, Volume 46, Issue 9, pp 1211–1219 | Cite as

Predominant role of reduced beta-cell sensitivity to glucose over insulin resistance in impaired glucose tolerance

  • E. Ferrannini
  • A. Gastaldelli
  • Y. Miyazaki
  • M. Matsuda
  • M. Pettiti
  • A. Natali
  • A. Mari
  • R. A. DeFronzo



Impaired glucose tolerance (IGT) is an insulin-resistant state and a risk factor for Type 2 diabetes. The relative roles of insulin resistance and insulin deficiency in IGT have been disputed.


In 40 IGT subjects and 63 sex-, age-, and weight-matched controls with normal glucose tolerance (NGT), we measured (i) indices of insulin sensitivity of fasting glucose production (by tracer glucose) and glucose disposal (M value on a 240 pmol·min–1·m–2 insulin clamp) and (ii) indices of beta-cell function (glucose sensitivity, rate sensitivity, and potentiation) derived from model analysis (Am J Physiol 283:E1159–E1166, 2002) of the insulin secretory response (by C-peptide deconvolution) to oral glucose.


In comparison with NGT, IGT were modestly insulin resistant (M=29±2 vs 35±2 µmol·min−1·kgFFM −1, p=0.01); insulin sensitivity of glucose production also was reduced, in approximate proportion to M. Despite higher baseline insulin secretion rates, IGT was characterized by a 50% reduction in glucose sensitivity [53 (36) vs 102 (123) pmol·min−1·m–2·mM–1, median (interquartile range), p=0.001] and impaired potentiation [1.6 (0.8) vs 2.0 (1.5) units, p<0.04] of insulin release, whereas rate sensitivity [1.15 (1.15) vs 1.38 (1.28) nmol·m–2·mM–1] was not significantly reduced. Glucose sensitivity made the single largest contribution (~50%) to the observed variability of glucose tolerance.


In IGT the defect in glucose sensitivity of insulin release quantitatively predominates over insulin resistance in the genesis of the reduced tolerance to oral glucose.


IGT beta-cell function insulin secretion insulin resistance mathematical modelling 

Impaired glucose tolerance (IGT) is a condition of high risk for the development of Type 2 diabetes. Early physiological studies consistently showed that IGT is an insulin-resistant state [1, 2, 3, 4, 5] and documented the presence of impaired insulin response to intravenous glucose [6, 7, 8, 9, 10, 11]. The general concept has evolved that in IGT, a mixture of insulin resistance and insulin deficiency is responsible for the abnormal rise in postglucose plasma glucose concentrations [12, 13]. However, many previous studies of IGT have based conclusions on the measurement of plasma insulin concentrations rather than true insulin secretion rates. More importantly, to what extent insulin response to intravenous glucose reflects the insulin secretory response to oral glucose or mixed meals remains uncertain. In studies using graded glucose infusions and the C-peptide deconvolution technique to reconstruct insulin secretory rates, one study [14] documented a slower rise in insulin secretion as a function of rising plasma glucose concentrations in IGT subjects in comparison with subjects with normal glucose tolerance. In those experiments, insulin secretion rates were normalised by the body mass index (BMI) to account for the effect of obesity. Again, the impact of such decreased sensitivity to intravenous glucose on oral glucose tolerance could only be inferred qualitatively. In particular, the phenomenon of potentiation, which in large measure depends on incretins [15, 16, 17], eluded investigation.

A detailed quantitative analysis of the principal characteristics of beta-cell function, and their relation to insulin resistance, has not been done. This study of a relatively large group of IGT and BMI-matched controls was therefore undertaken (i) to measure indices of insulin sensitivity of glucose disposal and glucose production, (ii) to derive indices of beta-cell function by using a newly developed mathematical model of glucose-stimulated insulin secretion and (iii) to quantitate the contribution of each set of indices to oral glucose tolerance.

Materials and methods


The study group included 103 subjects recruited at the Clinical Research Centre of the University of Texas Health Sciences Center at San Antonio, Tex., USA where all studies were conducted. A 75-g oral glucose tolerance test (OGTT) was carried out in each subject, who was then classified as having normal glucose tolerance (NGT, i.e., fasting glucose <6.1 mmol/l and 2-h glucose <7.8 mmol/l, n=63) or impaired glucose tolerance (IGT, i.e., fasting glucose <7.0 mmol/l and 2-h glucose=7.0–11.1 mmol/l, n=40) according to the criteria of the American Diabetes Association; seven of the IGT subjects also had impaired fasting glycaemia (IFG, i.e., fasting glucose=6.1–7.0 mmol/l); excluding these subjects from the analysis did not change any of the results. The NGT group was similar to the IGT group in sex, distribution, age, and BMI (Table 1). Similar proportions of IGT and NGT subjects were Mexican-American or Caucasian (p=ns by χ2 test). History of diabetes in first-degree relatives was obtained from all subjects, and was positive in 78% of IGT and 72% of NGT (p=ns). All study subjects had normal liver, heart, and kidney function, and none was taking any other drugs known to affect glucose tolerance. The study protocol was approved by the Institutional Review Board of the University of Texas Health Science Centre at San Antonio, and informed written consent was obtained from each subject before participation.
Table 1.

Clinical characteristics







Sex (M/W)




Age (years)




BMI (kg·m–2)




Fat mass (kg)




Fat-free mass (kg)




Waist (cm)




Waist-to-hip ratio




Serum triglycerides (mmol/l)




HDL-cholesterol (mmol/l)




Values are means±SEM; NGT = normal glucose tolerance; IGT = impaired glucose tolerance

* adjusted for sex by ANCOVA

Experimental protocol

The waist-to-hip circumference ratio (WHR) was determined by measuring the waist circumference at the narrowest part of the torso, and the hip circumference in a horizontal plane at the level of the maximal extension of the buttocks. Fat-free mass (FFM) was measured using electrical bioimpedance, and fat mass (FM) was calculated as the difference between body weight and FFM. For the OGTT, subjects were fasted overnight (after 10:00 p.m.) and the study was carried out at 8:00 a.m. on the following morning. Blood samples were collected at –30, –15, 0, 30, 60, 90, and 120 min for the measurement of plasma glucose, free fatty acids (FFA), C-peptide, and insulin concentrations. For the euglycaemic insulin clamp, subjects were admitted to the Clinical Research Centre at 7:00 a.m. following an overnight fast (after 10:00 p.m.). A polyethylene cannula was inserted into an antecubital vein for the infusion of all test substances. A second catheter was inserted retrogradely into an ipsilateral wrist vein on the dorsum of the hand for blood sampling and the hand was kept in a heated box at 65°C. A primed (740 kBq)-continuous (7.4 kBq/min) infusion of 3-[3H]glucose (DuPont-NEN, Boston, Mass., USA) was administered for 120 min. During the last 30 min of the basal equilibration period, plasma samples were taken at 5-to 10-min intervals for the determination of plasma glucose and insulin concentration and tritiated glucose specific activity. After the basal equilibration period, insulin was administered as a prime continuous infusion at the rate of 240 pmol·min–1·m–2 for 120 min. The plasma glucose concentration was clamped at approximately 5 mmol/l—with a coefficient of variation less than 5%—by a variable infusion of 20% glucose adjusted according to the negative feedback principle. Plasma samples were collected every 15 min from 0 to 90 min and every 5 to 10 min from 90 to 120 min for the determination of plasma glucose and insulin concentrations and tritiated glucose specific activity.

Analytical techniques

Plasma glucose was measured by the glucose oxidase reaction (Beckman Glucose Analyzer, Fullerton, Calif, USA). Plasma insulin and C-peptide concentrations were measured by radioimmunoassay using specific kits (Linco Research, St Louis, Mo., USA). Plasma FFA were measured spectrophotometrically (Wako, Neuss, Germany). Plasma 3-[3H]glucose radioactivity was measured in Somogyi precipitates as previously described [18].

Data analysis

Insulin-mediated glucose uptake (M value) was expressed per kg of FFM. During the baseline period, both the plasma glucose concentrations and plasma 3-[3H]glucose specific activity were stable during the last 30 min of tracer infusion in all subjects. Therefore, total endogenous glucose production (EGP) was calculated as the ratio of the 3-[3H]glucose infusion rate to the plasma 3-[3H]glucose specific activity (mean of 4–5 determinations). During the insulin clamp, total glucose rates of appearance (Ra) were calculated using Steele's equation [19, 20]. EGP was then obtained as the difference between Ra and the exogenous glucose infusion rate.

Since the fasting plasma insulin concentration is a strong inhibitory stimulus for EGP [21], and fasting hyperinsulinaemia could therefore mask hepatic insulin resistance, an index of insulin resistance of endogenous glucose production (IRGP) was calculated as the product of fasting EGP and fasting plasma insulin concentration. Experimental validation for the use of this index has been published [22]. It should be noted that, over the concentration range 36±6 to 66±12 to 132±12 pmol/l, the increment in plasma insulin is linearly related to the decline in EGP [23].

Post-hepatic insulin clearance was calculated as the ratio of the exogenous insulin infusion rate to the steady-state plasma insulin concentrations during the final 40 min of the clamp [24]. Areas under glucose (AUCG) or insulin (AUCI) concentration curves were calculated by the trapezoidal rule.


In 25 IGT and 43 NGT subjects (whose clinical characteristics were similar to those of the whole study group, data not shown), insulin secretion rates were calculated from plasma C-peptide concentrations by deconvolution [25]. Parameters of beta-cell function were derived from mathematical analysis of plasma glucose and C-peptide concentrations during the OGTT according to a previously developed model [26, 27]. The model describes the relationship between insulin secretion, S(t), and glucose concentration. Insulin secretion consists of two components, according to the equation:
$${{\rm{S}}({\rm{t}}) = {\rm{S}}_{{\rm{g}}} ({\rm{t}}) + {\rm{S}}_{{\rm{d}}} ({\rm{t}})}$$
The first component, Sg(t), represents the dependence of insulin secretion on the absolute glucose concentration (G) at any time point, and is characterised by a dose-response function, f(G), relating these variables. A characteristic parameter of the dose-response is its mean slope (in the observed glucose range), denoted as glucose sensitivity. The dose-response is modulated by a potentiation factor, P(t), which incorporates glucose-mediated and non-glucose-mediated potentiation (i.e., by non-glucose substrates, gastrointestinal hormones and neurotransmitters):
$${{\rm{Sg}}({\rm{t}}) = {\rm{P}}({\rm{t}}){\rm{f}}({\rm{G}})}$$
Potentiation is a time-dependent phenomenon [28]. The potentiation factor is therefore modelled as a positive function of time and averages one during the experiment. The potentiation parameter used for this analysis is the ratio of the potentiation factor at the end of the OGTT (100–120 min) to the one at the beginning of the OGTT (0–20 min).
The second insulin secretion component represents a dynamic dependence of insulin secretion on the rate of change of glucose concentration, expressed as
$$ {{\rm{Sd(t) = }}\left\{ {\matrix{ {{{\rm{k}}_{{\rm{d}}} {{{\rm{dG}}{\left( {\rm{t}} \right)}} \over {{\rm{dt}}}}{\rm{,}}} \hfill} & {{{{{\rm{dG}}{\left( {\rm{t}} \right)}} \over {{\rm{dt}}}}{\rm{ > 0}}} \hfill} \cr {{{\rm{0}}{\rm{,}}} \hfill} & {{{{{\rm{dG}}{\left( {\rm{t}} \right)}} \over {{\rm{dt}}}}{\rm{ \le 0}}} \hfill} \cr } } \right.} $$
This component is called the derivative component and is described by a single parameter (Kd), denoted as rate sensitivity. From these estimated model parameters [the parameters of the dose-response f(G) and the potentiation factor P(t)], total and basal insulin secretion were calculated. An empirical parameter of glucose-induced insulin release, namely, the incremental insulin/glucose concentration ratio at 30 min postglucose (or insulinogenic index, ∂I/∂G30), was also calculated.

Statistical analysis

Data are given as the mean±SE; insulin parameters, whose distribution is strongly skewed, are presented as median and interquartile range. Group values were compared by Mann-Whitney U test; adjustment for confounders was carried out with the use of ANCOVA. Categorical variables were compared by the χ2 test. Univariate associations were tested with the use of Spearman rank correlations. The contribution of multiple factors to glucose tolerance (as AUCG) was assessed by multivariate analysis; in this model, insulin sensitivity parameters were logarithmically transformed and the contribution of each independent variable was expressed as the standardized correlation coefficient. A p value of less than 0.05 was considered significant.


IGT subjects were matched to NGT subjects by sex, age, and BMI; however, they had more central fat distribution and higher serum triglyceride levels (Table 1).

Both fasting plasma glucose and insulin concentrations were slightly higher in IGT than NGT (Table 2). Fasting EGP was similar in the two groups, whereas the IRGP index, which estimates resistance to insulin of EGP, was higher in IGT than NGT. During the insulin clamp, EGP was suppressed to a similar extent in the two groups, whereas insulin-mediated glucose disposal was reduced by about 20% in IGT as compared to NGT. Post-hepatic insulin clearance was slightly depressed in IGT.
Table 2.

Metabolic data in the fasting state and during the clamp




Fasting plasma glucose (mmol/l)




Fasting plasma insulin (pmol/l)

79 [49]

52 [33]


EGP (µmol·min–1·kgFFM –1)




IRGP (mmol·min–1·kgFFM –1·pmol·l–1)




Fasting plasma FFA (mmol/l)




Clamp plasma insulin (pmol/l)




Post-hepatic insulin clearance (ml·min–1·m–2)




Clamp EGP (µmol·min–1·kgFFM –1)




M (µmol·min–1·kgFFM –1)




M/I (nmol·min–1·kgFFM –1·pmol·l–1)




EGP = endogenous glucose production; IRGP = insulin resistance of EGP (EGPxfasting insulin); M = insulin-mediated glucose disposal rate on the insulin clamp; M/I = M divided by the steady-state plasma insulin concentration during the clamp

During the OGTT, the total plasma insulin response, both absolute and incremental, was marginally higher in IGT than NGT (Table 3, Fig. 1), whereas the insulinogenic index (∂I/∂G30) was 15% lower in IGT.
Table 3.

Metabolic data during the OGTT




AUCG (mol/l·h)




AUCI (nmol/l·h)

64 [47]

47 [38]


∂I/∂G30 (pmol/mmol)

110 [103]

130 [188]


Mean OGTT plasma FFA (mmol/l)




AUCG = area under the glucose response to oral glucose; AUCI = area under the insulin response to oral glucose; ∂I/∂G30 = ratio of plasma insulin to plasma glucose increment at 30′ during the OGTT

Fig. 1.

Plasma glucose, insulin, and C-peptide concentrations in response to oral glucose in IGT and NGT subjects

On the pooled data from all subjects as well as in the NGT group alone, the two indices of insulin resistance were interrelated (Table 4). It is noteworthy that IGT and NGT subjects fell on the same regression line for this relationship (Fig. 2).
Table 4.

Interrelationships among indices of insulin secretion, insulin sensitivity, and glucose tolerance

Gluc. Sens.


Rate sens.




Glucose sensitivity














Rate sensitivity








0.38 (0.41)


0.54 (0.55)






0.31 (0.36)


–0.46 (–0.48)





–0.25 (–0.25)


0.28 (0.40)

–0.58 (–0.50)



–0.67 (–0.67)

–0.23 (ns)


–0.35 (–0.38)


0.27 (ns)

Entries are Spearman rank correlation coefficients significant at p≤0.05 for the whole dataset. Numbers in parenthesis are Spearman rank correlation coefficients for the NGT group alone

Fig. 2.

Reciprocal relationship between insulin resistance of glucose production (IRGP) and insulin sensitivity of glucose disposal (M). The regression lines for NGT and IGT subjects (not shown) are not statistically different

Basal rates of insulin secretion, as reconstructed by C-peptide deconvolution, were modestly increased in IGT as compared to NGT, whereas total postglucose insulin output was similar (Fig. 3, Table 5). Model-derived indices of beta-cell function differed between the two groups. Thus, glucose sensitivity was markedly impaired, and the potentiation factor was lower, in IGT than NGT (Fig. 4, Table 5), whereas rate sensitivity was slightly lower, but not significantly, in IGT than NGT. Of note is that the model derived dose-response functions (top panel in Fig. 4) essentially reproduced the positive relationship between insulin secretory rates and plasma glucose concentrations as measured at each time-point during the OGTT in the two groups (inset in Fig. 4,) i.e., model-dependent and empirical estimates were coherent.
Fig. 3.

Insulin secretion rates in the basal state and following glucose ingestion in IGT and NGT subjects

Table 5.

Model-derived insulin secretory parameters




Basal insulin secretion (pmol·min–1·m–2)

113 [49]

84 [52]


Total insulin secretion (nmol·m–2)

46 [17]

43 [17]


Glucose sensitivity (pmol·min–1·m–2·mM–1)

53 [36]

102 [123]


Rate sensitivity (nmol·m–2·mM–1)

1.15 [1.15]

1.38 [1.28]


Potentiation factor (fold)

1.6 [0.8]

2.0 [1.5]


Values are median [interquartile range]

Fig. 4.

Model-derived dose-response curve of glucose-induced insulin secretion (top), and time-course of potentiation factor during glucose loading (bottom) in IGT and NGT subjects. The inset shows the plot of insulin secretory rates vs plasma glucose concentrations as measured at each time-point during the OGGT in the two groups. Note the overall similarity with the model-derived dose-response functions

When assessing the relation of insulin sensitivity to insulin release, significant power relationships were found between the M value and fasting plasma insulin concentration, with IGT and NGT subjects falling along the same regression lines (Fig. 5); the same pattern emerged when fasting insulin concentrations were replaced by basal insulin secretion rates (data not shown). In contrast, ∂I/∂G30 was related to insulin sensitivity in both NGT and IGT, but the regression line was shifted downward in the latter (Fig. 5). Of the model-derived secretory parameters, neither glucose sensitivity nor rate sensitivity were related to either of the insulin sensitivity indices, whereas the potentiation ratio was related to both of them (Table 4).
Fig. 5.

Power relationship between insulin sensitivity of glucose disposal (M) and fasting plasma insulin concentrations (top) or ∂I/∂G30 (ratio of insulin to glucose increment above baseline at 30 min postglucose) (bottom). The two regression lines in the lower graph are significantly different from one another

Finally, glucose tolerance (as AUCG) was reciprocally related to both glucose sensitivity (Fig. 6) and the potentiation ratio.
Fig. 6.

Power relationship between glucose sensitivity of insulin secretion (as the mean slope of the dose-response for glucose-induced insulin secretion, cf. Fig. 4) and glucose area in IGT and NGT subjects. The fitting function is y=57·x–3.5. Obesity was defined as a BMI ≥30 kg·m–2


The novel information about the IGT state provided by these studies emerges from the model analysis of beta-cell function. With regard to this, it should be emphasized that the model applied here, if mathematically sophisticated, is physiologically basic in that it incorporates only the three main characteristics of beta-cell function: glucose sensitivity, rate sensitivity, and potentiation. The determination of physiologically meaningful beta-cell function parameters from a 5-sample OGTT is possible because the model describes characteristics of insulin secretion that are apparent also in a simple test. These characteristics are: the anticipated rise in insulin secretion (accounted for by the rate sensitivity parameter), the direct dependence of insulin secretion on glucose concentration (represented by the dose-response curve), and the time-dependence of secretion (higher secretion at the end than at the beginning of the OGTT for similar glucose concentrations) described as a potentiation factor. Each of these three features has been amply described in the isolated perfused/perifused pancreas or in cultured beta cells [15]. In vivo, glucose sensitivity and rate sensitivity have been modelled previously [29, 30], whereas modelling of potentiation has been limited to intravenous glucose tests [31]. Potentiation, however, is an important phenomenon, encompassing not only glucose potentiation [28, 31] but also incretin effects [15, 32, 33]. Originally developed for multiple meal tests [27], our model has proved applicable to the oral glucose test [26].

In absolute terms, rates of insulin secretion, both in the basal state and following oral glucose, were, if anything, higher in IGT than NGT subjects, confirming a wealth of previous reports based on plasma insulin concentrations [34, 35, 36]. Incidentally, we found that, due to the reduced metabolic clearance rate of insulin, postglucose hyperinsulinaemia was more pronounced (+30% in IGT vs NGT) than true insulin hypersecretion (+15%). More importantly, when viewed in the context of the respective plasma glucose concentrations (i.e., as glucose sensitivity), insulin secretion was markedly (by ~50%) impaired in IGT as compared to NGT, considerably more than the plasma insulin profile would suggest. Thus, at a plasma glucose concentration of 7.6 mmol/l (i.e., the average post-OGTT glucose concentration for all subjects), the insulin secretory rate in IGT subjects was half that of the NGT controls. Furthermore, positive modulation of glucose sensitivity by potentiation was reduced in IGT by 20% on average. Thus, the insulin secretory defect that characterizes IGT (independently of obesity) is a composite of reduced glucose sensitivity and impaired potentiation, whereas rate sensitivity, or the ability of the beta cell to anticipate the effect of glucose, is mostly preserved. Clearly, the multi-parametric characterization of insulin secretion is model-dependent, and therefore subject to the limitations discussed previously [26, 27]. However, it should be noted that the key finding, i.e., impaired glucose sensitivity in IGT, is confirmed by an empirical index calculated as the ratio of the integral of the insulin secretion increment above baseline to the corresponding glucose increment. This index was 141 [116] pmol·min–1·m–2·mM–1 in NGT vs 92 [48] in IGT (p<0.001).

In general, it seems important to discriminate tonic from phasic components of beta-cell function in order to identify sites of regulation and potential abnormalities in different disease states. The picture emerging from our analysis is that the baseline insulin secretory rate is the closest physiological correlate of insulin sensitivity of glucose disposal. The relationship between basal insulin secretion and insulin sensitivity predicts that halving M from 40 to 20 µmol·min–1·kgFFM –1 yields a 1.5-fold increase in baseline insulin output (from 69 to 103 pmol·min–1·m–2). In our non-diabetic cohort, the data of the IGT subjects fell on the same regression line as those of the NGT subjects, implying that adaptation of the basal secretory tone to the prevailing insulin resistance is not markedly impaired in IGT. In contrast, the phasic components of beta-cell function, i.e., glucose sensitivity and rate sensitivity, were both largely independent of insulin sensitivity of glucose disposal, suggesting that the ability of beta cells to respond to acute glucose changes after oral administration is controlled by a separate set of mechanisms not strongly dependent on insulin action outside the islets. These results lead to the construct that beta-cell adaptation to insulin resistance mainly involves the tonic component of beta-cell function, and that this compensation is largely preserved in IGT. The signals mediating this adaptation remain elusive, although small, chronic increases in circulating plasma glucose and/or FFA concentrations are plausible candidates.

These results can be contrasted with the information derived from the use of an empirical index of acute beta-cell response, the insulinogenic index (∂I/∂G30). In our dataset, ∂I/∂G30 was inversely related to insulin sensitivity of glucose disposal, as has been shown to be true of the acute insulin response to intravenous glucose (AIR) [37]. ∂I/∂G30 was found to be about 15% lower in IGT compared to NGT, suggesting a modest impairment in early-phase insulin release. However, ∂I/∂G30 seems to be a compound index of rate sensitivity (given the early time after the glucose challenge) and glucose sensitivity, and its reduction in IGT does not discriminate between loss of glucose sensitivity compared to changes in rate sensitivity. Furthermore, the insulinogenic index explained only 17% of the variability of glucose area in the whole dataset.

By contrast, beta-cell glucose sensitivity was largely independent of insulin sensitivity, clearly separated the IGT from the NGT subjects, and singly explained almost 50% of the variability of glucose area in this dataset. The power function in Fig. 5 indicates that a 10-fold decrement in glucose sensitivity (from 250 to 25 pmol·min–1·m–2·mM–1) is associated with a doubling of glucose area (from 0.64 to 1.28 mol/l·h). It is interesting to speculate on the meaning of this relationship. If glucose sensitivity of insulin secretion is viewed as a mechanism for preserving glucose tolerance, then the beta cell is endowed with a very large functional reserve, which in NGT subjects covers an about 100-fold span. This is reminiscent of the wide range of insulin sensitivity of glucose disposal seen in normal, lean individuals [38]. Viewed in reverse, the strong relationship suggests that small, chronic changes in glucose concentrations could profoundly affect the glucose sensitivity of beta cells. This phenomenon, if proven prospectively, would identify phasic beta-cell responses as a sensitive target of glucose toxicity. Of course, a cross-sectional study such as this cannot establish the physiological plausibility of this dual link. Furthermore, 90% of the subjects in our study were obese (BMI >25 kg/m2), which prevents ready extrapolation of the current findings to a lean population.

The finding that rate sensitivity was not impaired in IGT is at variance with the prevailing view [39], that early-phase insulin release is impaired in IGT. This discrepancy, however, stems in part from the uncertain definition of 'early phase' during a slow stimulus such as oral glucose compared with the first phase of insulin release that is seen in response to an intravenous glucose bolus as previously discussed [26, 27]. In fact, in the subgroup in which the model parameters were calculated, both rate sensitivity and ∂I/∂G30 were slightly decreased in IGT (by 19% and 23%, respectively), but the group difference did not reach statistical significance for either parameter. According to our analysis, the plasma insulin profile in IGT results from a combination of reduced glucose sensitivity and impaired potentiation of insulin secretion. The latter defect might reflect insufficient release of gastrointestinal hormones, or resistance to their secretagogic effects. With regard to this, we have previously described a close relationship between the time-course and height of the potentiation factor and circulating levels of GIP [26].

This study confirms and extends previous findings relating to IGT as an insulin-resistant state. Our IGT subjects had a moderate degree of insulin resistance of glucose disposal in comparison with a large group of NGT control subjects matched for age and BMI. In addition, these IGT subjects had insulin resistance of fasting EGP, in a degree that was well correlated with the degree of insulin resistance of glucose disposal. EGP during the OGTT was not measured in our experiments, but has been previously shown to be inappropriately increased in IGT as compared to normal subjects [40]. Furthermore, we have previously shown that EGP during glucose loading is proportional to fasting EGP, both in non-diabetic and Type 2 diabetic subjects [41, 42]. These findings stand in contrast to the conclusion reached by another study [5], in which IGT subjects had normal hepatic sensitivity to insulin. This conclusion, however, was based on the finding of similarly suppressed EGP during a euglycaemic clamp using exogenous insulin infusion rates ( equivalent to those used in this study) which are maximally suppressive for EGP and, therefore, cannot detect differences in EGP. Thus, we can assume that in our IGT subjects EGP was increased when viewed in the context of the prevailing insulinaemia also during the OGTT, in approximate proportion to the insulin resistance of EGP in the fasting state (i.e., IRGP). This would adequately explain the positive association between IRGP and the glucose area in the whole cohort. Finally, plasma FFA concentrations during the OGTT remained higher in IGT than NGT subjects; coupled with the higher plasma insulin levels, this finding stands for an impaired ability of insulin to suppress lipolysis in IGT.

In summary, in the IGT state skeletal muscle, adipose tissue and liver share in vivo insulin insensitivity, with little evidence of tissue selectivity. It should be added that, despite the similar BMI, our IGT group had a higher waist circumference and waist-to-hip ratio than the NGT controls, suggesting fat accumulation in the abdominal region. Given the strong association of visceral fat accumulation with insulin resistance [43], the insulin resistance of IGT might be linked with the abdominal obesity rather than result from the glucose intolerance itself (i.e., glucose toxicity). However, it cannot be excluded that visceral fat excess and glucose toxicity independently contribute to the impairment of glucose tolerance.

In conclusion, glucose sensitivity of beta-cell insulin response is the single strongest determinant of oral glucose tolerance. Multiple defects in insulin action and beta-cell function characterise the IGT state, but the dominant role in its pathogenesis is to be ascribed to the impaired ability of pancreatic islets to sense glucose as a stimulus for appropriate insulin release.



The authors wish to thank our nurses (M. Ortiz, D. Frantz, S. Mejorado, J. Shapiro, J. Kinkaid, J. King, N. Diaz, P. Wolf) for their assistance in carrying out the insulin clamp studies. This work was supported by an EFSD-Novo Nordisk Type 2 Programme Focused Research Grant, NIH grant DK24092, NIH GCRC grant MOI-RR-01346, a VA merit Award, funds from the VA Research Foundation, and funds from the Italian Ministry of University and Scientific Research (MURST prot. 2001065883_001).


  1. 1.
    Eriksson J, Franssila-Kallunki A, Ekstrand A et al. (1989) Early metabolic defects in persons at increased risk for non-insulin-dependent diabetes mellitus. N Engl J Med 321:337–343PubMedGoogle Scholar
  2. 2.
    Saccà L, Orofino G, Petrone A, Vigorito C (1984) Differential roles of splanchnic and peripheral tissues in the pathogenesis of impaired glucose tolerance. J Clin Invest 73:1683–1687PubMedGoogle Scholar
  3. 3.
    Bogardus C, Lillioja S, Howard B, Reaven G, Mott D (1984) Relationships between insulin secretion, insulin action, and fasting plasma glucose concentration in nondiabetic and noninsulin-dependent diabetic subjects. J Clin Invest 74:1238–1246PubMedGoogle Scholar
  4. 4.
    Reaven GM, Hollenbeck CB, Chen YD (1989) Relationship between glucose tolerance, insulin secretion, and insulin action in non-obese individuals with varying degrees of glucose tolerance. Diabetologia 32:52–55PubMedGoogle Scholar
  5. 5.
    Berrish TS, Hetherington CS, Alberti KG, Walker M (1995) Peripheral and hepatic insulin sensitivity in subjects with impaired glucose tolerance. Diabetologia 38:699–704CrossRefPubMedGoogle Scholar
  6. 6.
    Cerasi E, Luft R (1967) Further studies on healthy subjects with low and high insulin response to glucose infusion. Acta Endocrinol (Copenh) 55:305–329Google Scholar
  7. 7.
    Efendic S, Cerasi E, Luft R (1974) Quantitative study on the potentiating effect of arginine on glucose-induced insulin response in healthy, prediabetic, and diabetic subjects. Diabetes 23:161–171PubMedGoogle Scholar
  8. 8.
    Pimenta W, Mitrakou A, Jensen T, Yki-Jarvinen H, Daily G, Gerich J (1996) Insulin secretion and insulin sensitivity in people with impaired glucose tolerance. Diabet Med 13:S33–S36PubMedGoogle Scholar
  9. 9.
    Larsson H, Ahren B (1996) Failure to adequately adapt reduced insulin sensitivity with increased insulin secretion in women with impaired glucose tolerance. Diabetologia 39:1099–1107CrossRefPubMedGoogle Scholar
  10. 10.
    Van Haeften TW, Pimenta W, Mitrakou A et al. (2002) Disturbances in beta-Cell Function in Impaired Fasting Glycemia. Diabetes 51 Suppl 1:S265–S270Google Scholar
  11. 11.
    Ehrmann DA, Breda E, Cavaghan MK et al. (2002) Insulin secretory responses to rising and falling glucose concentrations are delayed in subjects with impaired glucose tolerance. Diabetologia 45:509–517CrossRefGoogle Scholar
  12. 12.
    Ferrannini E (1998) Insulin resistance versus insulin deficiency in non-insulin-dependent diabetes mellitus: problems and prospects. Endocr Rev 19:477–490PubMedGoogle Scholar
  13. 13.
    Kahn SE (2003) The relative contributions of insulin resistance and beta-cell dysfunction to the pathophysiology of Type 2 diabetes. Diabetologia 46:3–19PubMedGoogle Scholar
  14. 14.
    Polonsky KS (1999) Evolution of beta-cell dysfunction in impaired glucose tolerance and diabetes. Exp Clin Endocrinol Diabetes 107 (Suppl 4):S124–S127PubMedGoogle Scholar
  15. 15.
    Cook D, Taborsky G (1997) ß-cell function and insulin secretion. In: Sherwin R (ed.) Ellenberg and Rifkin's diabetes mellitus: theory and practice, 5th edn. Appleton and Lange, Stamford, pp 49–73Google Scholar
  16. 16.
    Sandberg E, Ahren B, Tendler D, Carlquist M, Efendic S (1986) Potentiation of glucose-induced insulin secretion in the perfused rat pancreas by porcine GIP (gastric inhibitory polypeptide, bovine GIP, and bovine GIP(1–39). Acta Physiol Scand 127:323–326PubMedGoogle Scholar
  17. 17.
    Tseng CC, Zhang XY, Wolfe MM (1999) Effect of GIP and GLP-1 antagonists on insulin release in the rat. Am J Physiol 276:E1049–E1054PubMedGoogle Scholar
  18. 18.
    Ferrannini E, Smith J, Cobelli C, Toffolo G, Pilo A (1985) Effect of insulin on the distribution and disposal of glucose in man. J Clin Invest 76:367–374Google Scholar
  19. 19.
    Steele R (1959) Influences of glucose loading and of injected insulin on hepatic glucose output. Ann NY Acad Sci 82:420–430Google Scholar
  20. 20.
    Steele R, Wall JS, De Bodo RC, Altszuler N (1956) Measurement of size and turnover rate of body glucose pool by the isotope dilution method. Am J Physiol 187:15–24Google Scholar
  21. 21.
    Sindelar DK, Chu CA, Venson P, Donahue EP, Neal DW, Cherrington AD (1998) Basal hepatic glucose production is regulated by the portal vein insulin concentration. Diabetes 47:523–529PubMedGoogle Scholar
  22. 22.
    DeFronzo RA, Ferrannini E, Simonson DC (1989) Fasting hyperglycemia in Non-Insulin-Dependent Diabetes Mellitus: contributions of excessive hepatic glucose production and impaired tissue glucose uptake. Metabolism 38:387–395PubMedGoogle Scholar
  23. 23.
    Groop LC, Bonadonna RC, DelPrato S et al. (1989) Glucose and free fatty acid metabolism in non-insulin-dependent diabetes mellitus. J Clin Invest 84:205–213PubMedGoogle Scholar
  24. 24.
    Ferrannini E, Cobelli C (1987) The kinetics of insulin in man. Role of the liver. Diabetes Metab Rev 3:365–397PubMedGoogle Scholar
  25. 25.
    Van Cauter E, Mestrez F, Sturis J, Polonsky KS (1992) Estimation of insulin secretion rates from C-peptide levels. Comparison of individual and standard kinetic parameters for C-peptide clearance. Diabetes 41:368–377PubMedGoogle Scholar
  26. 26.
    Mari A, Schmitz O, Gastaldelli A, Oestergaard T, Nyholm B, Ferrannini E (2002) Meal and oral glucose tests for the assessment of β-cell function: modeling analysis in normal subjects. Am J Physiol Endocrinol Metab 283:E1159–E1166PubMedGoogle Scholar
  27. 27.
    Mari A, Tura A, Gastaldelli A, Ferrannini E (2002) Assessing insulin secretion by modeling in multiple-meal tests: role of potentiation. Diabetes 51 [Suppl 1]:S221–S226Google Scholar
  28. 28.
    Cerasi E (1981) Differential actions of glucose on insulin release: re-evaluation of a mathematical model. In: Bergman R (ed.) Quantitative physiology and mathematical modelling. Wiley, Chichester, pp 3–22Google Scholar
  29. 29.
    Licko V (1973) Threshold secretory mechanism: a model of derivative element in biological control. Bull Math Biol 35:51–58PubMedGoogle Scholar
  30. 30.
    Toffolo G, De Grandi F, Cobelli C (1995) Estimation of beta-cell sensitivity from intravenous glucose tolerance test C-peptide data. Knowledge of the kinetics avoids errors in modeling the secretion. Diabetes 44:845–854PubMedGoogle Scholar
  31. 31.
    Cerasi E (1975) Potentiation of insulin release by glucose in man. Quantitative analysis of the enhancement of glucose-induced insulin secretion by pretreatment with glucose in normal subject. Acta Endocrinol (Copenh) 79:483–501Google Scholar
  32. 32.
    Tseng CC, Kieffer TJ, Jarboe LA, Usdin TB, Wolfe MM (1996) Postprandial stimulation of insulin release by glucose-dependent insulinotropic polypeptide (GIP). Effect of a specific glucose-dependent insulinotropic polypeptide receptor antagonist in the rat. J Clin Invest 98:2440–2445PubMedGoogle Scholar
  33. 33.
    Willms B, Werner J, Holst JJ, Orskov C, Creutzfeldt W, Nauck MA (1996) Gastric emptying, glucose responses, and insulin secretion after a liquid test meal: effects of exogenous glucagon-like peptide-1 (GLP-1)-(7–36) amide in type 2 (noninsulin-dependent) diabetic patients. J Clin Endocrinol Metab 81:327–332PubMedGoogle Scholar
  34. 34.
    Zimmet P, Whitehouse S, Alford F, Chisholm D (1978) The relationship of insulin response to a glucose stimulus over a wide range of glucose tolerance. Diabetologia 15:23–27PubMedGoogle Scholar
  35. 35.
    Berntorp K, Lindgarde F, Malmquist J (1984) High and low insulin responders: relations to oral glucose tolerance, insulin secretion and physical fitness. Acta Med Scand 216:111–117PubMedGoogle Scholar
  36. 36.
    Cretti A, Lehtovirta M, Bonora E et al. (2001) Assessment of beta-cell function during the oral glucose tolerance test by a minimal model of insulin secretion. Eur J Clin Invest 31:405–416PubMedGoogle Scholar
  37. 37.
    Kahn S, Prigeon R, Mcculloch D et al. (1993) Quantification of the relationship between insulin sensitivity and beta-cell function in human subjects. Evidence for a hyperbolic function. Diabetes 42:1663–1672PubMedGoogle Scholar
  38. 38.
    Ferrannini E, Natali A, Bell P, Cavallo-Perin P, Lalic N, Mingrone G (1997) Insulin resistance and hypersecretion in obesity. J Clin Invest 100:1166–1173PubMedGoogle Scholar
  39. 39.
    Gerich JE (1998) The genetic basis of type 2 diabetes mellitus: impaired insulin secretion versus impaired insulin sensitivity. Endocr Rev 19:491–503PubMedGoogle Scholar
  40. 40.
    Mitrakou A, Kelley D, Mokan M et al. (1992) Role of reduced suppression of glucose production and diminished early insulin release in impaired glucose tolerance. N Engl J Med 326:22–29PubMedGoogle Scholar
  41. 41.
    Ferrannini E, Bjorkman O, Reichard GA, Jr et al. (1985) The disposal of an oral glucose load in healthy subjects. Diabetes 34:580–588PubMedGoogle Scholar
  42. 42.
    Ferrannini E, Simonson DC, Katz LD et al. (1988) The disposal of an oral glucose load in patients with non-insulin-dependent diabetes. Metabolism 37:79–85PubMedGoogle Scholar
  43. 43.
    Bonora E, Del Prato S, Bonadonna RC et al. (1992) Total body fat content and fat topography are associated differently with in vivo glucose metabolism in nonobese and obese nondiabetic women. Diabetes 41:1151–1159PubMedGoogle Scholar

Copyright information

© Springer-Verlag 2003

Authors and Affiliations

  • E. Ferrannini
    • 1
    • 2
  • A. Gastaldelli
    • 1
  • Y. Miyazaki
    • 2
  • M. Matsuda
    • 2
  • M. Pettiti
    • 1
  • A. Natali
    • 1
  • A. Mari
    • 3
  • R. A. DeFronzo
    • 2
  1. 1.Metabolism Unit, C.N.R. Institute of Clinical Physiology and Department of Internal MedicineUniversity of Pisa School of MedicinePisaItaly
  2. 2.Diabetes DivisionUniversity of Texas Health Sciences Center at San AntonioSan AntonioUSA
  3. 3.C.N.R. Institute of Biomedical EngineeringPadovaItaly

Personalised recommendations