| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Unit on Growth and Obesity, Developmental Endocrinology Branch, National Institutes of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892-1862
Address all correspondence and requests for reprints to: Gabriel I. Uwaifo, M.D., Unit on Growth and Obesity, National Institutes of Health, 10 Center Drive, Building 10, Room 10N262, MSC 1862, Bethesda, Maryland 20892-1862. E-mail: . Uwaifog{at}mail.nih.gov
Abstract
The euglycemic clamp is the gold standard for estimating insulin sensitivity. The hyperglycemic clamp is easier to perform and is the gold standard for estimating ß-cell secretion. Reports in adults suggest that hyperglycemic clamps can estimate insulin sensitivity with results equivalent to euglycemic clamps. We investigated whether insulin sensitivity measures from both clamps are equivalent in children.
Thirty-one lean and obese children (mean body mass index, 25.1 ± 4.9 kg/m2; mean age, 8.7 ± 1.4 yr; 15 girls and 16 boys; 12 black and 19 white) were studied. All subjects underwent hyperglycemic clamps, then euglycemic clamps 2–6 wk later. Body composition was estimated by dual energy x-ray absorptiometry. Visceral and sc abdominal fat was estimated by abdominal magnetic resonance imaging.
Whole-body glucose disposal and insulin sensitivity (SI clamp) derived from both clamps and normalized for total or visceral fat and lean mass were significantly correlated (r, 0.45–0.65; P < 0.05). However, absolute SI clamp values were not equivalent. Bland-Altman comparisons found that SI clamp estimates from hyperglycemic clamps became less precise as SI clamp increased. There were significant correlations between indices of ß-cell secretion from the hyperglycemic clamp and mean C-peptide values from the euglycemic clamp (P < 0.05). However, no correlation was found between measures of total insulin clearance (derived from the euglycemic clamp) and surrogates of hepatic insulin clearance (derived from the hyperglycemic clamp).
In this cohort of diverse children, SI clamp values from euglycemic and hyperglycemic clamps were significantly correlated but were not equivalent, whereas the insulin clearance measures were not correlated. It cannot be assumed that the hyperglycemic clamp obviates the need for euglycemic clamp studies to accurately estimate insulin sensitivity in children.
DEFECTS IN INSULIN sensitivity and secretion are commonly found in obese individuals, and such defects presage the development of type 2 diabetes (1, 2) and other components of the metabolic syndrome X (3, 4, 5). The importance of insulin resistance and insulin secretion in these disorders has made their accurate quantification vital.
The glucose clamp technique first described by Andres and colleagues is the accepted gold standard for estimating insulin sensitivity (6, 7, 8, 9). Likewise, the hyperglycemic clamp is the gold standard for estimating pancreatic ß-cell secretory capacity (9). The euglycemic clamp method is a laboratory investigation that is labor intensive and expensive to do, involves fairly sophisticated equipment, and requires close monitoring by highly trained personnel for the duration of the test (10). As a result, there has been an ongoing search for indices derived from fasting blood (7, 11, 12) and for dynamic tests of insulin sensitivity that combine adequate validity and accuracy at reasonable cost and ease of execution (6, 7, 11, 13, 14, 15). Although generally easier to perform and less expensive, the insulin sensitivity measures thus derived are considered less accurate, particularly in the setting of obesity and/or type 2 diabetes mellitus (6, 10, 12, 13, 14).
The hyperglycemic clamp technique, in which a hyperglycemia-inducing bolus of dextrose is followed by dextrose infusion to maintain hyperglycemia, is another approach to measuring insulin action that can be used to estimate not only insulin sensitivity, but also noninsulin-dependent glucose uptake, glucose effectiveness, ß-cell secretory capacity, and hepatic insulin clearance (8, 16, 17, 18). Several studies have suggested that the insulin sensitivity measures derived from the hyperglycemic clamp are closely correlated with those derived from euglycemic clamps in normal, insulin-resistant, obese, and diabetic adults (6, 9, 17, 19, 20). On the basis of these reports, some investigators have chosen to estimate insulin sensitivity using the hyperglycemic clamp (21, 22, 23).
There are, to our knowledge, no published data on the validation of hyperglycemic clamp studies for insulin sensitivity in children. The increasing prevalence of obesity and type 2 diabetes in children has made the quantification of insulin resistance as important in children as in adults (24, 25). Therefore, we performed hyperglycemic and euglycemic, hyperinsulinemic clamp studies in a diverse population of children to compare estimates of insulin sensitivity, pancreatic ß-cell secretory capacity, and clearance.
Subjects and Methods
Subjects
Overweight children and the nonoverweight children of two overweight parents were recruited for metabolic studies through mailed notices to 6- to 12-yr-old children in the Montgomery County and Prince Georges County school districts of Maryland as well as the Washington D.C. area, and through local physician referrals and advertisements in local newspapers. All subjects had normal history and physical examinations as well as normal blood chemistry and hepatic and thyroid function. None of the subjects had any significant medical illness, and none were taking medications known to impact on insulin sensitivity. Nonobese children of two overweight parents were recruited because they are considered to have a condition, namely a predisposition to develop obesity, that can justify studying them with procedures that may constitute a minor increment over minimal risk. The clinical protocol was approved by the National Institute of Child Health and Human Development institutional review board. Informed consent and assent were obtained from parents and children.
Clinical protocol
Subjects were studied at the Warren Grant Magnuson Clinical Center of the National Institutes of Health (NIH). Each subject had a full history and physical examination and underwent anthropometric measurements and determination of skeletal age (26), abdominal magnetic resonance imaging for abdominal fat estimation, and whole-body dual energy x-ray absorptiometry for body composition analysis. Dual energy x-ray absorptiometry was performed using the Hologic QDR2000 instrument in the pencil beam mode (software version 5.64), as previously described (27). Magnetic resonance imaging areal estimates of visceral and sc fat were determined as previously described (28). Body mass index (BMI) SD scores were computed for each subject by using the following formula: BMI SD score = (Actual BMI - Mean BMI for age, race, and gender)/BMI SD for age, race, and gender based on established standards and norms (29). Socioeconomic status was estimated using the Hollingshead four factor index of social status (30). Breast development was recorded according to the stages of Tanner, whereas testicular volumes were measured according to Prader (31). Weight was measured to the nearest 0.1 kg using a calibrated digital scale (Scale-Tronix, Wheaton, IL), whereas height was measured in triplicate to the nearest 1 mm using a stadiometer calibrated before each set of measurements (Holtain Ltd., Crymych, UK). A hyperglycemic clamp study was subsequently performed, and 2–6 wk later subjects underwent a euglycemic clamp study. Hyperglycemic and euglycemic clamp studies were carried out using a modification of the methods described by Andres and colleagues (9). Each hyperglycemic clamp began at approximately 1000 h, after an overnight fast. Plasma glucose was acutely raised during a 2-min period to approximately 225 mg/dl (12.5 mmol/liter) by bolus infusion of 25% dextrose and was then maintained near a target of 200 mg/dl (11.1 mmol/liter) by continuous infusion of variable amounts of 20% dextrose for 120 min. Plasma glucose was measured every 2.5 min for the first 15 min and then every 5 min until the end of the study. Insulin was measured every 5 min for the first 15 min and then every 15 min from 15 to 120 min. Plasma FFA levels were measured during the clamp using an enzymatic colorimetric assay (Wako Laboratories, Richmond, VA). Plasma glucose was concurrently measured using a glucose analyzer (YSI, Inc., Yellow Springs, OH), calibrated to within 5% of multiple glucose standards (50, 100, 125, 150, 250, and 500 mg/dl) before each study using a Hitachi 736-30 analyzer (Roche Molecular Biochemicals, Inc., Indianapolis, IN). Plasma insulin was measured by the TOSOH two-site immunoenzymometric assay (Covance Laboratories, Inc., Vienna, VA). The cross-reactivity of the assay with both proinsulin and C-peptide is less than 1%, whereas the mean inter- and intra-assay coefficients of variation are 5.8 and 3.6%, respectively. First phase insulin was calculated as the mean of three measurements obtained during the first 15 min, second phase insulin as the mean of eight measurements from 15–120 min, and steady-state insulin (I) as the mean of the five measurements obtained from 60–120 min. Serum C-peptide was assayed during the same time points of the studies using the analyte-specific reagents immunochemiluminometric assay method (Mayo Medical Laboratories, Rochester, MN), and the first phase, second phase, and steady-state C-peptide levels were defined similarly to the corresponding insulin values. Whole-body glucose uptake was estimated as the metabolic rate (M), defined as the infusion rate of exogenous glucose administered, corrected for urinary glucose losses and the glucose space correction (9, 17, 32). As a measure of insulin sensitivity, (SI Hyper clamp), the ratio of metabolic rate to steady-state insulin (M/I) was calculated (9, 17, 32).
The euglycemic clamp studies also began at about 1000 h, after an overnight fast. An iv cannula was inserted into each arm. Blood from one of these was arterialized by placing the arm continuously in a warming blanket set at about 60 C. A continuous infusion of regular insulin (Humulin S, Eli Lilly \|[amp ]\| Co., Indianapolis, IN) was maintained throughout the 180 min duration of the test at a rate chosen to maintain plasma insulin at more than 200 µU/ml (1435 pmol/liter), so as to suppress endogenous hepatic glucose output (33). Blood glucose was maintained between 95 and 105 mg/dl (5.3–5.8 mmol/liter) using a continuous infusion of variable amounts of 20% dextrose. The infusion adjustments were made on the basis of the results of plasma glucose measurements done every 5 min using a glucose analyzer (YSI, Inc.) as previously described. Serial samples for glucose, insulin, C-peptide, and FFA were obtained every 10 min during the steady-state period (between the 120- and 180-min time points) and assayed using the previously described methods above. Whole-body glucose uptake was estimated as the metabolic rate (M), defined as the infusion rate of exogenous glucose administered, corrected for urinary glucose losses and the glucose space correction (9, 17, 19, 32). The metabolic clearance rate (MCR) for insulin was computed as the insulin infusion rate/increase in plasma insulin concentration above baseline (9). As a measure of insulin sensitivity (SIEug clamp), the ratio of metabolic rate to steady-state insulin (M/I) was calculated (9, 17, 19, 32). M, MCR, SI Eug clamp, and SI Hyper clamp measures were all normalized for fat free mass and fat mass (34).
Derived indices
The fasting, first phase, and steady-state insulin and C-peptide levels were derived from the hyperglycemic clamp study as indices of pancreatic ß-cell secretory capacity (9, 35, 36). The mean steady-state C-peptide levels were also derived from the euglycemic clamp study as a measure of pancreatic ß-cell endogenous secretion in the setting of the exogenously induced hyperinsulinemia (35, 36). The C-peptide to insulin molar ratios for the fasting, first phase, and steady-state phases of the hyperglycemic clamp study were derived as indices of fasting and dynamic hepatic insulin clearance (36, 37, 38, 39).
Statistical analysis
The derived data were analyzed using JMP IN version 3.2.1 software for Windows (1989–1997, SAS Institute, Inc., Duxbury Press, Cary, NC) and StatView version 5.0.1 for Windows (1992–1998, SAS Institute, Inc.). Standard tests of data symmetry using skewness and kurtosis were performed on all data, and normality was tested using the Shapiro-Wilkes test. Non-normal data were transformed by common log or other transformation procedures to achieve data symmetry and normality before use of parametric tests. Data that could not be normalized by transformation procedures were analyzed using equivalent nonparametric tests. Unless otherwise indicated, data are reported as mean ± SD.
Correlations between parameters were evaluated using the Spearman correlation coefficients. Comparison of equivalence of insulin sensitivity measures from the hyperglycemic and euglycemic clamp methods was performed using the Bland-Altman approach (40). This is a validated methods-comparisons test that involves plotting the difference between individual data points using the two different test methods against means of each data point using the two different test methods. Comparisons between groups of data (e.g. between obese and nonobese children or prepubertal and pubertal) were done using unpaired t tests, ANOVA, or analysis of covariance. P value less than 0.05 was considered significant for all of the data analyses.
Results
Thirty-one children (12 black and 19 white) with a mean age of 8.7 ± 1.4 yr (Table 1
) were studied. The cohort was largely comprised of obese, prepubertal children based on BMI of at least the 95th percentile for age-, gender- and race- established norms from the National Health and Nutrition Examination Survey (41). Eighty-one percent of the subjects were obese (25 obese and 6 nonobese). Among the subgroups, black children did have greater BMI SD (4.7 ± 2.5 vs. 2.4 ± 2.0; P < 0.05). All of the boys studied were prepubertal, with Tanner stage I or II pubic hair and testis volume no more than 4 cc; 6 of the 15 girls studied had evidence for gonadarche (Table 1).
|
details the parameters of insulin sensitivity, secretion, and clearance derived from the hyperglycemic and euglycemic clamp studies for the entire group. The insulin sensitivity, secretion, and clearance parameter subgroup analyses based on race and presence or absence of obesity are shown in Table 3. The black children had higher first phase and steady-state insulin as well as higher first phase C-peptide than the white children, independent of pubertal status (P < 0.05). Obese children had significantly higher pancreatic ß-cell secretory capacity reflected by higher fasting, first phase, and steady-state insulin and C-peptide levels (from the hyperglycemic clamp) (Table 3). They also had significantly lower hepatic insulin clearance (as reflected by C-peptide to insulin molar ratio) and lower glucose clearance, but comparable total insulin clearance (reflected by MCR for insulin).
|
|
; R between 0.45 and 0.65; all P < 0.05). The SI clamp and glucose clearance values from both studies were, however, not equivalent (Fig. 2, A and B). SI clamp and glucose clearance derived from the hyperglycemic clamp were both systematically overestimated compared with the values derived from the euglycemic clamp (P < 0.001) with a significant increase in the difference between the measures as the magnitude of the insulin sensitivity and clearance increased (R = -0.91; P < 0.01). These findings were corroborated by paired T tests of the individual SI and glucose clearance data (in each case the hyperglycemic clamp data were significantly higher with P < 0.05). The correlation coefficients remained significant among both black and white children as subgroups, although they were higher in black children. Apart from the insulin sensitivity data, there were also strong correlations between the insulin secretory parameters from the hyperglycemic clamp and the mean C-peptide values from the euglycemic clamp studies (Table 4). Similarly, there was no correlation between indices of hepatic insulin clearance derived from the hyperglycemic clamp (the C-peptide to insulin molar ratios) and total insulin clearance derived from the euglycemic clamp (Table 4).
|
|
|
In a diverse group of lean and obese black and white children, we found that insulin sensitivity measures derived from hyperglycemic clamps are significantly correlated with those derived from euglycemic clamps. However, the absolute values obtained are not equivalent, and the bias for the estimate increases in children who have greater insulin sensitivity. A significant correlation was also found between indices of pancreatic ß-cell secretory reserves derived from the hyperglycemic clamp and mean C-peptide values from the euglycemic clamp studies. However, there were no significant correlations between estimated hepatic insulin clearance from the hyperglycemic clamp and total whole-body insulin clearance from the euglycemic clamp.
Although many methods can be used to estimate insulin secretion and sensitivity in vivo, the glucose clamp approach, combining both euglycemic and hyperglycemic clamps, remains the gold standard (9, 11, 32, 34, 42); the euglycemic clamp method is considered the ideal means of measuring insulin sensitivity and total body insulin clearance (9, 43, 44), and the hyperglycemic clamp is considered the preferred method for measuring insulin secretory capacity (9). The euglycemic clamp methodology is, however, fairly difficult to perform and expensive, and it does have the theoretical risk of inducing hypoglycemia should iv access be lost. Because of these limitations, the euglycemic clamp methodology is used infrequently to assess insulin sensitivity and clearance in pediatric populations (19, 45). The utility of the hyperglycemic clamp for estimating insulin sensitivity has been validated in adult populations (6, 17, 19, 20), with significant correlations found between SI clamp measurements from hyperglycemic and euglycemic clamp studies for most, but not all (6), populations. The absolute values of SI clamp from euglycemic and hyperglycemic clamps have been found in some adult studies to be statistically different (6, 17). Therefore, it is perhaps not surprising that we also found that SI clamp estimates, examined on an absolute basis, were different in the two studies.
Apart from measures of insulin sensitivity, there was also a strong correlation between measures of insulin secretion based on both the insulin and C-peptide levels during the hyperglycemic clamp studies and the mean C-peptide levels during the euglycemic clamp studies (35, 39, 46). These measures of insulin secretion were, however, not equivalent because the exogenous supraphysiological insulin infusion inherent in the euglycemic clamp method suppresses endogenous ß-cell secretion of insulin and C-peptide. For this reason, whereas the hyperglycemic clamp provides accurate estimates of ß-cell secretory capacity, the euglycemic clamp does not. The C-peptide to insulin molar ratio has been widely used as a surrogate of hepatic insulin clearance/extraction (36, 38, 47) but was not found to correlate with the MCR for insulin derived from the euglycemic clamp studies in our cohort. The MCR for insulin is believed to be a more accurate estimation of total body insulin disposition (9, 43, 44, 48) and includes in its quantification the role of the kidney and other peripheral tissues in insulin clearance. This suggests that trends in hepatic insulin clearance may not reflect trends in total body insulin disposition. This finding may be important for the interpretation of the recent reports documenting significant differences in hepatic insulin clearance (estimated by the C-peptide to insulin molar ratios) between whites and blacks (49, 50, 51, 52, 53).
Insulin sensitivity, as measured by the hyperglycemic clamp, appeared to be greater than that found with the euglycemic clamp. Similar findings have been reported in adults (19). Greater apparent insulin sensitivity may be related to increased glucose effectiveness in the setting of the hyperglycemia associated with the hyperglycemic model. The relative contribution of glucose effectiveness to the total insulin sensitivity is expected to be greater in obese, insulin-resistant populations than in normal cohorts (18) but clamps such as those done in the current study, without radiolabeled tracers, are unable to estimate glucose effectiveness adequately.
Among the identified limitations to the use of the hyperglycemic clamp to estimate insulin sensitivity are: 1) subjects must have fasting blood glucose levels less than the level at which the glucose is clamped during the study; 2) the clamped glucose level should nevertheless be below the renal threshold to avoid the need for correcting glucose infusion rates for large urine glucose losses; 3) use of the hyperglycemic clamp method requires that the subject have significant insulin secretory capacity (>175 pmol/liter); and 4) because the hyperglycemic clamp uses endogenous insulin production for its estimates, the assay used for its insulin measurements must have minimal cross-reactivity with proinsulin or other insulino-mimetic compounds (17). None of these limitations pertain to our study population, because we used a specific assay for intact insulin, did not study diabetic subjects, and employed a steady-state glucose concentration of 200 mg/dl (11.1 mmol/liter), well below the renal threshold. Furthermore, we measured, and then corrected for, the minimal renal glucose losses observed.
Although our study is the first to address the question of hyperglycemic and euglycemic clamp equivalence in a childhood population, there are some potential limitations. The interval between the hyperglycemic and euglycemic clamp studies in our protocol was 2–6 wk, and the hyperglycemic clamp studies were done first in all subjects. Although the initial description of Defronzo et al. (9) had 3- to 4-wk intervals with no significant difference in obtained indices, and subsequent studies have used variable intervals between tests that exceed 1 month (17, 19, 20), it is unclear what effect if any the interval between tests would have on the reproducibility of the obtained results. In addition, our study protocol used a single, high-dose insulin infusion rate. The resulting significantly elevated ambient plasma insulin levels could conceivably reduce measured insulin clearance (by both receptor- and nonreceptor-mediated mechanisms) as well as increase measured insulin sensitivity. Others have noted that lower dose infusion protocols for euglycemic clamp studies may be more representative of insulin sensitivity in vivo (34, 42). Obesity, particularly visceral adiposity, significantly impacts insulin sensitivity and consequently insulin secretion and clearance. Despite the observed higher BMI SD among the black children, however, it is unlikely that this would result in any significant ethnic subgroup bias because both black and white children had comparable total fat and visceral fat mass. Finally, our cohort included only 31 subjects, a small number considering the many potential covariates that may affect insulin sensitivity, secretion, and clearance.
We conclude that although the insulin sensitivity and ß-cell function measures derived from hyperglycemic clamp studies are largely correlated with those derived from euglycemic clamp studies in children, the hyperglycemic clamp significantly overestimates insulin sensitivity. Caution thus needs to be exercised in assuming equivalence of insulin sensitivity measures from hyperglycemic and euglycemic clamp studies in pediatric populations. Performing the hyperglycemic clamp does not obviate the need for euglycemic clamp studies in children if actual insulin sensitivity must be known.
Acknowledgments
We acknowledge the efforts of all the nursing staff of the 9 West ward of the Warren Grant Magnuson Clinical Center, the clinical fellows who were involved in the clinical care of the subjects during their inpatient visits, and especially the children who participated in these studies.
Footnotes
This research project was funded by an intramural research grant from the Developmental Endocrinology Branch of the National Institute of Child Health and Human Development (NIH Grant HD-000641). J.A.Y. is also supported by the National Center on Minority Health and Health Disparities, NIH, and is a Commissioned Officer in the United States Public Health Service. There are no conflicts of interest to be disclosed by any of the authors.
Abbreviations: BMI, Body mass index; I, insulin; M, metabolic rate; MCR, metabolic clearance rate; SI, insulin sensitivity.
Received January 4, 2002.
Accepted March 5, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  J. Rotteveel, M. M. van Weissenbruch, J. W. R. Twisk, and H. A. Delemarre-Van de Waal Infant and Childhood Growth Patterns, Insulin Sensitivity, and Blood Pressure in Prematurely Born Young Adults Pediatrics, August 1, 2008; 122(2): 313 - 321. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J Rotteveel, M M van Weissenbruch, and H A Delemarre-Van de Waal Decreased insulin sensitivity in small for gestational age males treated with GH and preterm untreated males: a study in young adults. Eur. J. Endocrinol., June 1, 2008; 158(6): 899 - 904. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Velasquez-Mieyer, C. P. Neira, R. Nieto, and P. A. Cowan Review: Obesity and cardiometabolic syndrome in children Therapeutic Advances in Cardiovascular Disease, October 1, 2007; 1(1): 61 - 81. [Abstract] [PDF]
|
|
|
|
 |
|
 K. K. Trout, C. Homko, and N. C. Tkacs Methods of Measuring Insulin Sensitivity Biol Res Nurs, April 1, 2007; 8(4): 305 - 318. [Abstract] [PDF]
|
|
|
|
 |
|
 C. W. Yeckel, S. E. Taksali, J. Dziura, R. Weiss, T. S. Burgert, R. S. Sherwin, W. V. Tamborlane, and S. Caprio The Normal Glucose Tolerance Continuum in Obese Youth: Evidence for Impairment in {beta}-Cell Function Independent of Insulin Resistance J. Clin. Endocrinol. Metab., February 1, 2005; 90(2): 747 - 754. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Muzumdar, X. Ma, G. Atzmon, P. Vuguin, X. Yang, and N. Barzilai Decrease in Glucose-Stimulated Insulin Secretion With Aging Is Independent of Insulin Action Diabetes, February 1, 2004; 53(2): 441 - 446. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
