Advertisement

Use of proxies and reference quintiles obtained from minimal model analysis for determination of insulin sensitivity and pancreatic beta-cell responsiveness in horses

Kibby H. Treiber MS1, David S. Kronfeld PhD, DSc, MVSc2, Tanja M. Hess DVM, PhD3, Ray C. Boston PhD4, and Pat A. Harris VMB, PhD5
View More View Less
  • 1 Department of Animal and Poultry Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0306.
  • | 2 Department of Animal and Poultry Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0306.
  • | 3 Department of Animal and Poultry Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0306.
  • | 4 Department of Clinical Studies, New Bolton Center, Kennett Square, PA 19348.
  • | 5 Equine Studies Group, WALTHAM Centre for Pet Nutrition, Melton Mowbray, Leicestershire, LE14 4RT, UK.

Abstract

Objective—To develop proxies calculated from basal plasma glucose and insulin concentrations that predict insulin sensitivity (SI; L·min–1·mU–1) and beta-cell responsiveness (ie, acute insulin response to glucose [AIRg]; mU/L·min–1) and to determine reference quintiles for these and minimal model variables.

Animals—1 laminitic pony and 46 healthy horses.

Procedure—Basal plasma glucose (mg/dL) and insulin (mU/L) concentrations were determined from blood samples obtained between 8:00 AM and 9:00 AM. Minimal model results for 46 horses were compared by equivalence testing with proxies for screening SI and pancreatic beta-cell responsiveness in humans and with 2 new proxies for screening in horses (ie, reciprocal of the square root of insulin [RISQI] and modified insulin-to-glucose ratio [MIRG]).

Results—Best predictors of SI and AIRg were RISQI (r = 0.77) and MIRG (r = 0.75) as follows: SI = 7.93(RISQI) – 1.03 and AIRg = 70.1(MIRG) – 13.8, where RISQI equals plasma insulin concentration–0.5 and MIRG equals [800 – 0.30(plasma insulin concentration – 50)2]/(plasma glucose concentration – 30). Total predictive powers were 78% and 80% for RISQI and MIRG, respectively. Reference ranges and quintiles for a population of healthy horses were calculated nonparametrically.

Conclusions and Clinical Relevance—Proxies for screening SI and pancreatic beta-cell responsiveness in horses from this study compared favorably with proxies used effectively for humans. Combined use of RISQI and MIRG will enable differentiation between compensated and uncompensated insulin resistance. The sample size of our study allowed for determination of sound reference range values and quintiles for healthy horses. (Am J Vet Res 2005;66:2114–2121)

Abstract

Objective—To develop proxies calculated from basal plasma glucose and insulin concentrations that predict insulin sensitivity (SI; L·min–1·mU–1) and beta-cell responsiveness (ie, acute insulin response to glucose [AIRg]; mU/L·min–1) and to determine reference quintiles for these and minimal model variables.

Animals—1 laminitic pony and 46 healthy horses.

Procedure—Basal plasma glucose (mg/dL) and insulin (mU/L) concentrations were determined from blood samples obtained between 8:00 AM and 9:00 AM. Minimal model results for 46 horses were compared by equivalence testing with proxies for screening SI and pancreatic beta-cell responsiveness in humans and with 2 new proxies for screening in horses (ie, reciprocal of the square root of insulin [RISQI] and modified insulin-to-glucose ratio [MIRG]).

Results—Best predictors of SI and AIRg were RISQI (r = 0.77) and MIRG (r = 0.75) as follows: SI = 7.93(RISQI) – 1.03 and AIRg = 70.1(MIRG) – 13.8, where RISQI equals plasma insulin concentration–0.5 and MIRG equals [800 – 0.30(plasma insulin concentration – 50)2]/(plasma glucose concentration – 30). Total predictive powers were 78% and 80% for RISQI and MIRG, respectively. Reference ranges and quintiles for a population of healthy horses were calculated nonparametrically.

Conclusions and Clinical Relevance—Proxies for screening SI and pancreatic beta-cell responsiveness in horses from this study compared favorably with proxies used effectively for humans. Combined use of RISQI and MIRG will enable differentiation between compensated and uncompensated insulin resistance. The sample size of our study allowed for determination of sound reference range values and quintiles for healthy horses. (Am J Vet Res 2005;66:2114–2121)