HOME    »    PROGRAMS/ACTIVITIES    »    Annual Thematic Program
Talk Abstract
Mathematical Modeling of Insulin Action and In Vivo Estimates of Insulin Sensitivity

Michael J. Quon, M.D., Ph.D.
Senior Investigator
Hypertension-Endocrine Branch
Bethesda, MD 20892-1754

Diabetes, obesity, and hypertension are major inter-related public health problems that are all characterized, in part, by insulin resistance (decreased sensitivity or responsiveness to metabolic actions of insulin). Therefore, it is of great interest to develop tools to quantify insulin sensitivity in vivo. The "gold standard" hyperinsulinemic euglycemic glucose clamp method is labor intensive and not well suited to large studies. A well-accepted alternative for estimating insulin sensitivity in vivo is to analyze insulin and glucose data from a frequently sampled intravenous glucose tolerance test (FSIVGTT) using Bergman*s minimal model of glucose metabolism. This is less cumbersome than the glucose clamp but still requires at least 3 hours to complete. Estimates of insulin sensitivity (SIMM) derived from minimal model analysis correlate well with measurements of insulin sensitivity using the glucose clamp technique (SIClamp). In addition, SIMM has predictive power with respect to the development of diabetes. Nevertheless, we have previously shown that minimal model analysis systematically underestimates the effect of glucose on glucose disposal and therefore overestimates SIMM (Quon et al., Diabetes 43:890-896, 1994). Furthermore, we have recently shown that this error is due to an oversimplified single-compartment representation of glucose kinetics and is dependent on the dynamics of insulin secretion (Cobelli et al., Am J Phsyiol 38:E1031-E1036, 1998). Therefore, we have developed an alternative Quantitative Insulin-sensitivity Check Index (QUICKI). After analyzing data from both glucose clamp and FSIVGTT studies, we discovered that physiological steady-state values (i.e., fasting insulin (I0) and fasting glucose (G0)) contain important information related to insulin sensitivity and thus defined QUICKI as 1/[log (I0) + log (G0)]. Correlations of QUICKI with SIClamp were as good, or better, than correlations of SIMM with SIClamp. We conclude that QUICKI is a simple, accurate, and reliable insulin sensitivity index obtained from a single fasting blood sample that may be useful for clinical research and epidemiological studies related to diabetes and other insulin resistant states.

Back to Workshop Schedule

1998-1999 Mathematics in Biology