Stroke is an infrequent yet potentially devastating complication of acute myocardial infarction (MI). Previously published data show stroke occurrence rates to be 1% to 2% for hospitalized patients with acute MI, whether or not fibrinolytic therapy was administered.1-3 This contemporary, community-based, observational study examined temporal trends (1986—1999) in the incidence, risk, and outcomes of 6,325 metropolitan Worcester, Massachusetts, residents hospitalized with acute MI without prior stroke.
Patients and methods
Our investigation was part of the Worcester Heart Attack Study, an ongoing study analyzing temporal trends in Worcester metropolitan area residents with validated acute MI.4-7 Trained nurses and physicians reviewed medical records from the Worcester Standard Metropolitan Statistical Area hospitals for eight annual periods between 1986 and 1999. Demographic, clinical, and outcome data on the occurrence of acute stroke in patients with confirmed acute MI who had no history of stroke were compiled and analyzed.
Differences in the characteristics of patients with acute MI according to the diagnosis of acute stroke were analyzed using chi-square and t tests. Logistic regression analyses were performed to identify independent variables correlating with the development of stroke, temporal trends in the likelihood of the development of stroke, and factors associated with hospital death for patients with acute MI.
Of 6,325 patients with acute MI, 92 (1.5%) developed acute stroke. Data from 1991 to 1999 confirmed that 78% of strokes were ischemic or embolic, and the rest were hemorrhagic. There was no difference in stroke rates (1.5%) between patients who received fibrinolysis (n = 1,396) and patients who did not (n = 4,929).
Acute stroke factors. Compared with acute MI patients who did not develop stroke, those who did were more likely to be older than 75 years (59.8% versus 36.0%; P < .001), to be women (58.7% versus 41.1%; P < .001), and to have an anterior wall location of their acute MI (51.1% versus 39.9%; P < .05). In addition, the hospital course of patients who had a stroke was also more frequently complicated by atrial fibrillation (31.5% versus 14.8%; P < .001). These patients were less like-
ly to receive aspirin therapy (76% versus 66.3%; P < .05). Interventional procedures, including cardiac catheterization, percutaneous coronary intervention, coronary artery bypass graft (CABG) surgery, and treatment with fibrinolytic therapy, did not substantially correlate with acute stroke.
Factors associated with acute stroke were further stratified according to age (≥ 75 years or < 75 years). Acute MI patients older than 75 years were more likely to develop stroke in the presence of atrial fibrillation (38.2% versus 21.3%; P < .005) if they underwent CABG surgery (5.6% versus 2.1%) and if they had previous heart failure (43.6% versus 29.5%; P < .05); however, they were less likely to develop stroke if they had prior angina (16.4% versus 31.4%; P < .05). Younger patients (< 75 years) had an increased risk of stroke in
the presence of anterior wall acute MI (56.8% versus 38.3%; P < .05), atrial fibrillation (21.6% versus 11.2%;
P < .05), and cardiogenic shock (13.5% versus 5.5%; P < .05).
After adjusting for specific patient characteristics, age of 75 years or older (odds ratio [OR], 2.53; 95% confidence interval [CI], 1.05—6.11) and atrial fibrillation (OR, 1.91; 95% CI, 1.17–3.11) were the only factors that correlated with an increased risk of acute stroke. A trend toward increased rates of stroke was observed in women (OR, 1.51; 95% CI, 0.96–2.38) and patients with anterior wall acute MI (OR, 1.54; 95% CI, 1.00–2.36). The increased risk associated with advanced age was not altered by controlling for additional variables relating to therapies and procedures.
Time trends and the incidence of acute stroke. There was an initial decrease in the occurrence of stroke between 1986 and 1990; however, between 1991 and 1999, there was a continual increase over time (figure). The annual period of 1999 had the highest incidence of acute MI patients with acute stroke (3%).
After adjusting for certain patient demographic, historical, and clinical variables that were potentially confounding factors for the development of stroke,1,2,8 trends in the risk of stroke over time were reexamined for four 2-year periods. The odds of developing acute stroke increased significantly and progressively over time (table 1). Further analyses controlling for specific hospital therapies and procedures, hospital complications, length of stay, and hospital discharge status reinforced these findings (table 1). When temporal data were further stratified according to age, the increased risk of developing stroke between 1986 and 1988 and between 1997 and 1999 was greater for patients 75 years or older (OR, 9.93; 95% CI, 2.56—36.60) compared with patients younger than 75 years (OR, 4.63; 95% CI, 1.46–14.70).
Hospital case fatality rates and acute stroke. In-hospital death was significantly more likely to occur in acute MI patients with stroke than in those without stroke (32.6% versus 12.8%; P < .001). This held true for patients younger than 75 years (27% versus 7.5%) and for those 75 years or old-er (36.4% versus 22.3%). After advanced age and cardiogenic shock, multivariate-adjusted regression analysis identified stroke as the most potent predictor of in-hospital death (table 2).
Despite improvements in coronary revascularization techniques and adjunctive pharmacotherapies used for treatment of acute MI,
and despite controlling for specific high-risk demographic and clinical markers, our population-based study suggested that there was a progressive increase in the incidence of acute stroke in patients with validated acute MI between 1986 and 1999. Previously published studies from the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries-1 (GUSTO-1) trial,1,2 a meta-analysis of fibrinolysis trials by the Fibrinolytic Therapy Trialists (FTT) collaborative group,3 and the Second National Registry of Myocardial Infarction (NRMI-2)8 all report overall stroke rates of 1% to 2% after development of acute MI. Although the early incidence of stroke observed in our study was similar, data from our most recent study cohort (1999) suggest an approximate twofold increase in events. Furthermore, unlike previous studies, the majority (approximately 80%) of patients with validated acute MI in the Worcester Heart Attack Study did not receive fibrinolysis, suggesting that the majority of cerebrovascular events were nonhemorrhagic.
Risk factors for stroke complicating acute MI. The median age of acute MI patients in our study increased from 68 years in the 1986 to 1988 period to 71 years in the 1997 to 1999 period, and the prevalence of patients 75 years or older increased from 30% to 41%. This supports the hypothesis that advanced age contributes directly to the higher risk of stroke after acute MI in the community setting. In fact, advanced age was identified as one of the strongest predictors of post-MI stroke in our study, as well as in a number of clinical trials and observational studies.8-11
Interestingly, it appeared that other risk factors for stroke had a differential impact according to patient age. Within this community-based population, anterior wall location of acute MI was a significant predictor of stroke in patients younger than 75 years (57% with stroke versus 38% without stroke; P < .05), but not in patients 75 years or older (47% with stroke versus 42% without stroke). Commonly accepted as an important cause of stroke following an acute MI, anterior location of acute MI is often followed by 3 months of prophylactic anticoagulation therapy to prevent cardioembolic events. No prospective study has clearly established thromboemboli from the left ventricle as an important cause of acute stroke. In both a GUSTO-1 analysis2 and the Multicenter Diltiazem Postinfarction Trial (MDPIT), anterior location of acute MI was not a reliable predictor of stroke in hospitalized patients.12 It may be that left ventricular apical thrombi, which were an important cause of stroke in younger patients with acute MI in our study, play less of a role in older patients, who were heavily represented in these trials.
It is likely that the pathophysi-
ology of post-MI stroke in older pa-tients is more heterogeneous than in younger patients with acute MI. This is supported by our data, which suggest that atrial fibrillation, previous heart failure, and CABG surgery were all significantly associated with post-MI stroke. Accordingly, it is unlikely that targeting only elderly patients with anterior MI for prophylactic anticoagulation therapy will successfully reduce post-MI stroke. It is more likely that aggressive, coordinated management of a number of more general risk factors (atrial fibrillation, congestive heart failure, hypertension, etc) would be more efficacious.
An additional risk factor for stroke not addressed in the current study but worthy of mention, particularly in the elderly, is the presence of noncardiac vascular disease. Previous evidence of cerebrovascular disease has been identified as a potent predictor of acute MI—associated stroke in several studies. A history of stroke or transient ischemic attack was the strongest predictor of nonhemorrhagic stroke after acute MI in a multivariate analysis of the GUSTO-1 dataset.1 A history of pe-
ripheral vascular disease was independently associated with increased risk of stroke among elderly patients with acute MI in the Cooperative Cardiovascular Project.11 Although we excluded patients with prior stroke from our analyses, it is probable that many (if not most) patients aged 75 years and older in our population had some degree of “silent” cerebrovascular disease. This underlying condition, in conjunction with the heightened platelet activation and hypercoagulability associated with the onset of acute MI, could be a fertile substrate for de novo cerebrovascular events. The results of the recent Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering (MIRACL) trial, which documented a 50% reduction in nonfatal and fatal strokes in patients randomly assigned to receive high-dose early HMG-CoA reductase inhibitor (statin) therapy after acute MI, suggest that plaque stabilization in the noncoronary vasculature may be an unexpected benefit of this therapy.13
The results of this study suggest that the rate of stroke after acute MI has increased in the last decade. Mortality associated with this complication remains extraordinarily high. Increasing age and specific comorbidities were significantly associated with stroke risk in this
community-based population. These findings reinforce the need for aggressive identification and treatment of modifiable risk factors for stroke following acute MI. n