We assessed how well dobutamine stress echocardiography predicted morbidity and mortality in 2349 patients with diabetes mellitus over a follow-up period of 13.2 years. Results showed that age, failure to achieve the target heart rate, and the percentage of ischemic segments were independent predictors of both mortality and cardiovascular morbidity. Using clinical and stress echocardiographic parameters, a simple model for risk stratification was developed.
Coronary artery disease (CAD) is the leading cause of morbidity and mortality in patients with diabetes mellitus. Risk stratification and modification in this population, therefore, is of great clinical importance. Dobutamine (Dobutrex) stress echocardiography (DSE) has become recognized as a useful tool for diagnosis and risk stratification in patients with CAD. We evaluated how well DSE predicted morbidity and mortality in a large group of diabetic patients over a period of 13.2 years. We also created a model for risk assessment based on DSE and clinical information.1
Patients and methods
Our study included 2349 diabetic patients (1338 men) who were referred for clinically indicated DSE, had adequate images, and were available for follow-up. The mean age was 67 ± 11 years. Patients underwent DSE following established methods. A peak dose of 40 mg/kg/min of dobutamine was reached using 3-minute stages. If necessary, atropine sulfate (Atrosulf-1, Isopto, Sal-Tropine) was given, up to a maximum dose of 2 mg, to increase heart rate.2 Wall motion was assessed; if no wall motion abnormality was detected during stress or at rest, DSE was considered normal. New or worsening wall motion abnormality was considered to be inducible ischemia. The target heart rate was defined as 85% of age-predicted maximal heart rate. Cardiovascular morbidity, categorized as myocardial infarction (MI) and late coronary revascularization (> 3 months), and all-cause mortality were the end points of the study. Multivariable predictors of mortality and cardiovascular morbidity were used to develop risk scores derived by assigning weights to each multivariable parameter according to parameter estimates from the Cox proportional hazards model.
A total of 50% of patients underwent DSE for cardiac risk evaluation prior to surgery; 46% were assessed for known or suspected CAD, and 4% underwent DSE for other reasons. Forty-seven percent of patients were taking insulin, 32% were taking oral antidiabetic drugs, and 21% were using diet alone to control diabetes. Among those who had cardiovascular risk factors, 40% had a family history of CAD, 56% had hyperlipidemia, 62% were smokers, 73% had hypertension, and age was a risk factor in 91% of patients (≥ 55 and ≥ 45 years of age for women and men, respectively).
Patients with ischemic DSE were older, more likely to be male, to be on insulin therapy, and to have prior MI, prior coronary revascularization, hypertension, and a family history of CAD ( < .05). Of the 603 patients who failed to achieve the target heart rate, 262 were receiving beta-blocker therapy. Dobutamine stress echocardiography was normal in 844 patients (36%). A total of 1007 patients (43%) developed inducible ischemia.
For morbidity, the follow-up period was 3.9 ± 2.7 years (maximum, 13.2 years); for mortality, the follow-up period was 5.4 ± 2.2 (maximum, 13.2 years). Death and cardiovascular morbidity occurred in 1044 (44%) and 309 (13%) patients, respectively; 116 patients had late coronary revascularization, and 193 had MI. Those without inducible ischemia were less likely than those with inducible ischemia to undergo revascularization (9% vs 19%; < .001).
Survival probabilities at 1, 3, 5, and 8 years were 89%, 74%, 60%, and 44%, respectively. The cumulative mortality rate was higher in patients with abnormal, compared with patients with normal, DSE at 1 year (13% vs 7%), 3 years (30% vs 19%), 5 years (45% vs 31%), and 8 years (61% vs 48%; < .001). Patients receiving insulin had lower survival probabilities compared with patients who were receiving oral antidiabetic agents or diet alone ( = .005). In multivariate analysis, age, failure to achieve the target heart rate, and the percentage of ischemic segments were important predictors of both mortality and morbidity. The risk of mortality was increased in patients with inducible ischemia. As shown in Figure 1, those with more extensive ischemia had greater risk. In a stepwise multivariate model, DSE had incremental value over clinical and resting echocardiographic data in predicting mortality (model χ2 increased from 243 to 270; < .001) and cardiovascular morbidity (model χ2 increased from 3 to 78; < .001).
All-cause mortality risk score.
• For the mortality risk score, patients were given 3 points for impaired left ventricular systolic function, 2 points if the percentage of ischemic segments was more than 25%, and 1 point each for insulin therapy, smoking, or failure to achieve the target heart rate. Based on their risk scores, patients were assigned to 1 of 3 risk categories: 5 to 8, 2 to 4, or 0 to 1. As shown in Figure 2, estimated 5-year survival rates for patients in the 3 risk categories were 46%, 57%, or 70%, respectively ( < .001). Age group was used to further classify patients. Figure 3 shows that the prognostic value of the mortality risk score was independent of age ( < .001).
Cardiovascular morbidity risk score
• . For the cardiovascular morbidity risk score, patients were given 3 points if the percentage of ischemic segments was ≥ 25%, 2 points for failure to achieve the target heart rate, and 1 point each for hypertension and prior MI. Based on their risk scores, patients were assigned to 1 of 3 risk categories: 2 to 6, 1, or 0. The estimated 5-year predictive values of remaining free of an ischemic event (excluding early revascularization) were 76%, 87%, and 94%, respectively ( < .001).
The value of DSE in predicting mortality and cardiovascular morbidity among diabetic patients was shown by the results of this study. Patients with ischemic or abnormal test results had markedly greater risk of cardiovascular morbidity and mortality. Independent predictors of unfavorable outcomes were percentage of ischemic segments and failure to achieve the target heart rate, and were incremental to clinical and resting echocardiographic factors. Patients with diabetes mellitus represented a high-risk population as evidenced by a high prevalence of risk factors, prior MI, and adverse outcomes. We created simple risk scores to aid in risk stratification. However, even patients in the lowest risk group had substantial mortality.
For the evaluation of CAD in patients with diabetes, myocardial perfusion imaging is recommended by the American College of Cardiology and the American Diabetes Association.3 However, several recent studies have ascertained the diagnostic and prognostic value of stress echocardiography as an alternative to myocardial perfusion imaging for the evaluation of patients with diabetes mellitus.4-10 Elhendy and colleagues assessed the prognostic role of treadmill exercise echocardiography in 563 patients with diabetes mellitus and showed that the percentage of ischemic segments, or evidence of CAD in a multivessel distribution, was a strong predictor of hard cardiac events during a 3-year follow-up period.7 Furthermore, diabetic patients with normal exercise echocardiography had a low cardiac event rate. However, inability to exercise, exercise intolerance, and failure to achieve an adequate workload may limit the utility of exercise echocardiography in patients with diabetes mellitus. Safety, feasibility, and accuracy of DSE were comparable in diabetic and nondiabetic patients.8
The prognostic role of DSE was shown in 53 patients with juvenile-onset, insulin-dependent diabetes being considered for kidney or pancreas transplantation, or both.4 During a follow-up period of 418 ± 269 days, the cardiac event rate was significantly higher in patients with abnormal DSE (45% vs 6%; = .002). Bigi and colleagues studied 259 diabetic patients with known or suspected CAD who underwent dobutamine or dipyridamole (Persantine) stress echocardiography.5 Peak wall motion score index was the only independent predictor of cardiac events, and positive stress echocardiography was associated with significantly lower event-free survival during 2 years of follow-up.
In 937 patients with diabetes mellitus who underwent exercise or dobutamine stress echocardiography, the presence of ischemia was an independent predictor of death during a mean follow-up period of 3.9 ± 2.3 years.9 Sozzi and colleagues studied the role of DSE in 396 patients with diabetes mellitus10; the peak wall motion score index was incremental to clinical data for predicting mortality. Cortigiani and colleagues compared dipyridamole and dobutamine stress echocardiography in risk stratification of CAD among 5456 patients with and without diabetes mellitus during a median follow-up period of 31 months11; both tests were equally effective in stratifying diabetic and nondiabetic patients for risk, independently of age. The incremental value of stress echocardiography over clinical and resting echocardiographic data was greater in diabetic than in nondiabetic patients. Normal DSE was associated with a worse outcome in diabetic than in nondiabetic patients. The results of these studies and our findings show that the presence, extent, and severity of inducible ischemia and inadequate chronotropic response during stress echocardiography are independently associated with unfavorable outcomes in patients with diabetes mellitus.
Stress echocardiography has become a well-established diagnostic tool in diabetic patients. Our findings confirm the prognostic utility of DSE in a large cohort of patients with diabetes mellitus during long-term follow-up. Diabetic patients may be evaluated for risk of mortality and cardiovascular morbidity using the simple risk score developed in this study.