Study was performed at the College of Veterinary Medicine, University of Missouri, Columbia, MO.
Clinical Assessment of Blood Glucose Homeostasis in Horses: Comparison of a Continuous Glucose Monitoring System with a Combined Intravenous Glucose and Insulin Test Protocol
Article first published online: 8 DEC 2010
Copyright © 2010 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 25, Issue 1, pages 162–165, January/February 2011
How to Cite
Johnson, P.J., Wiedmeyer, C.E., LaCarrubba, A., Messer, N.T., Dingfelder, H.A., Cogswell, A.M., Amorim, J.R.R. and Ganjam, V.K. (2011), Clinical Assessment of Blood Glucose Homeostasis in Horses: Comparison of a Continuous Glucose Monitoring System with a Combined Intravenous Glucose and Insulin Test Protocol. Journal of Veterinary Internal Medicine, 25: 162–165. doi: 10.1111/j.1939-1676.2010.0643.x
- Issue published online: 11 JAN 2011
- Article first published online: 8 DEC 2010
- Submitted February 19, 2010; Revised August 11, 2010; Accepted October 7, 2010.
- Combined glucose insulin test;
- Continuous interstitial glucose monitor;
- Insulin resistance;
Background: The combined glucose-insulin test (CGIT) is helpful for evaluating insulin sensitivity. A continuous glucose monitoring system (CGMS) reports changes in interstitial glucose concentrations as they occur in the blood. Use of the CGMS minimizes animal contact and may be useful when performing a CGIT.
Hypothesis: Results obtained using a CGMS are useful for the evaluation of glucose responses during the evaluation of insulin sensitivity in equids.
Animals: Seven mature, obese ponies.
Methods: Ponies were equipped with CGMS for determination of interstitial glucose concentrations. Glucose (150 mg/kg, IV) and insulin (0.1 U/kg, IV) were administered and blood glucose concentrations determined at (minutes after time zero) 1, 5, 15, 25, 35, 45, 60, 75, 90, 105, and 120 with a hand-held glucometer. Blood chemistry results were compared with simultaneously obtained results using CGMS.
Results: Concordance coefficients determined for comparison of blood glucose concentrations determined by a hand-held glucometer and those determined by CGMS after the zero time point were 0.623, 0.764, 0.834, 0.854, and 0.818 (for delays of 0, 5, 10, 15, and 20 minutes, respectively).
Conclusions and Clinical Importance: Interstitial glucose concentrations obtained by the CGMS compared favorably to blood glucose concentrations. CGMS may be useful for assessment of glucose dynamics in the CGIT.
combined glucose-insulin test
continuous glucose monitoring system
pituitary pars intermedia dysfunction
Chronic conditions of horses that have been associated with decreased insulin sensitivity (insulin resistance [IR]) include laminitis,1 pituitary pars intermedia dysfunction (PPID),2 type 2 diabetes mellitus,3 hyperlipemia syndrome (hepatic lipidosis),4 and osteochondrosis.5 The combined glucose-insulin test (CGIT) represents a practical method for assessment of blood glucose homeostasis that incorporates the administration of both glucose and insulin and thus affords better and more integrated information than the administration of either glucose or insulin alone.6 Protocols that minimize patient handling and multiple blood sample acquisitions may improve the diagnostic accuracy of the CGIT because stress can contribute to decreased insulin sensitivity.6 Continuous glucose monitoring systems (CGMS), developed for monitoring glucose regulation in diabetic humans, have been employed in veterinary medicine.7 With CGMS, interstitial fluid glucose concentrations are assessed every 5 minutes (288 measurements/24 h), decreasing the need for patient restraint. The instrumentation remains in place for several days, enabling performance of CGIT with only slight disturbance to the animal. Performance of CGMS compared with serial blood sampling with a hand-held glucometer during CGIT has not been reported. We hypothesized that CGMS results would be similar to results of a glucometer for evaluating changes in glucose concentration during the CGIT.
Materials and Methods
Seven female ponies with a median age of 16 years (range, 13–20 years) were used. The body condition of all ponies was obese. Five ponies were affected with chronic laminitis and PPID was identified in 5 of the ponies at necropsy. Ponies were accommodated in stalls in the Veterinary Medical Teaching Hospital throughout the experimental period. The study protocol was approved by the IACUC at the University of Missouri.
Continuous Glucose Monitoring
The CGMSa was attached to each horse as described previously (Fig 1). Briefly, an approximately 3 × 3 cm patch of skin was clipped on the lateral neck region. The skin was wiped of excess hair using an alcohol soaked gauge and allowed to air dry. The flexible CGMS probe sensor was introduced into the skin using the attached 18-G needle stylet and secured to the skin with cyanoacrylate adhesive glue. Once the sensor was in place, the cord from the recording device was attached to the probe and further secured to the skin using adhesive. The recording device then was secured to the animal by taping it to the halter. The CGMS was initialized and calibrated according to the manufacturer's recommendations. Briefly, after placement of sensor, the CGMS was allowed a 1-hour initialization period followed by 3 blood glucose concentrations entered into the instrument per day for calibration. The CGMS remained attached to each horse and recorded interstitial glucose concentrations approximately 18–20 hours before the CGIT was performed. The CGMS recorded the interstitial glucose concentration every 5 minutes for a total of 288 recordings per 24 hours. After placement of the CGMS sensor, each animal was observed carefully for the development of abnormal behavior resulting from the equipment. Special attention was paid at the site of attachment for the development of any adverse skin reactions.
Combined Intravenous Glucose and Insulin Test Protocol (CGIT)
The CGIT that was used has been described previously.6 Briefly, after a 4-day period of hospitalized acclimation, during which time the ponies were fed an ad libitum grass hay ration and not subjected to any stressful handling procedures, an IV catheter was placed in the left jugular vein. On the test day (day after catheter placement), blood was collected for determination of the glucose concentration (time zero). Each pony then was immediately treated with glucoseb (150 mg/kg of body weight, administered quickly, IV) and insulinc (0.1 U/kg, IV). Blood samples were collected via the catheter at the following time intervals (minutes after time zero) for determination of the blood glucose concentration: 1, 5, 15, 25, 35, 45, 60, 75, 90, 105, and 120. Ponies were protected from stressful activities and allowed to eat hay and drink throughout the course of each experimental protocol.
Blood Glucose/CGMS Comparison
Glucometer selectiond was based on results of a previous study (that included coefficients of variation), in which the selected glucometer was demonstrated to yield blood glucose concentration data that were consistent with results obtained by use of an automated laboratory analyzer (gold standard) during the CGIT testing protocol.6 Results of the glucometerd method were obtained at time zero and then at 5, 15, 25, 35, 45, 60, 75, 90, 105, and 120 minutes. For the CGMS method, results were recorded every 5 minutes.
Analysis of Data
The concordance coefficient was used to address the question of agreement between 2 measurement methods.8 The concordance coefficient is more appropriate than a simple correlation coefficient because the latter is a measure of the strength of the linear relationship between the 2 measures but does not address whether they actually agree (ie, yield the same response). Data were evaluated for concordance with no time lag between the 2 methods (ie, readings taken at the same time) and for time lags of +5, +10, and +15 minutes. A regression model was fitted to the data at the optimal time lag (15 minutes) to determine if the fitted line had a slope or intercept that was significantly different from the ideal identity line. The mixed procedure in the statistical software SAS v9e was used because it allowed fitting a model with repeated observations on the same subject as well as heterogeneous variances.
Application of the sensor and the transmitter did not cause any adverse behavioral or skin reactions. A total of 77 blood glucose concentrations were compared with interstitial glucose concentrations. The 1-minute time point, obtained for the CGIT, was not used for comparison of the 2 methods because the CGMS records interstitial glucose every 5 minutes; therefore there was not a comparable result from the CGMS.
Administration of the CGIT protocol resulted in measurable changes in both the blood glucose and the interstitial glucose concentrations. Both detection methods indicated a rapid rise in glucose concentrations as the ponies were given glucose and insulin followed by progressive decreases in glucose concentrations over time. Both the interstitial glucose curve obtained from the CGMS and the curve generated from blood glucose concentrations showed similar results (Fig 2).
Concordance coefficients determined for comparison of results of glucose concentration measured by hand-held glucometer and those measured by CGMS after the zero time point were 0.623, 0.764, 0.834, 0.854, and 0.818 (for lags of 0, 5, 10, 15, and 20 minutes, respectively). The highest concordance coefficient was obtained when a 15-minute time lag was applied to CGMS data and compared with glucometer data.
As a measure of disagreement between readings, we evaluated the absolute difference between glucose concentrations determined by hand-held glucometer and those determined by CGMS. The smallest (best) mean absolute difference was identified for the lag time of 15 minutes. By fitting a regression line to the CGMS (+15 minute lag) results (with glucometer-determined glucose results as the dependent variable) by the procedure as defined above, the intercept did not differ significantly from zero (P= .1380). Moreover, the slope did not differ significantly from 1 (P= .49). The identity reference line had a 0 intercept and a slope of 1.
Results of this study demonstrated that glucose curves obtained using CGMS had high concordance with the concentration of glucose in blood. Optimal concordance was evident when CGMS data were compared with blood glucose concentrations with a 15-minute delay, suggesting that 15 minutes are required for CGMS detection of dynamic changes in the blood glucose concentration.
These results compare favorably with what previously has been reported in other species, and suggest that changes in blood glucose concentration are rapidly detected in the interstitial fluid.7,9 The mean delay between a change in blood glucose concentration versus interstitial fluid glucose concentration has been reported to be approximately 5, 10, and 11.4 minutes in humans,9 dogs,10 and cats.11 All of the ponies in this study exhibited suboptimal insulin sensitivity, suggesting that blood-to-interstitial fluid equilibration may have been relatively slow, as has been reported for older, diabetic cats.11 Future studies comparing CGMS assessment of glucose curves between insulin sensitive and IR-individuals should be considered. Conclusions regarding the usefulness of the CGMS for the detection of IR cannot be drawn from these results because we did not concomitantly evaluate insulin-sensitive controls.
Although we employed a hand-held glucometer device that previously had been shown to yield results comparable to those obtained in a clinical pathology laboratory, we did not directly compare our results to those from the laboratory. We were principally interested in determining whether CGMS technology would afford sufficient sensitivity to detect rapid changes in interstitial glucose concentration such as those expected to occur during the CGIT. Hand-held analyzers also are subject to greater variability, and greater agreement between the methods may have been identified if glucose concentrations had been compared with those obtained by a chemistry analyzer. The CGMS technology fails to detect concentrations of glucose >400 mg/dL, suggesting that it would not be practical for evaluation of dynamic glucose curves in diabetic horses.3
Our results suggest that a CGIT could be performed to complement a longer term evaluation of interstitial glucose concentration (up to 72 hours). During a 72-hour period of interstitial glucose evaluation, the CGIT could be incorporated at such times so as to obviate the need for human contact (aside from glucose and insulin administration). The use of CGMS decreases the need for patient handling during the CGIT because blood sampling for purposes of CGMS calibration can be done at times independent of the CGIT.
Diagnostic protocols that simply evaluate glucose responses may indicate normal CGMS and CGIT results in insulin-resistant patients because horses are adept at increasing pancreatic insulin secretion in response to decreased peripheral insulin sensitivity. Currently, CGIT testing for purposes of demonstrating IR also requires that serum insulin concentration should be measured at 0 and +45 minutes. Detection of higher insulin concentrations during the CGIT indicates increased compensation (presumed to be IR).12
Results of this study suggest that the evaluation of interstitial glucose concentrations as measured by the CGMS could be used to evaluate changes in blood glucose concentrations when the CGIT is performed. Although the use of CGMS decreases the need for blood sampling and the stress of animal handling during the CGIT, it does not completely eliminate either disturbance. The limitations associated with the use of CGMS must be considered and further investigations are warranted.
aMedtronic Minimed, Northridge, CA
bDextrose 50% Solution, Vedco, St Joseph, MO
cHumulin-R, Eli Lilly & Co, Indianapolis, IN
dMedisense Precision QID, Medisense Inc, Waltham, MA
eSAS Institute Inc, Cary, NC
This study was made possible by funds provided by the Animal Health Foundation of St Louis, Missouri. Dick Madsen in the Biostatistics Office of Medical Research at the University of Missouri's School of Medicine performed the statistical analysis. Howard Wilson helped with preparation of the graphs.
- 5Hyperglycemia/hyperinsulinemia after feeding a meal of grain to young horses with osteochondritis dissecans (OCD) lesions. Pferdeheilkunde 1996;12:320–322.