Physical activity and fitness levels are inversely associated with the occurrence of cardiovascular events and mortality in type 2 diabetes (T2D).1,2 Cardiac ischemia and hypoglycemia are 2 notable complications that may occur during exercise in these patients.3 High insulin levels, occurring in response to insulin resistance in T2D, may enhance exercise-associated peripheral glucose uptake, potentially inducing hypoglycemia.4,5 This may be a more critical issue among T2D patients treated with insulin secretagogues, which increase insulin levels, than among those treated with insulin sensitizers, which reduce insulin levels.
The American Diabetes Association (ADA) recommends that exercise stress testing be conducted among T2D patients older than 35 years planning a moderate/ high-intensity exercise program to detect unstable cardiovascular disease.3 Diabetic patients and their healthcare providers are advised to monitor glucose response to exercise and to adjust food intake and medications accordingly. However, there are currently no guidelines governing glucose monitoring at the time of stress testing. Such monitoring may be justified given that electrocardiographic changes6,7 and symptoms8 induced by hypoglycemia may be difficult to distinguish from those attributable to cardiac ischemia. Our aim was to compare the magnitude of change in capillary blood glucose (CBG) levels during maximal exercise stress testing among T2D patients treated with insulin secretagogues (sulfonylurea medication) and those treated with insulin sensitizers and/or diet and exercise.
METHODS
Subjects in this study were identified at the time of baseline assessment for participation in a trial evaluating the impact of supervised exercise on body weight in T2D. All evaluations were performed at the Cardiovascular Health Improvement Program (CHIP), McGill University's cardiac rehabilitation unit. Procedures were approved by the Institutional Review Board of McGill University. Participants provided written informed consent. Eligibility criteria included a diagnosis of T2D and a body mass index (BMI) of 28 kg/m2 or more. Patients on insulin therapy were excluded from participation as were patients with a chronic illness that could affect body weight or physical activity levels (eg, malignancy, inflammatory arthritis). Between May 2004 and September 2004, among 97 individuals who completed preliminary eligibility questionnaires, 50 candidates underwent more detailed evaluation and stress testing at CHIP. Most were recruited through the diabetes clinics of the McGill University Health Centre in Montreal, Quebec, Canada. Among those who underwent stress testing, 9 were excluded from the present analysis because they were evaluated at an alternative site where CBG assessment was not systematically performed (7 candidates) or had received a dextrose tablet prior to stress testing (2 candidates, see Glucose Testing). The remaining 41 participants constitute the subjects of the present analysis.
Exercise Testing
Following interview and physical examination by a study physician, subjects underwent stress testing following the Bruce protocol, with continuous electrocardiographic monitoring. Workload was increased by approximately 3 metabolic equivalents (METs) every 3 minutes and blood pressure was assessed immediately prior to each workload increment. The exercise stress test (EST) was terminated upon volitional exhaustion or any of the American College of Sports Medicine criteria for ending an EST.9 Use of a Bruce protocol was permitted as an alternative, at the discretion of study physicians. Prior to the EST, subjects were instructed to avoid alcohol consumption, caffeine intake, and cigarette smoking for 6 hours and to fast for a 2-hour period.
Glucose Testing
CBG testing was performed (One-Touch meter, Lifescan(R)) immediately preceding and following each EST. As prespecified by our protocol, 2.75 g of dextrose was administered in tablet form to participants with a CBG value of less than 5.5 mmol/L prior to stress testing, for the prevention of hypoglycemia.
Serum Parameters
Following an overnight fasting period, venous blood was sampled among subjects who were enrolled in the pilot supervised exercise trial. Total cholesterol, high-density lipoprotein cholesterol (HDL-C), and hemoglobin A1C were directly assessed using a Hitachi 717 analyzer (Roche Diagnostics, Canada). Hemoglobin A1C, the proportion of red blood cells with a glycosylated hemoglobin protein, reflects overall blood sugar control during the preceding 3 months (ie, the average life span of a red blood cell).
Anthropomorphic Measures
Weight in light clothes was ascertained to the nearest 0.1 kg using a standing beam balance scale. Height was assessed to the nearest 0.1 cm (shoes removed). BMI was calculated (weight in kilograms divided by the square of height in meters). Blood pressure was determined using a mercury sphygmomanometer and auscultation of the left brachial artery with participants seated, following a rest period of at least 5 minutes.
Analyses
Sulfonylurea users and nonusers were compared (t tests and [chi]2 tests as appropriate, SAS software version 8.2) in terms of pre-EST CBG, post-EST CBG, change in CBG, age, gender, diabetes duration, BMI, hemoglobin A1C, blood pressure, total cholesterol/HDL ratio, use of cardiac/glucose lowering medication, EST peak workload, and percent maximal predicted heart rate achieved during EST. Variables that differed between users and nonusers at the P < .25 were considered for inclusion in a multivariate linear regression model with change in CBG as the dependent variable. Final model selection was based on forward selection with a test for backward elimination and the likelihood ratio test.10
RESULTS
As indicated in Table 1, approximately half of the subjects reported use of sulfonylurea medication (80% glyburide, 20% gliclazide). Sulfonylurea users were significantly older, had been diagnosed as having T2D for a longer period of time, and were more likely to report use of thiazolodinedione medication. Most participants were at the stage 1 level of obesity (35-45 kg/m2). There were no current smokers in this cohort. Three participants were known for stable cardiovascular disease (2 sulfonylurea users, 1 nonuser).
Serum Parameters
Venous blood sampling was performed among 34 of the 41 subjects (83%). Among these 34 subjects, mean lipid parameters and hemoglobin A1C values were close to current ADA target levels (Table 1). The remaining 7 subjects did not undergo venous blood sampling because they were ineligible for the supervised exercise trial from which the present study cohort was derived. Reasons for ineligibility included anginal symptoms during EST (1 user), very low exercise capacity (3 users), exacerbation of knee pain secondary to osteoarthritis (1 user), BMI less than 28 kg/m2 (1 user), and withdrawal of consent for unspecified reasons (1 user).
EST Results
One subject (sulfonylurea user) underwent a modified Bruce protocol because of a history consistent with angina of recent onset. The EST was terminated because of chest pain symptoms that resolved with rest. This individual was started on [beta]-blocker and aspirin therapy, and referred to a nonstudy cardiologist. The remaining subjects underwent stress testing as per the Bruce protocol. Sulfonylurea users achieved a significantly lower workload than nonusers (7.6 +/- 2.0 vs 9.3 +/- 2.4 METs, P = .02) and were less likely to achieve 85% or more of the maximum predicted heart rate during exercise stress testing (14/21, 67% vs 19/20, 95%, P = .02).
Pre-EST glucose values ranged between 5.4 and 15.2 mmol/L among sulfonylurea users and between 4.9 and 23.2 mmol/L among nonusers. Unadjusted comparisons between users and nonusers were suggestive of higher pre-EST glucose levels (9.5 +/- 2.9 vs 8.0 +/- 4.0 mmol/L, P = .18) and greater decline in CBG levels during stress testing (-1.9 vs -0.6 mmol/L, P = .004, Figure 1) among users. Post-EST glucose values ranged from 4.9 to 12.2 mmol/L among users and 4.5 to 20.7 mmol/L among nonusers. Use of sulfonylurea medication was associated with a 1.4 mmol/L change in CBG (95% confidence interval, CI: -1.8 to -0.6 mmol/L) in a linear regression model adjusted for diabetes duration, use of thiazolodinedione medication, and pre-EST glucose level. In this model, each 1-mmol/L increment in pre-EST glucose level was associated with -0.2 mmol/L change in glucose during the EST (95% CI: -0.1 to -0.3 mmol/L). Each 5-year increment in diabetes duration associated with a 0.5 mmol/L increase in CBG (95% CI: 0.1 to 0.9 mmol/L). Use of thiazolodinedione medication was not significantly associated with change in CBG (change of 0.3 mmol/L, 95% CI: -0.5 to 0.7 mmol/L).
DISCUSSION
Most subjects in the present study reported use of the insulin sensitizer metformin. Those who were also taking sulfonylurea medication tended to have a longer history of diabetes, consistent with the decline in pancreatic islet cell function that tends to occur over time in T2D patients. A longer history of diabetes was associated with an increase in glucose levels during stress testing. However, compared to T2D patients controlled with an insulin sensitizer or without medication, subjects receiving sulfonylurea medication had a significantly greater decline in glucose levels during exercise stress testing. Changes in glucose levels were not related to the absolute workload achieved.
The greater decline in sulfonylurea users during stress testing may be the result of higher insulin levels in these patients and the synergistic effect of insulin and exercise on peripheral glucose uptake.4 Glucose is required only when lipolysis is inhibited by the insulin response to food intake during low-intensity exercise, but during bouts of intense exercise, as during exercise stress testing, glucose and glycogen are required for energy.11
The ADA advocates a premeal glucose target of 4 to 7 mmol/L. Given that hypoglycemic symptoms may occur at glucose levels of less than 4 mmol/L, it is plausible that a decrease in glucose levels of at little as 1 mmol/L could result in hypoglycemia-range blood glucose levels among diabetic patients with optimal glycemic control. The palpitations, diaphoresis, and dizziness associated with hypoglycemia8 may be difficult to distinguish from angina-related symptoms.12 There are case reports of severe hypoglycemia precipitating atrial fibrillation.13 Hypoglycemia may itself provoke cardiac ischemia.14
None of the subjects in the present study developed symptoms or electrocardiographic changes attributable to hypoglycemia. One possible explanation may be that the mean pretest glucose level was well above the hypoglycemic threshold of 4 mmol/L. Second, the study protocol dictated a dextrose tablet be administered prior to stress testing for subjects in whom CBG levels were less than 5.5 mmol/L during stress testing, further diminishing the likelihood of detecting symptomatic hypoglycemia. Finally, with only 20 sulfonylurea users in the present study, the sample size may not have been sufficiently large to detect symptomatic hypoglycemia among sulfonylurea users. In a study examining glucose changes in T2D patients following 1-hour exercise sessions in a cardiac rehabilitation setting, a postexercise CBG less than 3.3 mmol/L was documented during 1.8% of sessions.15
In conclusion, results indicate that sulfonylurea use and diabetes duration are determinants of glucose change during stress testing. Glucose assessment prior to stress testing should be considered at least among sulfonylurea-treated patients who report premeal glucose levels of 4 to 7 mmol/L. A larger investigation will be necessary to determine the incidence of clinically significant hypoglycemia among sulfonylurea-treated patients undergoing stress testing. Future studies could consider venous blood sampling in place of CBG testing, permitting blinded assessment of the relationship between prestress and poststress test glucose levels and occurrence of symptoms and electrocardiographic changes. The increasing incidence of T2D and greater emphasis on strict blood sugar control3 clearly justify further study concerning the frequency, detection, and prevention of hypoglycemia during exercise stress testing.
Acknowledgments
Ms Barbara Craig and Ms Carmela D'Avella, diabetes educators at the McGill University Health Centre, greatly facilitated recruitment by identifying candidates for our efficient recruitment assistant, Martine Lecomte. Ms Debbie Chan capably coordinated study procedures, and Mr Youssef Toubouti provided invaluable assistance with statistical analyses.
References