1st Page

Journal of Animal Science - Article



This article in

  1. Vol. 87 No. 1, p. 131-135
    Received: May 14, 2008
    Published: December 5, 2014

    2 Corresponding author(s):


Effect of a single dose of dexamethasone on glucose homeostasis in healthy horses by using the combined intravenous glucose and insulin test1

  1. J. C. Haffner*2,
  2. H. Eiler,
  3. R. M. Hoffman*,
  4. K. A. Fecteau and
  5. J. W. Oliver
  1. Department of Agribusiness and Agriscience, Horse Science Center, Middle Tennessee State University, Murfreesboro 37129; and
    Department of Comparative Medicine, College of Veterinary Medicine, The University of Tennessee, 2407 River Drive A205, Knoxville 37996


Sustained dexamethasone administration to horses results in insulin resistance, which may predispose them to laminitis. A single dose of dexamethasone is commonly used as a diagnostic aid, yet the effect of a single dose of dexamethasone on glucose homeostasis in horses is not well defined. The objective of this study was to characterize the change in glucose dynamics over time in response to a single dose of dexamethasone. A combined glucose-insulin tolerance test (CGIT) was performed on 6 adult geldings before and at 2, 24, and 72 h postdexamethasone (40 μg/kg of BW, i.v.); a minimum of 1 wk of rest was allowed between treatments. Before any treatment, the CGIT resulted in a hyperglycemic phase followed by a hypoglycemic phase. Dexamethasone affected glucose dynamics in 3 ways: 1) at 2 h, dexamethasone shortened the ascending branch of the negative phase (P < 0.001) of the test, indicating moderate insulin resistance; 2) at 24 h, dexamethasone impaired glucose clearance by extending the positive phase and eliminating the negative phase while insulin was elevated before the CGIT, indicating a decreased response to insulin; and 3) at 72 h, dexamethasone caused a deeper nadir value (P < 0.001) compared with predexamethasone, indicating an increased response to insulin. It was concluded that dexamethasone decreased the response to insulin as early as 2 h and maximally at 24 h. At 72 h, dexamethasone caused an increased response to insulin, which was unexpected.


Repeated administration of glucocorticoids to horses has been shown to result in increased blood glucose and insulin, and increased insulin resistance (Cartmill et al., 2003; Firshman et al., 2005; Ruzzin et al., 2005). Tiley et al. (2008) showed in horses that 21 d of dexamethasone treatment resulted in substantial insulin resistance and impaired glycogen synthetase kinase phosphorylation in muscle, whereas glucose transporter 4 content was not affected. Triamcinolone acetonide, given intravenously or intramuscularly, induced a prolonged period of hyperglycemia and hyperinsulinemia (French et al., 2000). Investigating the effects of a single dose is important because a dose of dexamethasone alone (Dybdal et al., 1994) or in combination with thyrotropin-releasing hormone (Eiler et al., 1997; Frank et al., 2006) is used in testing for equine Cushing’s disease. Administration of glucocorticoids, especially dexamethasone and triamcinolone, is suspected to be associated with increased risk of acute laminitic episodes in horses previously affected by laminitis from any cause, and particularly in horses that suffer from recurrent chronic laminitis caused by Cushing’s disease (Johnson et al., 2002, 2004; Schott, 2002); the physiological effect of a single dose of dexamethasone on glucose homeostasis in horses has not been established.

A new tool to evaluate glucose homeostasis in horses is an intravenous combined glucose-insulin test (CGIT) developed in our laboratory (Eiler et al., 2005). The CGIT combines 2 well-known procedures, the glucose tolerance and insulin sensitivity tests, into a single procedure in an effort both to increase diagnostic accuracy compared with a single test and to circumvent the complexity of more elaborate testing, such as the hyperinsulinemic-euglycemic clamp test (DeFronzo et al., 1979). The purpose of this study was to characterize the change in glucose dynamics over time in response to a single dose of dexamethasone. It was our hypothesis that a single dose of dexamethasone would affect glucose homeostasis in horses.


This study protocol was conducted with the approval of the Middle Tennessee State University Institutional Animal Care and Use Committee.


Six clinically healthy, mixed-breed adult geldings from The Horse Science Center at Middle Tennessee State University were used. Horses had a mean ± SD age of 9.8 ± 4.8 yr (range: 4 to 16 yr). Mean BW was 535 ± 13.8 kg (range 505 to546 kg), and BCS (Henneke et al., 1983) were between 5 and 6 (moderate) on a scale of 1 (very thin) to 9 (obese). Horses were housed in separate stalls during the test periods. They were kept on pasture when not being used. Geldings were fed prairie grass hay at approximately 1% of BW, divided twice daily, and 0.5 kg of a commercial grain mix once daily in the morning. A sample of the prairie grass hay submitted for analysis by Equi-Analytical Laboratories (Ithaca, NY) indicated 6.4% CP, 74.4% NDF, 46.1% ADF, 2.0% simple sugars, 3.6% starch, and 1.6% crude fat on a DM basis. Water was available for ad libitum intake.

Blood Analyses

Blood glucose was analyzed by using a validated (Eiler et al., 2005) handheld strip glucometer (Precision Xtra, MediSense, Abbott Laboratories, Alameda, CA) designed for use with a single drop of fresh capillary blood. Plasma insulin and cortisol were analyzed by RIA kits (Coat-A-Count Insulin and Coat-A-Count Cortisol from Diagnostic Products Corporation, Los Angeles, CA).

Testing Protocol

Five treatments were evaluated: spontaneous fluctuation of glucose over the testing period with no treatment; the effect of an intravenous CGIT on blood glucose response without dexamethasone; and the effect of dexamethasone on glucose and insulin tolerance as assessed by a CGIT at 2, 24, and 72 h after dexamethasone administration. All testing began at 0800 h. In all groups, blood samples were obtained and injections were performed by using jugular catheters. Catheters were flushed after use and kept patent with heparinized saline (0.9% NaCl plus 10 U/mL of heparin solution). Ten-milliliter blood samples were collected in 10-mL Vacutainer tubes with sodium heparin (Becton Dickinson and Company, Franklin Lakes, NJ) at time 0 (baseline) and at 5, 15, 25, 35, 45, 60, 75, 90, 105, 120, 135, and 150 min. Food was withheld overnight, beginning 12 h before testing, to standardize conditions for all horses. Water and mixed-grass hay were available at the beginning of the test for ad libitum consumption during the test to minimize stress on the horses. Previous experiments in our laboratory and by others indicated that water and grass hay did not affect glycemia during the test period (Stull and Rodiek, 1988; Ralston, 2002).

Intravenous CGIT Protocol

In this protocol (Eiler et al., 2005), a 50% dextrose solution (150 mg/kg of BW, i.v.) was administered rapidly (<1 min). Immediately thereafter (<6 s), a bolus (0.1 U/kg of BW) of human recombinant DNA insulin (Humulin-R, Eli Lily, Indianapolis, IN) diluted in isotonic saline solution (3.0 mL) was injected intravenously, and the catheter was flushed. Blood samples were collected as described above.

Trial 1. Spontaneous Blood Glucose Variability

The objective of this trial was to determine the spontaneous fluctuations of blood glucose concentrations in horses receiving no treatment within the experimental period of 150 min. Serial blood samples were collected from the 6 horses according to the described testing protocol, and blood glucose concentrations were determined.

Trial 2. Predexamethasone CGIT Profile.

The objective of this trial was to determine a control for the blood glucose profile when horses were subjected to the CGIT. For this, 6 horses were administered a CGIT as described above.

Trial 3. Effect of Dexamethasone on the CGIT Profile at 2, 24, or 72 h Postinjection.

The objective of this trial was to determine a time-response relationship of glucose and insulin profiles as affected by dexamethasone administration. For this trial, the 6 horses were injected intravenously with 40 μg of dexamethasone/kg of BW on 3 different occasions, with at least 1 wk between each test. A CGIT was performed at 24 h postdexamethasone for one test, then on a subsequent day at 2 h postdexamethasone, and on the last test day at 72 h postdexamethasone.

Statistical Analysis

Data were tested for normality by using the Shapiro-Wilk statistic. The effect of the 5 treatments on blood glucose was analyzed by using a mixed model with repeated measures (SAS Inst. Inc., Cary, NC), with horse within treatment as the repeated term and sources of variation including treatment, time of sampling (13 sample times per test), and the residual error of horse × treatment × time. The horse within treatment term was used as the error term to test the effects of treatment, and a Tukey-Kramer adjustment was used for comparison of least squares means.


Trial 1. Spontaneous Blood Glucose Variability

No significant variation was detected for spontaneous concentrations of blood glucose over a 150-min period (Figure 1). Throughout the experiment, overall mean (±SD, n = 6) concentration of blood glucose was 77 ± 3.2 mg/dL (range = 61 to 90 mg/dL). Within-horse CV varied between 5.9 and 13% (mean = 9.8%). Interassay CV varied between 5.6 and 2.1%. Sensitivity of the test was 20 mg/dL. No blood glucose concentration pattern was observed regardless of the availability of hay. This demonstrated that there was no effect of time or hay consumption for the duration of the experiments.

Trial 2. Predexamethasone CGIT Profile

Blood glucose profile in horses undergoing CGIT (Figure 1 and Table 1) was characterized by a positive phase (above the baseline) and a negative phase (below the baseline). Initial mean glucose concentration (time 0) was 89 ± 9 mg/dL. Mean peak value recorded at 5 min was 195 ± 18 mg/dL (P < 0.05); thereafter, glucose concentrations declined at the rate of approximately 0.83 mg/dL per min. The positive phase lasted approximately 26 ± 7.0 min. The positive phase continued without interruption into a negative phase. The negative phase reached a nadir of 40 ± 12 mg/dL (approximately 47% of the baseline value; P < 0.05) at 75 min, and then began an ascending trend at approximately 0.36 mg/dL per min to regain approximately 82% of the baseline value by 150 min. This portion of the experiment provided a standard for comparison with the other tests.

Trial 3. Effect of Dexamethasone on the CGIT Profile at 2, 24, or 72 h Postinjection

Dexamethasone treatment affected the CGIT profile in different ways depending on the time that elapsed between dexamethasone treatment and the CGIT (Figure 1 and Table 1).

Effect of Dexamethasone at 2 h.

At 2 h postdexamethasone, the positive and negative test phases were remarkably similar to predexamethasone up to 60 min. At 60 min, the negative phase mean values began a rapidly ascending trend to regain 100% of baseline values by 150 min, compared with 82% in predexamethasone.

Effect of Dexamethasone at 24 h.

At 24 h, the positive phase lasted 2.9-fold longer (75 ± 8.5 vs. 26 ± 7.0 min; P < 0.05) compared with predexamethasone. By 75 min, blood glucose concentrations returned to the baseline value and were sustained at the baseline value for up to 150 min. This profile reflects abnormal glucose dynamics and is consistent with insulin resistance.

Effect of Dexamethasone 72 h.

At 72 h postdexamethasone, the positive phase was similar to predexamethasone. The negative phase in the 72-h treatment demonstrated a deeper nadir than predexamethasone at 60 min (P = 0.015). After CV and least and greatest values (shown in parentheses) for each group were: no treatment, 9% (3 to 13); CGIT predexamethasone, 26% (9 to 39); 2 h, 22% (9 to 39); 24 h, 18% (10 to 25); and 72 h, 33% (12 to 49).

These results indicate that a single dose of dexamethasone affected glucose homeostasis moderately as early as 2 h, with a drastic effect (P < 0.05) occurring at 24 h postdexamethasone. From a clinical viewpoint, there is skepticism concerning whether a single dose of dexamethasone can trigger an episode of laminitis, because administration of glucocorticoids to horses (Johnson et al., 2004) and ponies (Cunningham et. al., 1996) rarely results in laminitis. The fact that a single dexamethasone injection affected whole-body glucose dynamics indicates that one dose of dexamethasone potentially can act as a physiological trigger for laminitis in horses at high risk. However, this study demonstrated that, although there was observable insulin resistance within 24 h, by 72 h there was a compensatory decrease in insulin resistance. Thus, the induction of laminitis in normal horses attributable to a single injection of dexamethasone is doubtful; however, time and dose variants for dexamethasone to affect hoof glucose metabolism have not been defined.

Effect of Dexamethasone on Baseline Concentrations of Insulin and Cortisol

Initial serum concentration of insulin during each testing period was considered to be the baseline concentration for that testing period (Table 2). The baseline value was increased (P < 0.05) at 24 h, but not at 2 or 72 h, compared with predexamethasone. Serum concentrations of cortisol were partially suppressed (P < 0.05) at 2 h and most were suppressed at 24 h. At 72 h, cortisol values were back to baseline and were not different from predexamethasone values.

This work demonstrated that glucose homeostasis in horses was affected by a single dose of dexamethasone as early as 2 h postinjection and effects were seen as long as 72 h. At 24 h postdexamethasone, significant insulin resistance was seen. Although dexamethasone is suspected of precipitating an episode of laminitis in horses, the decreased insulin resistance seen at 72 h may mitigate any laminitis-inducing effect of dexamethasone in normal horses. This work demonstrated in normal horses that dexamethasone induces observable insulin resistance within 24 h, but by 72 h, there is a compensatory decrease in insulin resistance.

Common diagnostic tests for diabetes mellitus in several species and glucose dynamics in equids include fasting glycemia, oral and intravenous glucose tolerance tests (Jeffcott et al., 1986), a frequent-sampling technique and minimal model analysis (Hoffman et al., 2003), insulin tolerance tests (Forhead and Dobson, 1997), and hyperinulinemic-euglycemic clamp tests (DeFronzo et al., 1979; Powell et al., 2002). Of these, the minimal model, frequent-sampling technique and the hyperinulinemic-euglycemic clamp tests seem to provide the best repeatability (Pratt et al., 2005) and have become the standard for horse research in this area. The advantages of these techniques are that they allow specific calculation of glucose effectiveness and insulin sensitivity. The limiting factors are that both techniques are technically demanding: one requires frequent sampling and special software, whereas the other requires proficiency in the simultaneous use of infusion pumps. To circumvent these limitations, predictor equations for the use of proxies and reference quintiles have been used to evaluate snapshot blood samples in horses for screening of insulin sensitivity (Treiber et al., 2005). Although these proxies may be useful in large populations, they do not provide specific information regarding the ability of horses to manage glucose. The CGIT provides a tool that may be used rather readily in the field, is easy for one person to perform, and presents results immediately. It provides more information than a single blood sample without requiring specific technical skills and special software. In the CGIT, glucose administration is followed by insulin administration, which mimics the natural physiological response of endogenous insulin. Thus, an added advantage of the CGIT is that it protects potentially diabetic animals against the negative effects of a glucose overload, because a decreased glucose dose is sufficient, and the insulin added earlier moderates the glucose load. Although it remains to be seen whether modeling techniques may be applied to the CGIT to estimate insulin sensitivity effectively in addition to its applications to glucose, the test currently provides a useful application for veterinarians and equine scientists who may not have the opportunity to apply the more rigid minimal model or insulin-glucose clamp techniques in the field.

View Full Table | Close Full ViewTable 1.

Characterization of combined glucose-insulin test phases before and at 2, 24, and 72 h after dexamethasone injection in horses1

Glycemia, mg/dL 89 ± 9 94 ± 22 106 ± 14 81 ± 10
Positive phase3 duration, min* 25 25 150 25
5-min peak glycemia, mg/dL (observed concentrations) 195 ± 18 189 ± 40 209 ± 34 187 ± 32
Glucose clearance rate, mean mg/dL per min 4.2 ± 1.0 4.7 ± 1.3 0.71 ± 0.1 5.3 ± 1.7
Negative phase4 duration, min* 125 120 Absent 135
Start to nadir interval, min* 35 35 Absent 35
Glucose clearance rate, mean mg/dL per min 1.4 ± 0.3 1.6 ± 0.3 Absent 1.7 ± 0.4
Nadir, mg/dL 40 ± 12 39 ± 9 Absent 23 ± 8

View Full Table | Close Full ViewTable 2.

Blood concentrations of insulin and cortisol before and at 2, 24, and 72 h after dexamethasone injection1

Insulin, μU/mL 8.6 ± 4.1a 4.6 ± 0.92a 30 ± 15.3b 6.8 ± 0.90a
Cortisol, ng/mL 67 ± 27.2a 23 ± 11.4b 2.8 ± 1.0b 49 ± 11.0a
Figure 1.
Figure 1.

Changes in blood glucose concentrations (mean ± SD, n = 6) in untreated horses (No Tx) and in response to a combined glucose-insulin test (CGIT) before dexamethasone administration (pre-dex) and at 2, 24, and 72 h after dexamethasone injection. Asterisks indicate differences (P < 0.001) between CGIT predexamethasone, No TX, and dexamethasone at 2 h (75 to 150 min interval), 24 h, and 72 h (25- to 75-min interval). For clarity, average SD are shown in each group with either the full or one-half of the dispersion bar. Notice a positive and a negative phase in the CGIT predexamethasone profile; the prolonged positive phase and the missing negative phase in the 24-h group; the early return of the negative phase to baseline in the 2-h group; and the exaggerated negative phase in the 72-h group.