Impact of Diabetes on Long-Term Prognosis in Patients With Unstable Angina and Non–Q-Wave Myocardial Infarction
Results of the OASIS (Organization to Assess Strategies for Ischemic Syndromes) Registry
Background—Although unstable coronary artery disease is the most common reason for admission to a coronary care unit, the long-term prognosis of patients with this diagnosis is unknown. This is particularly true for patients with diabetes mellitus, who are known to have a high morbidity and mortality after an acute myocardial infarction.
Methods and Results—Prospectively collected data from 6 different countries in the Organization to Assess Strategies for Ischemic Syndromes (OASIS) registry were analyzed to determine the 2-year prognosis of diabetic and nondiabetic patients who were hospitalized with unstable angina or non–Q-wave myocardial infarction. Overall, 1718 of 8013 registry patients (21%) had diabetes. Diabetic patients had a higher rate of coronary bypass surgery than nondiabetic patients (23% versus 20%, P<0.001) but had similar rates of catheterization and angioplasty. Diabetes independently predicted mortality (relative risk [RR], 1.57; 95% CI, 1.38 to 1.81; P<0.001), as well as cardiovascular death, new myocardial infarction, stroke, and new congestive heart failure. Moreover, compared with their nondiabetic counterparts, women had a significantly higher risk than men (RR, 1.98; 95% CI, 1.60 to 2.44; and RR, 1.28; 95% CI, 1.06 to 1.56, respectively). Interestingly, diabetic patients without prior cardiovascular disease had the same event rates for all outcomes as nondiabetic patients with previous vascular disease.
Conclusions—Hospitalization for unstable angina or non–Q-wave myocardial infarction predicts a high 2-year morbidity and mortality; this is especially evident for patients with diabetes. Diabetic patients with no previous cardiovascular disease have the same long-term morbidity and mortality as nondiabetic patients with established cardiovascular disease after hospitalization for unstable coronary artery disease.
Diabetes increases the risk of cardiovascular outcomes and mortality in patients with established cardiovascular disease (CVD). Patients with diabetes with no history heart disease have the same risk for future cardiovascular death as nondiabetic patients with a history of myocardial infarction (MI).1 Furthermore, patients with diabetes have not experienced the reduction in mortality rates that recently has been observed in nondiabetic people.2 These observations, plus the fact that the prevalence of diabetes is likely to double during the first quarter of the 21st century,3 suggest that the importance of diabetes as a cardiovascular risk factor will increase substantially.
Several previous studies have shown that patients with diabetes have higher mortality and morbidity rates than nondiabetic patients after an acute MI.4 5 6 This increased risk is also evident after implementation of both modern revascularization and pharmacological treatment strategies.7 Despite these observations, few studies have assessed the prognosis of diabetic patients after hospital admission with the most common reason for admissions to a coronary care unit: unstable angina and non–Q-wave MI. One case-control study found that patients with diabetes had higher 3-month and 1-year mortality rates than nondiabetic patients. Furthermore, patients with diabetes were less likely to have coronary angiography and treatment with β-blockers.8 Although most of these studies reported that diabetes predicts a worse outcome after an episode of unstable coronary artery disease,9 10 conflicting data exist.11 The reason for the conflicting results probably is that the available studies were small, involved relatively short follow-up, and were retrospective in nature.
We therefore compared the outcomes and treatment strategies of patients with and without diabetes mellitus in the large, prospective Organization to Assess Strategies for Ischemic Syndromes (OASIS) registry, which provides long-term information on patients with unstable coronary artery disease from 6 different countries and 95 hospitals.
A detailed description of methods and patients has been published elsewhere.12 Briefly, the OASIS registry included patients with acute ischemic chest pain within 48 hours of onset and suspected unstable angina or acute MI without initial ST elevation. Ninety-five hospitals in Australia, Brazil, Canada, the United States, Hungary, and Poland participated in the study. Each hospital enrolled consecutive eligible patients during their participation in the study between 1995 and 1996.
Data on clinical variables, procedures, and events were collected prospectively through the use of standardized forms during hospitalization. If patients were transferred from one hospital to another for a procedure, these data were also recorded. At 6, 12, and 24 months, data were collected on all interim procedures, occurrence of MI, rehospitalization for unstable angina, and death. All data were transmitted to the Canadian Cardiovascular Collaboration Project office located at the Preventive Cardiology and Therapeutics Research Program at the Hamilton Civic Hospitals Research Center. Extensive quality-control efforts were made to ensure completeness and accuracy of data. Data were complete in 99.9% of patients at hospital discharge, 96.0% at 6 months, 86.0% at 12 months, and 80.3% at 24 months. The protocol was approved by each hospital’s ethics committee. Informed consent was obtained from each patient.
The following prospective definitions were used. Diabetes was defined as a known history of diabetes mellitus treated with diet or use of oral glucose-lowering agents or insulin. MI associated with admission symptoms (onset before hospital admission) was defined as elevation of cardiac enzymes, eg, creatine kinase (CK) or diagnostic CK-MB within 12 hours of admission to at least twice the upper limit of the reference range. To identify new MI occurring beyond 24 hours after admission, we used 2 of 3 criteria of new onset of pain, new ECG changes (ST elevation of 2 mm in contiguous leads, new Q waves, or other persistent ECG changes), or new enzyme elevations: creatine kinase rise to more than twice the upper limit of the reference range (or >20% of the previous value of CK already elevated above the upper limit of the reference range) or a rise in CK-MB above the reference range. If CK or CK-MB results were not available, an increase in other less specific cardiac enzymes as for CK was accepted. Stroke was identified with a clinical diagnosis (with or without confirmation by CT scan or MRI) with definite symptoms persisting ≥24 hours. Major bleed was defined as a bleed considered fatal or life-threatening, an overt bleed requiring transfusion of ≥2 units of packed red blood cells (or equivalent), or a bleed requiring surgical intervention. Heart failure was identified through a clinical diagnosis of new-onset congestive heart failure with radiological evidence to support this diagnosis or ejection fraction measurement <45% (by echocardiography or radionuclide assessment). A patient was considered to have previous known cardiovascular disease if previous MI, stroke, congestive heart failure, PTCA, or CABG had been experienced. The combined end point was cardiovascular death, new MI, or stroke, whichever occurred first.
Baseline historical data, in-hospital management, and in-hospital outcome rates of all diabetic and nondiabetic patients enrolled in the registry were summarized. Continuous and categorical data were compared with the use of Student’s t test and Fisher’s exact test, respectively. Crude relative risks (RRs) and 95% CIs for in-hospital outcomes in diabetic patients compared with nondiabetic patients were calculated. To correct for in-hospital treatment dissimilarities, a logistic regression model was used.
Patients were followed up for 2 years after entering the registry. Kaplan-Meier event curves for diabetic and nondiabetic patients were constructed for total mortality, MI occurring after the first 24 hours of hospitalization, stroke, and new onset of congestive heart failure. Missing 2-year data were treated as censored. These curves were compared by the log-rank test. Univariate and multivariable relative hazards and 95% CIs for these outcomes were calculated with Cox regression models. Variables entered into the multivariable model included age, sex, diabetes, previous congestive heart failure, previous MI, previous stroke, previous PTCA, and previous CABG.
To determine the independent prognostic importance of various clinical factors on the 2-year prognosis after an episode of unstable angina or non–Q-wave MI, a second Cox regression model was used in which the above variables and in-hospital use of intravenous heparin, β-blockers, and calcium channel blockers were entered. In addition, interaction terms for diabetes and sex, heparin, β-blockers, and calcium channel blockers were entered. The only significant interaction was found in the model for total mortality and was seen for diabetes and sex. However, entering this interaction term into the model did not change the other included variables or their relation in the model. Relative hazards for long-term total mortality is therefore given from the model not including the interaction term. All statistical analyses were done with the use of SAS version 6.12.
The study included 8013 patients, of whom 1718 (21%) had diabetes (Table 1⇓). The prevalence of diabetes varied significantly between different countries, even after correction for differences in age and sex, with the highest prevalence in the United States (30%) and the lowest in Australia (16%).
Table 1⇑ gives the baseline characteristics. Diabetic patients were older and more likely to be female than nondiabetic patients. They also had significantly more previous cardiovascular events, including MI, congestive heart failure, and stroke, and more revascularization procedures. Before admission, more diabetic than nondiabetic patients were taking calcium channel blockers and ACE inhibitors, but prior use of β-blockers was the same in both groups.
In-Hospital Treatment and Evaluation
As outlined in Table 1⇑, after admission diabetic patients were less likely to be treated with β-blockers but significantly more likely to receive calcium channel blockers during their hospital stay than nondiabetic patients. This was also evident after correction for dissimilarities between the 2 groups with regards to age, sex, previous MI, previous stroke, and previous congestive heart failure in a multivariable logistic regression analyses. More diabetic patients received ACE inhibitor treatment than nondiabetics. Approximately three quarters of all patients received intravenous heparin treatment with no difference between the 2 groups.
The frequency of cardiac catheterization and in-hospital rates of PTCA did not differ between diabetic and nondiabetic patients. However, more diabetic patients underwent CABG surgery.
In-Hospital Morbidity and Mortality
There were no differences between patients with and without diabetes in the rate of confirmed admission MI, new in-hospital MI, or bleeding complications. Patients with diabetes had a significantly higher crude rate of in-hospital death, congestive heart failure, and stroke than nondiabetic patients (Table 2⇓). On average, patients with diabetes stayed 1 day longer in hospital than nondiabetics (11.4±8 versus 10.3±8 days, P<0.0001).
During follow-up, there were no differences in the rates of coronary angiography and PTCA between patients with or without diabetes. However, more diabetic patients (23%) than nondiabetic patients (20%) had CABG during follow-up (Table 3⇓). Cumulative actuarial event rates for the outcomes of mortality, new congestive heart failure, reinfarction, and stroke are given in Figure 1⇓. Patients with diabetes had a significantly worse long-term outcome in all event categories; the RR was 1.84 (95% CI, 1.60 to 2.19) for mortality and 1.44 (95% CI, 1.22 to 1.68) for reinfarction. This higher mortality and morbidity in patients with diabetes was also evident after adjustment for dissimilarities in baseline variables, including age, sex, and a previous history of MI, stroke, PTCA, CABG, and congestive heart failure. Absolute rates and both unadjusted and adjusted hazard ratios (by the Cox model) are presented in Table 3⇓. The negative impact of diabetes was consistent for all outcomes across all countries, and there was no interaction between diabetes and country in a multivariate analysis (data not shown).
Because patients with diabetes were more likely to have established CVD at baseline, we also performed stratified analyses according to prior CVD and reported diabetes status. Event curves for the stratified analyses are given in Figure 2⇓; corresponding crude event rates for all outcomes are given in relation to diabetic status and previous CVD history in Table 4⇓. A history of diabetes was associated with an increased event rate for all outcomes for patients both with and without established CVD. The 2-year mortality rate for diabetic patients was 20.3% for those with prior CVD compared with 13.0% for those without prior CVD. For nondiabetic patients, the corresponding figures were 12.8% and 6.9%, respectively. For all end-point categories, including the combined end point, the curves for patients with diabetes and no prior CVD and the curves for nondiabetic patients with prior CVD were nearly identical (Figure 2⇓). Indeed, a comparison of event rates for all outcomes in diabetic patients without prior CVD with nondiabetic patients with previous CVD in a Cox proportional-hazard model after adjustment for age and sex did not reveal any differences between the 2 groups (Table 5⇓). This indicates that after hospitalization for unstable angina or non–Q-wave MI, a previously healthy patient with diabetes has the same long-term prognosis as a nondiabetic patient with established CVD. Adjusted RRs between nondiabetic patients without prior CVD and the other strata are presented in Figure 2⇓.
Prognostic Importance of Clinical Baseline Factors
To determine the relative impact of clinical baseline factors on long-term prognosis, the Cox model for mortality was used. In this model, we also controlled for in-hospital treatment (use of β-blockade, calcium channel blockade, and heparin; Table 6⇓). The most powerful independent predictors were age (RR=1.58; 95% CI, 1.48 to 1.69, per 10 years), previous heart failure (RR=2.34; 95% CI, 2.01 to 2.73), and diabetes (RR=1.56; 95% CI, 1.35 to 1.79). Indeed, the fact that the RRs for 10 years of age and diabetes are almost identical suggests that the risk associated with diabetes is similar to the risk of a nondiabetic patient who is 10 years older. Previous treatment with PTCA was associated with increased survival (RR=0.68; 95% CI, 0.55 to 0.85; P<0.001), whereas previous CABG had a neutral effect (NS; P=0.27).
Prior MI, stroke, and male sex were also independently associated with long-term mortality. There was a positive interaction between diabetes and sex (P=0.002 for interaction). Diabetes increased the risk for men by 1.28 (95% CI, 1.06 to 1.56). In contrast, the risk was increased by 1.98 (95% CI, 1.60 to 2.44) among women (P<0.01 for the differences between the 2 RRs). Thus, diabetes was of greater prognostic importance for women than for men.
There are 4 major findings of this study. First, the 2-year morbidity and mortality rates after an acute episode of unstable angina or non–Q-wave MI are substantial and therefore require long-term treatment strategies. Second, the poor long-term outcome is especially evident among patients with diabetes mellitus. Third, diabetes per se has the same impact on future morbidity and mortality as already established cardiovascular disease after a new event. Fourth, the adverse impact on prognosis of diabetes is greater among women than men.
Several studies have previously shown that diabetic patients have increased short- and long-term mortality and morbidity after an acute MI.4 5 6 7 This increase has been attributed to an increased susceptibility to myocardial pump failure and reinfarction.4 13 14 In our study, patients with diabetes had a higher baseline risk factor burden than nondiabetic patients, but even after statistical correction for these baseline dissimilarities, diabetes turned out to be a major independent predictor for all event categories. An earlier study that examined this relationship was a case-control study that found a 2-fold increased mortality among diabetic patients compared with nondiabetics during 1 year of follow up, a difference largely accounted for by an increased early mortality.8 Another study showed that among patients admitted to hospital with a suspicion of MI, those with a history of diabetes had a 1-year mortality rate of 25% compared with 10% among nondiabetic patients.15 Our study showed a progressive increase in mortality and morbidity for patients both with and without diabetes. Moreover, there was a continuos widening of the event curves with time, an observation that supports the contention that an episode of unstable coronary artery disease is a marker of a chronic progressive disease, which is more advanced and rapidly progressive in diabetic than in nondiabetic patients. That patients with diabetes do have accelerated atherothrombotic disease is also supported by the fact that the diagnosis of diabetes per se had the same prognostic impact as an increase in age of 10 years.
The finding that patients with diabetes and no previous CVD experienced the same long-term prognosis as nondiabetic patients with a history of previous CVD is interesting and supports previous observations of a similar cardiovascular long-term prognosis for diabetic patients without established CVD as for nondiabetic persons who have suffered an acute MI.1 Our finding confirms this hypothesis in patients who have suffered an episode of unstable angina or non–Q-wave MI. Furthermore, because our result is derived from a large number of patients from 6 different countries and is very consistent for all cardiovascular outcomes, it also increases the generalizability of the concept, thereby reinforcing the need for meticulous care and modification of risk factors in patients with diabetes mellitus.
The relatively higher incidence of new congestive heart failure is consistent with previous studies in postinfarction patients and may be due to a more vulnerable nonischemic area, presumably because of vascular reasons or a particular concomitant diabetes-specific heart disease.16 In addition to hyperglycemia, diabetes is characterized by an increased turnover of free fatty acids, which may be one of the explanations for a diabetes-specific myocardial disease.17 Free fatty acids utilization increases myocardial oxygen consumption and may enhance intracellular accumulation of toxic intermediates.18 An encouraging finding was that patients with diabetes received the same care in terms of invasive evaluation and treatment as nondiabetic patients. In fact, they had a slightly higher rate of bypass procedures during follow-up. This presumably reflects a higher prevalence of multivessel disease in patients with diabetes compared with nondiabetic patients who have had previous angiographic studies.14 Earlier reports have found less frequent use of invasive procedures in patients with diabetes and have speculated that this may be a reason for a worse prognosis among diabetics after an unstable coronary event compared with nondiabetic patients.8 Furthermore, there was no increase in bleeding complications despite extensive use of both aspirin and heparin. Previous reports have indicated less use of anticoagulation and thrombolytic treatment among diabetic patients, presumably because of fear of treatment complications.19
We also found an interaction with diabetes and sex with respect to total mortality. The increased risk among diabetic woman compared with diabetic men has in several previous studies been interpreted as evidence that diabetic women have a worse prognosis than diabetic men.20 However, a more appropriate interpretation is that women with diabetes have the same poor long-term prognosis as diabetic men and that diabetes blunts the female protection for cardiovascular disease.
In conclusion, this study demonstrates that diabetes and unstable coronary artery disease are a common combination. It also reports that diabetes is an independent risk factor for long-term morbidity and mortality after an episode of unstable coronary artery disease even after implementation of modern evaluation and treatment strategies. The prevalence of diabetes is expected to increase as the population ages. Therefore, more knowledge regarding specific care and treatment in this fast-growing patient population is urgently needed.
Reprint requests to Professor Salim Yusuf, Preventive Cardiology and Therapeutics, Program, McMaster University, HGH-McMaster Clinic, 237 Barton St E, Hamilton, Ontario L8L 2X2 Canada.
- Received November 29, 1999.
- Revision received February 29, 2000.
- Accepted March 27, 2000.
- Copyright © 2000 by American Heart Association
King H, Aubert RE, Herman WH. Global burden of diabetes, 1995–2025: prevalence, numerical estimates, and projections. Diabetes Care. 1998;21:1414–1431.
Malmberg K, Rydén L. Myocardial infarction in patients with diabetes mellitus. Eur Heart J. 1988;9:256–264.
Herlitz J, Malmberg K, Karlsson B, et al. Mortality and morbidity during a five year follow-up of diabetics with myocardial infarction. Acta Med Scand. 1988;24:31–38.
Wilcox I, Freedman SB, Allman KC, et al. Prognostic significance of a predischarge exercise test in risk stratification after unstable angina pectoris. J Am Coll Cardiol. 1991;18:667–683.
Gowda MS, Vacek JL, Hallas D. One-year outcomes of diabetic versus nondiabetic patients with non-Q-wave acute myocardial infarction treated with percutaneous transluminal coronary angioplasty. Am J Cardiol. 1998;89:1067–1071.
Yusuf S, Flather M, Pouge J, et al, for the OASIS (Organisation to Assess Strategies for Ischemic Syndromes) Registry Investigators. Variations between countries in invasive cardiac procedures and outcomes in patients with suspected unstable angina or myocardial infarction without initial ST elevation. Lancet. 1998;352:507–514.
Stone P, Muller J, Hartwell T, et al. The effect of diabetes mellitus on prognosis and serial left ventricular function after acute myocardial infarction: contribution of both coronary disease and diastolic left ventricular dysfunction to the adverse prognosis. J Am Coll Cardiol. 1989;14:49–57.
Rodrigues B, Cam MC, McNeill JH. Metabolic disturbances in diabetic cardiomyopathy. Mol Cell Biol Biochem. 1998;180:53–57.
Ward H, Yudkin JS. Thrombolysis in patients with diabetes: withholding therapy is probably mistaken: patients should be given a choice. BMJ. 1995;310:3–4.