Diabetes & Metabolism 37 (2011) 179–188

Review

How can we measure insulin sensitivity/resistance?

B. Antuna-Puente a , E. Disse b,c,d,e,f,g , R. Rabasa-Lhoret h,i,j ,
M. Laville b,c,d,e,f,g , J. Capeau a,k,l , J.-P. Bastard a,∗,k,l
a Unité mixte de recherche S938, centre de recherche Saint-Antoine, Institut national de la santé et de la recherche médicale,
université Pierre-et-Marie-Curie Paris 6, 27, rue Chaligny, 75571 Paris cedex 12, France
b Centre de recherche en nutrition humaine Rhône-Alpes, université de Lyon, 69003Lyon, France
c Inserm, U870, IFR62, 69008 Lyon, France
d Inra, UMR1235, 69008 Lyon, France
e Insa-Lyon, RMND, 69621 Villeurbanne, France
f Faculté de médecine R.-Laennec, université Lyon 1, 69003 Lyon, France
g Hospices civils de Lyon, hôpital Edouard-Herriot, 69008 Lyon, France
h Department of nutrition, université de Montréal, Montréal, Québec, Canada
i Institut de recherches cliniques de Montréal (IRCM), Montréal, Québec, Canada
j Montreal Diabetes Research Center (MDRC), centre hospitalier de l’université de Montréal, Montréal, Québec, Canada
k Service de biochimie et hormonologie, AP–HP, hôpital Tenon, 75020 Paris, France
l Centre de recherche en nutrition humaine, Île-de-France, 75013 Paris, France

Received 27 September 2010; received in revised form 25 January 2011; accepted 27 January 2011
Available online 23 March 2011

Abstract
Insulin resistance represents a major public health problem, as it plays a major role in the pathophysiology of type 2 diabetes mellitus; it is also
associated with increased cardiovascular risk and atherogenic dyslipidaemia, and is a central component of the cluster of metabolic abnormalities
that comprise the metabolic syndrome. Thus, the development of tools to quantify insulin sensitivity/resistance has been the main objective of
a number of studies. Insulin resistance can be estimated with the use of several biological measurements that evaluate different aspects of this
complex situation. To that end, it requires various resources, ranging from just a single fasting blood sample for simple indices, such as the HOMA
or QUICKI, to a research setting in which to perform the gold-standard hyperinsulinaemic–euglycaemic clamp test. The choice of method for
evaluating insulin resistance depends on the nature of the information required (classification of individual subjects, group comparisons, precise
measurement of either global, muscle or liver insulin sensitivity/resistance) and on the available resources. The aim of this review is to analyze the
most frequently used assay methods in an attempt to evaluate when and why these methods may be useful.
© 2011 Elsevier Masson SAS. All rights reserved.

Keywords: Hyperinsulinaemic–euglycaemic glucose clamp; Insulin sensitivity; Insulin resistance; HOMA; QUICKI; OGTT; Review

Résumé
Comment mesurer la sensibilité/résistance à l’insuline?.
La résistance à l’insuline représente un problème majeur de santé publique puisqu’elle joue un rôle central dans la physiopathologie du diabète de
type 2, est associée à une augmentation du risque cardiovasculaire et est un élément central d’un ensemble d’anomalies métaboliques qui définissent
le syndrome métabolique. Ainsi, de nombreuses études se sont attachées à développer des outils pour apprécier et quantifier la sensibilité/résistance
à l’insuline. L’insulinorésistance peut être mesurée de plusieurs façons qui permettent d’en appréhender différents aspects et qui font appel à des
méthodes qui vont du simple prélèvement sanguin à jeun, qui permet le calcul des index simples HOMA ou QUICKI, à la méthode de référence
du clamp hyperinsulinémique euglycémique. Le choix de la méthode pour évaluer le niveau de résistance à l’hormone dépend de la nature des

∗ Corresponding author. Tel.: +33 1 56 01 66 76; fax: +33 1 56 01 60 17.

E-mail address: jean-philippe.bastard@tnn.aphp.fr (J.-P. Bastard).

1262-3636/$ – see front matter © 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.diabet.2011.01.002
180 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
informations requises (procéder à un classement de sujets, de groupes, mesurer précisément la sensibilité/résistance à l’insuline globale, musculaire
ou hépatique) et des outils disponibles. Le but de cette revue est d’analyser les méthodes le plus fréquemment utilisées afin de clarifier quand et
dans quelles situations, elles sont utiles.
© 2011 Elsevier Masson SAS. Tous droits réservés.
Mots clés : Clamp hyperinsulinémique euglycémique ; Insulinorésistance ; Insulinosensibilité ; HOMA ; QUICKI ; HGPO ; Revue
1. Introduction
Insulin is a peptide hormone that exerts a variety of effects,
mostly anabolic, on different cell types, but mainly in hepato-
cytes, myocytes and adipocytes. The hormone stimulates cell
growth and differentiation, and promotes the storage of sub-
strates in fat, liver and muscle by stimulating lipogenesis and
lipid storage, and glycogen and protein synthesis, while inhibit-
ing lipolysis, glycogenolysis and protein breakdown. Thus,
resistance to the hormone leads to hyperglycaemia and hyper-
lipidaemia, while a lack of insulin results in protein-wasting,
ketoacidosis and, ultimately, death [1]. Although insulin is cen-
tral to all intermediary metabolic processes, its main action is
related to glucose homoeostasis. Therefore, insulin resistance
is typically defined as a decrease of insulin-mediated glucose
disposal in insulin-sensitive tissues and increased hepatic glu-
cose production (HGP). Peripheral insulin resistance—in other
words, reduced insulin sensitivity—refers to a low capacity to
utilize glucose in target tissues, mainly skeletal muscle, where it
affects glucose transport and glycogen synthesis. At the level of
adipose tissue, the main feature of insulin resistance is increased
lipolysis via decreased antilipolytic insulin activity [2,3]. In
addition, hepatic insulin resistance refers to increased glucose
production by the liver via impaired inhibition of glycogenoly-
sis and stimulation of gluconeogenesis [3,4]. Overall, insulin
is implicated in numerous cellular mechanisms and in key
tissues, such as muscle, in which insulin regulates over 700
genes [5].
Since the definition of the metabolic syndrome in the 1980s
by Reaven [6], who first proposed that insulin resistance was
clustered together with numerous cardiometabolic abnormalities
(hyperglycaemia, elevated plasma triglycerides [TG], low levels
of high-density lipoprotein [HDL] cholesterol and hyperten-
sion) in association with the emergence of the obesity pandemic,
insulin resistance has gained in importance. Insulin resistance
is a major component of several significant cardiometabolic
abnormalities, including the metabolic syndrome, type 2 dia-
betes and cardiovascular disease (CVD) [7]. Their presence
strongly suggests evidence of an insulin-resistant state, but does
not either demonstrate its presence or quantify it. Some nonobese
subjects may present with all of the metabolic abnormalities,
whereas other obese subjects may not present with any metabolic
abnormality and be classified as metabolically normal [8,9].
Moreover, insulin resistance is not always pathological, and
may be observed in physiological states such as pregnancy and
puberty [10]. In addition, there is considerable overlap of the
insulin sensitivity/resistance distribution between the physio-
logical and pathological states [10]. Thus, the development of
tools to quantify insulin sensitivity/resistance has been the main
focus of several studies. Such studies can provide information
on the underlying pathophysiological mechanisms, and may be
associated with clinical and therapeutic outcomes in the field of
diabetes and CVD.
The different methods available for measuring insulin sen-
sitivity/resistance have been exhaustively reviewed elsewhere
[11–18]. For this reason, the aim of the present review is to dis-
cuss the most frequently used tests in clinical research and to
focus on the surrogate methods currently available.
1.1. The importance of insulin assay
In addition to most dynamic tests to evaluate insulin
resistance, all the simple surrogate indices for insulin sensitiv-
ity/resistance estimation mainly depend on insulin values. These
are particularly important in the fasting state, when insulin levels
are relatively low and a small difference in its measurement could
considerably impact the surrogate indices. Also, it is important
to consider that the insulin assay itself might be critical at both
the preanalytical and analytical levels. In addition to the well-
known importance of avoiding haemolyzed samples for insulin
assay, it is important to note that insulin levels are higher in
serum than in plasma samples, with differences that are pro-
portional to insulin concentrations [19]. Therefore, even while
using the same analytical procedures, when insulin values are
derived from both serum and plasma samples in a study, they
cannot be compared.
Furthermore, it is not possible to compare the results of sur-
rogate indices where different insulin assays have been used,
as may be the case in multicentre studies where insulin has not
been centrally tested. Indeed, a twofold difference in insulin val-
ues has been reported in a study investigating 11 different types
of human insulin assays [19]. Standardization of these insulin
assays may thus improve several variations that preclude such
comparisons between studies [20]. Recently, the Insulin Stan-
dardization Work Group recommended that concentrations of
insulin be reported in international units (SI, pmol/L), thereby
avoiding all references to traditional insulin units based on
insulin biological activity per milligram of standard preparation
[20,21].
On the other hand, the Work Group also showed that
most commercial insulin assays [21] can achieve consistent
performance with calibration traceability based on indi-
vidual serum samples, with insulin concentrations set by
isotope dilution-liquid chromatography/tandem mass spectrom-
etry measurement, a procedure that is calibrated using purified
recombinant insulin. They also observed that several, but not all,
insulin assays were acceptable for precision, accuracy and cross-
reactivity. However, performance of these insulin assays was not
similar at low concentrations, a critical point when determin-
 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
ing cut-off values between normal and insulin-resistant states.
Therefore, a serious and sustainable insulin assay standardiza-
tion programme needs to be established [20], as this will likely
improve the reliability of surrogate indices of insulin sensitiv-
ity.
2. Frequently used exploration tests in clinical research
2.1. Hyperinsulinaemic–euglycaemic glucose clamp
The hyperinsulinaemic–euglycaemic glucose (HIEG) clamp
test is accepted as the gold standard procedure for the assess-
ment of insulin sensitivity in clinical research. The procedure,
as described by DeFronzo et al. [22], consists of both a constant
intravenous infusion of insulin to create an artificially cons-
tant hyperinsulinaemic state and a variable glucose infusion to
maintain a euglycaemic state. For this procedure, intravenous
catheterization of one arm is necessary throughout the test for
both the glucose and insulin infusions. During the test, the arteri-
ovenous blood glucose difference increases in proportion to the
rate of insulin infusion and insulin sensitivity. Thus, adjusting
glucose infusion rates (GIR) from venous glycaemia may lead
to overestimation of insulin sensitivity.
To avoid such a drawback, arterial catheterization—or, at
least, arterialized blood—may be obtained via retrograde can-
nulation of a wrist vein warmed with a heating pad. This allows
the opening up of arteriovenous anastomoses as is required
for blood glucose measurement. However, in one comparative
study, Nauck et al. [23] showed that HIEG clamp experiments
using venous, arterialized venous and capillary euglycaemia
were all nearly equally useful for determination of insulin sen-
sitivity. Glucose levels were maintained (at 80–90 mg/dL or
4.45–5.00 mmol/L) by monitoring the glucose level every 5 or
10 min, and adjusting the infusion rate of a 20% dextrose solu-
tion. The constant insulin infusion produces a new steady state,
with a plateau of insulin concentration sufficiently above fast-
ing levels to suppress HGP, and to increase glucose disposal in
skeletal muscle and adipose tissue.
Under such a state of constant glycaemia, the GIR equals
the glucose disposal rate (M), which is the mean rate of glu-
cose infused in the last 30 min of the clamp test, which overall
lasts around 2 to 3 h. Thus, we can calculate the glucose dis-
posal per unit of plasma insulin concentration in a steady state
(M/I). Usually, M is expressed in mg/kg body weight/min. How-
ever, as most of the glucose uptake occurs in muscles and only
a small proportion in adipose tissue, the GIR expressed as M
(in mg/kg/min) could overestimate insulin resistance in obese
subjects [11]. For this reason, it has been proposed to normal-
ize M to fat-free mass (FFM) (mg/kgFFM /min) or to lean body
mass (mg/kgLBM /min), which should be more representative
of insulin sensitivity whatever the subject’s body mass index
(BMI).
However, this concept was recently criticized, as it was shown
that adipose tissue contributes notably to insulin-mediated glu-
cose uptake in morbidly obese subjects [24]. Thus, it has been
proposed that the M value should be normalized by kg body
weight instead of FFM in such subjects [24]. Nevertheless,
181
when normalization by FFM is required, body composition mea-
surement can be done by relatively simple measures such as
bioimpedance or the more precise dual X-ray absorptiometry
(DXA). Furthermore, the results can also be expressed as
SIClamp = M/(G × I), where G is the steady-state plasma glu-
cose (SSPG) level, and I = the difference between fasting and
steady-state plasma insulin levels [17].
The validity of HIEG clamp measurements depends on the
complete suppression of HGP by insulin. In cases where HGP
is not suppressed, M underestimates the total amount of glu-
cose metabolized. Suppression of HGP is achieved with an
insulin infusion rate of 40–60 mU/m2 /min (dose/body surface
area/time) in nonobese (BMI < 25 kg/m2 ) nondiabetic subjects
in most, but not all, cases [22]. In some situations (overweight,
obesity or type 2 diabetes) where HGP is not totally inhibited by
such an infusion rate, HGP can be measured using an infusion
of glucose labelled with stable isotopes and/or a higher rate of
insulin infusion (≥ 80 mU/m2 /min) to ensure a high probability
of HGP suppression [16].
The use of different insulin rates during the HIEG clamp, pro-
ducing different steady-state levels, allows determination of the
dose–response curves between increasing insulin concentrations
and increases in glucose disposal rate. In this way, Rizza et al.
[25] found that the maximum effect of insulin in control subjects
occurred at insulin concentrations of 200–700 ␮U/mL, while the
half-maximum effect was seen at 42–89 ␮U/mL, a level that can
be obtained with an insulin infusion rate of 40 mU/m2 /min or
1 mU/kg/min.
The HIEG clamp test requires that glycaemia be clamped at a
predetermined value in the normal range. This is easily achieved
in normoglycaemic subjects, but is more complex in diabetic
patients because of the variability of previous plasma glucose
levels. When plasma glucose is high after an overnight fast in
diabetic patients, intravenous insulin should be given to nor-
malize plasma glucose before starting the clamp. However, this
procedure, which may require some time to achieve, may induce
acute changes of glycaemia and, thus, alter insulin sensitivity.
Therefore, some investigators prefer clamping plasma glucose
at its fasting level (isoglycaemia). However, in this case, it is
necessary to correct the GIR for urinary loss of glucose, and the
clamp results are then more appropriately expressed as SIClamp
[17].
The HIEG clamp test has generated a vast amount of valuable
information. However, it has three major disadvantages: (1) it
requires two intravenous catheters, calibrated pumps and online
glucose-level determination; (2) it requires trained staff; and (3)
it is time-consuming, thereby precluding its use in large cohorts
[11]. In addition, the absence of standardization of important
parameters, such as duration of the procedure, insulin infusion
rate and normalization of the GIR, precludes comparisons across
studies [26].
The glucose clamp procedure is a reliable test for assess-
ing insulin secretion. To do this, the hyperglycaemic clamp
was developed. The procedure involves infusing glucose to
rapidly achieve hyperglycaemia. Then, as with the HIEG
clamp, the glucose infusion is adapted to maintain the level
of hyperglycaemia at usually 200 mg/dL (11.0 mmol/L). The
182 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
hyperglycaemic clamp allows evaluation of both the early and
late phases of insulin secretion, and permits the assessment of
insulin sensitivity/resistance by calculation of M/I, as with the
HIEG clamp [22]. However, this method is less precise than
the HIEG clamp for comparing insulin action between subjects
with wide variations in endogenous insulin secretion rates and/or
variable fasting glucose levels.
2.2. Insulin suppression test
The insulin suppression test (IST), described by Shen et al.
[27], is another method that directly measures insulin sensitiv-
ity/resistance. After an overnight fast, somatostatin (250 ␮g/h)
is intravenously infused to suppress the endogenous production
of insulin. At the same time, glucose (6 mg/kg body weight/min)
and insulin (50 mU/min) are infused over 150 min at a constant
rate. Glucose and insulin determinations are performed every
30 min for 2.5 h, then at 10 min intervals from 150 to 180 min
of the IST. The resulting SSPG concentration obtained during
the last 30 min of infusion represents an estimation of tissue
insulin sensitivity. The higher the SSPG concentration, the more
insulin-resistant the individual is. The IST was the first test to
use steady-state plasma insulin levels to promote disposal of a
glucose load.
However, as with the HIEG clamp, the IST is difficult to
apply in large epidemiological studies. Furthermore, there is a
risk of hypoglycaemia in insulin-sensitive subjects. Moreover,
the IST could provoke glycosuria in some subjects, such as type
2 diabetic patients, and could then lead to underestimation of
insulin resistance by SSPG.
2.3. Minimal model analysis of frequently sampled
intravenous glucose tolerance test
This method, described by Bergman et al. [28], gives an indi-
rect measurement of insulin sensitivity/resistance based on the
glucose and insulin values obtained during a frequently sampled
intravenous glucose tolerance test (FSIVGTT), and is associated
with a mathematical model that integrates the glucose–insulin
relationships. The minimal model uses the kinetics of increas-
ing circulating insulin and decreasing circulating glucose to
obtain two different indices—namely, the SiMM (insulin sen-
sitivity index) and SgMM (glucose effectiveness index) [14,29].
The SiMM index—which represents the link between insulin lev-
els and its effect, glucose disappearance from plasma—provides
information on peripheral and liver insulin sensitivity/resistance.
On the other hand, the SgMM gives information on the effects
of glucose on its own disappearance independent of any insulin
variation [29].
In addition to the insulin sensitivity index, the FSIVGTT also
derives two indices of insulin secretion (early phase and late
phase) from a single test. However, as with the HIEG clamp,
the minimal model is complex because it requires two venous
catheters, and blood must be sampled very frequently over 3–4 h,
even when a reduced sampling schedule has been applied [30].
The test also depends on specific software. Moreover, in individ-
uals with significantly impaired insulin secretion and/or major
insulin resistance, such as type 2 diabetic patients, the ‘stan-
dard’ minimal model is less reliable. In this case, the minimal
model analysis of FSIVGTT modified by exogenous insulin
infusion allows the assessment of insulin sensitivity/resistance
[31].
However, as these sophisticated methods are not easily
performed in large populations, surrogate indices for insulin
sensitivity/resistance assessment have been developed.
3. Simple surrogate indices of insulin
sensitivity/resistance
3.1. Indices derived from OGTT values
The oral glucose tolerance test (OGTT) is a simple test
that is widely used in clinical settings for the diagnosis of
glucose intolerance and type 2 diabetes. Whereas, for clinical
diagnosis, fasting and 2 h postload glucose values are suffi-
cient, additional samples for both plasma insulin and glucose
obtained every 30 min following an oral glucose load (75 g)
can allow estimation of insulin sensitivity and/or secretion [32].
The oral route of glucose delivery is clearly more physiolog-
ical than is either intravenous glucose injection or continuous
insulin infusion during an HIEG clamp. However, it remains
far from a classical meal situation. To estimate insulin resis-
tance, several surrogate indices incorporate plasma glucose and
insulin values during the OGTT into mathematical equations
while, in some cases, other parameters, such as weight, BMI
and glucose volume of distribution, are used [32–41]. The most
frequently used indices are summarized in Table 1. They have
usually been validated against the HIEG as the gold-standard
method.
The Matsuda index was first described by Matsuda and
DeFronzo [32] in subjects with a wide range of glucose toler-
ance. Fasting glucose and insulin values mainly reflect hepatic
insulin sensitivity, whereas mean OGTT values reflect insulin
sensitivity in peripheral skeletal muscle. More recently, the same
group proposed different formulas for more specifically evalu-
ating hepatic and muscle insulin resistance [42].
The Stumvoll index was derived from a multiple linear-
regression model evaluating the effect of different demographic
and OGTT parameters, and was also correlated with the clamp.
Using plasma insulin values at 120 min and plasma glucose val-
ues at 90 min during OGTT, and incorporating BMI into the
formula, Stumvoll et al. [33] obtained good correlation with the
clamp test.
In the same way, the Avignon index used glucose and insulin
concentrations at T0 and T120, incorporating estimates of glu-
cose volume of distribution (VD = 150 mL/kg body weight),
and was compared to the Bergman minimal model measure
of insulin sensitivity [36]. Belfiore et al. [35] used the area
under the curve (AUC) of insulin and glucose during OGTT,
and the Gutt index, adapted from the Cederholm index, omit-
ted the constant terms and, instead, used plasma glucose and
insulin concentrations at T0 and T120 [34,37]. The oral glu-
cose insulin sensitivity (OGIS) index, developed by Mari et al.
[38], is more complex, as it requires the use of two primary
 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
Table 1
Oral glucose tolerance test (OGTT)-derived indices.
Index Reference
ISI Avignon [36]
ISI Belfiore [35]
ISI Cederholm [34]
ISI Gutt [37]
ISI Matsuda [32]
ISI Stumvoll [33]
OGIS [38]
SIis OGTT [40]
VD: volume of distribution [150 (mL/kg) × body weight (kg)]; AUC: area under the curve; BMI: body mass index.
formulas that have to be incorporated into a third one. More-
over, the final calculation requires the incorporation of six
parameters—determined by the authors—that vary, depending
on the duration of the OGTT (2 or 3 h) and the units used to
express glycaemia (mg/dL or mmol/L). Nevertheless, it is pos-
sible to easily calculate the OGIS index through the website
given in Table 1.
More recently, Vangipurapu et al. [39] used regression anal-
ysis to develop a new index that combines insulin values during
an OGTT with HDL cholesterol and clinical measurements,
such as percentage of fat mass and BMI. This index, called the
‘liver insulin resistance’ (liver IR) index, correlated well with
the clamp and, more particularly, with the endogenous glucose
production measured by labelled glucose isotopes during the
HIEG clamp, which is representative of liver insulin resistance
[39].
These indices correlate well with the HIEG clamp and have
been validated in different populations. However, not only
insulin sensitivity per se, but other biological processes can
also affect plasma patterns of glucose and insulin after glu-
cose ingestion. These include the rate of glucose absorption,
and endogenous insulin secretion in response to glucose and
incretins, as well as insulin-independent effects of glucose
on glucose uptake and production [43]. Thus, in contrast to
insulin sensitivity measured by the HIEG clamp, insulin sen-
sitivity estimated from OGTT-derived indices could potentially
be confounded by physiological factors other than insulin resis-
tance itself. However, OGTT-derived indices depend on the
reproducibility of OGTT, which is variable due to interindi-
vidual variability of glucose absorption and interindividual
differences in incretin action on insulin secretion. Nevertheless,
183
Formula
{(0.137 × [108 /(fasting insulin␮IU/mL × fasting
glucosemg/dL × VD)]) + [108 /(120 min insulin␮IU/mL × 120 min
glucosemg/dL × VD)]}/2
2/[(INSp␮IU/mL ) × (GLYpmg/dL ) + 1], where INSp = AUC of insulin(␮IU/mL)
during OGTT divided by mean values of nondiabetic subjects as a unit, and
GLYp = AUC of glucose(mg/dL) during OGTT divided by mean values of
nondiabetic subjects as a unit
[75,000 + (fasting glucose–120 min
glucosemmol/L ) × 1.15 × 180 × 0.19 × body weightkg ]/[120 × (log mean
insulin␮IU/mL ) × mean glucosemmol/L ]
[{75,000 + [(fasting glucose–120 min glucosemg/dL ) × 0.19 × body
weightkg ]}/120]/[(fasting glucose + 2 h glucosemg/dL )/2]/log [(fasting
insulin + 120 min insulin␮IU/mL )/2]
104 /(fasting glucosemg/dL × fasting insulin␮IU/mL × mean glucose
OGTTmg/dL × mean insulin OGTT␮IU/mL )0.5
0.226–0032 × BMIkg/m 2 –0.0000645 × 120 min
insulinpmol/L –0.00375 × 90 min glucosemmol/L
For this calculation, go to: http://webmet.pd.cnr.it/ogis/index.php
1/{log [ glucose0+30+90+120 min (mmol/L) ] + log [ insulin0+30+90+120 min
(␮IU/mL) ]}
OGTT-derived indices have the advantage of being able to eval-
uate several parameters during the same test, including glucose
tolerance, insulin sensitivity and insulin secretion. However,
these indices are all more complex than fasting indices, and
require at least 2 h of testing.
More recently, the present team developed a new, additional
index derived from the OGTT (SIis OGTT) [40]. This involved
taking into consideration the data reported by Cheng et al. [44]
indicating that the log transformation of the sum of insulin con-
centrations, together with the sum of glucose concentrations
during OGTT, were highly correlated with the HIEG clamp
expressed as either M, M/I, MFFM or MFFM /I. Thus, by making
an analogy with the quantitative insulin-sensitivity check index
(QUICKI, see below), we proposed a simple formula (Table 1)
that correlates well with the HIEG clamp in the nondiabetic,
obese/overweight postmenopausal female population, however,
the clamp results are expressed [40]. The SIis OGTT index was
recently validated in a group of middle-aged sedentary men [45]
and in nonobese, nondiabetic healthy subjects [46].
Two recent studies systematically compared surrogate
indices, including OGTT and fasting-derived indices, with
the gold-standard HIEG clamp, and examined the rela-
tionship between these different assessments of insulin
sensitivity/resistance with CVD risk factors [47,48]. It was
demonstrated that the association between various surrogate
indices, including those derived from both fasting (HOMA,
QUICKI) and OGTT (Matsuda index, Stumvoll index and
SIis OGTT) measurements, and cardiometabolic risk factors can
vary widely from one index to another [48], although Lorenzo
et al. [47] found different results. They showed that surrogate
indices, including those derived from fasting measurements,
184 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
were reliable measures of insulin resistance and similar in
their relationships to metabolic abnormalities, including defini-
tive measures of fat distribution. However, three OGTT-derived
indices (Matsuda index, Avignon index and SIis OGTT) were
better correlated with the HIEG clamp [47]. Moreover, they
also concluded that surrogate indices, including those derived
from fasting measurements, are valid measures of insulin resis-
tance, and that OGTT with multiple sampling is not necessary
for estimating insulin sensitivity/resistance in both clinical and
epidemiological studies [47].
Although several indices derived from OGTT are currently
available and provide important information on insulin-
resistance status, the absence of reference range values for these
indices limits their use in clinical practice. Furthermore, most of
these indices still need to be validated in large cohorts of subjects
with wide ranges of insulin sensitivity and glucose tolerance
before considering their use in clinical practice.
3.2. Indices derived from fasting values
The fasting condition, in healthy subjects, represents a steady
state, with glycaemia tightly maintained between normal val-
ues as the result of the effect of insulin on HGP, which equals
whole-body glucose disposal upon fasting. Under such condi-
tions, plasma insulin levels will vary according to the degree
of insulin sensitivity. Therefore, surrogate fasting indices that
include both glycaemia and insulinaemia will permit estima-
tion of insulin resistance in a simple and cost-effective way,
although it primarily represents hepatic sensitivity [42]. Also,
glucose utilization under fasting conditions is mainly cerebral
and is noninsulin-dependent. Yet, what is surprising is that fast-
ing indices have been validated against the gold-standard clamp
method, which more specifically measures muscle, rather than
hepatic, insulin sensitivity. However, although fasting indices
primarily reflect hepatic insulin resistance, they also estimate
global insulin resistance. In addition, peripheral insulin resis-
tance and hepatic insulin resistance are closely related [49]. It is
possible with the clamp method to measure hepatic insulin sen-
sitivity/resistance, but this requires the use of labelled glucose,
which makes the test more complicated and laborious than the
classical HIEG clamp.
Another point that should be mentioned is that the use of
arterialized blood is recommended for blood glucose measure-
ment by glucose clamp. However, arterialized sampling is never
done in clinical practice for calculation of fasting indices and,
thus, might influence the relationship between fasting surrogate
indices and clamp results. Nevertheless, as mentioned above,
the results obtained from venous and arterialized blood mea-
surements are roughly comparable, and the differences are most
likely small enough to be comparable to those found in fasting
measurement variability and, thus, unlikely to alter the relation-
ship.
Homoeostasis model assessment (HOMA) and QUICKI are
the most widely used indices based on fasting parameters
[50,51]. The HOMA, first described in 1985 [50], is the result
of a computed model in which fasting glucose and insulin
values are plotted to allow assessment of the expected ␤-cell
response and insulin resistance with each pair of values. The
HOMA formula is a simplification of this mathematical model,
and can be described as the product of fasting glucose and
fasting insulin divided by a constant: HOMA = {[fasting insulin
(␮U/mL)] × [fasting glucose (mmol/L)]}/22.5. The denomina-
tor reflects normalization, based on normal fasting values as the
result of the product of 5 ␮U/mL insulin levels and 4.5 mmol/L
glucose levels.
As insulin secretion follows a pulsatile pattern, the use of the
mean of three samples taken at 5 min intervals, instead of a single
sample, to compute the HOMA score has been recommended
[50]. However, a single sample, which has often been used in
practice, provides an acceptable compromise and yields similar
results in large datasets [52]. Nevertheless, for a given individual,
the use of a single sample results in intrasubject coefficients of
variation that are higher than when three samples are used [52].
In this circumstance, the use of the mean insulin concentration
from three samples is preferable.
HOMA has proved useful in large epidemiological studies
by demonstrating good correlations with HIEG clamp results
in several populations. Interestingly, the HOMA of beta-cell
function (HOMA-␤) is a complementary formula that allows
estimation of insulin secretion [50], while the addition of anthro-
pometric measurements such as BMI is able to improve the
reliability of HOMA for estimating insulin sensitivity. Indeed,
to identify insulin-resistant patients, Stern et al. [53] developed
decision rules from measurements of obesity, fasting glucose,
insulin and other metabolic parameters in 2321 individuals,
who participated in an HIEG clamp study at several European
and USA sites. Distribution of whole-body glucose disposal
appeared to be bimodal, with an optimal insulin resistance cut-
off value of < 28 ␮mol/kgLBM /min. Using recursive partitioning,
the authors also developed several types of classification tree
models. One was of particular interest as it generated the simple
decision rule of an insulin-resistance diagnosis if any of the fol-
lowing conditions were present: BMI > 28.9 kg/m2 , HOMA-IR
(insulin resistance) > 4.65, or BMI > 27.5 kg/m2 and HOMA-
IR > 3.60, with a sensitivity of 84.9% and a specificity of 78.7%,
respectively [53].
However, the original contributors of the HOMA index
have pointed out some limitations of such simplification
of the mathematical model [52,54]. For this reason, the
HOMA software has been recently updated and is cur-
rently available as HOMA2 on the Oxford University website
(www.dtu.ox.ac.uk/homacalculator/index.php), and now allows
the estimation of steady-state beta-cell function (%B) and insulin
sensitivity (%S) as percentages of a normal reference population.
Although the use of HOMA-B and HOMA-S does not defini-
tively classify a subject as insulin-resistant or insulin-deficient,
it may offer information to help the clinician to decide which
first-line therapy to use. Moreover, such information may be
of interest in cases of severe insulin resistance syndromes by
identifying such a state.
Based on the same physiological principle, the QUICKI
index is similar to HOMA by being its inverse logarithm.
Described by Katz et al. [51] in a heterogeneous population
(nonobese, obese and type 2 diabetics), good correlations were
 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
observed between fasting values and HIEG clamp results with
QUICKI, as with HOMA [49]. As fasting insulin levels are
not normally distributed, their log transformation improves
linear correlation with the HIEG clamp. Thus, the QUICKI for-
mula is: 1/[log(I0 ␮U/mL ) + log(G0mg/dL )] [51]. QUICKI has been
extensively validated in different populations and is reasonably
correlated with the HIEG clamp [51,55–59].
Comparison of these two fasting indices has been the sub-
ject of many publications. Chen et al. [57] demonstrated, in
a cohort of normal, obese, hypertensive and diabetic subjects,
that QUICKI and log(HOMA) had better predictive power than
fasting insulin, different expressions of HOMA and the insulin
sensitivity index derived from the minimal model (SIMM ).
Indeed, log transformation of the HOMA formula can further
improve its linear correlation with the HIEG clamp [55]. On
the other hand, Vaccaro et al. [60] could find no difference
between the QUICKI and HOMA in evaluating insulin resis-
tance assessed by the FSIVGTT, nor in identifying subjects with
a metabolic syndrome.
Evidently, there is no apparent difference between these two
indices, as both are based on the same underlying physiologi-
cal principles. More specifically, HOMA and QUICKI exhibit
equivalent coefficients of variance when the HOMA is cor-
rectly log-transformed [61]. On the other hand, the usefulness
of the HOMA and QUICKI is limited in pathophysiological
situations such as insulinoma and primary hyperaldostero-
nism [62]. Their relevance has been reported to be weak in
nondiabetic subjects and in specific ethnic groups such as
Japanese subjects, in whom OGTT-derived indices appear to
better correlate with insulin-resistance measurements [63–65].
In addition, the reliability of these indices to estimate insulin
sensitivity may depend on the subject’s insulin-sensitive sta-
tus, as noted by Ferrara and Goldberg [66], who reported
weaker correlations for those with impaired glucose tolerance
(fasting glucose: < 5.6 mmol/L, 2 h glucose: 7.8–11.1 mmol/L)
than for subjects with normal glucose tolerance (fasting glu-
cose: < 5.6 mmol/L, 2 h glucose: < 7.8 mmol/L) in a middle-aged
(47–74 years) cohort.
However, in general, the relevance of fasting-derived indices
has been shown to be weaker in healthy populations, and sev-
eral authors have reported that fasting insulin is merely an
acceptable surrogate for estimating insulin sensitivity [67,68].
In fact, while healthy subjects may exhibit a wide range of
insulin-sensitivity levels, fasting glucose and insulin values are
maintained within narrow ranges, thus explaining the poorer
performance of fasting-derived surrogates in estimating insulin
sensitivity in clinically healthy subjects. This suggests that
small, but clinically relevant, variations in insulin sensitivity may
be overlooked by simple indices and, thus, need to be identified
by more sophisticated methods, such as the HIEG clamp or other
direct and dynamic measures of insulin sensitivity.
To enhance the sensitivity and specificity of fasting-derived
indices in estimating insulin sensitivity and identifying insulin-
resistant subjects, several groups have either developed new
indices, or proposed modified formulas for commonly used
indices by including additional fasting metabolic parameters,
especially lipids.
185
Perseghin et al. [69] added the log of the fasting value of
non-esterified fatty acids (NEFA) to the QUICKI formula. The
rationale for such modification was that the effect of insulin
on lipolysis occurs at lower levels than its effects on glucose
metabolism [2]. Thus, by estimating the antilipolytic effect of
insulin, fasting NEFA concentrations can reflect insulin resis-
tance earlier than do fasting glucose values.
This revised QUICKI formula improved its correlation with
HIEG clamp data in healthy subjects and in healthy type 2
diabetes offspring [69], and was further confirmed by two inde-
pendent teams [70,71]. The present authors also observed an
improved performance with the revised QUICKI in a wide spec-
trum of pathophysiological states [70]. Switching NEFA for
glycerol, another surrogate of lipolysis, similarly enhanced the
QUICKI formula performance [70]. However, despite its poten-
tial value, the addition of NEFA to the fasting-derived index
formula required a further biochemical measure, and NEFA con-
centrations may also be dependent on confounding parameters
such as dietary interventions and weight loss [72].
McAuley et al. [73] used regression analysis and bootstrap
procedures to identify fasting insulin and fasting TG as the most
useful variables in the prediction of insulin sensitivity. They
proposed an empirical formula—the McAuley index—derived
from the regression equation including these two variables: exp
{2.63–0.28 ln[fasting insulin (␮IU/mL)]–0.31 ln(fasting TG
(mmol/L)]} [73]. Following the same line of thinking, Guerrero-
Romero et al. [74] evaluated a similar simple index in subjects
with various degrees of glucose tolerance and body weight
using just fasting TG and glucose (TyG index = ln[fasting TG
(mg/dL) × fasting glucose (mg/dL)/2]). They observed good
correlation with the clamp and suggested that this index might be
helpful for the identification of subjects with insulin resistance
[74].
Recently, our group developed a new fasting-based index
in a population of healthy subjects (31 men and 39 women)
[75]. Using multiple forward-regression analysis, we determined
that fasting NEFA concentrations and the atherogenic index
were two reliable variables to further improve the estimation
of insulin sensitivity in healthy subjects. Indeed, among the
variables tested, fasting insulin, NEFA and the HDL choles-
terol/total cholesterol ratio explained 53% of the variation of
insulin sensitivity, expressed as MFFM /I. We also derived a
simplified formula from the regression equation, the Disse
index: 12 × {2.5 × [HDL-cholesterol (mmol/L)/total choles-
terol (mmol/L)]–NEFA (mmol/lL)}–insulinaemia (␮IU/mlmL)
[75]. The Disse index was highly correlated with MFFM /I
(r = 0.79, P < 0.001), and there was good agreement between the
two methods. Using a receiver operating characteristic (ROC)
curve analysis, we also demonstrated that this formula was the
best way to identify insulin-resistant subjects in an apparently
healthy population, as it was able to identify 88% of insulin-
resistant subjects with 77% specificity [75]. Interestingly, the
Disse index is also able to estimate both insulin sensitivity and
its improvement following a weight-loss programme in a popula-
tion of nondiabetic overweight or obese postmenopausal women
[76], and it is still relevant for estimating insulin sensitivity
in type 2 diabetic patients (personal data), despite the absence
186 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
of a glucose value in the formula. However, the impact of the
widely prescribed lipid-lowering medications on the accuracy of
formulas that include lipid parameters remains to be determined.
4. Choosing a method to measure or estimate insulin
sensitivity
Numerous factors have to be taken into account when select-
ing a way to measure or estimate insulin sensitivity. Among these
factors, the nature of the study population and the question that is
being asked are the main indicators. However, different methods
also require, for example, different time considerations, sam-
ple numbers and costs that will ultimately influence the choice.
In addition, the different methods allow the investigation of
different aspects of glucose homoeostasis and different compo-
nents of insulin sensitivity/resistance. Thus, knowledge of the
underlying mechanisms that can explain the results obtained
with the different methods is of major importance. Moreover,
it should be remembered that numerous factors are thought to
influence insulin sensitivity, such as recent physical exercise and
food intake. Therefore, it is important to develop a method of
standardization to control dietary intake, physical activity and
duration of fasting prior to the use of any methods (surrogate or
direct measures of insulin sensitivity/resistance) to assess insulin
resistance [77].
When exploring insulin sensitivity in research protocols,
the HIEG clamp—and, to a lesser extent, its hyperglycaemic
version—remain the gold-standard method for the direct mea-
surement of insulin sensitivity. Depending on the clamp protocol
used (insulin infusion rate, use of stable metabolic tracers), it
is possible to assess tissue-specific insulin sensitivity in mus-
cle, liver and adipose tissue [11–18]. In addition, the HIEG
clamp can be used in all types of populations, including dia-
betic patients, although the situation is more complex in such
patients and calls for a choice between clamping at euglycaemia
or isoglycaemia, as already discussed in the clamp section above.
The IST also allows the assessment of insulin resistance in var-
ious populations, including diabetic patients, but is not able to
assess the tissue specificity of insulin sensitivity. The FSIVGTT,
which allows assessment of both insulin sensitivity and glucose-
stimulated insulin secretion, is impractical in subjects with major
impaired insulin secretion (type 1 and 2 diabetic patients), does
not differentiate the tissue specificity of insulin sensitivity and
requires technical resources as major as those for an HIEG
clamp.
OGTT-derived indices offer several reliable types of informa-
tion within a simple test: glucose tolerance, insulin resistance
and insulin secretion. However, when specifically exploring
insulin sensitivity, the OGTT is not as precise as the HIEG. In
addition, although OGTT-derived indices can be used in larger
populations than the sophisticated methods described above, the
number of OGTT-derived formulas without a well-accepted ref-
erence formula limits comparisons between studies. Indeed, no
one formula appears to be significantly superior to the others
[48], despite some of them being better correlated with the HIEG
clamp [47].
In spite of their wide use in research, fasting-derived indices
remain imperfectly validated in numerous populations and situa-
tions. However, they are easy to calculate, which explains their
popularity. For large groups of subjects, such as in epidemiolog-
ical studies, they offer an easy-to-use tool for determining any
differences in insulin-sensitivity status between groups. They
may also be relevant for estimating global cardiometabolic risk,
as insulin sensitivity/resistance assessed by these indices has
been shown to correlate with the intima–media thickness of the
carotid artery [78], a well-accepted surrogate of atherosclerosis.
They are also relevant in the prediction of type 2 diabetes [79]
and the occurrence of cardiovascular events [80].
Nevertheless, it is important to note that, in clinical prac-
tice, assessing the level of insulin sensitivity using the different
methods presented in this review is not currently recommended.
Indeed, the identification of the classical risk factors (such as
glycaemia, blood pressure, lipid profile, BMI and waist circum-
ference) included in the various definitions of the metabolic
syndrome remains, at present, satisfactory for guiding any ther-
apeutic approaches.
5. Funding source
B. Antuna-Puente received a scholarship from Consejo
Nacional de Ciencia y Tecnología (CONACyT). R. Rabasa-
Lhoret holds the J.-A. De Sève Chair in clinical research and
receives a scholarship from Fonds de recherche en santé du
Québec (FRSQ). This manuscript was in part supported by
Canadian Institute for Health Research (CHIR) and the MONET-
SOMET programmes awarded to R.R.L.; the EGIR-RISC Study
was partly supported by EU grant QLG1-CT-2001-01252.
Disclosure statement
The authors have nothing to disclose.
References
[1] Saltiel A, Kahn R. Insulin signalling and the regulation of glucose and lipid
metabolism. Nature 2001;414:799–806.
[2] Groop LC, Bonadonna RC, DelPrato S, Ratheiser K, Zyck K, Ferrannini
E, et al. Glucose and free fatty acid metabolism in non-insulin-dependent
diabetes mellitus. J Clin Invest 1989;84:205–13.
[3] Fritsche L, Weigert C, Haring HU, Lehmann R. How insulin receptor sub-
strate regulate the metabolic capacity of the liver- implications for health
and disease. Curr Med Chem 2008;15:1316–29.
[4] Cho H, Mu J, Kim JK, Thorvaldsen Jl, Chu Q, Crenshaw 3rd EB, et al.
Insulin resistance and diabetes mellitus-like syndrome in mice lacking the
protein kinase Akt2 (PKB␤). Science 2001;292:1728–31.
[5] Rome S, Clément K, Rabasa-Lhoret R, Loizon E, Poitou C, Barsh GS, et al.
Microarray profiling of human skeletal muscle revels that insulin regulates
approximately 800 genes during a hyperinsulinaemic clamp. J Biol Chem
2003;278:18063–8.
[6] Reaven GM. Banting lecture 1988. Role of insulin resistance in human
disease. Diabetes 1988;37:1595–607.
[7] Bloomgarden Z. Insulin resistance concepts. Diabetes Care
2007;30:1320–6.
[8] Messier V, Karelis AD, Prud’homme D, Primeau V, Brochu M, Rabasa-
Lhoret R. Identifying metabolically healthy but obese individuals in
sedentary postmenopausal women. Obesity 2010;18:911–7.
 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
[9] Conus F, Rabasa-Lhoret R, Péronnet F. Characteristics of metaboli-
cally obese normal-weight (MONW) subjects. Appl Physiol Nutr Metab
2007;32:4–12.
[10] Wallace TM, Matthews DR. The assessment of insulin resistance in man.
Diabet Med 2002;19:527–34.
[11] Ferranini E, Mari A. How to measure insulin sensitivity. J Hypert
1998;16:895–906.
[12] Radziuk J. Insulin sensitivity and its measurement: structural com-
monalities among the methods. J Clin Endocrinol Metab 2000;85:
4426–33.
[13] Rabasa-Lhoret R, Laville M. How to measure insulin sensitivity in clinical
practice? Diabetes Metab 2001;27(2 Pt 2):201–8.
[14] Scheen AJ. Comment apprécier chez l’homme l’action de l’insuline
en recherche et en pratique. Ann Endocrinol (Paris) 1999;60:
179–87.
[15] Scheen AJ. Évaluation de l’insulinosécrétion et de l’insulinosensibilité chez
l’homme. Therapie 2007;62:311–8.
[16] Choukem SP, Gautier JF. How to measure hepatic insulin resistance? Dia-
betes Metab 2008;34:664–73.
[17] Muniyappa R, Lee S, Chen H, Quon MJ. Current approaches for assess-
ing insulin sensitivity and resistance in vivo: advantages, limitations, and
appropriate usage. Am J Physiol Endocrinol Metab 2008;294:E15–26.
[18] Matsuda M. Measuring and estimating insulin resistance in clinical and
research settings. Nutr Metab Cardiovasc Dis 2010;20:79–86.
[19] Manley SE, Stratton IM, Clark PM, Luzio SD. Comparison of 11 human
insulin assays: implications for clinical investigation and research. Clin
Chem 2007;53:922–32.
[20] Staten MA, Stern MP, Miller WG, Steffes MW, Campbell SE, and Insulin
Standardization Workgroup. Insulin assay standardization: leading to mea-
sures of insulin sensitivity and secretion for practical clinical care. Diabetes
Care 2010;33:205–6.
[21] Miller WG, Thienpont LM, Van Uytfanghe K, Clark PM, Lindstedt P,
Nilsson G, et al. Toward standardization of insulin immunoassays. Clin
Chem 2009;55:1011–8.
[22] DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a
method for quantifying insulin secretion and resistance. Am J Physiol
1979;237:E214–23.
[23] Nauck MA, Blietz RW, Qualmann C. Comparison of hyperinsulinaemic
clamp experiments using venous, “arterialized” venous, or capillary eug-
lycaemia. Clin Physiol 1996;16:589–602.
[24] Gniuli D, Castagneto-Gissey G, Iaconelli A, Leccesi L, Mingrone G. Fat
mass largely contributes to insulin mediate glucose uptake in morbidly
obese subjects. Int J Obes (Lond) 2010;34:1726–32.
[25] Rizza RA, Mandarino LJ, Gerich JE. Dose-response characteristics for
effects of insulin on production and utilization of glucose in man. Am J
Physiol 1981;240:E630–9.
[26] Karelis AD, Henry JF, Malita F, St-Pierre DH, Vigneault I, Poehlman ET,
et al. Comparison of insulin sensitivity values using the hyperinsulinemic
euglycemic clamp: 2 vs 3 hours. Diabetes Metab 2004;30:413–4.
[27] Shen SW, Reaven GM, Farquhar JW. Comparison of impedance to insulin-
mediated glucose uptake in normal subjects and subjects with latent
diabetes. J Clin Invest 1970;49:2151–60.
[28] Bergman RN, Prager R, Volund A, Olefsky JM. Equivalence of the insulin
sensitivity index in man derived by the minimal model method and the
euglycaemic glucose clamp. J Clin Invest 1987;79:790–800.
[29] Bergman RN, Phillips LS, Cobelli C. Physiologic evaluation of factors
controlling glucose tolerance in man. Measurement of insulin sensitivity
and Bcell glucose sensitivity from the response to intravenous glucose. J
Clin Invest 1981;68:1456–67.
[30] Cobelli C, Ruggeri A. A reduced sampling schedule for estimating the
parameters of the glucose minimal model from a labeled IVGTT. IEEE
Trans Biomed Eng 1991;38:1023–9.
[31] Welch S, Gebhart SS, Bergman RN, Phillips LS. Minimal model analysis
of intravenous glucose tolerance test-derived insulin sensitivity in diabetic
subjects. J Clin Endocrinol Metab 1990;71:1508–18.
[32] Matsuda M, DeFronzo R. Insulin sensitivity indexes obtained from oral
glucose tolerance testing, comparison with the euglycaemic insulin clamp.
Diabetes Care 1999;22:1462–70.
187
[33] Stumvoll M, Mitrakou A, Pimenta W, Jenssen T, Yki-Jarvinen H, Van
Haeften T, et al. Use of the oral glucose tolerance test to assess insulin
release and insulin sensitivity. Diabetes Care 2000;23:295–301.
[34] Cederholm J, Wibell L. Insulin release and peripheral sensitivity at the oral
glucose tolerance test. Diabetes Res Clin Pract 1990;10:167–75.
[35] Belfiore F, Iannello S, Volpicelli G. Insulin sensitivity indexes calculated
from basal and OGTT-induced insulin, glucose, and FFA levels. Mol Gen
Metab 1998;63:134–41.
[36] Avignon A, Boegner C, Mariano-Goulart D, Colette C, Monnier L. Assess-
ment of insulin sensitivity from plasma insulin and glucose in the fasting
or post oral glucose-load state. Int J Obes 1999;23:512–7.
[37] Gutt M, Davis CL, Spitzer SB, Llabre MM, Kumar M, Czarnecki EM,
et al. Validation of the insulin sensitivity index (ISI(0,120)): comparison
with other measures. Diabetes Res Clin Pract 2000;47:177–84.
[38] Mari A, Pacini G, Murphy E, Ludvik B, Nolan JJ. A model-based method
for assessing insulin sensitivity from the oral glucose tolerance test. Dia-
betes Care 2001;24:539–48.
[39] Vangipurapu J, Stancaková A, Kuulasmaa T, Paananen J, Kuusisto J,
et al. A novel surrogate index for hepatic insulin resistance. Diabetologia
2011;54:540–3.
[40] Bastard JP, Vandernotte JM, Faraj M, Karelis AD, Messier L, Malita FM,
et al. Relationship between the hyprinsulinaemic-euglycaemic clamp and
a new simple index assessing insulin sensitivity in overweight and obese
postmenopausal women. Diabetes Metab 2007;33:261–8.
[41] Soonthornpun S, Setasuban W, Thamprasit A, Chayanunnukul W,
Rattarasarn C, Geater A. Novel insulin sensitivity index derived
from oral glucose tolerance test. J Clin Endocrinol Metab 2003;88:
1019–23.
[42] Abdul-Ghani MA, Matsuda M, Balas B, DeFronzo RA. Muscle and liver
insulin resistance indexes derived from the oral glucose tolerance test.
Diabetes Care 2007;30:89–94.
[43] Hücking K, Watanabe RM, Stefanovski D, Bergman RN. OGTT-derived
measures of insulin sensitivity are confounded by factors other than insulin
sensitivity itself. Obesity 2008;16:1938–45.
[44] Cheng C, Campbell KL, Kushner H, Falkner BE. Correlation of
oral glucose tolerance test-derived estimates of insulin sensitivity with
insulin clamp measurements in an African-American cohort. Metabolism
2004;53:1107–12.
[45] Pigeon E, Riou ME, St-Onge J, Couillard E, Tremblay A, Marette
A, et al. Validation of a simple index (SIisOGTT) of insulin sen-
sitivity in a population of sedentary men. Diabetes Metab 2009;35:
398–403.
[46] Antuna-Puente B, Disse E, Lavoie ME, Rabasa-Lhoret R, Laville M, Bas-
tard JP. Evaluation of the SIisOGTT index in a healthy population with
normal glucose tolerance. Diabetes Metab 2009;35:236–7.
[47] Lorenzo C, Haffner SM, Stancakova A, Laakso M. Relation of direct
and surrogate measures of insulin resistance to cardiovascular risk fac-
tors in nondiabetic Finnish offspring of Type 2 diabetic individuals. J Clin
Endocrinol Metab 2010;95:5082–90.
[48] Malita FM, Messier V, Lavoie JM, Bastard JP, Rabasa-Lhoret R, Karelis
AD. Comparison between several insulin sensitivity indexes and metabolic
risk factors in overweight and obese postmenopausal women: a MONET
study. Nutr Metab Cardiovasc Dis 2010;20:173–9.
[49] Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, Zenere
MB, et al. Homeostasis model assessment closely mirrors the glucose
clamp technique in assessment of insulin sensitivity: studies in subjects
with various degrees of glucose tolerance and insulin sensitivity. Diabetes
Care 2000;23:57–63.
[50] Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner
RC. Homeostasis model assessment: insulin resistance and beta cell func-
tion from fasting plasma glucose and insulin concentrations in man.
Diabetologia 1985;28:412–9.
[51] Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, et al.
Quantitative insulin sensitivity check index: a simple, accurate method
for assessing insulin sensitivity in humans. J Clin Endocrinol Metab
2000;85:2402–10.
[52] Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modelling.
Diabetes Care 2004;27:1487–95.
188 B. Antuna-Puente et al. / Diabetes & Metabolism 37 (2011) 179–188
[53] Stern SE, Williams K, Ferrannini E, DeFronzo RA, Bogardus C, Stern MP.
Identification of individuals with insulin resistance using routine clinical
measurements. Diabetes 2005;54:333–9.
[54] Levy J, Matthews D, Hermans MP. Correct homeostasis model assess-
ment (HOMA) evaluation uses the computer program. Diabetes Care
1998;21:2191–2.
[55] Bastard JP, Robert JJ, Jardel C, Bruckert E, Grimaldi A, Hainque B. Is
quantitative insulin sensitivity check index, a fair insulin sensitivity index
in humans? Diabetes Metab 2001;27:69–70.
[56] Katsuki A, Sumida Y, Gabazza EC, Murashima S, Urakawa H, Morioka K,
et al. QUICKI is useful for following improvements in insulin sensitivity
after therapy in patients with type 2 diabetes mellitus. J Clin Endocrinol
Metab 2002;87:2906–8.
[57] Chen H, Sullivan G, Yu L, Katz A, Quon MJ. QUICKI is a useful index of
insulin sensitivity in subjects with hypertension. Am J Physiol Endocrinol
Metab 2003;284:E804–12.
[58] Chen H, Sullivan G, Quon MJ. Assessing the predictive accuracy of
QUICKI as a surrogate index for insulin sensitivity using a calibration
model. Diabetes 2005;54:1914–25.
[59] Bastard JP, Rabsa-Lhoret R, Maachi M, Vidal H, Laville M. What kind
of simple fasting index should be used to measure insulin sensitivity in
human? Diabetes Metab 2003;29:285–8.
[60] Vaccaro O, Masulli M, Cuomo V, Albarosa Riveliese AA, Uusitupa M,
Vessby B, et al. Comparative evaluation of simple indexes of insulin resis-
tance. Metabolism 2004;53:1522–6.
[61] Antuna-Puente B, Faraj M, Karelis AD, Garrel D, Prud’homme D, Rabasa-
Lhoret R, et al. HOMA or QUICKI: is it useful to test the reproducibility
of formulas? Diabetes Metab 2008;34:294–6.
[62] Skrha J, Haas T, Sindelka G, Prázný M, Widimský J, Cibula D, et al. Com-
parison of the insulin action parameters from hyperinsulinaemic clamps
with homeostasis model assessment and QUICKI indexes in subjects with
different endocrine disorders. J Clin Endocrinol Metab 2004;89:135–41.
[63] Yeni-Komshian H, Carantoni M, Abbasi F, Reaven GM. Relationship
between several surrogate estimates of insulin resistance and quantification
of insulin-mediated glucose disposal in 490 healthy nondiabetic volunteers.
Diabetes Care 2000;23:171–5.
[64] Kanauchi M, Tsujimoto N, Hashimoto T. Validation of simple indexes to
assess insulin sensitivity based on the oral glucose tolerance test in Japanese
population. Diabetes Res Clin Pract 2002;55:229–35.
[65] Kanauchi M, Yamano S, Kanauchi K, Saito Y. Homeostasis model assess-
ment of insulin resistance, quantitative insulin sensitivity check index, and
oral glucose insulin sensitivity index in nonobese, nondiabetic subjects with
high-normal blood pressure. J Clin Endocrinol Metab 2003;88:3444–6.
[66] Ferrara C, Goldberg A. Limited value of the homeostasis model assessment
to predict insulin resistance in older men with impaired glucose tolerance.
Diabetes care 2001;24:245–9.
[67] Malita FM, Karelis AD, St-Pierre DH, Garrel D, Bastard JP, Tardif A, et al.
Surrogate indexes vs. euglycaemic-hyperinsulinemic clamp as an indicator
of insulin resistance and cardiovascular risk factors in overweight and obese
postmenopausal women. Diabetes Metab 2006;32:251–5.
[68] Ruige JB, Mertens IL, Bartholomeeusen E, Dirinck E, Ferrannini E, Van
Gaal LF. Fasting-based estimates of insulin sensitivity in overweight and
obesity: a critical appraisal. Obesity 2006;14:1250–6.
[69] Perseghin G, Caumo A, Caloni M, Testolin G, Luzi L. Incorporation of the
fasting plasma FFA concentration into QUICKI improves its association
with insulin sensitivity in non obese individuals. J Clin Endocrinol Metab
2001;86:4776–81.
[70] Rabasa-Lhoret R, Bastard JP, Jan V, Ducluzeau PH, Andreelli F, Guebre
F, et al. Modified quantitative insulin sensitivity check index is better cor-
related to hyprinsulinaemic glucose clamp than other fasting-based index
of insulin sensitivity in different insulin-resistant states. J Clin Endocrinol
Metab 2003;88:4917–23.
[71] Brady LM, Gower BA, Lovegrove SS, Williams CM, Lovegrove JA.
Revised QUICKI provides a strong surrogate estimate of insulin sensi-
tivity when compared with the minimal model. Int J Obes Relat Metab
Disord 2004;28:222–7.
[72] Bastard JP, Jan V, Maachi M, Rabasa-Lhoret R, Jardel C, Bruckert E,
et al. Incorporation of non esterified Fatty Acids into QUICKI is not rele-
vant in Obese Subjects during diet inducing weight loss. Diabetes Metab
2002;28:333–4.
[73] McAuley KA, Williams SM, Mann JI, Walker RJ, Lewis-Barned NJ, Tem-
ple LA, et al. Diagnosing insulin resistance in the general population.
Diabetes Care 2001;24:460–4.
[74] Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-
Abundis E, Ramos-Zavala MG, Hernández-González SO, et al. The product
of triglycérides and glucose, a simple measure of insulin sensitivity.
Comparison with the clamp euglycemic-hyperinsulinemic clamp. J Clin
Endocrinol Metab 2010;95:3347–51.
[75] Disse E, Bastard JP, Bonnet F, Maitrepierre C, Peyrat J, Louche-Pelissier
C, et al. A lipid-parameter-based index for estimating insulin sensitivity
and identifying insulin resistance in healthy populations. Diabetes Metab
2008;34:457–63.
[76] Antuna-Puente B, Disse E, Faraj M, Lavoie ME, Laville M, Rabasa-
Lhoret R, et al. Evaluation of insulin sensitivity with a new lipid-based
index in non-diabetic postmenopausal overweight and obese women
before and after a weight loss intervention. Eur J Endocrinol 2009;161:
51–6.
[77] Bastard JP, Antuna-Puente B, Rabasa-Lhoret R. Is it important to reduce
inter- and intra-individual variations to explore insulin sensitivity? Diabetes
Metab 2010;36:249 [Epub].
[78] Rajala U, Laakso M, Päivänsalo M, Pelkonen O, Suramo I,
Keinänen-Kiukaanniemi S. Low insulin sensitivity measured by both
quantitative insulin sensitivity check index and homeostasis model
assessment method as a risk factor of increased intima-media thick-
ness of the carotid artery. J Clin Endocrinol Metab 2002;87:
5092–7.
[79] Hanley AJ, Williams K, Gonzalez C, D’Agostino Jr RB, Wagenknecht
LE, Stern MP, et al. Prediction of Type 2 diabetes using sim-
ple measures of insulin resistance: combined results from the
San Antonio Heart Study, the Mexico City Diabetes Study, and
the Insulin Resistance Atherosclerosis Study. Diabetes 2003;52:
463–9.
[80] Robins SJ, Rubins HB, Faas FH, Schaefer EJ, Elam MB, Anderson JW,
et al. Insulin resistance and cardiovascular events with low HDL choles-
terol: the Veterans Affairs HDL Intervention Trial (VA-HIT). Diabetes Care
2003;26:1513–7.