Vascular Cognitive Impairment and Dementia (VCID): Clinical Features, Neuropathology and Biomarkers

Abstract

Vascular cognitive impairment and dementia (VCID), i.e., cognitive impairment secondary to cerebrovascular disease (CeVD), is the second most common form of dementia after Alzheimer’s disease (AD), accounting for 15-20% of all cases. CeVD, in fact, contributes to dementia alongside other neuropathologies in up to 75% of dementia cases. CeVD and AD not only frequently co-occur in the brain, but they may also interact, and some VCID risk factors (midlife hypertension and diabetes) also increase AD risk. Since CeVD and cardiovascular disease (CVD) share risk factors and pathophysiology, the cardiovascular clinician is likely to encounter both in the clinic. Moreover, common cardiac disorders, such as atrial fibrillation, heart failure, acute coronary syndrome and valvular disease, increase VCID risk. There have been recent developments in the diagnostic criteria for VCID, with advances in risk biomarkers, treatment and prevention of cognitive impairment and dementia. The diagnosis of VCID is a two-step process, with the initial identification of a cognitive syndrome followed by the establishment of a predominantly vascular etiology, guided by clinical history and examination and substantiated by neuroimaging, preferably MRI. Clinical presentations include an acute onset, a stepwise decline, a fluctuating course if due to multiple strokes, or a gradual slow progression if attributable to cerebral small vessel disease. Cognitive deficits can be found in several domains, such as information-processing speed, attention, executive function, and emotional lability, sometimes referred to as the subcortical syndrome, often seen in the early stages of VCID without cortical infarcts. The diagnosis is supported by the identification of large and small infarcts, lacunes, white matter hyperintensities, dilated perivascular spaces and cerebral microbleeds using MRI. This part 1 of a 2-part JACC review series describes the clinical features, pathophysiology and biomarkers of VCID for cardiovascular clinicians who have a critical role in its early identification, management and prevention in their patients.

CONDENSED ABSTRACT:

Vascular cognitive impairment and dementia (VCID) is the second most common form of dementia, accounting for 15-20% of all cases. However cerebrovascular disease is often comorbid in neurodegenerative disorders and contributes to 75% of all dementia cases. The risk factors and pathophysiology of VCID are shared with cardiovascular disease, and many cardiac disorders increase the risk of VCID. The cardiovascular clinician therefore has a key role in the early identification of VCID as well as its prevention.

Central Illustration:

The pathophysiology of VCID illustrating the role of risk factors and the contributions of both large and small vessel diseases. Vascular factors such as hypertension, diabetes, smoking and high cholesterol increase the risk of both cerebrovascular (CeVD) and cardiovascular diseases (CVD). Some CVDs independently increase the risk of CeVD. (see Box Table 3 for details). Cerebral small and large vessel diseases and CVD lead to hypoxic/anoxic and haemorrhagic injury to the brain, leading to structural damage and functional brain dysfunction through various mechanistic pathways. The final cognitive-neurobehavioral outcomes are influenced by the possible contributions of concomitant neurodegenerative disease and the modulatory effect of cognitive reserve mechanisms.

Media for promotion:

1. Perminder Sachdev in X (Twitter): @sachdevps

2. LinkedIn: https://www.linkedin.com/showcase/unsw-cheba/

3. Facebook: http://www.facebook.com/CHeBACentreforHealthyBrainAgeing

4. YouTube: https://www.youtube.com/channel/UCjcd8l3LtGfj8uYu4Tthtig

5. Instagram: https://www.instagram.com/unswcheba/

INTRODUCTION

Vascular cognitive impairment and dementia (VCID) is an umbrella term for a heterogeneous group of disorders in which an individual suffers cognitive decline attributable to ischemic and/or hemorrhagic damage to the brain due to vascular disease (Central Illustration). It is the second most common cause of dementia after Alzheimer’s disease (AD)¹ The concept of VCID has evolved historically and several terms continue to be used to describe it [Supplemental Table 1s]. The term multi-infarct dementia (MID) was introduced in 1974² to indicate that cognitive impairment was the consequence of multiple strokes and became synonymous with vascular dementia (VaD) for the next two decades. In the early 1990s, two sets of formal diagnostic criteria^3,4 were developed which expanded the concept to include not only multiple cortical and/or subcortical infarcts, but also strategic single infarcts, white matter lesions, hemorrhages, and hypoperfusion as possible causes of VaD. VCID elaborates the concept to include the entire spectrum of cognitive impairment of vascular etiology, from individuals at high risk but with no cognitive deficit (‘brain-at-risk stage’) to vascular mild cognitive impairment (VaMCI) to vascular dementia (VaD).

VCID is commonly encountered in the cardiology clinic and is perhaps under appreciated. Since cardiovascular disease (CVD) and cerebrovascular disease (CeVD) share risk factors such as hypertension, diabetes, smoking, hyperlipidemia, sendetariness and obesity, they frequently co-exist⁵ and management of both is necessary for optimal clinical outcomes. Cardiac disorders such as atrial fibrillation (AF), acute myocardial infarction (MI), heart failure and valvular disease are associated with increased risk of stroke as well as cognitive impairment. The presence of cognitive impairment may, moreover, compromise the management of heart disease.⁶ The cardiovascular clinician should therefore be alert to the possible presence of cognitive impairment in their patient, identification of which at an early stage is important to minimize future decline. The cognitive symptoms may become more obvious during an acute illness or hospitalization or following a cardiac procedure. Since many cardiology patients are older, cognitive impairment may be due to multiple brain pathologies, with concomitant cerebrovascular and Alzheimer’s pathologies being most common. In this Part 1 of a 2-part JACC State-of-the-Art review series, we examine the clinical features of VCID, the approach to diagnosis, and its pathophysiology, risk factors, neuropathology and biomarkers for a comprehensive assessment of a patient from a multidisciplinary perspective. The review serves as a critical primer for cardiovascular clinicians to recognize and mitigate cognitive complications in their patients.

EPIDEMIOLOGY

VaD accounts for 15-20% of all dementia cases.^1,7 In 11 population-based European studies conducted in the 1990s, the age-standardized prevalence of VaD was estimated at 1.6% compared to 4.4% for AD,¹ with VaD accounting for 15.8% of all cases of dementia. A 2025 Scientific Statement from the American Heart Association estimated VaD prevalence and incidence in those older than 65, to be 1.4 - 4.2% and 0.27 – 1.1% respectively.⁸ The Canadian Study of Health and Ageing used the broader concept of VCID and estimated that approximately 5% of people older than 65 years had VCID, with 2.4% having VaMCI, 0.9% having mixed dementia (vascular and neurodegenerative), and 1.5% having VaD.⁷ The incidence of VaD ranges from 6 to 12 cases/1000/year in those older than 70 years.⁹ In developing nations, including China, VaD has previously accounted for around 30% of dementia cases^10–12 recently revised to 17% of all dementia cases,¹³ possibly reflecting recent socioeconomic and demographic changes. VCID increases exponentially with age, with VaD affecting globally an estimated 1.5% of individuals between 70-84 and 5.4% of those older than 85.¹⁴

The prevalence of VaD is higher in poststroke patients. Three months following a stroke, 6 - 32% of individuals are diagnosed with VaD. Overall, approximately 10% of pre-stroke individuals have dementia and a further 10% or more will develop dementia in the months following stroke.¹⁵ If we include VaMCI, 44% of stroke patients have cognitive impairment 2-6 months after stroke.¹⁵ Vascular changes are common in autopsy studies of dementia cases, with greater than 75% of cases exhibiting some evidence of vascular pathology^16,17 and approximately one third showing significant vascular pathology.^18,19

The high prevalence of VCID in cardiovascular patients is summarized in Table 1. Several salient epidemiological studies of VCID are presented in Supplemental Table 2s.

Table 1.

The prevalence of VaD and VaMCI in the general population and patient subgroups

Group	Cognitive Impairment	Dementia
General Population179	-	VaD: 0.10% - 0.14%
Age Groups
65+180	VaMCI - 2.4%	Mixed dementia: 0.9% VaD: 1.5%
70 – 84181	-	VaD: 1.5%
85+181	-	VaD: 5.4%
Vascular diseases
Post-Stroke182	CI - 44%	Dementia: 20%
AF183	CI – 5.6%	Dementia: 2.8%
Heart Failure 64	CI - 41.4%	Dementia: 19.8%
Coronary artery disease 65		Relative risk of dementia 1.27 [1.07-1.50]
Acute coronary syndrome 66	CI – 16-20%	-
Severe Aortic stenosis 67	CI - 13%	-

CI: any cognitive impairment; Dementia: all-cause dementia

VaMCI: vascular mild cognitive impairment; VaD: vascular dementia

CLINICAL FEATURES AND DIAGNOSIS

VCID diagnosis requires evidence of cognitive decline from a previous level of functioning with the determination of a predominantly, if not exclusively, cerebrovascular etiology. Cognitive decline is judged on the basis of a subjective concern by the patient or a knowledgeable informant, and supported by objective evidence of cognitive impairment on testing. If the impairment is judged to be to the level that it impairs independent functioning by the individual, a diagnosis of dementia is made. A search for likely etiology follows, which in the case of VCID involves detection of the presence of significant CeVD based on history, physical examination and imaging biomarkers, and ruling out other possible neurodegenerative etiologies. In many cases, the likelihood of multiple underlying pathologies for MCI or dementia must be considered.

Subjective decline

The ‘subjective’ clinical concern of decline may stem from the patient or a knowledgeable other (e.g., family member or friend) or a professional who knows the patient well. Deficits reported may include patient reliance on others to plan or make decisions, repetition in conversation, need for frequent reminders, significant difficulties with finding words or expression, difficulty in navigating familiar environments, or difficulties in reading, writing or dealing with numbers. Eliciting concerns often requires careful questioning as these may not be voiced spontaneously.

Objective deficits

To make a diagnosis, the physician is required to document cognitive deficits by using objective measures. Screening with sub-tests such as the five-word immediate and delayed recall, six-item orientation task, and phonemic fluency tests from the Montreal Cognitive Assessment (MoCA)²⁰ is currently recommended,²¹ with administration of the entire MoCA, Trail Making Test, and semantic fluency test if time allows. The Mini-Mental State Examination (MMSE) has been used as an alternative, although it is less-sensitive for milder cases of VCID.²² A score of <26 on the MoCA or <25 on the MMSE suggests cognitive impairment. A very time-poor physician may use a briefer screening instrument such as the Mini-Cog²³ which has two items (recall of three words and drawing a clock) with a cut-off of ≤2, but its sensitivity and specificity are both low, limiting its utility.

On neuropsychometric testing, generally performed by a neuropsychologist, deficits are noted in several cognitive domains, with the widespread nature of cerebral vascular disease leading to the heterogeneous nature of deficits.²⁴ Cortical infarcts are associated with characteristic cognitive profiles including impairment in executive functioning, memory, language, visuospatial functioning, praxis (integrating simple and complex movements) and gnosis (recognizing form and nature through sensory modalities). Lesions largely restricted to subcortical structures characteristically present with abnormalities of information-processing speed, attention, executive function, and emotional lability, sometimes referred to as the subcortical syndrome, often seen in the early stages of VCID without cortical infarcts.^25,26 In contrast with AD, memory is often spared in the early stages of VCID, and when affected, the pattern of impairment is different. Whereas the pattern in AD is typically that of rapid forgetting, the individual with VCID with difficulty in retrieving memorized information is aided by cuing, suggesting that the information has not been lost but the retrieval process is inefficient. However, posterior cortical territory strokes may involve the medial temporal lobes and lead to memory impairment resembling that in AD.²⁷ Cognitive test batteries recommended for 10, 20 and 45-minute assessments of VCID developed recently from an international Delphi survey are listed in Supplemental Table 3s.

Establishing a predominantly vascular etiology

Vascular etiology is guided by clinical history and examination and substantiated by neuroimaging. The nature of onset and course, history of cerebrovascular events (strokes and transient ischemic attacks - TIAs) preceding the development or worsening of cognitive deficits, the presence of vascular risk factors, neurological signs suggestive of past strokes, and the nature of the cognitive deficits, with clear evidence of deficits in processing speed and executive function rather than episodic memory, all assist with the diagnosis (see Table 2– VASCOG-2-WSO criteria). Evidence of significant large vessel disease and cSVD is best based on neuroimaging features such as large and small infarcts, lacunes, atrophy, white matter changes, dilated perivascular spaces, cerebral microbleeds, hippocampal sclerosis, and cortical siderosis (see below). Since some degree of vascular pathology is common in the brains of older adults, expert clinical judgment is frequently necessary to determine if the observed lesions on imaging adequately support the diagnosis. Some guidelines have been presented in the various diagnostic criteria,^3,4,28–30 such as two or more cortical infarcts with at least one outside the cerebellum,³ single strategic infarcts (e.g., in the thalamus, angular gyrus or basal forebrain)⁴ or extensive confluent white matter lesions,²⁹ but no consensus exists and imaging abnormalities generally explain only a small proportion of the variance in cognitive function.³¹ Efforts to develop indices of total cerebrovascular disease burden in the brain by incorporating multiple lesion types have had limited success.³²

Table 2.

VASCOG-2-WSO Criteria for mild vascular cognitive impairment (VaMCI) and vascular dementia (VaD) (adapted from reference ¹⁸⁴, with permission)

Part A. Proposed criteria for MCI and dementia
Mild cognitive impairment (both A and B are necessary)   (A) Acquired decline from a documented or inferred previous level of performance in one or more cognitive domains as evidenced by the following:      a. Concerns of the person, knowledgeable informant or a clinician of mild levels of decline      b. Evidence of modest deficits on objective cognitive assessment in ≥ 1 cognitive domain   (B) The cognitive deficits are not sufficient to interfere with independence Dementia (both A and B are necessary)   (A) Evidence of substantial cognitive decline from a documented or inferred previous level of performance in one or more cognitive domains, based on:      a. Concerns of the person, a knowledgeable informant, or the clinician, of significant decline;      b. Significant deficits in objective assessment based in ≥ 1 cognitive domain.   (B) The cognitive deficits are sufficient to interfere with independence
Part B. Evidence for predominantly vascular etiology of cognitive impairment
(A) One of the following clinical features (A.1. or A.2.)   A.1. The onset of the cognitive deficits is temporally related to ≥ 1 clinical strokes.   B.2. If no history or signs of stroke, cognitive decline has more gradual onset and course, typically predominant in some combination of attention and processing speed, and/or executive functioning (B) Presence of significant neuroimaging MRI (preferable) or CT evidence of cerebrovascular disease (at least one of the following)   1. Multiple infarcts or a single extensive or strategically placed infarct   2. Multiple lacunes outside the brainstem   3. Extensive and confluent WMH   4. Multiple intracerebral hemorrhages, or one large or strategically placed hemorrhage. (C) Features that May Suggest an Alternative or Additional Etiology    1. Clinical features:    a) Insidious early onset of deficits in the absence of corresponding focal vascular lesions    b) Early and prominent movement disorder suggestive of Lewy body disease    c) Features strongly suggestive of another primary neurological disorder.    2. Absent or minimal cerebrovascular lesions on CT or MRI.    3. The presence of biomarkers of Alzheimer’s disease or other neurodegenerative disease.

Two levels of certainty are generally used for the diagnosis – possible or probable – with neuroimaging considered necessary for a probable VCID diagnosis. Neuroimaging also assists in distinguishing VCID from other neurodegenerative diseases such as AD or frontotemporal degeneration based on the patterns of atrophy, and less common causes of cognitive impairment, such as brain tumor or normal pressure hydrocephalus. Positron emission tomography has a limited role in establishing vascular etiology but can be helpful in identifying AD and other etiologies of dementia. This is discussed in more detail in the Biomarkers section (see below).

Neuropsychiatric Symptoms

Neuropsychiatric or neurobehavioral symptoms, in particular apathy, depression and agitation, are common in VCID, given that frontal-subcortical systems are often affected, and are more likely to occur in the later stages. The point prevalence of apathy in VaD has been estimated to be 33.8%, and its prevalence in post-stroke patients ranges from 22.5% to 56.7%.³³ Depression is also common in VCID, and in post-stroke cases, figures of 21.6% for major and 21% for minor depression have been cited.³⁴ The concept of vascular depression, with cerebrovascular disease as causal, has been proposed, although not fully accepted.³⁵ An additional challenge presented by major depression is that it is associated with cognitive impairment, the features of which are similar to those in VCID, thereby having an additive effect. Psychotic symptoms are common in VaD in later stages, with one review reporting that 37% of patients experienced psychotic symptoms, of which 19-50% experienced delusions, 14-60% visual hallucinations, and 19-30% delusional misidentification.³⁶

Course and prognosis

The classic description of the progression of VCID is that of an acute onset and a subsequent stepwise or fluctuating decline, with intervening stability or even some improvement.^2,37 This pattern is characteristic of VCID due to multiple strokes, in which the decline is temporally associated with cerebrovascular events of infarction, hemorrhage or vasculitis, and is at its peak soon after the event and persists beyond 3-6 months thereafter,³ although some improvement may continue up to a year or beyond.³⁸ Many patients with VCID do not demonstrate this pattern and instead show a gradual onset with slow progression,³⁹ generally attributable to cSVD with lesions in the white matter, basal ganglia, and/or thalamus. The progression in these cases may be punctuated by acute events, which leave subtle neurologic deficits.⁴⁰ Because of its heterogeneity, the long-term prognosis of VCID shows considerable variability. Risk factors for progression and functional decline include age, previous cognitive impairment, polypharmacy, hypotension during acute stroke, depression, and medial temporal atrophy. Compared with AD, those with VCID may show more prominent progression of affective symptoms such as depression.⁴¹

Longitudinal studies of white matter lesions have shown substantial progression in the majority, but with regression reported in up to one-third of post-stroke patients.⁴² In normal aging, white matter lesions increase by about 0.56 mL/year (95% CI 0.06-1.06),⁴² while in those with early confluent to confluent lesions, the rate of progression ranges from 0.23-1.33 mL/year.⁴³ In individuals with cognitive disorder, depression or vascular risk factors, the rate is higher,⁴² and it is much higher in CADASIL (Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy), a rare monogenic form of VCID. The frequency of new lacunes ranges from 1.6-19.0% over 3 to 5 years in different studies,⁴⁴ with the Dijon Stroke Registry showing an incidence of lacunes of 16.7% over 5 years.⁴⁴ The incidence of microbleeds varies strongly with age. The progression of cSVD as assessed by neuroimaging markers has not, however, been clearly correlated with the progression of cognitive deficits.

SUBTYPES OF VCID

Subcategories of Vascular Cognitive Disorders have been described, but they have been used inconsistently because of frequent overlap.

1. Post-stroke cognitive impairment or dementia: Since cognitive deficits are common and often salient in stroke patients, and stroke is estimated to nearly double the risk of dementia in people over 65,^45,46 many studies have focused on this group of patients, with post-stroke dementia comprising any incident dementia following stroke, usually manifesting within 6-months.⁴⁷ A recent review of 44 studies reported post-stroke dementia prevalence at 22.3% (18.8-39.7%), and 16.% (10.4-25.1%) if pre-stroke dementia was removed.⁴⁸ The data largely relate to ischemic stroke, and the impairment in cognition is not exclusively vascular in origin as other contributors such as AD are not excluded. Another study of population-based cohorts showed that a stroke leads to an average acute decline of 0.25 standard deviation (SD) in global cognitive sores, and accelerates the subsequent decline by 0.038 SD per year⁴⁶ (Supplemental Figure 2s). These data highlight the potential of stroke prevention and early and effective treatment for the prevention of dementia, but also the importance of cognitive rehabilitation in stroke patients.

2. Multi-infarct dementia (MID) is characterized by multiple cortical and/or subcortical infarcts and is an older term now included under the umbrella of VCID.

3. Cortico-subcortical and subcortical VCID: Most patients with VCID have a combination of cortical and subcortical vascular lesions; hence the use of the term cortico-subcortical VCID. VCID may however be exclusively attributable to subcortical vascular lesions,^29,40,49 due to small vessel disease causing multiple lacunes, white matter lesions, and non-infarct white matter changes. Subcortical VCID usually has a gradual onset and slowly progressive course, much like AD, although neuropsychological assessment reveals predominant deficits in speed of information processing and executive function, with episodic memory being relatively spared, and if impaired, not showing the pattern of rapid forgetting characteristic of AD.⁵⁰ There may be associated gait disturbance (e.g., parkinsonian or magnetic gait), incontinence, emotional lability, apathy or depression, with pseudobulbar features in the advanced stage.⁵¹ The symptoms are attributable to lesions in the prefrontal subcortical circuit (including the prefrontal cortex, caudate, pallidum, and thalamus) or thalamocortical circuit. The historical term Binswanger disease has sometimes been applied to progressive subcortical VCID but the usage of this term is controversial and it is rarely applied in contemporary practice.⁵²

RISK FACTORS

Vascular risk factors

Since cerebrovascular disease underpins the pathophysiology of VCID, its risk factors will understandably also be risk factors for VCID.⁵³ Hypertension is the leading risk factor for stroke and therefore an important contributor to VCID. Strong evidence highlights that mid-life hypertension predisposes to increased risk of both VaD⁵⁴ and AD.⁵⁵ A recent meta-analysis estimated that late-life hypertension doubled the risk of VaD (RR=2.12, 95% CI 1.50–2.99), while there was no association with AD.⁵⁶ Diabetes is strongly associated with VaD (RR=2.49, 2.09-2.97)⁵⁷ as is smoking (RR=1.78, 1.28- 2.47).⁵⁸ Mid-life, but not late-life, hyperlipidemia has been associated with elevated VaD⁵⁹ and AD risk.⁶⁰ Interestingly, high mid-life BMI (>30 kg/m² compared to normal BMI [range 20-25]) has been strongly associated with VaD (RR=4.04, 2.59-6.31) and AD (RR = 2.13, 1.75-2.59), whereas in late-life there is no association with AD and the relationship with VaD is reversed in that low BMI predicts increased VaD risk (RR=2.18, 1.18-4.02) and a higher BMI may be slightly protective (RR=0.73, 0.61-0.86).⁶¹

Cardiovascular disease

The presence of various cardiovascular diseases has been associated with increased prevalence of cognitive impairment and dementia, although attribution specifically to vascular etiology in these cases has not been generally reported. An exception is atrial fibrillation, which is a strong risk factor for stroke and VaD (OR=1.7, 1.2-2.3), with conflicting evidence whether it increases AD risk.^62,63 In congestive heart failure, overall prevalence of any cognitive impairment at 41.4% and dementia at 19.8% was reported in a systematic review.⁶⁴ The presence of coronary artery disease increases future dementia risk by 27%.⁶⁵ The prevalence of cognitive impairment in acute coronary syndrome is high, 16-20% in one systematic review.⁶⁶ Cognitive impairment is also commonly associated with valvular disease, with a prevalence of 13% in patients with severe aortic stenosis undergoing aortic valve replacement.⁶⁷ There has been considerable debate on the incidence of cognitive impairment after coronary revascularization. A recent systematic review and meta-analysis of 23 studies showed that the rates of cognitive impairment after coronary artery bypass grafting (CABG) were high, 36% (95%CI 28.2-44.5%), improved over 6 months, but were again higher after 12 months, 39.1% (21.7-58.4%).⁶⁸ The presence of hypertension and post-operative delirium were associated with worse cognition. Comparisons of different procedures, including CABG, percutaneous coronary intervention and endoscopic CABG have shown inconsistent results.⁶⁹ More research is needed in this field, but the current evidence suggests the high prevalence of cognitive impairment relates to vascular risk factors or post-procedural complications rather than the procedures per se.⁷⁰

Sociodemographic factors

Other factors influence the development of cognitive impairment in individuals with cerebrovascular disease. Age is a well-recognized major risk factor for all-cause dementia. The effect of age represents an accumulation of deficits over time and the body’s response in terms of reparative processes. In this context, the cumulative deficit model of frailty has been applied to dementia.⁷¹ While higher levels of education appear strongly protective against AD and all-cause dementia, study results have been less consistent for VaD.⁷² However, studies have generally found that low education is a predictor of post-stroke dementia.⁷³

Lifestyle factors

Late-life physical activity has been consistently associated with lower VaD risk, with a 2010 meta-analysis (24 studies, n=10,482) finding that high levels of physical activity were associated with a 38% lower risk of VaD (OR = 0.62, 0.42-0.92)⁷⁴ with more recent studies similarly supporting protective effects.⁶¹ Complex cognitive activity has been explored as a possible protective factor, and while it is clearly linked to reductions in all-cause dementia, its mechanism is likely non-specific and it has not been found to be specifically associated with reduced VaD risk.⁷⁵ Previous meta-analyses have found that poor social engagement is associated with general cognitive decline;⁷⁶ one meta-analysis found no association between VaD and loneliness.⁷⁷ However, a recent UK Biobank analysis (n= 492,322) found that loneliness was a stronger predictor of VaD (HR = 1.82, 1.62-2.03) than AD (HR = 1.40, 1.28-1.53).⁷⁸ Multiple meta-analyses have found associations between Mediterranean diets and lower all-cause dementia and AD risk,^79,80 but no meta-analysis on the association between the Mediterranean diet and VaD has been published. The largest study to date involved a Swedish cohort (n=28,025) with two decades of follow up, and found no association between the Mediterranean diet and reduced VaD risk.⁸¹ In a recent meta-analysis, there was a moderate association between ‘heavy’ versus ‘light’ alcohol use and VaD,⁶¹ a fact not reported in previous meta-analyses.⁸² Considering other dietary/metabolic factors, high homocysteine⁸³ has been shown to increase the risk of white matter hyperintensities but data are lacking to demonstrate a strong association with clinically diagnosed VCID.

Other

Of the other risk factors, depression has been associated with VaD⁶¹ (RR = 2.57, 2.16-3.07) although it seems likely that this relationship is bidirectional as stroke and cSVD are known to predispose to depression. Similarly, a recent meta-analysis⁸⁴ found that anxiety predicts both risk of VaD (OR=1.88, 1.05-3.36) and AD (HR=1.53, 1.16-2.01), although given the frequent comorbidity of anxiety and depression it is unclear if this association is driven independently by the anxiety disorder. Insomnia has been associated with a higher incidence of AD and all-cause dementia but not VaD, whereas sleep disordered breathing (frequently caused by obstructive sleep apnea) has been linked to VaD, AD and all-cause dementia.⁸⁵ Periodontal disease, chronic inflammation and chronic kidney disease have also been linked to increased risk.⁸⁶ Additional risk factors including vitamin D deficiency, sensory loss, especially hearing, and traumatic brain injury have been associated with all-cause dementia but their role in VaD has not be carefully examined.

Genetic factors

Monogenic disorders

The genetic risk of VCID has been significantly informed by monogenic cSVDs (Supplemental Table 4s).⁸⁷ These conditions affect arterial function, primarily the small cerebral vessels, and lead to the development of VCID, which often arises in midlife. While spontaneous cerebrovascular disruption occurs frequently, monogenic cSVDs are comparatively rare, with the most common, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), estimated to have a minimal prevalence of 2–5 individuals per 100,000 globally.⁸⁸ CADASIL is caused by primarily cysteine-altering pathogenic variants in the NOTCH3 gene, resulting in common symptoms of migraine with aura, recurrent subcortical strokes, progressive cognitive decline and mood disturbance.⁸⁹ Other monogenic causes of VCID⁹⁰ are listed in Supplemental Table 4s and include CARASIL (cerebral autosomal recessive arteriopathy with subcortical infarcts and leukoencephalopathy), Fabry disease, cathepsin-A related arteriopathy strokes and leukoencephalopathy (CARASAL), Mitochondrial Encephalopathy, Lactic acidosis and Stroke-like episodes (MELAS), retinal vasculopathy with cerebral leukoencephalopathy and systemic manifestations (RVCL-S), and familial cerebral amyloid angiopathy (FCAA).

Polygenic associations

The small number of genome wide association studies (GWAS) of VCID have resulted in limited insight into the genetic polymorphisms of VCID.⁹¹ A GWAS meta-analysis has suggested several variants associated with VCID.⁹⁰ These included APOE4, the major risk gene for sporadic AD, which has a range of functions involving amyloid, neurodegeneration, and cerebrovascular neuropathology, and novel loci (SPRY2, FOXA2, AJAP1, and PSMA3) associated with vascular risk including hypertension and diabetes, and neuronal maintenance. An additional meta-analysis has suggested associations of VCID with MTHFR, a homocysteine regulator, PON1, which has antioxidant function, TGF-β1 and TNF-α.⁹¹ Given stroke and cSVD are major causes of VCID, GWAS in these fields can be extrapolated to understand the genetic risk of VCID. Meta-analytic findings have suggested over 80 risk loci for stroke, with common cross-ancestry susceptibility.^92,93 A large-scale GWAS of neuroimaging features related to cSVD including WMH, fractional anisotropy and mean diffusivity suggested 31 associated loci.⁹⁴ Table 3 summarizes the risk factors for VCID, some of which are shared with AD (Supplementary Table 5s).

Table 3. Risk factors for Vascular Cognitive Impairment and Dementia.

Cardiologists must pay special attention to patients with some cardiac disorders and those with vascular risk factors.

Category	Risk factor	Strength of evidence	Relevance for cardiovascular clinician
Demographic	Age	++	Enquire about cognitive concerns in all older patients
	Sex	+/−
	Education	+/−
Genetic	Autosomal dominant	+ (uncommon)	Suspect genetic cause if young-onset stroke, migraine with aura, strong family history
	Apolipoprotein E ε4	+
	Polygenic risk	+
Lifestyle factors	Physical inactivity	++	Encourage regular exercise (aerobic and resistance)
	Lack of complex cognitive activity	-	Encourage cognitively challenging activities
	Social isolation/Loneliness	+/−	Engage social worker; social prescribing
	Sub-optimal diet	+/−	Encourage balanced diet; referral to dietician
	Alcohol use	-	Encourage moderation
Vascular risk factors	Mid-life hypertension	++	Detect early; treat according to guidelines 185
	Late-life hypertension	+/−	Treat according to guidelines
	Diabetes/hyperglycemia	++	Treat adequately; involve endocrinologist; prevention (weight, diet, drugs)
	Dyslipidemia	++	Diet and drugs
	Smoking	++	Recommend cessation
	Stroke/TIA	++	Active physical and cognitive rehabilitation; prevent relapse
	Chronic kidney disease	+	Liaise with nephrologist
	Mid-life BMI	+	Manage early
	Late-life BMI	+	Manage early
Cardiovascular disease	Atrial fibrillation	++	Treat early
	Congestive heart failure	++	Treat as per guidelines 186
	Coronary artery disease	++	Treat as per guidelines 187
	Acute coronary syndrome	++	Treat as per guidelines 188
	Valvular disease (e.g., aortic stenosis)	++	Treat as per guidelines 189
Other disorders	Frailty	+	Liaise with geriatrician/endocrinologist
	Depression	+	Detect early; liaise with psychiatrist/psychologist
	Anxiety	+/−	Detect early; liaise with psychologist/psychiatrist
	Systemic inflammation	+	Liaise with geriatrician
	Sleep disturbance	+/−	Advise sleep hygiene
	Vitamin D Deficiency	+/−	Recommend supplementation
	Sensory loss	+/−	Recommend correction
	TBI	+/−	Advise preventive measures

AD: Alzheimer’s disease; BMI: Body mass index; TBI: Traumatic brain injury; VaD: Vascular dementia. (++ established and strong; + established; +/− conflicting or insufficient data; - no association)

PATHOPHYSIOLOGY

VCID results from abnormalities in large and small blood vessels supplying the brain, leading to ischemic (acute or chronic) or hemorrhagic injury, or due to hypoperfusion from other causes, such as cardiac failure or disease. These lead to cumulative tissue damage, disruption of neural networks and cognitive impairment.⁷⁷ Processes such as endothelial cell dysfunction, disruption of the neurovascular unit (NVU) with abnormal blood-brain barrier permeability, neuroinflammation and oxidative stress are involved. The vulnerability of the brain to hypoxic damage stems from its high energy demand and lack of fuel reserves.⁴⁷ Brain hypoxia may result from cerebral large or small vessel disease as well as other causes.

Large vessel disease

Disease in the anterior cerebral circulation, which supplies about 70% of the brain, or the posterior circulation which supplies the remainder, leads to stroke, TIAs or chronic cerebral hypoperfusion. To appreciate the mechanisms underlying the cerebral injury that resulting from large vessel disease-induced ischemia or hemorrhage, it is necessary to understand the etiology and pathophysiology of stroke.⁹⁵ Most strokes are ischemic in origin and result from arterial blockage due to thromboembolism, with the embolus originating from an atherosclerotic plaque in a large artery or in the heart due to cardiac disease such as AF. Arteriosclerotic disease of small cerebral vessels leads to lacunar strokes, and there are other less common causes of stroke (see below).⁹⁶ About 20% of strokes are hemorrhagic, either subarachnoid or intracerebral, which could be traumatic or non-traumatic, the latter related to cerebral aneurysm, vascular malformation, tumor, cerebral amyloid angiopathy, uncontrolled hypertension, anticoagulant use and coagulopathy.⁹⁷ The interruption of blood supply leads to a necrotic core and an ischemic penumbra which sustains viability for a variable period of time. The site and extent of brain damage is determined by the arterial territory affected, the extent of collateral blood flow, timely interventions and rehabilitation. The pathophysiology of cerebral injury due to reduced blood flow is discussed below, with endothelial cell damage, breakage of the blood brain barrier (BBB), inflammation, oxidative stress and excitotoxicity all playing a role. The pathophysiological mechanisms of TIA are similar to stroke, with reversibility of clinical symptoms not always associated with complete reversal of any tissue injury.⁹⁸ While many of the downstream effects of cerebral hemorrhage are shared with ischemia, it also produces tissue injury due to its mass effect, edema and inflammation, and toxic biochemical and metabolic effects of the clot’s constituent parts.⁹⁸

Cerebral small vessel disease (cSVD)

Pathologies affecting the small perforating arteries, arterioles, capillaries and venules are major VCID contributors. These pathologies are multiple, as described below, but most commonly it is arteriolosclerosis related to vascular risk factors such as hypertension, diabetes, etc. or cerebral amyloid angiopathy (CAA) due to age-related deposition of β-amyloid in cerebral blood vessels. cSVD is generally sporadic and multifactorial, although monogenic abnormalities are described in <5% of cases (see above). Arteriolosclerosis leads to stiffening of the arterial wall, impaired vasodilatory capacity, and progressive narrowing or occlusion of the lumen, resulting in reduced cerebral blood flow (CBF), lacunar infarction, or chronic hypoperfusion. Reduced autoregulation increases vulnerability to low or fluctuating blood pressure, predisposing to ischemic injury. cSVD also disrupts endothelial cell function, BBB integrity and neurovascular coupling. It also compromises the glymphatic system and the dynamic flow of interstitial fluid in the brain. cSVD is therefore the most prominent pathomechanism for VCID. Its presence is recognized on MRI by characteristic features, in particular white matter hyperintensities, lacunes, cerebral microbleeds and dilated perivascular spaces, which will be discussed later.

Other causes of hypoxic injury

Cerebral hypoxic injury may result from several cardiac disorders listed above, in particular embolism from AF or patent foramen ovale with paradoxical embolism, as well as aortic dissection, cerebral vasculitis and hematological disorders.

We now describe the downstream effects of cerebral hypoperfusion and hypoxia.

The neurovascular unit (NVU)

CBF is tightly regulated in tandem with arterial pressure to meet the metabolic needs of the brain. Moreover, the metabolic needs of any brain region change with its level of activity, and these varying energy needs are met through a close structural and functional relationship between brain tissue and the brain vasculature, the basic element of which is the NVU.⁹⁹ A drop in cerebral blood flow to <18-20 ml/100g/min results in ischemia and <8-10 ml/100g/min in tissue death, with ischemia triggering a cascade of biochemical changes that can damage neurons. Chronic cerebral hypoperfusion, resulting from carotid artery stenosis, hypotension, or heart failure, can cause ischemic injury to the white matter and deep gray matter structures, important neural substrates of cognition.

The NVU (Supplemental Figure 2s) is present at the microvasculature level and comprises cells from the vessel wall (pericytes, endothelial cells and smooth muscle cells), glia (astrocytes, microglia and oligodendrocytes) and neurons.¹⁰⁰ The pericytes and the endothelial cells, with their tight junctions, maintain the integrity of the blood-brain barrier (BBB), which is important to maintain the internal milieu of the brain for optimal function. The NVU is important to maintain the hemodynamic response, and the cells of the NVU also have roles in BBB permeability, inflammation, angiogenesis, and the afferent and efferent arms of the immune system in the brain. It is becoming more evident that the NVU interacts bidirectionally and dynamically with systemic biology,¹⁰¹ which has an important impact on NVU integrity and brain health.

Endothelial dysfunction and BBB integrity

Endothelial dysfunction, brought about by exposure to different vascular risk factors (e.g., hypertension, diabetes, or smoking) leads to abnormalities in the production or bioavailability of endothelial-derived paracrine factors (nitric oxide, prostaglandins, etc.), important for smooth muscle vasomotion and platelet activation. This causes deleterious changes in vascular reactivity, including alterations in the regulation of local vascular tone and redox balance, hemostasis and thrombosis, and inflammation in arterial walls.²⁵ The structural and functional integrity of the cerebral microcirculation is further affected by hypertension, fostering cerebral microvascular endothelial dysfunction and neurovascular uncoupling, leading to brain hypoperfusion. It also disrupts the BBB, causing neuroinflammation and enhancing amyloid pathology.²⁶ Indeed, the dysfunction of the BBB due to chronic hypertension, inflammation and cSVD allows the entry of inflammatory cells, neurotoxic substances and plasma proteins, contributing to further neuroinflammation and neuronal damage. Although cerebral endothelial cells are ultimately responsible for BBB integrity, the development and maintenance of its barrier properties also depend on the interaction with other vascular-associated cells, in particular, astrocytes, pericytes and smooth muscle cells. The dysfunction of any or all of these cell types, and others, may contribute to VCID.²⁷ In humans, cerebral microvessel endothelial dysfunction manifestations include BBB dysfunction, impaired vasodilation, vessel stiffening, dysfunctional blood flow and interstitial fluid drainage, white matter rarefaction, ischemia, inflammation, myelin damage and also secondary neurodegeneration, which collectively lead to brain damage in cSVD (see below).²⁰

Chronic inflammation, oxidative stress and excitotoxic mechanisms

Cardiovascular risk factors, sub-optimal lifestyle, and aging are associated with a chronic low-grade, systemic inflammatory status (inflammaging”), with multiple common cellular and molecular mechanisms leading to vascular pathology (endothelial dysfunction, arterial stiffness, vasoreactivity alterations, hypoperfusion, BBB breakdown) including development of atherosclerosis, immune-thrombosis and alterations in hematopoietic function.^{102 103–105} These induce the release of proinflammatory cytokines and the excessive production of reactive oxygen and nitrogen species that, at the brain level, can damage neuronal and glial cells as well as cellular components, ultimately impairing neuronal function and viability. In this manner, chronic inflammation is likely to act as a driver of brain health-to-disease transition, with cognitive impairment a plausible end-result.

The brain’s fluid transport and clearance system

The intense metabolic activity of the brain generates cellular debris and waste which must be cleared for homeostasis to be maintained and efficient functioning to continue.¹⁰⁶ However, unlike other organs, the brain lacks a traditional lymphatic system. Several mechanisms have been proposed for waste clearance which include uptake by microglia and astrocytes, enzymatic degradation, transcellular transport through the blood–brain barrier or the blood–CSF barrier, and a convective transport by interstitial fluid exchanging with cerebrospinal fluid.¹⁰⁷ Two relatively recent discoveries must be mentioned in this regard. The first is the discovery of the glial-lymphatic or glymphatic system, where the cerebrospinal fluid from the subarachnoid space runs into the deep brain along the perivascular spaces between the arterial basement membrane and the glia limitans, where it is propelled through the glia limitans end-feet to encounter the interstitial fluid and the waste products across the parenchyma. This movement is pushed by arteriole pulsatility and cerebrospinal fluid pressure, and depends on water-channel aquaporin-4 (AQP4) expressed in the astrocytic end-feet. Along this stream, waste products and fluids are driven towards the perivenous spaces and other pathways, exiting the brain via meningeal and cervical lymphatic vessels along cranial and spinal nerves, towards the general circulation.²¹ The second discovery, that of a bona fide lymphatic system in the dura mater, supports this proposed mechanism.¹⁰⁸ Although physically detached, the glymphatic and meningeal lymphatic systems are functionally connected and together constitute an interconnected glymphatic-lymphatic fluid transport system.²² A role for meningeal lymphatics in AD for removal of protein aggregates of β-amyloid and tau has also been shown by recent studies.²⁹ However, this has been disputed in other recent work,¹⁰⁹ and the further research is needed to fully understand the brain’s clearance mechanisms¹¹⁰ In the context of VCID, it is important to state that widening of perivascular spaces, a biomarker of cSVD, has been proposed to be associated with the presence of perivascular cell debris and other waste products that form part of a vicious cycle involving impaired cerebrovascular reactivity, BBB dysfunction, perivascular inflammation and ultimately compromised clearance of waste proteins from the interstitial fluid space, leading to accumulation of toxins, hypoxia, and tissue damage.²⁸ All these observations together suggest that a defective brain drainage could be a common pathway in VCID and AD, as well as some other neurodegenerative diseases, even though the exact mechanisms remain to be conclusively determined.

Altered neurotransmitter systems

VCID is associated with disruptions in cholinergic, serotonergic and dopaminergic neurotransmitter pathways, crucial for cognitive processes such as attention, memory, and executive function as well as mood.³⁰ The greater focus has been on the cholinergic system, with the hypothesis that vascular lesions affecting the brain stem cholinergic nuclei and their white matter projection compromise cholinergic activity¹¹¹ Cholinesterase inhibitors such as donepezil, galantamine and rivastigmine, commonly used for AD, have been shown to have a beneficial effect in VaD, albeit modest and with doubtful clinical significance.¹¹²

Recent developments using “omics”

Single-cell transcriptomics, proteomics, and other omics approaches have transformed VCID research by mapping the diverse cell types of the brain vasculature, identifying disease-associated endothelial, pericyte, and microglial states, and linking genetic risk variants to specific vascular and immune pathways.^113,114 Proteomic and metabolomic studies have uncovered circulating and CSF biomarkers (e.g., BBB breakdown proteins, complement, lipid signatures) that correlate with small vessel disease burden and cognitive decline, offering new targets for diagnosis and therapy.¹¹⁵ Together, these advances have the potential to shift VCID from a “radiological” entity to a molecularly defined disease process.

NEUROPATHOLOGICAL FEATURES

Spectrum of cerebrovascular pathology in VCID

Few studies have recorded the entire spectrum of ischemic, edematous and hemorrhagic lesions (Figure 1 and Table 6s) induced by pathological changes in the brain circulation or perfusion and consequences in the parenchyma that are to be associated with VCID.³² In different studies where VCID was diagnosed clinically, 78% of the cases revealed cortical and subcortical infarcts suggesting other vascular pathologies involving incomplete infarction or borderzone infarcts could be contributory factors. Among other lesions, 25% of cases had cystic infarcts whereas 50% showed lacunar infarcts or microinfarcts. Severe CAA was present in 10%. VCID is also associated with leukoencephalopathy, large infarcts, lacunar infarcts and higher vascular burden (combined macroscopic score).³³ Subcortical macroscopic infarcts³⁴ and thalamic lacunar infarcts appear to be important predictors of VCID.³⁵

Figure 1: Neuropathological substrates of VCID.

A-C, Coronal blocks from VCID diagnosed subjects showing A) recent large cortical infarct (white arrow), B, lacunes in the Putamen (black arrow) and C) ~1cm intracerebral hemorrhage (white arrow) and a lacune (remote cavity, black arrow). D-I, Pathological changes in the cerebral vasculature. D, intracranial atheromatous disease in a branch of the middle cerebral artery (asterisk shows wall of atheroma), E, fibrinoid necrosis in intracerebral arteriole, F, moderately hyalinized arteriole still retaining some vascular smooth muscle cell nuclei, G, hyalinized vessel in rarefied white matter, H-I, cerebral vessels with CAA (H, H&E; I, antibodies to β-amyloid), Inset in H show hemosiderin surrogate of microbleed. J-L, Parenchymal changes associated with vessel occlusion. J, lacunar infarct in the putamen with clear edge (black arrow), K, infarcted tissue with severe necrosis, and L, a microinfarct in the basal ganglia. Magnification bar =100μm, applies to panels E, F, G, H, I, J, K and L. [Images provided by RK]. Abbreviations: CAA: Cerebral amyloid angiopathy; VCID: Vascular cognitive impairment and dementia.

Microvascular infarcts (lacunar infarcts and microinfarcts) appear central to the most common cause of VCID and predict poor outcomes in the elderly. Cortical microinfarcts associated with severe CAA may be the primary pathological substrate in a significant proportion of VCID cases.³⁶ It is proposed that changes in hemodynamics, e.g. hypotension and atherosclerosis, may play a role in the genesis of cortical watershed microinfarcts. Greater putamen hemosiderin was significantly associated with indices of small vessel ischemia including microinfarcts, arteriolosclerosis, enlarged perivascular spaces, and lacunes in any brain region, but not with large vessel disease or whole brain measures of neurodegenerative pathology.³⁸

Hippocampal sclerosis and cell atrophy, which may be caused by remote ischemic injury, was apparent in 55% of cases in one study with a clinical diagnosis of ischemic VCID. Hippocampal neurons in the Sommer’s sector are highly vulnerable to disturbances in the cerebral circulation or hypoxia caused by systemic vascular disease. Focal loss of CA1 neurons in ischemic VCID, post-stroke dementia and mixed dementia have been related to lower hippocampal volumes and memory scores,^39,40 not explained by the presence of any neurodegenerative pathology. A vascular basis of hippocampal degeneration and atrophy is supported by VCID and post-stroke dementia cohorts and population-based studies.^41,42 The intensity of astrocytic gliosis and microgliosis, is an important consideration in judging the degree and age of infarction. Whilst borderzone or watershed infarcts and laminar necrosis are also found in cases with VCID but there is no clear evidence to what extent they contribute to cognitive impairment.

Diffuse and focal white matter lesions are a hallmark of VCID but also occur in approximately 30% of patients with AD or dementia with Lewy bodies.⁴³ Cognitive impairment is associated with diffuse white matter demyelination,⁴⁴ periventricular demyelination ⁴⁴, and arteriolosclerotic small vessel disease.^46,47 While white matter changes focus on the arterial system, narrowing and, in many cases, occlusion of veins and venules by collagenous thickening of the vessel walls also occur. The thickening of the walls of periventricular veins and venules by collagen (collagenosis) increases with age, and perivenous collagenosis is increased further in brains with leukoaraiosis.

In terms of actual vessel changes, ~15% of VCID cases involve occlusion of large extracranial arteries such as the internal carotid artery and the main intracranial arteries of the circle of Willis, including the middle cerebral artery, leading to multiple-infarcts and dementia. Typical atherosclerosis or microatheromatous disease in the meningeal and smaller vessels, beyond the circle of Willis involving proximal segments of the middle and anterior cerebral arteries, is generally rare but may be found in very old subjects.⁴⁹ cSVD pathology in terms of arteriolosclerosis comprising fibroid necrosis and hyalinization, enlarged perivascular spaces and white matter attenuation both in the periventricular and deep regions, with associated astrocytic gliosis, is probably the most common feature and is present in the majority of VCID cases. Variable degrees of inflammation may also be present, shown by the presence of perivascular lymphocytes or macrophages, not necessarily a function of brain ischemia. In the oldest cSVD subjects, there often is evidence of remote hemorrhage in the form of perivascular hemosiderin.⁴⁹ Perivascular edema and thickening, inflammation and disintegration of the arteriolar wall are common, whereas vessel occlusion is rare.⁵⁰ Small cerebral arteries and arterioles are also affected in many other diseases including hereditary angiopathies, inflammatory and infective vasculitides and toxic disorders, all of which can cause irreversible cognitive impairment.

Neuropathological criteria

Attempts have been made to develop neuropathological criteria for the diagnosis of clinically defined VCID. The Oxford Project to Investigate Memory and Ageing study developed a simple, novel, image-matching scoring system⁴⁶ relating the extent of cSVD to cognitive function in 70 cases with insufficient pathology to meet criteria for an AD diagnosis. Severity of cSVD pathology was inversely related to cognitive scores and 43% of the cases with high cSVD scores were designated to have dementia. A staging system related to the natural history of cerebrovascular pathology and an algorithm for the neuropathological quantification of the cerebrovascular disease burden in dementia was also proposed.⁴⁹ While the staging system requires further validation against cognitive function scores, it can be used in large-scale studies to understand clinicopathological relationships even in post-stroke survivors who fulfil the NINDS-AIREN criteria¹² for probable VaD. According to the Newcastle scheme,⁶⁰ there are two neuropathological diagnostic groups: probable VaD is based on the exclusion of a primary neurodegenerative disease known to cause dementia plus the presence of cerebrovascular pathology that defines one or more of the VaD subtypes; possible VaD is designated when the brain contains vascular pathology that does not fulfil the criteria for one of the subtypes but where no other explanation for dementia is found.

The Vascular Cognitive Impairment Neuropathology Guidelines (VCING) were developed to reach a consensus for scoring cerebrovascular disease in relation to VCID.⁵¹ It was concluded that a combination of three main determinants - moderate/severe occipital leptomeningeal CAA, at least one large infarct (>4cm), and moderate/severe arteriolosclerosis in the occipital white matter – could be used to assign low, intermediate or high likelihood that cerebrovascular disease contributed to cognitive impairment in an individual case. The presence of 0, 1, 2 or 3 of these features resulted in predicted probabilities of VCID of 16%, 43%, 73% or 95%, respectively. Thus, arteriolosclerosis of white matter is an important determinant of VCID. However, the VCING collaboration also found that pathologies associated with VCID include microinfarcts, lacunar infarcts, large infarcts, arteriolosclerosis, perivascular space dilation, myelin loss and leptomeningeal CAA. Microinfarcts were found in all brain regions and in agreement with a systematic review⁵² those in the parietal cortex and putamen were predictive of VCID.⁵¹

BIOMARKERS OF VCID

Neuroimaging markers are currently the mainstay for the establishment of CeVD as the likely basis of cognitive impairment, an important step in the diagnosis of VCID. Both computed tomography (CT) and magnetic resonance imaging (MRI) are used, although MRI offers advantages in terms of sensitivity, range of pathologies being identified, and the derivation of functional as well as structural information. Neuroimaging also helps differentiate VCID from neurodegenerative diseases. The characteristic pattern on MRI in AD is generalized brain atrophy, with emphasis on the parietal and temporal cortices, and particularly the medial temporal lobes, including the hippocampus and entorhinal cortex. Dementia with Lewy Bodies (DLB), on the other hand has a hippocampal-sparing pattern of atrophy, with the posterior parietal and occipital cortices, and subcortical structures such as dorsal midbrain and substantia innominata differentially affected. In frontotemporal dementia (FTD), the emphasis is on atrophy in the frontal and temporal lobes, often with asymmetrical involvement.

Functional imaging using F¹⁸-deoxyglucose positron emission tomography (FDG PET) typically shows scattered areas of hypometabolism in cortical and subcortical regions in VCID, unlike the deficits in bilateral parietotemporal association cortices, the posterior cingulate and the precuneus in early AD, the parieto-occipital hypometabolism with relative preservation in the posterior cingulate in DLB, and the often asymmetrical involvement of the frontal and anterior temporal lobes. Similar patterns of reduced regional CBF are seen on single photon emission computed tomography (SPECT), albeit with less sensitivity than FDG PET. Molecular PET imaging is increasingly used to diagnose AD, in particular amyloid and tau PET which help visualize the specific pathologies of plaques and tangle in the AD brain.¹¹⁶

Plasma and CSF measurements of AD pathophysiological markers have emerged as accepted biomarkers of AD; these include Aβ42/40, Aβ40 and Aβ42, and tau, both phosphorylated tau (p-tau) 217, 181 and 231 and total tau.¹¹⁷ This is in contrast with VCID, for which molecular biomarkers are still in development and not ready for clinical use. CSF or plasma pTau217, pTau181, Aβ42/40 ratio and pTau217/ Aβ42 are increasingly used to establish the presence of AD-specific amyloid pathology in the brain. Neurofilament light chain (NfL) in CSF or plasma is a useful marker of neuroaxonal injury and can be reliably measured, but it is a non-specific marker of neurodegeneration and is raised in VCID, AD and other neurodegenerative diseases.^118–120 Glial fibrillary acidic protein (GFAP) in the CSF or blood is a marker of neuroinflammation and astrocytic activation and is raised in both AD and VCID, although it differentially reflects amyloid pathology.¹²¹ In amyloid-negative patients, GFAP can help discriminate between VCID and non-vascular cognitive impairment.¹²² The integration of neuroimaging with CSF and blood biomarkers is currently being examined to increase the accuracy of disease discrimination.¹²³

Ocular and retinal biomarkers of VCID and AD are in development and more research and longitudinal studies are needed before these can be adopted for clinical practice.^124–127 The biomarkers of VCID are summarized in table 4, with the key biomarkers highlighted. The steps needed to differentiate it from AD, or to indicate the presence of both pathologies in mixed dementia, are indicated in Box 2. A more detailed description of some important biomarkers is given below.

Table 4. Biomarkers of VCID.

T1-weighted (T1w) and T2-weighted fluid attenuated inversion recovery (FLAIR) MRI are the two commonly used protocols for routine assessment, with susceptibility-weighted imaging (SWI) added for cerebral microbleeds. Other protocols are for special clinical situations or for research. Fluid and retinal biomarkers for VCID are mainly in the research domain. Positron emission tomography (PET) is primarily for differential diagnosis.

Neuroimaging	Comment
Large vessel disease
Infarct/hemorrhage	Non-contrast CT or MRI
Arterial health	CT/MRI angiography; perfusion imaging; digital subtraction angiography; intracranial vessel wall MRI
Blood flow and pulsatility	Transcranial doppler
Small vessel disease	MRI preferred
WMHs	T2w FLAIR MRI (best); CT/T1w MRI; rated visually on Fazekas scale 0-3 (2 or 3 suggests high burden)
Lacunes	T2w FLAIR MRI; number and location important
Dilated perivascular spaces	T2w FLAIR MRI; T1-weighted MRI
Cerebral microbleeds (CMB)	SWI (best); T2*w GRE imaging; multiple CMB (≥3) significant for cognition
Cerebral microinfarcts	Very high field scanner (best)
Brain atrophy	Whole brain; corpus callosum, central white matter and mid-brain; hippocampal atrophy can occur
Diffusion weighted MRI	Mean diffusivity (MD); fractional anisotropy (FA); peak skeletonized mean diffusivity (PSMD); differences in distribution functions (DDF)
	Free water imaging
Functional imaging
Cerebral blood flow	Contrast perfusion MRI (best); arterial spin labelling MRI (alternative); CT perfusion; transcranial doppler; O15-water PET
Cerebrovascular reactivity	Change in CBF using CO2 challenge
BBB integrity	Dynamic contrast enhanced MRI
PET
18F-deoxy glucose (FDG)	For differential diagnosis from other dementias
Amyloid PET	To rule AD in or out
Fluid biomarkers 1	Still in development
CSF	MMP-2, MMP-9, IL-6, VEGF, PlGF, NfL, GFAP, Lipocalin 2
Blood proteins	ICAM-1, VCAM-1, E-selectin, ADMA, PlGF, MR-proADM, MMP-3, Lp-PLA2, MDA, IL-6, IL-1β, fibrinogen, NfL, GFAP
Circulating MicroRNAs	miR-130b-3p, miR-10b*, miR-146a, miR-222, miR-29a
Exosomal MicroRNAs	exo-miR-154-5p, exo-miR-223-3p
Ocular and retinal	Still in development
Retinal vessel density	Using optical coherence tomography (OCT) angiography
Retinal vessel caliber	Fundus imaging
Retinal structural layers	Retinal nerve fiber layer, ganglion cell-inner plexiform layer (using OCT)
Visual function	Visual acuity, color vision

Neuroimaging Biomarkers

Structural imaging

Size of an infarct or hemorrhage and their brain location are both important, and either non-contrast CT or MRI is sufficient for defining these features. For large arterial health, CT/MRI angiography and perfusion imaging are commonly applied to assess luminal compromise due to atheroma, thrombus, or stenosis, and aneurysms or arterio-venous malformations.¹²⁸ Digital subtraction angiography visualizes vascular structures free of the surrounding tissue, to identify aneurysms, arterio-venous malformations, carotid stenosis and other similar pathologies.¹²⁹ Transcranial doppler is used to image the major intracranial blood vessels and monitor cerebral blood flow (CBF) and vessel pulsatility with good temporal resolution.¹³⁰ Intracranial vessel wall MRI is a newer technique used alongside conventional cerebral angiographic techniques to differentiate the various causes of arterial narrowing and to identify symptomatic non-stenotic disease of intracranial arteries.¹³¹

cSVD includes a range of pathological processes affecting cerebral small arteries, arterioles, venules, and capillaries. Using MRI, several imaging abnormalities that signify cSVD have been studied, which include white matter hyperintensities (WMHs), lacunes, dilated perivascular spaces, cerebral microbleeds, cortical microinfarcts and cortical superficial siderosis. These are usefully described in the Standards for Reporting Vascular changes in Neuroimaging (STRIVE) criteria^14,132 which have helped standardize terminology, definitions, image acquisition and analysis, and reporting standards for imaging biomarkers of cSVD (Figure 2). The salient features of the MRI markers of cSVD are summarized in Table 4. Recommendations for MRI protocols can be found in Supplemental Table 7s.

Figure 2: MRI-based neuroimaging markers of Vascular Cognitive Impairment and Dementia.

Abbreviations: CBF, cerebral blood flow (arterial spin labelling); CMB, cortical microbleeds; CVR, cerebrovascular reactivity; FA, fractional anisotropy (on diffusion imaging); MD, mean diffusivity; Network, brain networks using functional MRI or diffusion tensor imaging; PVS, perivascular spaces; WMH, white matter hyperintensities.

Of the various imaging markers of cSVD, WMHs, seen as bright (hyperintense) on T2-weighted MRI and dark (hypointense) on T1-weighted MRI or CT images^14,133 and have often been referred to as white matter lesions or leukoaraiosis, with WMH now the preferred term.¹⁴ While WMHs can result from several etiologies such as inflammation, demyelination, edema and axonal loss, there is converging evidence that the WMHs seen in middle and older age are largely caused by ischemia dues to cSVD.²⁰ When mild, they are often dismissed as “normal for age”, but there is persuasive evidence that they should not be overlooked. While researchers tend to measure the volume of WMH and normalize that to brain or intracranial volume,¹³⁴ a simple and clinician-friendly approach is to rate them using the Fazekas scale – absent (0), punctate foci or pencil-thin lining (1), beginning confluence or smooth ‘halo’ (2), and large confluent areas of irregular periventricular signal (3). High burden (Fazekas 2 or 3) is associated with a 2-fold increased dementia risk and a 3-fold stroke risk.(^118,119) (Supplemental Figure 3s).

Lacunes are round or ovoid, subcortical, fluid-filled (cerebrospinal fluid; CSF-like) cavities, 3-15 mm in diameter, and are consistent with a previous small deep cerebral infarct or hemorrhage in a perforating arteriole territory.¹⁴ Cross-sectional and longitudinal studies have associated lacunes with lower performance in executive function and processing speed.^135,136 Strategically located lacunes, e.g., in the thalamus or striatum, relate with lower cognitive performance, independent of other markers of cSVD.^135,137 The significance of the number of lacunes has also been examined, with a recent meta-analysis suggesting the presence of three or more lacunes as a risk factor for post-stroke dementia.¹³⁸

Perivascular spaces (PVS) (or Virchow-Robin spaces) are commonly microscopic, becoming apparent on MRI when enlarged (dilated PVS), generally <3 mm in diameter, when they are markers of pathology. A systematic review of VaD studies showed an association between dementia and increased dPVS, with inconsistent results in studies of AD.¹³⁹

Cerebral microbleeds (CMBs) are defined as generally 2-5 mm in diameter (but up to 10 mm) areas of signal void with associated blooming using MRI sequences sensitive to susceptibility effects.¹⁴ They are best visualized on susceptibility-weighted MRI compared to the conventional T2*-weighted gradient-recalled echo imaging often used to image them. The two main pathological causes of CMBs appear to be geographically distinct: cerebral amyloid angiopathy generally results in lobar CMBs, and arteriolosclerosis, often associated with hypertension, typically results in deep brain CMBs. CMBs may also occur due to traumatic brain injury, radiation injury, infective endocarditis and other causes.¹⁴⁰ A systematic review did not find a relationship between CMBs and incident dementia or AD.¹⁴¹ This may be because of the fact that the relationship between CMB and cognition is significant only in individuals with multiple CMBs, this being ≥3 CMBs in one study.¹⁴²

Cerebral microinfarcts are brain tissue infarctions that are microscopic at neuropathological examinations and not readily visible on gross examination or conventional MRI. They are common in older people at autopsy, with rates of 16-42%,¹⁴³ and may be visualized using 7T MRI, although some studies using 3T scanners,¹⁴⁴ have reported detecting larger microinfarcts of up to 5 mm.^52,143 Diffusion-weighted imaging is useful in identifying acute cerebral microinfarcts.¹⁴³

Both cross-sectional and longitudinal studies have shown a positive relationship between cSVD biomarkers and brain atrophy,¹⁴⁵ both whole brain atrophy as well as atrophy in the corpus callosum, central white matter and mid brain.¹⁴⁶. Mediotemporal and hippocampal atrophy, once considered exclusive to AD, can also result exclusively from vascular pathology.^147,148

Diffusion weighted MRI

Studies have shown a significant relationship between diffusion tensor imaging (DTI) measures and cognition in cSVD.¹⁴⁹ Computational biomarkers based on DTI measures, such as peak skeletonized mean diffusivity¹⁵⁰ and differences in distribution functions,¹⁵¹ have shown excellent sensitivity and specificity to cSVD. Through modelling an additional free-water compartment in diffusion-weighted MRI data, studies have shown that mean free water is a sensitive biomarker of cognitive decline in cSVD.^152,153 Longitudinal studies have also linked diffusion-weighted MRI-based structural network disruption with cognitive decline and mortality in cSVD.^154,155 These techniques are yet to find their way to clinic practice.

Quantifying burden of disease with structural imaging

Visual rating scales have been applied to assess WMH,¹⁵⁶ dPVS,^157,158 CMBs,^159,160 and cerebral microinfarcts.¹⁴³ Advances in computer vision and artificial intelligence have enabled computer-aided detection and segmentation,¹⁶¹ but generalizable and reliable tools for automated segmentation of small cSVD lesions (e.g. dPVS, CMBs and lacunes) are still unavailable. Attempts to construct composite cSVD indices by considering all available cSVD lesions on MRI¹⁶² are yet to determine whether composite indices outperform counts of a single lesion type, especially WMHs.¹⁵⁶

Functional imaging

Functional imaging techniques are generally used to assess cerebrovascular abnormalities preceding lesions and the alterations in functional connectivity once such lesions have occurred. Their clinical application is, however, limited. Reduced cerebral blood flow (CBF) has been associated with cognitive decline.¹⁶³ Techniques available include O¹⁵-water PET, dynamic susceptibility contrast perfusion MRI, SPECT, CT perfusion, and transcranial Doppler ultrasound, but most recent work in VCID has used arterial spin labelling MRI.^164,165 Cerebrovascular reactivity, which quantifies the ratio of the change in CBF to the change in vasodilatory stimulus such as carbon dioxide, is emerging as a useful biomarker. Dynamic contrast enhanced MRI is the most widely used approach to image BBB integrity, with other methods including PET and SPECT.¹⁶⁶ The brain’s functional connectivity is disturbed by vascular pathology, and this can be studied using resting state functional networks.¹⁶⁷ Newer techniques include the measurement of blood flow velocity and pulsatility of perforating vessels, myelin water imaging,¹⁶⁸ and oxygen extraction fraction. These techniques remain mostly under investigation. Ultra-high field MRI (e.g. 7 Tesla) and the application of machine learning and artificial intelligence are other developments being evaluated.

Neuroimaging markers – conclusion

In clinical practice, structural MRI is the most commonly used technique, with a focus on large infarcts, WMH, lacunes, CMBs and dPVS, most frequently with visual reads of the images. The recommended protocol should use, at a minimum, T1-weighted and T2-weighted FLAIR of the whole brain. For a more thorough examination, diffusion-weighted and susceptibility-weighted sequences are highly recommended, with the protocol becoming more extensive for research purposes (Suppl table 7s). Neuroimaging is also providing surrogate markers for treatment studies of VCID.¹⁶⁹ It is likely that clinical practice will be influenced by the application of newer techniques in the near future, and quantitative assessments assisted by AI will become routine practice.

Molecular biomarkers for VCID,_summarized in supplemental Table 8s and Figure 4s, are still in development and not ready for clinical use.

Coexisting Vascular and Neurodegenerative Diseases and Mixed Dementia

Both cerebrovascular and Alzheimer’s type of pathologies are common in the brains of older adults, and mixed pathologies, with additional contributions from α-synuclein and TDP-43 (TAR DNA binding protein 43), account for most dementia cases in community-dwelling older people,^170,171 exceeding pure AD and pure VCID. This could simply be the concurrence of two common pathologies, with an additive effect on cognitive function, as was shown in the Nun Study, where the clinical expression of AD pathology was greatly enhanced by the presence of infarcts.¹⁷² However, there is some evidence that suggests interactive and reciprocally amplifying effects of vascular and Alzheimer pathologies.¹⁷³ While the evidence that small or large vessel diseases promote amyloid deposition is weak,¹⁷⁴ vascular risk factors such as midlife hypertension and diabetes have been implicated in the pathogenesis AD, and the breakdown of the BBB, inflammation and oxidative stress are shared mechanisms.¹⁷¹ Cerebrovascular function is reduced in the early AD and in those with MCI at risk of developing dementia due to AD. Accumulating evidence supports hypoperfusion and hypoxia promotion of β-amyloid production and amyloid plaque formation.¹⁷⁵ The vascular pathway is also important for the clearance of β-amyloid from the brain, as it is transported along the perivascular pathway through a transvascular transport system.¹⁰⁰ Recent evidence demonstrated accelerated tau pathology in transgenic Alzheimer mice with vascular lesions.¹⁷⁵ Apolipoprotein E ε4 (APOE*4), a major genetic risk factor for AD, disrupts the BBB through proinflammatory activation.¹⁷¹ According to the amyloid hypothesis of AD, β-amyloid oligomers are neurotoxic. It is possible these oligomers are independently toxic to the pericytes and/or endothelial cells, thereby compromising the neurovascular unit and the BBB.

Several factors link cerebral blood vessels with grey and white matter damage in an interactive process. The joint role of the two pathologies in producing cognitive dysfunction is summarized in Supplemental Figure 5s.

The occurrence of dementia due to the additive effects of two or more pathologies has been referred to as “Mixed Dementia”, with AD and CeVD pathologies being the two most likely to co-occur. The availability of biomarkers for AD, either through amyloid PET or CSF/blood biomarkers, has made it possible to establish the presence of AD pathology alongside neuroimaging evidence of CeVD. A challenge is presented by the fact that WMHs can occur in AD in the absence of independent CeVD.¹⁷⁶ A significant contributor to this is cerebral amyloid angiopathy, which is present in nearly 50% of AD cases based on pathology,¹⁷⁷ although WMH may also be related to axonal injury due to Wallerian degeneration as well as direct amyloid toxicity.¹⁷⁸

CONCLUSION AND FUTURE DIRECTIONS

Concerns about cognitive decline are commonly encountered in the clinical practice of cardiologists since the risk factors for cardiac disease also increase the risk of VCID and some of these may increase the risk of AD. Many cardiac diseases also increase the risk of VCID. Cardiologists should therefore be alert in recognizing concerns and evaluating them appropriately for their management and disposition. Cardiologists also have an important role in the prevention of cognitive disorders in older people. With recent developments in early diagnosis, confirmation using biomarkers, and management of cognitive disorders, the role of cardiologists, in collaboration with neurologists and primary care physicians, has become especially important.

The future is likely to be marked by continuing improvements in diagnostic techniques and biomarkers for early detection and diagnosis, important as disease-modifying medications become increasingly available for the dementias. The development of blood-based biomarkers will be a breakthrough in diagnostics. The availability of large-scale omics data will catalyze drug-discovery, aided by artificial intelligence techniques. For disease modification, multiple drugs will most likely be needed, requiring personalized approaches in prevention and treatment of VCID and other neurocognitive disorders. We may be at the threshold of such a transformative change.

BOX 1: Recognizing VCID in the Cardiology Clinic.

Step 1: Cardiologist is alerted to the presence of cognitive impairment

• Patient is older (>65 years)

• Patient has difficulty explaining their problems or understanding the instructions

• Patient repeatedly misses appointments or arrives at wrong time

• Poor control of cardiac problems owing to non-adherence or erratic use of medications

• Patient repeatedly defers to a support person when queried about their problems

• Patient has recently suffered a stroke or transient ischemic attack

• Patient belongs to very high risk category: atrial fibrillation, heart failure, recent cardiac surgery

Step 2: Cardiologist seeks evidence for concern of cognitive decline

• Careful questioning from the patient and, if possible, a knowledgeable informant, about recent decline in memory (e.g., difficulty remembering a grocery list or the plot of a television program) or executive function (e.g., planning a meal, using an appliance, checking a bill) or language (e.g., recalling names of common objects or close friends, making uncharacteristic spelling or grammatical errors)

• The difficulties represent a change and not a lifelong pattern.

Step 3: Obtaining objective evidence of cognitive deficits

• Administer a screening instrument (e.g. MoCA¹, 16 items, 10 min, which is preferred, or MMSE², 11 items, 8 min), with cu-off of <26 (MoCA) or <25 (MMSE suggesting cognitive impairment.

• As a minimum, do a Mini-Cog (2 items, 3 min, cut-off ≤2), but with low sensitivity and specificity¹⁹⁰

Step 4: Assessment of functional impairment due to cognitive deficits

• Assess independence in instrumental activities of daily life (IADL) (e.g., using a telephone, managing money, preparing a meal, etc.). A rating scale (e.g. Lawton and Brody’s IADL scale¹⁹¹ may be used. Assessment is best done with the help of an informant.

MoCA: Montreal Cognitive Assessment¹⁹²

MMSE: Mini-Mental State Examination¹⁹³

BOX 2: Key Clinical Messages.

1. Alzheimer’s disease (AD) and VCID are the two most common causes of dementia, with cerebrovascular disease making some contribution to the majority of dementia cases.

2. Risk factors for cerebrovascular disease (CeVD) are shared with those for cardiovascular disease (CVD). Heart and bran health are therefore closely aligned.

3. Cardiac patients are at increased risk of cognitive impairment, especially in the presence of atrial fibrillation, congestive heart failure, coronary artery disease, acute coronary syndrome and valvular disease.

4. Cognitive impairment can in turn compromise the management of cardiac conditions.

5. Cardiovascular clinicians should have a high index of suspicion of the presence of cognitive impairment in their patients, especially those in the older age group, or with risk factors.

6. The assessment of a patient for cognitive impairment is a stepwise process, as described in Box 1. The clinician should be versant with these steps.

7. The presence of CeVD is suggested by history and physical examination and supported by neuroimaging, in particular MRI.

8. The preferred imaging for examining CeVD is an MRI scan, minimally with T1-weighted and T2-weighted FLAIR whole brain sequences. Other sequences may be indicated for special clinical applications.

9. MRI is also useful in differentiating VCID from some other causes of dementia. FDG PET and SPECT scans are used less frequently for this purpose.

10. If AD is suspected as a cause of cognitive impairment, amyloid PET imaging can be used to confirm or rule it out.

11. Blood and cerebrospinal fluid biomarkers for VCID are not ready for clinical application. These are however available for AD, and should be considered when AD is being considered as an alternative or concomitant pathology.

12. The cardiologist needs to work closely with a neurologist, geriatrician and/or family physician to provide optimal outcome for patients with VCID.

ABBREVIATIONS

AD

Alzheimer’s Disease

BBB

Blood Brain Barrier

CAA

Cerebral Amyloid Angiopathy

CADASIL

Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy

CBF

Cerebral Blood Flow

CMB

Cerebral Microbleed

CeVD

Cerebrovascular disease

CVD

Cardiovascular Disease

cSVD

Cerebral Small Vessel Disease

MCI

Mild Cognitive Impairment

MID

Multi-infarct Dementia

NVU

Neurovascular unit

PVS

Perivascular Spaces

VaD

Vascular Dementia

VCI

Vascular Cognitive Impairment

VCID

Vascular Cognitive Impairment and Dementia

WMH

White Matter Hyperintensities