Silent Brain Infarction in Patients With Asymptomatic Carotid Artery Atherosclerotic Disease
Background and Purpose—The relationship between carotid atherosclerosis and ipsilateral silent brain infarction (SBI) is unclear. We tested the hypothesis that extracranial internal carotid artery (ICA) stenosis is associated with a greater prevalence of SBI in the cerebral hemisphere ipsilateral to ICA disease compared with the unaffected, contralateral side.
Methods—We identified patients with unilateral extracranial ICA stenosis ≥50% on angiography by standard imaging criteria. We included patients with recent brain magnetic resonance imaging who had no previous history of stroke or transient ischemic attack. Blinded readers ascertained the presence of anterior circulation SBIs. SBI was defined as either a cavitary lacunar infarction in the white or deep gray matter or cortical infarction defined by T2 hyperintense signal in cortical gray matter. The Wilcoxon signed-rank test was used to compare SBI in the cerebral hemispheres and Cohen κ to assess inter-rater reliability of SBI evaluation.
Results—Among 104 patients, we found a higher prevalence of SBIs ipsilateral to ICA disease (33%) compared with the contralateral side (20.8%; P=0.0067). There was no significant difference in the prevalence of lacunar SBIs (including both white and deep gray matter) between hemispheres (P=0.109), but there was a significantly higher prevalence of cortical SBIs occurring downstream from ICA disease (P=0.0045). High inter-rater reliability was observed (κ=0.818).
Conclusions—Patients with asymptomatic ICA disease demonstrate a higher prevalence of SBI downstream from their ICA atherosclerotic disease compared with the contralateral side but only of the cortical and not lacunar SBI subtype.
Silent brain infarctions (SBIs) are small, radiographically detected infarctions in patients without a clinical history of stroke whose presence confers an ≈2-fold increased risk of incident stroke.1 Although several studies have investigated the relationship between the carotid stenosis and the presence of any SBI2–4 in either cerebral hemisphere, the role of carotid atherosclerotic lesions in the pathogenesis of SBI is unclear. In the Cardiovascular Health Study (CHS), no association was found between carotid stenosis and ipsilateral SBI prevalence.2 Given that only a small proportion (≈5%) of the CHS population–based sample included individuals with ≥50% stenosis measured on ultrasound and that older generation magnetic resonance imaging (MRI) equipment was used to detect SBI, it is unclear how these findings are applicable to patients with carotid artery stenosis strictly defined with modern angiographic techniques and in whom brain imaging is performed using higher field magnet strengths. For this reason, we evaluated patients with asymptomatic carotid artery disease for the prevalence of SBI on MRI in the cerebral hemisphere ipsilateral to known carotid artery disease compared with contralateral cerebral hemisphere.
Materials and Methods
We screened consecutive magnetic resonance angiography studies or computed tomographic angiography studies of the neck performed at our institution from January 2008 through December 2014 to identify patients meeting the following inclusion criteria: (1) unilateral extracranial internal carotid artery stenosis ≥50% according to standard measurement criteria5; (2) adequate information in the medical record to exclude a history of previous stroke or transient ischemic attack and to allow for the collection of pre-existing vascular risk factors; and (3) MRI brain performed within 1 month of the index magnetic resonance angiography or computed tomographic angiography demonstrating extracranial internal carotid artery disease.
MRI studies were performed on either 1.5T or 3T Signa (GE Healthcare, Milwaukee, WI) scanners with the use of standard brain coils. Each brain MRI was assessed for SBI by a neuroradiologist blinded to clinical information, including the side of carotid artery disease (Figure). Detailed MRI protocol and SBI imaging assessment technique and representative MRIs of SBI are given in Figures I and II in the online-only Data Supplement.
A neurologist retrospectively collected data on patient demographics, including symptomatic status, and standard vascular risk factors (description of clinical data collection and definitions is given in the online-only Data Supplement). We used McNemar test to compare the prevalence of any SBI and the Wilcoxon signed-rank test to compare the total number of SBIs ipsilateral versus contralateral to extracranial internal carotid artery stenosis.
A total of 104 patients were included for analysis whose demographic characteristics and indications for MRI are available in Table I in the online-only Data Supplement.
The prevalence of any SBI ipsilateral to the side of the carotid atherosclerosis was 33.3% (34/102 patients) compared with a prevalence of 20.8% (21/102 patients) on the contralateral side (P=0.0067; Table). When stratified by SBI subtype, there was no significant difference in the prevalence of lacunar SBI ipsilateral compared with contralateral to the carotid stenosis (P=0.1088), whereas there was a significant difference in the prevalence of cortical SBI ipsilateral versus contralateral to the side of carotid stenosis (P=0.0045).
Similarly, in our analysis of total SBI events, in which any given patient could have multiple SBI events in each hemisphere, there were a total of 58 SBIs in the hemisphere ipsilateral to carotid artery disease compared with 30 events in the contralateral hemisphere (P=0.0002; Table). There was no significant difference in the total number of the lacunar events (including both white and deep gray matter) between ipsilateral and contralateral hemispheres (P=0.10). There was a significant side-to-side difference in the total number of cortical SBI events, with a higher number of cortical infarctions ipsilateral to the stenotic carotid artery when compared with the contralateral side (P<0.001). The results were robust to sensitivity analyses performed comparing patients with stenosis versus occlusion (Results section in the online-only Data Supplement). There was a high interobserver reliability of SBI detection with a κ of 0.82 (95% confidence interval, 0.67–0.96; P<0.001).
In patients with asymptomatic carotid atherosclerotic disease causing ≥50% luminal stenosis, we found an increased prevalence of SBI ipsilateral to the side of the carotid disease compared with the same patient’s contralateral side. This between-hemisphere asymmetry was present for cortical infarctions but not for lacunar infarcts. Our findings suggest that carotid steno-occlusive disease is associated with increased incidence of SBI in the downstream anterior circulation.
Although previous work has shown that carotid artery disease is associated with the presence of SBI in any cerebral hemisphere,3 it is not clear from this previous work whether carotid disease is a marker of generalized, elevated cardiovascular risk or if carotid artery lesions are themselves the cause of a significant proportion of these ipsilateral silent thromboembolic events. A previous study showed no association between carotid stenosis and ipsilateral SBI compared with the contralateral side but was performed >20 years ago with subjects scanned on MRI field strengths as low as 0.35T.2 The improved resolution of modern, higher field-strength MRI equipment used in this study (performed at 1.5T and 3.0T) MRs may have contributed to differences in detectability of SBIs between studies.1 It is also conceivable that improvements in medical therapy since this previous study may have resulted in a decrease in the risk of SBIs attributable to systemic vascular risk factors, especially hypertension, thereby making the risk attributable to specific carotid lesions more apparent in our current study.
Infarctions resulting from carotid disease are likely caused by a combination of hypoperfusion secondary to the flow-limiting stenosis and by artery-to-artery thromboembolic events from unstable plaque elements.6 Our findings demonstrate that there are increased cortical infarctions in patients with asymptomatic carotid disease. Although our study did not include brain perfusion imaging to evaluate the hemodynamic effect of carotid stenosis, our findings do suggest that thromboembolic infarcts may occur more frequently than infarcts from flow-limiting stenosis because cortical gray matter is more commonly affected in infarctions of embolic cause compared with those occurring secondary to hypoperfusion.7 This is in keeping with other prospective studies demonstrating that asymptomatic patients with carotid stenosis who have higher rates of asymptomatic embolization on transcranial Doppler are more likely have future ischemic events.8
Our study has some limitations. First, because of the retrospective nature of the study, it is possible that the heterogeneity in our cohort may have introduced the presence of significant confounding vascular risk factors that may have mediated the relationship between carotid disease and SBI. We minimized this risk of bias by focusing on a within-subject comparison between cerebral hemispheres, thereby reducing the influence of unmeasured confounding factors that might occur in between-patient comparisons. In addition, we screened ≈20 000 vascular imaging examinations to ensure a relatively homogeneous group of patients in terms of rigorously defined stenosis severity (all ≥50%), brain imaging tests performed, and adequate history to accurately determine the presence of previous ischemic symptoms. Second, not all patients were receiving the most robust medical therapy, including a rate of ≈70% for antiplatelet and antilipid medications. Although this is below the universal utilization of optimal medical therapy currently recommended by guidelines, it does likely reflect a real-world practice pattern and may, in part, be because of the fact that ≈45% of our cohort received a first-time diagnosis of carotid artery stenosis at the time of MRI. Third, ≈25% of the patients included in our study received brain MRI for surveillance of known carotid stenosis, which may call into question whether they were truly asymptomatic. However, because the presence of infarcts is associated with a higher risk of future stroke in patients with asymptomatic carotid artery stenosis,8 clinicians at our medical center routinely use brain imaging to guide decisions about treatment intensity even in patients who are completely asymptomatic. In addition, all of the patients included in our study had sufficiently detailed medical records to be deemed to be asymptomatic by a neurologist who reviewed their medical records. Although we believe that our approach reasonably minimized the risk of selection bias, we acknowledge that future prospective, population-based studies are required to confirm our findings.
Although it is known that carotid stenosis represents a marker of generalized, elevated cardiovascular risk, our study demonstrates that this form of large-artery atherosclerosis is associated with an asymmetrical burden of SBI in asymptomatic patients. This finding was particularly evident in cortical infarcts, supporting the hypothesis that asymptomatic carotid disease itself is associated with increased risk of ipsilateral covert brain infarctions from artery-to-artery embolism. Taken together, our findings bring into question current treatment guidelines of patients with asymptomatic carotid stenosis. Although patients may be asymptomatic based on self-reported history, these patients may harbor a greater burden of SBI ipsilateral to their carotid disease that may be silent only because they are located in noneloquent areas. Further studies are now warranted to further assess whether asymptomatic carotid stenosis actually entails subclinical artery-to-artery brain infarction and whether targeted vascular risk factor management can mitigate this risk.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.116.013193/-/DC1.
- Received February 18, 2016.
- Revision received February 18, 2016.
- Accepted March 7, 2016.
- © 2016 American Heart Association, Inc.
- Gupta A,
- Giambrone AE,
- Gialdini G,
- Finn C,
- Delgado D,
- Gutierrez J,
- et al
- Manolio TA,
- Burke GL,
- O’Leary DH,
- Evans G,
- Beauchamp N,
- Knepper L,
- et al
- Kakkos SK,
- Sabetai M,
- Tegos T,
- Stevens J,
- Thomas D,
- Griffin M,
- et al