Extracranial carotid atherosclerosis is associated with increased neurofibrillary tangle accumulation.

Abstract

Objective:

To determine whether extracranial carotid atherosclerotic disease (ECAD) is associated with increased key neurodegenerative pathology such as neurofibrillary tangle (NFT), beta-amyloid plaque or cerebral amyloid angiopathy (CAA) accumulation, findings associated with Alzheimer’s disease (AD) and other dementias.

Methods:

Our prospective longitudinal clinicopathological study, the Arizona Study of Aging and Neurodegenerative Disorders/Brain and Body Donation Program, records the presence or absence of clinically-diagnosed ECAD and does semi-quantitative density estimates of NFT, beta amyloid plaques and CAA at death. After adjusting for potential confounding factors determined by logistic regression analysis, histopathology density scores were evaluated in individuals with ECAD (n=66) and individuals without ECAD (n=125).

Results:

We found that the presence of ECAD is associated with a 21% greater NFT burden at death compared to no ECAD (P=.02). Anatomically, increased NFT burden was seen throughout the brain regions evaluated but was only significant in the temporal lobe (P <.05) and the entorhinal cortex (P=.02). Complimentary to this finding, we found that subjects with carotid endarterectomy (CEA), the surgical treatment of ECAD (n=32), had decreased NFT densities compared to those who had ECAD without CEA (n=66) (P=.04). In contrast to NFT, we found that ECAD was not associated with beta-amyloid plaque or CAA density.

Conclusion:

These findings indicate that ECAD is associated with NFT burden in the temporal lobe and entorhinal cortex, which has clinical significance for AD and non-AD dementias and cognitive dysfunction. Further understanding of whether ECAD increases risk for neurodegenerative brain changes is highly relevant because ECAD is a treatable disease that is otherwise not evaluated for or specifically treated as a dementia risk factor.

Graphical Abstract

Extracranial carotid atherosclerotic disease is associated with increased neurofibrillary tangle accumulation at death and patients with carotid endarterectomy demonstrated comparatively decreased NFT accumulation in this prospective study of 1,234 subjects who underwent neuropathological evaluation. This indicates that extracranial carotid atherosclerotic disease is associated with neurodegenerative changes seen in Alzheimer’s disease and other dementias.

INTRODUCTION

NFTs and beta-amyloid (Aβ) plaques are two key histopathological hallmarks of AD,¹ and are found in other types of dementias and forms of cognitive dysfunction.^2,3 Moreover, both have been identified in early clinical stages of cognitive dysfunction such as mild cognitive impairment.^4-6 NFTs are comprised of hyperphosphorylated tau, a microtubule-associated protein.¹ Histologically, NFTs are seen as fibrillary intracellular deposits in the cell bodies and apical dendrites of neurons.² Their accumulation causes axonal transport dysfunction, loss of synapses, and mitochondrial and cytoskeletal impairment.⁷ NFTs spread following a well-defined pattern, according to the Braak staging: Stages I-II: lesions are limited to the transentorhinal and entorhinal regions of the brain, Stages III-IV lesions extend to the limbic region, including the hippocampus; Stages V-VI: lesions expand to neocortical regions. ⁸Aβ peptide is a proteolytic fragment of the transmembrane amyloid precursor protein,⁹ which are deposited in the brain parenchyma as amyloid plaques or as CAA in the cerebrovasculature under pathological conditions.¹⁰

Risk for NFT accumulation is multifactorial and includes genetic and non-genetic risks. Non-genetic risks linked to hyperphosphorylation of tau include cerebral ischemia, which can be associated with vascular diseases including diabetes, hypertension and intracranial vascular disease.^11,12 Extracranial carotid atherosclerotic disease (ECAD), refers to atherosclerotic plaque build-up at the common carotid artery bifurcation and the internal carotid artery (ICA) causing arterial stenosis. ECAD had approximately 21% prevalence in subjects aged 30-79 years in the general population in 2020.¹³ ECAD prevalence increases with age.¹⁴ Although the ICA provides blood supply to the majority of the brain and ECAD has been associated with decreased cerebral perfusion¹⁵, cognitive decline¹⁶, and some markers associated with carotid atherosclerosis such as common carotid intima/media thickness and carotid plaque number are associated with incident dementia,^17-19 it is unknown whether ECAD is associated with NFT accumulation or other neurodegenerative brain changes associated with dementia. This has clinical relevance because ECAD has specific treatment options that are reserved primarily for stroke prevention without clinical consideration of cognitive or neurodegenerative outcomes such as NFT burden or dementia risk.

In the present study, we evaluated NFT, Aβ plaque and CAA burden in 1,234 post-mortem brains from the Arizona Study of Aging and Neurodegenerative Disorders (AZSAND) and Brain and Body Donation Program (BBDP) by histopathology. We compared brain pathology in subjects with and without ECAD while controlling for potential confounding factors defined by a logistic regression analysis.

METHODS:

Subject Recruitment:

This prospective study, quantified brain pathology at death in subjects in the Banner-Sun Health Research Institute Brain Donation Program, a large, IRB-approved clinicopathologic study of aging and neurodegenerative disease across metropolitan Phoenix, Arizona.²⁰ Participants were enrolled from year 2000 to 2018 and had an annual follow-up visit until death. At death, they underwent autopsy and neuropathological brain assessment.²⁰

Neuropathological assessment:

A complete description of brain dissection, fixation and histological processing has been previously described.²⁰ Briefly, pathological scoring is performed blinded to clinical and neuropathological diagnosis. Amyloid plaque and neurofibrillary tangle density are graded and staged at standard coronal levels in large, thick (80 mm) sections encompassing complete frontal, temporal, and parietal lobes, as well as hippocampus CA1 region and entorhinal cortex, based on the aggregate impression from the 80 mm sections stained with thioflavin S, Campbell- Switzer and Gallyas methods.²¹ The total plaque score, considering all types of plaques (cored, neuritic and diffuse) together, is predominantly derived from the Campbell- Switzer stain while the Gallyas and thioflavin S stains are especially useful for estimating neuritic plaque densities. All three stains show neurofibrillary changes and therefore this score is estimated after viewing slides stained with all three. Both total and neuritic plaque densities are rated as none, sparse, moderate and frequent, using the published CERAD (Consortium to Establish a Registry for Alzheimer Disease) templates.²² These templates have been shown to give more consistent interobserver estimates than actual counts.²³ Conversion of the descriptive terms to numerical values give scores of 0–3 for each area, with a maximum score of 15 for all five areas combined.

Subject population:

Primary Analysis: Evaluation of ECAD association with brain pathology. Groups were defined according to the presence or absence of ECAD history in the participants’ medical records, data acquired through an extensive chart review (Figure 1). Subjects were excluded who had last follow-up visit more than 2 years prior to death, history of carotid endarterectomy or missing data. Secondary Analysis: Evaluation of CEA association with NFT pathology. Groups were defined as those with ECAD with or without CEA. Subjects were excluded who had last follow-up visit more than 2 years prior to death or missing data.

Figure 1. Study design and Patient breakdown.

Abbreviations: CEA, Carotid Endarterectomy; PAD, Peripheral Artery Disease; ECAD, Extracranial Carotid Atherosclerotic Disease; CVD, Cardiac Valvular Disease; TIA, transient ischemic attack.

Statistical analyses:

Patient demographics (Table I ) and comparisons between groups were analyzed using Wilcoxon Rank Sum test or Fisher’s Exact test, as applicable, to test the significance of the differences between continuous and categorical variables; a P value < .05 was considered statistically significant. For the primary analysis, confounding factors were identified using a logistic regression considering ECAD status as outcome variable. Significant confounding variables (P value<.05) were used to generate a matched population using 2:1 matching procedure as described by Mortensen et al.²⁴ The secondary analysis was carried on with univariate analysis in an unadjusted population. Statistical analyses were conducted with SAS Studio v3.8.

Table I.

Baseline characteristics for Pre-match and Post-match patients with and without history of ECAD.

	Pre-match			Post-match a
	No ECAD	ECAD		No ECAD	ECAD
	n (%)	n (%)		n (%)	n (%)
Number of Patients	705	66		125	66
Age			P-value			P-value
Mean ± SD	83.0 ± 8.71	85.5 ± 7.74	0.02	85 ± 6.74	85.5 ± 7.74	0.65
Gender
Female	274 (38.9)	22 (33.3)	0.45	52 (41.6)	22 (33.3)	0.34
Male	431 (61.1)	44 (66.7)		73 (58.4)	44 (66.7)
Co-morbidities
CVD	114 (16.2)	23 (34.8)	<0.001	44 (35.2)	23 (34.8)	1.00
PAD	88 (12.2)	20 (30.3)	<0.001	33 (26.4)	20 (30.3)	0.69
Stroke	133 (18.9)	26 (39.4)	<0.001	46 (36.8)	26 (39.4)	0.85
TIA	127 (18)	31 (47)	<0.001	55 (44)	31 (47)	0.81

Abbreviations: CVD, cardiac valvular disease; PAD, peripheral artery disease; TIA, transient ischemic attack.

^aMatching procedure in 2:1 ratio was performed using SAS Studio 3.8v.

RESULTS

Of the 1,234 total subjects, 803 met the inclusion and exclusion criteria of our study. After the match process, 191 patients were available for analysis (Figure 1). Subject groups were determined according to the ECAD status. Demographics and significant comorbidities are shown in Table I. In the unmatched population, 66 individuals (mean [SD] age, 85.5 [7.74] years) had a diagnosis of ECAD, whereas 705 individuals (mean [SD] age, 83.0 [8.71] did not have a diagnosis of ECAD. Gender did not differ between groups. The logistic regression analysis determined as potential confounders associated with ECAD the following comorbidities: cardiac valvular disease, peripheral artery disease, stroke, and transient ischemic attack. Table I shows demographic characteristics of the pre and post-match populations.

In the matched cohorts, ECAD was associated with a 21% increase in NFT burden compared to the no ECAD group (P = .02, see Figure 2A). Beta-amyloid plaque (Figure 2B) and CAA (Figure 2C) scores were not statistically different between groups. Although increased NFT burden was observed throughout the brain in subjects with ECAD, differences observed in the temporal lobe and the entorhinal cortex were significant (P <.05, P =.02, respectively) (Figure 3).

Figure 2. NFT burden is significantly greater in subjects with ECAD compared to those without ECAD.

Subjects with No ECAD (n=125) and with ECAD without CEA (n=66) were matched for peripheral artery disease, cardiac valvular disease, stroke, and TIA. Histopathology scores for NFT (A), Aβ plaques (B) and CAA (C) are shown with mean and SD.

Figure 3. NFTs densities are significantly larger in the temporal lobe and the entorhinal cortex in subjects with ECAD.

The NFTs densities are graded in the parietal, frontal, temporal lobes, as well as the Hippocampal CA1 region and the entorhinal cortex. NFT scores are shown with mean and SD.

In a secondary analysis, ECAD surgical treatment was evaluated. Subjects with ECAD and a history of CEA was compared to the cohort of subjects with ECAD without CEA (Figure 4). We found NFT score is significantly decreased (P = .04) in subjects with ECAD with CEA compared to those with ECAD without CEA. Based on regression analysis, age, cardiac implantable electronic devices, statins and anti-platelet therapy were found to be potential confounders associated with ECAD+CEA (Table II). However, there were inadequate subjects to generate a matched cohort for the above-mentioned factors.

Figure 4. In subjects with ECAD, those with CEA have decreased NFT burden compared to those without CEA.

Subjects with ECAD with CEA (n= 32) and with ECAD without CEA (n=66) were evaluated for NFT burden at death. NFT scores are shown with mean and SD.

Table II.

Baseline characteristics for ECAD patients with and without history of CEA.

Pre-match
	ECAD w/o CEA	ECAD w/ CEA
	n (%)	n (%)
Number of Patients	66	32
Age			P-value
Mean ± SD	85.5 ± 7.74	85.8 ± 5.53	0.86
Gender
Female	22 (33.3)	11 (34.4)	1.00
Male	44 (66.7)	21 (65.6)
Co-morbidities
Statin therapy	21 (31.8)	18 (56.2)	0.04
Antiplatelet therapy	34 (54.5)	26 (81.2)	0.01
CIEDs	7 (10.6)	9 (21.9)	0.06

Abbreviations: CIEDs, cardiac implantable electronic devices.

DISCUSSION

We found that NFT burden at death is higher in individuals with a history of ECAD than in individuals who did not have ECAD (Figure 2A) after controlling for potential confounders. Complimentary to this, we found that having surgical removal of the carotid plaque through CEA is associated with decreased NFT burden at death compared to those with ECAD without CEA. These findings raise a new concept that ECAD could contribute to specific neuropathology that is relevant in AD as well as other forms of cognitive impairment and that treatment with CEA could mitigate this effect.

Although a potential role of ECAD in NFT accumulation has not been established, there are growing data that implicate ECAD in cognitive decline ^25-27 and less so in AD risk¹⁸. The majority of studies assessing carotid atherosclerosis related to dementia risk use intima/media ratios of the common carotid artery or plaque counting methods that likely have more to do with systemic atherosclerosis than carotid stenosis.^17-19 Our study is unique because it provides initial evidence that ECAD is associated with specific change in NFT pathophysiology in humans, with significant NFT accumulation in the temporal lobe and entorhinal cortex, the areas of the brain first affected in tau aggregation during asymptomatic AD stages^28,29. PET-based studies in humans have found decreased cerebral perfusion associated with higher tau depositions in the entorhinal cortex, independent of Aβ pathology³⁰. In addition, animal-based studies demonstrate that decreased cerebral perfusion is associated with increased tau phosphorylation without affecting amyloid pathology in the hippocampus and cortex^31,32. This correlates with our results, which suggest specificity of ECAD association with NFT accumulation and raises the hypothesis that hypoperfusion associated with ECAD may contribute to NFT accumulation.

Symptomatic ECAD, meaning those with ECAD and stroke or transient ischemic attach (TIA) within the previous 3-6 months, could not be confidently defined in contrast to those with asymptomatic ECAD in our patient population. However, in our primary analysis, all subjects with CEA were excluded and the two compared groups (with and without ECAD) were matched for stroke and TIA prevalence. This suggests that the observed ECAD association with increased NFT burden is independent of the known increased risk for stroke or TIA in patients with ECAD. As such, we hypothesize that the observed associations are relevant for patients with asymptomatic ECAD.

Our study has key limitations. 1) It is not randomized as to who has or does not have ECAD. As such, it is impossible to account for all potential confounders. However, to address this, the cohorts were matched according to potential confounding factors which were selected based on logistic regression analysis. 2) Subjects did not have carotid disease severity (percent stenosis) defined or independently confirmed by the study. As such, the cohort likely represents a mixed group of subjects with moderate and severe carotid stenosis and we cannot rule out that subjects with <50% stenosis were included. However, inclusion of subjects with minimal carotid stenosis in the ECAD cohort would tend to dilute differences between the cohorts based on ECAD status. This would result in decreased ability to find differences based on ECAD status. It may be that increasing carotid stenosis severity results in worsening NFT accumulation. 3) Because of limited subjects with ECAD with or without CEA, our ability to control this part of the study for potential confounders was limited. Despite this, these data are complementary to our hypothesis that ECAD is associated with increased NFT burden, yet follow-up is needed with larger cohorts optimized for analysis of carotid disease interventions. 4) Our study population had few subjects who underwent carotid artery stenting and no meaningful comparison could be made in this population.

CONCLUSION

Our data suggest that ECAD is associated with a 21% increased burden of NFT burden at death. NFT burden is highly relevant in a broad group of AD and non-AD dementias.² These findings are important because whether and how ECAD is related to cognitive decline in general and dementia specifically is poorly understood. It is particularly compelling because ECAD is a treatable condition, yet is not clinically evaluated or treated in the context of dementia risk or cognitive outcomes and is currently offered to only a small percentage of patients with ECAD.³³ Understanding the role of ECAD in tau pathology genesis and progression could increase understanding of vascular contributions to dementia and create additional impetus for research leading to evaluation and treatment options in patients at risk for dementia. Additional prospective studies may help confirm and extend these results.

ARTICLE HIGHLIGHTS.

Type of Research:

Prospective study.

Key Findings:

Extracranial carotid atherosclerotic disease (ECAD) is associated with 21% increased neurofibrillary tangle (NFT) burden at death compared to controls. In patients with ECAD those who underwent carotid endarterectomy (CEA) had significantly lower neurofibrillary densities compared to those without CEA.

Take home Message:

ECAD is associated with increased NFT accumulation, a key histopathological hallmark of Alzheimer’s disease and other dementias. In agreement, CEA was associated with decreased NFT burden compared to those with ECAD without CEA.