Sage Journals: Discover world-class research

Abstract

Background

Insulin-like growth factor binding protein-1 (IGFBP-1) is a hepatically synthesised protein which can be used as a marker of insulin resistance. We hypothesised that the fall in serum IGFBP-1 at 2 h following a glucose load (ΔIGFBP-1_0–2h) would be a more robust measure of hepatic insulin resistance than a fasting level alone.

Methods

All subjects had a standard 75 g oral GTT including IGFBP-1 and insulin measurement to enable calculation of ΔIGFBP-1_0–2h, ΔInsulin_0–2h, and insulin sensitivity indices. At a second visit, subjects had a frequently sampled intravenous glucose tolerance test (FSIVGTT) in order to determine the insulin sensitivity index, Si.

Results

Twenty-two individuals had normal glucose tolerance (NGT) and 9 impaired fasting glucose (IFG). ΔIGFBP-1_0–2h correlated with Si in total subjects (r = 0.49, p = 0.005) and NGT subjects (r = 0.50, p = 0.02) but not in IFG subjects (r = 0.43, p = 0.24). ΔInsulin_0–2h/ΔIGFBP-1_0–2h correlated significantly with Si in total subjects (r = −0.68, p < 0.001) and in NGT subjects (r = −0.57, p = 0.005). Multivariate analysis was fitted using different models while keeping other explanatory variables constant. The fasting IGFBP-1 model was a better predictor of Si (β = 0.431, p < 0.0001) than the ΔInsulin_0–2h/ΔIGFBP1_0–2h model (β = −0.185, p = 0.004).

Conclusions

The dynamic indices ΔIGFBP-1_0–2h and ΔInsulin_0–2h/ΔIGFBP-1_0–2h appear suitable markers of (hepatic) insulin resistance with potential clinical utility.

Keywords

FSIVGTT HOMA-IR IGFBP-1 insulin

Introduction

Insulin resistance predisposes to cardiovascular disease and type 2 diabetes.¹ Although not routinely measured in clinical settings, it has been widely measured for research purposes. It can be estimated using the hyperinsulinaemic euglycaemic clamp (HEC) which is the gold standard technique or by the frequently sampled intravenous glucose tolerance test (FSIVGTT) which is a simpler alternative. It can also be estimated using simple indices such as homeostasis model assessment (HOMA-IR)² and fasting plasma insulin (FPI).³ Whilst simpler to apply than the gold standard techniques, these all suffer from limitations which have been reviewed previously.⁴ Among novel parameters of insulin resistance is fasting serum insulin-like growth factor binding protein-1 (IGFBP-1).⁵ It correlates strongly with measurements of insulin sensitivity made by HEC and with HOMA-IR.^6,7 We have previously shown, by means of validation against the FSIVGTT, that in subjects with normal glucose tolerance (NGT) fasting serum IGFBP-1 could be used as a simple and reliable marker of insulin sensitivity.⁸ Moreover, its concentrations do not fluctuate as rapidly or as widely as those of insulin.

A difficulty in interpreting the various indices of insulin resistance is that they reflect different biological functions. The HEC, which usually involves suppressing hepatic glucose output (HGO), predominantly reflects peripheral glucose disposal and peripheral insulin resistance. Si is the insulin sensitivity index derived from the FSIVGTT. It represents the net fractional glucose clearance rate per unit change in the serum insulin concentration following a glucose load (min⁻¹·µU⁻¹·mL⁻¹).⁹ It should be distinguished from glucose effectiveness (Sg) which represents the ability of glucose to promote its own disposal and does not concern us in this study. Si reflects both hepatic and peripheral insulin resistance whereas HOMA-IR, a parameter based on a single fasting specimen, largely reflects hepatic insulin resistance. The problem with this in practice is that peripheral and hepatic insulin resistance may differ. Many individuals with near normal whole body insulin sensitivity have significantly impaired hepatic insulin sensitivity and vice versa.¹⁰ This difference can manifest itself clinically. Where hepatic insulin resistance predominates, there is elevated HGO resulting in elevated fasting glucose levels. Where peripheral insulin resistance predominates, post-load glucose levels tend to be higher reflecting impaired peripheral glucose disposal. It is appropriate therefore to consider how novel indices reflect insulin resistance at these different sites.¹¹

Serum IGFBP-1 levels vary reciprocally with those of insulin in individuals who are insulin resistant tending to have higher fasting insulin levels and lower IGFBP-1 levels than normal subjects. The changes in serum levels of IGFBP-1 lag behind those of insulin due to the suppressive action of insulin on hepatic IGFBP-1 gene transcription and the longer serum half life of IGFBP-1. This has been reported to be 89 min, considerably longer than that of insulin, which is only a few minutes.^12,13 Individuals in whom hepatic insulin resistance predominates would be expected to have a lesser fall in IGFBP-1 as insulin rises following a caloric load. Furthermore, it is recognised that dynamic tests can be more discriminating than single measurements. We therefore hypothesised that the fall in serum IGFBP-1 at 2 h following a glucose load (ΔIGFBP-1_0–2h) would be a more robust measure of hepatic insulin resistance than a fasting level alone with a lesser fall indicative of greater hepatic insulin resistance. Here we compared ΔIGFBP-1_0–2h to Si and other insulin resistance indices in non-diabetic subjects.

Subjects and methods

Subjects

The subjects in this study were the same cohort of subjects used for our previously published work on IGFBP-1.⁸ They were recruited from amongst hospital staff and patients attending the lipid out-patient clinics between February 2005 and July 2008. Inclusion criteria were age 20–65 years and BMI 20–35 kg/m². Clinical and biochemical characteristics of the subjects in the different glycaemic categories are shown in Table 1. Some of this data, excluding the dynamic indices, has been presented previously.¹⁴ Exclusion criteria were treatment with oral hypoglycaemic agents or insulin and any drugs likely to modify insulin sensitivity including steroids. Subjects were also excluded if they had had an acute illness within the previous six weeks. They attended the Clinical Investigation Unit at the Royal Surrey County Hospital, Guildford on two occasions. A standard 75 g oral glucose tolerance test (GTT) was carried out at the first visit and glycaemic status classified according to WHO criteria.¹⁰ Subjects had 5 mL of blood taken at baseline into plain tubes and 2 mL into fluoride oxalate. The same set of specimens was taken at 2 h following the glucose load. Plasma samples were analysed the same day for glucose while serum aliquots for insulin and IGFBP-1 assays were separated and stored at −80℃ until analysis.

Table 1.

Clinical and biochemical characteristics of subjects in the different glycaemic categories.

NGT (n = 22)	IFG (n = 9)	p
TBF (%)	25.5 ± 2.3	31.8 ± 3.3	0.17
WHR*	0.87 ± 0.02	0.95 ± 0.02	0.01
SBP* (mmHg)	116 ± 3	131 ± 4	0.007
DBP* (mmHg)	69 ± 2	87 ± 4	<0.001
Total cholesterol (mmol/L)	4.5 ± 0.3	4.6 ± 0.3	0.78
HDLC (mmol/L)	1.33 ± 0.06	1.19 ± 0.08	0.29
LDLC (mmol/L)	2.73 ± 0.24	2.74 ± 0.14	0.78
Triglycerides (mmol/L)	1.3 ± 0.2	1.5 ± 0.2	0.20
Urate (µmol/L)	330 ± 14	339 ± 20	0.42
HbA1c* (mmol/mol)	36 ± 1	42 ± 1	0.003
FSI (pmol/L)	76.9 ± 9.2	106.4 ± 18.3	0.10
Fasting IGFBP-1 (ng/mL)	6.7 ± 1.2	3.9 ± 1.0	0.16
Si* (min⁻¹.uU⁻¹.ml⁻¹))	9.7 ± 2.3	3.5 ± 0.7	0.015
ΔIGFBP-1_0–2h (ng/mL)	3.7 ± 0.6	2.6 ± 0.6	0.52
ΔInsulin_0–2h* (pmol/L)	200 ± 40	524 ± 61	0.003

Data are mean ± SEM. *Means between NGT and IFG categories are significantly different (p < 0.05).

BMI: body mass index; DBP: diastolic blood pressure; FSI: fasting serum insulin; HbA1c: glycated haemoglobin; HDLC: high density lipoprotein cholesterol; IFG: impaired fasting glucose; IGFBP-1: insulin-like growth factor binding protein-1; LDLC: low density lipoprotein cholesterol; NGT: normal glucose tolerance; SBP: systolic blood pressure; TBF: total body fat; WHR: waist hip ratio.

All subjects had waist hip ratio (WHR) measured. Blood pressure (BP) was measured using a Critikon Dinamap TS BP machine (Tampa, Florida, USA). Percentage total body fat (TBF) was estimated by bioimpedance analysis using a Tanita BF-680 tetra-polar bioelectric impedance device (Tokyo, Japan). All subjects gave informed consent for participation and the study was approved by the South West Surrey Local Research Ethics Committee, Guildford, UK (Study number 05/Q1909/72).

Frequently sampled intravenous glucose tolerance test and simple indices

At a second visit, subjects underwent a standard FSIVGTT. The mean interval between the first and second visits was 10 days. Following placement of two intravenous cannulae, baseline blood samples for insulin and glucose measurement were taken at 15, 20, 25 and 30 min. Glucose (0.3 g/kg of 50%) was then injected over 1 min and samples drawn at intervals up to 180 min as previously described.⁸ Subjects had 2 mL of blood taken at each time point, 1 mL into fluoride oxalate and 1 mL into a plain plastic tube for insulin analysis. All specimens were separated promptly and stored on ice until the end of the test. Glucose analysis was carried out on the fluoride oxalate plasma later the same day. Serum for insulin assay was stored at −80℃ until analysis. Data analysis was carried out using the software package Millennium 6.02 purchased from MinMod Inc (Pasadena, California, USA).¹⁵

Assays

Routine laboratory analyses were carried out on a Siemens Advia 1650 auto-analyser (Camberley, Surrey, UK). Glucose was assayed by the hexokinase method and urate by a uricase method.¹⁶ Cholesterol measurement was by an enzymatic method using cholesterol esterase,¹⁷ HDL-cholesterol by a direct method using a colorimetric endpoint based on Trinder reaction¹⁸ and triglycerides by a method based on a Fossati three-step enzymatic reaction.¹⁹ LDL-cholesterol was estimated by calculation using the Friedewald formula.²⁰ Glycated haemoglobin (HbA1c) analysis was by high pressure liquid chromatography on a Biorad Variant-II instrument. Insulin was measured by solid phase two-site ELISA (Mercodia, Uppsala, Sweden) with detection limit of 7.0 pmol/L and mean analytical CV of 4.9% within and between runs. IGFBP-1 was measured by in-house ELISA using a biotinylated monoclonal antibody and other R&D analytical components. The maximum analytical CVs were 7.0% and 9.1% for intra- and inter-assay precision, respectively, and sensitivity was 0.01 ng/mL. The mean values for quality control sera (within and between IGFBP-1 assays) were 1.3, 4.2, 23.2 ng/mL for low, medium and high levels, respectively.

The binding of IGFBP-1 to capture antibody was validated using a western blotting technique which showed no cross reactivity with other binding proteins.¹²

Simple insulin resistance indices

The index ISI-gly (insulin sensitivity index-glycaemia) was calculated using the formula

ISI-gly = 2 / [(insulin ρ \times glucose ρ) + 1]

where insulin

ρ

and glucose

ρ

represent the sum of measurements of insulin (mU/L) and glucose (mg/dL) taken at baseline and 2 h during a GTT. Conversion between units for insulin was by pmol/L ÷ 6 = mU/L.

HOMA-IR was calculated using the formula:

HOMA-IR = insulin (mU / L) \times glucose (mmol / L) / 22.5

and HOMA2-IR determined using a computer programme.²¹ QUICKI (quantitative insulin sensitivity check index) was determined from fasting glucose and insulin measurements using the formula:

QUICKI = 1 / [loginsulin (mU / L) + logglucose (mg / dL)]

Further details of these indices can be found in a previous publication.⁴

Statistics

Data were analyzed using SPSS version 17 (SPSS Inc, Chicago, Illinois, USA). Data were transformed using natural logarithm (Ln) in the subjects to approximate a normal distribution. Associations between data were examined using Pearson’s correlation coefficients and a two-tailed t-test. Comparisons of the means were performed using Mann-Whitney and unpaired Student’s t-tests. Statistical significance was taken as p < 0.05. Multiple linear regression analysis was used to model the data and LnSi used as a function of the explanatory variables. Model fit was assessed using the F test. Multicollinearity was assessed using the variance inflation factor (VIF). The level of significance was set at 5%.

Results

There were 22 subjects with normal glucose tolerance (NGT) (15 males, seven females, aged 38.5 ± 13.6 years and BMI 26.2 ± 4.7 kg/m²) and nine subjects with impaired fasting glucose (IFG) (six males, three females, aged 49 ± 8.7 years and BMI 30.3 ± 3.1 kg/m²). A comparison between the groups is shown in Table 1. Individuals with IFG were older, had higher systolic and diastolic blood pressures and higher HbA1_C than those with NGT. Furthermore, ΔInsulin _0–2h after a glucose load was much higher in the group with IFG. ΔIGFBP-1_0–2h was not significantly different between the two groups. Fasting IGFBP-1 was strongly correlated with Si in total subjects but not in subjects with IFG as previously reported by ourselves in these subject groups.⁸

Correlation analysis

Correlation analysis of Si and the delta (Δ) dynamic indices is shown in Table 2. ΔIGFBP-1_0–2h correlated with Si in total subjects (r = 0.49, p = 0.005) and NGT subjects (r = 0.50, p = 0.017). In other words, the fall in IGFBP-1 was greater in insulin sensitive patients. There was no significant correlation in individuals with IFG (r = 0.43, p = 0.24), but this may relate to the small size of the IFG group. In subjects with NGT the correlation between Si and ΔInsulin_0–2h/ΔIGFBP-1_0–2h ratio was significant (r = −0.57, p = 0.005) and in total subjects (following Ln transformation) the correlation between Si and the ratio was highly significant (r = −0.68, p < 0.001). Correlation analysis of ΔIGFBP-1_0–2h and simple insulin resistance indices is shown in Table 3. ΔIGFBP-1_0–2h did not correlate significantly with any of the other indices in any subject group. The strong positive correlation between ΔIGFBP-1_0–2h and Si for all subjects is shown in Figure 1 and the strong negative correlation between Si and ΔInsulin_0–2h/ΔIGFBP-1_0–2h ratio in Figure 2.

Figure 1.

Plot of the correlation between Si and ΔIGFBP-1_0–2h in all subjects after Ln transformation. The glycaemic status of subjects was normal glucose tolerance (NGT = O; n = 22) and impaired fasting glucose (IFG = •; n = 9).

Figure 2.

Plot of the correlation between Si and ΔInsulin_0–2h/ΔIGFBP-1_0–2h ratio in all subjects after Ln transformation. The glycaemic status of subjects was normal glucose tolerance (NGT = O; n = 22) and impaired fasting glucose (IFG = •; n = 9).

Table 2.

Correlation analysis of FSIVGTT insulin sensitivity index (Si) with ΔIGFBP-1_0–2h, ΔInsulin_0–2h and ΔInsulin_0–2h/ΔIGFBP-1_0–2h ratio in subjects with normal glucose tolerance (NGT), impaired fasting glycaemia (IFG) and total subjects.*

NGT (n = 22)		IFG (n = 9)		Total (n = 31)*
r	P	r	p	r	p
ΔInsulin_0–2h	−0.39	0.072	−0.72	0.030	−0.54	0.002
ΔInsulin_0–2h/ ΔIGFBP-1_0–2h	−0.57	0.005	−0.40	0.286	−0.68	<0.001
ΔIGFBP-1_0–2h	0.50	0.017	0.43	0.244	0.49	0.005

Ln transformed data were used for Si, ΔInsulin_0–2h, ΔInsulin_0–2h/ΔIGFBP-1_0–2h, ΔIGFBP-1_0–2h.

Table 3.

Correlation analysis of ΔIGFBP-1_0–2h and insulin resistance indices in subjects with normal glucose tolerance (NGT) and impaired fasting glucose (IFG).

Insulin resistance index	ΔIGFBP-1_0–2h
	NGT (n = 22)	IFG (n = 9)	Total (n = 31)
	r	r	r
ISI-gly	0.13	0.38	0.21
QUICKI	0.16	0.58	0.24
HOMA-IR	−0.15	−0.58	−0.26
HOMA2-IR	−0.08	−0.58	−0.21
FSI	−0.09	−0.58	−0.20

All p values were >0.05.

Regression analysis

Further statistical analysis was carried out in order to determine whether a fasting IGFBP-1 or ΔInsulin_0–2h/ΔIGFBP-1_0–2h model was a stronger predictor of Si. The rationale for using the ratio (ΔInsulin _0–2h/ΔIGFBP-1_0–2h) in this analysis was that it had already been demonstrated to correlate more strongly with Si than ΔInsulin_0–2h or ΔIGFBP-1_0–2h. After Ln transformation of Si, IGFBP-1, FSI (fasting serum insulin), and ΔInsulin_0–2h/ΔIGFBP-1_0–2h, multiple linear regression analysis was used to fit the data. Two models were constructed. In the two models the Si values were used as a function of the explanatory variables for age, systolic BP, HDL-cholesterol, FSI, fasting glucose, WHR, and BMI. In each model IGFBP-1 and ΔInsulin_0–2h/ΔIGFBP-1_0–2h were entered independently.

In model 1, IGFBP-1 was a highly significant predictor of Si (β = 0.431, p < 0.0001). This means that for every one unit increase in IGFBP-1 there is an increase of 0.431 in the predicted value of Si while other explanatory variables remain constant. In this model, the percentage of variance explained was R² = 75.7%, F = 12.71, p < 0.001.

In model 2, ΔInsulin_0–2h/ΔIGFBP-1_0–2h was a significant predictor of Si (β = −0.185, p = 0.004). This means that for every one unit increase in ΔInsulin_0–2h/ΔIGFBP-1_0–2h there is a decrease of 0.185 in the predicted value of Si while other explanatory variables remain constant. In this model, the percentage of variance explained was R² = 68.9%, F = 9.30, p < 0.001. Regression values for all predictors are listed in Table 4.

Table 4.

Multiple regression models for predictors of Si.

	β	SE (β)	95% CI		t	p
	β	SE (β)	Lower	Upper	t	p
Model 1
Constant	7.996	1.493	4.899	11.093	5.354	<0.0001
Age	−0.003	0.008	−0.020	0.015	−0.306	0.763
SBP	−0.022	0.007	−0.038	−0.007	−3.050	0.006
HDLC	−0.222	0.380	−1.011	0.566	−0.585	0.565
FSI	−0.437	0.154	−0.756	−0.118	−2.839	0.010
FPG	−0.281	0.176	−0.646	0.084	−1.598	0.124
WHR	−2.282	1.639	−5.681	1.117	−1.393	0.178
BMI	0.060	0.032	−0.007	0.126	1.851	0.078
Fasting IGFBP-1	0.431	0.098	4.899	11.093	4.401	<0.0001
Model 2
Constant	8.415	1.692	4.906	11.925	4.973	<0.0001
Age	0.003	0.009	−0.016	0.022	0.289	0.775
SBP	−0.018	0.008	−0.035	−0.001	−2.180	0.040
HDLC	−0.109	0.430	−0.999	0.782	−0.253	0.802
FSI	−0.486	0.173	−0.846	−0.126	−2.803	0.010
FPG	−0.342	0.197	−0.751	0.068	−1.732	0.097
WHR	−0.351	1.789	−4.062	3.360	−0.196	0.846
BMI	0.020	0.035	−0.051	0.092	0.592	0.560
ΔInsulin_0–2h/ΔIGFBP-1_0–2h	−0.185	0.058	−0.305	0.065	−3.201	0.004

Data for Si, fasting IGFBP-1, FSI, and ΔInsulin_0–2h/ΔIGFBP-1_0–2h were Ln transformed.

BMI: body mass index; FPG: fasting plasma glucose; FSI: fasting serum insulin; HDLC: high density lipoprotein cholesterol; SBP: systolic blood pressure; WHR: waist hip ratio.

Discussion

In studying ΔIGFBP-1_0–2h, we chose to confine the investigation to subjects with NGT and IFG. This was because in previous work we observed that IGFBP-1 was a poorer marker of insulin resistance in subjects with deteriorating glycaemic status, a finding that likely reflects the decline in insulin secretion, particularly in subjects with type 2 diabetes.⁷ In addition, the reproducibility of serum IGFBP-1, like that of other simple indices of insulin resistance, has been reported to decline with deteriorating glucose tolerance.¹⁴

We have demonstrated that the degree of fall in serum IGFBP-1 concentration in relation to a standard glucose load may provide additional important information about an individual’s insulin sensitivity, particularly hepatic insulin sensitivity. Although the ΔInsulin_0–2h after a glucose load was much higher for IFG, ΔIGFBP-1_0–2h was lower than for individuals with NGT in keeping with greater insulin resistance at the liver in IFG and consequently less suppression of IGFBP-1 production post glucose load. The lack of any significant relation between ΔIGFBP-1_0–2h, ΔInsulin_0–2h/ΔIGFBP-1_0–2h and Si for individuals with IFG suggests that there is already significant insulin resistance at the liver in IFG, which would impair the suppression of hepatically synthesised IGFBP-1 by the increase in insulin levels at the liver in response to calorie intake.⁷ Numbers in the IFG group were small, but the trend was similar to the NGT group, indicating an attenuated but still present influence of the increase in insulin post-glucose load on hepatic IGFBP-1 production. We anticipate that if the number of subjects in the IFG group had been sufficiently large, the correlation between these novel parameters and Si would have reached statistical significance.

The strong correlations between ΔIGFBP-1_0–2h, ΔInsulin_0–2h/ΔIGFBP-1_0–2h and Si in this study (Table 2) are consistent with their representing markers of insulin resistance, both peripheral and hepatic, in subjects with NGT at least and possibly in other glycaemic categories as well. In the IFG category, the higher Δinsulin_0–2h in subjects with IFG compared to the lower ΔIGFBP-1_0–2h in the same category suggests that changes in insulin concentrations are a more reliable marker of insulin resistance in this glycaemic category. This finding of a higher Δinsulin_0–2h in subjects with IFG compared to those with NGT is comparable with the finding of Haffner et al.²² who observed increased Δinsulin_0–30min in subjects with impaired glucose tolerance. The lack of any relationship between ΔIGFBP-1_0–2h or the ΔInsulin_0–2h/ΔIGFBP-1_0–2h ratio and Si in individuals with IFG may relate to the small numbers studied (see Figures 1 and 2).

The independent associations observed in the multiple regression analysis between ΔIGFBP-1_0–2h and various features of metabolic syndrome also support the contention that it represents a marker of insulin resistance. However, the correlations are rather weaker than that between fasting serum IGFBP-1 and Si previously reported by ourselves in subjects with NGT (r = 0.79, p < 0.001).⁸ As is the case for fasting serum IGFBP-1, it would be necessary to measure ΔIGFBP-1_0–2h and ΔInsulin_0–2h in the context of a GTT if these were to be used as parameters of insulin resistance, as it would be important to demonstrate normal glucose tolerance. This is not necessarily a limitation of these indices, as their clinical usefulness, as for other simple markers of insulin resistance, is likely to be in the early stages of insulin resistance before other parameters become significantly deranged.

The absence of a significant correlation between ΔIGFBP-1_0–2h and other simple insulin resistance indices (Table 3) suggests that they provide different information. The present study was limited in examining IGFBP-1 levels at only two time points. Two hours may not be the optimal time for the second IGFBP-1 measurement. It could be anticipated that the correlation with the standard method would be stronger at maximal ΔIGFBP-1, i.e. at trough IGFBP-1 levels, reflecting the maximal suppressive effect of insulin. Other researchers who have derived insulin resistance indices used insulin and glucose measurements at 30 min intervals, which would be expected to be a more robust approach.²³ This is something that we plan to investigate in a further study.

In an effort to further investigate which factors best predicted Si, multivariate models were constructed using the parameters available, one including fasting IGFBP-1 and a second model including ΔInsulin_0–2h/ΔIGFBP1_0–2h. Model one explained more of the variance, suggesting that, on the basis of the available data, fasting IGFBP-1 is a stronger predictor of Si. As above however, this assessment is limited by the possibility that 2 h is not the optimal time point for the second insulin and IGFBP-1 measurements.

IGFBP-1 has much to recommend it as a marker of insulin resistance, being less labile than insulin and therefore less subject to artefactually low results where specimen processing is delayed. ΔIGFBP-1 is worthy of further investigation as a relatively simple marker of insulin resistance in combination with ΔInsulin. When measured at different intervals following a glucose load, these indices may potentially reflect different combinations of peripheral and hepatic insulin resistance. Investigation of this would require sampling of IGFBP-1 at multiple intervals during a GTT to ascertain the optimal time for sampling. It would require comparison against the HEC using a glucose tracer to evaluate hepatic and peripheral insulin sensitivity components directly, as has been done by other researchers.²⁴ Given that ΔIGFBP-1 is at least in part determined by hepatic insulin resistance for a given Δinsulin, it has the potential to give important information about hepatic insulin resistance.

In summary, we propose that ΔIGFBP-1_0–2h, i.e. the fall in serum IGFBP-1 in the context of a GTT and the ΔInsulin_0–2h/ΔIGFBP-1_0–2h ratio are markers of (hepatic) insulin resistance worthy of fuller investigation with potential clinical utility.

Footnotes

Declaration of conflicting interests

None.

Funding

The authors gratefully acknowledge funding support from the Saudi Government represented by the Saudi National Guard Health Affairs at King Abdullah International Medical Research Center – Western Region (KIMRC-WR).

Ethical approval

The study was approved by the South West Surrey Local Research Ethics Committee, Guildford, UK (05/Q1909/72).

Guarantor

CL.

Contributorship

All authors made a substantial contribution to this work. AB carried out the practical work, data analysis and contributed to drafting. CL contributed to study design, practical work, data analysis and drafting. AHH has an extensive background in IGF research which was crucial to data analysis in this study and contributed to drafting. YO carried out regression analysis and verification of other data analysis. GF contributed to study design, practical work and drafting.

References

Turner

Clapham

. Insulin resistance, impaired glucose tolerance and non-insulin-dependent diabetes, pathologic mechanisms and treatment: current status and therapeutic possibilities. Prog Drug Res 1998; 51: 33–94.

Matthews

Hosker

Rudenski

. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412–419.

Ferrannini

Mari

. How to measure insulin sensitivity. J Hypertens 1998; 16: 895–906.

Borai

Livingstone

Ferns

. The biochemical assessment of insulin resistance. Ann Clin Biochem 2007; 44: 324–342.

Brismar

Hall

. Clinical applications of IGFBP-1 and its regulation. Growth Regul 1993; 3: 98–100.

Maddux

Chan

Mandarino

. IGF-binding protein-1 levels are related to insulin-mediated glucose disposal and are a potential serum marker of insulin resistance. Diabetes Care 2006; 29: 1535–1537.

Heald

Cruickshank

Riste

. Close relation of fasting insulin-like growth factor binding protein-1 (IGFBP-1) with glucose tolerance and cardiovascular risk in two populations. Diabetologia 2001; 44: 333–339.

Borai

Livingstone

Zarif

. Serum insulin-like growth factor binding protein-1: an improvement over other simple indices of insulin sensitivity in the assessment of subjects with normal glucose tolerance. Ann Clin Biochem 2009; 46: 109–113.

Bergman

. Lilly lecture 1989. Toward physiological understanding of glucose tolerance. Minimal-model approach. Diabetes 1989; 38: 1512–1527.

10.

Alberti

Zimmet

. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998; 15: 539–553.

11.

Borai

Livingstone

Kaddam

. Selection of the appropriate method for the assessment of insulin resistance. BMC Med Res Methodol 2011; 11: 158–158.

12.

Conover

Lee

Kanaley

. Insulin regulation of insulin-like growth factor binding protein-1 in obese and nonobese humans. J Clin Endocrinol Metab 1992; 74: 1355–1360.

13.

Lee

Conover

Powell

. Regulation and function of insulin-like growth factor-binding protein-1. Proc Soc Exp Biol Med 1993; 204: 4–29.

14.

Borai

Livingstone

Mehta

. Biological variation in fasting serum insulin-like growth factor binding protein-1 (IGFBP-1) among individuals with a varying glucose tolerance. Clin Biochem 2009; 42: 1270–1274.

15.

Boston

Stefanovski

Moate

. MINMOD Millennium: a computer program to calculate glucose effectiveness and insulin sensitivity from the frequently sampled intravenous glucose tolerance test. Diabetes Technol Ther 2003; 5: 1003–1015.

16.

Fossati

Prencipe

Berti

. Use of 3,5-dichloro-2-hydroxybenzenesulfonic acid/4-aminophenazone chromogenic system in direct enzymic assay of uric acid in serum and urine. Clin Chem 1980; 26: 227–231.

17.

Allain

Poon

Chan

. Enzymatic determination of total serum cholesterol. Clin Chem 1974; 20: 470–475.

18.

Warnick

Mayfield

Benderson

. HDL cholesterol quantitation by phosphotungstate-Mg2+ and by dextran sulfate-Mn2+-polyethylene glycol precipitation, both with enzymic cholesterol assay compared with the lipid research method. Am J Clin Pathol 1982; 78: 718–723.

19.

Fossati

Prencipe

. Serum triglycerides determined colorimetrically with an enzyme that produces hydrogen peroxide. Clin Chem 1982; 28: 2077–2080.

20.

Friedewald

Levy

Fredrickson

. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502.

21.

Levy

Matthews

Hermans

. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care 1998; 21: 2191–2192.

22.

Haffner

Miettinen

Gaskill

. Decreased insulin action and insulin secretion predict the development of impaired glucose tolerance. Diabetologia 1996; 39: 1201–1207.

23.

Matsuda

DeFronzo

. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care 1999; 22: 1462–1470.

24.

DeFronzo

Simonson

Ferrannini

. Hepatic and peripheral insulin resistance: a common feature of type 2 (non-insulin-dependent) and type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1982; 23: 313–319.

Delta insulin-like growth factor binding protein-1 (ΔIGFBP-1): a marker of hepatic insulin resistance?

Abstract

Background

Methods

Results

Conclusions

Keywords

Introduction

Subjects and methods

Subjects

Frequently sampled intravenous glucose tolerance test and simple indices

Assays

Simple insulin resistance indices

Statistics

Results

Correlation analysis

Regression analysis

Discussion

Footnotes

Declaration of conflicting interests

Funding

Ethical approval

Guarantor

Contributorship

References