Independent measures of insulin secretion and insulin sensitivity during the same test: the glucagon insulin tolerance test

Transcription

1 Original Article doi: /j x Independent measures of insulin secretion and insulin sensitivity during the same test: the glucagon insulin tolerance test M. Dorkhan 1, D. Tripathy 2, G. Malm 3, H. Asgharian 4, T. Tuomi 5,6 & L. Groop 1 From the 1 Division of Diabetes & Endocrinology, Department of Clinical Sciences, Lund University, Malmö University Hospital, Malmö, Sweden, 2 Division of Diabetes, University of Texas Health Sciences Center at San Antonio, San Antonio, TX, USA, 3 Primary Health Care, Skåne Southwest, Health Care Centre Eden, Malmö; 4 Department of Economics, Lund University, Lund; Sweden; 5 Department of Medicine, Helsinki University Central Hospital; and 6 Folkhalsan Genetic Institute, Folkhalsan Research Center and Research Program for Molecular Medicine, University of Helsinki; Helsinki, Finland Abstract. Dorkhan M, Tripathy D, Malm G, Asgharian H, Tuomi T, Groop L (Malmö University Hospital, Malmö, Sweden, University of Texas Health Sciences Center at San Antonio, San Antonio, TX, USA, Health Care Centre Eden, Malmö; Lund University, Lund; Sweden; Helsinki University Central Hospital and Folkhalsan Research Center University of Helsinki, Helsinki; Finland). Independent measures of insulin secretion and insulin sensitivity during the same test: the glucagon insulin tolerance test. J Intern Med 2008; 264: Background. To validate a test for independent assessment of insulin secretion and insulin sensitivity during the same occasion for metabolic studies in clinical practice, i.e. combined glucagon-stimulated C-peptide test and insulin tolerance test (GITT). Subjects and methods. We measured C-peptide response to 0.5 mg of intravenous glucagon followed 30 min later by administration of 0.05 U kg )1 insulin (insulin tolerance test, ITT). Ten subjects with normal glucose tolerance participated on different days in an ITT, glucagon-c-peptide test, ITT followed by glucagon-c-peptide test and glucagon-c-peptide test followed by ITT to establish whether and how the tests could be combined. The test was then repeated in nine patients with type 2 diabetes to investigate its reproducibility. In 20 subjects with varying degrees of glucose tolerance, the test was compared with the Botnia clamp (an intravenous glucose tolerance test combined with a euglycaemic hyperinsulinemic clamp). Results. When ITT preceded the glucagon test, C-peptide response was blunted. Therefore, we first administered glucagon and then insulin (GITT). The K ITT from the GITT was reproducible (CV = 13 %) and correlated strongly with the glucose disposal rate from the Botnia clamp (r = 0.87, r 2 = 0.75, P < 0.001). The C-peptide response to glucagon was reproducible (CV = 13 %). The disposition index, providing a measure of beta-cell function adjusted for insulin sensitivity, calculated from the GITT showed good discrimination between individuals with varying degrees of glucose tolerance. Conclusions. The GITT provides simple, reproducible and independent estimates of insulin sensitivity and secretion on the same occasion for metabolic studies in individuals with normal and abnormal glucose tolerance. Keywords: disposition index, euglycaemic clamp, insulin secretion, insulin sensitivity, insulin tolerance test, type 2 diabetes. 62 ª 2008 Blackwell Publishing Ltd

2 Introduction Assessment of the degree of insulin resistance and insulin deficiency is often required in metabolic studies of the prediabetic and diabetic state [1 3]. Given that insulin secretion is highly dependent upon the degree of insulin sensitivity [4, 5], these parameters should be measured on the same occasion. The gold standard methods for assessing insulin sensitivity (euglycaemic hyperinsulinaemic clamp) and beta-cell function (hyperglycaemic clamp) are time-consuming and labour-intensive and difficult to apply in largescale studies. Consequently, there is a need for simpler methods. The frequently sampled intravenous glucose tolerance test provides useful estimates of both insulin secretion and insulin sensitivity with the limitation that insulin sensitivity is derived from the insulin values during the test and thereby vulnerable to errors during conditions of impaired beta-cell function [6 8]. One way to circumvent the problem is to add a bolus of insulin or tolbutamide, which results in somewhat poorer reproducibility [9]. Another common approach is to estimate insulin sensitivity from oral glucose tolerance test (OGTT). The OGTT has relatively poor reproducibility and estimates of insulin sensitivity and beta-cell function from OGTT do not seem to reproduce the hyperbolic relation [10] which is required for calculating the disposition index (DI) [11]. Homeostasis model assessment (HOMA) uses the fasting values of insulin and glucose for estimating insulin sensitivity and beta-cell function [12]. The test therefore describes better basal than glucose-stimulated insulin sensitivity in the postprandial state. In addition, the HOMA insulin resistance index does not perform well in subjects with impaired glucose tolerance or impaired fasting glucose (IFG) [13] and the HOMA beta-cell index is a very crude measure of beta-cell function. To allow the assessment of insulin secretion and insulin sensitivity on the same day for phenotyping individuals participating in the Botnia family study [14], Tripathy et al. [15] developed the Botnia clamp which combines a regular intravenous glucose tolerance test (IVGTT) followed 60 min later by a euglycaemic hyperinsulinaemic (45 mu m )2 ) clamp for another 120 min. Although the Botnia clamp provides reproducible estimates of insulin secretion and action in prediabetic individuals with preserved first phase insulin secretion (FPIR), it does not perform as well for diabetic patients who lack FPIR. Common to tests which rely on stimulation of endogenous insulin secretion is that they may not perform well in individuals with impaired insulin secretion or the insulin-treated patients. There is still in need for a simple test, which can be applied to diabetic patients on a larger scale. In the current study, we evaluated how a combined glucagon-stimulated C-peptide test [16, 17] and insulin tolerance test (ITT) [18] perform in estimating insulin secretion and insulin sensitivity. The tests were performed for the same subjects both on separate days and during the same day. The sequence of the glucagon C-peptide test followed 30 min later by the ITT (GITT) was compared with the Botnia clamp, and its reproducibility was evaluated by repeating the test after one week. The calculated DI, which provides a measure of beta-cell function adjusted for insulin sensitivity, was evaluated for its capacity to discriminate between individuals with varying degrees of glucose tolerance. Research design and methods Subjects The studies were conducted at the Clinical Research Center at the Department of Endocrinology, Malmö University Hospital, Malmö, Sweden and at one of the Botnia centres (Närpes, Finland). Informed consent was obtained from all subjects and the local ethics committees approved the studies. Subjects were classified into different stages of glucose tolerance according to The World Health Organization (WHO) 1999 criteria [19]. ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264;

3 Protocol 1, validation of the combined glucagonstimulated C-peptide and insulin tolerance test. Ten subjects with normal glucose tolerance (NGT) underwent in random order: glucagon-stimulated C-peptide test, ITT, ITT followed by the glucagonstimulated C-peptide test and GITT. These tests were carried out on four separate days at least 1 week apart. The number of subjects participitating in different tests is shown in Table 1. Protocol 2, reproducibility of GITT. Nine patients with type 2 diabetes (T2D) participated in GITT twice at 1-week intervals. Protocol 3, comparison of GITT and the Botnia clamp. Twenty subjects (10 T2D, two IFG and eight NGT) participated in a Botnia clamp and a GITT 1 week apart. Protocol 4, application of GITT for the assessment of beta-cell function adjusted for insulin sensitivity in individuals with varying degrees of glucose tolerance. Eight subjects with NGT were compared with six drug-naïve patients with T2D and 11 patients with T2D and secondary drug failure, defined as HbA1c >6.5% in spite of treatment with metformin and an insulin secretagogue in doses >50% of maximum recommended doses. Protocol 5, reduced sampling test. In 17 patients with T2D, the K ITT was calculated using all timepoint samples and also using only a reduced number of samples during the GITT in order to determine if the number of samples and the duration of the test could be decreased. Methods Botnia clamp. This test was designed to obtain independent measures of insulin secretion and insulin sensitivity from the same test. In brief, 0.3 g kg )1 body weight of a 20% glucose solution was given at time 0. Blood samples for the measurement of serum insulin and plasma glucose were obtained at )10, 0, 2, 4, 6, 8, 10, 20, 40, 50, 60, 120 and 180 min. The incremental trapezoidal area during the first 10 min of the test was called FPIR. After 60 min, a priming dose of insulin was given followed by an infusion (infusion rate 45 mu m )2 ) of short-acting human insulin (Actrapid; Novo Nordisk, Gentofte, Denmark) and continued for 120 min. Blood samples for the measurement of plasma glucose were obtained at 5-min intervals throughout the clamp. A variable infusion of 20% glucose was started to maintain plasma glucose concentration unchanged at 5.5 mmol L )1 with a coefficient of variation (CV) of 6%. Insulin sensitivity (M-value) was calculated from the glucose infusion rates during the last 60 min of the euglycaemic clamp and expressed as mg kg )1 FFM min [15]. Glucagon-stimulated C-peptide test. Following basal sampling ()5, 0 min), 0.5 mg of intravenous glucagon was given at time 0, and plasma glucose, serum C-peptide and insulin values were measured at )5, 0 and 6 min. ITT. Short-acting insulin (0.05 U kg )1 of Actrapid 100 IU ml )1 ; Novo Nordisk, Gentofte), diluted to 10 IU ml )1 with saline was given at time 0 and the plasma glucose concentrations measured at 0, 1, 3, 5, 7, 9, 11, 13, 15 and 20 min. ITT followed by a glucagon-c-peptide test on the same day. The ITT was performed as described above, but at 20 min the intravenous injection of glucagon was given. Blood samples for the measurement of plasma glucose were taken at )5, 0, 1, 3, 5, 7, 9, 11, 13, 15 and 20 min and for serum C-peptide and insulin at 0, 20 and 26 min. Glucagon-stimulated C-peptide test followed by ITT on the same day. A bolus of 0.5 mg glucagon was given intravenously at time 0 and a blood sample for the measurement of plasma glucose and serum C- peptide and insulin was taken 6 min later. At 30 min, a bolus injection of 0.05 U kg )1 of Actrapid insulin was given and blood samples for the measurement of plasma glucose were taken at 30, 31, 33, 35, 37, 39, 41, 43, 45, 50 and 60 min (Fig. 1). This test is a combination of the assessment of C-peptide response to intravenous glucagon followed by an ITT. The incremental C-peptide response at 6 min was used as a 64 ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264; 62 71

4 Table 1 Individual characteristics and glucagon insulin tolerance test (GITT) and Botnia clamp indices from protocols 1 to 5 Protocol 1: validity tests Glucose Test n (M F) tolerance Age (years) BMI (kg m )2 ) Glucose (mmol L )1 ) Insulin (mu L )1 ) Basal C-peptide (nmol L )1 ) a 6-min response (nmol L )1 ) KITT c M-value (mg kg )1 (% min )1 b ) DI FFM min) G 10 (5 5) NGT 45 ± ± ± ± 0.17 ITT 9 (5 4) ± 0.34 ITT G 8 (4 4) 4.4 (1.5) 5.5 (85) 0.28 ± 0.09 (0.15) )0.05 ± ± 0.36 G ITT 9 (5 4) 4.4 (6.6) ± ± ± CV 13% 13% Protocol 2: reproducibility tests GITT 1 9 T2D 62 ± 7 32 ± ± ± ± GITT ± ± ± CV 19% 23% 22% Protocol 3: GITT versus Botnia clamp 8(2 6) NGT 43 ± ± (5.2) 0.60 ± ± ± ± 2.0 2(2 0) IFG 55 ± ± (5.7) 0.81 ± ± ± ± (7 3) T2D 58 ± 8 30 ± (10) 0.84 ± ± ± ± 3.2 Protocol 4: application of GITT 8(2 6) NGT 43 ± ± ± ± ± (5 1) Drug naïve 56 ± 8 28 ± ± ± ± (6 5) Secondary 58 ± 8 32 ± ± ± ± drug failure Protocol 5: reduced sampling test 0 30 min 17 (11 6) T2D 60.0 ± 9 29 ± ± ± ± , 5, 11 and 1.18 ± min CV 11% G, glucagon test; ITT, insulin tolerance test; CV, coefficient of variation; NGT, normal glucose tolerance; IFG, impaired fasting glucose; T2D, type 2 diabetes; DI, disposition index; BMI, body mass index. a The 6-min response represents the incremental C-peptide response to glucagon 6 min after intravenous administration of glucagon. Data are mean ± SD. The C-peptide response was totally blunted when the glucagon test was preceded by ITT. The CV for C-peptide response is calculated between the glucagon test (G) and G ITT. b The DI is calculated as the product of the 6 min C-peptide response and K ITT. c The M-value is the measure of insulin sensitivity derived from the euglycaemic hyperinsulinaemic clamp. The values in parenthesis in protocol 1 are mean values just before the glucagon test in ITT G and before ITT in G ITT. The insulin values in parenthesis in ITT G are after intravenous administration of insulin. The values in parenthesis in protocol 3 are mean fasting glucose values at the start of the clamp. ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264;

5 Glucose (mmol l 1 ) Time (min) C-peptide (nmol l 1 ) C-peptide response nmol l 1 Glucagon (0.5 mg) Insulin 0.05 U kg K ITT (% min 1 ) Fig. 1 Schematic figure of combined glucagon-stimulated C-peptide test and insulin tolerance test (GITT); glucagon is injected intravenously at 0 min and insulin at 30 min. Black circles = C-peptide; black squares = glucose. marker of beta-cell function. For insulin sensitivity, the first-order rate constant for glucose disappearance, K ITT, was estimated from the slope of regression line of ln plasma glucose compared with the time using the formula K ITT =ln2 T from 0 to 30 min after the insulin bolus [18]. As a hyperbolic relationship between insulin sensitivity and beta-cell function is a prerequisite for estimating DI [5], we investigated the existence of the hyperbolic relationship between insulin sensitivity and beta-cell function in GITT using complete GITT data from 24 subjects with NGT (Fig. 2). DI was then calculated as the product of K ITT and incremental C-peptide response at 6 min. Anthropometric measurements Height, weight, waist to hip ratio and fat free mass (FFM) were measured on the first test day, FFM with a bioelectrical impedance method [20]. Assays Plasma glucose was measured with a glucose oxidase method, using a Beckman glucose analyzer II (Beckman Instruments, Fullerton, CA, USA). Serum insulin concentrations were measured with a double antibody ELISA (DAKO, Cambridgeshire, UK) with an interassay CV of 8 %. Plasma C-peptide concentrations were measured by radioimmunoassay (LINCO, St. Charles, MO). The lower limit of detection for Fig. 2 The hyperbolic relationship between beta-cell function and insulin sensitivity in 24 subjects with normal glucose tolerance; the correlation between the log-transformed C-peptide responses and the log-transformed K ITT was equal to )0.59 (P = 0.004). The slope of the relationship between the log-transformed variables was equal to )1.37 ± 0.82 and it was not significantly different from )1 (P = 0.366). C-peptide was 0.03 nmol L )1 ; the intra-assay CV was 4.5% and the inter-assay CV 3.2%. Statistical analysis Data are expressed as mean ± SD. Statistical analyses included Mann Whitney test for comparison between group means, and Spearman rank correlation for testing of relations. Data for insulin were log transformed for normality. Statistical analyses were carried out using the ncss statistical software (Number cruncher statistical system, version 6, Cork, Ireland) and spss version The CVs for the measures of insulin sensitivity (K ITT ) and insulin secretion (C-peptide) for each subject were calculated individually from each set of repeated tests. To investigate whether the relationship between C-peptide response and K ITT was hyperbolic (X Y = constant), we estimated the ln(c-peptide response) as a linear function of ln(k ITT ) using regression [21]. Results Protocol 1: validation of the glucagon-stimulated C-peptide and insulin tolerance test The C-peptide response to glucagon and K ITT values from four different tests are shown in Table 1. The 66 ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264; 62 71

6 K ITT measured on three different occasions did not differ regardless of whether the ITT preceded or followed the glucagon test. The mean CV for K ITT calculated from three different days was 13%. On the other hand, the C-peptide response to glucagon was significantly lower when it was preceded by the ITT. Protocol 2: reproducibility of the glucagon-stimulated C-peptide and insulin tolerance test The K ITT values and C-peptide response to glucagon at two different tests are shown in Table 1. The mean CV for K ITT calculated from two different test days was 22% and for the C-peptide response 23% in nine patients with T2D. Protocol 3: comparison of the Botnia clamp with the glucagon-stimulated C-peptide and insulin tolerance test relationship between indices of insulin sensitivity and secretion The K ITT correlated strongly with the M-value obtained during the Botnia clamp (r = 0.87, r 2 = 0.75, P < 0.001, n = 20) in both diabetic and nondiabetic subjects (Fig. 3). The C-peptide response KITT (% min 1 ) M (mg kg 1 FFM.min) Glucose tolerance IFG NGT T2D Fit line for Total R Sq Linear = Fig. 3 Correlation between indices of insulin sensitivity (K ITT obtained during GITT and M obtained during Botnia clamp) across the different degrees of glucose tolerance, n = 20 (10 T2D d, two IFG e, eight NGT s) r = 0.87, r 2 = 0.75, P < for all individuals (in T2D r = 0.85, in NGT r = 0.80, in NGT + IFG r = 0.83) to glucagon did not correlate with the FPIR measured during the Botnia clamp (r = 0.09, P = NS). This is not surprising because patients with T2D have practically absent FPIR to glucose but preserved response to stimulation with glucagon. Protocol 4: application of the GITT to discriminate between different stages of glucose tolerance In the 24 NGT subjects, the C-peptide responses to glucose plotted against the K ITT values showed the expected hyperbolic relationship as reported previously [5, 21] (Fig. 2). The correlation between the log-transformed C-peptide responses and the logtransformed K ITT values was equal to )0.59 with P = The slope of the curve, when we plotted the log-transformed C-peptide responses against the log-transformed K ITT, was equal to )1.37 ± 0.82 and we could not reject the null hypothesis of the coefficient being equal to )1 (P = 0.366), supporting a hyperbolic relationship between the two variables [21]. We therefore used these measurements to calculate the DI as the product of K ITT and C-peptide response at 6 min. The DI was significantly lower in patients with T2D and secondary drug failure than in drug-naïve T2D patients who, in turn, had lower values compared with individuals with NGT (Fig. 4). Protocol 5: reduced sampling test To test whether the number of glucose measurements needed during the test could be reduced, we recalculated the K ITT values in 17 patients with T2D using different approaches. K ITT values calculated based on measurements at 0, 1, 3, 5, 7, 9, 11, 13, 15, 20 and 30 min were compared with values calculated using fewer measurements (Fig. 5). There was no benefit of extending the sampling beyond 20 min and the number of glucose samples could also be reduced to four without any significant changes in K ITT values (0 30 vs. 0, 5, 11 and 20 min = 1.18 ± 0.15 vs ± 0.16, P = 0.96, CV = 11%). ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264;

7 (a) K ITT (% min 1 ) n = 8 n = 6 n = 11 K ITT , 5, 11, 30 0, 5, 11, 20 Calculated time points (b) NGT Drug naive Sec. drug. T2D fail. n = 8 n = 6 n = 11 Fig. 5 The reduced sampling test for calculation of K ITT. The solid line represents average K ITT values when using all glucose samples at 0, 1, 3, 5, 7, 9, 11, 13, 15, 20 and 30 min. The dashed lines indicate the corresponding 95% confidence limits. The dots represent average K ITT values when using reduced number of glucose samples. C-peptide response (nmol l 1 ) (c) DI NGT Drug naive Sec. drug T2D fail n = 8 n = 6 n = 11 NGT Drug naive Sec. drug T2D fail. Fig. 4 (a) K ITT, (b) C-peptide response and (c) the disposition index (DI) calculated from GITT in eight subjects with normal glucose tolerance (NGT), six drug-naïve patients with type 2 diabetes (T2D) and 11 patients with T2D and secondary drug failure as a product of C-peptide response and K ITT. The P-value for differences in DI between groups was significant after adjustment for age, sex and BMI (NGT versus drug-naïve T2D = 0.038, drug-naïve T2D versus secondary drug failure = 0.022). Discussion The present study validates a new tool for the independent assessment of insulin sensitivity and beta-cell function from the same test, the combined GITT. Measurements of insulin sensitivity and beta-cell function have not been routinely advocated in clinical practice. There might be several reasons; simple tools have not been available and this knowledge would not have influenced the choice of treatment. This might now change with the increasing knowledge of heterogeneity of diabetes and with the introduction of new drugs with specific effects on insulin sensitivity and insulin secretion. We therefore tried to build such a test from a test of insulin sensitivity (ITT) and betacell function, the glucagon C-peptide test (GITT). The short ITT has been reported to have good reproducibility with CV of 9 13% in NGT subjects [22]. In subjects with glucose tolerance varying from normal to diabetic, the CV was as high as 30% [23]. The K ITT value in this study showed good reproducibility with a CV of 13% in nine subjects with NGT and 22% in nine patients with T2D. The K ITT also correlated strongly with the insulin sensitivity measured during the Botnia clamp (r = 0.87, P < 0.001) regardless of glucose tolerance of the subjects. The relationship between the indices of beta-cell function and insulin sensitivity derived from the GITT was hyperbolic, allowing calculation of DI. The 68 ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264; 62 71

8 importance of this is demonstrated in Table 1. Although the incremental C-peptide response to glucagon was similar in both NGT and T2D in protocols 1 and 2, the disposition indices were markedly different (3.02 vs ) indicating that the insulin sensitive NGT subjects can maintain NGT with much less insulin than the insulin resistant subjects with T2D. This, in our view, is a strong argument for the test and adjusting beta-cell function for the degree of insulin resistance. We chose to perform the C-peptide test before the ITT because the C-peptide response was markedly blunted when preceded by the ITT. Exogenous insulin is known to suppress the endogenous insulin response and in subjects with NGT, the pancreatic beta-cells may need more than 30 min to recover. We chose to use only 0.5 mg of glucagon instead of the original 1 mg as this half dose gives the same C-peptide response but fewer side effects [24]. Serum C-peptide levels measured 6 min after intravenous administration of glucagon have been used as a marker of pancreatic beta-cell reserves and have been reported in some studies to correlate with FPIR [16, 17]. However, in this study, there was no significant correlation between the absolute or incremental C-peptide concentrations at 6 min and FPIR. This is not surprising, as most patients with T2D have lost their FPIR despite a retained response to other stimuli such as glucagon. Not until the disease has progressed to late stages is there a significant reduction in the response to other stimuli including glucagon [25, 26]. The reason why we chose the low insulin dose of 0.05 U kg )1 was to avoid hypoglycaemia and thereby stimulation of the counter regulatory mechanisms, which is known to occur after about 20 min following intravenous insulin injection [27]. In our study, almost all individuals with NGT developed hypoglycaemia at min but none of the individuals with abnormal glucose tolerance did. A comparison between the shorter test (20 min) and the full-length test (30 min) as shown in Fig. 5 gave virtually identical results in patients with T2D. We thus think that the GITT provides a simple tool for the assessment of insulin sensitivity and beta-cell function in the clinic. It takes about 60 min, is relatively cheap and the number of glucose samples can be reduced to four (0, 5, 11 and 20 min) without any loss of reproducibility. The test also performs well in patients with T2D who have lost their FPIR to glucose stimulation. The DI calculated from the C-peptide response to glucagon and K ITT clearly distinguished between patients newly diagnosed with T2D and patients with T2D and secondary drug failure. Assessment of beta-cell function has been problematic in patients with T2D as both glucose levels and insulin resistance tend to enhance insulin secretion. Therefore, the treatment that lowers glucose concentrations and enhances insulin sensitivity can give an erroneous picture of deteriorated beta-cell function unless insulin sensitivity and glucose concentrations are considered. The main indication for GITT is assessment of insulin secretion and action in diabetic patients with impaired insulin secretion. Further studies are needed to determine whether it can be useful for choice of therapy and assessment of changes in beta-cell function over time. GITT could be particularly useful in this regard as it allows to correct the assessment of beta-cell function for changes in insulin sensitivity, which are likely to occur in response to therapy. In conclusion, we present and validate a simple independent measure of insulin sensitivity and beta-cell function obtained from the same test, the GITT, providing a feasible and reproducible tool for assessing insulin sensitivity and beta-cell function in metabolic studies and clinical practice, particularly when monitoring the effect of interventions which change both beta-cell function and insulin sensitivity in patients with T2D. Conflict of interest statement TT is a member of the Cardiovascular Advisory Board (Finnish and Nordic) of Novartis, which is not directly connected with the contents of this manuscript. The other authors declare no conflict of interest. ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264;

9 Acknowledgements The skilful assistance by the Botnia Research group and the research nurses Agneta Kastensson, Gertrud Ahlqvist, Hanna Söderling and Ylva Wessman (Malmö, Sweden) as well as Monika Gullström (Närpes, Finland) is gratefully acknowledged. Peter Almgren and Esa Laurila are acknowledged for statistical support and analysing C-peptide and insulin. The study was financially supported by grants from the Sigrid Juselius Foundation, Folkhälsan Research Foundation, Academy of Finland, Swedish Medical Research Council, Finnish Diabetic Research Foundation, Swedish Diabetic Research Foundation, Skane County Council Research & Development Foundation and the Novo Nordisk Foundation. References 1 Beck-Nielsen H, Groop LC. Metabolic and genetic characterization of prediabetic states. Sequence of events leading to noninsulin-dependent diabetes mellitus. J Clin Invest 1994; 94: Cavaghan MK, Ehrmann DA, Polonsky KS. Interactions between insulin resistance and insulin secretion in the development of glucose intolerance. J Clin Invest 2000; 106: Eriksson J, Franssila-Kallunki A, Ekstrand A et al. Early metabolic defects in persons at increased risk for non-insulin-dependent diabetes mellitus. N Engl J Med 1989; 321: Tripathy D, Carlsson AL, Lehto M, Isomaa B, Tuomi T, Groop L. Insulin secretion and insulin sensitivity in diabetic subgroups: studies in the prediabetic and diabetic state. Diabetologia 2000; 43: Kahn SE, Prigeon RL, McCulloch DK et al. Quantification of the relationship between insulin sensitivity and beta-cell function in human subjects. Evidence for a hyperbolic function. Diabetes 1993; 42: Bergman RN, Ider YZ, Bowden CR, Cobelli C. Quantitative estimation of insulin sensitivity. Am J Physiol 1979; 236: E Korytkowski MT, Berga SL, Horwitz MJ. Comparison of the minimal model and the hyperglycemic clamp for measuring insulin sensitivity and acute insulin response to glucose. Metabolism 1995; 44: Saad MF, Anderson RL, Laws A et al. A comparison between the minimal model and the glucose clamp in the assessment of insulin sensitivity across the spectrum of glucose tolerance. Insulin Resistance Atherosclerosis Study. Diabetes 1994; 43: Saad MF, Steil GM, Riad-Gabriel M et al. Method of insulin administration has no effect on insulin sensitivity estimates from the insulin-modified minimal model protocol. Diabetes 1997; 46: Albareda M, Rodriguez-Espinosa J, Murugo M, de Leiva A, Corcoy R. Assessment of insulin sensitivity and beta-cell function from measurements in the fasting state and during an oral glucose tolerance test. Diabetologia 2000; 43: Bergman RN, Ader M, Huecking K, Van Citters G. Accurate assessment of beta-cell function: the hyperbolic correction. Diabetes 2002; 51(Suppl. 1): S Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: Ferrara CM, Goldberg AP. Limited value of the homeostasis model assessment to predict insulin resistance in older men with impaired glucose tolerance. Diabetes Care 2001; 24: Groop L, Forsblom C, Lehtovirta M et al. Metabolic consequences of a family history of NIDDM (the Botnia study): evidence for sex-specific parental effects. Diabetes 1996; 45: Tripathy D, Wessman Y, Gullstrom M, Tuomi T, Groop L. Importance of obtaining independent measures of insulin secretion and insulin sensitivity during the same test: results with the Botnia clamp. Diabetes Care 2003; 26: Scheen AJ, Castillo MJ, Lefebvre PJ. Glucagon-induced plasma C-peptide response in diabetic patients. Influence of body weight and relationship to insulin requirement. Diabetes Metab 1996; 22: Scheen AJ, Castillo MJ, Lefebvre PJ. Assessment of residual insulin secretion in diabetic patients using the intravenous glucagon stimulatory test: methodological aspects and clinical applications. Diabetes Metab 1996; 22: Akinmokun A, Selby PL, Ramaiya K, Alberti KG. The short insulin tolerance test for determination of insulin sensitivity: a comparison with the euglycaemic clamp. Diabet Med 1992; 9: Zimmet PZ. Diabetes epidemiology as a tool to trigger diabetes research and care. Diabetologia 1999; 42: Segal KR, Van Loan M, Fitzgerald PI, Hodgdon JA, Van Itallie TB. Lean body mass estimation by bioelectrical impedance analysis: a four-site cross-validation study. Am J Clin Nutr 1988; 47: Utzschneider KM, Prigeon RL, Carr DB et al. Impact of differences in fasting glucose and glucose tolerance on the hyperbolic relationship between insulin sensitivity and insulin responses. Diabetes Care 2006; 29: Hirst S, Phillips DI, Vines SK, Clark PM, Hales CN. Reproducibility of the short insulin tolerance test. Diabet Med 1993; 10: Hermans MP, Levy JC, Morris RJ, Turner RC. Comparison of insulin sensitivity tests across a range of glucose tolerance from normal to diabetes. Diabetologia 1999; 42: Ahren B, Nobin A, Schersten B. Insulin and C-peptide secretory responses to glucagon in man: studies on the dose-response relationships. Acta Med Scand 1987; 221: ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264; 62 71

10 25 Hermans MP, Levy JC, Morris RJ, Turner RC. Comparison of tests of beta-cell function across a range of glucose tolerance from normal to diabetes. Diabetes 1999; 48: Weir GC, Laybutt DR, Kaneto H, Bonner-Weir S, Sharma A. Beta-cell adaptation and decompensation during the progression of diabetes. Diabetes 2001; 50(Suppl. 1): S Gerich J, Cryer P, Rizza R. Hormonal mechanisms in acute glucose counterregulation: the relative roles of glucagon, epinephrine, norepinephrine, growth hormone, and cortisol. Metabolism 1980; 29: Correspondence: Mozhgan Dorkhan, Division of Diabetes & Endocrinology, Department of Clinical Sciences, Lund University, Malmö University Hospital, Malmö S-20502, Sweden. (fax: ; mozhgan.dorkhan@med.lu.se). ª 2008 Blackwell Publishing Ltd Journal of Internal Medicine 264;