Different Mechanisms for Impaired Fasting Glucose and Impaired Postprandial Glucose Tolerance in Humans

From 1986 to 2005, data were systematically collected from all individuals not known to have type 2 diabetes (n = 402); these subjects underwent a standard OGTT, as recommended by the American Diabetes Association, and a hyperglycemic clamp, as described below. All subjects gave informed written consent after the protocol had been approved by local institutional review boards. All participants were in good health and had normal physical examinations and routine laboratory tests. None of the subjects were taking any medications known to affect glucose metabolism.

Subjects were divided into different categories based on fasting plasma glucose concentrations (mean of two measurements at least 1 week apart) and 2-h plasma glucose concentrations during the standard 75-g OGTT. IFG was defined as fasting plasma glucose concentrations between ≥100 and <126 mg/dl and IGT as 2-h postchallenge plasma glucose concentrations between ≥140 and <200 mg/dl (1). The data of subjects who had fasting plasma glucose concentrations ≥126 mg/dl or 2-h postprandial glucose concentrations ≥200 mg/dl (n = 37) were eliminated for the purpose of the present study. This stratification resulted in four different groups of subjects: normal fasting glucose (NFG)/NGT (n = 240), IFG/NGT (n = 21), NFG/IGT (n = 61), and IFG/IGT (n = 43). Of these subjects, 197 were studied in the U.S., 83 in Brazil, 30 in the Netherlands, 21 in Italy, 20 in Greece, and 14 in Finland. The results of some of these subjects have been included in previous reports (14–17).

For 3 days before the study, all subjects were on a weight-maintaining diet containing at least 200 g carbohydrate and had abstained from alcohol and strenuous exercise. Subjects were admitted to a clinical research unit the evening before the experiments and given a standard dinner (10 kcal/kg: 50% carbohydrate, 35% fat, and 15% protein) between 6:30 and 7:00 p.m. Apart from water ad lib, subjects fasted thereafter until the experiments were completed. The following morning, a dorsal hand vein was cannulated in a retrograde fashion and kept in a thermoregulated Plexiglass box at 65°C for sampling arterialized venous blood (18). An antecubital vein was cannulated for infusion of 20% glucose. After a 30-min baseline period, plasma glucose concentrations were increased to 180 mg/dl (10 mmol/l) and subsequently maintained at this level for 180 min using the glucose clamp technique (19). During the experiments, blood samples for plasma insulin determinations were obtained at −30, −15, 0, 2.5, 5, 7.5, 10, 20, 40, 60, 80, 100, 120, 140, 160, and 180 min.

Plasma glucose concentrations were determined by a glucose analyzer (YSI Glucose Analyzer; Yellow Springs Instruments, Yellow Springs, OH, or Beckman Glucose Analyzer; Beckman Coulter, Fullerton, CA). Plasma insulin concentrations were determined using human insulin–specific radioimmunoassays (Pharmacia Insulin RIA; Kabi Pharmacia Diagnostics, Piscataway, NJ, or Linco Insulin RIA; Linco Research, St. Charles, MO).

Basal insulin secretion was assessed by HOMA-%B, which was calculated as (fasting plasma insulin [pmol/l] × 3.33)/(fasting plasma glucose [mmol/l] − 3.5) (20). First-phase insulin secretion was considered the average incremental plasma insulin concentration from 2.5 to 10 min of the hyperglycemic clamp divided by the average incremental plasma glucose concentration during the same interval. Second-phase insulin release was taken as the average incremental plasma insulin concentration during the last hour of the hyperglycemic clamp divided by the average incremental plasma glucose concentrations during the same interval. Insulin sensitivity was assessed by using HOMA-IR, calculated as (fasting plasma insulin [pmol/l] × fasting plasma glucose [mmol/l])/135 (20) and directly determined by dividing the average glucose infusion rate during the last hour of the hyperglycemic clamp by the average plasma insulin concentration during the same interval (referred to as the insulin sensitivity index [ISI]) (19). The ISI was multiplied by the ratio of 180 mg/dl to the actual plasma glucose concentration (in milligrams per deciliter) during the hyperglycemic clamp to account for small differences in glycemia. In healthy individuals with normal glucose homeostasis, a decrease in insulin sensitivity is fully compensated for by an increase in insulin secretion, so that the product of both of these processes, referred to as the disposition index (DI), remains constant (11). We therefore evaluated the appropriateness of β-cell function in relation to insulin sensitivity by calculating the DI of the first- and second-phase insulin response by multiplying the respective insulin responses by the ISI (11).

BMI, waist-to-hip ratio (WHR), fasting plasma insulin, HOMA-%B, HOMA-IR, first- and second-phase insulin responses, ISI, and the DIs were log transformed to obtain a normal distribution of the data for statistical analyses. Since our primary goal was to compare the pathophysiology of IFG and IGT, data were compared among the NFG/NGT, IFG/NGT, and NFG/IGT groups. Data of the IFG/IGT group were used to test for interaction between IFG and IGT. Continuous variables were compared among groups using ANOVA followed by the least-significant-difference test for variables for which the F test was significant; categorical variables, i.e., sex, were compared using the χ² test. Metabolic parameters were compared among the NFG/NGT, IFG/NGT, and NFG/IGT groups with and without adjustments for age, sex, BMI, and WHR. Normally distributed variables are given as means ± SD. For skewed data, geometric means and 95% CIs are given. All analyses were conducted using SAS (version 9.1.2; SAS Institute, Cary, NC). P values <0.05 were considered statistically significant.

Age, sex, BMI, and WHR were significantly different among the NFG/NGT, IFG/NGT, and NFG/IGT groups, as indicated by ANOVA and χ² tests (all P < 0.01). Compared with the NFG/NGT subjects (age 40.4 ± 11.8 years, n = 64 men and 176 women, BMI 25.9 kg/m² [95% CI 25.5–26.4], and WHR 0.81 [0.80–0.82]), IFG/NGT (age 45.0 ± 9.7 years, P = 0.08) and NFG/IGT (age 47.0 ± 12.1 years, P < 0.0001) subjects were older, IFG/NGT (n = 11/10, P = 0.013) but not NFG/IGT (n = 20/41, P > 0.3) subjects had a greater male-to-female ratio, NFG/IGT (27.7 kg/m² [26.5–28.9], P < 0.003) but not IFG/NGT (26.6 kg/m² [25.1–28.2], P > 0.4) subjects had a greater BMI, and IFG/NGT (0.88 [0.85–0.91]) and NFG/IGT (0.85 [0.83–0.88]) subjects had a greater WHR (both P < 0.0001). None of the physical characteristics were, however, significantly different between the IFG/NGT and NFG/IGT groups.

By definition, fasting plasma glucose concentrations were greater in the IFG/NGT group (5.82 ± 0.25 mmol/l) than in the NFG/NGT (4.89 ± 0.35) and NFG/IGT groups (5.16 ± 0.25) (both P < 0.0001), and 2-h postchallenge plasma glucose concentrations were greater in the NFG/IGT group (8.81 ± 0.80 mmol/l) than in the NFG/NGT (5.75 ± 1.05) and IFG/NGT groups (6.15 ± 1.17) (both P < 0.0001). Nevertheless, fasting plasma glucose concentrations were slightly but significantly greater in the NFG/IGT than the NFG/NGT group (P < 0.0001).

ANOVA indicated significant differences among the NFG/NGT, IFG/NGT, and NFG/IGT groups for HOMA-%B, first-phase plasma insulin response, second-phase plasma insulin response, ISI, both DIs (all P < 0.01), and HOMA-IR (P < 0.05). Baseline plasma insulin concentrations were not significantly different among the three groups despite the fasting hyperglycemia in the IFG/NGT subjects (Figs. 1 and Table 1).

Accordingly, compared with the NFG/NGT group, HOMA-%B was decreased ∼35% in the IFG/NGT group (P < 0.002), indicating impaired basal insulin secretion, but was not reduced in the NFG/IGT group (P > 0.5). HOMA-IR was significantly increased in the IFG/NGT group (by ∼30%, P < 0.05) but not the NFG/IGT group (P = 0.11) (Table 1).

During the hyperglycemic clamp, plasma glucose increased to comparable levels in all groups (P > 0.6) (Fig. 1). Compared with the NFG/NGT group, both the IFG/NGT and NFG/IGT groups had ∼30% reduced first-phase insulin responses (P < 0.018 and <0.0003, respectively). In contrast, NFG/IGT but not IFG/NGT subjects had significant decreases in second-phase insulin responses (by ∼30%, P < 0.0001) and the ISI (by ∼15%, P < 0.03) (Figs. 1 and Table 1). Accordingly, in IFG/NGT subjects, the DI of the first-phase insulin response was reduced ∼30% (P < 0.006) and the DI of the second-phase insulin response was normal (P > 0.4), whereas in NFG/IGT subjects, both DIs were reduced ∼40–50% (both P < 0.0001) (Table 1).

Adjustment for age, sex, BMI, and WHR had little influence on baseline plasma insulin concentrations, HOMA-%B, and first- and second-phase plasma insulin responses; therefore, the statistical results of these parameters remained similar (Table 2). Adjustment for age and sex also had little influence on HOMA-IR and the ISI (data not shown). However, after further adjustment for BMI and WHR, HOMA-IR (P > 0.5) and the ISI were no longer significantly different among the NFG/NGT, IFG/NGT, and NFG/IGT groups, suggesting that the differences in insulin sensitivity were largely due to differences in obesity.

When subjects with isolated IFG and isolated IGT were compared with one another, the IFG group had greater reductions in basal β-cell function as assessed by HOMA-%B (∼−35 vs. −5%, P < 0.012), comparable reductions in first-phase insulin release (both ∼−30%, P > 0.9), and less reductions in second-phase insulin release (∼+10 vs. −30%, P < 0.005) and the DI of the second-phase insulin response (∼+8% vs. −45%, P < 0.0001). Differences in HOMA-IR, the ISI, and the DI of the first-phase insulin response did not reach statistical significance (Table 1). The statistical results remained similar after adjustment for age, sex, BMI, and WHR (Table 2).

There was no significant interaction between IFG and IGT (all P > 0.2), such that their associated defects in insulin secretion and action were additive in subjects with IFG/IGT (data not shown).

The present study was undertaken to compare the pathophysiology of IFG and IGT in a more comprehensive and standardized fashion than has hitherto been done by assessing basal insulin release and first- and second-phase insulin release, as well as insulin sensitivity, in a large number of individuals with isolated IFG, isolated IGT, or combined IFG and IGT and in healthy control subjects, using the hyperglycemic clamp technique. We found that 1) in isolated IFG, basal insulin secretion and glucose-stimulated first-phase insulin secretion are impaired but second-phase insulin secretion and (peripheral, mainly representing muscle) insulin sensitivity are normal; 2) in isolated IGT, basal insulin secretion is normal but glucose-stimulated first- and second-phase insulin secretion and (peripheral) insulin sensitivity are reduced; and 3) there is no significant interaction between IFG and IGT, such that their associated defects in insulin secretion and sensitivity are additive in individuals who have combined IFG and IGT. These results hence support previous suggestions (3–10) that IFG and IGT are distinct metabolic abnormalities.

Our findings are consistent with those from most but not all previous studies that insulin resistance, as assessed by FSIVGTTs or euglycemic clamps, plays a lesser role in IFG than IGT (3,4,7). In contrast, when evaluated by HOMA-IR, individuals with IFG were found to be at least similarly, if not more, insulin resistant than individuals with IGT in the present and most previous studies (3,5,6,8–10). These apparent contradictory findings may be due to a bias of HOMA-IR toward finding lower insulin sensitivity in individuals with IFG, since their increased fasting plasma glucose concentrations contribute to an increase in HOMA-IR. Alternatively, it is possible that HOMA-IR, FSIVGTTs, and euglycemic as well as the present hyperglycemic clamps measured different processes.

In the fasting state, insulin regulation of EGP is the major factor determining plasma glucose concentrations, since most tissue glucose uptake occurs by insulin-independent mechanisms (2). HOMA-IR, which is solely based on fasting plasma glucose and insulin concentrations, may therefore primarily be an index of the resistance of hepatic and renal glucose release to suppression by insulin (21), whereas the studies using FSIVGTTs or glucose clamp experiments (3,4,7) probably primarily measured the ability of insulin to stimulate muscle glucose uptake.

With respect to β-cell function, we found that basal insulin release was impaired in IFG but not IGT individuals, as indicated by the lack of increased plasma insulin concentrations despite fasting hyperglycemia and decreased HOMA-%B. This observation is consistent with all HOMA-%B data of earlier similar studies (6–8). In contrast to basal insulin release, first-phase insulin responses were comparably reduced in IFG and IGT in the present study, as has been found in one (4) but not the other (3) prior study using FSIVGTTs. Furthermore, we found that second-phase insulin responses were reduced in IGT but not IFG; this had not been previously examined.

Regarding the differences in pathophysiologies between IFG and IGT, our results indicate that decreases in first-phase insulin responses were not involved, as first-phase insulin responses were comparably reduced in both groups. The fact that basal insulin release was decreased in IFG but not IGT suggests that impaired basal insulin release may be an important element in the pathogenesis of IFG. Conversely, the fact that second-phase insulin responses were reduced in IGT but not IFG suggests that impaired second-phase insulin release may be an important element in the pathogenesis of IGT. Furthermore, our data indicate that there may be differences in hepatorenal and muscle insulin sensitivity between IFG and IGT that may be critical factors.

As mentioned above, HOMA-IR largely reflects resistance of glucose production by liver and kidney to suppression by insulin, whereas hyperglycemic clamp–determined insulin sensitivity largely reflects the sensitivity of muscle glucose uptake to stimulation by insulin (22). Our finding that in IFG, HOMA-IR was increased but hyperglycemic clamp–determined insulin sensitivity was not decreased lead us therefore to postulate that in IFG individuals, there may be selective or preferential hepatorenal insulin resistance. The observation that in overnight-fasted IFG Pima Indians, rates of EGP were increased despite increased plasma insulin concentrations supports this hypothesis (4). Conversely, our finding that in IGT, hyperglycemic clamp–determined insulin sensitivity was reduced but HOMA-IR was not increased suggests that in IGT individuals, there may be preferential insulin resistance in muscle.

Accordingly, considering the abnormalities in insulin sensitivity and β-cell function, the present study suggests that in IFG, reduced hepatorenal insulin sensitivity, combined with impairments in basal insulin secretion and first-phase insulin release, causes fasting hyperglycemia, whereas in IGT, peripheral insulin resistance, combined with impairments in first- and second-phase insulin responses, causes postprandial hyperglycemia. Expressed differently, normal or near-normal second-phase insulin responses and muscle insulin sensitivity may prevent IFG individuals from having postprandial hyperglycemia. In contrast, normal or near-normal basal insulin release and hepatorenal insulin sensitivity may prevent IGT individuals from having fasting hyperglycemia.

The present work was supported in part by Division of Research Resources–General Clinical Research Center Grant 5M01 RR-00044, National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-20411, and an American Diabetes Association Career Development Award to C.M.

We thank Becky Miller for her excellent editorial assistance and the nursing and laboratory staff for their superb help.