The aim of the present study was to determine the effects of a 4-week exercise training intervention on blood glucose, insulin sensitivity, BMI (body mass index) and cardiorespiratory fitness in patients with Type 2 diabetes, and to identify and establish criteria for patients who are more likely to improve their blood glucose from short-term exercise training. A randomized, controlled trial of exercise training, comprising two supervised and one non-supervised sessions of individualized cardiorespiratory and resistance exercise per week, was performed in 132 healthy patients with Type 2 diabetes (exercise training group, n=68), with the aim of accumulating a minimum of 150 min of moderate-intensity exercise for 4 weeks. BMI, waist circumference, blood pressure, blood lipid profile, blood glucose, insulin, insulin sensitivity [calculated by HOMAIR (homoeostasis model assessment of insulin resistance) and QUICKI (quantitative insulin check index)], β-cell function (calculated by HOMAβ-Cell), HbA1c (glycated haemoglobin) and V̇O2max (maximal oxygen consumption) were measured at baseline and at 4 weeks. The exercise training group had significant improvements in V̇O2max, BMI and triacylglycerols (triglycerides). There were no significant changes in blood glucose, HOMAIR, QUICKI or HOMAβ-Cell. Decreases in blood glucose were significantly predicted by baseline blood glucose and HbA1c, with these variables accounting for 15.9% of the change in blood glucose (P<0.001). ROC (receiver operator characteristic) curve analysis revealed that patients with a blood glucose >8.85 mmol/l (sensitivity=73%, specificity=78%) and HbA1c >7.15% (sensitivity=79%, specificity=60%) were more likely to achieve a clinically significant decrease in blood glucose. In conclusion, in apparently healthy patients with Type 2 diabetes, a 4-week exercise intervention improved cardiorespiratory fitness, BMI and triacylglycerols. Elevated blood glucose and HbA1c predicted improvements in blood glucose.

INTRODUCTION

Meta-analyses have found that exercise training in patients with Type 2 diabetes has a small-to-moderate beneficial effect on glycaemic control [1,2]. The original investigations on this topic have generally used interventions longer than 8 weeks, with the meta-analysis stating that the average exercise training time is approx. 60 h [1]. Shorter-term interventions offer potential for close supervision, and short-term success may increase the likelihood of ongoing lifestyle changes [3]. Furthermore, they offer a greater opportunity to individualize the exercise prescription, which is beneficial to exercise motivation and compliance [4].

The current position statements by exercise and diabetes associations do not discuss whether certain patients may benefit more from exercise training based on their baseline health status [5,6]. If particular patients with Type 2 diabetes do respond better to exercise training, developing methods to identify these individuals has clinical appeal and economic merit.

Therefore the aims of the present study were to determine the effects of a 4-week exercise training intervention on blood glucose, insulin sensitivity, BMI (body mass index) and cardiorespiratory fitness in patients with Type 2 diabetes, and to identify and establish criteria for patients who are more likely to improve their blood glucose from short-term exercise. Given the short duration of the present study, changes in HbA1c (glycated haemoglobin) would not be representative and were therefore not chosen as an outcome measure. We hypothesized that the 4-week programme would decrease blood glucose and BMI, increase insulin sensitivity and cardiorespiratory fitness and that improvements in glycaemic control would be negatively associated with baseline glycaemic control.

MATERIALS AND METHODS

Patient selection

Patients with Type 2 diabetes from hospital clinics and the community were eligible for inclusion if they were between 18 and 75 years of age, excluding those with a serious co-morbidity (life expectancy <6 months), pregnancy or known cardiovascular disease. Recruited patients (n=248) underwent screening for occult coronary artery disease using exercise echocardiography, which excluded 25 patients. With the goal of achieving blood glucose levels <7 mmol/l in patients with a baseline blood glucose of 8.4±2.6 mmol/l, 60 patients per group provided 83% power.

After baseline testing, patients were randomly allocated to either usual care (n=112) or exercise training groups (n=111) by an independent statistician using random number generation. A total of 13 groups of 15–20 patients were randomized every 3 months between 2003 and 2006, and randomization was stratified by group intake.

After randomization to exercise training or usual care, five declined the intervention and, over the 4-week period, five patients withdrew from the study and 13 did not attend for follow-up. Complete data were available for the present study from 132 patients of whom 68 were in the exercise training group.

The trial was approved by the Research Ethics Committees of the Princess Alexandra Hospital and The University of Queensland, and informed consent was obtained from all participants.

Usual care

All participants received standard risk factor intervention for Type 2 diabetes, including support to maintain a target BP (blood pressure) of <130/80 mmHg, smoking cessation and attainment of lipid targets [LDL (low-density lipoprotein)-cholesterol <2.6 mmol/l; triacyglycerols (triglycerides) <1.7 mmol/l; and HDL (high-density lipoprotein)-cholesterol >1.0 mmol/l] [7]. Usual care was co-ordinated through the Diabetes Clinic of Princess Alexandra Hospital, a teaching hospital in Brisbane, Australia. Control patients received usual care only.

Exercise training

In addition to usual care, patients randomized to the exercise training group received individualized exercise training. An accredited exercise physiologist supervised the exercise intervention, which consisted of a 4-week gym-based training programme. The aims of the programme were to (i) initiate training, (ii) achieve a minimum of 150 min/week of at least moderate-intensity exercise, and (iii) educate patients in preparation for the second stage of the intervention, ensuring they had the knowledge and skills to exercise independently with telephone counselling supervised by an accredited exercise physiologist. During each of the 4 weeks, patients attended two exercise sessions of 1 h in duration in a hospital-based gymnasium, consisting of individualized moderate-intensity cardiorespiratory and resistance training, and were asked to perform an additional 30 min of exercise at home. The programmes were designed according to fitness strengths and weaknesses, co-morbidities and complications, likes and dislikes and barriers to exercise. Cardiorespiratory training was usually chosen based on patient preference (cycling, walking, jogging, stepping or rowing). Moderate intensity was defined as a rating of perceived exertion of 12–13 (Borg 20 point scale). Resistance exercises were initially performed using gym equipment (both machine and free weight) and progressed to exercises modified to equipment that patients could take home for the second stage of the intervention (Thera-Bands® and Fitballs). Participants were encouraged to reach failure (not able to complete another repetition) from 12–15 repetitions for each exercise for two to three sets. The home-based session generally consisted of walking, jogging, swimming and home-based resistance exercises at moderate intensity.

Clinical and laboratory measurements

Height, body mass (Seca; Vogel & Aalke), waist circumference (KDS), resting supine BP (Baumanometer; W.A. Baum), fasting blood lipid profile, glucose, insulin and HbA1c were measured in all patients prior to and after the 4-week intervention period. Blood analyses were conducted using standard procedures by the Queensland Health Pathology Service (Brisbane, Australia), and all measurements were taken after an overnight fast and at least 24 h after the last exercise session. Insulin sensitivity [HOMAIR (homoeostasis model assessment of insulin resistance) and QUICKI (quantitative insulin check index)] and β-cell function [HOMAβ-Cell (homoeostasis model assessment of β-cell function)] were determined using formulae derived previously [8,9].

Cardiorespiratory fitness

Cardiorespiratory fitness was assessed by indirect calorimetry (Vmax29c; SensorMedics), measuring V̇O2max (maximal oxygen consumption) during a graded exercise test to exhaustion. BP and cardiac status (using a 12-lead ECG) were monitored during the exercise test (CASE; GE Medical Systems).

Statistical analysis

All data was tested for normality with the Kolmogorov–Smirnov test and the Normal Q-Q plot, where appropriate. The following data required log transformation: BMI, glucose, HbA1c, HOMAIR, HOMAβ-Cell, QUICKI, triacyglycerols, HDL, SBP (systolic BP), DBP (diastolic BP) and V̇O2max. General linear modelling with repeated measures ANOVA was used to assess group and time effects. A group–time interaction was used to establish differences between groups. Paired Student's t tests were used to assess within-group changes. The mean absolute change was determined from the individual change. Pearson correlations were used to assess associations, and multiple regression analysis was used to determine significant independent predictors, with P<0.10 in bivariate testing used as a requirement to enter the model. Significance was assumed if P<0.05. ROC (receiver operator characteristic) curves were used to determine cut-off values associated with improvements in blood glucose. A clinically significant improvement in blood glucose was selected as ≥1 mmol/l [10]. All baseline variables that correlated significantly with a change in blood glucose were plotted to determine the specificity and sensitivity of the measure in predicting a clinically significant change in blood glucose. Statistical analyses were conducted using standard statistical software (SPSS version 15.0 for Windows).

RESULTS

Patient characteristics

Table 1 provides a comparison of baseline demographic, clinical and laboratory results in the exercise training and control groups, and their change with exercise training. The exercise training group had a significantly (P<0.05) higher SBP and lower V̇O2max compared with the control group at baseline.

Table 1
Comparison of demographic, clinical, physical and medication data at baseline and after 4 weeks between patients in the exercise training and control groups

Values are means±S.D. P value represents the group×time effect from the general linear model with repeated measures ANOVA. *Significantly different to baseline in the control group (P<0.05); †Significant within group difference (P<0.05). ACE, angiotensin-converting enzyme

Exercise trainingControl
BaselineChangeBaselineChangeGroup×time P value
Age (years) 57±11  65±8   
Duration of diabetes (years) 5±5  4±5   
Gender (n) (male/female) 36/32  34/30   
BMI (kg/m232.3±5.4 −0.3±1.3 30.9±5.8 0.2±1.1† 0.033 
Waist circumference (cm) 108.5±11.9 −1.0±2.3† 100.7±10.3 −0.6±3.1 0.551 
Blood glucose (mmol/l) 8.08±2.61 −0.02±1.95 8.38±2.62 −0.24±1.80 0.504 
Insulin (m-units/l) 15.3±8.0 0.4±5.5 13.4±9.6 0.3±5.8 0.921 
HOMAIR 5.8±4.0 −0.1±3.0 4.8±3.4 0.2±2.2 0.486 
HOMAβ-Cell 86.3±67.8 1.4±30.6 76.2±80.0 3.8±27.1 0.518 
QUICKI 0.308±0.027 −0.001±0.019 0.319±0.037 −0.004±0.022 0.388 
HbA1c (%) 7.4±1.3 −0.1±0.8 7.4±1.4 −0.1±0.8 0.701 
Total cholesterol (mmol/l) 4.9±1.0 −0.14±0.62 4.8±0.9 0.01±0.48 0.143 
HDL-cholesterol (mmol/l) 1.3±0.3 0.01±0.17 1.4±0.5 −0.02±0.17 0.342 
LDL-cholesterol (mmol/l) 2.7±0.8 −0.1±0.6 2.6±0.7 −0.03±0.5 0.441 
Triacylglycerols (mmol/l) 2.1±2.8 −0.2±0.8 1.7±1.1 0.1±0.5 0.033 
SBP (mmHg) 137.1±18.1* −6.9±15.1† 129.2±15.3 −5.4±12.1† 0.558 
DBP (mmHg) 79.8±8.5 3.5±9.2† 76.9±9 3.2±10† 0.888 
V̇O2max (ml·kg−1 of body weight·min−120.7±5.7 * 1.4±4.4† 24.4±7 −0.7±4.2 0.011 
Medication (n     
 Statins 30 26 −2  
 β-Blocker  
 ACE inhibitor 21 15  
 Angiotensin receptor antagonists 14 10  
 Calcium channel blocker  
 Metformin 41 −1 36  
 Sulfonylurea 13 19  
 Thiazolidinedione  
 Insulin  
Exercise trainingControl
BaselineChangeBaselineChangeGroup×time P value
Age (years) 57±11  65±8   
Duration of diabetes (years) 5±5  4±5   
Gender (n) (male/female) 36/32  34/30   
BMI (kg/m232.3±5.4 −0.3±1.3 30.9±5.8 0.2±1.1† 0.033 
Waist circumference (cm) 108.5±11.9 −1.0±2.3† 100.7±10.3 −0.6±3.1 0.551 
Blood glucose (mmol/l) 8.08±2.61 −0.02±1.95 8.38±2.62 −0.24±1.80 0.504 
Insulin (m-units/l) 15.3±8.0 0.4±5.5 13.4±9.6 0.3±5.8 0.921 
HOMAIR 5.8±4.0 −0.1±3.0 4.8±3.4 0.2±2.2 0.486 
HOMAβ-Cell 86.3±67.8 1.4±30.6 76.2±80.0 3.8±27.1 0.518 
QUICKI 0.308±0.027 −0.001±0.019 0.319±0.037 −0.004±0.022 0.388 
HbA1c (%) 7.4±1.3 −0.1±0.8 7.4±1.4 −0.1±0.8 0.701 
Total cholesterol (mmol/l) 4.9±1.0 −0.14±0.62 4.8±0.9 0.01±0.48 0.143 
HDL-cholesterol (mmol/l) 1.3±0.3 0.01±0.17 1.4±0.5 −0.02±0.17 0.342 
LDL-cholesterol (mmol/l) 2.7±0.8 −0.1±0.6 2.6±0.7 −0.03±0.5 0.441 
Triacylglycerols (mmol/l) 2.1±2.8 −0.2±0.8 1.7±1.1 0.1±0.5 0.033 
SBP (mmHg) 137.1±18.1* −6.9±15.1† 129.2±15.3 −5.4±12.1† 0.558 
DBP (mmHg) 79.8±8.5 3.5±9.2† 76.9±9 3.2±10† 0.888 
V̇O2max (ml·kg−1 of body weight·min−120.7±5.7 * 1.4±4.4† 24.4±7 −0.7±4.2 0.011 
Medication (n     
 Statins 30 26 −2  
 β-Blocker  
 ACE inhibitor 21 15  
 Angiotensin receptor antagonists 14 10  
 Calcium channel blocker  
 Metformin 41 −1 36  
 Sulfonylurea 13 19  
 Thiazolidinedione  
 Insulin  

Efficacy of the exercise training

Exercise training led to significant improvements in V̇O2max, BMI and triacylglycerols compared with controls (Table 1). Waist circumference significantly (P<0.05) decreased within the exercise training group; however, exercise training had no effect on blood glucose, insulin sensitivity (QUICKI and HOMAIR) or β-cell function (HOMAβ-Cell).

Correlates of changes in blood glucose and insulin sensitivity

As there were no significant (P>0.05) group differences for blood glucose and markers of insulin sensitivity, results from both groups were pooled to determine patients who were more likely to improve their blood glucose and insulin sensitivity. Analysis of the results from within the intervention group only (results not shown) showed similar results to the pooled analysis. Table 2 shows the significant bivariate associations with changes in blood glucose, HOMAIR and QUICKI. The change in blood glucose was significantly (P<0.05) negatively correlated with baseline blood glucose and baseline HbA1c, and significantly (P<0.05) positively correlated with a change in total cholesterol and triacylglycerols. All four variables were entered into a multiple regression model which revealed that baseline blood glucose (β=−0.677, P<0.001), baseline HbA1c (β=−0.288, P=0.019) and changes in triacyglycerols (β=0.191, P=0.021) were significant independent predictors of the change in blood glucose, with the combination of these variables accounting for 30% of the change in blood glucose (P<0.001). Furthermore, when including only baseline blood glucose and HbA1c in the model, it still accounted for 15.9% (P<0.001) of the change in glucose (β=−0.580, P<0.001, and β=−0.262, P=0.038 respectively) (Figure 1).

Table 2
Baseline and absolute change in variables associated with changes in blood glucose and markers of insulin sensitivity

Change in blood glucoseChange in HOMAIRChange in QUICKI
VariableBivariate (r)P valueBivariate (r)P valueBivariate (r)P value
Baseline waist circumference (cm)   −0.208 0.018 0.235 0.007 
Baseline blood glucose (mmol/l) −0.379 <0.001     
Baseline insulin (m-units/l)   −0.184 0.035 0.414 <0.001 
Baseline HbA1c (%) −0.183 0.036     
Baseline QUICKI     −0.454 <0.001 
Baseline HOMAIR   −0.383 <0.001   
Change in BMI (kg/m2  0.369 <0.001 −0.243 0.006 
Change in total cholesterol (mmol/l) 0.181 0.038 0.272 0.002   
Change in triacylglycerols (mmol/l) 0.182 0.037 0.172 0.049 −0.186 0.033 
Change in blood glucoseChange in HOMAIRChange in QUICKI
VariableBivariate (r)P valueBivariate (r)P valueBivariate (r)P value
Baseline waist circumference (cm)   −0.208 0.018 0.235 0.007 
Baseline blood glucose (mmol/l) −0.379 <0.001     
Baseline insulin (m-units/l)   −0.184 0.035 0.414 <0.001 
Baseline HbA1c (%) −0.183 0.036     
Baseline QUICKI     −0.454 <0.001 
Baseline HOMAIR   −0.383 <0.001   
Change in BMI (kg/m2  0.369 <0.001 −0.243 0.006 
Change in total cholesterol (mmol/l) 0.181 0.038 0.272 0.002   
Change in triacylglycerols (mmol/l) 0.182 0.037 0.172 0.049 −0.186 0.033 

Associations between the change in blood glucose and baseline blood glucose (a) and HbA1c (b)

Figure 1
Associations between the change in blood glucose and baseline blood glucose (a) and HbA1c (b)

The horizontal lines indicate a clinically significant improvement in blood glucose. The vertical lines indicate the proposed baseline cut-off values determined by ROC curve analyses.

Figure 1
Associations between the change in blood glucose and baseline blood glucose (a) and HbA1c (b)

The horizontal lines indicate a clinically significant improvement in blood glucose. The vertical lines indicate the proposed baseline cut-off values determined by ROC curve analyses.

The change in HOMAIR was significantly (P<0.05) negatively correlated with baseline HOMAIR, waist circumference and baseline insulin and significantly (P<0.05) positively correlated with the change in BMI, total cholesterol and triacylglycerols. These variables were entered into a multiple regression model which showed that baseline HOMAIR (β=−0.307, P<0.001), a change in BMI (β=0.343, P<0.001) and a change in total cholesterol (β=0.198, P=0.021) were significant independent predictors of a change in HOMAIR, with the combination of these variables accounting for 30% of the change in HOMAIR (P<0.001). Furthermore, when including only baseline HOMAIR, it still accounted for 15% (P<0.001) of the change in HOMAIR (β=−0.388, P<0.001).

The change in QUICKI was significantly (P<0.05) negatively correlated with baseline QUICKI, the change in BMI and triacylglycerols, and significantly (P<0.05) positively correlated with baseline waist circumference and baseline insulin. Multiple regression with these variables revealed that baseline QUICKI (β=−0.442, P<0.001), the change in BMI (β=0.227, P=0.007) and the change in triacylglycerols (β=0.168, P=0.007) were significant independent predictors of the change in QUICKI, with the combination of these three accounting for 39% of the change in QUICKI (P<0.001). Furthermore, when including only baseline QUICKI in the model, it still accounted for 23% (P<0.001) of the change in QUICKI (β=−0.476, P<0.001).

Determining thresholds of the benefit for exercise training

ROC curve analyses were conducted to identify patients who would have the most benefit on blood glucose from exercise training. In this analysis, a decrease equivalent to 1 mmol/l was used to define a clinically significant improvement in blood glucose [10]. Baseline blood glucose and HbA1c (the only significant baseline predictors of a change in blood glucose) were plotted to determine the specificity and sensitivity of optimal cut-offs in predicting a clinically significant improvement in blood glucose. The cut-off values determined for baseline blood glucose and HbA1c were 8.85 mmol/l (sensitivity=74%, specificity=78%) and 7.15% (sensitivity=79%, specificity=60%) respectively. Baseline blood glucose had a comparable AUC (area under the curve) of 0.80 (P<0.001) and sensitivity and a higher specificity compared with baseline HbA1c (AUC=0.75, P<0.001). Patients who had baseline blood glucose above the cut-off value had significantly decreased blood glucose compared with those patients below the cut-off (change in glucose, −0.82±0.38 mmol/l compared with 0.24±0.13 mmol/l respectively; P<0.002).

DISCUSSION

In the present study, an individualized 4-week exercise training intervention, using a combination of cardiorespiratory and resistance training exercises, in patients with Type 2 diabetes was effective in producing a small, but significant, increase in cardiorespiratory fitness and small, but significant, decreases in BMI and triacylglycerols, although the intervention did not significantly effect blood glucose or markers of insulin sensitivity. However, patients with a higher blood glucose, HbA1c and HOMAIR and a lower QUICKI score were more likely to have an improvement in their glycaemic control and markers of insulin sensitivity over 4 weeks. The degree of improvement in BMI and blood lipid profile was associated with improvements in blood glucose and insulin sensitivity. Patients with Type 2 diabetes and a baseline blood glucose >8.85 mmol/l and HbA1c >7.15% derived the greatest benefit.

Treatment effects

The exercise training in the present trial resulted in a significant increase in V̇O2max in the exercise training group compared with controls. This is consistent with findings from a meta-analysis of an increase of 2.1 ml·kg−1 of body weight·min−1 in patients with Type 2 diabetes undergoing aerobic exercise training [3].

Changes in BMI and body composition reported previously are mixed, with some studies reporting improvements [1113] or no improvements [1418] in body composition, usually measured by changes in BMI or body mass. Our present results were consistent with this pattern, with small significant decreases in BMI. The varying modes of exercise (resistance compared with cardiorespiratory) and assessment techniques in more homogeneous patients may be responsible for such findings. Although there was no group difference, there was a significant decrease of 1 cm in waist circumference within the intervention group from baseline to post-training. It is therefore likely that the decrease observed in BMI is related to decreases in fat mass.

Triacylglycerols were also significantly decreased in the intervention group compared with controls. Some similar decreases have been reported from exercise training studies in Type 2 diabetes patients utilizing three sessions each week for training periods varying from 6 weeks to 1 year [1923], although a greater range and even increases have been shown [24]. This variation is likely to be a result of dietary changes that were not controlled during this 4-week training period or in other exercise training studies.

Contributions to blood glucose and insulin sensitivity effects

Small, non-significant, changes were observed over the 4-week period in blood glucose, HOMAIR, QUICKI and HOMAβ-Cell respectively.

There are several potential reasons why exercise training had no significant effect on blood glucose or markers of insulin sensitivity. First, it is possible that the exercise prescription may not have been long enough, had sufficient volume of training or was intense enough to yield these improvements. Previous exercise intervention studies have shown improvements in HbA1c and markers of insulin sensitivity (insulin resistance and glucose disposal rate) from cardiorespiratory training [15,17,25], cardiorespiratory and resistance training [12,18,26] or resistance training [14,21,22,27], with all of these interventions utilizing three or more sessions a week for an average of 58 h of training and prescribing >150 min/week of moderate-intensity exercise [1]. In comparison, the present study utilized 8 h of supervised training and 4 h of unsupervised training. Higher volumes (min/week) and intensities of exercise have been shown to produce greater improvements in glycaemic control and insulin resistance [2729]; furthermore, the shorter duration (4 weeks) may not have been sufficient. Nonetheless, the intervention used in the present study was based on current recommendations, which raises concerns regarding the appropriateness of these, especially for short-term results.

A second explanation for the lack of effect on blood glucose and markers of insulin resistance may relate to patient selection. We recruited patients from a diabetes clinic and the general community, resulting in a heterogeneous population that we believe maybe more representative of the overall Type 2 diabetes patient population. Patients with poor glycaemic control and markers of insulin sensitivity (HOMAIR and QUICKI) were significantly more likely to have an improvement in their glycaemic control. Baseline blood glucose, HbA1c, HOMAIR and QUICKI were not only negatively correlated, but were all significant independent predictors of changes in glycaemic control and markers of insulin sensitivity. ROC curve analysis revealed that the cut-off values for improvements in blood glucose were 8.85 mmol/l for baseline blood glucose and 7.15% for baseline HbA1c (Figure 1). Current exercise training guidelines [5,6,30] do not specify whether all or selected patients with Type 2 diabetes would respond better to exercise training. Using the ROC curve analysis, 35% of the patients in the present study were above the proposed cut-off value for blood glucose and 50% for HbA1c. Indeed, a sub-group analysis showed that patients who had baseline blood glucose above the proposed cut-off value significantly improved their blood glucose more than those patients below the cut-off. These thresholds may be considered as selection criteria for patients with Type 2 diabetes for priority to receive short-term exercise training. This conclusion is not suggesting that only selected patients with Type 2 diabetes should be targeted, rather, given the highly resource-dependent nature of exercise interventions, priority could be directed towards these selected patients. These findings are concordant with previous studies showing that patients with poorer baseline glycaemic control and body composition were more likely to improve their glycaemic control to both exercise [18] and pharmaceutical [31] interventions. However, improvements in insulin sensitivity were previously only shown in non-obese patients with Type 2 diabetes [17]. However, even the non-obese patients recruited for that particular study had a 15% higher blood glucose and a 26% higher HbA1c compared with the intervention group in the present study.

Table 3 shows a comparison of the variation of patients with Type 2 diabetes used in exercise intervention studies. Seven of the 12 studies reported no effects of exercise training on blood glucose. In comparison with patients with Type 2 diabetes in other investigations, the patients enrolled in the present study had the lowest blood glucose, a comparatively low HbA1c and the largest variation in age and cardiorespiratory fitness. This is probably a result of subject recruitment with many of the previous studies recruiting patients who were overweight, non-obese or were within a specific age group. As explained in the Materials and methods section, recruitment for the present study comprised apparently healthy patients from the diabetes clinic and the community.

Table 3
Baseline characteristics of patients from exercise training studies and the effect on blood glucose compared with the present study

Values are means±S.D.

StudyEffect on blood glucosenPatient characteristicsAge (years)Blood glucose (mmol/l)Insulin (pmol/l)QUICKIHOMAIRHbA1c (%)Weight (kg)BMI (kg/m2)V̇O2max (ml·kg−1 of body weight·min−1)
Present study Negative 132  57±11 8.08±2.6 106.3±55.6 0.308±0.03 5.8±4 7.4±1.3 91.7±16.6 32.3±5.4 20.7±5.7 
Boudou et al. [25Negative 20 Men, non-smokers and no hypertension 46.8±7.7 8.85±1.6 134.7±49.3   8.0±1.7 58.3±14.1 28.1±4.5 22.7±3.5 
Burns et al. [34Negative 13 Subjects 15–30 years of age with early onset Type 2 diabetes 25.8±1.2 10.3±0.4    8.8±0.3 109±5 25.8±1.2  
Castaneda et al.[22Negative 62 Latino patients >55 years of age 66±2 8.8±0.5    8.7±0.3 79.3±3.2 30.9±1.1  
Dunstan et al. [27Negative 36 Overweight and sedentary 67.6±5.2 9.5±0.8 132.9±63  17.7±6.5 8.1±1.0 88.7±10.9 31.5±3.7  
Mourier at al. [15Negative 24 Good metabolic control 45±2 9.2±0.4 141.8±15.3   8.5±0.6 84.4±4.5 30.4±0.8  
Poirier et al. [17Negative 13 Men  11.0±0.7 110±21   9.9±0.6 87.0±4.9  31.2±1.2 
    Six obese 45±3 10.4±0.6 145±29   10.4±0.6 95.4±7.8  28.8±1.5 
    Seven non-obese 49±3 9.3±1.2 69±8   9.3±1.2 77.2±2.2  34.1±1.4 
Tessier et al. [18Negative 39 Subjects >65 years of age 69.3±4.2 8.8±2.7 147±87   7.5±1.2  78.6±19.3  
Baldi et al. [21Positive 18 Inactive men 46.5±2.1 12.0±0.9 268.1±35.4   8.9±1.2    
Dunstan et al. [20Positive 49 Non-smokers 52.3±8.3 9.6±3.3 78.2±47.2   8.3±1.5 85.6±10.6 29.1±10.6 22.3±3.1 
Katsuki et al. [32Positive 60 Japanese 54.4±11.3 8.7±2.7 43.2±28.8 0.343±0.036  9.6±2.3  23.7±3.2  
Kim et al. [13Positive 58 No insulin therapy 55±8.1 9.1±2.1   2.6±1.4 8.5±1.4 65.7±13.5 25.8±3.8  
Maiorana et al. [26Positive 16 Non-smokers without hypertension or hypercholesterolaemia 52±2 12.0±0.5    8.5±0.4 88.7±4.4 29.6±3.4 23.1±0.1 
StudyEffect on blood glucosenPatient characteristicsAge (years)Blood glucose (mmol/l)Insulin (pmol/l)QUICKIHOMAIRHbA1c (%)Weight (kg)BMI (kg/m2)V̇O2max (ml·kg−1 of body weight·min−1)
Present study Negative 132  57±11 8.08±2.6 106.3±55.6 0.308±0.03 5.8±4 7.4±1.3 91.7±16.6 32.3±5.4 20.7±5.7 
Boudou et al. [25Negative 20 Men, non-smokers and no hypertension 46.8±7.7 8.85±1.6 134.7±49.3   8.0±1.7 58.3±14.1 28.1±4.5 22.7±3.5 
Burns et al. [34Negative 13 Subjects 15–30 years of age with early onset Type 2 diabetes 25.8±1.2 10.3±0.4    8.8±0.3 109±5 25.8±1.2  
Castaneda et al.[22Negative 62 Latino patients >55 years of age 66±2 8.8±0.5    8.7±0.3 79.3±3.2 30.9±1.1  
Dunstan et al. [27Negative 36 Overweight and sedentary 67.6±5.2 9.5±0.8 132.9±63  17.7±6.5 8.1±1.0 88.7±10.9 31.5±3.7  
Mourier at al. [15Negative 24 Good metabolic control 45±2 9.2±0.4 141.8±15.3   8.5±0.6 84.4±4.5 30.4±0.8  
Poirier et al. [17Negative 13 Men  11.0±0.7 110±21   9.9±0.6 87.0±4.9  31.2±1.2 
    Six obese 45±3 10.4±0.6 145±29   10.4±0.6 95.4±7.8  28.8±1.5 
    Seven non-obese 49±3 9.3±1.2 69±8   9.3±1.2 77.2±2.2  34.1±1.4 
Tessier et al. [18Negative 39 Subjects >65 years of age 69.3±4.2 8.8±2.7 147±87   7.5±1.2  78.6±19.3  
Baldi et al. [21Positive 18 Inactive men 46.5±2.1 12.0±0.9 268.1±35.4   8.9±1.2    
Dunstan et al. [20Positive 49 Non-smokers 52.3±8.3 9.6±3.3 78.2±47.2   8.3±1.5 85.6±10.6 29.1±10.6 22.3±3.1 
Katsuki et al. [32Positive 60 Japanese 54.4±11.3 8.7±2.7 43.2±28.8 0.343±0.036  9.6±2.3  23.7±3.2  
Kim et al. [13Positive 58 No insulin therapy 55±8.1 9.1±2.1   2.6±1.4 8.5±1.4 65.7±13.5 25.8±3.8  
Maiorana et al. [26Positive 16 Non-smokers without hypertension or hypercholesterolaemia 52±2 12.0±0.5    8.5±0.4 88.7±4.4 29.6±3.4 23.1±0.1 

Very few studies have reported changes in HOMAIR or QUICKI as a result of exercise training. One study reported 10000 steps/day for 6 weeks had positive effects on QUICKI and blood glucose levels, although the lack of a control group limits this finding [32]. Another study showed weekly lifestyle education (including diet and exercise) for 16 weeks achieved an average of 8.6 h/per week of activity, which resulted in no change in HOMAIR but a significant decrease in blood glucose [13]. On the basis of previous studies, it appears that findings on the effect of exercise training on glycaemic control and insulin sensitivity greatly depend on the measures used to assess these changes. It is well recognized that HbA1c is a robust indictor of glycaemic control; however, as mentioned previously, due to the short duration of our present intervention, changes in HbA1c were not chosen as an outcome measure.

Changes in blood lipid profile (total cholesterol and triacylglycerols) and BMI were also associated with and predictive of changes in blood glucose. Multiple regression models containing baseline blood glucose, baseline HbA1c and changes in blood lipid profile and BMI were able to significantly predict 30% of the change in blood glucose, 30.3% of the change in HOMAIR and 28.7% of the change in QUICKI respectively. Previous studies have linked decreases in BMI or fat loss to improvements in glycaemic control [2,15], although some reports have suggested that these changes are independent of each other [14,33]. However, the results of the present study do support a link between decreases in BMI and improvements in glycaemic control, which is likely to be, at least in part, due to the use of cardiorespiratory training in the exercise programme.

Potential limitations of the present study were the use of the outcome measures, blood glucose, HOMAIR and QUICKI, to represent glycaemic control and insulin sensitivity, although these techniques have been used previously in studies investigating the effects of exercise in patients with Type 2 diabetes [13,27,32]. More sensitive measures, such as euglycaemic clamp tests, would give further insight into the patients' glycaemic control and insulin sensitivity; however, the time required for using this technique was prohibitive in such a large number of patients. An additional limitation may have been due to the non-control of medication. However, importantly, medication profiles did not change over the intervention period and, therefore, it is likely that medication did not influence the results.

Conclusions

Exercise training for 4 weeks had no effect on blood glucose or markers of insulin sensitivity in patients with Type 2 diabetes. The short-term training did result in a small, but significant, increase in cardiorespiratory fitness and decreases in BMI and triacylglycerols. The intervention followed current guidelines (≥150 min of moderate intensity exercise per week) which are less than those used in previous studies that have reported a positive result, particularly over short durations. In addition, we provide evidence that the metabolic health of the patients recruited may be responsible for the lack of an effect. Pooled data indicated that elevated blood glucose and HbA1c were associated with, and predictive of, improvements in blood glucose. If resources to deliver exercise training to patients with Type 2 diabetes are limited, then priority should be directed to patients with poor metabolic health. It is proposed that patients with blood glucose >8.85 mmol/l and a HbA1c >7.15% benefit more.

Abbreviations

     
  • AUC

    area under the curve

  •  
  • BMI

    body mass index

  •  
  • BP

    blood pressure

  •  
  • DBP

    diastolic BP

  •  
  • HbA1c

    glycated haemoglobin

  •  
  • HDL

    high-density lipoprotein

  •  
  • HOMAIR

    homoeostasis model assessment of insulin resistance

  •  
  • HOMAβ-Cell

    homoeostasis model assessment of β-cell function

  •  
  • LDL

    low-density lipoprotein

  •  
  • QUICKI

    quantitative insulin check index

  •  
  • ROC

    receiver operator characteristic

  •  
  • SBP

    systolic BP

  •  
  • V̇O2max

    maximal oxygen consumption

This work was supported in part by a Clinical Centre of Research Excellence Award from the National Health and Medical Research Council, Canberra, Australia.

References

References
1
Snowling
 
N. J.
Hopkins
 
W. G.
 
Effects of different modes of exercise training on glucose control and risk factors for complications in type 2 diabetic patients: a meta-analysis
Diabetes Care
2006
, vol. 
29
 (pg. 
2518
-
2527
)
2
Norris
 
S. L.
Zhang
 
X.
Avenell
 
A.
, et al 
Long-term effectiveness of lifestyle and behavioral weight loss interventions in adults with type 2 diabetes: a meta-analysis
Am. J. Med.
2004
, vol. 
117
 (pg. 
762
-
774
)
3
Boule
 
N. G.
Kenny
 
G. P.
Haddad
 
E.
Wells
 
G. A.
Sigal
 
R. J.
 
Meta-analysis of the effect of structured exercise training on cardiorespiratory fitness in Type 2 diabetes mellitus
Diabetologia
2003
, vol. 
46
 (pg. 
1071
-
1081
)
4
Lafortuna
 
C. L.
Resnik
 
M.
Galvani
 
C.
Sartorio
 
A.
 
Effects of non-specific vs individualized exercise training protocols on aerobic, anaerobic and strength performance in severely obese subjects during a short-term body mass reduction program
J. Endocrinol. Invest.
2003
, vol. 
26
 (pg. 
197
-
205
)
5
American Diabetes Association
Standards of medical care in diabetes–2007
Diabetes Care
2007
, vol. 
30
 
Suppl. 1
(pg. 
S4
-
S41
)
6
Albright
 
A.
Franz
 
M.
Hornsby
 
G.
, et al 
American College of Sports Medicine position stand. Exercise and type 2 diabetes
Med. Sci. Sports Exercise
2000
, vol. 
32
 (pg. 
1345
-
1360
)
7
American Diabetes Association
Standards of medical care for patients with diabetes mellitus
Diabetes Care
2002
, vol. 
25
 (pg. 
213
-
229
)
8
Matthews
 
D. R.
Hosker
 
J. P.
Rudenski
 
A. S.
Naylor
 
B. A.
Treacher
 
D. F.
Turner
 
R. C.
 
Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man
Diabetologia
1985
, vol. 
28
 (pg. 
412
-
419
)
9
Katz
 
A.
Nambi
 
S. S.
Mather
 
K.
, et al 
Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans
J. Clin. Endocrinol. Metab.
2000
, vol. 
85
 (pg. 
2402
-
2410
)
10
UK Prospective Diabetes Study (UKPDS) Group
Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33)
Lancet
1998
, vol. 
352
 (pg. 
837
-
853
)
11
Dunstan
 
D. W.
Puddey
 
I. B.
Beilin
 
L. J.
Burke
 
V.
Morton
 
A. R.
Stanton
 
K. G.
 
Effects of a short-term circuit weight training program on glycaemic control in NIDDM
Diabetes Res. Clin. Pract.
1998
, vol. 
40
 (pg. 
53
-
61
)
12
Cuff
 
D. J.
Meneilly
 
G. S.
Martin
 
A.
Ignaszewski
 
A.
Tildesley
 
H. D.
Frohlich
 
J. J.
 
Effective exercise modality to reduce insulin resistance in women with type 2 diabetes
Diabetes Care
2003
, vol. 
26
 (pg. 
2977
-
2982
)
13
Kim
 
S. H.
Lee
 
S. J.
Kang
 
E. S.
, et al 
Effects of lifestyle modification on metabolic parameters and carotid intima-media thickness in patients with type 2 diabetes mellitus
Metab. Clin. Exp.
2006
, vol. 
55
 (pg. 
1053
-
1059
)
14
Ishii
 
T.
Yamakita
 
T.
Sato
 
T.
Tanaka
 
S.
Fujii
 
S.
 
Resistance training improves insulin sensitivity in NIDDM subjects without altering maximal oxygen uptake
Diabetes Care
1998
, vol. 
21
 (pg. 
1353
-
1355
)
15
Mourier
 
A.
Gautier
 
J. F.
De Kerviler
 
E.
, et al 
Mobilization of visceral adipose tissue related to the improvement in insulin sensitivity in response to physical training in NIDDM. Effects of branched-chain amino acid supplements
Diabetes Care
1997
, vol. 
20
 (pg. 
385
-
391
)
16
Honkola
 
A.
Forsen
 
T.
Eriksson
 
J.
 
Resistance training improves the metabolic profile in individuals with type 2 diabetes
Acta Diabetol.
1997
, vol. 
34
 (pg. 
245
-
248
)
17
Poirier
 
P.
Tremblay
 
A.
Broderick
 
T.
Catellier
 
C.
Tancrede
 
G.
Nadeau
 
A.
 
Impact of moderate aerobic exercise training on insulin sensitivity in type 2 diabetic men treated with oral hypoglycemic agents: is insulin sensitivity enhanced only in nonobese subjects?
Med. Sci. Monit.
2002
, vol. 
8
 (pg. 
CR59
-
CR65
)
18
Tessier
 
D.
Menard
 
J.
Fulop
 
T.
, et al 
Effects of aerobic physical exercise in the elderly with type 2 diabetes mellitus
Arch. Gerontol. Geriatr.
2000
, vol. 
31
 (pg. 
121
-
132
)
19
Ligtenberg
 
P. C.
Hoekstra
 
J. B.
Bol
 
E.
Zonderland
 
M. L.
Erkelens
 
D. W.
 
Effects of physical training on metabolic control in elderly type 2 diabetes mellitus patients
Clin. Sci.
1997
, vol. 
93
 (pg. 
127
-
135
)
20
Dunstan
 
D. W.
Mori
 
T. A.
Puddey
 
I. B.
, et al 
The independent and combined effects of aerobic exercise and dietary fish intake on serum lipids and glycemic control in NIDDM. A randomized controlled study
Diabetes Care
1997
, vol. 
20
 (pg. 
913
-
921
)
21
Baldi
 
J. C.
Snowling
 
N.
 
Resistance training improves glycaemic control in obese type 2 diabetic men
Int. J. Sports Med.
2003
, vol. 
24
 (pg. 
419
-
423
)
22
Castaneda
 
C.
Layne
 
J. E.
Munoz-Orians
 
L.
, et al 
A randomized controlled trial of resistance exercise training to improve glycemic control in older adults with type 2 diabetes
Diabetes Care
2002
, vol. 
25
 (pg. 
2335
-
2341
)
23
Balducci
 
S.
Leonetti
 
F.
Di Mario
 
U.
Fallucca
 
F.
 
Is a long-term aerobic plus resistance training program feasible for and effective on metabolic profiles in type 2 diabetic patients?
Diabetes Care
2004
, vol. 
27
 (pg. 
841
-
842
)
24
Wing
 
R. R.
Epstein
 
L. H.
Paternostro-Bayles
 
M.
Kriska
 
A.
Nowalk
 
M. P.
Gooding
 
W.
 
Exercise in a behavioural weight control programme for obese patients with Type 2 (non-insulin-dependent) diabetes
Diabetologia
1988
, vol. 
31
 (pg. 
902
-
909
)
25
Boudou
 
P.
De Kerviler
 
E.
Vexiau
 
P.
Fiet
 
J.
Cathelineau
 
G.
Gautier
 
J.
 
Effects of a single bout of exercise and exercise training on steroid levels in middle-aged type 2 diabetic men: relationship to abdominal adipose tissue distribution and metabolic status
Diabetes Metab.
2000
, vol. 
26
 (pg. 
450
-
457
)
26
Maiorana
 
A.
O'Driscoll
 
G.
Goodman
 
C.
Taylor
 
R.
Green
 
D.
 
Combined aerobic and resistance exercise improves glycemic control and fitness in type 2 diabetes
Diabetes Res. Clin. Pract.
2002
, vol. 
56
 (pg. 
115
-
123
)
27
Dunstan
 
D. W.
Daly
 
R. M.
Owen
 
N.
, et al 
High-intensity resistance training improves glycemic control in older patients with type 2 diabetes
Diabetes Care
2002
, vol. 
25
 (pg. 
1729
-
1736
)
28
Houmard
 
J. A.
Tanner
 
C. J.
Slentz
 
C. A.
Duscha
 
B. D.
McCartney
 
J. S.
Kraus
 
W. E.
 
Effect of the volume and intensity of exercise training on insulin sensitivity
J. Appl. Physiol.
2004
, vol. 
96
 (pg. 
101
-
106
)
29
Mourier
 
A.
Bigard
 
A. X.
de Kerviler
 
E.
Roger
 
B.
Legrand
 
H.
Guezennec
 
C. Y.
 
Combined effects of caloric restriction and branched-chain amino acid supplementation on body composition and exercise performance in elite wrestlers
Int. J. Sports Med.
1997
, vol. 
18
 (pg. 
47
-
55
)
30
Sigal
 
R. J.
Kenny
 
G. P.
Wasserman
 
D. H.
Castaneda-Sceppa
 
C.
White
 
R. D.
 
Physical activity/exercise and type 2 diabetes: a consensus statement from the American Diabetes Association
Diabetes Care
2006
, vol. 
29
 (pg. 
1433
-
1438
)
31
Thrainsdottir
 
I. S.
Von Bibra
 
H.
Malmberg
 
K.
Ryden
 
L.
 
Effects of trimetazidine on left ventricular function in patients with type 2 diabetes and heart failure
J. Cardiovasc. Pharmacol.
2004
, vol. 
44
 (pg. 
101
-
108
)
32
Katsuki
 
A.
Sumida
 
Y.
Gabazza
 
E. C.
, et al 
QUICKI is useful for following improvements in insulin sensitivity after therapy in patients with type 2 diabetes mellitus
J. Clin. Endocrinol. Metab.
2002
, vol. 
87
 (pg. 
2906
-
2908
)
33
Klein
 
S.
Fontana
 
L.
Young
 
V. L.
, et al 
Absence of an effect of liposuction on insulin action and risk factors for coronary heart disease
N. Engl. J. Med.
2004
, vol. 
350
 (pg. 
2549
-
2557
)
34
Burns
 
N.
Finucane
 
F. M.
Hatunic
 
M.
, et al 
Early-onset type 2 diabetes in obese white subjects is characterised by a marked defect in β-cell insulin secretion, severe insulin resistance and a lack of response to aerobic exercise training
Diabetologia
2007
, vol. 
50
 (pg. 
1500
-
1508
)