Isolated Impaired Fasting Glucose and Peripheral Insulin Sensitivity: Not a simple relationship

Study subjects included all individuals who had previously participated in our research studies (1990–1998) approved by the institutional review board at Stanford University. All study participants provided informed consent. Individuals considered for inclusion were nondiabetic (13) and in good general health with no history of coronary artery, kidney, or liver disease. All had completed the following assessments: height and weight; systolic and diastolic blood pressure; oral glucose tolerance testing; lipid assessment; and IST, a quantitative evaluation of insulin resistance described below. Initially, 490 individuals were identified; 44 were removed for missing data. Of the 446 remaining individuals, 318 had NGT, 63 had isolated IFG, 33 had isolated IGT, and 32 had combined IFG/IGT.

All metabolic testing was performed in the general clinical research center after fasting for 12 h. During the oral glucose tolerance test, plasma glucose and insulin concentrations were measured before (fasting) and 2-h after oral ingestion of 75 g of glucose (14). Using these results, individuals were classified as having NGT (fasting glucose <5.6 mmol/l, 2-h glucose <7.8 mmol/l), IFG (fasting glucose 5.6–7.0 mmol/l, 2-h glucose <7.8 mmol/l), IGT (fasting glucose <5.6 mmol/l, 2-h glucose between 7.8–11.1 mmol/l), and IFG/IGT (fasting glucose 5.6–7.0 mmol/l, 2-h glucose 7.8–11.1 mmol/l) (1). Lipid measurements were performed by the core laboratory at Stanford University Medical Center by standardized methods approved by the Centers for Disease Control. This included total cholesterol, triglyceride, and HDL and LDL cholesterol concentrations.

Insulin resistance was directly measured with the modified version (15) of the IST, initially introduced and validated by our research group (16,17). The values for insulin sensitivity obtained with this approach are highly correlated (r > 0.9) with the euglycemic clamp technique (17). Briefly, after an overnight fast, an intravenous catheter was placed in each of the subjects’ arms. One arm was used for the administration of a 180-min infusion of octreotide (0.27 μg per m²/min), insulin (32 mU per m²/min) and glucose (267 mg per m²/min); the other arm was used for collecting blood samples. Blood was drawn at 10-min intervals from 150 to 180 min of the infusion to determine the steady-state plasma glucose (SSPG) and insulin concentrations. Because steady-state insulin concentrations are similar in all subjects, the SSPG concentration provides a direct measure of the ability of insulin to mediate disposal of an infused glucose load; therefore, the higher the SSPG concentration, the more insulin resistant the individual. Insulin resistance was also calculated using the formula for HOMA-IR [fasting glucose (in mmol/l) × fasting insulin (in mU/l)/22.5] (18).

Because there is no objective method to classify individuals as being insulin resistant or sensitive, we used an operational definition based on results of previously published prospective studies demonstrating a statistically significant increase in clinical syndromes in individuals in the highest tertile of SSPG concentration and lack of these outcomes in those in the lowest tertile (19,20). Thus, we quantified the percentage in the lowest, middle, and highest tertile of SSPG among each of the four glucose tolerance groups using the SSPG values in the NGT group.

Statistical analyses were performed using SPSS (version 12 for Windows; SPSS, Chicago, IL). Triglyceride and insulin concentrations were log transformed to obtain a more normal distribution for statistical tests; geometric means (21) are presented in the results section. Statistical differences between groups were assessed by one-way ANOVA followed by Bonferroni post hoc pairwise comparisons when significant (P < 0.05). P values were not otherwise adjusted for multiple comparisons made throughout the study. Indexes for insulin resistance (SSPG and HOMA-IR) were adjusted for age, sex, and BMI. The χ² test was used for categorical variables.

Table 1 compares the demographic and metabolic characteristics of the four glucose tolerance groups. The NGT group was younger, less heavy, and had lower blood pressures than the three pre-diabetic groups. However, there were no significant differences in these variables within the IFG, IGT, or combined IFG/IGT groups.

Concerning lipid metabolism, triglyceride concentrations were significantly lower and HDL cholesterol concentrations higher in the NGT group compared with the three pre-diabetic groups. Total and LDL cholesterol concentrations were also lower in individuals with NGT than in those with IFG. However, there were no differences in any lipid and lipoprotein concentrations among the three pre-diabetic groups.

Individuals with NGT had lower fasting and 2-h plasma glucose concentrations and lower insulin concentrations than the three pre-diabetic groups. Within the pre-diabetic groups, fasting and 2-h glucose concentrations differed based on their classification: IFG and IFG/IGT groups had higher fasting glucose than the IGT group, and the IFG group had lower 2-h glucose concentrations than the other two groups. Finally, fasting plasma insulin concentrations were significantly higher in the IFG/IGT groups than in those with either IFG or IGT alone, and 2-h insulin concentrations were lower in the IFG group than in the IGT and IFG/IGT groups.

SSPG concentrations increased progressively from the NGT to IFG/IGT groups (Table 2). The NGT group had significantly lower SSPG concentrations than the three pre-diabetic groups. In addition, the IFG group had lower SSPG concentrations than both the IGT and IFG/IGT groups. Adjusting for age, sex, and BMI tended to lower the SSPG concentrations in the pre-diabetic groups and accentuated the difference between the IFG and IGT groups. Of note, steady-state plasma insulin concentrations (geometric means [95% CI]) were not significantly different between the four groups (P = 0.1 by ANOVA) (NGT 347 pmol/l [336–358], IFG 368 pmol/l [334–405], IGT 393 pmol/l [323–480], and IFG/IGT 421 pmol/l [344–517]).

HOMA-IR was lowest in the NGT group and highest in the IFG/IGT group. However, in contrast to SSPG concentrations, the increase in HOMA-IR was not progressive across the four groups. Thus, HOMA-IR values were higher in the IFG group than in the NGT group and lower than in the IFG/IGT group. On the other hand, values were not lower in the IFG group compared with the IGT group. Adjusting for age, sex, and BMI attenuated the difference between the NGT and IGT groups.

Figure 1 shows the distribution of insulin resistance (SSPG concentration) in the four groups and demonstrates a progressive increase in the median SSPG concentrations from the NGT to the IFG/IGT groups. While there is considerable overlap between the groups, the length of the boxes (depicting the interquartile range) was greatest in the IFG group and least in the IFG/IGT group. In addition, for the IFG group, the median was centered between the interquartile range. This was not true of the other three groups, with the extreme example being those with IFG/IGT, demonstrating a much narrower distribution in SSPG values above the median than below.

To quantify the proportion of individuals in each of the four groups based on their insulin action, we divided them into tertiles based on the SSPG concentrations of the NGT group (Table 3): lowest, tertile 1 (insulin sensitive); middle, tertile 2; and highest, tertile 3 (insulin resistant). There was a significant difference in the percentage of individuals in each SSPG tertile among the four glucose tolerance groups (P < 0.001 by χ² test). Thus, the percentage of individuals in tertile 3 (insulin resistant) increased progressively from NGT (33%) to IFG (57%) to IGT (76%) to IFG/IGT (94%); the percentage of individuals in tertile 1 (insulin sensitive) decreased from NGT (33%) to IFG (13%) to IGT (3%) to IFG/IGT (3%). As indicated in the research design and methods section, the decision to use tertiles of SSPG concentrations in the NGT group to classify individuals as being insulin resistant or sensitive was based on results of prospective studies showing that these cut points identified individuals who did, or did not, develop clinical syndromes associated with insulin resistance (19,20). However, as seen in Table 3, the relative proportion of insulin-resistant individuals in the IFG group remained high even when the upper quartile (53%) or quintile (44%) of SSPG in the NGT group was used as the definition. Finally, limiting the analysis to the non-Hispanic White population did not meaningfully alter the results (not shown).

The results presented emphasize the heterogeneity of peripheral insulin sensitivity in individuals with NGT, as well as those classified as having pre-diabetes. Although we arbitrarily used the SSPG tertiles in the NGT group to define insulin sensitivity (lowest tertile) and insulin resistance (highest tertile), this definition has been shown in two prospective studies to predict disease outcomes including type 2 diabetes and cardiovascular disease (19,20). Using this definition, the IFG group was the most heterogeneous of the pre-diabetes group, with 57% who were insulin resistant and 13% who were insulin sensitive. The thrust of these findings remained when individuals with IFG were divided into SSPG quartiles (53% insulin resistant, 6% insulin sensitive) or quintiles (44% insulin resistant, 6% insulin sensitive).

In the most general sense, our findings are similar to several others in that pre-diabetic individuals tend to be older and heavier, with higher blood pressure and with dyslipidemia (2,3,5–8,10,11)—changes seen in the three pre-diabetes group in our study. Controversy, however, remains concerning the degree of peripheral insulin resistance in individuals with isolated IFG. Specifically, how do values of insulin sensitivity in individuals with IFG differ from those with NGT, IGT, or IFG/IGT? Prior publications disagree in their conclusions; some report that individuals with IFG are more insulin resistant than those with NGT (5–11), while others have found that IFG individuals are similar to those with NGT (2–4) and more insulin sensitive than those with IGT and IFG/IGT (2,4). It is not possible to review in detail all previous publications addressing this issue, but it seems likely that some of the discordance results from use of surrogate markers, rather than direct measurements of insulin sensitivity. For example, HOMA-IR tends to overemphasize the degree of peripheral insulin resistance in the IFG group. This is due to the HOMA-IR formula, which is simply based on fasting glucose and insulin: [fasting glucose (in mmol/l) × fasting insulin (in mU/l)/22.5]. By definition, the IFG group has higher fasting glucose than the NGT or IGT groups; therefore, they will tend to have higher HOMA-IR than either group. Thus, Table 2 indicates that the IFG group tends to have higher HOMA-IR than the IGT group, although by a specific measure of insulin resistance (SSPG), the IFG group is less insulin resistant. This discrepancy has been seen in other studies where both HOMA-IR and a more specific measure of insulin resistance were reported (2,3,7).

Limiting consideration to studies utilizing direct measurements of insulin sensitivity does not remove uncertainty as to the degree of peripheral insulin resistance in IFG. Thus, Weyer et al. quantified insulin sensitivity using the euglycemic-hyperinsulinemic clamp technique in Pima Indians and found IFG individuals to have significantly lower insulin sensitivity values compared with NGT individuals and values similar to those of IGT or IFG/IGT individuals (11). In contrast, three other studies using the hyperinsulinemic clamp technique (2–4) found no significant differences in insulin sensitivity between the IFG and NGT groups. In two of these studies (2,4), individuals with IFG were more insulin sensitive than those with NGT (although not statistically significant). Comparison with the IGT group yielded different results; in two studies, insulin sensitivity values were similar between IFG and IGT groups (3,11) and in two other studies IFG individuals were more insulin sensitive than those with IGT (2,4).

A possible explanation for such discordant results when similar techniques are used to address the same question is that it results from a combination of the heterogeneity of insulin sensitivity values in IFG individuals and the relatively few such individuals in the reports discussed above (2–4,11). Specifically, in those four studies, measurements of insulin sensitivity were performed in 10 (2), 21 (3), 10 (4), and 11 (11) individuals with IFG; this represents a total of 52 patients with IFG as compared with the 63 individuals with IFG in this study. Furthermore, while 57% of IFG individuals were insulin resistant, there were 43% who were in the most insulin sensitive and intermediate tertiles. Therefore, it seems obvious that conclusions about the degree of peripheral insulin resistance in individuals with IFG could vary dramatically as a function of which members of this heterogeneous population were enrolled in experimental groups of ∼10 individuals.

In conclusion, our results are consistent with many previous reports in that our subjects with IFG, IGT, and IFG/IGT tend to be older and heavier and to have more metabolic abnormalities than those with NGT (2,3,5–8,10,11). Our results are also in general agreement that peripheral insulin resistance is increased in individuals with IGT and IFG/IGT compared with those with NGT (2–12). However, they are somewhat in conflict with the recent American Diabetes Association (ADA) statement (22) that “people with isolated IFG predominantly have hepatic insulin resistance and normal muscle insulin sensitivity.” Although hepatic glucose output is likely to be suppressed during measurement of insulin sensitivity (SSPG) (16), we cannot comment upon that aspect of the ADA statement, as we did not specifically quantify hepatic glucose output. Nonetheless, at the simplest level we have shown that insulin sensitivity values are heterogeneous in all four glucose tolerance groups (Table 3), and there are both insulin sensitive and resistant individuals with NGT, IFG, IGT, and IFG/IGT; what distinguishes these four groups from one another is the relative number of individuals who fall into each category. The greatest degree of homogeneity is seen in the IFG/IGT group (94% insulin resistant, 3% insulin sensitive) and IGT group (76% insulin resistant, 3% insulin sensitive), explaining why there is no disagreement about the presence of insulin resistance in either group. In contrast, there is considerable heterogeneity in the NGT and IFG groups, and this very likely contributes to different findings when these groups are compared. However, despite this heterogeneity, and irrespective of how the population is divided, the majority of IFG individuals are insulin resistant and the minority are insulin sensitive when compared with NGT individuals. Based on these findings, it appears that considerable caution needs to be expressed when generalizing about peripheral insulin resistance in IFG individuals based on the findings in relatively few patients.

This research was supported by the National Institutes of Health (RR-00070 to the general clinical research center).