Short-Term Exercise Improves β-Cell Function and Insulin Resistance in Older People with Impaired Glucose Tolerance

Abstract

Background: There is a high prevalence of diabetes and impaired glucose tolerance (IGT) in the older population. Normal aging is associated with insulin resistance and impaired insulin secretion, with greater defects in people with IGT. Short-term exercise has been found to increase insulin sensitivity, but little is known about acute exercise effects on β-cell function in older people with IGT.

Methods: We assessed the effects of 7 consecutive days of supervised aerobic exercise (1 h/d at 60–70% heart rate reserve) in 12 sedentary older people with IGT. Screening included oral glucose tolerance test, stress/maximal O₂ uptake test, and dual-energy x-ray absorptiometry scan. Participants had a frequently sampled iv glucose tolerance test at baseline and 15–20 h after the seventh exercise session. Insulin sensitivity (S_I), glucose disappearance constant (Kg, a measure of iv glucose tolerance), acute insulin response to glucose (AIRg), and disposition index (AIRg × S_I), a measure of β-cell function in relation to insulin resistance, were calculated.

Results: Exercise was well tolerated. Body weight, fasting glucose, fasting insulin, and iv glucose tolerance were unchanged with exercise. S_I increased by 59%, AIRg decreased by 12%, and disposition index increased by 31%. There was no significant change in fasting lipid, catecholamine, leptin, or adiponectin levels.

Conclusions: Short-term exercise not only improved insulin resistance but also significantly enhanced β-cell function in older people with IGT. These effects of short-term exercise on β-cell function cannot be explained by changes in body weight or circulating levels of lipids, leptin, adiponectin, or catecholamines.

The prevalence of type 2 diabetes and impaired glucose tolerance (IGT) increases with age (1). More than 45% of the older population meet current diagnostic criteria for type 2 diabetes and IGT (2, 3). Multiple risk factors for type 2 diabetes associated with aging including increased adiposity and decreased physical activity predispose older people to develop glucose intolerance and increased insulin resistance. The progression from normal glucose tolerance to IGT and type 2 diabetes is characterized by progressive defects in β-cell function or impaired β-cell compensation for insulin resistance (4). Impaired insulin secretion has been demonstrated, even with normal aging (1, 5, 6). Longitudinal studies indicate that progressive impairments in β-cell function are critical in the progression from normal glucose tolerance to IGT and diabetes (4). Impairments in β-cell sensitivity to glucose and β-cell compensation for insulin resistance have been found in older people with normal glucose tolerance with greater defects in older people with IGT (6).

Recent diabetes prevention studies including the Diabetes Prevention Program (DPP) in the United States have included older people with IGT. In the DPP, 20% of 3234 participants were age 60 yr or older. In this trial, lifestyle (diet and exercise) intervention significantly reduced the overall incidence of diabetes by 58% vs. placebo (7). The lifestyle intervention was particularly effective in older people, with a 71% reduction in diabetes incidence. Lifestyle treatment was felt to improve both insulin sensitivity and insulin secretion (independent of weight reduction) to a greater degree than placebo or metformin treatment per gross assessment based on fasting and 30-min glucose/insulin levels in a subset of DPP participants at yr 1 (8). However, the mechanisms of lifestyle intervention to prevent diabetes, and in particular the effects on β-cell function in the older population, remain unclear.

Exercise alone has been shown to improve insulin sensitivity in numerous studies (9–13). Six- and 9-month endurance exercise in healthy older people (with significant weight loss) increased insulin action and, as expected, decreased glucose-stimulated insulin secretion (9, 14), given the hyperbolic relationship between insulin sensitivity and insulin secretion (15, 16). However, exercise effects on β-cell function (in relation to changes in insulin sensitivity) in older people with IGT and without confounding effects of weight loss have not been examined.

Seven-day aerobic exercise studies have also been performed with findings of increased insulin sensitivity, including in healthy older people (12, 17), but without associated changes in body weight/composition or maximal O₂ uptake (12, 17–19). However, these prior 7-d exercise studies did not define glucose tolerance and/or did not use current diagnostic criteria for glucose tolerance and did not examine β-cell function or older people with IGT. In this project, we examined the acute effects of 7 d of aerobic exercise on β-cell function and insulin sensitivity in 12 sedentary older people with IGT.

Subjects and Methods

Participants

This protocol was approved by the University of Michigan Institutional Review Board and performed in accordance with the Declaration of Helsinki. Healthy, community-dwelling older men and women were recruited by advertisement to participate in a study of exercise, glucose metabolism, and aging. After the nature of the study was explained in detail, informed consent was obtained from all participants. Health status, activity/diet history, and glucose tolerance were assessed by screening history, physical examination, blood chemistries, complete blood count, electrocardiogram, and 75-g oral glucose tolerance test (OGTT). The degree of physical activity was evaluated by the self-report Community Healthy Activities Model Program for Seniors Physical Activity Questionnaire for Older Adults (20).

Study design

Sedentary older people (age ≥ 60 yr) with IGT (fasting glucose level < 126 mg/dl and 2-h OGTT glucose level 140–199 mg/dl) were eligible for enrollment. Participants had no evidence of coronary artery disease; tobacco use; participation in exercise/diet programs; or use of systemic steroids, diabetes treatments, β-blockers, or thiazides. Eligible participants underwent a screening maximal stress/O₂ uptake (VO₂) test. Assessment of insulin sensitivity, insulin secretion, and β-cell function by frequently sampled iv glucose tolerance test (FSIGT) was performed: at baseline before exercise and on completion of the 7-d supervised aerobic exercise period. The baseline FSIGT was performed a minimum of 1 wk after the screening stress test and after exercise FSIGT 15–20 h after the seventh exercise session. This time period of assessment was selected because a study goal was to examine and isolate acute, exercise-induced effects on insulin sensitivity and β-cell function without potential confounders such as weight loss or changes in body composition that may occur with a longer-term exercise intervention.

All participants met with the University of Michigan General Clinical Research Center dietitians before randomization and were instructed on a weight-maintenance diet for the duration of the study. Each participant also underwent total body dual-energy x-ray absorptiometry scanning for assessment of body composition upon enrollment.

Study protocols

Screening maximal treadmill/VO₂ test

A modified Balke protocol was used for safety purposes before enrollment in the exercise program as well as to determine maximum heart rate for training purposes and peak VO₂.

Exercise program

Participants performed supervised aerobic exercise for 7 consecutive days including: 5-min warm-up; three 20-min periods of alternating treadmill walking, stationary cycling, or NuStep recumbent cross-trainer (NuStep, Inc., Ann Arbor, MI) at 60–70% heart rate reserve (maximum heart rate-resting heart rate); 5-min rest in between periods; and 5-min cool-down. Participants wore heart rate monitors to ensure that they were reaching target heart rate levels.

Assessments of glucose tolerance, insulin sensitivity, and β-cell function were performed in the morning after 12-h overnight fasts. OGTT and insulin-assisted FSIGT were performed on separate study days.

OGTT

After a baseline sample for fasting plasma glucose and insulin levels was obtained, participants ingested 75 g glucose. Blood samples were obtained for glucose and insulin every 30 min for 120 min.

FSIGT

The insulin-assisted FSIGT was performed as described by Bergman (21) with the addition of insulin to enhance precision of the estimates of insulin action (22). Participants consumed a diet containing a minimum of 150 g/d of carbohydrate for 3 d before studies. The hand/wrist with the blood sampling iv catheter was placed in a hot hand box to obtain arterialized blood samples. After three baseline samples for fasting glucose and insulin levels were drawn, 50% dextrose (0.3 g/kg) was injected iv over 30 sec followed by injection of insulin (0.02 U/kg) over 30 sec at time 20 min. Twenty-nine blood samples were collected according to a protocol schedule over 180 min. Fasting baseline samples were also drawn for lipid profile, free fatty acid (FFA), and adipocytokines.

The sensitivity to insulin index was calculated from a least-squares fitting of the temporal pattern of glucose and insulin throughout the FSIGT using the minimal model of glucose kinetics (21). Glucose disappearance constant (Kg), a measure of iv glucose tolerance, was calculated as the least square slope of the natural log of absolute glucose concentration between 10 and 19 min after the glucose bolus. Glucose effectiveness (Sg) is the capacity of glucose to mediate its own disposal. The acute insulin response to iv glucose (AIRg) was calculated as the mean rise in plasma insulin above baseline between 2 and 10 min after iv glucose administration. The relationship between two independent measurements of insulin secretion and insulin sensitivity has been found to be hyperbolic, allowing calculation of the product of AIRg × insulin sensitivity, or the disposition index (DI) (16, 23, 24). DI provides an indirect assessment of whether insulin secretion is appropriate for the level of insulin resistance (β-cell compensation for insulin resistance or β-cell function).

Assays

Serum was stored at −80 C until analysis. Plasma glucose levels were measured using a hexokinase method with an interassay coefficient of variation of 3.1% (Roche Diagnostics Corp., Indianapolis, IN). Plasma insulin was quantified using a double-antibody human-specific RIA with an interassay coefficient of variation (CV) of 3.4% and an intraassay variability of 2.5% (Linco Research, Inc., St. Charles, MO). Hemoglobin A1c was measured by HPLC with a normal range of 3.8–6.4%. Total cholesterol, triglycerides, high-density lipoprotein, and low-density lipoprotein were measured using standard reagents (Roche Diagnostics). FFA concentrations were measured by an original, enzymatic, colorimetric assay with an interassay CV of 4.7% and an intraassay variability of 3.6% (Wako Chemicals USA, Inc., Richmond, VA). High-sensitivity C-reactive protein (hsCRP) was measured using a latex immunoturbidimetric assay with an intraassay variation at 4.2 mg/liter of 1.29% (Equal Diagnostics, Exton, PA); plasminogen activator inhibitor, type 1, activity (PAI-1) by the chromolize PAI-1 assay with an intraassay precision of 2.6% (Trinity Biotech, Bray, Ireland); adiponectin by RIA with a sensitivity of 1.0 ng/ml and an interassay CV of 12% (Linco Research); and leptin by RIA with a sensitivity of 0.5 ng/ml and an interassay CV of 2.0% (Linco Research). Plasma catecholamines (epinephrine and norepinephrine) were measured by single-isotope enzymatic assay (25).

Statistical analysis

Data are presented as means ± se, with the exception of participant characteristics, which are presented as means ± sd. FSIGT parameters including insulin sensitivity, AIRg, and DI were log transformed to approximate a normal distribution. Differences between postexercise and baseline were assessed by Student’s paired t test. P < 0.05 was considered statistically significant.

Results

Screening participant characteristics

A total of 12 healthy, older men and women (mean age 68 ± 5 yr) with IGT were enrolled. Eleven of 12 participants also had impaired fasting glucose. Screening total caloric expenditure in leisure time physical activity per week as estimated by the Community Healthy Activities Model Program for Seniors Physical Activity Questionnaire for Older Adults was 1920 ± 270. Screening caloric expenditure per week in moderate intensity exercise-related activities (MET values ≥ 3) was 759 ± 144. Race of the study population was primarily white with one Hispanic person and one Asian American person. The screening characteristics are displayed in Table 1. Body mass index (BMI), elevated waist circumference, and increased body fat percentage per dual-energy x-ray absorptiometry was in the obese range representative of the community. One participant had a BMI of 24 kg/m², and the remaining 11 participants had a BMI ranging between 27 and 37 kg/m². Fasting and 2-h OGTT glucose levels were in the impaired fasting glucose and IGT range.

TABLE 1.

Screening characteristics of study participants

	Older participants
n	12
Age (yr)	68 ± 5
Gender (female/male)	7/5
Weight (kg)	92 ± 16
BMI (kg/m2)	32 ± 4
Waist circumference (cm)	108 ± 8
Body fat (%)	41.4 ± 7.6
Peak VO2 (ml/kg·min)	21.2 ± 4.1
Fasting glucose (mg/dl)	106 ± 6
2-h OGTT glucose (mg/dl)	160 ± 18
Hemoglobin A1c (%)	5.6 ± 0.2

Mean ± sd. To convert glucose from milligrams per deciliter to millimoles per liter, multiply by 0.05551.

Ten of 12 participants met clinical criteria (≥ 3 Adult Treatment Panel III and/or World Health Organization criteria) for the metabolic syndrome. The remaining two enrollees met two criteria for the metabolic syndrome. Urinary albumin was not measured. Three participants were taking a single antihypertensive medication including an angiotensin receptor blocker, calcium channel blocker, or α-1 blocker. Four participants were taking statin agents for hyperlipidemia.

Exercise program

All 12 participants completed the study and the 7 d of consecutive exercise. There were no dropouts or injuries. All participants exercised for a total 60 min daily with an average intensity of 65.6 ± 0.6% of heart rate reserve with no change in body weight, as displayed in Table 2.

TABLE 2.

Effect of exercise on FSIGT measures

	Baseline	After exercise	P value
n	12	12
Weight (kg)	91.8 ± 4.6	91.7 ± 4.8	0.9
FSIGT
Fasting glucose (mg/dl)	101 ± 3	103 ± 3	0.2
Fasting insulin (μU/ml)	16 ± 2	15 ± 1	0.2
Kg (min−1)	1.1 ± 0.2	1.2 ± 0.2	0.7
Sg (10−2/min −1)	1.4 ± 0.2	1.2 ± 0.2	0.1
SI (10−4·min−1/μU·ml)	1.7 ± 0.4	2.6 ± 0.3	0.002
AIRg (μU/ml)	40 ± 5	35 ± 5	0.01
DI (AIRg × SI)	70 ± 18	90 ± 17	0.02

Mean ± se. To convert fasting insulin and AIRg from microunits per milliliter to picomoles, multiply by 6. To convert S_I from minutes⁻¹ per microunits per milliliter to minutes⁻¹ per picomole, divide by 6. S_I, Insulin sensitivity.

Glucose tolerance and fasting insulin

Fasting glucose, iv glucose tolerance (Kg), and fasting insulin levels obtained during FSIGT studies before and after 7 d of exercise are summarized in Table 2. There was no significant change in fasting glucose, fasting insulin, Kg, or Sg after exercise, compared with baseline.

Insulin sensitivity

As shown in Table 2 and Fig. 1A, insulin sensitivity significantly increased with exercise, compared with baseline (P = 0.002).

Fig. 1.

Profiles of effects of 7 d of aerobic exercise training on insulin sensitivity (S_I), AIRg, and DI (AIRg × S_I or β-cell function in relation to insulin resistance) from FSIGT in 12 older people with IGT (A–C). Median values for baseline and postexercise parameters are as follows: S_I, 1.5, 1.8; AIRg, 42, 33; DI, 63, 85. Exercise (vs. baseline) significantly increased S_I (P = 0.002), decreased AIRg (P = 0.01), and increased DI (P = 0.02).

Insulin secretion

AIRg and DI (AIRg × insulin sensitivity or β-cell compensation for insulin resistance) from the FSIGT studies at baseline and after exercise are displayed in Table 2 and Fig. 1, B and C. There was a significant exercise effect for a decrease in AIRg (P = 0.01) and an increase in DI (P = 0.02).

Fasting lipid, adipocytokine, and catecholamine levels

Fasting lipid, adipocytokine, and catecholamine levels obtained before the FSIGT studies at baseline and after exercise are displayed in Table 3. The baseline lipid profiles were not elevated in these older people with IGT. There was no significant change in fasting lipid or FFA levels after exercise, except for a trend for a reduction in triglyceride levels (P = 0.07). There was no significant change in adipocytokine or catecholamine levels, except for a trend for a decline in leptin (P = 0.09).

TABLE 3.

Effect of exercise on fasting lipid, adipocytokine, and catecholamine levels

	Baseline	After exercise	P value
n	12	12
Cholesterol (mg/dl)	178 ± 12	175 ± 12	0.70
HDL (mg/dl)	51 ± 5	52 ± 5	0.21
Triglycerides (mg/dl)	107 ± 15	88 ± 10	0.07
LDL (mg/dl)	117 ± 10	115 ± 9	0.70
FFA (mEq/liter)	0.77 ± 0.07	0.90 ± 0.10	0.12
hsCRP (mg/liter)	2.2 ± 0.6	2.8 ± 1.1	0.55
PAI-1 (IU/ml)	23 ± 6	24 ± 6	0.86
Adiponectin (μg/ml)	12.4 ± 1.3	12.6 ± 1.3	0.49
Leptin (ng/ml)	21.5 ± 4.5	20.3 ± 4.5	0.09
Norepinephrine (pg/ml)	265 ± 30	267 ± 22	0.95
Epinephrine (pg/ml)	73 ± 11	84 ± 8	0.22

Mean ± se. To convert cholesterol, HDL, and LDL from milligrams per deciliter to millimoles, multiply by 0.0259. To convert triglycerides from milligrams per deciliter to millimoles, multiply by 0.0113. To convert FFA from milligrams per deciliter to grams per liter, multiply by 0.01. To convert hsCRP from milligrams per liter to nanomoles per liter, multiply by 8.45. To convert norepinephrine from pg/ml to nanomoles per liter, multiply by 0.0059. To convert epinephrine from picograms per milliliter to picomoles per liter, multiply by 5.458. HDL, High-density lipoprotein; LDL, low-density lipoprotein.

Discussion

The study goal was to examine acute effects of exercise on β-cell function in older people with IGT without potential confounding effects of weight loss or changes in body composition. Single bouts of acute exercise have been shown to significantly increase insulin sensitivity in healthy people (26, 27), insulin-resistant people with obesity or a family of type 2 diabetes as well as people with type 2 diabetes (28–30). However, previous studies have not assessed acute exercise effects on β-cell function in older people with IGT.

The new finding in our study was the significant improvement in β-cell function in older people with IGT. As expected, insulin sensitivity also improved after exercise, and we found baseline insulin resistance and impairments in insulin secretion, compared with previously studied young people (6, 31). Seven days of exercise training increased insulin sensitivity by 53%. Although AIRg significantly decreased as expected with improved insulin action, the DI (AIRg × insulin sensitivity) during FSIGT, an indirect measure of β-cell compensation of insulin resistance, increased by 28%. Although DI provides an indirect, calculated measure of β-cell function in relation to insulin resistance, it has been widely used in numerous studies to assess β-cell function in humans, given the need to examine insulin secretion in relation to insulin resistance (32, 33). There was no change in body weight; fasting glucose; iv glucose tolerance; or circulating levels of lipid, catecholamine, leptin, or adiponectin levels.

Three-month exercise training in combination with diet has been shown to improve insulin sensitivity in older people with IGT (34–36). Other work has also suggested, as in our study, that insulin resistance and β-cell function are closely related in people at risk for diabetes (37). Insulin secretion as assessed by hyperglycemic clamp (without assessment of insulin sensitivity) has been shown to increase after 14 d of inactivity, compared with 16 h after exercise in healthy, well-trained individuals (38). However, these effects were not evaluated in relation to the inferred increase in insulin resistance with cessation of exercise.

Kahn et al. (39) showed that 6 months of exercise training (with significant weight loss) in healthy older men improved insulin sensitivity to similar levels of younger men. However, the trained older men had decreased maximal insulin secretory capacity with glucose/arginine stimulation, compared with the young men, highlighting age-related impairments in insulin secretion, even with normal aging. Dela et al. (40) found that 3 months of exercise training in middle-aged people with type 2 diabetes increased peak VO₂ and may have improved β-cell response to hyperglycemia and arginine stimulation in those who were classified as being moderate C-peptide secretors vs. low secretors. This suggests that people with mild to moderate impairments in β-cell function, such as people with IGT (and not overt type 2 diabetes), may respond to therapies to improve β-cell function, although this requires confirmation with prospective studies.

The mechanisms contributing to the age-related decline in β-cell function and how exercise may improve β-cell function cannot be determined from the current study. Aging as well as development of glucose intolerance may be associated with loss of β-cell mass and/or impaired insulin secretion. Adipose tissue may play a role via FFAs. Although FFAs have acute effects to stimulate insulin secretion and maintain basal insulin secretion in healthy people (41, 42), sustained increases in plasma FFAs impair insulin secretion in people at risk for diabetes (43, 44). A single session of acute exercise has been shown to prevent fatty acid-induced insulin resistance in young, lean people (45). Adipocytokines have been strongly associated with insulin resistance, although they have not been established to play a role in β-cell dysfunction (46). Although adiponectin receptors were found to be expressed in human islets, adiponectin was not associated with insulin secretion in humans (47). Exercise training has not been shown to increase adiponectin levels (48, 49). Chronic exposure of human islets to leptin has been shown to lead to impaired insulin secretion and β-cell apoptosis, although this has not been a consistent finding (50). The current data from our 7-d exercise study in older people shows no change in fasting FFAs, lipid levels, adiponectin, or catecholamines, although there was a trend for a decline in triglyceride and leptin levels.

Limitations of this study include the short duration of the exercise training and the lack of a nonexercising control group or young comparison group. However, the short duration of exercise in this study allowed us to isolate acute exercise effects in older people with IGT with maintenance of energy balance without changes in body weight or body composition. The results may be generalized to both men and women, but the study group included primarily a white population, and the findings may not apply to people of other racial groups.

In summary, short-term exercise significantly improved insulin resistance and β-cell function in older people with IGT. These effects of short-term exercise cannot be explained by changes in body mass or of circulating lipids, adipocytokines, or catecholamines. Longer-term exercise training studies are required and are currently in progress to evaluate further exercise training effects on β-cell function in age-related glucose intolerance.

Acknowledgments

We thank the study participants for their cooperation and commitment. We thank Dr. Neil Alexander and Mr. Eric Pear for their expertise in exercise testing and training and use of the University of Michigan Geriatrics Center Mobility Research Center (National Institutes of Health Grant AG024824) and Ms. Marla Smith for expert technical assistance. We are grateful to Dr. Jeffrey Halter for constructive criticism regarding the manuscript.

This work was supported by a Department of Veterans Affairs Clinical Science Research and Development Career Development Award, the Veterans Education and Research Association of Michigan, the University of Michigan Claude D. Pepper Older Americans Independence Center (National Institutes of Health Grant AG024824), the Michigan Diabetes Research and Training Center (National Institutes of Health Grant DK20572), the John A. Hartford Foundation, and the University of Michigan General Clinical Research Center (National Institutes of Health Grant RR0042).

Disclosure Statement: The authors have nothing to disclose.

Abbreviations

 
• AIRg,

  Acute insulin response to glucose;

 
• BMI,

  body mass index;

 
• CV,

  coefficient of variation;

 
• DI,

  disposition index;

 
• DPP,

  Diabetes Prevention Program;

 
• FFA,

  free fatty acid;

 
• FSIGT,

  frequently sampled iv glucose tolerance test;

 
• hsCRP,

  high-sensitivity C-reactive protein;

 
• IGT,

  impaired glucose tolerance;

 
• Kg,

  glucose disappearance constant;

 
• OGTT,

  oral glucose tolerance test;

 
• PAI-1,

  plasminogen activator inhibitor type 1;

 
• Sg,

  glucose effectiveness;

 
• VO₂,

  O₂ uptake.