Racial Differences in Prevalence and Risk for Intracranial Atherosclerosis in a US Community-Based Population

Key Points

Question  Are there racial differences in the prevalence of and risk factors for intracranial atherosclerotic disease?

Findings  In this Atherosclerosis Risk in Communities cohort study of 1752 elderly adults, black men had the highest prevalence of intracranial atherosclerotic disease and the highest frequency of multiple plaques. Midlife smoking and diabetes were associated with late-life intracranial atherosclerotic disease in black individuals only, whereas midlife hypertension and hyperlipidemia were associated with late-life intracranial atherosclerotic disease in both races.

Meaning  Racial differences in intracranial atherosclerotic disease prevalence may help to explain stroke rates in the United States, and differences in intracranial atherosclerotic disease risk factors could offer insight into preventive risk-factor management strategies.

Importance  Intracranial atherosclerotic disease (ICAD) is an important cause of stroke; however, little is known about racial differences in ICAD prevalence and its risk factors.

Objective  To determine racial differences in ICAD prevalence and the risk factors (both midlife and concurrent) associated with its development in a large, US community-based cohort.

Design, Setting, and Participants  Analysis of 1752 black and white participants recruited from the Atherosclerosis Risk in Communities (ARIC) cohort study who underwent 3-dimensional intracranial vessel wall magnetic resonance imaging from October 18, 2011 to December 30, 2013; data analysis was performed from October 18, 2011 to May 13, 2015.

Exposures  Midlife and concurrent cardiovascular risk factors.

Main Outcomes and Measures  Intracranial plaque presence, size (maximum normalized wall index) and number were assessed by vessel wall magnetic resonance imaging. Midlife and concurrent vascular risk factor associations were determined by Poisson regression (plaque presence), negative binominal regression (plaque number), and linear regression (plaque size), and compared between races.

Results  Of the 1752 study participants (mean [SD] age, 77.6 [5.3] years; range, 67-90 years), 1023 (58.4%) were women and 518 (29.6%) were black. Black men had the highest prevalence (50.9% vs 35.9% for black women, 35.5% for white men, and 30.2% for white women; P < .001) and the highest frequency (22.4% vs 12.1% for black women, 10.7% for white men, and 8.7% for white women; P < .01) of multiple plaques. Prevalence increased with age, reaching 50% before ages 68, 84, and 88 years in black men, white men, and white women, respectively (ICAD prevalence remained <50% in black women). Midlife hypertension and hyperlipidemia were associated with 29% (prevalence ratio [PR], 1.29; 95% CI, 1.08-1.55) and 18% (PR, 1.18; 95% CI, 0.98-1.42), respectively, increased ICAD prevalence with no significant differences between races. Midlife hypertension was also associated with larger plaques (1.11 higher mean maximum normalized wall index; 95% CI, 0.21-2.01). Midlife smoking and diabetes were associated with increased risk of ICAD in black individuals (102% [PR, 2.02; 95% CI, 1.12-3.63] and 57% [PR, 1.57; 95% CI, 1.13- 2.19], respectively), but not in white participants (P < .05 interaction by race).

Conclusions and Relevance  The prevalence of ICAD was highest in black men. Midlife smoking and diabetes were strongly associated with late-life ICAD in blacks only, whereas midlife hypertension and hyperlipidemia were associated with late-life ICAD in both races. These associations may help to explain racial differences in US stroke rates and offer insight into preventive risk-factor management strategies.

Strokes are reported more frequently in black than white individuals.¹ Intracranial atherosclerotic disease (ICAD) is a major contributor to ischemic stroke,² but it remains to be established if ICAD is also more prevalent in black individuals in the general population. Our current understanding of ICAD distribution and its risk factors is based on stenosis measurements primarily from symptomatic cohorts^3-6; however, as many as 75% of patients who develop strokes have not had prior strokes.⁷ A large study of intracranial calcifications in internal carotid arteries in white European descendants by computed tomography reported more than 80% prevalence⁸; however, this likely overestimates ICAD because Mönckeberg arteriosclerosis accounts for calcifications here.⁹ Furthermore, previous studies^10-12 identified ICAD as lumen narrowing of intracranial arteries. Lumen narrowing, however, is a poor indicator of plaque burden when vessels accommodate plaque formation by compensatory dilatation (remodeling).¹³ The recent development of 3-dimensional (3-D) vessel wall magnetic resonance imaging (MRI) has made possible the identification and characterization of ICAD, providing highly reliable intracranial vessel wall measurements.^14,15 Previous studies also involved older adults in whom traditional risk factors (eg, blood pressure, cholesterol level) may have been altered by medical treatment or lifestyle modifications, but lacked information on midlife risk factor levels that may be more relevant to atherosclerotic burden later in life.^8,11 Identifying these midlife risk factors and understanding their implications based on race and sex is an important step in preventive treatment strategies, recognizing that the effect of these strategies may not be perceived for decades.

The objective of this study was to determine racial differences in ICAD prevalence and the risk factors (both midlife and concurrent) associated with its development in a large community-based study, the Atherosclerosis Risk in Communities Study (ARIC).

The ARIC study is a community-based, prospective investigation of 15 792 black and white participants recruited in middle age (45-64 years) from 4 US communities (Forsyth County, North Carolina; Jackson, Mississippi; suburban Minneapolis, Minnesota; and Washington County, Maryland).¹⁶ The baseline examination (visit 1) took place from 1987 to 1989, with 4 subsequent visits.

Of the 6538 participants who attended ARIC visit 5 (from October 18, 2011, to December 30, 2013), 1980 were selected for brain MRI using a probability sampling mechanism with oversampling of participants with cognitive impairment and those who had a prior brain MRI. A total of 1752 black and white participants (mean [SD] age, 77.6 [5.3] years; range, 67-90 years; 1023 [58.4%] women; 518 [20.6%] black) had complete vascular MRI examinations with adequate or excellent image quality and MRI protocol adherence and were included in this study. The study was approved by the institutional review boards at each site (protocol name: The Intracranial Atherosclerotic Disease and Cognitive Impairment Study). All participants provided written informed consent and received financial compensation.

The MRI protocol has been previously described.¹⁴ Briefly, all MRI scans were performed on 3.0T Siemens scanners. High-resolution vascular sequences were acquired at the end of a standardized brain MRI protocol and consisted of a 3-D time-of-flight magnetic resonance angiography (MRA) and 3-D high–isotropic resolution black blood MRI with acquired resolutions of 0.50 × 0.50 × 0.55 mm³ and 0.50 × 0.50 × 0.50 mm³, respectively. The MRI images were analyzed by 7 certified readers at the MRI reading center without knowledge of participant characteristics.

An atherosclerotic plaque was defined as wall thickening on reconstructed black blood MRI images with or without luminal stenosis on MRA.¹⁵ Vascular territories were categorized as right and left intracranial internal carotid artery (cavernous and supraclinoid segments), right and left middle cerebral artery (M1-M3 segments), right and left posterior cerebral artery (P1-P3 segments), right and left anterior cerebral artery (A1-A3 segments), basilar artery, and right and left vertebral artery (V4 segment). All plaques were recorded for each vascular territory. The degree of stenosis was measured according to the criteria established in the Warfarin-Aspirin Symptomatic Intracranial Disease trial,¹⁷ and categorized as no detectable stenosis, less than 50%, 51% to 70%, 71% to 99%, and occlusion. The area and normalized wall index (NWI) (wall area / outer wall area) were measured for the most stenotic plaque identified in each territory. The NWI uses wall area to standardize measures of plaque size and enables comparison of plaques in vessel segments of varying size. Reliability estimates for plaque detection and measurements were good to excellent as previously reported.¹⁵

Cardiovascular risk factors measured at visits 1 and 5 were used in this study and included smoking history (current, past, or never), systolic and diastolic blood pressures, antihypertensive medication use, plasma low-density lipoprotein cholesterol level, plasma high-density lipoprotein cholesterol level, plasma triglyceride level, cholesterol-lowering medication use, fasting and nonfasting glucose levels, antidiabetic medication use, history of coronary heart disease, and history of stroke. We defined hypertension as systolic blood pressure of 140 mm Hg or higher, diastolic blood pressure of 90 mm Hg or higher, or use of antihypertensive medications; diabetes as fasting glucose level of 126 mg/dL or higher, nonfasting glucose level of 200 mg/dL or higher (to convert to millimoles per liter, multiply by 0.0555), or use of antidiabetic medications; and hyperlipidemia as total cholesterol level higher than 240 mg/dL, high-density lipoprotein cholesterol level lower than 40 mg/dL, or low-density lipoprotein cholesterol level higher than 160 mg/dL (to convert all cholesterol levels to millimoles per liter, multiply by 0.0259), triglyceride level higher than 200 mg/dL (to convert to millimoles per liter, multiply by 0.0113), or use of cholesterol-lowering medications.

Data analysis was performed from October 18, 2011, to May 13, 2015. All analyses were performed using Stata, version 12.1 (Stata Inc). Stata svy commands with sampling weights were used to account for oversampling of participants with cognitive impairment and provide estimates referable to the overall ARIC population from which participants were sampled (ARIC visit 5, n = 6538). In addition, we used inverse probability of attrition weighting to adjust for potential selection bias due to the underrepresentation of individuals who died or were lost to follow-up between ARIC visits 1 and 5¹⁸ (eMethods in the Supplement). The prevalence of ICAD was estimated in the overall ARIC population and stratified by sex and race. Demographic and risk factor variables were compared between races using unpaired, 2-tailed t tests for continuous variables and χ² tests for categorical variables.

ICAD prevalence by age was estimated from Poisson regression models using restricted cubic splines with knots at the 10th, 50th, and 90th percentiles of the age distribution and stratified by sex and race. Prevalence ratios (PRs) for plaque presence for visit 1 and visit 5 risk factors were calculated separately using multivariable Poisson regression. A full model was then established including all risk factors from both visits. We also used Poisson regression models to estimate marginally adjusted prevalence of ICAD as a function of risk factor levels after controlling for use of antihypertensive, cholesterol-lowering, and antidiabetic medications. All models were adjusted for age, sex, and race, and repeated with race stratification. Negative binomial regression was used to estimate the ratio of the number of plaques per participant by risk factor levels (combined risk factors from both visits). In addition, multivariable linear regression was used to estimate mean differences in maximum NWI within a participant (a marker of plaque burden) as a function of risk factor levels (combined risk factors from both visits). Racial differences in associations between risk factors and ICAD were estimated using interaction terms between race and each risk factor. P < .05 was considered significant.

Most black participants were younger than white participants and were more likely to have hypertension and diabetes but less likely to have hyperlipidemia at both midlife and late-life visits (Table 1). A higher prevalence of cardiovascular disease risk factors was observed in late life compared with midlife in both the white and black groups.

Of the 1752 participants, 637 (36.4%) had ICAD with at least 1 intracranial plaque, and 69 of 637 individuals (10.8%) had intracranial plaques without detectable stenosis. The distribution of plaques across intracranial vascular territories is reported in eTable 1 in the Supplement. ICAD prevalence in the ARIC cohort at visit 5 was 34.4% and was highest for black men (50.9% for black men vs 35.9% for black women, 35.5% for white men, and 30.2% for white women, P < .001). Black men also had the highest frequency of multiple plaques (22.4% of black men had >3 plaques vs 12.1% for black women, 10.7% for white men, and 8.7% for white women; P < .01) (Figure 1). The prevalence of ICAD increased with age for each group, reaching 50% in black men before age 68 years, white men by 84 years, and white women by 88 years (Figure 2; eFigure 1 in the Supplement). The percentage of participants older than 77 years was 38%, 47%, 53%, and 56% for black women, white women, black men, and white men, respectively. Black women did not reach 50% even by 90 years. A rise in prevalence began at approximately 77 years for all groups except black women, who showed no significant change above this age (Figure 2; eFigure 1 in the Supplement). After excluding participants with a history of stroke (n = 60), the weighted ICAD prevalence was 33.3% (49.6% for black men, 34.8% for black women, 34.6% for white men, and 29.3% for white women, P < .001).

Midlife (visit 1) hypertension and hyperlipidemia were independently associated with plaque presence, after controlling for age, sex, race, and other midlife cardiovascular risk factors (Table 2). No associations were identified for visit 5 risk factors using a comparable model. In a combined model including all risk factors from both visits, midlife hypertension and hyperlipidemia were associated with a 29% (PR, 1.29; 95% CI, 1.08-1.55) and 18% (PR, 1.18; 95% CI, 0.98-1.42), respectively, increased prevalence of having ICAD by visit 5 irrespective of hypertensive or hyperlipidemic status at visit 5 (Table 3), with no evidence for differences in effect between black and white participants (Table 3). Midlife smoking and diabetes were associated with 102% (PR, 2.02; 95% CI, 1.12-3.63) and 57% (PR, 1.57; 95% CI, 1.13-2.19), respectively, increased prevalence of ICAD by visit 5 in black participants irrespective of smoking or diabetic status at visit 5, but associations were not present for white participants (Table 3).

When risk factors were analyzed as continuous variables, the prevalence of ICAD was significantly increased with concurrent (visit 5) measures of systolic blood pressure as well as fasting glucose and total cholesterol levels, and decreased with high-density lipoprotein cholesterol levels after adjusting for medication use (eFigure 2 in the Supplement). Associations were strongest in black men.

Midlife hypertension and hyperlipidemia were associated with a 39% (ratio 1.39; 95% CI, 1.07-1.80) and a 33% (ratio, 1.33; 95% CI, 1.04-1.71), respectively, increase in plaque number by visit 5 irrespective of risk status at visit 5 (eTable 2 in the Supplement). Midlife hypertension was also associated with a higher mean maximum NWI (mean difference comparing hypertensive with nonhypertensive status, 1.11; 95% CI, 0.21-2.01) irrespective of hypertensive status at visit 5 (eTable 3 in the Supplement). Midlife smoking was associated with a 57% (ratio, 1.57; 95% CI, 1.06-2.33) increase in plaque number in black participants irrespective of visit 5 smoking status, but no association was seen in white participants. Midlife smoking, diabetes, and hyperlipidemia were not associated with plaque size by visit 5 in black or white individuals. We repeated the regression models after excluding participants with a history of stroke and noted no change in the above associations.

To our knowledge, this is the first study to report racial differences in prevalence and risk factors for ICAD in a large, US community-based population. Black participants had a higher ICAD prevalence than white participants, and we also observed racial differences in risk factors for ICAD. Midlife smoking and diabetes were strongly associated with late-life ICAD in black but not white individuals, whereas midlife hypertension and hyperlipidemia were associated with late-life ICAD in both races. These associations were observed irrespective of late-life risk status and highlight the importance of early risk-factor management.

To our knowledge, racial differences in ICAD prevalence have not been previously reported in general population studies. An autopsy study of cerebral artery specimens from 190 decedents (age, 65-69 years) in New Orleans, Louisiana,¹⁹ identified ICAD in 7.6% of white men, 8.5% of white women, 32.6% of black men, and 20.0% of black women. However, prevalence estimates in these studies may be biased because of a greater risk of death in patients with a higher ICAD prevalence or more advanced lesions. There have been reports of intracranial stenosis prevalence in asymptomatic populations using transcranial Doppler ultrasonography^10,11,20 or MRA,¹² but transcranial Doppler ultrasonography and MRA rely on luminal narrowing for detection of ICAD and can miss nonstenotic plaques in vessels with outward remodeling. Our estimated ICAD prevalence was higher than that reported for transcranial Doppler ultrasonography (6.9%¹⁰ to 8.3%¹¹) or MRA (31%¹²). We found that as many as 10.8% of our participants had ICAD without detectable stenosis based on black blood MRI and these cases would have eluded MRA detection, limiting analyses based on stenosis alone.¹²

Our findings may have clinical implications, as they might help to explain the increased stroke risk secondary to ICAD in black men in the United States^21-23 as a consequence of their higher prevalence of plaques or their higher frequency of multiple plaques. The rise in prevalence in white individuals beginning at approximately 77 years (Figure 2) may also explain the crossover in stroke mortality rates reported at 85 years, when mortality rates in white individuals begin to exceed rates in black individuals.²¹

Racial differences were also observed with risk factors for ICAD, most notably, the associations between ICAD with midlife smoking and diabetes seen in black but not white participants. The mechanisms for racial differences in ICAD prevalence and associated risk factors could be complex and might include genetic and ecosocial factors.^24,25 Furthermore, at midlife, black participants had a higher frequency of current smokers and a lower frequency of former smokers, which might have influenced associations with late-life ICAD in the black but not white cohort. The differences likely reflect a combination of the relatively small number of white individuals who smoked in midlife plus selection factors associated with continuing or starting smoking at an advanced age. Established risk factors for stroke in black persons (ie, hypertension, smoking, and diabetes) coincide with our observed risk factors for ICAD in black participants.^21,23,26,27 When the races were combined, the overall midlife risk factors were more strongly associated with ICAD than late-life measures, as reported in carotid and coronary studies.^28,29 In particular, midlife hypertension and hyperlipidemia demonstrated strong associations with ICAD presence and number, with hypertension contributing further to plaque size. Stronger associations with midlife compared with late-life risk factors are expected since earlier measures may reflect a long-term cumulative effect with less influence by treatment.²⁸

The weak associations between ICAD and late-life risk factor measurements could partly stem from reverse causation³⁰ due to changes in lifestyle (eg, smoking cessation), the high prevalence of some risk factors in late life (eg, hypertension), or the definitions used to establish the presence of risk factors, which include treatment as part of the definition. When late-life risk factors were analyzed as continuous variables, increasing systolic blood pressure as well as fasting glucose and total cholesterol levels, and decreasing high-density lipoprotein cholesterol levels were associated with increased ICAD presence, suggesting that ICAD depends on risk factor degree rather than exceeding a threshold.

The major strengths of our study include the use of advanced 3-D black blood MRI to reliably identify and quantify ICAD in a large, nationally representative sample, with access to a wide range of vascular risk factors measured over 20 years. This new MRI technique allowed us to identify ICAD even in the absence of luminal narrowing. The large black population enabled discriminating racial differences in ICAD prevalence and its associated risk factors.

Some limitations, however, need to be considered in the interpretation of our findings. Our precision for estimating ICAD prevalence for black women older than 77 years was lower than for the other groups (Figure 2) so we primarily focused on black and white racial differences. Although we used inverse probability of attrition weighting to compensate for the effect of attrition from death or dropout since ARIC visit 1, attrition may have been disproportionate between the races since more black participants died before visit 5.¹⁸ We cannot discard residual biases, particularly for race comparisons at advanced age. Treatment decisions may also have affected risk factor associations. For example, participants with better risk factor management (eg, cholesterol lowering or blood pressure control) since midlife may have fewer or smaller ICAD lesions in late life. These patterns could dilute associations between visit 5 risk factors and ICAD. In addition, future studies should include other racial groups (eg, Hispanic or Asian) to better reflect an urban population. Finally, the lack of ICAD measures at the baseline visit prohibits an investigation of ICAD progression, which may have different risk factors. Longitudinal studies of ICAD would be needed to estimate these associations.

In this community-based population study of black and white elderly adults, the prevalence of ICAD was higher in the blacks than the white participants, with the highest prevalence and highest frequency of multiple lesions seen in black men. Midlife smoking and diabetes were associated with late-life ICAD presence in black but not white individuals, whereas both races demonstrated associations between midlife hypertension and hyperlipidemia with ICAD presence and intracranial plaque number, and hypertension further contributed to intracranial plaque size. These racial differences may contribute to racial differences in stroke rates in the United States, providing insight into stroke risk and potential strategies for risk reduction.

Accepted for Publication: September 13, 2017.

Corresponding Author: Bruce A. Wasserman, MD, The Russell H. Morgan Department of Radiology & Radiological Science, The Johns Hopkins University School of Medicine, 367 East Park Bldg, 600 N Wolfe St, Baltimore, MD 21287 (bwasser@jhmi.edu).

Published Online: November 1, 2017. doi:10.1001/jamacardio.2017.4041

Author Contributions: Drs Qiao and Suri contributed equally to the study. Drs Qiao and Wasserman had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Qiao, Suri, Guallar, Wasserman.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Qiao.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Qiao, Zhang, Liu, Guallar, Wasserman.

Obtained funding: Suri, Wasserman.

Administrative, technical, or material support: Guallar, Wasserman.

Study supervision: Qiao, Alonso, Wasserman.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Drs Qiao and Wasserman have a patent (13/922,111) for the magnetic resonance imaging (MRI) technique used in this study. Dr Wasserman reports grants from National Heart, Lung and Blood Institute (NHLBI), National Institutes of Health (NIH) during the conduct of the study; in addition, Dr Wasserman has a patent null issued. Dr Gottesman reports other grants from the American Academy of Neurology outside the submitted work. Dr Alonso reports grants from the NIH during the conduct of the study. No other disclosures were reported.

Funding/Support: This study was supported by NIH grants RO1HL105930, RO1HL105626, and K99/R00HL106232. The Atherosclerosis Risk in Communities (ARIC) study is part of a collaborative study supported by contracts HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C from the NHLBI. Neurocognitive data are collected by grants U01 HL096812, HL096814, HL096899, HL096902, HL096917 from the NHLBI and the National Institute of Neurological Disorders and Stroke, and with previous brain MRI examinations funded by grant R01-HL70825 from the NHLBI.

Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank the staff and participants in the Atherosclerosis Risk in Communities study for their important contributions.