| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Internal Medicine (A.M.C., M.A.S.) and Geriatric Research, Education, and Clinical Center (M.A.S.), Veterans Affairs Ann Arbor Healthcare System, Ann Arbor, Michigan 48105; and Department of Internal Medicine (A.M.C., M.J.S., C.J.B., A.T.G., J.B.H., M.A.S.), University of Michigan, Ann Arbor, Michigan 49109
Address all correspondence and requests for reprints to: Annette M. Chang, M.D., University of Michigan, 5570 MSRB II, 1150 West Medical Center Drive, Ann Arbor, Michigan 48109-0678. E-mail: annchang{at}umich.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: We compared measures of insulin sensitivity and ß-cell function from the frequently sampled iv glucose tolerance test (FSIGT) to HOMA models.
Design/Patients/Setting/Intervention: Two hundred fourteen young and old with normal glucose tolerance (NGT) and old with impaired glucose tolerance (IGT) participated in a retrospective analysis of FSIGT data in a university medical setting.
Main Outcome Measure: Sensitivity to insulin (SI) and acute insulin response to glucose (AIRg) from FSIGT were compared with HOMA models.
Results: SI and HOMA-IR measures identified similar patterns of increasing insulin resistance in the two older groups, compared with younger people with NGT, with the greatest degree of insulin resistance in older people with IGT (P < 0.05 vs. young and old NGT for both SI and HOMA-IR). Agreement between HOMA-IR and SI was moderate (weighted kappa = 0.51). AIRg was similar in young and old NGT but was markedly decreased in old IGT (P < 0.05 vs. young and old NGT). HOMA-ß-cell was similar in the three groups. Agreement between HOMA ß-cell and AIRg was weak (weighted kappa = 0.35).
Conclusions: HOMA-IR may detect age-related insulin resistance when comparing large populations of older people. However, dynamic testing appears to be necessary to quantitate diminished insulin secretion in older people.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The homeostasis model assessment (HOMA) is a commonly used method to assess insulin resistance and ß-cell function and requires only fasting glucose and insulin levels. The HOMA model of insulin resistance (HOMA-IR) has been found to correlate with the hyperinsulinemic glucose clamp and the minimal model measures of insulin action and ß-cell function, primarily in younger and middle-aged people with normal glucose tolerance (3, 4, 5) and smaller groups of middle-aged people with type 2 diabetes (3, 5, 6).
Previous studies have suggested that HOMA-IR may not be a good predictor of insulin resistance when compared with the euglycemic clamp in older men with IGT (7) and older men and women with diabetes (8). Evaluation of HOMA measures of ß-cell function in older people, particularly those with IGT, has not been performed. In the present study, we compared measures of insulin sensitivity and ß-cell function from the frequently sampled iv glucose tolerance test (FSIGT) minimal model to the HOMA models in a cohort of 214 young and older men and women with normal glucose tolerance (NGT) and older people with IGT.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
All protocols were approved by the University of Michigan Institutional Review Board and performed in accordance with the Declaration of Helsinki. Healthy, community-dwelling men and women were recruited by advertisement to participate in these studies of glucose metabolism and aging. After the nature of the study was explained in detail, written informed consent was obtained from all participants.
Health status and glucose tolerance were assessed by screening medical history; physical exam; blood chemistries (liver enzymes, creatinine, electrolytes, and glucose); complete blood count; electrocardiogram; and a 75-g oral glucose tolerance test (OGTT). A total of 214 younger people (age younger than 36 yr) with NGT and older people (age 60 yr or older) with NGT and IGT were enrolled. None of the participants were taking medications known to affect glucose metabolism. This was a retrospective analysis of available data from several previous studies that reported the FSIGT data for these participants (9, 10, 11, 12, 13, 14).
Study protocols
Assessments of glucose tolerance, insulin sensitivity, and ß-cell function were performed in the morning after 12-h overnight fasts. OGTT and insulin-assisted FSIGT were performed on 2 separate study days.
OGTT
Participants reported to the University of Michigan General Clinical Research Center after a 12-h overnight fast. An antecubital iv line was placed. After one baseline sample for plasma glucose and insulin levels, subjects ingested 75 g glucose. Blood samples were obtained for glucose and insulin every 30 min for 120 min. Incremental OGTT glucose and insulin area under the curve (AUC) over the 120-min time interval was estimated for each subject by using the trapezoidal rule.
FSIGT
The insulin-assisted iv glucose tolerance test (FSIGT) was performed in all volunteers as described by Bergman (15) with the addition of insulin to enhance precision of the estimates of insulin action (16). Participants consumed minimum 150 g/d carbohydrate diet for 3 d before studies. Subjects reported after an overnight (12 h) fast. An iv catheter was placed in each arm for insulin and glucose infusions and blood sampling. After three baseline samples for fasting glucose and insulin levels were drawn, 50% dextrose (0.3 g/kg) was injected iv over 30 sec followed by injection of insulin (0.02 U/kg) over 30 sec at time 20 min. Twenty-eight blood samples were collected according to a protocol schedule over the 180 min after the glucose bolus.
The sensitivity to insulin index (SI) was calculated from a least-squares fitting of the temporal pattern of glucose and insulin throughout the FSIGT using the minimal model of glucose kinetics (15). The acute insulin response to iv glucose (AIRg or first phase insulin secretion) was calculated as the mean rise in plasma insulin above baseline at 3, 4, and 5 min after iv glucose administration. The relationship between two independent measurements of insulin secretion and insulin sensitivity has been found to be hyperbolic, allowing calculation of the product of AIRg x SI, or the disposition index (DI) (17, 18, 19). DI provides an assessment of whether insulin secretion is appropriate for the level of insulin resistance (ß-cell compensation for insulin resistance or ß-cell function). This hyperbolic relationship has been validated in older people (20) and people at risk for diabetes including those with IGT and a history of gestational diabetes (21, 22).
HOMA
HOMA assessments were calculated using the average of the three baseline fasting glucose and insulin levels from FSIGT studies. The original HOMA model of insulin resistance (HOMA1-IR) and ß-cell function (HOMA1-ß-cell) were calculated as described by Matthews et al. (23):
HOMA1-IR = [fasting insulin (microunits per milliliter) x fasting glucose (millimoles)]/22.5 and
HOMA1-ß-cell = [20 x fasting insulin (microunits per milliliter)]/[fasting glucose (millimoles) – 3.5]
The updated HOMA2 (or computer) model with nonlinear solutions, which also uses paired fasting glucose and insulin values, were calculated using the computer model (HOMA Calculator version 2.2) as described by Wallace et al. (24), yielding the following assessments: HOMA2-IR and HOMA2-%ß-cell function (HOMA2%ß), where 100% is normal ß-cell function.
Assays
All blood samples were centrifuged, and serum was stored at –20 C until analysis by the Chemistry Laboratory of the Michigan Diabetes Research and Training Center. Plasma glucose levels during the FSIGT studies were measured using a hexokinase method with an interassay coefficient of variation of 3.1% (Roche Diagnostics Corp., Indianapolis, IN). Plasma insulin was quantified using a highly specific and sensitive double-antibody RIA with an interassay coefficient of variation of 3.4% and an intraassay variability of 2.5% (Linco Research, Inc., St. Charles, MO).
Statistical analysis
Data are presented as means ± SE, with the exception of subject characteristics, which are presented as means ± SD. AIRg, SI, and HOMA measures were log transformed to approximate a normal distribution. Differences between groups were assessed by ANOVA. P < 0.05 was considered statistically significant. Two approaches were used to assess agreement between FSIGT and HOMA measures. The quantitative relationship of HOMA and FSIGT measures was evaluated by Pearson correlation coefficients and multiple regression analysis. HOMA and FSIGT measures were also stratified into tertiles of insulin resistance and insulin secretion in the three groups, and weighted kappa was computed (25). The kappa statistic estimates the closeness of agreement for FSIGT and HOMA methods to classify participants based on level of insulin resistance and insulin secretion. Kappa is always less than or equal to 1. A value of 1 implies perfect agreement between the two methods. A value less than 1 implies less than perfect agreement with a value of 0.00 indicating full disagreement between the two methods.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A total of 214 healthy men and women (aged 19–88 yr) were included in this analysis. Ethnicity of the study population was largely Caucasian (191 of 214 or 89%), but also included 14 African-American, five Hispanic, two Native American, and two Asian-American individuals. As shown in Table 1
, the 214 participants were classified into young (19–35 yr) and older (60–88 yr) groups and then by glucose tolerance during OGTT (NGT or IGT). Body weight was not significantly different in the three groups. Body mass index (BMI) was higher in the two older groups, compared with young NGT, and was in the overweight, preobese range, representative of the average, age-matched BMI in the community. The younger group was also representative of the community including college students but mainly postcollege-aged younger people. Fasting, 2-h OGTT glucose levels and 2-h OGTT glucose AUC were progressively and significantly increased in the older groups. Associated with their higher glucose levels during OGTT, 2-h OGTT insulin AUC trended toward being higher in the older NGT vs. young NGT group and was significantly higher in the older IGT vs. young NGT group.
|
SI from FSIGT, HOMA1-IR, and HOMA2-IR as measures of insulin sensitivity in the three groups are displayed in Table 2. SI and HOMA1 and -2 measures identified similar patterns of increasing insulin resistance in the two older groups, compared with younger people with NGT, with the greatest degree of insulin resistance in older people with IGT.
|
AIRg and DI from FSIGT, HOMA1 ß-cell, and HOMA2%B as measures of ß-cell function in the three groups are summarized in Table 3. Insulin secretion as assessed by AIRg was similar in young and older people with NGT, although it was markedly reduced in older people with IGT, compared with both groups with NGT. Adjustment of insulin secretion for insulin sensitivity, or the DI (AIRg x SI), as an assessment of ß-cell compensation for insulin resistance was significantly lower in older compared with younger people with NGT and even lower in older people with IGT, compared with both young and old groups with NGT. In contrast, there were no statistically significant differences in HOMA1 and HOMA2 measures of ß-cell function among the three groups.
|
As displayed in Fig. 1, there was a significant relationship between log-transformed HOMA1-IR and log-transformed SI when considering the entire cohort of young and older people (r = –0.68, P < 0.0001). The correlations between log HOMA1-IR and log SI were significant within all three groups: young NGT r = –0.50, P < 0.0001; old NGT r = –0.72, P < 0.0001; old IGT r = –0.64, P < 0.0001. Similar relationships were observed in men and women in all three groups as well as with log-transformed HOMA2-IR and log-transformed SI.
|
Agreement between FSIGT and HOMA measures
The agreement in the categorization of level of insulin sensitivity when individual HOMA1-IR and SI values were stratified into tertiles of insulin sensitivity was moderate with: weighted kappa = 0.51 in the entire cohort (the two measures agreed in 125 of 214 total participants), weighted kappa = 0.37 in young NGT, weighted kappa = 0.49 in old NGT, and weighted kappa = 0.45 in old IGT. Kappa values were similar in men (kappa = 0.51) and women (kappa = 0.50). When considering tertile 1 as insulin sensitive (higher SI and lower HOMA1-IR values) and tertile 3 as insulin resistant (lower SI and higher HOMA1-IR values) per both measures, the categorization of five of 214 (2.3%) participants disagreed by two tertiles as insulin sensitive in one method and insulin resistant in the second method. However, when considering tertile 2 (midrange) levels of insulin resistance, the categorization of 42 of 71 (59%) participants disagreed.
The agreement in the categorization of ß-cell function when individual HOMA1-ß-cell and AIRg values were stratified into tertiles of ß-cell function was weak with: weighted kappa = 0.35 in the entire cohort (the two measures agreed in 115 of 214 total participants), weighted kappa = 0.10 in young NGT, weighted kappa = 0.33 in old NGT, and weighted kappa = 0.35 in old IGT. Kappa agreement was moderate in women (kappa = 0.45) and weak in men (kappa = 0.25). When considering tertile 1 as highest degree of ß-cell function (higher AIRg and HOMA1-ß-cell values) and tertile 3 as lowest level of ß-cell function (lower AIRg and HOMA1-ß-cell values) per both measures, the categorization of 24 of 214 (11%) participants disagreed by two tertiles as highest degree of ß-cell function in one method and lowest degree of ß-cell function in the second method. When considering tertile 2 (midrange) levels of ß-cell function, the categorization of 38 of 72 (53%) participants disagreed.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In this study, SI and HOMA1 and -2 measures identified similar patterns of increasing insulin resistance in the two older groups, compared with younger people with NGT, with the greatest degree of insulin resistance in older people with IGT. Log-transformed HOMA1-IR and HOMA2-IR correlated significantly with SI within each of the groups. Overall agreement between the two measures of insulin resistance was moderate.
ß-Cell function was clearly impaired in the older group with IGT as quantified by AIRg. Furthermore, DI from FSIGT as an assessment of ß-cell function was significantly different in the three groups and was able to identify decreasing ß-cell function in the older group with NGT as well as those with IGT. These findings confirm previous reports of impaired ß-cell function in human aging (1, 2, 19, 28). HOMA1 and HOMA2 measures of ß-cell function were similar in the three groups. However, the similar HOMA ß-cell measures suggest a lack of appropriate ß-cell compensation for the increasing insulin resistance identified by both SI and HOMA-IR. A limitation of the HOMA measures is the inability to calculate DI because the hyperbolic relationship between insulin secretion and insulin sensitivity requires two independent measurements (19). In addition, overall agreement between AIRg and HOMA measures of ß-cell function was weak.
The significant relationships between HOMA1-IR and SI in older people with both NGT and IGT differ from two previous studies examining HOMA1-IR in older people. Ferrara and Goldberg (7) found no significant correlation between HOMA1-IR and euglycemic clamp-derived measures of insulin sensitivity in older men with IGT. Katsuki et al. (8) also found no significant association between HOMA1-IR and insulin sensitivity per euglycemic clamp in older Japanese people with type 2 diabetes. Our findings may have differed due to the much smaller samples sizes of 24–28 older people in these previous studies, compared with 185 older people in the current study, as well as the inclusion of women in our study. However, no significant gender effect related to measures of insulin resistance was noted in this study.
Previous studies have not specifically examined use of HOMA1-ß-cell or HOMA2 measures of insulin resistance and ß-cell function in the older population. HOMA1 and HOMA2 measures of ß-cell function have been found to have significant correlations with ß-cell function assessment by FSIGT and hyperglycemic clamp in primarily younger and middle-aged people (5, 29). The results of the current study suggest that HOMA1 and HOMA2 measures of ß-cell function may have limited use, even when examining large populations of older people with both NGT and IGT. HOMA1-ß-cell and HOMA2%B were unable to detect the absolute impairment in insulin secretion, which AIRg from FSIGT was able to detect. However, these two methods are measuring different aspects of insulin secretion: basal insulin secretion and acute insulin response to iv glucose
Fasting C-peptide levels can be used in the updated HOMA2 model, thereby alleviating concerns regarding hepatic and peripheral clearance of insulin. It would be of interest to assess whether C-peptide would provide a more accurate HOMA assessment of insulin secretion. However, C-peptide levels were not available in the current study. In addition, C-peptide levels cannot be used in the original HOMA1 model; thus, prior studies reporting HOMA results have used fasting insulin.
Both the original and updated HOMA models are based on a feedback loop between the liver and ß-cell in the basal state. The model assumes that fasting glucose is regulated by hepatic glucose production, which is insulin dependent, and that fasting insulin is dependent on ß-cell response to glucose (24). The HOMA2 computer model was updated to allow assessment with glucose levels up to 25 mM and assumed reduced suppression of hepatic glucose production and increased insulin secretion in response to glucose levels greater than 10 mM (24).
The previously described impaired insulin secretion, decreased ß-cell sensitivity to glucose with normal aging (1, 2, 19, 28) with more severe defects in older people with IGT (30) as well as a small decrease in insulin clearance in older people (1) may explain the limited use of HOMA-ß-cell, compared with dynamic testing with FSIGT (AIRg), to identify the absolute impairment in insulin secretion in older people with IGT. In particular, the underlying assumptions of the HOMA model ß-cell response curve (basal glucose production rate, insulin space, and plasma insulin half-life) may not be valid in older people. The HOMA model assumes that fasting glucose and insulin are indicative of normal insulin response to increasing glucose levels, which may not be valid in people with IGT. The significantly higher fasting glucose levels in older people with IGT, compared with both younger and older people with NGT, may also have contributed to the similar HOMA-ß-cell levels in the three groups. In contrast, AIRg represents insulin response to a maximal, controlled iv glucose challenge without need for glucose matching (18). Impairment in first-phase insulin secretion has been found to be critical in the progression to type 2 diabetes (31). The variability of fasting insulin levels, although reduced in the current study by using the average of three baseline samples, can be considerable and is an additional limitation to the HOMA method.
A limitation of this study is the smaller sample size in the young group, although this sample size was similar to or greater than that involved in many other studies examining HOMA and other measures of insulin action and ß-cell function. The results may be generalized to both men and women, unlike other studies, which mainly involved men. However, our study group included predominantly a Caucasian population, and the findings may not apply to people of other ethnic groups.
In summary, HOMA measures of insulin resistance were able to detect increased insulin resistance in older people with IGT as confirmed by SI from FSIGT. Overall agreement between the two measures of insulin resistance was moderate. HOMA measures of ß-cell function were similar in the three groups, and overall agreement between FSIGT and HOMA measures of ß-cell function was weak. In contrast, an absolute impairment of AIRg was present in the older IGT group.
In conclusion, when the sample size is large enough, HOMA-IR may be able to detect insulin resistance in older men and women with IGT but still requires caution in its use due to dependence on insulin response to glucose in the basal state and the variability of fasting insulin levels. However, for assessment of ß-cell function, even with large populations of older people with NGT and IGT, dynamic testing appears to be necessary to quantitate age-related impairment of ß-cell function.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online November 29, 2005
Abbreviations: AIRg, Acute insulin response to glucose; AUC, area under the curve; BMI, body mass index; DI, disposition index; FSIGT, frequently sampled iv glucose tolerance test; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA1-IR, original HOMA model of insulin resistance; HOMA1-ß-cell, original HOMA model of ß-cell function; HOMA2, updated (or computer) model; HOMA2%ß, HOMA2-%ß-cell function; IGT, impaired glucose tolerance; NGT, normal glucose tolerance; OGTT, oral glucose tolerance test; SI, sensitivity to insulin.
Received August 10, 2005.
Accepted November 22, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. G. Pittas, S. S. Harris, M. Eliades, P. Stark, and B. Dawson-Hughes Association between Serum Osteocalcin and Markers of Metabolic Phenotype J. Clin. Endocrinol. Metab., March 1, 2009; 94(3): 827 - 832. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Drummond-Barbosa Stem Cells, Their Niches and the Systemic Environment: An Aging Network Genetics, December 1, 2008; 180(4): 1787 - 1797. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. V. Matveyenko, J. D. Veldhuis, and P. C. Butler Adaptations in pulsatile insulin secretion, hepatic insulin clearance, and {beta}-cell mass to age-related insulin resistance in rats Am J Physiol Endocrinol Metab, October 1, 2008; 295(4): E832 - E841. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Bloem and A. M. Chang Short-Term Exercise Improves {beta}-Cell Function and Insulin Resistance in Older People with Impaired Glucose Tolerance J. Clin. Endocrinol. Metab., February 1, 2008; 93(2): 387 - 392. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. G. Pittas, S. S. Harris, P. C. Stark, and B. Dawson-Hughes The Effects of Calcium and Vitamin D Supplementation on Blood Glucose and Markers of Inflammation in Nondiabetic Adults Diabetes Care, April 1, 2007; 30(4): 980 - 986. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Chang, M. J. Smith, A. T. Galecki, C. J. Bloem, and J. B. Halter Impaired {beta}-Cell Function in Human Aging: Response to Nicotinic Acid-Induced Insulin Resistance J. Clin. Endocrinol. Metab., September 1, 2006; 91(9): 3303 - 3309. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
