measure of pancreatic insulin response). Overall MIGRP was significantly different between groups, but it could not accurately separate individuals with differing predisposition to laminitis.

Seasonal differences were apparent in both groups of ponies. In NP, the highest RISQI and QUICKI values were measured in spring and the lowest in winter, with summer intermediate. A different trend was apparent in PLP, where RISQI and QUICKI were significantly higher in summer than in spring. This would seem to indicate that both groups of ponies are more sensitive to insulin in summer than in winter. This may reflect an adaptive mechanism to deal with an abundance of food, particularly sugars, in summer pasture and a scarcity in winter. However, unlike NP, PLP do not also show this increased insulin sensitivity in spring, which may increase their susceptibility to laminitis in spring. However, it must be noted that despite their low insulin sensitivity in spring, none of the ponies in the current study developed laminitis at that time.

Group and season did not significantly affect MIRG. However, MIGRP was significantly lower in NP than in PLP in spring, indicating a lower requirement for pancreatic insulin secretion and hence higher glucose tolerance in NP at this time. This fits with the significant difference in RISQI and QUICKI and thus insulin sensitivity between groups in spring. In both groups of ponies, seasonal increases or decreases in RISQI and QUICKI were generally paralleled by inverse changes in MIGRP, indicating pancreatic compensation for changes in insulin sensitivity.

In the current study, there were no significant effects on the proxies when the ponies were adapted to eating pasture or hay. This suggests that any seasonal changes in insulin sensitivity are due to a longer-term adaptation to environmental and nutritional conditions and are less affected by a short-term change in diet to hay.

One of the limitations of the current study is that a smaller number of samples were obtained in the spring than in summer and winter, and the ponies adapted to eating hay were kept outside in spring (in a dirt paddock), whereas they were housed for this dietary adaptation in summer and winter. The ponies eating hay in spring would thus have been more subject to environmental influences such as weather conditions, temperature and day length. Dietary adaptation to hay may have had a different effect on the proxies had the ponies been housed while eating hay in the spring. However, the pasture adaptation was the same for all seasons, suggesting that measurements taken from ponies at pasture in spring may be most likely to yield differences in proxies, as found in the current study.

Proxy measurements of insulin sensitivity have been previously reported in a group of 40 NP and 40 PLP, which included some of the same individuals as in the current study [10]. In contrast to the results reported here, values for RISQI in both groups were lower in summer than in winter. RISQI and MIRG were significantly different between groups in summer but not in winter [10]. This was hypothesised to represent a relative increase in insulin resistance in PLP in summer [10]. Reasons for the different seasonal pattern of results from those in the study reported here are difficult to explain. Differences in management practices of the pony herds to prevent obesity and the development of laminitis by limiting access to grazing during the summer months may have affected the insulin sensitivity of the ponies in the present study.

RISQI and MIRG were also calculated in 54 PLP and 106 NP at pasture in March and May [9]. RISQI was significantly lower in PLP than in NP in both months studied, and RISQI was higher in May in both groups. MIRG was significantly higher in PLP than in NP in both months. In May, a subset (n = 13) of the PLP had developed clinical laminitis and their RISQI was significantly decreased and MIRG significantly increased [9].

In conclusion, RISQI, QUICKI and MIGRP distinguished between NP and PLP in the current study. However, none of the proxies could accurately predict predisposition to laminitis in individual animals. Seasonal differences were apparent, suggesting seasonal changes in insulin sensitivity and insulin secretory response. Both groups in the current study were relatively insulin resistant, despite their differing predisposition to laminitis. However, given the large variation in measurements, both within and between individuals, categorisation of ponies as normal or laminitis prone based on threshold values is unlikely to be accurate. Amendments to the formulae, particularly MIGRP, for use in hyperinsulinaemic individuals is worthy of further investigation in a larger cohort of individuals of different susceptibilities to laminitis.

Source of funding

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

K.E.B. is funded by a BBSRC Case Studentship in association with the Waltham Centre for Pet Nutrition.

Manufacturers' addresses

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References

a Clinical Pathology Laboratory, Royal Veterinary College, London, UK.

b Insulin RIA, Coat-A-Count, Siemens, Surrey, UK.

c Norit, Sigma-Aldrich Company Ltd. Dorset, UK.

d Anachem Instruments Ltd. Bedfordshire, UK.

e IBM, Chicago, Illinois, USA.

f GraphPad Software, San Diego, California, USA.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Manufacturers' addresses
  11. References
  • 1
    Frank, N., Geor, R.J., Bailey, S.R., Durham, A.E. and Johnson, P.J. (2010b) Equine metabolic syndrome. J. vet. intern. Med. 24, 467-475.
  • 2
    Kronfeld, D.S., Treiber, K.H. and Geor, R.J. (2005) Comparison of nonspecific indications and quantitative methods for the assessment of insulin resistance in horses and ponies. J. Am. vet. med. Ass. 226, 712-719.
  • 3
    Firshman, A.M. and Valberg, S.J. (2007) Factors affecting clinical assessment of insulin sensitivity in horses. Equine vet. J. 39, 567-575.
  • 4
    Pratt, S.E., Siciliano, P.D. and Walston, L. (2009) Variation of insulin sensitivity estimates in horses. J. equine vet. Sci. 29, 507-512.
  • 5
    Frank, N., Elliott, S.B., Chameroy, K.A., Toth, F., Chumbler, N.S. and McClamroch, R. (2010a) Association of season and pasture grazing with blood hormone and metabolite concentrations in horses with presumed pituitary pars intermedia dysfunction. J. vet. intern. Med. 24, 1167-1175.
  • 6
    Antuna-Puente, B., Faraj, M., Karelis, A.D., Garrel, D., Prud'homme, D., Rabasa-Lhoret, R. and Bastard, J.P. (2008) HOMA or QUICKI: is it useful to test the reproducibility of formulas? Diabetes Metab. 34, 294-296.
  • 7
    Muniyappa, R., Lee, S., Chen, H. and Quon, M.J. (2008) Current approaches for assessing insulin sensitivity and resistance in vivo: advantages, limitations, and appropriate usage. Am. J. Physiol. Endocrinol. Metab. 294, E15-26.
  • 8
    Treiber, K.H., Kronfeld, D.S., Hess, T.M., Boston, R.C. and Harris, P.A. (2005) Use of proxies and reference quintiles obtained from minimal model analysis for determination of insulin sensitivity and pancreatic beta-cell responsiveness in horses. Am. J. vet. Res. 66, 2114-2121.
  • 9
    Treiber, K.H., Kronfeld, D.S., Hess, T.M., Byrd, B.M., Splan, R.K. and Staniar, W.B. (2006) Evaluation of genetic and metabolic predispositions and nutritional risk factors for pasture-associated laminitis in ponies. J. Am. vet. med. Ass. 228, 1538-1545.
  • 10
    Bailey, S.R., Habershon-Butcher, J.L., Ransom, K.J., Elliott, J. and Menzies-Gow, N.J. (2008) Hypertension and insulin resistance in a mixed-breed population of ponies predisposed to laminitis. Am. J. vet. Res. 69, 122-129.
  • 11
    Henneke, D.R., Potter, G.D., Kreider, J.L. and Yeates, B.F. (1983) Relationship between condition score, physical measurements and body fat percentage in mares. Equine vet. J. 15, 371-372.
  • 12
    Borer, K.E., Berhane, Y., Menzies-Gow, N.J., Harris, P.A. and Elliott, J. (2009) Measurement of concentrations of insulin in equine serum that exceed the working range of radioimmunoassay kits. In: Proceedings of the 48th British Equine Veterinary Association congress. Equine Veterinary Journal Ltd, Fordham. p 103.
  • 13
    Jeffcott, L.B., Field, J.R., McLean, J.G. and O'Dea, K. (1986) Glucose tolerance and insulin sensitivity in ponies and Standardbred horses. Equine vet. J. 18, 97-101.
  • 14
    Abdul-Ghani, M.A., Matsuda, M., Balas, B. and DeFronzo, R.A. (2007) Muscle and liver insulin resistance indexes derived from the oral glucose tolerance test. Diabetes Care 30, 89-94.
  • 15
    Abdul-Ghani, M.A., Tripathy, D. and DeFronzo, R.A. (2006) Contributions of beta-cell dysfunction and insulin resistance to the pathogenesis of impaired glucose tolerance and impaired fasting glucose. Diabetes Care 29, 1130-1139.
  • 16
    Tinworth, K.D., Wynn, P.C., Harris, P.A., Sillence, M.N. and Noble, G.K. (2009) Optimising the Siemens Coat-A-Count Radioimmunoassay to measure insulin in equine plasma. J. equine vet. Sci. 29, 411-413.