Abstract
Background and Purpose—The impact of asymptomatic coronary artery disease on the risk of major vascular events in patients with cerebral infarction is unknown.
Methods—Four hundred five patients with acute cerebral infarction underwent carotid, femoral artery, thoracic, and abdominal aorta ultrasound examination. Of 342 patients with no known coronary heart disease, 315 underwent coronary angiography. We evaluated the 2-year risk of major vascular events (myocardial infarction, resuscitation after cardiac arrest, hospitalization for unstable angina or heart failure, stroke, or major peripheral arterial disease events) in patients with known coronary heart disease (n=63), and in the no known coronary heart disease group (n=315) as a function of coronary angiographic status (n=315).
Results—At 2 years, the estimated risk of major vascular events was 11.0% (95% confidence interval, 8.2–14.7). According to baseline coronary angiography, estimated risk was 3.4% in patients with no coronary artery disease (n=120), 8.0% with asymptomatic coronary artery stenosis <50% (n=113), 16.2% with asymptomatic coronary artery stenosis ≥50% (n=81), and 24.1% with known coronary heart disease (P<0.0001). Using no coronary artery disease as the reference, the age- and sex-adjusted hazard ratio (95% confidence interval) of vascular events was 2.10 (0.63–6.96) for asymptomatic coronary stenosis <50%, 4.36 (1.35–14.12) for asymptomatic coronary stenosis ≥50%, and 6.86 (2.15–21.31) for known coronary artery disease.
Conclusions—In patients with nonfatal cerebral infarction, presence and extent of asymptomatic stenoses on coronary angiography are strong predictors of major vascular events within 2 years.
Introduction
Patients with a previous ischemic stroke have a well-documented risk of myocardial infarction, ranging from 4% within 3 months of stroke onset to 1% per year thereafter.1–5 Indeed, the prevalence of asymptomatic coronary artery disease (CAD) in patients with stroke or transient ischemic attack is high in autopsy series of fatal stroke, with up to 80% of patients having coronary plaques and 40% having coronary stenosis ≥50%.6 Likewise, recent coronary angiography studies of patients with nonfatal stroke/transient ischemic attack found a consistent prevalence of coronary stenoses ≥50% in the range of 20% to 25%.7,8
The impact of the presence and the severity of CAD on the risk of future vascular events has never been evaluated prospectively in patients with cerebral infarction. We previously reported the prevalence of CAD in a cohort of 405 consecutive patients with acute ischemic stroke on MRI, of whom 16.6% had a known history of coronary heart disease (CHD).7 Among these 405 patients, imaging of extracranial arteries, including carotid and vertebral arteries, thoracic and abdominal aorta, femoral arteries, as well as coronary angiography (only in patients with no known CHD), was performed to detect the presence, severity, and extent of asymptomatic atherosclerosis, particularly CAD.7 Prospectively collected 2-year follow-up of this cohort allows the correlation of the presence and the extent of symptomatic and asymptomatic CAD to the risk of cardiovascular events.
Methods
Study Population
The design, methods, and baseline results of the Asymptomatic Myocardial Ischemia in Stroke and Atherosclerotic Disease (AMISTAD) study have been published previously.7 AMISTAD is a prospective, single-center registry of patients with acute ischemic stroke, designed to assess the prevalence of CAD and its impact on short- and long-term prognosis, as well as other locations of atherosclerotic disease. All patients aged ≥18 years, consecutively admitted with acute ischemic stroke documented by neuroimaging and a Rankin Scale <5, were offered participation, and were enrolled within 10 days of symptom onset, after providing written informed consent. Pregnant women, patients with other nonvascular diseases associated with a life expectancy of <30 months, and patients with cerebral infarction caused by carotid or vertebral artery dissection, or secondary to a revascularization procedure, were excluded. All included patients underwent coronary angiography except if they had a history of CAD (defined as acute coronary syndrome, myocardial infarction, or prior coronary revascularization). Among the 785 patients consecutively assessed for eligibility between June 2005 and December 2008, 405 were enrolled (patients not enrolled either had exclusion criteria, eg, had no documented cerebral infarction on neuroimaging, were bedridden with a Rankin Scale >4, or refused to sign informed consent). Of these, 27 patients with no known CAD were excluded from the analysis because they either refused (n=24) or had a contraindication (n=3) to undergo coronary angiography.
The research protocol was approved by the Ethics Committee of Paris Bichat-Claude Bernard (n°2004/11, 05/09/2004) and of Ile de France No1 Hotel Dieu (Amendment n°1, 0611369, 14/09/2006; and amendment n°2, 0611445, 24/11/2006).
Risk Factor Evaluation
All patients had an evaluation of their demographic characteristics and risk factors after a face-to-face interview using a structured questionnaire. At inclusion, blood pressure was measured after 5 minutes of rest in a supine position; blood samples were collected in fasting conditions before any change in lipid-lowering treatment. Arterial hypertension was defined as treated hypertension at discharge, diabetes mellitus as treated diabetes mellitus at discharge, hypercholesterolemia as previous lipid-lowering treatment or admission low-density lipoprotein-cholesterol (LDL-C) >4.14 mmol/L (>160 mg/dL). Smoking history was classified as current (any smoking in the past 9 months), past, or never smokers. Family history (first-degree relative) of stroke or myocardial infarction was recorded.
Evaluation of Coronary Artery Atherosclerosis
In patients with no known history of CHD, coronary angiography was performed at a median 8 days after stroke onset (interquartile range, 6–11 days) to evaluate the presence and the severity of coronary atherosclerosis. Each coronary artery was classified as normal (ie, absence of plaque), plaque with stenosis <50%, 50% to 69% stenosis, 70% to 99% stenosis, or complete occlusion. The most severe coronary artery lesion was considered to categorize patients as having no CAD, asymptomatic stenosis <50%, and asymptomatic stenosis ≥50%. Readers of coronary angiography were blinded to other investigations. Patients were also categorized according to the number of atherosclerotic coronary artery lesions.
Follow-up
Follow-up visits were scheduled between 3 and 6 months after enrollment and every year thereafter. The current report is based on the database-lock in December 2010, when the last patient completed a 2-year follow-up. At each follow-up visit, the patient’s treatment, blood pressure, lipid profile (except at the 6-month visit), and any occurrence of clinical events or hospitalizations were recorded.
The main outcome was the time to the first major vascular event, including vascular death, nonfatal cardiac events, nonfatal stroke, or major peripheral arterial event. Vascular death included fatal stroke, fatal MI, and other cardiovascular death. Other cardiovascular death included any sudden death, including unobserved and unexpected death (eg, while sleeping) unless proven otherwise by autopsy; death after a vascular surgery, vascular procedure, or amputation (except for trauma or malignancy); death ascribed to heart failure; death after a visceral or limb infarction. Any myocardial infarction or stroke followed by death, whatever the cause, in the subsequent 28 days was considered as a fatal myocardial infarction or fatal stroke. Cardiac events included myocardial infarction, resuscitation after cardiac arrest, and hospitalization for unstable angina or cardiac insufficiency. Major peripheral events included all events related to noncoronary or cervicocephalic arterial disease leading to hospitalization or revascularization (eg, new or worsening of claudication leading to revascularization, surgery for ruptured aneurysm, cholesterol emboli syndrome).
Study Conduct and Funding
Screening and enrollment were performed by 2 full-time clinical research assistants, supervised by a data manager, who were in charge of verifying enrollment criteria; ensuring completeness of the case report forms; collecting risk factor data and laboratory test results; arranging neuroimaging, carotid, aortic, and femoral ultrasonography examinations and coronary angiography; and collecting the reports. Data audit was performed independently by another clinical research assistant. Carotid ultrasonography examinations were performed by senior ultrasonographers (P.C.L., E.M., L.C., V.A., S.R., J.K., and P.-J.T.), after the same protocol. Other carotid ultrasound visits were performed as close as possible to baseline and 1-year follow-up for an ancillary study by a technical research assistant, which is not part of this report. Abdominal aorta echography and femoral artery ultrasonography examinations were performed by 2 ultrasonographers (J.K. and S.R.). Coronary angiography was performed by senior interventional cardiologists (G.D., J.-M.J., L.F., and P.G.S.), and results were stored digitally for off-line review. For all these investigations, specific case report forms were completed immediately and collected by clinical research assistants. The database concerning cross-sectional data were locked in April 2009.
Follow-up information was obtained by neurologists in the outpatient clinic during face-to-face interviews. Patient missing appointments for these visits underwent a structured telephone interview. If the patient could not be contacted, a close relative or family doctor was interviewed. When available, medical records for all outcome events, including death, were reviewed by medical specialists blinded to clinical data. For remaining patients with no follow-up information, vital status was ascertained using the Register for Births, Marriages, and Deaths from the city of birth (or from the city of Nantes in the case of birth outside of France). Strokes were validated by experienced neurologists (P.C.L. and P.A.). Cardiac events were adjudicated by an experienced cardiologist (P.G.S.). Peripheral arterial events were validated by consensus among P.C.L., P.A., and P.G.S.
Funding and sponsorship of the study were provided by SOS-ATTAQUE CEREBRALE Association (a nonprofit stroke survivor association).
Statistical Analysis
All analyses were performed in patients eligible for 2-year follow-up, defined as all patients with cross-sectional data on coronary atherosclerosis status who completed ≥1 postbaseline follow-up visit (Figure I in the online-only Data Supplement). Continuous variables are expressed as mean (SD) or median (interquartile range). Categorical variables are expressed as frequencies and percentages. We compared the mean baseline and follow-up blood pressure levels and lipid concentrations using the paired Student t test; median triglyceride concentrations were compared using the Wilcoxon signed-rank test.
We estimated and compared the 2-year major vascular event rates according to the 4 baseline CAD subgroups (no disease, asymptomatic coronary stenosis <50%, asymptomatic coronary stenosis ≥50%, and known symptomatic disease) using the Kaplan–Meier method and log-rank test. Patients who died from nonvascular causes were censored at the time of death. Using patients without CAD as the reference, we estimated the relative risk of cardiac and cerebrovascular recurrence for each CAD subgroup. Estimates were calculated from a Cox proportional hazards model adjusted for age and sex. A further adjustment was done for hypertension, dyslipidemia, current smoking, familial history of CHD, atrial fibrillation, and ipsilateral intracranial or extracranial arterial stenosis ≥50%. Our primary analyses concerned the entire study group. Further analyses were restricted to patients with no known history of CHD. Adjusted relative risks associated with presence and severity of asymptomatic CAD were calculated, first using patients with no coronary plaque as the reference group, and second using the patients with no coronary stenosis ≥50%. Differences in 2-year major vascular event rates according to the number of diseased coronary arteries were tested using a trend test in a Cox proportional hazards model. Finally, we performed a sensitivity analysis excluding the patients enrolled with stroke history, because the prognosis of prevalent and incident strokes could be different.
Statistical testing was done at the 2-tailed α level of 0.05. Data were analyzed using the SAS software package, release 9.2 (SAS Institute, Cary, NC).
Results
Among 405 patients with cerebral infarction enrolled in the AMISTAD study between June 2005 and December 2008, all 378 patients included in the cross-sectional analysis7 were entered in the follow-up phase. In December 2010, only 1 patient had no follow-up information and was excluded from the 2-year interim analysis (Figure I in the online-only Data Supplement). The median follow-up for the remaining 377 patients was 36 months (interquartile range, 26–48 months). Thirty-two patients did not complete the 2-year follow-up, including 20 who died and 12 with shorter follow-up (3 with 3-month, 2 with 6-month, and 7 with 12-month follow-up). The baseline demographic and clinical characteristics of the 377 patients included in the 2-year interim analysis are shown in Table I in the online-only Data Supplement.
Medication Use and Change in Main Risk Factor Profile
Secondary prevention medications used at discharge and at follow-up visits within 1 year of stroke onset are listed overall (Table II in the online-only Data Supplement) and by baseline CAD subgroups (Tables III–VI in the online-only Data Supplement). At discharge, 81.7% of patients received blood pressure–lowering therapy (56.2% were treated by dual or triple therapy), 89.9% received statins, and 97.1% received antithrombotic therapy (84.1% received antiplatelet agents, and 27.6% received anticoagulants). According to baseline CAD subgroups, blood pressure–lowering drugs and anticoagulant therapy were more frequently used in patients with versus those without CAD. There was no relevant difference in the use of lipid-lowering and antidiabetic agents. Overall, as well as in each baseline CAD subgroup, rates of self-reported medication adherence at 3 months were very high and did not change thereafter. As shown in Table VII in the online-only Data Supplement, significant changes in the main modifiable cardiovascular risk factors were found 3 months after discharge and remained stable thereafter. The mean systolic blood pressure level achieved at 1 year was 132 mm Hg and ranged from 129 to 134 mm Hg across baseline CAD subgroups (Figure 1A). The mean LDL-C concentration achieved at 1 year was 2.0 mmol/L (78 mg/dL) and ranged from 1.9 to 2.2 mmol/L (73–86 mg/dL) across baseline CAD subgroups (Figure 1B), with an absolute average reduction of 1 mmol/L (40 mg/dL).
Mean baseline and follow-up systolic blood pressure (SBP) levels (A) and low-density lipoprotein-cholesterol (LDL-C) concentrations (B) by baseline coronary artery disease subgroup; 95% confidence intervals are plotted.
Baseline CAD Subgroup and Risk of Recurrent Events
During 2-year follow-up, there were 7 vascular deaths (including 5 from cardiac disease), 13 first nonfatal coronary events (3 nonfatal myocardial infarctions, 2 resuscitated cardiac arrests, and 8 hospitalizations for unstable angina or heart failure), 15 nonfatal strokes, and 8 peripheral arterial disease events. Nine nonvascular deaths and 4 deaths of unknown cause also occurred. A total of 41 patients experienced ≥1 major vascular events, giving an overall estimated 2-year risk of 11.0% (95% confidence interval [CI], 8.2–14.7). As shown in Figure 2, the Kaplan–Meier event curves as a function of time from stroke symptom onset for combined major vascular events differed significantly across the 4 baseline CAD subgroups (log-rank, P<0.0001). Two-year Kaplan–Meier estimates were 3.4% (95% CI, 1.3–8.9) in patients with no CAD, 8.0% (95% CI, 4.2–14.8) with asymptomatic coronary stenosis <50%, 16.2% (95% CI, 9.7–26.3) with asymptomatic coronary stenosis ≥50%, and 24.1% (95% CI, 15.3–36.8) with known CHD. Using patients without CAD as the reference, the age- and sex-adjusted hazard ratio of cardiac and cerebrovascular events was 2.10 (95% CI, 0.63–6.96) for patients with asymptomatic coronary stenosis <50%, 4.36 (95% CI, 1.35–14.12) for patients with asymptomatic coronary stenosis ≥50%, and 6.86 (95% CI, 2.15–21.91) for patients with known CHD. Similar results were found after additional adjustment for hypertension, dyslipidemia, current smoking, family history of CHD, atrial fibrillation, and ipsilateral intra- or extracranial arterial stenosis ≥50%, or after restriction of the analyzes to the patients enrolled with incident stroke (Table 1).
Risk of Major Vascular Events at 2-Years by Baseline Coronary Artery Disease Subgroups
Cumulative incidence curves of the composite end point of major vascular events in the baseline coronary artery disease subgroups.
Prognostic Values of Presence and Severity of Asymptomatic CAD
In an analysis restricted to patients with no known CHD at baseline, the presence of asymptomatic coronary plaque was associated with an increased risk of major vascular events at 2-year follow-up (age- and sex-adjusted hazard ratio, 3.73; 95% CI, 1.22–11.44; Table 2). Moreover, the 2-year risk of major vascular events was also related to the number of coronary arteries affected by plaque, ranging from 6.4% in patients with 1-vessel disease to 14.9% in patients with 3-vessel disease (age- and sex-adjusted P for trend, 0.003). Similar results were found on the presence and the extent of asymptomatic coronary stenosis ≥50% (age- and sex-adjusted P for trend, 0.002; Table 2). In our fully adjusted multivariate analysis (including intra- and extracranial ipsilateral stenosis and atrial fibrillation), hazard ratio for combined vascular events was 4.03 (95% CI, 1.21–13.39) for asymptomatic coronary plaques and 3.77 (95% CI, 1.44–9.87) for asymptomatic coronary stenosis. In sensitivity analysis restricted to patients with incident stroke, the fully adjusted hazard ratios were 3.85 (95% CI, 1.13–13.14) and 4.09 (95% CI, 1.48–11.27), respectively.
Despite the evident lack of statistical power, when outcome measure was restricted to major coronary events (cardiac death, nonfatal myocardial infarction, and resuscitation after cardiac arrest), a nonsignificant difference was found between patients with and without asymptomatic coronary plaque (2.1% versus 0.8%; log-rank, P=0.40), or between patients with and without asymptomatic coronary stenosis ≥50% (3.7% versus 0.9%; log-rank, P=0.076).
Discussion
We found that asymptomatic CAD as diagnosed on coronary angiography in patients with a recent nonfatal cerebral infarction was associated with a higher 2-year risk of major vascular events than in those without CAD. The sex- and age-adjusted 2-year major vascular event risk increased gradually with the severity of asymptomatic CAD <50% or ≥50% or with known CHD. The 2-year major vascular risk also increased with the number of coronary arteries involved irrespective of the degree of stenosis. Importantly, these events occurred despite achieving optimal secondary prevention targets with a very high proportion of patients at guideline-recommended LDL-C concentration and blood pressure, and with a high rate of use of preventive medications.9 In patients with known CHD, asymptomatic CAD, or no CAD, the average LDL-C concentration was well <2.5 mmol/L (100 mg/dL) after discharge, with an absolute average difference of 1.0 mmol/L (40 mg/dL) between baseline and follow-up measurements (Figure 1). Similarly, average blood pressure was well <140/90 mm Hg, with an average reduction of 9/2 mm Hg between baseline and follow-up measurements (Figure 1). These absolute differences in LDL-C and blood pressure levels partly explain the low risk of major vascular events observed. Patients with asymptomatic CAD ≥50% were also more frequently treated by dual antiplatelet therapy (at discharge 29% used 2 antiplatelet agents in the asymptomatic CAD and known CHD groups versus 14% in the group with no CAD), β-blockers (59% versus 18% in the group with no CAD), and angiotensin-converting enzyme inhibitors (27% versus 13%). In addition, and despite implementation of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial results in our center, patients with asymptomatic CAD ≥50% had a more aggressive treatment, with 9 asymptomatic patients undergoing preventive revascularization (3 patients had main trunk artery stenosis), which has likely further decreased the risk of coronary events.10,11 Despite this, the group with asymptomatic CAD ≥50% had a 2-year risk of 16.2% (95% CI, 9.7–26.3), which was 3-fold higher than in patients with no CAD. This risk was even higher in patients with multiple-vessel disease ≥50% (20.7%; P=0.003 versus no CAD). Practical implications may be important on the necessity or not to detect asymptomatic CAD in patients with cerebral infarction. Although the COURAGE trial has shown that asymptomatic CAD can be addressed only with best medical therapy, residual risk in our study was markedly increased in comparison with patients without CAD, despite optimal medical therapy, suggesting there may be a role for additional interventions, including revascularization in patients with cerebral infarction and asymptomatic CAD. This, however, remains to be tested. Epidemiological studies and long-term (4–5 years) clinical trials have shown that the most frequent cause of death after an index stroke is cardiac death.1–3,12,13 Neuroprotective trials that randomized patients very early after stroke onset found a 4.1% risk of fatal myocardial infarction 90 days after the index stroke.4 Therefore, detection of asymptomatic CAD in patients with cerebral infarction may help to stratify the risk of future vascular or cardiac events, but this remains to be tested in a randomized controlled trial. Pending demonstration of a clinical benefit, routine angiography in patients with cerebral infarction would, however, be premature. Our findings strengthen the American Heart Association/American Stroke Association statement that in stroke/transient ischemic attack patients, routine noninvasive testing for CHD should be considered in the presence of significant carotid disease.14
Association of 2-Year Major Vascular Risk With Presence and Severity of Asymptomatic CAD Diagnosed at Baseline
Our study had strengths and limitations. Our study was not population based, but it was hospital based in 1 of the largest stroke units in Paris, France, working as a primary care referral center for acute stroke admissions with a dedicated catchment area. The baseline characteristics of our population are in line with what can be expected in the general stroke population and seem representative of patients with stroke. Although single-center cohort studies tend to be less externally valid than population-based multicenter studies, this is less of a concern when examining internal correlations with outcomes collected during prospective follow-up than it is in cross-sectional analyses. Therefore, our study gives a fair evaluation of the relationship between asymptomatic CAD and future, incident, or recurrent vascular events. Second, we have excluded the most severe stroke patients (Rankin Scale, 5; ie, bedridden patients), who are also likely to die from their index stroke, but may also be at risk for an early recurrent fatal vascular event. However, we have already reported on the prevalence of CAD in fatal stroke in a large autopsy series.6 In this study, we found a higher prevalence of severe (>50% stenosis) CAD than in the present series (37.5% versus 25.7%).6 If anything, exclusion of patients with Rankin Scale 5 has rather decreased the relationship between asymptomatic CAD and future cardiac and cerebrovascular events. Finally, the observed 2-year risk of 11.0% (95% CI, 8.2–14.7) for incident major vascular events during the follow-up was lower than expected given the large composite end point used. As previously mentioned, this may be, in part, because of the best medical care applied to these patients, as exemplified by the LDL-C concentrations maintained throughout the follow-up, as obtained in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels trial, and the further 9-mm Hg decrease in systolic blood pressure also maintained during the follow-up, similar to that obtained in the Perindopril Protection Against Recurrent Stroke Study trial.15,16 Also, many of patients detected with a severe coronary stenosis underwent percutaneous angioplasty, one third of them having main trunk coronary stenosis, a subgroup for which a beneficial preventive effect of revascularization has not been excluded in the COURAGE trial.11,17 Finally, given the small number of major coronary events, we lacked statistical power to detect a difference in this hard end point, and despite our multivariate-adjusted analysis we cannot exclude the possibility of residual confounding. One final theoretical limitation was the use of coronary angiography rather than coronary intravascular ultrasound to detect plaque in coronary artery. However, intravascular ultrasound is limited by only a partial evaluation of coronary arteries, as compared with coronary angiography that evaluates the entire length of the arteries.
Our study also had strengths. We included all patients admitted with cerebral infarction regardless of ischemic stroke subtype and underlying disease, given the fact that our autopsy study found no difference in the prevalence of coronary artery atherosclerosis and its severity across stroke subtypes.6 Consequently, our population was not limited to ischemic stroke due to atherosclerosis, and was representative of the entire population with cerebral infarction. We used gold standard conventional angiography for coronary artery stenosis diagnosis. Follow-up rate was good (96.7%) and secondary prevention treatment was nearly optimal, with a remarkably proportion of patients reaching guidelines-recommended targets: controlled blood pressure at baseline (<140/90 mm Hg) and further decrease by 9 mm Hg for systolic blood pressure, and an on-treatment LDL-C well <100 mg/dL with an average decrease of LDL-C from baseline of 1 mmol/L (40 mg/dL; Figure 1). Such an effective prevention is a strength of our study.
In conclusion, this is the first study to correlate coronary angiographic findings with subsequent cardiovascular events in patients with cerebral infarction. In patients with nonfatal cerebral infarction, asymptomatic CAD is highly predictive of future major vascular events, even in patients receiving optimal secondary prevention and achieving high blood pressure and LDL-C targets. Risk increased with the extent and the severity of coronary arteries and the extent of extracranial atherosclerotic disease. These results suggest that the benefits of detection of asymptomatic CAD and subsequent preventive interventions in patients with cerebral infarction should be tested in randomized controlled trials.
Disclosure
None.
Acknowledgments
We are indebted to all nurses of Bichat Stroke Center and to the interventional cardiology personnel for their continuous support in this research. Aimee Grosz, Nassima Schmoll, Evelyne Herinomenjanahary, Hugo Brandao, and Genevieve Pétré were appointed clinical research assistants and were funded by SOS-ATTAQUE CEREBRALE. Sophie Rushton-Smith, PhD, provided editorial assistance in editing the English of the final draft of this manuscript and was funded by SOS-ATTAQUE CEREBRALE association.
Sources of Funding
The study was funded by SOS-ATTAQUE CEREBRALE.
Footnotes
Bo Norrving, MD, was the guest editor for this article.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.000142/-/DC1.
- Received November 13, 2012.
- Accepted February 25, 2013.
- © 2013 American Heart Association, Inc.
References
Jump to
This Issue
Article Tools
Share this Article
- Coronary Artery Disease and Risk of Major Vascular Events After Cerebral Infarction
