Author + information
- Received August 24, 2000
- Revision received January 9, 2001
- Accepted January 24, 2001
- Published online May 1, 2001.
- Abdou Elhendy, MD, PhD∗,* (, )
- Adelaide M Arruda, MD∗,
- Douglas W Mahoney, MSc† and
- Patricia A Pellikka, MD, FACC∗
- ↵*Reprint requests and correspondence: Dr. Patricia A. Pellikka, Division of Cardiovascular Diseases, Mayo Clinic, 200 1st Street SW, Rochester, Minnesota 55905
The aim of this study was to assess the incremental value of exercise echocardiography for the risk stratification of diabetic patients.
There are currently insufficient outcome data in diabetic patients to define the role of stress echocardiography as a prognostic tool.
We studied the prognostic value of exercise echocardiography in 563 patients with diabetes mellitus (mean age 64 ± 11 years, 336 men) and known or suspected ischemic heart disease (IHD).
Cardiac events occurred in 50 patients (cardiac death in 23 and nonfatal myocardial infarction [MI] in 27) during a median follow-up period of three years. Event rate was lower in patients with normal as compared to those with abnormal exercise echocardiography at one year (0% vs. 1.9%), three years (1.8% vs. 11.9%), and five years (7.6% vs. 23.3%), respectively (p = 0.0001). Patients with multivessel distribution of echocardiographic abnormalities had the highest event rate (2.9% at one year, 15.2% at three years, and 32.8% at five years). In an incremental multivariate analysis model, exercise echocardiography increased the chi-square of the clinical and exercise ECG model from 29 to 44.8 (p = 0.0001).
Exercise echocardiography provides incremental data for risk stratification of diabetic patients with known or suspected IHD. Patients with a normal exercise echocardiogram have a low event rate. Patients with multivessel distribution of exercise echocardiographic abnormalities are at the highest risk of cardiac events, as one-third of these patients experience cardiac death or nonfatal MI during the five years following exercise echocardiography.
Diabetes mellitus is a major risk factor for cardiovascular morbidity and mortality (1,2). Ischemic heart disease (IHD) is the leading cause of death in diabetic patients, and the related risk is independent of other conventional risk factors for IHD (3–5). Additionally, diabetic patients were shown to have a higher morbidity after acute myocardial infarction (MI) than nondiabetic patients (6). Therefore, the noninvasive diagnosis and risk stratification of IHD in diabetic patients is important for the selection and optimization of therapeutic interventions, which may improve survival and reduce complications of IHD in this population (1,7–10). Exercise electrocardiography is the most widely used test for the diagnosis and functional evaluation of IHD (11,12). However, the value of the test may be limited by the presence of baseline ST-segment abnormalities as well as by the presence of exercise-induced myocardial ischemia. Because of these limitations, the American Diabetes Association and the American College of Cardiology recommended the use of exercise myocardial perfusion imaging for prognostic stratification of diabetic patients, based on the published data of the incremental prognostic value of this method in patients with and without diabetes mellitus (1,13). It was concluded that there are currently insufficient outcome data in diabetic patients following stress echocardiography to define its role as a prognostic tool (1). Therefore, the aim of this study was to evaluate the incremental value of exercise echocardiography for the risk stratification of diabetic patients with known or suspected IHD, relative to clinical data.
The study included 631 patients with diabetes mellitus who underwent exercise echocardiography at the Mayo Clinic, Rochester, Minnesota, between January 1990 and December 1995. Diabetes mellitus was defined by the presence of a fasting blood glucose ≥140 mg/dl on at least two occasions or requirement for insulin or oral hypoglycemic agents. The Institutional Review Board approved the study. Twelve patients (2%) were excluded because they declined participation in research or refused to provide follow-up status, and 33 patients (5%) were excluded because of inadequate echocardiographic images. Twenty-three patients (4%) were lost to follow-up. The final population of the study consisted of 563 patients.
Exercise echocardiography protocol
Patients underwent symptom-limited treadmill exercise testing according to the Bruce protocol in 82%, Naughton protocol in 10%, and modified Bruce protocol in 8%. Both standard blood pressure and 12-channel electrocardiographic (ECG) monitoring were performed. Two-dimensional echocardiographic images were obtained from the standard parasternal and apical windows before and immediately after exercise. Both digitized and videotape-recorded images were used for interpretation of the studies (14). Ejection fraction at rest was measured using a previously validated modification of the method of Quinones et al. (15)or by visual estimation (16). Regional wall motion was assessed semiquantitatively by an echocardiographer, blinded to clinical information. Wall motion at rest and with exercise was scored 1 through 5 according to a 16-segment model (17). Wall motion score index was determined at rest and exercise as the sum of the segmental scores divided by the number of visualized segments. The difference between exercise and rest wall motion score index was reported as Δ wall motion score index. The development of new or worsening wall motion abnormalities (WMA) was considered indicative of myocardial ischemia. A WMA present at rest and unchanged with exercise was classified as fixed. Exercise echocardiography results were defined as abnormal if there was ischemia or fixed WMA (18).
To assess the number of vascular regions with echocardiographic abnormalities, the anterior, apical, anteroseptal, and midinferoseptal segments were assigned to the left anterior descending coronary artery, the anterolateral and inferolateral segments to the left circumflex, and the inferior and basal inferoseptal segments to the right coronary artery. The exercise ECG was considered positive for ischemia if there was horizontal or downsloping ST-segment depression of ≥1 mm at 80 ms after the J-point, nondiagnostic if the baseline ST-segment was abnormal or if the patient was receiving digitalis, and negative for ischemia in the absence of the criteria described in the previous text. Workload was measured by metabolic equivalents (METs).
Follow-up was obtained by mailed questionnaires and scripted telephone interviews. Events were verified by contacting the patients’ primary physician and reviewing medical records and death certificates. The end points considered were hard cardiac events defined as nonfatal MI and cardiac death. Sudden unexpected death occurring without another explanation was included as cardiac death. Coronary revascularization procedures during the follow-up period were also noted. Patients who had revascularization before other events were censored at the time of revascularization.
Continuous variables were reported as mean ± SD, and comparisons between groups were based on the Wilcoxon rank-sum test. Categorical variables were summarized as percentages, and group comparisons were based on the chi-square test. Survival free of the end point of interest was estimated by the Kaplan-Meier method. Univariable and multivariable associations of clinical and exercise echocardiographic variables with the end points were assessed in the Cox proportional hazards framework. Variables were selected in a stepwise forward selection manner with entry and retention set at a significance level of 0.05. Results of these analyses were summarized as risk ratios with corresponding 95% confidence intervals. The incremental value of exercise echocardiographic information over clinical and exercise ECG data was assessed in four modeling steps. The first step consisted of fitting a multivariable model of only clinical data. Variables selected from the first step were then used as baseline risk factors, and exercise ECG and hemodynamic variables were added in a stepwise forward selection manner. Variables selected from the first two steps were then used as a baseline risk model, and rest echocardiographic variables were added to this model in a stepwise forward selection manner. In the final step, exercise echocardiographic abnormalities were added.
The mean age of the study group was 64 ± 11 years. There were 336 men and 227 women. Two hundred twenty-five patients (40%) were receiving insulin (with or without concomitant oral hypoglycemic agents), 261 patients (46%) were receiving oral hypoglycemic agents, and 77 (14%) were treated by diet alone. One hundred sixty-eight patients (30%) had previous MI and 155 (30%) had previous myocardial revascularization (coronary artery bypass grafting in 60, angioplasty in 79, and both procedures in 16 patients). Hypertension was present in 362 patients (64%), hypercholesterolemia in 337 patients (60%), and smoking in 318 patients (57%). Reasons for termination of the exercise stress test were fatigue in 344 patients (61%), dyspnea in 111 (20%), angina in 33 (6%), arrhythmias in 4 (0.7%), ST-segment changes in 10 (2%), and leg distress in 61 (11%).
Eighty-four patients underwent coronary revascularization prior to any cardiac event and were censored. Revascularization was performed early (<3 months) in 27 patients and late in 57 patients. Patients who underwent revascularization were more likely to have had ischemia by echocardiography (63% vs. 37%, p < 0.001), exercise-induced angina (24% vs. 7%, p < 0.001), and ECG evidence of ischemia (45% vs. 20%, p < 0.0001).
During a median follow-up of three years (maximum eight years), 50 hard cardiac events occurred at a median of 2.5 years (range five days to seven years) after the test. These consisted of cardiac death in 23 patients and nonfatal MI in 27 patients. Event-free survival was 99% at one year, 97% at three years, and 83% at five years. Table 1demonstrates the clinical features of patients with and without cardiac events.
Resting WMA were detected in 241 patients (43%). Among these patients, 134 (56%) had a history of previous MI, 104 (43%) had Q-waves on baseline ECG, and 87 (36%) had resting WMA in absence of a history of MI or Q-waves on the baseline ECG. The exercise echocardiogram was abnormal in 340 patients (60%). Ischemia (exercise-induced WMA) was detected in 229 patients (41%). Of these patients, 130 (57%) also had resting WMA. Among the 229 patients with exercise-induced WMA, 81 (35%) had a history of typical angina, 34 (15%) had angina during the test, and 89 (39%) had a positive exercise ECG. No significant differences were seen among patients receiving insulin, oral hypoglycemic agents, and those treated by diet alone with regard to the incidence of an abnormal stress echocardiogram (60%, 59%, and 66%) or the incidence of ischemia by exercise echocardiography (42%, 39%, and 44%), respectively.
Hemodynamic and echocardiographic features of patients with and without events
Table 2demonstrates the hemodynamic and exercise echocardiographic data in patients with and without hard cardiac events. No significant differences were detected between patients with and without events regarding heart rate, systolic blood pressure, and rate pressure product at rest or at maximal exercise. Patients with cardiac events achieved a lower workload. The ECG was more frequently nondiagnostic in patients with cardiac events.
Patients with cardiac events had a higher wall motion score index, both at rest and with exercise, a lower resting ejection fraction, a larger difference between rest and exercise wall motion score index, and a larger percentage of ischemic segments as compared to patients without cardiac events (Table 2). Clinical, exercise stress and echocardiographic variables associated with an increased risk of hard cardiac events in the univariate analysis are demonstrated in Table 3.
Event rates according to exercise echocardiographic abnormalities
Hard cardiac events occurred in 45 (13%) of 340 patients with an abnormal exercise echocardiogram and in 5 (2%) of 223 patients with a normal exercise echocardiogram. Event rate was lower in patients with normal as compared to those with abnormal exercise echocardiography at one year (0% vs. 1.9%), three years (1.8% vs. 11.9%), and at five years (7.6% vs. 23.3%), overall p = 0.0001. Event rates according to the presence and severity of exercise echocardiographic abnormalities are presented in Figures 1 and 2. ⇓⇓Patients with multivessel distribution of echocardiographic abnormalities had the highest event rate (32.8% at five years).
Predictors of cardiac events in the multivariate analysis models
In the multivariate analysis of clinical, exercise ECG, and echocardiographic parameters combined, independent predictors of hard events were a history of MI (chi-square = 2.1, [confidence interval] CI: 1.2–3.8, p = 0.01) and percentage of abnormal segments with exercise (chi-square = 2.2, CI: 1.4–2.1, p = 0.0001).
Using the stepwise incremental model, independent variables predictive of hard cardiac events are demonstrated in Table 4. Exercise echocardiography provided incremental value for risk stratification, additional to clinical and exercise stress data as indicated by the increase of the chi-square of the incremental model by the use of echocardiographic data (p = 0.0001). The most significant echocardiographic parameters in this model were resting ejection fraction and percentage of ischemic segments.
Recognition of the presence and severity of IHD—and, more importantly, the risk of future cardiac events in diabetic patients—is important for the optimization of treatment (1,2,7–10). The value of exercise echocardiography in the prognostic stratification of patients with known or suspected IHD has been studied (19–22). However, there are no previous studies describing the utility of exercise echocardiography in the risk stratification of diabetic patients.
The current study
In this study, 563 diabetic patients with known or suspected IHD underwent exercise echocardiography and were followed for a median of three years. Fifty patients (9%) experienced hard events (23 cardiac deaths and 27 nonfatal MI) during follow-up. The event rate was higher in patients with abnormal as compared to patients with normal exercise echocardiogram (11.9% vs. 1.8%, respectively, at three years). Patients with an exercise echocardiogram showing a multivessel distribution of abnormalities were at the highest risk for cardiac events. Approximately one-third of these patients experienced cardiac death or nonfatal MI during the five years following the exercise echocardiogram. In the Bypass Angioplasty Revascularization Investigation (BARI) trial (10), the five-year mortality rate after coronary artery bypass grafting in diabetic patients with multivessel disease was 8.2%. This is much lower than the spontaneous event rate in diabetic patients with multivessel distribution of exercise echocardiographic abnormalities in the present study. During the first few years of follow-up in the present study, the event rate was low for patients with normal exercise echocardiography; none of these patients had cardiac events at the end of the second year. However, the event rate increased gradually after two years to reach 7.6% by the end of the fifth year. This may be related to the progression of IHD. A repeat stress test after the second year may be beneficial in reassessing the risk status of patients with an initially negative study.
Patients with an exercise echocardiogram indicative of single-vessel IHD had a risk of hard events intermediate between patients with a normal test and those with abnormalities in multivessel distribution. No cardiac events occurred in the first year following the test in these patients. However, in contrast to patients with a normal test, the event rate began to increase after the first year. These patients may benefit from aggressive treatment of risk factors. If revascularization is not performed, repeating the test after one year may be useful to detect progression of functional abnormalities.
Independent predictors of cardiac events
In an incremental multivariate analysis model, a history of previous MI was the only clinical variable associated with increased risk of hard events. Among exercise stress test variables, workload was additive to clinical parameters. Exercise echocardiography provided data incremental to clinical and exercise stress parameters and increased the chi-square of the clinical and exercise ECG model from 29 to 44.8 (p = 0.0001). The most powerful echocardiographic predictors were resting ejection fraction and the percentage of ischemic segments with exercise. These findings indicate that the prognosis in diabetic patients is not only related to resting left ventricular function but also to the extent of exercise-induced ischemia.
Symptoms and ECG changes
The occurrence of cardiac death or nonfatal MI was poorly predicted by a history of angina or angina during the test. This may be related to the high frequency of silent ischemia in diabetic patients (23,24). Among patients with exercise-induced ischemia (new or worsening WMA) in our study, only 15% had angina during the stress test, and 35% had a history of typical angina. These data indicate that the absence of symptoms should not be interpreted as an indicator of a low risk status in diabetic patients. Similarly, exercise-induced ST-segment depression was not predictive of hard cardiac events. This may be explained by the high prevalence (29% in this study) of nondiagnostic exercise ECG in diabetic patients. This high prevalence of resting ECG abnormalities may reflect the trend to the selection of patients with such abnormalities for an exercise stress test with imaging.
In our study, 36% of patients with resting WMA had no history or ECG findings suggestive of previous MI. The higher prevalence of silent MI in diabetic patients and the possible occurrence of left ventricular dysfunction attributed to diabetic cardiomyopathy may explain this (25). These findings indicate that exercise echocardiography can play an important role in the clinical assessment of diabetic patients by evaluation of myocardial function at rest as well as during exercise.
Exercise capacity, expressed as METs, was related to improved outcome. This indicates that in diabetic patients able to exercise, an exercise stress test may be preferable to a pharmacologic stress test because of the prognostic information provided by the exercise capacity. However, acquisition of the images after exercise may underestimate the severity of abnormalities due to possible rapid reversibility of ischemic changes. Pharmacologic stress echocardiography does not suffer from this limitation (26). Nonetheless, the present study showed that the extent of WMA immediately after exercise was directly related to the event rate. This study selected a population with a relatively lower risk profile, as patients were able to exercise. Inability to exercise is a known independent predictor of adverse outcome. This should be of particular consideration in diabetic patients, as the ability to exercise also indicates the absence of severe peripheral vascular disease, with its known impact on prognosis.
Bates et al. (27)studied the role of dobutamine stress echocardiography in the stratification of 53 patients with juvenile onset, insulin-dependent diabetes mellitus who were considered for kidney and/or pancreas transplantation. Dobutamine stress echocardiography accurately stratified patients who were at risk of future cardiac events. Kang et al. (13)studied 1,271 patients with known or suspected IHD who underwent thallium-201/stress technetium-99m sestamibi dual-isotope myocardial perfusion single-photon emission computed tomography with exercise or adenosine stress testing. Patients were followed for 23.7 ± 7.7 months. The incidence of hard events was 4.3% per year. Nuclear testing added incremental value over clinical variables.
Exercise or pharmacologic stress myocardial perfusion imaging has been recommended by the American Heart Association for evaluation of IHD in diabetic patients. Echocardiography has the advantages of wider availability and lower cost, and it avoids the injection of radioactive material. In addition, left ventricular systolic function is assessed. The current study indicates that exercise echocardiography is a useful alternative to myocardial perfusion imaging in the evaluation of diabetic patients.
The type and the duration of diabetes mellitus were not defined in this study. Therefore, the impact of these parameters on the outcome of patients and their correlation with echocardiographic abnormalities could not be evaluated.
Exercise echocardiography is useful for risk stratification of diabetic patients with known or suspected IHD, and it provides prognostic information incremental to clinical and exercise ECG variables. Patients with a normal exercise echocardiogram have a low cardiac event rate and can be excluded from further diagnostic evaluation for the next two years. The prognosis of diabetic patients with known or suspected IHD is dependent on both resting left ventricular function and the extent of exercise-induced myocardial ischemia. Patients with multivessel distribution of exercise echocardiographic abnormalities are at the highest risk of cardiac events; one-third of these patients experience cardiac death or nonfatal MI within five years.
☆ This study was supported by the Mayo Foundation.
- confidence interval
- electrocardiogram, electrocardiographic
- ischemic heart disease
- metabolic equivalents
- myocardial infarction
- wall motion abnormalities
- Received August 24, 2000.
- Revision received January 9, 2001.
- Accepted January 24, 2001.
- American College of Cardiology
- ↵(1998) Consensus development conference on the diagnosis of coronary heart disease in people with diabetes: 10–11 February 1998, Miami, Florida. American Diabetes Association. Diabetes Care 21:1551–1559.
- Grundy S.M,
- Benjamin I.J,
- Burke G.L,
- et al.
- Mak K.H,
- Moliterno D.J,
- Granger C.B,
- et al.
- Pyorala K,
- Pedersen T.R,
- Kjeksus J,
- Faergeman O,
- Olsson A.G,
- Thorgeirsson G
- Antiplatelet Trialists’ Collaboration
- Bypass Angioplasty Revascularization Investigation (BARI) investigators
- Detre K.M,
- Guo P,
- Holubkov R,
- et al.
- Ellestad M,
- Wan M
- Dujardin K.S,
- Enriquez-Sarano M,
- Rossi A,
- Bailey K.R,
- Seward J.B
- Schiller N.B,
- Shah P.M,
- Crawford M,
- et al.
- Armstrong W.F,
- Pellikka P.A,
- Ryan T,
- Crouse L,
- Zoghbi W.A
- Marwick T.H,
- Mehta R,
- Arheart K,
- Lauer M.S
- McCully R.B,
- Roger V.L,
- Mahoney D.W,
- et al.
- Olmos L.I,
- Dakik H,
- Gordon R,
- et al.
- Marchant B,
- Umachandan V,
- Stevensen R,
- Kopelman P.G,
- Timmis A.D
- ↵Annonu AHM, Abdel Fattah A, Mokhtar SM, Ghareeb S, Elhendy A. Left ventricular systolic and diastolic functional abnormalities in asymptomatic patients with noninsulin-dependent diabetes mellitus. J Am Soc Echocardiogr. In press.
- Elhendy A,
- van Domburg R.T,
- Poldermans D,
- et al.