Impaired plasma glucose clearance is a key determinant of fasting hyperglycemia in people with obesity

Abstract

Objective:

To evaluate the relative importance of the basal rate of glucose appearance (Ra) in the circulation and the basal rate of plasma glucose clearance (CR) in determining fasting plasma glucose concentration in people with obesity and different fasting glycemic status.

Methods:

We evaluated basal glucose kinetics in 33 lean people with normal fasting glucose (<100 mg/dl; Lean<100) and 206 people with obesity and: i) normal fasting glucose (Ob<100 group, n=118), ii) impaired fasting glucose (100–125 mg/dl; Ob100–125 group, n=66), or iii) fasting glucose diagnostic of diabetes (≥126 mg/dl; Ob≥126 group, n=22).

Results:

Although there was a large (up to three-fold) range in glucose Ra within each group, the ranges in glucose concentration in the Lean<100, Ob<100, and Ob100–125 groups were small, because of a close relationship between glucose Ra and CR. However, the CR at any Ra value was lower in the hyperglycemic than the normoglycemic groups. In the Ob≥126 group, plasma glucose concentration was primarily determined by glucose Ra, because CR was markedly attenuated.

Conclusions:

Fasting hyperglycemia in people with obesity represents a disruption of the precisely regulated integration of glucose production and clearance.

TWITTER SUMMARY

Fasting hyperglycemia in people with obesity represents a disruption of the precisely regulated integration of glucose production and clearance, presumably due to insulin resistance in conjunction with an inadequate compensatory increase in beta-cell insulin secretion.

Graphical Abstract

INTRODUCTION

The plasma glucose concentration during basal, postabsorptive conditions (evaluated after about a 12 h overnight fast) accounts for more than 75% of the total 24-hour plasma glucose concentration (basal glucose plus postprandial increases in plasma glucose) in people with and without type 2 diabetes (1, 2). Basal/fasting hyperglycemia is used to diagnose prediabetes (fasting plasma glucose 100–125 mg/dl) and type 2 diabetes (fasting plasma glucose ≥126 mg/dl) and the degree of hyperglycemia predicts the risk for complications such as atherogenic dyslipidemia, cardiovascular diseases, and nonalcoholic fatty liver disease (3). During basal conditions, the liver produces most of the glucose needed to meet the energy demands of tissues that require glucose as a fuel, namely the brain, peripheral nervous system, red blood cells and the renal medulla, which take up glucose via insulin-independent facilitated diffusion and together account for about 75% of total basal glucose disposal (4, 5). The remaining ~25% of basal glucose disposal occurs in skeletal muscles and other tissues via a combination of insulin-dependent and insulin-independent mechanisms (5–7). It is generally believed that hepatic glucose production is the primary determinant of basal plasma glucose concentration and basal hyperglycemia is caused by resistance to insulin’s ability to suppress glucose production by the liver with a subsequent increase in hepatic glucose production (8). However, plasma glucose concentration is determined by the rate of glucose production (i.e., appearance rate in plasma) and the plasma glucose clearance rate (i.e., the volume of plasma cleared of glucose per unit time). Therefore, a decrease in glucose clearance rate could also contribute to fasting hyperglycemia.

The goal of this study was to evaluate the relative importance of glucose production rate and plasma glucose clearance rate in regulating basal plasma glucose concentration to determine the physiological mechanisms responsible for fasting hyperglycemia. To this end, we evaluated basal plasma glucose concentration, glucose appearance rate in the circulation (assessed by using stable isotope-labeled glucose tracer infusion), and plasma glucose clearance rate in people who were lean and normoglycemic (fasting plasma glucose <100 mg/dl, Lean<100 group) and people with obesity with different categories of fasting glycemia, including normal fasting glucose (<100 mg/dl, Ob<100 group), impaired fasting glucose (100 mg/dl to 125 mg/dl, Ob100–125 group), and fasting glucose diagnostic of type 2 diabetes (≥126 mg/dl, Ob≥126 group) (3). We hypothesized that a low plasma glucose clearance rate in relation to the glucose production rate, rather than a high glucose production rate alone, is the major determinant of hyperglycemia.

RESEARCH DESIGN AND METHODS

Study participants

The data included in this paper represent a secondary analysis of data obtained from 239 men and women [33 lean, defined as body mass index (BMI) ≥18.5 and <25.0 kg/m²; 206 with obesity, BMI ≥30.0 and <50.0 kg/m²] who participated in a series of research studies (Clinicaltrials.gov numbers NCT02207777, NCT01299519, NCT02706262; NCT04131166; NCT01977560; NCT02994459; NCT03408613; NCT00981500) that included an evaluation of basal (overnight 12-h fast) glucose kinetics. All participants provided written informed consent before enrolling in the studies, which were approved by the Institutional Review Board of Washington University in St. Louis, MO, and completed a screening evaluation that included a medical history and physical examination, and standard blood tests after they fasted for 12 h overnight. Potential participants were excluded if they: i) had a medical condition or received treatment that could affect the study outcome measures; ii) consumed excessive amounts of alcohol (men, >21 drinks/week; women, >14 drinks/week); iii) had ≥3% body weight change within the past six months; or iv) participated in structured exercise for more than 90 minutes/week.

Body composition and metabolic testing

Participants were admitted to the Washington University Clinical and Translational Research Unit the day before the glucose kinetics study. Total body fat-free mass and fat mass were assessed by using dual-energy X-ray absorptiometry (iDXA, GE Healthcare). Basal glucose kinetics were determined in the morning, by using a ~210-min primed, constant, intravenous infusion of [6,6-²H₂]- or [U-¹³C]-labeled glucose (9), after participants fasted and rested in bed for ~12 h overnight. One-hundred and fifty participants (14 lean, 136 with obesity) received the [²H₂]glucose tracer and 89 participants (19 lean, 70 with obesity) received the [¹³C]glucose tracer. In a series of quality control studies, we found that both tracers provide identical measures of glucose kinetics when infused simultaneously (r = 0.9996). Plasma glucose, insulin, and C-peptide concentrations and glucose tracer enrichment in plasma were measured before and every 7–10 min during the last 20–30 min of the tracer infusion. A hyperinsulinemic-euglycemic clamp procedure (insulin infusion rate: 50 mU/m² body surface area/min; target plasma glucose concentration: 100 mg/dl) in conjunction with [6,6-²H₂]- or [U-¹³C]-labeled glucose infusion was conducted to determine insulin sensitivity of glucose disposal (10, 11). Participants with diabetes were instructed to stop taking GLP-1 receptor agonists for two weeks, oral diabetes medications for three days, and insulin for one day before the glucose kinetics study.

Calculations

Glucose appearance rate (Ra) into the systemic circulation, which equals the glucose disposal rate from the circulation during basal steady state conditions, was calculated by dividing the glucose tracer infusion rate by the average glucose tracer-to-tracee ratio in plasma during the last 30 min of the tracer infusion rate (9). Plasma glucose clearance rate (the volume of plasma that is cleared of glucose per minute) was calculated by dividing the glucose disappearance rate (in μmol/min) by the plasma glucose concentration (in μmol/l) (9). Hepatic sinusoidal insulin concentration was calculated as described previously, assuming hepatic plasma flow is 0.567 liters/min/m² of body surface area, composed of 80% portal vein and 20% hepatic artery plasma flow, and insulin is secreted into the portal vein where it mixes with insulin that is returned to the liver from the systemic circulation (12, 13). The hepatic insulin resistance index was calculated as the product of basal glucose Ra and basal arterial insulin concentration (14) as well as the product of basal glucose Ra and basal hepatic sinusoidal insulin concentration. Insulin-stimulated glucose disposal rate, which occurs primarily in skeletal muscles and provides an index of skeletal muscle insulin sensitivity (11), was calculated as the sum of the glucose (labeled and unlabeled) infusion rate and endogenous glucose production rate during the hyperinsulinemic-euglycemic clamp procedure.

Statistical analysis

Summary data are presented as mean ± SEM. One-way analysis of variance in conjunction with Fischer’s LSD test was used to evaluate differences in participant characteristics and metabolic variables among the groups. The relationship between two variables was evaluated by using Pearson’s correlation coefficient. Differences between groups in the relationships between: i) glucose Ra or plasma glucose clearance rate and fat-free mass, a measure of metabolically active body mass, ii) glucose Ra and plasma glucose concentration, and iii) glucose Ra and plasma glucose clearance rate were evaluated by using a general linear model with: i) group as fixed factor, glucose Ra or plasma glucose clearance rate as dependent variable and fat-free mass as a covariate or ii) group as fixed factor, glucose concentration or plasma glucose clearance rate as the dependent variable and glucose Ra as a covariate, respectively. A p-value ≤0.05 with multiple comparison adjusted post-hoc analysis (Bonferroni) as needed was considered statistically significant. Statistical analyses were performed by using Excel (Microsoft) and SPSS (v. 26, IBM).

RESULTS

Participants’ glycemic status, sex, race, age, and adiposity

By design, fasting plasma glucose in all participants who were lean was <100 mg/dl. Among the 206 participants with obesity, 118 had normal fasting glucose (Ob<100 group), 66 had impaired fasting glucose (Ob100–125 group), and 22 had a fasting plasma glucose diagnostic of type 2 diabetes (Ob≥126 group) (Table 1). The relative number of participants who were male and female were not different between groups; about two-thirds of the participants were female and one-third male (Table 1). Most participants in all groups were White and about one-third were Black and other races (Table 1). Participants in the Ob100–125 and Ob≥126 groups were about 10 years older than participants in the Ob<100 and Lean<100 groups (Table 1). The percentage of body weight as fat was not different among the groups with obesity (Table 1).

Table 1.

Participants’ sex, age, body composition, basal plasma glucose and insulin concentrations, and glucose kinetics

	Lean<100	Ob<100	Ob100–125	Ob≥126
N (M/F)	33 (11/22)	118 (21/97)	66 (16/50)	22 (5/17)
Race (White/Black/Other) (%)	79/6/15	64/32/4	71/26/3	64/27/9
Age (years)	38 ± 2	40 ± 1	48 ± 1*ƒ	51 ± 2*ƒ
BMI (kg/m2)	22.7 ± 0.3	39.3 ± 0.6*	42.0 ± 0.8*ƒ	38.6 ± 1.1*‡
Body mass (kg)	64 ± 1	110 ± 2*	119 ± 3*ƒ	107 ± 4*
Fat-free mass (kg)	45 ± 1	56 ± 1*	61 ± 2*ƒ	57 ± 3*‡
Body fat (%)	29 ± 1	48 ± 1*	48 ± 1*	47 ± 1*
Basal plasma glucose (mg/dl)	87 ± 1	89 ± 1	109 ± 1*ƒ	167 ± 6*ƒ‡
Basal plasma insulin (pmol/l)	41 ± 3	113 ± 5*	146 ± 8*ƒ	116 ± 13*‡
Basal hepatic sinusoidal insulin (pmol/l)	155 ± 8	360 ± 13*	482 ± 25*ƒ	363 ± 31*‡
Basal glucose Ra (μmol/min)	623 ± 13	800 ± 15*	945 ± 25*ƒ	1,155 ± 61*ƒ‡
Basal glucose Ra (μmol/kg FFM/min)	14.0 ± 0.3	14.4 ± 0.2	15.8 ± 0.3*ƒ	20.8 ± 1.0*ƒ‡
Basal glucose clearance rate (ml/min)	125 ± 3	158 ± 3*	159 ± 4*	124 ± 5ƒ‡
Basal glucose clearance rate (ml/kg FFM/min)	2.81 ± 0.07	2.83 ± 0.04	2.64 ± 0.05*ƒ	2.25 ± 0.09*ƒ‡
HIRI (basal glucose Ra × arterial insulin) (μmol/kg FFM/min)×(pmol/l)	565 ± 46	1,637 ± 80*	2,282 ± 123*ƒ	2,256 ± 188*ƒ
HIRI (basal glucose Ra × hepatic sinusoidal insulin) (μmol/kg FFM/min)×(pmol/l)	2,143 ± 125	5,057 ± 208*	7,246 ± 366*ƒ	7,123 ± 459*ƒ
Muscle insulin sensitivity (clamp glucose Rd/Insulin) (nmol/kg FFM/min)/(pmol/l)	132 ± 9	67 ± 5*	47 ± 3*ƒ	42 ± 4*ƒ

Values are mean ± SEM.

^*Value significantly different from value in the Lean<100 group, p ≤ 0.05.

^ƒValue significantly different from value in the Ob<100 group, p ≤ 0.05;

^‡Value significantly different from value in the Ob100–125 group, p < 0.05. Abbreviations: FFM, fat-free mass; HIRI hepatic insulin resistance index; Ra, rate of appearance; Rd, rate of disposal.

Plasma glucose and insulin concentrations

Plasma glucose concentration was not different between the Lean<100 and the Ob<100 groups and progressively increased from the Lean<100 and Ob<100 groups to the Ob100–125 and the Ob≥126 groups (Table 1). Arterial and hepatic sinusoidal insulin concentrations were higher in all groups with obesity compared with the Lean<100 group, and higher in the Ob100–125 group than the Ob<100 and the Ob≥126 groups without a difference between the Ob<100 and the Ob≥126 groups (Table 1).

Glucose kinetics, hepatic insulin resistance, and plasma glucose concentration

Glucose Ra expressed as μmol/min increased from the Lean<100 group to the Ob<100 to the Ob100–125 to the Ob≥126 groups and varied by less than 2-fold among participants in the Lean<100 group (from 462 to 737 μmol/min) and by 2.5 to 3-fold in the Ob<100 (from 503 to 1,349 μmol/min), Ob100–125 (from 496 to 1,532 μmol/min), and the Ob≥126 (from 738 to 1,620 μmol/min) groups (Figure 1A). Glucose Ra correlated directly with FFM in all groups and glucose Ra in relationship to FFM was not different between the Lean<100 and the Ob<100 groups and increased progressively from the Ob<100 to the Ob100–125 to the Ob≥126 groups (Table 1 and Figure 1A). There was no significant correlation between plasma glucose concentration and either total glucose Ra expressed in μmol/min or glucose Ra normalized to FFM (expressed as μmol/kg FFM/min) in the Lean<100, Ob<100, and Ob100–125 groups, and the plasma glucose concentration for any glucose Ra value was higher in both the Ob100–125 compared with the Ob<100 and the Lean<100 groups (Figures 1B and 1C). In contrast, plasma glucose concentration correlated directly (p<0.05) with total glucose Ra (expressed as μmol/min) and tended to correlate with glucose Ra per kg FFM (p = 0.08) in the Ob≥126 group (Figures 1B and 1C). Total glucose Ra (μmol/min) and glucose Ra per kg FFM in the Ob≥126 group were higher than in all other groups, and total glucose Ra explained about 50% (p < 0.001) of the variation in plasma glucose concentration among participants in the Ob≥126 group.

Figure 1.

Relationships between basal glucose rate of appearance into the systemic circulation and fat-free mass (A), plasma glucose concentration and total glucose appearance rate (B), plasma glucose concentration and glucose appearance rate normalized to fat-free mass (C), total plasma glucose clearance rate and fat-free mass (D), total plasma glucose clearance rate and total glucose appearance rate (E), plasma glucose clearance rate normalized to fat-free mass and glucose appearance rate normalized to fat-free mass (F), plasma glucose concentration and the hepatic insulin resistance index (G), and plasma glucose concentration and skeletal muscle insulin sensitivity, assessed as insulin-stimulated glucose disposal normalized to fat-free mass and plasma insulin concentration (H) in lean people with normal fasting plasma glucose (Lean<100 group) and people with obesity and different fasting glycemic status, including normal fasting glucose (Ob<100 group), impaired fasting glucose (Ob100–125 group), and fasting glucose within the diabetic range (Ob≥126 group). Abbreviations: FFM, fat-free mass; HIRI, hepatic insulin resistance index; IS, insulin sensitivity; Ob, obesity; r, Pearson correlation coefficient; Ra, rate of appearance. *p < 0.05; ^ƒp = 0.08.

Plasma glucose clearance rate varied by less than 2-fold in the Lean<100 (from 84 to 154 ml/min) and Ob≥126 (from 90 to 165 ml/min) groups and by 2.5 to 4-fold in the Ob<100 (from 100 to 245 ml/min) and the Ob100–125 (from 73 to 271 ml/min) groups (Figure 1D). On average, total plasma glucose clearance rate (expressed as ml/min) was higher in the Ob<100 and Ob100–125 groups compared with both the Lean<100 and the Ob≥126 groups, with no differences between the Ob<100 and Ob100–125 groups or between the Lean<100 and the Ob≥126 groups (Table 1). Plasma glucose clearance rate (ml/min) correlated directly with FFM in all groups. The clearance rate for any FFM value was not different between the Lean<100 and the Ob<100 groups, and decreased progressively from the Ob<100 to the Ob100–125 to the Ob≥126 groups (Table 1 and Figure 1D). Plasma glucose clearance rate also correlated directly with glucose Ra in all groups, both when the data were expressed as either total clearance rate (ml/min) and total Ra (μmol/min) or when clearance rate and Ra were normalized to FFM (Figures 1E and 1F). There was no difference in the relationship between glucose Ra and plasma glucose clearance rate (expressed as either total Ra and clearance rate or Ra and clearance rate per kg FFM) between the Ob<100 and the Lean<100 groups (Figures 1E and 1F). However, plasma glucose clearance rate at any glucose appearance rate value was lower in the groups with fasting hyperglycemia (Ob100–125 and Ob≥126) than in the Ob<100 and Lean<100 groups, and lower in the Ob≥126 than in the Ob100–125 group (Figures 1E and 1F).

The hepatic insulin resistance index was higher in all groups with obesity compared with the Lean<100 group and was higher in the groups with fasting hyperglycemia (Ob100–125 and Ob≥126) compared with the Ob<100 group, without a difference between the hyperglycemic groups (Table 1). There was no correlation between hepatic insulin resistance and plasma glucose concentration, and plasma glucose concentration at any hepatic insulin resistance index value was higher in the hyperglycemic groups (Ob100–125 and Ob≥126) compared with the Ob<100 and Lean<100 groups, without a difference between the Ob<100 and Lean<100 groups (Figure 1G). Moreover, plasma glucose concentration at any hepatic insulin resistance index value was higher in the Ob100–125 group than the Ob<100 group and even higher in the Ob≥126 group (Figure 1G).

Skeletal muscle insulin sensitivity was higher in the Lean<100 group than all groups with obesity and was higher in the Ob<100 group compared with the Ob100–125 and the Ob≥126 groups without a difference between the Ob100–125 and the Ob≥126 groups (Table 1). There was no correlation between skeletal muscle insulin sensitivity and plasma glucose concentration, and plasma glucose concentration at any skeletal muscle insulin sensitivity value was higher in the hyperglycemic groups (Ob100–125 and Ob≥126) compared with the Ob<100 and Lean<100 groups, without a difference between the Ob<100 and Lean<100 groups (Figure 1H).

DISCUSSION

During basal conditions, plasma glucose concentration represents the balance between the rates of endogenous glucose production and plasma glucose clearance. In this study, we evaluated the relationships among fasting plasma glucose concentration and several key factors involved in regulating basal glucose metabolism in four groups: people with normal weight and normal fasting glucose (Lean<100); people with obesity and normal fasting glucose (Ob<100); and two groups of people with obesity and fasting hyperglycemia (Ob100–125 and Ob≥126). We found that glucose Ra correlated directly with FFM, the primary site of glucose disposal, in all groups. The relationship between glucose Ra and FFM was not different between the Lean<100 group and the Ob<100 group, but glucose Ra at any FFM value was higher in the groups with fasting hyperglycemia compared with the Lean<100 and Ob<100 groups and was higher in the Ob≥126 group than the Ob100–125 group. However, glucose Ra did not determine plasma glucose concentration in the groups with fasting plasma glucose concentrations below 126 mg/dl, because the rate of plasma glucose clearance progressively increased in concert with progressive increases in glucose Ra. In contrast, glucose concentration correlated directly with glucose Ra in the Ob≥126 group because of a marked attenuation in plasma glucose clearance rate. In addition, the severity of hepatic insulin resistance was not a determinant of plasma glucose concentration in any of our groups of participants. These data demonstrate that impaired plasma glucose clearance relative to glucose Ra, rather than an increase in glucose Ra itself or hepatic or skeletal muscle insulin resistance, are responsible for fasting hyperglycemia in people with obesity.

Our data demonstrate a large (up to a three-fold) range in glucose Ra in all groups with obesity with considerable overlap in glucose Ra values among the groups. In contrast, the ranges in plasma glucose concentration in the Ob<100 and Ob100–125 groups were very small because of a close relationship between glucose Ra and plasma clearance rates (i.e., a high glucose Ra was associated with a correspondingly high plasma glucose clearance rate and vice versa). Nonetheless, basal plasma glucose concentration was greater in the Ob100–125 group than in the Ob<100 group because the plasma glucose clearance rate at any glucose Ra value was less than in the Ob<100 group. In the Ob≥126 group, plasma glucose concentration correlated directly with glucose Ra, because of a markedly blunted glucose clearance rate relative to glucose Ra. These results extend the results from previous studies (15–17), and demonstrate basal glucose Ra does not determine basal plasma glucose concentration unless there is a defect in plasma glucose clearance. The close association between basal glucose Ra and plasma glucose clearance rate demonstrates a remarkable integration among the metabolic processes that regulate plasma glucose concentration. This association is presumably mediated by plasma glucose itself, whereby a small change in plasma glucose concentration due to a change in glucose production or a change in plasma glucose clearance triggers a series of metabolic events that involve alterations in insulin and glucagon secretion, and both insulin/glucagon-mediated and insulin/glucagon-independent alterations in glucose production and glucose clearance to maintain euglycemia (18–21). However, this finely tuned process is disrupted and results in fasting hyperglycemia when there is concomitant insulin resistance and beta-cell dysfunction, as observed in our Ob100–125 and Ob≥126 groups.

Our study cannot determine the precise mechanism responsible for the impairment in basal plasma glucose clearance in our Ob100–125 and Ob≥126 groups. However, the results from our study and those from previous studies suggest insulin resistance in both skeletal muscle and splanchnic tissues, including the liver, is likely involved. About 75% of basal glucose disposal occurs via insulin-independent facilitated diffusion in tissues that require glucose as an energy source (primarily the central and peripheral nervous systems), while most of the remaining 25% of basal glucose disposal occurs in skeletal muscles and splanchnic tissues via both insulin-dependent and insulin-independent mechanisms (5–7). The results from previous studies have shown that insulin-independent glucose disposal is normal (6) and the ability of an increase in plasma glucose to stimulate glucose disposal while plasma insulin is clamped is not impaired (22) in people with type 2 diabetes and fasting hyperglycemia. On the other hand, type 2 diabetes is associated with insulin resistant glucose uptake in both skeletal muscle and splanchnic tissues, including the liver (8, 23–25).Together, these data suggest that impaired insulin-mediated skeletal muscle and splanchnic glucose clearance was responsible for the blunted increase in glucose clearance relative to glucose Ra observed in our Ob100–125 and Ob≥126 groups. The lower glucose clearance rate in the Ob≥126 than the Ob100–125 group, even though both groups had nearly the same values for skeletal muscle and hepatic insulin sensitivity, was likely due to lower plasma insulin concentrations in the Ob≥126 than the Ob100–125 group, which demonstrates an important interaction between beta-cell function and insulin sensitivity in determining on basal glucose clearance.

CONCLUSION

Fasting hyperglycemia in people with obesity is caused by an abnormally low basal rate of plasma glucose clearance in relation to the basal rate of endogenous glucose production. The precise mechanism responsible for the disruption of this carefully regulated balance in glucose homeostasis is not clear, but is presumably related to an inadequate compensatory increase in plasma insulin concentration needed to overcome insulin resistance.

RESEARCH HIGHLIGHTS/STUDY IMPORTANCE.

What is already known about this subject?

• Fasting plasma glucose concentration is determined by the rate of glucose production (i.e., appearance rate in plasma) and the plasma glucose clearance rate (i.e., the volume of plasma cleared of glucose per unit time).

• The relative importance of glucose production and plasma glucose clearance in determining fasting plasma glucose concentration in people with different fasting glycemic status is unknown.

What are the new findings in this manuscript?

• An attenuation in glucose clearance is primarily responsible for fasting hyperglycemia in people with prediabetes and type 2 diabetes.

• Fasting hyperglycemia is caused by a disruption of the precisely regulated integration of glucose production and clearance, presumably due to insulin resistance with an inadequate compensatory increase in insulin secretion.

How might the results change the direction research of the focus of clinical practice?

• The results from our study support a focus on both endogenous glucose production and plasma glucose clearance as targets for alleviating fasting hyperglycemia.