Postprandial Insulin Response and Clearance Among Black and White Women: The Federal Women’s Study

Abstract

Context

Postprandial hyperinsulinemia might be an important cardiometabolic risk determinant in black compared with white women. However, the contributions of insulin clearance and β-cell function to racial differences in postprandial insulin response are unknown.

Objective

To compare, by race and menopause, early insulin response to oral and intravenous glucose and to measure postprandial intact glucagon-like peptide 1 (GLP-1) concentrations, insulin clearance, and β-cell function.

Design and Participants

119 federally employed women without diabetes [87 premenopausal (52 black, 35 white) and 32 postmenopausal (19 black, 13 white)] underwent an oral glucose tolerance test, insulin-modified frequently sampled intravenous glucose test (IM-FSIGT), and mixed meal tolerance test (MMTT).

Outcome Measures

Early insulin response was measured as follows: (i) insulinogenic index (oral glucose tolerance test); (ii) acute insulin response to glucose (IM-FSIGT); and (iii) ratio of incremental insulin/glucose area under the curve in the first 30 minutes of the MMTT. Insulin clearance was assessed during the IM-FSIGT and MMTT. During the MMTT, intact GLP-1 was measured and β-cell function assessed using the insulin secretion rate and β-cell responsivity indexes.

Results

Black pre-menopausal and postmenopausal women had a greater insulin response and lower insulin clearance and greater dynamic β-cell responsivity (P ≤ 0.05 for all). No differences were found in the total insulin secretion rates or intact GLP-1 concentrations.

Conclusions

Greater postprandial hyperinsulinemia in black pre-menopausal and postmenopausal women was associated with lower hepatic insulin clearance and heightened β-cell capacity to rapid changes in glucose, but not to higher insulin secretion. The relationship of increased β-cell secretory capacity, reduced insulin clearance, and ambient hyperinsulinemia to the development of cardiometabolic disease requires further investigation.

A greater postprandial insulin response, relative to glucose, in black compared with white women was associated with lower insulin clearance but not increased insulin secretion.

In the United States, black women have a twofold increased risk of type 2 diabetes and a 1.3-fold increased risk of heart disease compared with white women (1). Multiple socioenvironmental factors might contribute to this greater cardiometabolic disease risk. However, emerging evidence has suggested that hyperinsulinemia (high insulin relative to ambient glucose concentrations) might be an important mediating risk determinant (2). Hyperinsulinemia is a compensatory response to increasing insulin resistance and results from increased insulin secretion (enhanced β-cell responsiveness) or reduced insulin clearance, or both. Whether hyperinsulinemia directly mediates the greater cardiometabolic disease risk in black women is uncertain but is of high clinical significance because hyperinsulinemia, relative to insulin resistance, is a prominent feature in black individuals of African ancestry (3). For example, using intravenous (nonphysiologic) or oral glucose challenges, black individuals (African Americans and black Africans) had greater insulin concentrations compared with individuals of European or Asian ancestry (3–11). However, the underlying etiology for the greater insulin response in black compared with white adults under physiologic conditions is unclear, because only a few studies have examined the insulin response after a meal, and these studies were performed in pubertal youth and not in an adult population with a greater risk of cardiovascular disease (12–14). Clarifying racial differences in the insulin response to physiologic and nonphysiologic conditions could help to elucidate the mechanisms contributing to greater cardiometabolic disease risk in black populations.

Reduced hepatic insulin clearance is an important contributor to hyperinsulinemia, because the liver and regulation of insulin-degrading enzyme (IDE) accounts for most of the insulin elimination from the body (15, 16). The greater insulin response in black compared with white individuals has mostly been attributed to reduced insulin clearance rather than an increase in insulin secretion (9, 10, 17, 18). However, reduced insulin clearance has largely been observed in the context of overall lower insulin resistance in black individuals, and the study conclusions could have been confounded by a greater prevalence of abnormal glucose tolerance in the black compared with the white groups (8, 10, 18). The metabolic phenotype in women of African ancestry is also characterized by less visceral and hepatic fat content compared with white women, despite a similar body mass index (19). A lower hepatic fat content has been more commonly associated with an improved metabolic profile (and greater hepatic insulin clearance). However, whether this relationship of hepatic fat with insulin clearance differs by race is unknown.

An increased β-cell secretory capacity has also been posited as a major determinant of hyperinsulinemia in blacks, and increased insulin secretion has been observed using the hyperglycemic clamp (the reference standard measure of insulin secretory capacity) (5, 20). Further evidence supporting greater β-cell responsivity in blacks has come from studies measuring gastric inhibitory polypeptide (GIP) and glucagon-like peptide 1 (GLP-1), gastrointestinal incretin hormones that mediate 50% to 80% of postprandial insulin secretion (21). The greater insulin concentrations in blacks compared with whites were associated with a 30% greater increase in GLP-1 concentrations in black adults (22, 23). However, a consistent relationship between incretin levels and hyperinsulinemia is lacking, with little to no increase in GLP-1 or GIP concentrations found in youth after an oral or meal glucose challenge (12, 24, 25). Because age and menopause are associated with increasing insulin resistance and alterations in incretin response, these contradictory findings raise the question of whether age- and/or menopause-related increases in insulin resistance could be confounding the assessment of racial differences in incretin and insulin response (26, 27).

To fully appreciate the potential racial differences in hyperinsulinemia and its relationship to cardiometabolic disease risk, it is also important to measure the insulin response under physiologic conditions [i.e., after a meal when the enteroinsular axis has been optimally activated and physiologic insulin-mediated suppression of free fatty acid (FFA) concentrations, a marker of adipose-tissue insulin resistance], can be assessed. Previously, in a small group of black compared with white premenopausal women, we observed a multiphasic pattern of glucose concentrations and a trend toward higher insulin and total FFA concentrations after a mixed-meal tolerance test (MMTT) (14). In the present study, we sought to elucidate the spectrum of insulin response to both oral and intravenous glucose stimuli and used a MMTT to maximally stimulate the incretin response. Our objectives were to compare the insulin response to three separate glucose challenges [oral glucose tolerance test (OGTT), insulin-modified intravenous glucose test (IM-FSIGT), and MMTT] in black women of African ancestry (African American and African immigrant) with white women and evaluate incretin concentrations, insulin clearance, and β-cell function [insulin secretion rate (ISR) and β-cell responsivity] postprandially.

Research Design and Methods

Study cohort

We recruited women to the Federal Women’s Study (ClinicalTrials.gov identifier, NCT01809288), a trial designed to examine the racial/ethnic variations in the risk of diabetes and heart disease in African, African-American, and white women who were federally employed and lived in the Washington DC metropolitan area (28). Recruitment was achieved by newspaper advertisement, flyers, and referral by previous participants. The participant flow diagram is shown in Supplemental Fig. 1: 147 were screened and 119 women were enrolled. Black women of African ancestry were defined as African American (parents and participant identified as African American born in the United States) or African (parents and participant born in Africa). White women were those who had identified themselves and both parents as white. The participants self-identified as healthy, did not report a history of diabetes, and were not taking medications that influenced glucose or lipid metabolism. They had also undergone a history, a physical examination, urinalysis, electrocardiography, and routine laboratory tests to exclude chronic illnesses and anemia. Menopause was defined as having irregular periods and/or an FSH >21 U/L. The National Institute of Diabetes, Digestive and Kidney Diseases institutional review board approved the present study. All enrollees gave written informed consent before participation and were seen at the National Institutes of Health Clinical Center (Bethesda, MD).

Study protocol

After the screening visit, the women returned to the National Institutes of Health Clinical Center after a 10- to 12-hour overnight fast for each of the three visits (OGTT, IM-FSIGT, and MMTT), 7 to 14 days apart, and completed within a 4- to 6-week period. The OGTT, IM-FSIGT, and MMTT were completed consecutively, except for in 10 women (five black and five white) who had undergone the MMTT before the IM-FSIGT because of scheduling difficulties. During the standard 75-g OGTT, the plasma glucose, insulin, and C-peptide concentrations were measured at −15, 0, 30, 60, 90, and 120 minutes after glucose ingestion. Glucose tolerance status was defined according to the American Diabetes Association guidelines. Prediabetes was defined as fasting glucose ≥100 but <126 mg/dL and/or 2-hour glucose level ≥140 but <200 mg/dL (29). During the IM-FSIGT visit, one intravenous catheter was placed in each arm near or in the antecubital veins. Baseline blood samples were obtained, and dextrose (0.3 g/kg) was administered intravenously. Insulin was given as a bolus at 20 minutes (0.03 U/kg/min). As previously described, plasma glucose and insulin concentrations were obtained at 32 time points between baseline and 180 minutes (7). For the MMTT, the test meal consisted of a bagel with cream cheese, a cheese omelet, and orange juice to meet 33% of the estimated daily energy needs (52% carbohydrate, 15% protein, 33% fat), which was calculated using the Mifflin St. Jeor equation plus an activity factor of 1.5 (30). All women completed the breakfast meal within 20 minutes (mean, 13.8 ± 4.9 minutes). Blood samples were taken for glucose, insulin, and C-peptide concentrations at 0, 10, 20, 30, 40, 50, 60, 90, and 120 minutes. In a subgroup of 64 premenopausal women (40 black and 24 white) and 18 postmenopausal women (9 black and 9 white), additional samples were collected for intact GLP-1, GIP, and glucagon analysis.

Calculations and variables

OGTT

The early insulin response was calculated using the insulinogenic index (31).

IM-FSIGT

The acute insulin response to glucose (AIRg) was defined as the insulin area under the curve (AUC) greater than basal at 0 to 10 minutes (32). The insulin sensitivity index (S_I) was calculated via the minimal model (MinMOD Millenium, version 6.02) (32), and the disposition index was calculated as the product of the AIRg and insulin S_I. Hepatic insulin clearance was estimated as the basal and total fractional hepatic insulin extraction using a model for clearance from the liver and the periphery (33), combined with the classic C-peptide deconvolution method to estimate the ISR (34).

MMTT

The early insulin response was calculated as the ratio of insulin/glucose concentrations for the incremental area under the respective curves for the first 30 minutes during the MMTT $\left(\frac{Insulin iAUC30}{Glucose iAUC30}\right)$. Insulin clearance was calculated using two methods: (i) the molar ratio of C-peptide/insulin concentrations for the incremental area under the respective curves during the first 30 minutes of the test $\left(\frac{C-peptide iAUC30}{Insulin iAUC30}\right)$ (35); and (ii) the fractional hepatic insulin extraction from the three-compartment model for insulin kinetics [liver, plasma, and extravascular; model VII in the study by Piccinini et al. (36)]. The total ISR (ISR_total) during the MMTT was computed using the standard two-compartment model for C-peptide kinetics (34). The β-cell responsivity indexes were estimated from plasma glucose and C-peptide concentrations measured during the MMTT using the oral C-peptide minimal model (36). In brief, this C-peptide secretion model provides two β-cell responsivity indexes of pancreatic secretion: (i) Ф_d, the dynamic component that likely represents secretion of promptly releasable insulin and is proportional to the glucose rate of increase through the constant; and (ii) Ф_s, the static component that is proportional to delayed glucose concentration and likely represents the new insulin entering into the releasable pool (36). The ISR_total was partitioned into its three components: basal, dynamic, and static, as previously described (36). All modeling calculations were performed using MATLAB R2017b (Natick, MA).

Because of technical difficulties with sample collection, results were unavailable for the following: insulinogenic index (n = 2), S_I (n = 8), IM-FSIGT insulin clearance (n = 22), MMTT fractional hepatic extraction (n = 10), and β-cell responsivity indexes (n = 16).

Biochemical analyses

Glucose, insulin, and C-peptide concentrations were measured in serum using the Roche Cobas 6000 analyzer (Roche Diagnostics, Indianapolis, IN). Samples for incretin, glucagon, and total FFA analysis were collected in EDTA tubes with protease inhibitor, immediately centrifuged, and the plasma was stored at −80°C until analysis. Total GIP and intact GLP-1 concentrations were measured with Mesoscale ELISA kits [Gaithersburg, MD; GIP detection range, 73.4 to 2500 pg/mL; intact GLP-1-(7-36) detection range, 0.9 to 1000 pg/mL]. Glucagon was measured using a Mercodia kit (Winston-Salem, NC; glucagon detection range, 1.5 to 120 pmol/L). The total FFA concentrations were measured using an in vitro enzymatic colorimetric assay (Wako Diagnostics, Mountain View, CA).

Imaging studies

Whole body composition measurements were performed using dual-energy X-ray absorptiometry (Hologic Discovery, Bedford, MA). Hepatic fat content was measured using proton magnetic resonance spectroscopy at 3T (MAGNETOM Verio; Siemens, Tarrytown, NY) (37). Additionally, the visceral fat measurement was obtained using T1-weighted images at the level of L2-L3 (38). Because of difficulties with data acquisition, we were unable to obtain visceral fat in 14 black and 4 white women and hepatic fat in 17 black and 8 white women.

Statistical analysis

The data are presented as the mean ± SD, except as indicated otherwise. An early response was defined as the incremental AUC during the first 30 minutes of the MMTT (iAUC₃₀), calculated using a cubic spline to determine the integral (STATA, version 15.1; StataCorp, College Station, TX). The Student t test and Kruskal-Wallis test were used to compare the continuous parametric and nonparametric variables, respectively. The χ² test was used to compare categorical variables. Spearman correlations were performed for continuous variables. Mixed effects models were used to analyze the glucose, hormone, and FFA response during the MMTT (within-subject factor of time; between-subject factor of race; time × race interaction). Statistical analyses were performed with STATA, version 15.1 (StataCorp), and P values < 0.05 were considered statistically significant.

Results

Of the 119 women enrolled, 87 were premenopausal (13 African immigrant, 39 African American, and 35 white) and 32 were postmenopausal (two African immigrant, 17 African American, and 13 white; Supplemental Fig. 1). Except for body mass index, African-immigrant and African-American women had similar demographic and metabolic characteristics (P > 0.05 for all; Supplemental Table 1) and therefore were combined into one group for the subsequent analyses.

Participant characteristics

The participant characteristics stratified by race and menopausal status are listed in Table 1. Overall, black women had lower triglyceride concentrations and lower visceral fat and tended to have lower hepatic fat. To determine the effect of menopause on metabolic characteristics, we compared the black pre- and postmenopausal women and white pre-menopausal and postmenopausal women (Table 1). Black premenopausal women had a lower fasting glucose, 2-hour glucose, triglyceride concentrations, and a lower prevalence of prediabetes compared with black postmenopausal women (P ≤ 0.01 for all). A similar metabolic pattern was observed between white pre-menopausal and postmenopausal women (Table 1). Therefore, subsequent analyses were stratified by menopausal status.

Table 1.

Participant Characteristics Stratified by Race and Menopausal Status

Characteristic	Entire Cohort (n = 119)			Premenopausal Women (n = 87)			Postmenopausal Women (n = 32)			P Value (Pre- vs Postmenopausal)
										Black (n = 71)	White (n = 48)
	Black (n = 71)	White (n = 48)	P Value	Black (n = 52)	White (n = 35)	P Value	Black (n = 19)	White (n = 13)	P Value
Age, y	42.3 ± 9	43.3 ± 11	0.63	39 ± 7	38 ± 7	0.70	53 ± 4	57 ± 4	< 0.01a	< 0.01a	< 0.01a
BMI, kg/m2	30.6 ± 5.6	29.3 ± 4.9	0.20	30.7 ± 6.0	29.9 ± 5.0	0.48	30.1 ± 4.6	27.7 ± 4.4	0.15	0.66	0.17
Obesity, n (%)	35 (49)	23 (48)	0.88	28 (54)	19 (54)	0.97	7 (37)	4 (31)	0.72	0.21	0.15
Prediabetes, n (%)	23 (33)	13 (27)	0.54	10 (19)	6 (17)	0.81	13 (68)	7 (54)	0.40	< 0.01a	0.01a
Total fat mass, %	33.5 ± 13.2	32.3 ± 9.6	0.62	37.2 ± 6.9	39.1 ± 7.0	0.21	38.6 ± 5.4	37.2 ± 6.1	0.50	0.44	0.39
Visceral fat L2-3, cm2	67 (42–111)	95 (67–138)	0.01a	58 (34–104)	97 (66–132)	0.02a	89 (58–128)	92 (68–149)	0.54	0.04a	0.81
Subcutaneous fat L2-3, cm2	265 (197–407)	291 (193–344)	0.96	265 (177–407)	298 (191–358)	0.69	277 (218–375)	240 (194–337)	0.48	0.64	0.48
Hepatic fat, %)	0.7 (0.4–1.5)	1.0 (0.6–2.2)	0.09	0.6 (0.5–1.3)	1.0 (0.7–2.5)	0.04a	1.2 (0.4–3.7)	0.9 (0.5–1.7)	0.90	0.45	0.58
Fasting lipid profile, mg/dL
Total cholesterol	168 ± 33	171 ± 28	0.63	162 ± 30	163 ± 23	0.94	183 ± 36	192 ± 32	0.47	0.02a	0.01a
LDL cholesterol	95 ± 29	99 ± 26	0.51	91 ± 29	93 ± 23	0.75	107 ± 29	114 ± 29	0.46	0.04a	0.01a
HDL cholesterol	60 ± 17	56 ± 14	0.21	60 ± 17	55 ± 14	0.18	62 ± 18	60 ± 16	0.78	0.69	0.30
Triglycerides	57 (41–69)	70 (58–98)	< 0.01a	53 (41–67)	68 (56–99)	< 0.01a	61 (52–78)	75 (67–94)	0.06	0.04a	0.27
OGTT
Fasting glucose, mg/dL	92 ± 7	92 ± 7	0.67	90 ± 7	91 ± 7	0.55	95 ± 7	95 ± 7	0.89	0.01a	0.12
2-h Glucose, mg/dL	125 ± 24	127 ± 28	0.65	121 ± 24	124 ± 27	0.60	135 ± 21	136 ± 32	0.99	0.03a	0.22
IM-FSIGT
SI (mU/L)−1min−1 × 10−4	2.3 (1.7–3.2)	3.0 (2.1–5.2)	< 0.01a	2.3 (1.6–3.2)	2.8 (1.8–4.4)	0.06	2.4 (1.7–3.0)	4.5 (2.4–6.0)	0.04a	0.95	0.29
Disposition index	1676 (1118–2776)	1331 (972–1758)	0.02a	1867 (1255–2780)	1481 (972–1817)	0.02a	1240 (930–1967)	1234 (970–1557)	0.47	0.08	0.28

Data presented as mean ± SD or median (25th to 75th percentile).

Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein.

^aStatistically significant.

Among the premenopausal women, the prevalence of obesity and prediabetes, glucose concentrations, and fasting cholesterol profile were similar in black and white women (Table 1). Black premenopausal women had lower fasting triglyceride concentrations and lower visceral and hepatic fat content compared with the white women (P < 0.05 for all). In the postmenopausal women, blacks were 4.8 ± 1.5 years younger and tended to have lower fasting triglycerides but otherwise had similar rates of obesity and prediabetes, a similar body composition, and similar glucose concentrations.

S_I was lower in both premenopausal (P = 0.06) and postmenopausal (P = 0.04) black compared to white women (Table 1). The disposition index was greater in black premenopausal women only (P < 0.02; Table 1).

During the MMTT, black women compared with white women had greater insulin and lower glucose concentrations (at 20 to 40 minutes), regardless of menopausal status [Fig. 1(a)–(d)]. No differences were found in the C-peptide concentrations [Fig. 1(e) and 1(f)]. Intact GLP-1 and GIP concentrations were similar in black and white pre-menopausal and postmenopausal women (Table 2; Supplemental Fig. 2). Glucagon concentrations were similar in premenopausal women (P = 0.93); however, postmenopausal white women had lower glucagon concentrations during the first 30 minutes of the MMTT (P = 0.03; Table 2). The postprandial FFA concentrations were greater in the black pre- and postmenopausal women; however, this was only statistically significant at 30 minutes after the meal (Supplemental Fig. 3).

Figure 1.

The change in (a, b) glucose, (c, d) insulin, and (e, f) C-peptide concentrations during the MMTT in black (solid squares and lines) and white (white circles and dotted lines) pre-menopausal and postmenopausal women. Data presented as mean ± SEM.

Table 2.

Early Incretin and Glucagon Response in Pre-menopausal and Postmenopausal Women Stratified by Race During the MMTT

Variable	Premenopausal (n = 64)			Postmenopausal (n = 18)
	Black (n = 40)	White (n = 24)	P Value	Black (n = 9)	White (n = 9)	P Value
Intact GLP-1 iAUC30, pg/mL/min	101 (62–167)	92 (32–144)	0.41	53 (32–84)	77 (33–115)	0.57
Total GIP iAUC30, pg/mL/min	2600 (1355–3636)	1687 (1238–3044)	0.34	3184 (1862–3654)	1872 (1338–2941)	0.38
Glucagon iAUC30, pmol/L/min	47 (26–74)	45 (16–75)	0.93	41 (26–86)	17 (6–23)	0.03

Early response was defined as the iAUC₃₀ of the meal test.

Data presented as mean ± SD or median (25th to 75th percentile).

Insulin response

Black women had a greater insulin response, relative to glucose, during all three tests (Fig. 2). The insulinogenic index (OGTT) and AIRg (IM-FSIGT) was 40% to 50% and 80% greater in black pre-menopausal and postmenopausal women, respectively [P ≤ 0.01; Fig. 2(a) and (2b)]. Similarly, during the MMTT, black compared with white women had 2.5-fold greater early postprandial insulin to glucose response [Fig. 2(c)].

Figure 2.

Early insulin response to three different glucose stimuli. Early response was defined as (a) insulinogenic index during OGTT, (b) AIRg during IM-FSIGT, and (c) insulin iAUC₃₀/glucose iAUC₃₀ during MMTT. Data are presented as Tukey box and whiskers plots. Black circles represent any value that lie outside the range of the whiskers.

Insulin clearance, secretion, and β-cell responsivity

Using modeling parameters, the basal and total fractional hepatic insulin extraction was ~11% and 9% lower in black women during the IM-FSIGT and the MMTT, respectively (P ≤ 0.01; Fig. 3). During the MMTT, the insulin clearance ratio (C-peptide iAUC₃₀/insulin iAUC₃₀) was also 20% lower in the black compared with white, pre-menopausal and postmenopausal women and correlated positively with both basal fractional hepatic insulin extraction (IM-FSIGT: r = 0.35, P < 0.001; MMTT: r = 0.75, P < 0.001; Supplemental Fig. 4). The absolute ISR was not different when stratified by race or menopausal status [Fig. 4(a) and 4(b)]. Dynamic β-cell responsivity (Ф_d) and dynamic ISR was greater in black women [P < 0.01; Fig. 4(c) and 4(e)]. No differences were found in static β-cell responsivity (Ф_s), static ISR, or basal ISR [Fig. 4(d) and 4(e)].

Figure 3.

Basal and total fractional hepatic insulin extraction during IM-FSIGT and MMTT in black (black bars) and white (white bars) women. Data presented as mean ± SEM.

Figure 4.

Model estimates of insulin secretion and β-cell function during the MMTT. ISR in black (blue lines) and white (green lines) (a) premenopausal and (b) postmenopausal women. β-cell responsivity in black (gray boxes) and white (white boxes) women by menopausal status: (c) dynamic and (d) static. (e) The total insulin secretion AUC rate during the MMTT partitioned into its components: ISR_basal (black box), ISR_static (gray box) and ISR_dynamic (white box). Data presented as (a, b) line graphs, (c, d) Tukey Box and whisker plots, and (e) bar graphs. Black circles represent any value that lie outside the range of the whiskers. *P < 0.01 premenopausal white vs black women.

Correlations

Hepatic fat correlated with the MMTT early insulin response regardless of race or menopausal status (r = 0.33 and P < 0.001 for all). The relationship of MMTT basal hepatic insulin extraction with hepatic fat differed by race. The strong negative correlation observed in whites was absent in black women (Fig. 5). The early GIP and intact GLP-1 response correlated with insulin secretion (r = 0.42, P < 0.001 and r = 0.30, P < 0.001, respectively) but not with Ф_d or Ф_s (r ≤ 0.10 and P ≥ 0.17 for all). The FFA level at 30 minutes did not correlate with the insulin response (r = 0.16, P = 0.11).

Figure 5.

The relationship of basal fractional hepatic insulin extraction during the MMTT and hepatic fat in (a) black and (b) white women. Data presented as individual data points for black (black squares) and white (white circles) women with the corresponding regression line.

Discussion

Greater hyperinsulinemia in black compared with white pre-menopausal and postmenopausal women

The present study has provided insight into racial differences in insulin response in a large group of women stratified by race and menopausal status. We have demonstrated that blacks have a greater insulin response to oral (liquid glucose drink and solid mixed meal) and intravenous glucose stimuli compared with white women and confirmed that racial differences in insulin response are detectable in both pre-menopausal and postmenopausal women. Although greater hyperinsulinemia had been previously observed during intravenous glucose tests, the postprandial insulin response was less clear (5, 12, 14, 22–24, 27). Our study has extended these observations by confirming that hyperinsulinemia is also prominent under physiologic conditions in response to a mixed meal, regardless of menopausal status. Additionally, by including two groups of women of African descent, we found that greater insulin response is likely related to intrinsic biological factors that occur across the African diaspora. Clinically, our findings imply that interventions to target greater hyperinsulinemia in blacks in the United States could be applicable to other populations of African descent.

Postprandial hyperinsulinemia is related to racial differences in insulin clearance

To elucidate the physiologic mechanisms leading to greater hyperinsulinemia in black women, we estimated the fractional hepatic insulin extraction, ISR, and β-cell responsivity (a measure of secretory capacity) after the solid breakfast meal. Black women had lower basal and total hepatic insulin extraction estimated by two independent measures during the IM-FSIGT and MMTT. These findings confirmed the lower hepatic insulin extraction during the IM-FSIGT in a large group of black pre-menopausal and postmenopausal women and are in agreement with previous studies (5, 17, 39). Moreover, by also demonstrating lower basal and total hepatic insulin extraction in black women during the MMTT, we found that a lower insulin clearance was also an important physiologic feature after a solid meal (Fig. 3). As expected, the lower insulin clearance in black women was in the context of a lower S_I. However, the absolute racial difference in insulin clearance was only 10%, which would not fully explain the variance in hyperinsulinemia observed in pre-menopausal and postmenopausal women.

Because greater insulin secretion has also been proposed as a key contributor to the racial differences in hyperinsulinemia (5, 20, 25, 27), we also modeled the ISR and partitioned pancreatic secretion into its three components: basal, dynamic, and static. The total ISR and the pattern of insulin secretion were similar when stratified by race and menopause status (Fig. 4). However, the glucose concentrations were lower in the black women during the first 40 minutes of the MMTT, implying that they had greater β-cell responsivity. The dynamic insulin secretion was also higher in black women but no difference was found in basal or static insulin secretion. The higher dynamic responsivity (Ф_d), defined as the β-cell’s secretory capacity proportional to the rate of increase of glucose concentrations, posits that black women have a larger quantity of insulin in the rapidly releasable pool compared with white women. These findings are in keeping with the greater AIRg after an IM-FSIGT in black women compared with white women, a well-established phenotypic characteristic that has been replicated in the present study. The lack of difference by race in static responsivity implies that the capacity to deliver new insulin granules to the plasma membrane in response to maintained increases in glucose is similar. This component is the dominant source of the cumulative release of insulin during the MMTT. These findings have corroborated the greater β-cell function observed in black compared with white pubertal youth using an alternative C-peptide and glucose mathematical model during an OGTT (25) but argue against greater total insulin secretion as a major determinant of the greater postprandial hyperinsulinemia in black women.

Postprandial hyperinsulinemia is not related to racial differences in incretin response

In contrast to our hypothesis, the greater insulin response and greater dynamic β-cell responsivity in black women were not related to greater postprandial incretin concentrations. Previous studies that found a greater insulin and intact GLP-1 response were limited by small sample sizes and had been performed in response to an OGTT in which incretin secretion might not have been maximally stimulated (22, 23). Rather, our results are in agreement with the findings from youth, in which a greater early insulin response in black compared with white youth was associated with lower or no differences in incretin responses after either an oral glucose load or a mixed macronutrient meal (12, 24, 25).

Relationship of postprandial insulin response to hepatic fat and FFA

Because the liver accounts for ~80% of insulin clearance (15), increased hepatic and visceral adiposity are typically close correlates of hyperinsulinemia and reduced insulin clearance (40, 41). This lower insulin clearance in blacks has been paradoxically associated with lower visceral and hepatic fat accumulation (19), a finding replicated in our cohort. The relationship of hepatic insulin clearance with hepatic fat differed by race and was most prominent in premenopausal women. The expected inverse correlation of hepatic fat with insulin extraction was observed in white women but not in the black women (Fig. 4). Therefore, black women had lower hepatic insulin extraction across a wide range of hepatic fat content compared with white women. Other intrinsic factors, such as racial differences in IDE activity, may play a role in the differential regulation of hepatic insulin clearance (17). IDE is an evolutionarily conserved zinc metalloprotease whose activity is regulated by the liver, suppressed in obesity, and associated with reduced insulin clearance (16, 42, 43). Polymorphisms in the gene encoding IDE have been variably linked to an increased risk of type 2 diabetes, although these studies have been in largely Asian and white populations (43–47). The effects of obesity and liver fat accumulation on IDE activity and whether these factors mediate racial variations among individuals of African descent remain to be elucidated.

Greater postprandial hyperinsulinemia was also observed in the context of lower overall insulin sensitivity (S_I) and could indicate greater tissue-specific insulin resistance. The higher FFA concentrations during the MMTT in black women (Supplemental Fig. 3), likely reflected resistance to insulin-mediated suppression of adipose tissue lipolysis, because the FFA rate of appearance from the diet is negligible during the first 2 hours of the MMTT (14). These observations validate previous findings, in a smaller biracial group of 15 black and 13 white women, in which the modeling of FFA kinetics demonstrated reduced suppression of lipolysis in black women (14) and underscore the importance of assessing hormonal markers in the physiologic state, rather than during intravenous glucose tests (7, 14).

Overall, the cumulative effect of ambient hyperinsulinemia on long-term cardiometabolic risk remains to be elucidated. However, the greater postprandial insulin response in black women was not associated with greater glucose, triglyceride, or cholesterol concentrations or a greater prevalence of obesity or prediabetes in the present study and a previous cross-sectional study found no relationship of impaired postprandial endothelial function in insulin-resistant women (black or white) with an exaggerated insulin response (48). Nevertheless, the collective effect of ambient hyperinsulinemia, in the context of reduced insulin clearance and increased β-cell capacity to respond to rapid increases in glucose, is unknown. Recent murine studies have found that hyperinsulinemia alone, without insulin resistance, increases heart failure risk by inducing cardiac hypertrophy (49, 50).

Strengths and weaknesses

To the best of our knowledge, the present study is the first to systematically compare by race the contribution of insulin secretion and clearance to postprandial hyperinsulinemia in women who were similar in glucose tolerance status and stratified by menopause. Detailed modeling analyses of fractional hepatic extraction were used in two separate tests, and the magnitude of lower insulin clearance in black women was similar. The study limitations included the cross-sectional design and relatively smaller sample of postmenopausal women. Additionally, we did not measure gastric emptying, an important factor for the rate of glucose absorption and postprandial insulin response; however, the meal completion time was similar when stratified by race and menopausal status.

Conclusions

Compared with white women, black pre-menopausal and postmenopausal women had a greater insulin response to oral and intravenous glucose loads. Lower insulin clearance and a heightened β-cell capacity to secrete insulin in response to rapid changes in glucose concentrations were associated with this greater postprandial hyperinsulinemia, and the etiological mechanisms and cardiovascular implications of this response require further investigation.

Glossary

Abbreviations:

AIRg

acute insulin response to glucose

AUC

area under the curve

FFA

free fatty acid

iAUC₃₀

incremental area under the curve during the first 30 minutes of the mixed meal tolerance test

GLP-1

glucagon-like peptide 1

IDE

insulin-degrading enzyme

IM-FSIGT

insulin-modified frequently sampled intravenous glucose test

ISR

insulin secretion rate

ISR_total

total insulin secretion rate

MMTT

mixed meal tolerance test

OGTT

oral glucose tolerance test

S_I

insulin sensitivity index