Carotid Intima-Media Thickness and Amyloid-β Deposition: The ARIC-PET Study

INTRODUCTION

Dementia remains a public health concern as Alzheimer’s disease (AD) is the sixth leading cause of death in the U.S. [1] Changes in the brain may occur years prior to the development of clinical AD. Biomarkers, such as amyloid-β (Aβ) and tau, are believed to be markers of disease progression, and have been utilized to enhance understanding of AD pathophysiology and identifying AD risk factors. Aβ deposition likely occurs prior to that of tau [2]. Therefore, determining risk factors for elevated Aβ deposition may identify those at higher risk of AD.

Vascular risk factors, many of which are potentially modifiable, are believed to play a role in AD development, especially when considered in midlife [3]. The Atherosclerosis Risk in Communities (ARIC) study previously reported midlife vascular risk factors, including body mass index (BMI), hypertension, and diabetes, are linked to greater Aβ deposition in late life [4]. Elevated carotid intima-media thickness (cIMT), a marker of atherosclerosis that may reflect the cumulative contribution of vascular risk factors [5], has been prospectively associated with increased dementia risk [6, 7]. Atherosclerosis may lead to brain hypoxia, which could then result in increased Aβ deposition [8]; however, few studies have evaluated the relationship between cIMT and Aβ deposition [7, 10]. Cross-sectional studies reported null results [9, 10]. A prospective study found elevated cIMT was associated with abnormal Aβ levels, measured in cerebrospinal fluid (CSF), though this was not independent of vascular risk factors [7]. Therefore, additional research is warranted, particularly in nondemented populations. Utilizing data from the ARIC-PET study, we hypothesized that higher cIMT in midlife is prospectively associated with elevated Aβ deposition in late life.

METHODS

Study population

The ARIC study is a community-based prospective cohort study [11]. At baseline (1987–1989), 15,792 participants were recruited from four US communities: Forsyth County, NC; Jackson, MS; Washington County, MD; northwest suburbs of Minneapolis, MN. Participants have attended several follow-up visits and are continuously followed for hospitalizations. At visit 5 (2011–2013), 346 participants had a PET scan. We excluded those whose race was other than Black or White, as well as those with missing cIMT or covariate measures. After exclusions, 332 participants were included in this analysis (Fig. 1). Institutional review boards at each center approved the study and participants gave written informed consent at each study visit.

OPEN IN VIEWER

Fig. 1

Study sample flowchart.

cIMT measurements

At visit 2 (1990–1992), trained technicians conducted ultrasounds using Biosound 2000 II duplex scanners. Scans were read at the ARIC Ultrasound Reading Center [12]. cIMT was assessed in three segments of the extracranial carotid arteries: the internal carotid, carotid bifurcation, and common carotid [12]. A total of 11 measurements of the far wall were attempted at each segment in 1-mm increments and the mean was calculated. Measurements were made regardless of presence of focal plaque. Reliability coefficients for the mean far-wall thickness at the internal carotid, carotid bifurcation, and common carotid were 0.73, 0.77, and 0.70 [12].

At visit 5 (2011–2013), a subset of participants without dementia from 3 centers (Jackson, Mississippi; Washington County, Maryland; Forsyth County, North Carolina) who had neurocognitive assessments and a brain MRI were recruited into the ARIC-PET study. ARIC-PET study exclusion criteria have been described [13].

ARIC brain MRI imaging and florbetapir PET scan protocols have been described [13, 14]. Briefly, MRI imaging was conducted using 3T scanners, and magnetization-prepared rapid gradient echo was used for coregistration of PET images. Florbetapir PET scans were performed within one year of the MRI. Images were reviewed and quantified for standardized uptake value ratios (SUVRs). The global cortical measure of Aβ (weighted average of the orbitofrontal, prefrontal, and superior frontal cortices, lateral temportal, partietal, and occipital lobes, precuneus, and anterior and posterior cingulates) was used for this analysis. As done in prior ARIC studies, because SUVR was highly skewed, SUVR > 1.2, the sample median, was considered elevated Aβ [4, 15].

Covariate measurements

Covariates assessed from visit 2 included age, sex, self-reported race, APOE ɛ4, BMI, systolic blood pressure, antihypertensive medications, smoking status, and diabetes. Self-reported education was obtained from visit 1. Blood pressure and BMI were defined according to standard ARIC procedures [16]. Diabetes was defined as a fasting serum glucose ≥126 mg/dl, non-fasting serum glucose ≥200 mg/dl, self-reported physician diagnosis of diabetes, or antidiabetic medication use in the past 2 weeks. Medication use was recorded by technicians via review of medication bottles that participants brought to clinic visits. APOE ɛ4 genotyping was done via TaqMan assay (Applied Biosystems, Foster City, CA, USA) [17].

Statistical analysis

Participant characteristics, stratified by cIMT tertiles, were described using means and standard deviation (SD) for continuous variables and frequencies and percentages for categorical variables. Logistic regression was used to assess the relationship between cIMT and elevated Aβ deposition. cIMT was assessed in tertiles and per 1-SD increment for this analysis. Multivariable models were adjusted as follows: model 1 adjusted for age, sex, race, and APOE ɛ4; model 2 further adjusted for BMI, smoking status, and education; model 3 additionally adjusted for systolic blood pressure, antihypertensive medication use, and diabetes. Inverse probability weighting was used to account for potential selection bias due to inclusion in the brain MRI sample and attrition due visit non-attendance [18]. Multiplicative interactions by age (median split), sex, race, and APOE ɛ4 were assessed by including cross-product terms in the model. Analyses were conducted using SAS v9.4 (SAS Institute Inc., Cary, NC, USA).

3 RESULTS

We included 332 participants in this analysis. At baseline, participants were a mean (SD) age of 55 (5) years; 56% female and 41% Black. Median follow-up time was 22 years, and 170 participants had elevated Aβ. Compared to the lower tertiles, participants in the highest cIMT tertile were more likely to be older, male, diabetic, have higher systolic blood pressure and BMI, and identify as Black (Table 1).

After adjustments for demographics and APOE ɛ4, participants in the highest cIMT tertile had greater odds of elevated Aβ (model 1 OR [95% CI]: 2.17 [1.15–4.11]) compared to those in the lowest tertile. This association was no longer significant with further adjustments for vascular risk factors (model 3 OR [95% CI]: 1.87 [0.95–3.68]), though the measure of effect remained in the hypothesized direction (Table 2). In addition, an interaction by APOE ɛ4 was observed (p = 0.02). Among participants with ≥1 APOE ɛ4 allele, those in the highest cIMT tertile had 6.94 (95% CI: 1.32–36.53) greater odds of elevated Aβ than those in the lowest tertile after full model adjustments, including vascular risk factors. No significant association was present among those without APOE ɛ4 (highest versus lowest tertile OR [95% CI]: 1.32 [0.60–2.91]). No interactions by age, sex, or race were noted (p > 0.10).

In this cohort of individuals without dementia, participants with greater cIMT at midlife had higher odds of elevated Aβ deposition in late life, independent of demographics and APOE ɛ4. Additionally, an APOE ɛ4 interaction was present. In participants with ≥1 APOE ɛ4 allele, greater cIMT was significantly associated with elevated Aβ deposition in late life. No association was observed among those with no APOE ɛ4 alleles. However, results should be interpreted with caution as precision was poor for these APOE ɛ4 subgroup analyses.

Midlife cIMT is associated with dementia [6, 7]; however, few studies have assessed the relationship between cIMT and Aβ accumulation. The Swedish BioFINDER subcohort (n = 330) found the highest cIMT quartile (> 0.81 mm) was prospectively associated with greater odds of abnormal Aβ, measured in CSF, compared to the lowest quartile (OR [95% CI]: 2.26 [1.03–4.95]) 20 years later and after adjusting for age; however, this association was not independent of sex, vascular risk factors, and APOE ɛ4 [7]. Two cross-sectional studies, however, reported no association [9, 10].

The concept that cIMT may be linked to elevated Aβ burden is supported by studies that have linked specific cardiovascular risk factors, which are also associated with cIMT, to Aβ deposition. For instance, higher blood pressure [19] and ankle brachial index [20] have been associated with elevated Aβ burden. In prior ARIC analyses, however, no association was observed between lipids and future Aβ burden [21], or cross-sectionally between intracranial atherosclerosis and Aβ burden [15].

A previous ARIC analysis indicated that having two or more vascular risk factors in midlife, but not late life, is associated with elevated Aβ [4]. Similarly, in the present analysis, cIMT, which may represent the cumulative effect of vascular risk factors [5], was associated with elevated Aβ. These findings highlight the potential role of vascular disease in the development of AD. It is possible that preventing or effectively managing vascular risk factors in early and midlife could potentially decrease the risk of elevated Aβ deposition and AD.

Our results indicate an APOE ɛ4 interaction, in which carriers have a higher risk for Aβ deposition than noncarriers. APOE, a cholesterol and lipid carrier, binds and transports Aβ. Each APOE isoform has a different binding affinity for Aβ, with APOE ɛ4 being less efficient at promoting Aβ clearance than APOE ɛ2 or ɛ3. APOE ɛ4 initiates and accelerates Aβ accumulation and deposition in the brain, which then likely increases the risk of developing AD. Additionally, it has been suggested that APOE ɛ4 combines synergistically with atherosclerosis, further contributing to an increased risk of AD [22]. Similar atherosclerosis and APOE ɛ4 interactions have been previously observed [20]. This, in combination with our results, suggests that those with markers of atherosclerosis who are also APOE ɛ4 carriers may have a greater risk of elevated Aβ deposition.

A prior meta-analysis of 119 clinical trials indi-cated that reducing cIMT progression reduces the risk for cardiovascular disease [23]. In addition, the Framingham Offspring study reported that cIMT progression is associated with lower hippocampal volume, suggesting that cIMT progression may increase the risk for neurodegeneration [24]. Although our results suggest greater cIMT in middle age may be a risk factor for Aβ deposition in later life, future studies should assess whether changes in cIMT over time affects levels of Aβ deposition. There has also been evidence that antihypertensive medication use may reduce the rate of cIMT progression [25], further indicating that additional prospective research is warranted to assess cIMT progression and potential treatments, such as antihypertensive medications.

Strengths of this study include the prospective design, relatively large sample with PET scans, and the inclusion of white and Black participants. However, limitations exist. Participants who obtained brain MRI and PET scans tended to be healthier than the full cohort. Although we utilized inverse probability weighting to account for attrition due to visit non-attendance or selection into the brain MRI study, it is possible that selection bias remains. Additionally, participants with dementia, who may have greater cIMT and/or Aβ levels, were excluded from the ARIC-PET study, which likely biases our results towards the null.

In conclusion, higher cIMT at midlife was prospectively associated with greater Aβ deposition in late life, independent of demographics and APOE ɛ4. Furthermore, among APOE ɛ4 carriers, higher cIMT was associated with significantly greater odds of Aβ deposition, independent of vascular risk factors. These findings suggest markers of carotid atherosclerosis in middle age may be a risk factor for Aβ deposition in later life, especially among APOE ɛ4 carriers.