Franklin K, Goldberg RJ, Spencer F, Klein W, Budaj A, Brieger D, Marre M, Steg PG, Gowda N, Gore JM, for the GRACE Investigators. Implications of Diabetes in Patients With Acute Coronary SyndromesThe Global Registry of Acute Coronary Events. Arch Intern Med. 2004;164(13):1457-1463. doi:10.1001/archinte.164.13.1457
There are limited data describing the presenting characteristics, management, and outcomes of diabetic and nondiabetic patients with an acute coronary syndrome (ACS).
To examine differences in these factors, patients with ST-segment elevation acute myocardial infarction, non–ST-segment elevation acute myocardial infarction, and unstable angina were enrolled in a large multinational coronary disease registry.
The Global Registry of Acute Coronary Events is a prospective observational study of patients hospitalized with an ACS at 94 hospitals in 14 countries. The study sample consisted of 5403 patients with ST-segment elevation acute myocardial infarction, 4725 with non–ST-segment elevation acute myocardial infarction, and 5988 with unstable angina.
Approximately 1 in 4 patients presented to participating hospitals with a history of diabetes. Patients with diabetes were older, more often women, with a greater prevalence of comorbidities, and they were less likely to be treated with effective cardiac therapies than nondiabetic patients. Patients with diabetes who developed an ACS were at increased risk for each hospital outcome examined including heart failure, renal failure, cardiogenic shock, and death. These differences remained after adjustment for potentially confounding prognostic factors.
A considerable proportion of patients with an ACS has diabetes and is at increased risk for adverse outcomes compared with patients without diabetes. There are certain proven therapeutic strategies that remain underused in the diabetic population. A more widespread awareness of this increased risk and a more diligent use of proven cardiac treatment approaches are indicated for patients with diabetes who develop an ACS.
Coronary artery disease (CAD) is the leading cause of morbidity and mortality in patients with diabetes mellitus.1- 3 Patients with diabetes are more likely to experience acute myocardial infarction (AMI) and heart failure, and are at greater risk for dying after an acute cardiac event, than patients without diabetes.4- 14 These differences may be related to the severity and extent of coronary heart disease in diabetic patients, the extent of left ventricular remodeling, and the presence of significant ventricular dysrhythmias.4- 6,9,15- 20
Over the past several years, patients with diabetes have not enjoyed the same decline in CAD-related mortality as nondiabetic individuals in the United States. The poor prognosis associated with diabetes after AMI has been observed in several studies despite adjustment for age,5,9,10,13 sex,5,8,9,11 additional comorbidities,21 and coronary risk factors.6,12 Conflicting reports exist regarding the morbidity and mortality of diabetic patients treated with insulin vs oral hypoglycemic agents or diet after AMI.5,11,20- 22 It is also unclear whether the negative prognostic implications of diabetes apply equally to patients with different manifestations of acute coronary disease, including unstable angina, non–ST-segment elevation AMI (NSTEMI), and ST-segment elevation AMI (STEMI). In addition, since patients with diabetes who develop an acute coronary syndrome (ACS) appear to sustain worse outcomes than those without diabetes, it is important to determine if they are reliably receiving proven cardiac interventions under current practices.
The purpose of this study was to describe differences in the presenting characteristics, management, and hospital outcomes of diabetic and nondiabetic patients with ACSs using data from the multinational observational Global Registry of Acute Coronary Events (GRACE). This is one of the first large international studies of diabetes across the full spectrum of ACSs, including STEMI, NSTEMI, and unstable angina. Men and women of all ages hospitalized with an ACS in 14 countries throughout the world were included in this investigation.
Patients hospitalized with an ACS in 94 hospitals in 14 countries in North and South America, Europe, Australia, and New Zealand represent the sample of interest. The methods developed for GRACE have been described previously.23,24 In brief, information was obtained about patient demographic characteristics, medical history, onset of acute coronary symptoms, hospital presentation, clinical characteristics, electrocardiographic findings, use of cardiac medications, interventional procedures, and hospital outcomes.
Systematic surveillance approaches were used for the identification of patients from selected study populations and medical centers who were hospitalized with an ACS. Geographic clusters were chosen to represent populations with varying demographic, clinical, and treatment characteristics, as well as hospital systems of different sizes and treatment and diagnostic capabilities.
A total of 16 116 patients hospitalized between April 1999 and September 2001 were included in this report. The study sample consisted of 5403 patients with STEMI, 4725 with NSTEMI, and 5988 with unstable angina. The criteria for the diagnosis of STEMI, NSTEMI, and unstable angina developed for this investigation take into account clinical presentation, biochemical makers of acute ischemic injury, and electrocardiographic findings. Standardized criteria were used to define the occurrence of several key clinical outcomes including heart failure, cardiogenic shock, and renal failure during the acute hospitalization. Patients were classified as having diabetes based on the review of medical records. Information pertaining to method of diagnosis, severity and duration of diabetes, and degree of glycemic control was not available. Information regarding diabetic treatment modality (insulin requirement vs oral hypoglycemic agents and/or diet) was collected.
Differences in demographic information, medical history, and clinical characteristics among the comparison groups were examined using χ2 and t tests for discrete and categorical variables, respectively. Hospital case-fatality rates as well as heart failure, cardiogenic shock, and renal failure in diabetic and nondiabetic patients were calculated in a standard manner. Logistic multivariable regression analysis was carried out to examine differences in the risk of dying and the development of hospital complications between patients with and without diabetes while controlling for potentially confounding prognostic factors. These factors included age, sex, history of cardiovascular disease, initial creatinine and serum cholesterol levels, body mass index, smoking, blood pressure levels, and geographic region. They were controlled for because of their prognostic importance and because they differed among the comparison groups. Separate regression analyses were carried out for patients with STEMI, NSTEMI, and unstable angina. Because this observational study was nonrandomized, we did not control for the use of coronary interventional procedures or medical therapies in our multivariable-adjusted regression models.
Among patients with STEMI, NSTEMI, and unstable angina, 21%, 26%, and 25% had a history of diabetes, respectively (Table 1). Patients with diabetes who had an ACS were older and more likely to be women, and they had a higher body mass index than those without diabetes. Patients with diabetes were more likely to have additional comorbidities and coronary risk factors. They were less likely to be current cigarette smokers, however, and they presented with lower initial serum cholesterol levels.
Regarding prehospital delay, diabetic patients with STEMI were significantly more likely to present for medical attention 6 or more hours after the onset of acute coronary symptoms, while those without diabetes were more likely to present within 2 hours of onset (Table 1).
Diabetic patients with an ACS were less likely to be treated with aspirin, β-blockers, and thrombolytic agents during the index hospitalization (Table 2), and they were more likely to be treated with angiotensin-converting enzyme inhibitors, calcium channel antagonists, and oral nitrates than patients without diabetes. Statins were less often used in diabetic patients with STEMI. Coronary artery bypass surgery was performed significantly more often in diabetic patients with STEMI, although no significant difference was observed between diabetic and nondiabetic patients with NSTEMI or unstable angina. Similar numbers of diabetic and nondiabetic patients with STEMI underwent percutaneous coronary intervention (PCI), but diabetic patients with NSTEMI were less likely to undergo PCI during the acute hospitalization.
Regardless of the type of ACS, patients with diabetes were significantly more likely to develop heart failure during the acute hospitalization (Table 3). While the statistical significance of the associations observed differed according to the type of ACS, patients with diabetes were also more likely to experience cardiogenic shock and/or renal failure, and to die during the acute hospitalization. Irrespective of prehospital delay (<2 hours, 2-5.9 hours, or ≥6 hours), diabetic patients were at increased risk for each of the adverse outcomes examined compared with patients without diabetes.
In a series of subgroup analyses, we examined differences in our principal study outcomes by age, sex, and type of therapy for glycemic control (Table 3). Patients with diabetes were at increased risk for each of the outcomes studied, irrespective of age and sex, compared with those without diabetes. Younger patients with diabetes were at particularly increased risk for dying. Patients requiring insulin who presented with STEMI were at increased risk for each of the observed clinical complications compared with those treated with oral hypoglycemic agents or diet.
Finally, we examined differences in our principal study outcomes for each type of ACS while controlling for the previously described potentially confounding factors (Table 4). Across the spectrum of ACSs, diabetic individuals remained at significantly increased risk for heart failure, renal failure, cardiogenic shock, and death during the acute hospitalization compared with nondiabetic patients even after controlling for potential confounders.
The results of this large multinational observational study suggest that diabetic patients with an ACS present with a greater prevalence of cardiovascular risk factors, receive different treatments, and have a higher rate of adverse hospital outcomes than patients without diabetes. With few exceptions, these patterns were observed in patients with STEMI, NSTEMI, and unstable angina.
Approximately one quarter of patients had diabetes, irrespective of their type of ACS. This prevalence rate of diabetes is consistent with prior investigations of patients with an ACS.4,14
The findings of GRACE are consistent with those of prior studies showing that diabetic patients with an ACS are older,4,8,9,12- 14,25 more often women,9,14,18,25 more likely to be hypertensive,3,4,8 and more likely to have a history of angina4,8,14 and congestive heart failure8,14,25 than patients without diabetes. They are less likely to be smokers1,8,20,26 but more likely to be overweight.3,4,9,13,14 However, findings from regression analyses controlling for these and additional factors have shown that the increased morbidity and mortality of ACS when associated with diabetes cannot be explained by a worse risk profile alone.3 Factors such as diabetic cardiomyopathy, small vessel disease, increased platelet activity, decreased fibrinolysis, and autonomic neuropathy leading to ventricular arrhythmia14,15 may account for the poor prognostic outlook of diabetic patients.
Patients with diabetes have been shown to benefit to a greater extent than nondiabetic patients from the use of β-blockers in the immediate postmyocardial infarction period.4,9 However, our results confirm prior reports that fewer diabetic patients with an ACS are treated with β-blocking agents than nondiabetic patients with an ACS.8,9,14,25- 27 Reasons for this underuse may include the possibility of masking hypoglycemic symptoms and exacerbating impaired glucose tolerance.
Our data showed that approximately two thirds of diabetic patients were treated with angiotensin-converting enzyme inhibitors during their acute hospitalization, a percentage greater than that for nondiabetic patients. These encouraging findings are consistent with other recent data.8,9,14,25,26 Angiotensin-converting enzyme inhibitors slow myocardial remodeling and the progression of diabetic nephropathy, and they should be initiated and titrated aggressively after the onset of acute coronary disease. In the GRACE population, calcium channel antagonists were used in a greater proportion of diabetic than nondiabetic patients. This may be because our diabetic population included a greater percentage of patients with hypertension and angina at baseline. Nitrates were also more commonly used in diabetic patients, probably for similar reasons.
Several investigations have examined the use of thrombolytic therapy in STEMI and have shown a significant benefit for diabetic patients.7,19,26 However, our results suggest that thrombolytic agents are used significantly less often in patients with diabetes. Other recent investigations of patients with STEMI who presented within 12 hours of symptom onset confirm that those with diabetes are less likely to receive appropriate reperfusion therapy (thrombolysis or PCI).28 The lesser use of thrombolytic therapy in diabetic patients may be related to delayed presentation and more atypical symptoms.9,13,14,19 Our data show that those with STEMI (ie, those eligible for acute reperfusion strategies), present significantly later for hospital care. However, on further examination of outcomes stratified according to time from symptom onset to hospital presentation, diabetic patients continued to fare worse regardless of extent of prehospital delay. The underuse of thrombolytic agents in patients with diabetes and an ACS may also reflect the concern for an increased risk of retinal or intracranial hemorrhage in the diabetic population.26 A very small number of patients without ST elevation also received thrombolytic agents; they probably represent those suspected of STEMI at admission, but who eventually were diagnosed with NSTEMI or unstable angina by hospital discharge.
The frequency of PCI performed in diabetic patients in GRACE was similar to that for nondiabetic patients in spite of current trends toward earlier revascularization for high-risk diabetic patients. Early studies of PCI showed higher rates of restenosis for patients with diabetes.19 With the addition of various stents and new antiplatelet agents, however, PCI is used in diabetic patients with increasing success.29- 31
Coronary artery bypass surgery was more frequently performed in diabetic patients with STEMI in our study, which may reflect the more diffuse nature of atherosclerotic disease in these high-risk individuals. The Bypass Angioplasty Revascularization Investigation (BARI) suggested that diabetic patients derive greater long-term benefit from a surgical approach than from PCI if they have multivessel disease.32,33 However, diabetic patients with NSTEMI and unstable angina were no more likely to undergo coronary artery bypass grafting in GRACE, in contrast to the results of the OASIS Registry and other studies.25,27
Nearly twice as many diabetic as nondiabetic patients with an ACS experienced heart failure during their index hospitalization. This increased risk has been previously observed and has not been attributable to conditions typically associated with heart failure such as decreased ejection fraction, hypertension, obesity, or prior myocardial infarction. Differences in the risk of heart failure are also not adequately explained by older age, sex, duration of coronary disease, or uncontrolled hyperglycemia.4,13,18 Some investigators have invoked the role of diabetic cardiomyopathy as a potential causative factor in the development of heart failure in patients with diabetes and an ACS. Similar mechanisms likely increase the risk of cardiogenic shock in patients with diabetes and an ACS.
Diabetic patients in GRACE had elevated serum creatinine levels when they were admitted to participating hospitals, which increased their risk of acute renal failure when they also had an ACS. In the index hospital period, renal failure was diagnosed nearly twice as often in diabetic patients as in nondiabetic patients. This trend persisted across the spectrum of ACSs.
Our results suggest that diabetic patients treated with insulin were at greater risk for adverse outcomes after an ACS than those who had a form of diabetes not requiring insulin. Recent investigations that have controlled for adverse risk factors in diabetic individuals requiring insulin (advanced age, longer duration of disease, more severe disease, and a greater prevalence of coronary risk factors) have found that insulin dependence no longer portended a worse outcome once these variables were controlled for.26
In the present study, hospital case-fatality rates for diabetic patients with an ACS were almost twice as high as those for nondiabetic patients. These findings have been previously reported.26,27,34 Our study demonstrated particularly high death rates among diabetic patients younger than 65 years. These higher death rates may reflect differential use of evidence-based therapies as well as increased rates of cardiogenic shock, stroke,25 renal failure, and heart failure in younger diabetic patients. Aggressive and immediate medical management of diabetic patients with proven cardiac therapies and revascularization strategies may help bridge this gap during both the acute hospitalization and a long-term follow-up.
The strengths of this investigation include its multinational perspective, the complete spectrum of ACSs experienced by the large number of patients studied, the use of standardized criteria for defining ACS and hospital outcomes, and the use of regression approaches to control for the influence of other factors in examining the relation of diabetes to several important hospital outcomes. Several limitations must be kept in mind in interpreting the results of this study, however. These include that the diagnosis of diabetes was based on the review of medical records alone; no information was available about diabetes duration or adequacy of control, and patients with undiagnosed diabetes may have been misclassified. In this case, an underestimation of our study outcomes in diabetic patients should have occurred. We were also unable to examine the effects of other potential confounders, such as alcohol consumption, on our principal study end points.
The results of this large multinational observational study suggest that patients with diabetes who develop an ACS are at increased risk for adverse hospital outcomes and receive different therapeutic interventions than those without diabetes. Given the high-risk profile of these patients, further rationale is provided to increase the use of effective cardiac medications and coronary interventions in patients with diabetes who develop acute coronary disease.
Correspondence: Kristen Franklin, MD, Department of Medicine, Division of Cardiovascular Medicine, University of Massachusetts Medical School, 55 Lake Ave N, Worcester, MA 01655 (email@example.com).
Accepted for publication February 25, 2003.
This project was supported by an unrestricted educational grant from Aventis Pharma (Bridgewater, NJ). We thank the hospitals participating in the GRACE study for their cooperation (further information is available at http://www.outcomes.org/grace).