Prevalence of High-risk Plaques and Risk of Stroke in Patients With Asymptomatic Carotid Stenosis: A Meta-analysis

Question  Is it relevant and feasible to use multimodal neurovascular imaging to perform a risk-oriented selection for revascularization in patients with asymptomatic carotid stenosis?

Findings  In this meta-analysis of 64 studies that enrolled 20 751 participants, high-risk plaques were common in patients with asymptomatic carotid stenosis, and the associated annual incidence of ipsilateral ischemic events (4 events per 100 person-years) was higher than the currently accepted estimates.

Meaning  This study’s findings indicate that extending the assessment of asymptomatic carotid stenosis beyond the grade of stenosis is needed in routine practice to improve risk stratification and optimize therapy; clinical trials using multimodal neurovascular imaging for risk stratification before randomization are warranted to evaluate optimal strategies for stroke prevention in patients with asymptomatic carotid stenosis.

Importance  There is an ongoing debate regarding the management of asymptomatic carotid stenosis. Previous studies have reported imaging features of high-risk plaques that could help to optimize the risk-benefit ratio of revascularization. However, such studies have not provided an accurate estimate of the prevalence of high-risk plaques and the associated annual incidence of ipsilateral ischemic cerebrovascular events to inform the design of clinical trials using a risk-oriented selection of patients before randomization.

Objective  To assess the relevance and feasibility of risk-oriented selection of patients for revascularization.

Data Sources  A systematic search of PubMed and Ovid Embase from database inception to July 31, 2019, was performed.

Study Selection  Prospective observational studies that reported prevalence of high-risk plaques and incidence of ipsilateral ischemic cerebrovascular events were included.

Data Extraction and Synthesis  Aggregated data were pooled using random-effects meta-analysis. Data were analyzed from December 16, 2019, to January 15, 2020.

Main Outcomes and Measures  Prevalence of high-risk plaques and annual incidence of ipsilateral ischemic events.

Results  Overall, 64 studies enrolling 20 751 participants aged 29 to 95 years (mean age range, 55.0-76.5 years; proportion of men, 45%-87%) were included in the meta-analysis. Among all participants, the pooled prevalence of high-risk plaques was 26.5% (95% CI, 22.9%-30.3%). The most prevalent high-risk plaque features were neovascularization (43.4%; 95% CI, 31.4%-55.8%) in 785 participants, echolucency (42.3%; 95% CI, 32.2%-52.8%) in 12 364 participants, and lipid-rich necrotic core (36.3%; 95% CI, 27.7%-45.2%) in 3728 participants. The overall incidence of ipsilateral ischemic cerebrovascular events was 3.2 events per 100 person-years (22 cohorts with 10 381 participants; mean follow-up period, 2.8 years; range, 0.7-6.5 years). The incidence of ipsilateral ischemic cerebrovascular events was higher in patients with high-risk plaques (4.3 events per 100 person-years; 95% CI, 2.5-6.5 events per 100 person-years) than in those without high-risk plaques (1.2 events per 100 person-years; 95% CI, 0.6-1.8 events per 100 person-years), with an odds ratio of 3.0 (95% CI, 2.1-4.3; I² = 48.8%). In studies focusing on severe stenosis (9 cohorts with 2128 participants; mean follow-up period, 2.8 years; range, 1.4-6.5 years), the incidence of ipsilateral ischemic cerebrovascular events was 3.7 events per 100 person-years (95% CI, 1.9-6.0 events per 100 person-years). The incidence of ipsilateral ischemic cerebrovascular events was also higher in patients with high-risk plaques (7.3 events per 100 person-years; 95% CI, 2.0-15.0 events per 100 person-years) than in those without high-risk plaques (1.7 events per 100 person-years; 95% CI, 0.6-3.3 events per 100 person-years), with an odds ratio of 3.2 (95% CI, 1.7-5.9; I² = 39.6%).

Conclusions and Relevance  High-risk plaques are common in patients with asymptomatic carotid stenosis, and the associated risk of an ipsilateral ischemic cerebrovascular event is higher than the currently accepted estimates. Extension of routine assessment of asymptomatic carotid stenosis beyond the grade of stenosis may help improve risk stratification and optimize therapy.

Over the past 3 decades, numerous randomized clinical trials have evaluated the benefits of carotid endarterectomy and stenting for stroke prevention in patients with asymptomatic carotid stenosis.^1,2 In accordance with the results of those clinical trials, current international guidelines recommend considering revascularization in patients with severe asymptomatic carotid stenosis, provided that the periprocedural risk of ipsilateral ischemic cerebrovascular events is less than 3% (class 2a).^3,4 However, critics assert that owing to improvements in best medical therapy, the risk of an ipsilateral ischemic cerebrovascular event among patients with asymptomatic carotid stenosis could now be lower than 1%, suggesting that revascularization may be harmful in some patients.^5,6 As a consequence, the management of asymptomatic carotid stenosis has become a matter of debate, with some clinicians advocating for no revascularization outside of ongoing clinical trials.^7-10 Several studies, including meta-analyses, have reported that extending the assessment of the features of carotid plaque beyond the degree of stenosis can help to select patients with asymptomatic carotid stenosis who have an increased risk of stroke and could therefore benefit from revascularization.^11-24 However, owing to small samples and a low number of events, such studies have not provided data to indicate that the annual incidence of ipsilateral ischemic cerebrovascular events in patients with high-risk plaques exceeds the periprocedural risk of ipsilateral ischemic cerebrovascular events. Moreover, the studies have not provided an accurate estimate of the prevalence of high-risk plaque features to inform the design of clinical trials using a risk-oriented selection of patients before randomization. In this study, we provided estimates of the prevalence of plaques with high-risk features and the annual incidence of ipsilateral ischemic cerebrovascular events in patients with asymptomatic carotid stenosis.

We searched the PubMed and Ovid Embase databases to identify all prospective studies that reported the prevalence of plaques with high-risk features and the associated risk of ipsilateral ischemic cerebrovascular events in participants with asymptomatic carotid stenosis from database inception to July 31, 2019. The search strategy is presented in eTable 1 in the Supplement. For each study, we relied on the definitions used by the authors to quantify the grade of stenosis or identify the specific high-risk features, provided that those definitions were scientifically valid. The definitions are reported in eTable 2 in the Supplement, and details of the study selection criteria are provided in eFigure 1 in the Supplement. This report followed the reporting guidelines for Meta-analysis of Observational Studies in Epidemiology (MOOSE).

Two of us (J.K.-T. and J.J.N.) independently screened the titles and abstracts of the records retrieved from database searches. The full texts of potentially eligible articles were obtained and further assessed for final inclusion in the meta-analysis. Methodological quality and risk of bias were independently assessed by 2 of us (J.K.-T. and J.J.N.) using an adapted version of the Risk of Bias Tool for Prevalence Studies (eTable 3 in the Supplement).²⁵ We aimed to include only studies with a low risk of bias. The interrater agreement for study selection was assessed using a nonweighted Cohen κ.^26-28 Disagreements regarding study inclusion were resolved through discussion and consensus between the 2 assessors.

Aggregated data were extracted using a predesigned standard form (list of variables provided in eTable 2 in the Supplement). The definition of ipsilateral ischemic cerebrovascular events was consistent across studies and corresponded with any acute stroke or transient ischemic attack occurring during follow-up and affecting a brain region supplied by the index carotid as ascertained through imaging (computed tomography, magnetic resonance imaging, or both) and clinical review by at least 1 experienced physician. In our analysis, retinal infarction and amaurosis fugax were considered equivalent to stroke and transient ischemic attack, respectively. For studies reported in 2 or more articles, only the most comprehensive report with the largest sample was considered.

The number of person-years in each cohort was obtained by multiplying the sample size or number of patients with high-risk features by the mean duration of follow-up. The prevalence of high-risk features and the incidence of ipsilateral ischemic cerebrovascular events (in the overall sample and in patients with high-risk plaques) were evaluated by pooling study-specific estimates using random-effects meta-analysis after stabilizing the variance of each study with the Freeman-Tukey double arcsine transformation.^29-32 Publication bias was assessed by inspecting funnel plots and performing the Egger test.³³ In our analysis, the odds ratio (OR) was preferred as the measure of association between plaque features and stroke risk because adjusted hazard ratios (HRs) were reported in only 7 of 22 cohorts eligible for the meta-analysis. Nevertheless, we also performed a meta-analysis of the reported adjusted HRs to verify that the association remained the same (eFigure 2 in the Supplement).

For studies that included a subset of more than 30 patients with symptomatic carotid atherosclerosis, we also computed the prevalence of each high-risk plaque feature. Although it was not the primary objective of this study, this ancillary analysis was used to verify the hypothesis that the prevalence of high-risk plaque features would be higher in patients with symptomatic carotid stenosis compared with patients with asymptomatic carotid stenosis who were examined under the same conditions (period of recruitment, medical management, imaging technique used, and expertise of the investigators).

Subgroup analyses were performed to identify parameters associated with the prevalence of high-risk plaques and the associated risk of ipsilateral ischemic cerebrovascular events. Of special interest was the quantification of the stroke risk associated with the presence of high-risk plaque in studies that enrolled only patients with severe stenosis. To define the subgroups, the levels of the factor variable were used for categorical parameters (eg, decade of publication, type of high-risk feature, or grade of stenosis), while the median across relevant studies was used as the cutoff for continuous parameters (eg, mean age of participants or proportion of participants receiving statin therapy). This methodological approach is standard to increase transparency and avoid arbitrary or biased cutoff selection. Heterogeneity between studies and subgroups was assessed using the χ² test on the Cochran Q statistic and was quantified by the I² index.³⁴ Values of I² were interpreted as follows: less than 25% indicated low heterogeneity, 25% to 75% indicated medium heterogeneity, and more than 75% indicated high heterogeneity.

Univariable random-effects meta-regression models were performed to test the difference of pooled prevalence, incidence, or ORs between subgroups. This testing was conducted by first recoding the subgroups as numerical ordinal variables guided by the observed pattern in prevalence, incidence, or OR. A linear meta-regression of pooled prevalence, incidence, or OR (dependent variable) over the ordinal variable (independent variable) was then performed, and the hypothesis that the slope of the fitted regression line would differ from 0 was tested. The meta-regression analysis accounted for the weight of each study in the initial meta-analysis. All statistical tests were 2-sided and unpaired with a significance threshold of P ≤ .05. Data analyses were performed using Stata software, version 13 (StataCorp LLC). Data were analyzed from December 16, 2019, to January 15, 2020.

Overall, 68 studies enrolling 21 210 participants aged 29 to 95 years (mean age range, 55.0-76.5 years; proportion of men, 45%-87%) were included in the qualitative synthesis (Table 1). The individual characteristics of the included studies are presented in eTable 2 in the Supplement. There was 99.2% agreement between investigators (J.K.-T. and J.J.N.) for study inclusion (κ = 0.86). A subset of 64 studies was included in the meta-analysis (eFigure 1 in the Supplement).

The pooled prevalence of high-risk plaques was 26.5% (95% CI, 22.9%-30.3%) in 20 751 participants with asymptomatic carotid stenosis (Table 2). The most prevalent high-risk plaque features were neovascularization (43.4%; 95% CI, 31.4%-55.8%) in 785 participants, echolucency (42.3%; 95% CI, 32.2%-52.8%) in 12 364 participants, and lipid-rich necrotic core (36.3%; 95% CI, 27.7%-45.2%) in 3728 participants. The prevalence of other specific high-risk plaque features was as follows: for ulceration, 13.1% (95% CI, 3.5%-27.1%) in 2086 participants; for microembolic signals, 14.3% (95% CI, 10.0%-19.2%) in 1648 participants; for intraplaque hemorrhage, 19.1% (95% CI, 13.8%-25.0%) in 3245 participants; for silent brain infarction, 21.9% (95% CI, 15.6%-28.8%) in 2226 participants; for thin or ruptured fibrous cap, 24.1% (95% CI, 12.0%-38.7%) in 670 participants; for impaired cerebrovascular reserve, 29.2% (95% CI, 15.1%-45.7%) in 348 participants; and for American Heart Association (AHA) lesion type 4, 5, or 6 plaque, 30.8% (95% CI, 15.6%-48.4%) in 168 participants (Table 2; eFigures 3-12 in the Supplement). The visual inspection of the funnel plot suggested that no publication bias existed, as indicated by the Egger test (Egger intercept, −0.25; Egger P = .58) (Table 2; eFigure 13 in the Supplement).

The prevalence of high-risk plaques was not associated with the demographic characteristics (mean age and proportion of men) of the study population, the grade of stenosis, or the circumstances of enrollment (setting and planned endarterectomy) (eTable 4 in the Supplement). The prevalence of high-risk plaques was significantly higher in studies in which 78% of participants or more were receiving antiplatelet therapy (34.6%) compared with studies in which less than 78% of participants were receiving antiplatelet therapy (17.8%; P = .002) (eTable 4 in the Supplement). A higher (but statistically nonsignificant) prevalence of high-risk plaques was observed in studies published after 2000 and in studies with a higher proportion of patients with hypertension and diabetes (eTable 4 in the Supplement).

A total of 18 cross-sectional studies and 2 cohort studies also provided relevant data on 1652 patients with symptomatic carotid stenosis. In these 20 studies, the pooled prevalence of high-risk plaques was 43.3% (95% CI, 33.6%-53.2%) in patients with symptomatic carotid stenosis vs 19.9% (95% CI, 14.5%-25.8%) in patients with asymptomatic carotid stenosis (eTable 5 in the Supplement). The OR of a high-risk plaque in a patient with symptomatic carotid stenosis vs a patient with asymptomatic carotid stenosis was 3.4 (95% CI, 2.5-4.6).

A total of 22 cohort studies that enrolled 10 381 participants with asymptomatic carotid stenosis provided relevant data for the meta-analysis of the risk of ipsilateral ischemic cerebrovascular events associated with high-risk plaque features. The mean duration of follow-up was 2.8 years (range, 0.7 to 6.5 years) (Table 1). The incidence of ipsilateral ischemic cerebrovascular events was 3.2 events per 100 person-years (95% CI, 2.2-4.3 events per 100 person-years) in the overall population of patients with asymptomatic carotid stenosis (eFigure 14 and eFigure 15 in the Supplement). There was evidence of publication bias (Egger intercept, 0.64; Egger P < .001). Studies with a higher number of person-years reported lower incidence rates (eFigure 16 in the Supplement).

The incidence of ipsilateral ischemic events was higher in patients with high-risk features (4.3 events per 100 person-years; 95% CI, 2.5-6.5 events per 100 person-years) (Figure 1^35-51; eFigure 17 in the Supplement) than in those without high-risk features (1.2 events per 100 person-years; 95% CI, 0.6-1.8 events per 100 person-years) (eFigure 18 in the Supplement), with a corresponding OR of 3.0 (95% CI, 2.1-4.3; I² = 48.8%) (Figure 2^35-51). This association between the presence of high-risk plaque and the risk of ipsilateral ischemic cerebrovascular events was also observed in pooling data from the 7 studies that reported adjusted HRs (pooled HR, 3.0; 95% CI, 1.8-4.2; I² = 0%) (eFigure 2 in the Supplement). This increased risk was also observed specifically for ischemic stroke (OR, 2.0; 95% CI, 1.5-2.7; I² = 0%) and transient ischemic attack (OR, 2.4; 95% CI, 1.2-4.9; I² = 13.0%) (Table 3).

A higher (but statistically nonsignificant) risk of ipsilateral ischemic events was found in studies with a greater proportion of participants who smoked, had hypertension or diabetes, or were receiving statin or antiplatelet therapy (Table 3). The presence of AHA lesion type 4, 5, or 6 plaque was the greatest indicator of the risk of ipsilateral ischemic events (OR, 28.7; 95% CI, 1.6-513.3), followed by the presence of microembolic signals (OR, 5.6; 95% CI, 2.0-15.3; I² = 68.0%) (Table 3). The incidence of ipsilateral ischemic events in patients with asymptomatic carotid stenosis with high-risk features is presented for each type of high-risk feature in eFigure 19 in the Supplement and for each decade since 1990 in eFigure 17 in the Supplement.

In the subgroup of studies focusing on participants with severe stenosis only (9 cohorts with 2128 participants; mean follow-up period, 2.8 years; range, 1.4-6.5 years), the incidence of ipsilateral ischemic cerebrovascular events was 3.7 events per 100 person-years (95% CI, 1.9-6.0 events per 100 person-years). The incidence of ipsilateral ischemic cerebrovascular events was also higher in patients with high-risk features (7.3 events per 100 person-years; 95% CI, 2.0-15.0 events per 100 person-years) (Figure 1^35-51) than in those without high-risk features (1.7 events per 100 person-years; 95% CI, 0.6-3.3 events per 100 person-years) (eFigure 18 in the Supplement), with an OR of 3.2 (95% CI, 1.7-5.9; I² = 39.6%) (Table 3).

The incidence of ipsilateral ischemic events in patients with high-risk plaques was not modified by participant characteristics, including the mean age and the proportion of men, the frequency of various cardiovascular risk factors, and the use of statin or antiplatelet therapy (eTable 6 in the Supplement). The incidence of ipsilateral ischemic stroke and transient ischemic attack in the overall population of patients with asymptomatic carotid stenosis and in those with high-risk plaques is provided in eFigures 20-23 in the Supplement.

To our knowledge, this is the first summary of data on the prevalence of high-risk plaque and the associated risk of stroke in the specific population of patients with asymptomatic carotid stenosis. The prevalence of high-risk features was not associated with the grade of stenosis, suggesting that high-risk features reflect the underlying pathomechanism of plaque formation, remodeling, and destabilization, which is the same across all grades of stenosis.^52,53 This finding also indicates that, although revascularization is not beneficial in all patients with mild stenosis,⁵⁴ there may be a subset of patients in this group who could benefit from specific interventions in addition to best medical therapy.

The prevalence data reported in this study are important for sample size calculations in future interventional clinical trials in which patients with asymptomatic carotid stenosis with a higher risk of stroke are randomized to receive therapy. The low prevalence of well-validated high-risk features, such as microembolic signals³⁵ and intraplaque hemorrhage,²¹ suggests that a combination of high-risk features in a multimodal imaging approach might be necessary to optimize a screening strategy relying on imaging biomarkers for risk stratification. This assumption is consistent with the higher prevalence of high-risk plaques found in studies using the AHA classification system. By pooling AHA lesion type 4, 5, and 6 plaques, those studies combined features that were equivalent to lipid-rich necrotic core, thin or ruptured fibrous cap, ulceration, intraplaque hemorrhage, and mural thrombus.^52,53 The combination of imaging modalities is already a part of routine clinical practice in many medical centers, but the combination approach adds cost and time. Therefore, a simple alternative may be desirable, such as the use of a blood-based biomarker that aligns with vascular imaging findings and is associated with the risk of stroke.⁵⁵ Such a biomarker still needs to be developed and may include a panel of markers reflecting the risk of plaque rupture and thromboembolism.

A key finding of this meta-analysis is that the risk of ipsilateral ischemic events among the overall population of patients with asymptomatic carotid stenosis (3.2%) and among the subsets of patients with high-risk plaque features (4.3%) and without high-risk plaque features (1.2%) is greater than the commonly accepted rate of 1%.^5,6 The latter rate was computed using the 10-year follow-up data from the Asymptomatic Carotid Surgery Trial 1 (ACST-1; ISRCTN26156392).⁵⁶ In this clinical trial, the decrease in stroke incidence among patients randomized to deferral of carotid intervention has been associated solely with improvements in best medical therapy. However, it is also probable that the composition of the clinical trial population became progressively skewed toward a predominance of patients without high-risk plaque features as those with high-risk plaques experienced a stroke earlier or underwent surgery. The hypothesis of the presence of bias underlying the report of low incidence rates in long-lasting closed cohorts is consistent with the findings of our meta-analysis of incidence data.

Current data indicate that patients with asymptomatic carotid stenosis have a lower periprocedural risk of stroke and better outcomes after revascularization compared with patients with symptomatic carotid stenosis.^57,58 Thus, in future clinical trials of carotid revascularization, it may be relevant to perform additional subgroup analyses based on the presence of high-risk features at baseline (eg, the Carotid Revascularization and Medical Management for Asymptomatic Carotid Stenosis 2 [CREST-2; NCT02089217]; the European Carotid Surgery Trial 2 [ECST-2; ISRCTN97744893]; the Asymptomatic Carotid Surgery Trial 2 [ACST-2; NCT00883402]; and the Endarterectomy Combined With Optimal Medical Therapy vs Optimal Medical Therapy Alone in Patients With Asymptomatic Severe Atherosclerotic Carotid Artery Stenosis at Higher-Than-Average Risk of Ipsilateral Stroke [ACTRIS; NCT02841098] clinical trials).^22,59

In the subgroup analyses, we observed a higher (but statistically nonsignificant) prevalence of high-risk features and a higher incidence of ipsilateral ischemic cerebrovascular events in studies enrolling a greater proportion of patients with hypertension or diabetes. The risk of an ipsilateral ischemic cerebrovascular event was also higher in studies with a greater proportion of smokers. These findings highlight the role of cardiovascular risk factors in atherosclerotic plaque progression and destabilization. They also emphasize the importance of vascular risk factor control in the management of carotid stenosis.^1,3,4

The decrease in the incidence of ipsilateral ischemic events after 2000, despite an increase in the prevalence of high-risk features, may also be associated with improvements in medical therapy over time. A complementary explanation could be that more revascularization procedures were performed after the release of results from the Asymptomatic Carotid Atherosclerosis Study,⁶⁰ contributing to further decreases in the incidence of ipsilateral ischemic cerebrovascular events. With the available data, it was not possible to explore whether the change in the prevalence of high-risk plaque features before and after 2000 was a true increase owing to various changes in lifestyle and environmental factors or whether the increase was associated with increases in the availability of vascular imaging and improvements in the reporting of high-risk plaque features.

An unexpected finding was the higher risk of stroke in studies including a greater proportion of patients who were receiving statin and antiplatelet drugs. This higher risk might be associated with the fact that patients with high-risk plaques were more likely to be offered these treatments, and the authors recorded the prescriptions for statin and antiplatelet drugs before and after the identification of the high-risk features without distinction. This hypothesis is consistent with our finding that the prevalence of high-risk plaques was significantly higher in studies that included a greater proportion of patients receiving antiplatelet drugs. A complementary hypothesis could be that, although useful, statin and antiplatelet therapies remain insufficient to suppress the higher risk of stroke associated with high-risk plaques.⁶¹ Testing such a hypothesis would require the collection of individual patient data regarding the nature and doses of statin and antiplatelet drugs, the control status of cardiovascular risk factors, the adherence to treatment, and the resistance to antiplatelet drugs.^1,2,62-65

This study has several limitations. First, despite the study’s large sample and rigorous methods, there was substantial heterogeneity in the prevalence of high-risk features among studies included in the meta-analysis, which is explained by differences across studies regarding the definition criteria and the imaging modalities used. Moreover, the prevalence of high-risk plaques was obtained by pooling the specific prevalence of each high-risk feature based on the pragmatic assumption that high-risk plaques were equal, regardless of how they were identified. Heterogeneity in stroke risk between high-risk features is likely. However, high-risk features do have an association with each other, as they represent aspects of the same atherosclerotic disease. Combining different views provides a better assessment of the underlying biological factors. Consensus imaging recommendations are needed to help decrease the heterogeneity of carotid imaging data across studies and to facilitate international clinical research collaborations.

Second, it is possible that some of the ipsilateral ischemic cerebrovascular events reported were not owing to atheroembolism from the index asymptomatic carotid stenosis. However, the proportion of ipsilateral ischemic cerebrovascular events owing to other factors is likely small because most studies excluded patients with atrial fibrillation, and ipsilateral ischemic cerebrovascular events deemed to be associated with other factors were also excluded from the analyses.^36,37,66-68

This study provides data to indicate that, in patients with asymptomatic carotid stenosis, high-risk plaques are common, and the associated risk of ipsilateral ischemic cerebrovascular events is higher than the currently accepted estimates. Therefore, a routine assessment of asymptomatic carotid stenosis that extends beyond the grade of stenosis could help to identify patients at a higher risk of stroke who require an intensification of medical therapy for the control of cardiovascular risk factors. Additional clinical trials using multimodal neurovascular imaging for risk stratification before randomization are warranted to evaluate the optimal strategy for stroke prevention in patients with asymptomatic carotid stenosis.

Accepted for Publication: May 15, 2020.

Corresponding Author: Joseph Kamtchum-Tatuene, MD, Neuroscience and Mental Health Institute, Faculty of Medicine and Dentistry, University of Alberta, 114 St and 87 Ave Northwest, 4-065 Katz Group Bldg, Edmonton, AB T6G 2E1, Canada (kamtchum@ualberta.ca).

Published Online: August 3, 2020. doi:10.1001/jamaneurol.2020.2658

Author Contributions: Dr Kamtchum-Tatuene had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Kamtchum-Tatuene, Saqqur, Jickling.

Acquisition, analysis, or interpretation of data: Kamtchum-Tatuene, Noubiap, Wilman, Shuaib.

Drafting of the manuscript: Kamtchum-Tatuene, Saqqur, Jickling.

Critical revision of the manuscript for important intellectual content: Kamtchum-Tatuene, Noubiap, Wilman, Shuaib, Jickling.

Statistical analysis: Kamtchum-Tatuene.

Obtained funding: Jickling.

Administrative, technical, or material support: Kamtchum-Tatuene.

Supervision: Saqqur, Shuaib, Jickling.

Conflict of Interest Disclosures: Dr Kamtchum-Tatuene reported receiving grants from the Banque of Montreal Financial Group and the Graduate Excellence Scholarships of the government of Alberta, Canada, during the conduct of the study. Dr Noubiap reported receiving grants from Adelaide Scholarship International of the University of Adelaide, Australia, outside the submitted work. Dr Jickling reported receiving grants from the Canada Foundation for Innovation, the Canadian Institutes of Health Research, the Heart and Stroke Foundation, the National Institutes of Health, and the University Hospital Foundation outside the submitted work. No other disclosures were reported.

Additional Contributions: The University of Alberta Library provided access to resources, and the Interlibrary Loan and Document Delivery Unit supplied the full text of all articles not included in the university collection or not available through open access repositories.