Apolipoprotein E Genotype and Incident Ischemic Stroke
The Atherosclerosis Risk in Communities Study
Background and Purpose— A relationship between the apolipoprotein E (apoE) genotype and ischemic stroke has been inconsistently reported. We explored this relation in the Atherosclerosis Risk in Communities Study (ARIC).
Methods— The ARIC cohort involves 15 792 men and women, aged 45 to 64 years at baseline and sampled from 4 U.S. communities. Between 1987 and 2001, 498 incident ischemic strokes occurred.
Results— After stratifying by race and sex and adjusting for other nonlipid risk factors, there was no significant relation between the apoE genotype and incident stroke, except in black women (hazard ratio for ε2 genotype relative to ε3/ε3=0.53; 95% CI, 0.28 to 0.99).
Conclusions— For the most part, in this middle-aged sample, apoE was not a risk factor for incident ischemic stroke.
Apolipoprotein E (apoE) genotype ε2 alleles are associated with lower blood cholesterol levels, whereas ε4 alleles are consistently associated with higher levels.1 A meta-analysis estimated that ε4 carriers versus noncarriers had an odds ratio of 1.42 (95% CI, 1.26 to 1.61) for coronary heart disease.2 In contrast, the relation between the apoE genotype and ischemic stroke has remained unclear despite many published articles. A recent meta-analysis found no relation between stroke and apoE, contradicting a previous meta-analysis that reported a positive association between stroke and apoE.3,4
The lack of a consistent association between the apoE genotype and ischemic stroke may be in harmony with little prospective relation between serum cholesterol and ischemic stroke,5 except at younger ages.6 We hypothesized that the incidence of ischemic stoke is not significantly different between apoE groups consisting of ε2 (2/2, 2/3, and 2/4), ε3 (3/3), and ε4 (3/4 and 4/4). A secondary question was whether the apoE genotype interacts with race, hypertension, and other risk factors for stroke.
The Atherosclerosis Risk in Communities (ARIC) study cohort consists of a population sample of 15 792 individuals, aged 45 to 64 years old at baseline when recruited between 1987 and 1989, from 4 communities: Jackson, Miss; Forsyth County, NC; Washington County, Md; and the northwestern suburbs of Minneapolis, Minn. The Jackson sample was comprised of black residents only. Previous reports have detailed ARIC measurements.7
At the baseline examination, blood samples were obtained and plasma, serum, and DNA isolated. Blinded apoE genotyping was done using the TaqMan assay (Applied Biosystems).
ARIC study participants were contacted annually by phone, and all of the hospitalizations and deaths during the previous year were identified. In addition, local hospitals provided lists of cardiovascular disease discharges, which were examined for participant hospitalizations. Hospital reports were reviewed for evidence of acute stroke if the discharge diagnosis included a cerebrovascular disease code (International Classification of Diseases, 9th revision, code 430 to 438), if a cerebrovascular procedure was mentioned in the summary, or if the computed tomography or magnetic resonance report showed evidence of cerebrovascular disease. ARIC adapted National Survey of Stroke criteria for its stroke definition.8 These criteria require the following: (1) evidence of sudden or rapid onset of neurological symptoms that persist for >24 hours or lead to death; (2) absence of a nonstroke cause (such as brain tumor); and (3) presence of ≥1 major neurological deficit (eg, aphasia or hemiparesis) or 2 minor neurological symptoms (eg, diplopia and dysarthria). Strokes were classified as ischemic if neuroimaging revealed acute infarct or in the absence of hemorrhage. Approximately 98% of stroke records had either computed tomography or MRI.
This report is based on follow-up through 2001. We included whites and blacks who had no previous stroke or transient ischemic attack at baseline and had consented to DNA analysis. This excluded 631 with previous events, 44 without consent, and 438 missing DNA information, leaving 14 679 participants. For the primary analysis, apoE genotypes were divided into 3 categories based on low-density lipoprotein cholesterol gradient: genotype ε2 (2/2, 2/3, and 2/4), genotype ε3 (3/3, the reference), and genotype ε4 (3/4 and 4/4). Prevalences or mean values of baseline risk factors were compared among the genotype groups using general linear models (SAS Institute). We used Cox regression to assess race- and sex-specific hazard ratios (HRs) for the apoE genotype and ischemic stroke incidence.
ApoE was in Hardy–Weinberg equilibrium in both blacks (P=0.19) and whites (P=0.49). As shown in Tables 1 and 2⇓, blacks had a higher proportion in the ε2 and ε4 groups than did whites. As Table 2 illustrates, baseline low-density lipoprotein cholesterol and total cholesterol increased from the ε2 group to the ε4 group and high-density lipoprotein cholesterol followed an opposite pattern. Because of the large size of the ARIC cohort, many of the other Table 2 relations were statistically significant but were small or exhibited no trend among genotypes. Those who had missing genotypes or did not consent did not have meaningfully different risk factors from those in Table 2.
In over 183 569 person-years of follow-up, there were 498 incident ischemic strokes (114 black women, 97 black men, 107 white women, and 180 white men).
As shown in the first 3 rows of Table 3, black women (significantly) and white men and women (nonsignificantly) showed decreased HRs of incident ischemic stroke for the ε2 group compared with those with the ε3/ε3 genotype (HRs 0.53, 0.68, and 0.86, respectively). Black and white women also had nonsignificantly increased HRs for the ε4 group (HRs 1.42 and 1.17, respectively). The association of the apoE genotype with ischemic stroke was in the opposite direction for black males: relative to the ε3 group, the ε2 group had a HR >1 and the ε4 group <1. The apoE group did not interact with age, prevalent cardiovascular disease, body mass index, hypertension, diabetes, smoking status, education level, or total cholesterol. Furthermore, adjustment for possible confounding by these risk factors (data not shown) did not alter the results.
For comparison with published reports, we performed our analysis by other apoE groupings (Table 3). In some of these comparisons, black women had increased HRs for ε4 relative to various groupings, as well as the decreased HRs of stroke for ε2.
This is the largest cohort study to date on apoE genotype and ischemic stroke incidence, as well as the only study that examined the relation for blacks and whites. Although black women who carried the ε2 allele had a decreased incidence of ischemic stroke relative to ε3/ε3, we did not find any appreciable relation between the apoE category and incident stroke in the ARIC cohort for either blacks males or whites. This latter result matches the majority of other studies.
Our main study limitation was the limited power to detect a very weak, but real, relation between ischemic stroke and apoE genotype and to fully evaluate the sex interaction we observed for blacks. Additionally, we did not discriminate between ischemic stroke subtypes. There exists the slight possibility of population stratification bias, but this possibility is attenuated by our race-specific reporting.
Studies of the stroke-apoE association have differed in their modeling of the apoE polymorphism, often using heterogeneous categorizations that misrepresent the biological effects of apoE. As shown in Table 3, some comparisons that are less biologically based are more likely to produce statistically significant results, although we still observed mostly null results.
Most studies, including ARIC, reported little relation between ischemic stroke and blood lipids.5 As shown in Table 2, the apoE genotype did correlate with blood lipid levels,but this did not affect stroke incidence. In conclusion, apoE genotype appears to be, at most, a weak risk factor for ischemic stroke.
The ARIC Study was funded by National Heart, Lung, and Blood Institute awards N01-HC-55015, N01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC-55020, N01-HC-55021, N01-HC-55022, T32-HL07779, and R01-HL073366. The authors acknowledge the staff and participants in the ARIC Study for their important contributions.
- Received June 8, 2005.
- Revision received June 8, 2005.
- Accepted July 18, 2005.
Davignon J, Gregg RE, Sing CF. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis. 1988; 8: 1–21.
McCarron MO, Delong D, Alberts MJ. apoE genotype as a risk factor for ischemic cerebrovascular disease: a meta-analysis. Neurology. 1999; 43: 1308–1311.
Shahar E, Chambless LE, Rosamond WD, Boland LL, Ballantyne CM, McGovern PG, Sharrett AR. Plasma lipid profile and incident ischemic stroke: ARIC. Stroke. 2003; 34: 623–631.
ARIC Investigators: The Athersoclerosis Risk in Communities (ARIC) Study: design and objectives. Am J Epidemiol. 1989; 129: 687–702.