Content uploaded by Kaname Nakatani
Author content
All content in this area was uploaded by Kaname Nakatani on Jan 28, 2014
Content may be subject to copyright.
COMMENT
Neither Homeostasis Model Assessment nor Quantitative
Insulin Sensitivity Check Index Can Predict Insulin
Resistance in Elderly Patients with Poorly Controlled
Type 2 Diabetes Mellitus
AKIRA KATSUKI, YASUHIRO SUMIDA, HIDEKI URAKAWA, ESTEBAN C. GABAZZA,
SHUICHI MURASHIMA, KOHEI MORIOKA, NAGAKO KITAGAWA, TAKASHI TANAKA,
RIKA ARAKI-SASAKI, YASUKO HORI, KANAME NAKATANI, YUTAKA YANO, AND
YUKIHIKO ADACHI
Third Department of Internal Medicine (A.K., Y.S., H.U., E.C.G., K.M., N.K., T.T., R.A.-S., Y.H., Y.Y., Y.A.) and
Departments of Radiology (S.M.) and Laboratory Medicine (K.N.), Mie University School of Medicine, 514-8507 Mie, Japan
To clarify whether homeostasis model assessment (HOMA IR)
and quantitative insulin sensitivity check index (QUICKI)
may be indicators of insulin resistance in elderly patients
with type 2 diabetes mellitus, their relationship with the glu-
cose infusion rate during the euglycemic hyperinsulinemic
clamp study (clamp IR) was assessed. This study comprised 56
Japanese patients with type 2 diabetes mellitus; of these, 28
were 70 yr of age or older (group 1) and 28 were less than 70
yr of age (group 2). Their blood sugars were in poor control
(fasting plasma glucose levels: group 1, 9.0 ⴞ 2.6 mmol/liter;
group 2, 8.9 ⴞ 2.3 mmol/liter; hemoglobin A1c: group 1, 9.5 ⴞ
2.0%; group 2, 9.2 ⴞ 1.7%).
Log-transformed HOMA IR was significantly correlated
with the clamp IR in group 2 patients (r ⴝⴚ0.51, P < 0.01), but
not in group 1 patients (r ⴝⴚ0.28, P ⴝ 0.15). There was a
significant positive correlation between QUICKI and clamp
IR in group 2 patients (r ⴝ 0.50, P < 0.01). However, no sig-
nificant correlation was observed between QUICKI and clamp
IR in group 1 patients (r ⴝ 0.31, P ⴝ 0.12). There was a signif-
icant correlation between log-transformed HOMA IR (r ⴝ
ⴚ0.37, P < 0.01) or QUICKI (r ⴝ 0.37, P < 0.01) and clamp IR
when both groups were combined.
In conclusion, neither HOMA IR nor QUICKI should be used
as an index of insulin resistance in elderly patients with
poorly controlled type 2 diabetes mellitus. The results of this
study suggest the need for developing a new noninvasive
method for evaluating insulin resistance in those patients.
(J Clin Endocrinol Metab 87: 5332–5335, 2002)
A
GING IS ASSOCIATED with insulin resistance (1). In
Japan, elderly patients with type 2 diabetes mellitus
are gradually increasing due to a long life span (2). The
number of patients needing insulin-sensitizing agents is also
increasing. Thus, a simple, accurate, and reproducible index
is necessary for the diagnosis of insulin resistance in elderly
patients with type 2 diabetes mellitus.
To date, several methods for evaluating insulin resistance
have been developed (3–7). Of these, homeostasis model
assessment (HOMA IR) has been reported to be unable to
evaluate insulin resistance in elderly individuals at risk of
having impaired glucose tolerance (8), however, its applica-
bility in elderly patients with type 2 diabetes mellitus has not
been studied as yet. The quantitative insulin sensitivity check
index (QUICKI) has been recently reported to be a useful
marker of insulin resistance in patients with type 2 diabetes
mellitus (5, 9), but its usefulness in elderly patients with type
2 diabetes mellitus has not been evaluated as yet.
In the present study, we investigated whether HOMA IR
or QUICKI is correlated with the index of the euglycemic
hyperinsulinemic clamp study (clamp IR) in elderly patients
with type 2 diabetes mellitus.
Subjects and Methods
Subjects
This study comprised 28 Japanese patients with type 2 diabetes mel-
litus (group 1) who were 70 yr of age or older (range, 70 –81 yr). Data
obtained in 28 patients with type 2 diabetes mellitus (group 2) who were
less than 70 yr of age were used as control (range, 30–68 yr). The
duration of diabetes mellitus, obesity, and glycemic control were
matched with those of group 1 patients (Table 1). We recruited patients
treated with only diet and exercise to exclude the influence of oral
hypoglycemic agents.
Body mass index (BMI) was calculated as the body weight (in kilo-
grams) divided by the square of the height (in meters).
Type 2 diabetes mellitus was diagnosed in our patients at a local clinic
14.8 ⫾ 9.1 (group 1) or 14.2 ⫾ 5.4 (group 2) years before the beginning
of this study. All patients received dietary (1,440–1,720 kcal/d) and
exercise (walking 7,000 –10,000 steps/d) therapy. Their blood sugar was
in good control [hemoglobin A1c (HbA1c), ⱕ6.5%] at the beginning of
their disease, but thereafter their blood sugar gradually increased
(HbA1c, ⱖ8%) because of overeating and inactivity about several
months before admission. They were admitted in our clinical depart-
ment for glycemia control. We categorized them as having type 2 dia-
Abbreviations: ALT, Alanine aminotransferase; AST, aspartate ami-
notransferase; AUC, area under the curve; BMI, body mass index; clamp
IR, euglycemic hyperinsulinemic clamp study; HbA1c, hemoglobin A1c;
HDL, high-density lipoprotein; HOMA, homeostasis model assessment;
QUICKI, quantitative insulin sensitivity check index.
0013-7227/02/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 87(11):5332–5335
Printed in U.S.A. Copyright © 2002 by The Endocrine Society
doi: 10.1210/jc.2002-020486
5332
betes mellitus based on the diagnostic criteria of the American Diabetes
Association (10). In group 1, there were 16 patients with peripheral
neuropathy, 12 with background diabetic retinopathy, 14 with macro-
proteinuria, 7 with lacunar stroke, 9 with hypertension (of these, 5 were
receiving 5 mg/d enalapril maleate and 4 were receiving 5 mg/d am-
lodipine besilate), and 7 with hyperlipidemia treated with 10 mg/d
pravastatin sodium. None of them had ischemic heart disease or arte-
riosclerosis obliterans. In group 2, there were 16 patients with peripheral
neuropathy, 14 with background diabetic retinopathy, 10 with macro-
proteinuria, 5 with hypertension (of these, 3 were receiving 5 mg/d
enalapril maleate and 2 were receiving 5 mg/d amlodipine besilate), and
3 with hyperlipidemia treated with 10 mg/d pravastatin sodium. None
of them had macrovascular complications.
Informed consent was obtained from all subjects before the beginning
of the study.
Study protocol and methods
The blood levels of glucose, HbA1c, insulin, total cholesterol, triglyc-
eride, high-density lipoprotein (HDL) cholesterol, aspartate aminotrans-
ferase (AST), alanine aminotransferase (ALT), and creatinine and the
values of HOMA IR, QUICKI, clamp IR, body fat area, and blood
pressure were measured in all patients with type 2 diabetes mellitus. To
evaluate the insulin secretion ability, a 75-g oral glucose tolerance test
(Trelan G 75; Shimizu, Shizuoka, Japan) was performed in 20 group 1
(age, 73.1 ⫾ 2.4; male/female, 13/7; BMI, 21.2 ⫾ 2.3) and 20 group 2 (age,
53.8 ⫾ 9.7; male/female, 13/7; BMI, 22.3 ⫾ 1.2) patients.
The plasma glucose level was measured by an automated enzymatic
method. The HbA1c (normal value, 4.3–5.8%) was measured by HPLC.
Serum insulin was measured using an immunoradiometric assay kit
(Insulin Riabead II kit; Dainabot, Tokyo, Japan). The intra- and inter-
assay coefficients of variation of the assay were 2.0% and 2.1%, respec-
tively. No significant cross-reactivity or interference was observed be-
tween insulin and proinsulin, C-peptide, glucagon, secretin, or gastrin-I.
Serum levels of total cholesterol, triglyceride, HDL cholesterol, AST,
ALT, and creatinine were measured by enzymatic methods using an
autoanalyzer (TBA60M; Toshiba, Tokyo, Japan).
HOMA IR and QUICKI were calculated from the fasting concentra-
tions of insulin and glucose using the following formula: HOMA IR ⫽
fasting serum insulin (
U/ml) ⫻ fasting plasma glucose (mmol/liter)/
22.5 (4), QUICKI ⫽ 1/[log fasting serum insulin (
U/ml) ⫹ log fasting
plasma glucose (mg/dl)] (5). clamp IR was evaluated by the euglycemic
hyperinsulinemic clamp technique using an artificial pancreas (STG-22;
Nikkiso, Tokyo, Japan) (3, 7, 11, 12). At 0800 h, a priming dose of insulin
(Humulin R; Eli Lilly Japan, Kobe, Japan) was administered during the
initial 10 min in a logarithmically decreasing manner to rapidly raise
serum insulin to the desired level (1200 pmol/liter); this level was then
maintained by continuous infusion of insulin at a rate of 13.44 pmol/
kg䡠min for 120 min. The mean insulin level from 90 –120 min after
starting the clamp study was stable (group 1, 1380.0 ⫾ 426.0 pmol/liter;
group 2, 1339.8 ⫾ 371.4 pmol/liter). The clamp study was started at the
fasting levels of glucose (group 1, 9.0 ⫾ 2.6 mmol/liter; group 2, 8.9 ⫾
2.3 mmol/liter). Blood glucose was monitored continuously and de-
creased to normoglycemic levels within 50 min; thereafter it was main-
tained at the target clamp level (5.24 mmol/liter) by infusing 10% glu-
cose. The mean amount of glucose given during the last 30 min was
defined as the glucose infusion rate , and it was used as a clamp IR.
The capacity to secrete insulin in response to oral glucose stimulation
was estimated by the ratio of ⌬I
30
to ⌬G
30
(⌬I
30
/⌬G
30
; calculated as:
increment of serum insulin from 0–30 min/increment of plasma glucose
from 0–30 min), and the insulin area under the curve (AUC). Blood was
taken at 0, 30, 60 and 120 min.
Body fat area was evaluated by a previously described method (13).
The intra-abdominal visceral fat and sc fat areas were measured in all
subjects by abdominal computed tomography scanning taken at the
umbilical level. Any ip regions having the same density as the sc fat layer
were defined as a visceral fat area; this area was measured by tracing
object contours on films using a computerized planimetric method.
In addition, we measured blood pressure in the supine position after
a rest of 5 min.
Statistical methods
Data were expressed as the means ⫾ sd. Differences between groups
were determined by unpaired t tests after checking the normal distri-
bution of the data. The relationship of HOMA IR or QUICKI with clamp
IR was evaluated by univariate regression analysis. Unpaired t tests and
correlations were carried out using the StatView 4.0 software program
(Abacus Concepts, Berkeley, CA) for the Macintosh computer. A P value
less than 0.05 was considered as statistically significant.
Results
Type 2 diabetic patients were in poor glycemic control
(fasting plasma glucose levels: group 1, 9.0 ⫾ 2.6 mmol/liter;
TABLE 1. Clinical and laboratory characteristics of patients with type 2 diabetes mellitus
Group 1
(age ⭌70 yr)
Group 2
(age ⬍70 yr)
No. 28 28
Sex (M/F) 15/13 17/11
Age (yr) 73.3 ⫾ 3.0 55.4 ⫾ 8.9
a
Duration of diabetes (yr) 14.8 ⫾ 9.1 14.2 ⫾ 5.4
BMI (kg/m
2
)
21.2 ⫾ 2.2 22.0 ⫾ 1.4
Visceral fat area (cm
2
)
82.2 ⫾ 42.1 82.9 ⫾ 24.8
Subcutaneous fat area (cm
2
)
109.6 ⫾ 46.1 111.1 ⫾ 36.7
HbA1c (%) 9.5 ⫾ 2.0 9.2 ⫾ 1.7
Fasting plasma glucose (mmol/liter) 9.0 ⫾ 2.6 8.9 ⫾ 2.3
Fasting serum insulin (pmol/liter) 34.2 ⫾ 22.2 39.0 ⫾ 34.2
Total cholesterol (mmol/liter) 5.0 ⫾ 1.0 5.0 ⫾ 1.0
Triglyceride (mmol/liter) 1.4 ⫾ 0.5 1.7 ⫾ 0.8
HDL cholesterol (mmol/liter) 1.3 ⫾ 0.5 1.2 ⫾ 0.4
AST (IU/liter) 20.0 ⫾ 4.9 21.3 ⫾ 5.6
ALT (IU/liter) 21.8 ⫾ 6.8 19.0 ⫾ 7.1
Creatinine (
mol/liter) 85.2 ⫾ 17.1 80.8 ⫾ 14.9
Systolic blood pressure (mm Hg) 135.0 ⫾ 17.8 135.2 ⫾ 15.7
Diastolic blood pressure (mm Hg) 73.7 ⫾ 10.4 76.4 ⫾ 9.4
Clamp IR (
mol/kg䡠min) 36.1 ⫾ 13.0 41.6 ⫾ 17.5
HOMA IR 2.3 ⫾ 1.9 2.4 ⫾ 1.8
Log-transformed HOMA IR 0.256 ⫾ 0.306 0.300 ⫾ 0.268
QUICKI 0.353 ⫾ 0.039 0.347 ⫾ 0.031
Data are means ⫾
SD. M, Male; F, female.
a
P ⬍ 0.01; group 1 vs. group 2.
Katsuki et al. • Comments J Clin Endocrinol Metab, November 2002, 87(11):5332–5335 5333
group 2, 8.9 ⫾ 2.3 mmol/liter; HbA1c: group 1, 9.5 ⫾ 2.0%;
group 2, 9.2 ⫾ 1.7%; Table 1).
Among all diabetic patients (n ⫽ 56), clamp IR was sig-
nificantly correlated with log-transformed HOMA IR (r ⫽
⫺0.37, P ⬍ 0.01) and QUICKI (r ⫽ 0.37, P ⬍ 0.01). Evaluation
by age group showed that clamp IR is significantly correlated
with log-transformed HOMA IR (r ⫽⫺0.51, P ⬍ 0.01; Fig. 1,
E, solid line) and QUICKI (r ⫽ 0.50, P ⬍ 0.01; Fig. 2, E, solid
line) in group 2. However, neither log-transformed HOMA
IR (r ⫽⫺0.28, P ⫽ 0.15; Fig. 1, F, dotted line) nor QUICKI (r ⫽
0.31, P ⫽ 0.12; Fig. 2, F, dotted line) was significantly corre-
lated with clamp IR in group 1.
Log-transformed HOMA IR was strongly correlated with
QUICKI in group 1 (r ⫽⫺0.99, P ⬍ 0.0001), group 2 (r ⫽⫺1.0,
P ⬍ 0.0001), and in all patients (r ⫽⫺1.0, P ⬍ 0.0001).
There was a significant correlation between fasting serum
levels of insulin and clamp IR in group 2 patients (r ⫽⫺0.39,
P ⬍ 0.05) but not in group 1 patients (r ⫽⫺0.08, P ⫽ 0.67).
Significant negative correlations were observed between
clamp IR and visceral fat areas both in group 1 (r ⫽⫺0.66,
P ⬍ 0.001) and group 2 (r ⫽⫺0.62, P ⬍ 0.001) patients.
The ability of insulin secretion in group 1 patients (⌬I
30
/
⌬G
30
, 14.3 ⫾ 10.6; AUC, 10,868.4 ⫾ 5,905.2) was significantly
decreased compared with group 2 patients (⌬I
30
/⌬G
30
,
17.6 ⫾ 10.4, P ⬍ 0.03; AUC, 16,903.8 ⫾ 6,527.4, P ⬍ 0.02).
Serum levels of AST, ALT, and creatinine in all patients
were within the normal range (AST, 0 – 40 IU/liter; ALT,
0–35 IU/liter; creatinine, 53.0–123.8
mol/liter; Table 1).
Discussion
The present study demonstrated that neither HOMA IR
nor QUICKI is useful for assessing insulin resistance in
elderly patients with poorly controlled type 2 diabetes
mellitus.
Many studies have reported the usefulness of HOMA IR
and QUICKI in type 2 diabetic patients (14–16). The values
of HOMA IR and QUICKI depend on the fasting concentra-
tions of insulin and glucose, and they are thought to estimate
mainly hepatic insulin resistance (14). On the contrary, clamp
IR reflects essentially peripheral insulin resistance (14). Pe-
ripheral insulin resistance is generally associated with de-
creased ability of insulin to suppress adipose tissue lipolysis
(17). Insulin stimulation of lipolysis is associated with an
increased flux of free fatty acids from fatty tissue to the liver
and, at the same time, with decreased insulin-mediated sup-
pression of hepatic glucose output (18). In the present study,
we demonstrated a significant delayed and decreased insulin
secretion in elderly patients with type 2 diabetes mellitus.
This decreased ability to secrete insulin may alter the relation
between peripheral lipolysis and hepatic glucose output.
This alteration might explain the lack of correlation of clamp
IR with HOMA IR or QUICKI in our elderly diabetic patients.
In the present study, fasting plasma levels of glucose
(group 1, 9.0 ⫾ 2.6 mmol/liter; group 2, 8.9 ⫾ 2.3 mmol/liter)
were relatively high compared with those reported in pre-
vious studies [6.6 ⫾ 1.5 mmol/liter (4), 7.3 ⫾ 2.3 mmol/liter
(15), 7.8 ⫾ 2.4 mmol/liter (19)]. It has been reported that
fasting plasma levels of glucose may influence the correlation
between HOMA IR and clamp IR; HOMA IR is suitable for
evaluating insulin sensitivity in obese type 2 diabetic patients
whose fasting glucose levels are 9.4 mmol/liter or less (20).
Compared with data reported in previous studies [r ⫽ 0.92,
P ⬍ 0.0001 (4); r ⫽⫺0.75, P ⬍ 0.0001 (15); r ⫽ 0.75, P ⬍ 0.0001
(19)], the relationship between HOMA IR and clamp IR was
relatively weak (r ⫽⫺0.51, P ⬍ 0.01) in patients of group 2).
The relatively high levels of fasting glycemia might have
affected correlation between HOMA IR and clamp IR in
patients of group 1.
The glucose clamp study may be performed based on
different criteria; in our present study, the target blood glu-
cose level of the clamp study was lower than the fasting level.
Therefore, acute changes in insulin sensitivity judged based
on changes in glucose levels may cause misinterpretation of
the glucose clamp results; based on our present protocol,
clamp IR may be overestimated. This factor may also explain
the lack of correlation between clamp IR and HOMA IR or
QUICKI in our elderly patients.
Patients of group 2 included 10 patients with macropro-
teinuria, but with normal serum levels of creatinine. HOMA
IR has been reported to be useful in type 2 diabetic patients
FIG. 1. Correlations between log-transformed HOMA IR and clamp
IR in patients with type 2 diabetes mellitus. There was a significant
negative correlation in group 2 (r ⫽⫺0.51, P ⬍ 0.01, E, solid line) but
not in group 1 (r ⫽⫺0.28, P ⫽ 0.15, F, dotted line) patients.
FIG. 2. Correlations between QUICKI and clamp IR in patients with
type 2 diabetes mellitus. A significant positive correlation was ob-
served in group 2 (r ⫽ 0.50, P ⬍ 0.01, E, solid line) but not in group
1(r⫽ 0.31, P ⫽ 0.12, F, dotted line) patients.
5334 J Clin Endocrinol Metab, November 2002, 87(11):5332–5335 Katsuki et al. • Comments
with overt proteinuria and renal failure (15, 21). The defect
of insulin secretion in these diabetic patients may be com-
pensated by decreased renal clearance of insulin, and this
phenomenon may contribute to maintain the relationship
between hepatic and peripheral insulin resistance. Addi-
tional studies should be performed to clarify these points.
Association between these insulin resistance indicators
should also be assessed in patients with other vascular
complications.
In conclusion, we found a limited value of HOMA IR and
QUICKI for elderly patients with poorly controlled type 2
diabetes mellitus. A different index of insulin resistance
should be developed for this group of patients.
Acknowledgments
Received March 27, 2002. Accepted July 23, 2002.
Address all correspondence and requests for reprints to: Akira Kat-
suki, M.D., Third Department of Internal Medicine, Mie University
School of Medicine, 2-174 Edobashi, Tsu, 514-8507 Mie, Japan. E-mail:
katuki-a@clin.medic.mie-u.ac.jp.
References
1. DeFronzo RA 1979 Glucose intolerance and aging: evidence for tissue insen-
sitivity to insulin. Diabetes 28:1095–1101
2. Kuzuya T 1994 Prevalence of diabetes mellitus in Japan compiled from liter-
ature. Diabetes Res Clin Pract 24(Suppl 1):S15–S21
3. DeFronzo RA, Tobin JD, Andres R 1979 Glucose clamp technique: a method
for quantifying insulin secretion and resistance. Am J Physiol 237:E214 –E223
4. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC
1985 Homeostasis model assessment: insulin resistance and

-cell function
from fasting glucose and insulin concentrations in man. Diabetologia 28:
412–419
5. 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
6. Radziuk J 2000 Insulin sensitivity and its measurement: structural common-
alities among the methods. J Clin Endocrinol Metab 85:4426 – 4433
7. Katsuki A, Sumida Y, Murashima S, Fujii M, Ito K, Tsuchihashi K, Murata
K, Yano Y, Shima T 1996 Acute and chronic regulation of serum sex hormone-
binding globulin levels by plasma insulin concentrations in male noninsulin-
dependent diabetes mellitus patients. J Clin Endocrinol Metab 83:2515–2519
8. Ferrara CM, Goldberg AP 2001 Limited value of the homeostasis model
assessment to predict insulin resistance in older men with impaired glucose
tolerance. Diabetes Care 24:245–249
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 re-
sistance: implications for research applications. J Clin Endocrinol Metab 86:
5457–5464
10. The Expert Committee on the Diagnosis and Classification of Diabetes
Mellitus 1997 Report of the Expert Committee on the Diagnosis and Classi-
fication of Diabetes Mellitus. Diabetes Care 10:1183–1197
11. Kawamori R, Morishima T, Ikeda M, Kubota M, Kubota M, Kishimoto M,
Shiba Y, Matsuhisa M, Kodama M, Watarai T, Kamada T 1994 Effect of strict
metabolic control on glucose handling by the liver and peripheral tissues in
non-insulin-dependent diabetes mellitus. Diabetes Res Clin Pract 23:155–161
12. Katsuki A, Sumida Y, Gabazza EC, Murashima S, Furuta M, Araki-Sasaki
R, Hori Y, Yano Y, Adachi Y 2001 Homeostasis model assessment is a reliable
indicator of insulin resistance during follow-up of patients with type 2 dia-
betes. Diabetes Care 24:362–365
13. Tokunaga K, Matsuzawa Y, Ishikawa K, Tarui S 1983 A novel technique for
the determination of body fat by computed tomography. Int J Obes 7:437–445
14. Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, Zenere MB,
Monauni T, Muggeo M 2000 Homeostasis model assessment closely mirrors
the glucose clamp technique in the assessment of insulin sensitivity. Diabetes
Care 23:57– 63
15. Emoto M, Nishizawa Y, Maekawa K, Hiura Y, Kanda H, Kawagishi T, Shoji
T, Okuno Y, Morii H 1999 Homeostasis model assessment as a clinical index
of insulin resistance in type 2 diabetic patients treated with sulfonylureas.
Diabetes Care 22:818 – 822
16. Quon MJ 2001 Editorial: Limitations of the fasting glucose to insulin ratio as
an index of insulin sensitivity. J Clin Endocrinol Metab 86:4615– 4617
17. Groop LC, Bonadonna RC, DelPrato S, Ratheiser K, Zyck K, Ferrannini E,
DeFronzo RA 1989 Glucose and free fatty acid metabolism in non-insulin-
dependent diabetes mellitus: evidence for multiple sites of insulin resistance.
J Clin Invest 84:205–213
18. Rebrin K, Steil GM, Mittelman SD, Bergman RN 1996 Causal linkage be-
tween insulin suppression of lipolysis and suppression of liver glucose output
in dogs. J Clin Invest 98:741–749
19. Ikeda Y, Suehiro T, Nakamura T, Kumon Y, Hashimoto K 2001 Clinical
significance of the insulin resistance index as assessed by homeostasis model
assessment. Endocr J 48:81– 86
20. Ono T, Shiga N, Taneda Y, Umemura S 1999 The fasting-plasma glucose range
in which insulin resistance measured by homeostasis model assessment cor-
relates with euglycemic clamping. J Jpn Diabetes Soc 42:1005–1011
21. Shoji T, Emoto M, Nishizawa Y 2001 HOMA index to assess insulin resistance
in renal failure patients. Nephron 89:348–349
Katsuki et al. • Comments J Clin Endocrinol Metab, November 2002, 87(11):5332–5335 5335