Materials and Methods

Horses—Six mature Standardbreds (2 mares and 4 geldings) were used in the study. Horses were 4 to 5 years old with a mean ± SD body weight of 443.2 ± 4.7 kg. Horses had been allowed to rest in a paddock for 1 month preceding the study.

During the study, horses were housed in box stalls. Each day, horses were allowed access to a small paddock for approximately 4 hours. While in the paddock, a muzzle was applied to each horse to restrict grazing. Throughout the study, horses were fed approximately 12 kg of grass hay daily and provided access to fresh water and a salt block at all times. Because an increase in dietary hydrolyzable carbohydrate is 1 factor that may contribute to altered Si,^10,11 grain-concentrate was not fed. All procedures were approved by the Animal Care Committee of the University of Guelph.

Experimental design—The design of the study was based on that of another study⁷ in which investigators examined the effects of dexamethasone treatment on Si in horses with polysaccharide storage myopathy. A balanced crossover design was used. Horses received IV administrations of dexamethasone^a (0.08 mg/kg, IV) or the equivalent volume of saline (0.9% NaCl) solution (control treatment) every 48 hours for 3 weeks (total of 11 treatments). After a 3-week washout period, the experimental protocol was repeated with horses receiving the other treatment.

Horses were randomly assigned to initial treatments (3 horses initially received saline solution, and 3 horses initially received dexamethasone). Treatments were administered at 8 AM. First day of each treatment period was designated as day 0.

Blood samples were collected at 7:30 AM 2 days before and 1, 7, 14, and 21 days after the start of each treatment for measurement of plasma glucose and serum nonesterified fatty acids, triglyceride, and insulin concentrations. Serum cortisol concentrations were measured in blood samples obtained on days −2, 7, and 21.

FSIGT—A 3-hour FSIGT was administered 2 days after the end of each treatment period. Feed was withheld for 12 hours before the procedure. At 7 AM on the day of an FSIGT, horses were weighed to the nearest 0.5 kg by use of an electronic scale.^b The skin overlying one of the jugular veins was aseptically prepared and injected locally with anesthetic, and a catheter^c was inserted. During a subsequent 30-minute rest period, baseline blood samples were obtained (30, 15, and 1 minute before glucose injection). Then, a glucose solution (50% dextrose [wt:vol]) was injected (0.3 g/kg) during a period of 30 seconds via the catheter inserted in the jugular vein. Additional blood samples were collected at 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 14, 16, 19, 22, 24, 25, 27, 30, 35, 40, 50, 60, 70, 80, 90, 100, 120, 150, and 180 minutes after glucose administration for measurement of plasma glucose and serum insulin concentrations. Twenty minutes after injection of the glucose bolus, insulin^d (20 mU/kg) was administered via the catheter inserted in the jugular vein. Horses were restrained in standing position in stocks for the first 40 minutes of the FSIGT, after which they were loosely restrained in a 3.5 × 4-m stall.

Analysis of samples—Blood samples for subsequent assay of plasma glucose concentrations were placed into evacuated tubes that contained EDTA, whereas samples for measurement of serum insulin, cortisol, nonesterified fatty acid, and triglyceride concentrations were placed into evacuated tubes that contained no additive. All blood samples were centrifuged at 1,600 × g for 15 minutes, after which plasma or serum was harvested and stored at −20°C until analysis.

Plasma glucose concentrations were spectrophotometrically measured in triplicate by use of a microplate reader and commercially available kit.^e Serum triglyceride and nonesterified fatty acid concentrations were measured at a commercial laboratory^f by use of enzymatic colorimetric methods and an automated analyzer.^g Serum insulin and cortisol concentrations were measured in duplicate by use of commercially available radioimmunoassay kits^h,i that have been validated for use in samples obtained from horses.¹² The intra-assay and interassay coefficients of variation (calculated as [SD/ mean] × 100) were 2.9% and 4.8%, respectively, for the insulin assay and 3.3% and 5.7%, respectively, for the cortisol assay.

Values for Si, Sg, disposition index, and AIRg were calculated from the plasma glucose and serum insulin concentration data from each FSIGT test by use of commercially available software.^j Values for Si and Sg were calculated by minimal model analysis, as described elsewhere.^10,11,13 The Si represents the capacity for insulin to promote glucose disposal, whereas Sg represents the capacity of glucose to mediate its own disposal independent of a change in circulating insulin concentrations. The AIRg was estimated by calculation of the incremental area under the insulin curve generated from data obtained during the first 10 minutes after glucose injection before insulin administration. The disposition index (calculated as Si × AIRg) represents the appropriateness of the pancreatic beta-cell response relative to the degree of insulin resistance in the tissues.

Resting values for plasma glucose and serum insulin concentrations measured on days −2, 1, 7, 14, and 21 were used to calculate proxies for Si (ie, RISQI) and pancreatic beta-cell responsiveness (ie, MIRG), as described elsewhere.¹⁴ The value for RISQI was calculated by use of the following equation:

View Expanded

The value for MIRG was calculated by use of the following equation:

View Expanded

Statistical analysis—Values were expressed as mean ± SEM. Effects of dexamethasone on variables for the minimal model were evaluated by use of paired Student t tests. Results for basal plasma glucose concentration; serum insulin, cortisol, triglyceride, and nonesterified fatty acids concentrations; proxies (RISQI and MIRG); and glucose and insulin concentrations during the FSIGT were compared by use of an ANOVA with repeated measures. When a significant value for the F ratio was obtained, least square means were compared by use of the Bonferroni t test for multiple comparisons. Spearman rank correlation coefficients were calculated to evaluate relationships between variables for the minimal model and values for the proxies obtained on day 21. Statistical analyses were conducted by use of computer software programs.^k,l For all analyses, significance was designated at a value of P < 0.05.

Results

Mean serum cortisol concentration did not differ between control and dexamethasone-treated horses on day −2, but it was significantly (P < 0.001) lower in dexamethasone-treated horses than in control horses on days 7 and 21 (Table 1). Dexamethasone treatment significantly affected basal plasma glucose (P = 0.01) and serum insulin (P < 0.001) concentrations during the 21-day treatment period (Figure 1). Compared with control horses, mean serum insulin concentration was significantly (P = 0.01) higher in dexamethasonetreated horses on days 7, 14, and 21. Similarly, mean plasma glucose concentration was significantly (P = 0.03) higher in dexamethasone-treated horses than in control horses on day 14; the plasma glucose concentration in dexamethasone-treated horses was higher but not significantly different on days 7 (P = 0.09) and 21 (P = 0.07). Mean serum triglyceride concentrations did not differ between dexamethasone-treated and control horses on any day. Mean serum nonesterified fatty acid concentration was significantly higher in dexamethasone-treated horses than in control horses on day −2 only.

View Full Size

Mean ± SEM plasma glucose (A) and serum insulin (B) concentrations in 6 healthy horses at various days of a 21day period during IV administration of dexamethasone (0.08 mg/kg; white bars) or an equivalent volume of saline (0.9% NaCl) solution (control treatment; black bars). Day of initial administration was designated as day 0. *Within a day, the plasma glucose concentration differs significantly (P = 0.03) between dexamethasone-treated and control horses. †Within a day, serum insulin concentration differs significantly (P = 0.01) between dexamethasone-treated and control horses.

Citation: American Journal of Veterinary Research 68, 7; 10.2460/ajvr.68.7.753

• Download Figure
• Download figure as PowerPoint slide

Mean plasma glucose concentration was not significantly different between treatments at any time point during the 3-hour FSIGT (Figure 2). The mean value of the mean serum insulin concentration in samples obtained before the FSIGT (ie, values obtained at −30, −15, and −1 minute were averaged) was significantly higher (P = 0.01) in dexamethasone-treated horses (mean ± SEM, 22.4 ± 3.5 μU/mL) than in control horses (5.8 ± 2.1 μU/mL). In FSIGT samples, mean serum insulin concentrations were also significantly higher for dexamethasone-treated horses from 0 to 20 minutes and from 40 to 150 minutes.

View Full Size

Mean ± SEM plasma glucose (A) and serum insulin (B) concentrations during the insulin-modified FSIGT for 6 horses administered dexamethasone (diamonds) or saline solution (control treatment; squares). Serum insulin concentrations differ significantly (P < 0.05) between dexamethasone-treated and control horses from −30 to 20 minutes and from 40 to 150 minutes (horizontal bars).

Citation: American Journal of Veterinary Research 68, 7; 10.2460/ajvr.68.7.753

• Download Figure
• Download figure as PowerPoint slide

The minimal model was successfully applied to the FSIGT glucose and insulin data for all horses during both treatments (r² = 0.971 ± 0.004). Compared with control horses, the Si was significantly (P < 0.001) lower and the AIRg was significantly (P < 0.01) higher in dexamethasone-treated horses (Table 2). There was no significant difference in Sg (P = 0.22) or disposition index (P = 0.09) between treatments.

The proxy for insulin sensitivity (ie, RISQI) was significantly lower in dexamethasone-treated horses than in control horses on days 7, 14, and 21 of treatment (Figure 3). On day 21, RISQI was significantly (P = 0.01) lower (approx half) in dexamethasone-treated horses (mean ± SEM, 0.214 ± 0.028 [mU/L]^−0.5), compared with the value in control horses (0.413 ± 0.049 [mU/L]^−0.5). Conversely, MIRG, the proxy for beta-cell responsiveness, was significantly higher in dexamethasone-treated horses than in control horses on days 7, 14, and 21. Values for RISQI concentration versus minimal model Si and for MIRG concentration versus AIRg were plotted (Figure 4). Spearman rank correlation analysis revealed significant relationships between RISQI and Si (r = 0.846; P < 0.001) and between MIRG and AIRg (r = 0.683; P = 0.014).

View Full Size

Mean ± SEM values for proxies of insulin sensitivity (RISQI; A) and pancreatic beta-cell responsiveness (MIRG; B) for 6 horses during administration of dexamethasone (white bars) and saline solution (control treatment; black bars), as determined on the basis of basal plasma data. Day of initial administration was designated as day 0. *Within a day, values differ significantly (P < 0.05) between dexamethasone-treated and control horses. mU_ins = Milliunits of insulin. mg_glu = Milligrams of glucose.

Citation: American Journal of Veterinary Research 68, 7; 10.2460/ajvr.68.7.753

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Plots of RISQI versus Si (A) and MIRG versus beta-cell response (AIRg; B) for the minimal model analysis in 6 horses administered dexamethasone (diamonds) or saline solution (control treatment; squares). Both relationships were significantly correlated (RISQI and Si, r = 0.846; P < 0.001; MIRG and AIRg, r = 0.683; P = 0.014; Spearman rank correlations). See Figure 3 for key.

Citation: American Journal of Veterinary Research 68, 7; 10.2460/ajvr.68.7.753

• Download Figure
• Download figure as PowerPoint slide

Discussion

In the study reported here, we examined the effects of dexamethasone administration (0.08 mg/kg, IV) during a 21-day period on Si and glucose dynamics in Standardbreds. The main findings of the study were a decrease in Si and increase in AIRg (as measured by minimal model analysis of an FSIGT), an approximately 2-fold increase in basal serum insulin concentration following dexamethasone treatment, and a decrease in RISQI and increase in MIRG (as measured by proxy analysis) by day 21 of dexamethasone administration. These observations are consistent with the development of partially compensated insulin resistance in dexamethasone-treated horses.

Minimal model analysis was used to quantify insulin sensitivity. This model is a physiologic compartmental representation of the glucose-insulin regulatory system and includes 4 indices (Si, Sg, AIRg, and disposition index) derived from analysis of the glucose and insulin data. In humans, this method is used to understand the composite effects of insulin secretion and insulin sensitivity on glucose tolerance and risk for developing type 2 diabetes mellitus.¹⁵ In horses, this technique has been used to assess glucose and insulin dynamics in response to changes in exercise and diet^6,10,14,16 and assess insulin resistance in equids predisposed to development of laminitis.^6,14 This is the only method that differentiates between the action of insulin on glucose uptake (Si) and the capacity of tissues for glucose uptake independent of insulin stimulation (glucose effectiveness, [ie, Sg]).⁶ The AIRg provides a measure of insulin secretion during the first 10 minutes after glucose injection, whereas disposition index (a product of Si and AIRg) reflects the combined effect of these 2 factors (insulin efficiency and insulin secretion) to minimize hyperglycemia.

In the study reported here, minimal model analysis revealed a significant decrease in Si after dexamethasone administration. Despite a decrease in Si, there was no significant change in Sg or disposition index, with the latter most likely reflecting an increase in insulin secretion (ie, compensation for a decrease in insulin sensitivity). Values for variables in the minimal model were within the 95% reference interval, which was determined on the basis of data from a population of 48 healthy Thoroughbreds.¹⁴ However, Si values for all dexamethasone-treated horses were in the lowest quintile, whereas all values for AIRg were within the highest quintile.¹⁴

To assess changes in insulin sensitivity and insulin response during dexamethasone administration, proxy analysis of basal data for plasma glucose and insulin concentrations (a method developed for use in horses and ponies^5,14) was applied. Results of this analysis further supported the finding of compensated insulin resistance derived from the minimal model analysis, with evidence of insulin resistance and increased pancreatic insulin response after 7 days of dexamethasone administration. Moderate correlations between the 1-sample basal proxies and the minimal model variables are consistent with results of studies^6,14 in healthy horses and in ponies predisposed to laminitis.

The decrease in Si after dexamethasone administration in the study reported here is similar to the results of a study⁷ in Quarter Horses with polysaccharide storage myopathy in which administration in accordance with a similar treatment protocol resulted in a reduction in whole-body insulin-stimulated glucose uptake measured by use of a euglycemic-hyperinsulinemic clamp. The increase in resting serum insulin concentration observed after 7 days of dexamethasone administration is also consistent with the results of another study⁸ in which investigators reported the effects of acute administration of triamcinolone acetonide on glucose and lipid metabolism in horses. When administered IV or IM at a dosage of 0.05 mg/kg, triamcinolone acetonide induced hyperglycemia, hyperinsulinemia, and hypertriglyceridemia that persisted for 3 or 4 days.⁸ In our study, plasma glucose concentrations were higher on days 7, 14, and 21 of dexamethasone administration, although the concentration was significantly higher only on day 14. However, increases in triglyceride or nonesterified fatty acid concentrations were not evident after dexamethasone administration. The more extensive changes in lipid metabolism observed after triamcinolone treatment, compared with changes after dexamethasone administration, may be attributable to the type of glucocorticoid administered. Interestingly, investigators in 1 study⁷ reported an increase in the sensitivity of peripheral tissues to administration of exogenous insulin in Standardbreds after a single dose of hydrocortisone. An apparent increase in glucose use as a result of cortisol excess has also been reported in dogs with hyperadrenocorticism and in healthy humans.^17–19

Results of the study reported here provide evidence for a major impact of glucocorticoid treatment on aspects of insulin sensitivity and glucose dynamics in horses. The extent to which such glucocorticoid-associated insulin refractoriness could contribute to an increased risk for development of laminitis in horses is unknown,¹ and little quantitative evidence at the wholeanimal level exists to support possible mechanisms for development of laminitis in horses in response to exogenous administration of glucocorticoids. After a 21-day dexamethasone treatment regimen, there was no evidence for the development of laminitis in the horses used in the study, and this is consistent with observations from other studies^7,8,20 in which dexamethasone or triamcinolone acetonide was administered. Interestingly, in another study,⁴ dexamethasone treatment of horses resulted in an apparent decrease in vascular perfusion of the skin, which was evident as a decrease in skin temperature. Furthermore, dexamethasone treatment potentiated decreases in skin temperature in response to dermal administration of the A₁-adrenoceptor agonist phenylephrine.⁴ The authors of that study hypothesized that similar alterations in the function of digital blood vessels after glucocorticoid administration could contribute to the pathogenesis of laminitis, particularly in animals with disease states (eg, sepsis or colic) that are characterized by peripheral hypoperfusion, increased circulating concentrations of catecholamines, or both.

In the study reported here, we quantified a significant decrease in insulin sensitivity in Standardbreds after 21 days of dexamethasone administration. The decrease in Si was compensated for by an increase in beta-cell response and insulin secretion such that no change in the disposition index was evident. The significant decrease in RISQI and increase in MIRG detected following long-term dexamethasone administration further corroborated the compensated insulin resistance determined by use of minimal model analysis of FSIGT data.