QUICKI Is Not a Useful and Accurate Index of Insulin Sensitivity following Exercise Training

Glen E. Duncan, Alan D. Hutson and Peter W. Stacpoole

Departments of Medicine (Division of Endocrinology and Metabolism) (G.E.D., P.W.S.), Biostatistics (A.D.H.), and Biochemistry and Molecular Biology (P.W.S.), University of Florida, Gainesville, Florida 32610

Address requests for reprints to: Glen E. Duncan, Ph.D., RCEPSM, Box 100226 JHMHSC, University of Florida, Gainesville, Florida 32610-0226. E-mail: gduncan{at}ufl.edu

To the editor:

We appreciate the comments by Dr. Quon regarding our study (1). The major criticism is based on a perceived problem with the interpretation of the data due to "not having a reference standard for insulin sensitivity to compare with Quantitative Insulin Sensitivity Check Index (QUICKI) and SIMM." Although our use of an unmodified frequently sampled iv glucose tolerance test and a reduced sampling schedule might be, as Dr. Quon states, "suboptimal for minimal model analysis," any error introduced by this technique would be constant, and, therefore, it is unbiased with respect to determining changes in SIMM and QUICKI that are induced by exercise training. With this in mind, our study showed that changes in SIMM were not correlated with changes in QUICKI (r = 0.24, P = 0.39) (1).

Dr. Quon remarks that "when compared with a direct measure of insulin sensitivity such as that derived from the glucose clamp, the correlation of QUICKI with SIClamp is significantly stronger than the correlation between SIMM and SIClamp." However, our inspection of the data (2) on which this statement is based yields a different interpretation. In this study, Katz et al. (2) reported that the overall correlation between QUICKI and SIClamp is r = 0.78, whereas the overall correlation between SIMM and SIClamp is r = 0.57, when groups of nonobese, obese, and diabetic subjects are combined. The subgroup correlations between QUICKI and SIClamp are r = 0.49 for nonobese, r = 0.89 for obese, and r = 0.70 for diabetics. The subgroup correlations between SIMM and SIClamp are essentially identical for the nonobese (r = 0.48) and obese (r = 0.82) groups, however, the low correlation between SIMM and SIClamp for diabetics (r = 0.51) attenuates the overall correlation between these measures. This is because the insulin response to iv glucose is diminished in diabetics, which mitigates the use of SIMM for this group. Indeed, 7 of the 15 diabetic subjects studied by Katz et al. (2) were eliminated from the analysis because of nonsensical SIMM values. Thus, the significantly stronger correlation of QUICKI and SIClamp is due to the low correlation of SIMM and SIClamp in the diabetic group. Our study did not suffer from this flaw because all of our subjects were nondiabetic and had an adequate insulin response to iv glucose.

Dr. Quon cites six reports (3, 4, 5, 6, 7, 8) of excellent correlations between QUICKI and reference glucose clamp measurements in diverse populations. Unfortunately, all of these studies are cross-sectional and cannot be used to argue for the use of QUICKI in examining changes in insulin sensitivity over time. Of the two prospective studies cited, one (9) is listed as "in press" and, thus, we are not able to comment on its findings. The abstract by Freemark and Bursey (10) indicates that insulin sensitivity, assessed by the ratio of fasting insulin to glucose concentrations, QUICKI, and HOMA indices, increased slightly, whereas SIMM did not, following metformin treatment of obese adolescents who had fasting hyperinsulinemia. Although QUICKI was able to discern significant improvements in insulin sensitivity after metformin therapy, this study also failed to include a reference standard for insulin sensitivity (i.e. glucose clamp) to compare with QUICKI and SIMM, which is the major criticism posed by Dr. Quon with respect to our paper. Thus, none of these citations lends support to the notion that QUICKI is useful for determining changes in insulin sensitivity induced by some metabolic perturbation (i.e. exercise training or pharmacoptherapy).

The major problem with using QUICKI to assess changes in insulin sensitivity is that it relies on static measures of fasting insulin and glucose and is, thus, not able to discern changes in dynamic glucose and insulin kinetics that are assessed by the glucose clamp or SIMM. This statement is supported by several independent investigations. For example, Houmard et al. (11) showed that 14 wk of exercise training significantly increased SIMM (2.1 ± 0.5 to 3.4 ± 0.7 x 105 min/pM, P < 0.05) but did not change either fasting glucose or insulin. Similarly, SIMM increased after training in groups of young and old men and women, whereas fasting glucose and insulin did not change in any group with training (12). Using a continuous iv infusion of epinephrine, propranolol, glucose, and insulin, Lampman et al. (13) found that 9 wk of training increased insulin-mediated glucose uptake, whereas fasting plasma glucose and insulin levels did not change. Furthermore, a recent report (14) demonstrated that QUICKI was unable to detect differences in insulin sensitivity between groups of nonobese individuals with or without a family history for type 2 diabetes, compared with measures by the standard glucose clamp. The authors concluded that QUICKI is less powerful when applied to nonobese subjects because the fasting glucose and insulin levels are both within narrow limits, making it difficult for QUICKI to span with accuracy the wide spectrum of insulin sensitivity characterizing normal individuals (14). Together, these studies demonstrate that a simple measure of fasting insulin and glucose applied in a mathematical formula (i.e. QUICKI) does not adequately reflect the dynamic state across a wide spectrum of insulin sensitivity values.

In summary, based on evidence that 1) changes in SIMM and QUICKI due to exercise are not correlated, using two different methods with constant error (1), 2) QUICKI is not a robust measure of insulin sensitivity across a wide spectrum of insulin sensitivity values (2, 14), and 3) discordance exists between changes in SIMM and changes in fasting glucose and insulin levels following exercise training (11, 12, 13), we believe it is correct to conclude that QUICKI does not accurately reflect changes in insulin sensitivity with exercise. Despite the use of QUICKI as an appropriate index of insulin sensitivity in a variety of static situations, it does not seem plausible that QUICKI may also be useful for assessing changes in insulin sensitivity due to a metabolic perturbation, such as exercise.

Received November 1, 2001.

References

  1. Duncan GE, Hutson AD, Stacpoole PW 2001 QUICKI does not accurately reflect changes in insulin sensitivity with exercise training. J Clin Endocrinol Metab 86:4115–4119[Abstract/Free Full Text]
  2. Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, Quon MJ 2000 Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab 85:2402–2410[Abstract/Free Full Text]
  3. Bastard JP, Robert JJ, Jardel C, Bruckert E, Grimaldi A, Hainque B 2001 Is quantitative insulin sensitivity check index a fair insulin sensitivity index in humans? Diabetes Metab 27:69–70[Medline]
  4. Rabasa-Lhoret R, Laville M 2001 How to measure insulin sensitivity in clinical practice? Diabetes Metab 27:201–208[Medline]
  5. Kirwin JP, Huston-Presley L, Kalhan SC, Catalano PM 2001 Clinically useful estimates of insulin sensitivity during pregnancy: validation studies in women with normal glucose tolerance and gestational diabetes mellitus. Diabetes Care 24:1602–1607[Abstract/Free Full Text]
  6. Escobar-Morreale HF, Calvo RM, Sancho J, San Millan JL 2001 TNF-{alpha} and hyperandrogenism: a clinical, biochemical, and molecular genetic study. J Clin Endocrinol Metab 86:3761–3767[Abstract/Free Full Text]
  7. Silfen ME, Manibo AM, McMahon DJ, Levine LS, Murphy AR, Oberfield SE 2001 Comparison of simple measures of insulin sensitivity in young girls with premature adrenarche: the fasting glucose to insulin ratio may be a simple and useful measure. J Clin Endocrinol Metab 86:2863–2868[Abstract/Free Full Text]
  8. Dixon JB, Bhathal PS, O’Brien PE 2001 Nonalcoholic fatty liver disease: predictors of nonalcoholic steatohepatitis and liver fibrosis in the severely obese. Gastroenterology 121:91–100[Medline]
  9. Mather KJ, Hunt AE, Steinberg HO, Paradisi G, Hook G, Katz A, Quon MJ, Baron AD 2001 Repeatability characteristics of simple indices of insulin resistance: implications for research applications. J Clin Endocrinol Metab 86:5457–5464[Abstract/Free Full Text]
  10. Freemark M, Bursey D 2001 The effects of metformin on body mass index and glucose tolerance in obese adolescents with fasting hyperinsulinemia and a family history of type 2 diabetes. Pediatrics 107:E55
  11. Houmard JA, Shinebarger MH, Dolan PL, Leggett-Frazier N, Bruner RK, McCammon MR, Israel RG, Dohm GL 1993 Exercise training increases GLUT-4 protein concentration in previously sedentary middle-aged men. Am J Physiol 264:E896–E901
  12. Cox JH, Cortright RN, Dohm GL, Houmard JA 1999 Effect of age on response to exercise training in humans: skeletal muscle GLUT-4 and insulin sensitivity. J Appl Physiol 86:2019–2025[Abstract/Free Full Text]
  13. Lampman RM, Santinga JT, Savage PJ, Bassett DR, Hydrick CR, Flora Jr JD, Block WD 1985 Effect of exercise training on glucose tolerance, in vivo insulin sensitivity, lipid and lipoprotein concentrations in middle-aged men with mild hypertriglyceridemia. Metabolism 34:205–211[Medline]
  14. Perseghin G, Caumo A, Caloni M, Testolin G, Luzi L 2001 Incorporation of the fasting plasma FFA concentration into QUICKI improves its association with insulin sensitivity in nonobese individuals. J Clin Endocrinol Metab 86:4776–4781[Abstract/Free Full Text]