Vascular dementia and vascular cognitive impairment—current treatments and future prospects

Abstract

Vascular cognitive impairment (VCI) and vascular dementia (VaD) represent a substantial and growing global health burden, yet remain under-recognized and comparatively under-researched. Two broad categories are post-stroke cognitive impairment, and VCI secondary to chronic cerebrovascular disease of which the most common cause is cerebral small vessel disease (SVD). In addition, cerebrovascular pathology frequently coexists with Alzheimer’s disease (AD), increasing the probability that AD pathology will manifest as clinical dementia. Current management focuses primarily on prevention through modification of cardiovascular risk factors and lifestyle interventions. Intensive blood-pressure control has shown the strongest evidence, reducing future all-cause dementia while slowing progression of white-matter abnormalities. Multidomain lifestyle interventions incorporating exercise, diet, cognitive training, and vascular risk monitoring may also preserve cognition, although results across trials remain variable. In established VCI and VaD, treatment options are limited; cholinesterase inhibitors and memantine lack convincing evidence of benefit, and no pharmacological therapy has definitively demonstrated symptomatic improvement or delayed disease progression. Advances in understanding SVD and VCI pathophysiology are, however, identifying promising therapeutic targets. Emerging evidence highlights endothelial dysfunction, impaired cerebral autoregulation, increased blood–brain barrier permeability, inflammation, and glymphatic dysfunction as possible mechanisms underlying SVD. Current trials are evaluating treatment approaches based on this framework including cerebral vasodilators such as cilostazol, isosorbide mononitrate and phosphodiesterase-5 inhibitors, anti-inflammatory therapies, glucagon-like peptide-1 receptor agonists, remote ischemic conditioning, and strategies targeting glymphatic function. Improved mechanistic understanding, appropriately targeted interventions, and adequately powered long-term clinical trials are required to translate these emerging approaches into effective treatments for VCI and VaD.

It has been estimated that 8.5 million people worldwide suffer from pure vascular dementia (VaD), and a further 9.1 million from mixed dementia to which vascular pathology contributes, while the global burden of total VaD (i.e. pure VaD and mixed dementia) will reach 42.7 million by 2050. ¹ Despite this, VaD remains under-recognized and under-researched compared to other forms of dementia, and there are few proven treatments. This issue of the International Journal of Stroke (IJS), published simultaneously with the 2026 World Stroke Congress, brings together a collection of articles on VaD and vascular cognitive impairment (VCI), and its underlying causes, particularly cerebral small vessel disease (CSVD). It highlights exciting new insights into pathogenesis which may lead to new treatment approaches.

The heterogeneity of VCI

VCI is increasingly used in place of VaD to refer to the spectrum of cerebrovascular pathologies that contribute to any degree of cognitive impairment, ranging from mild cognitive impairment to dementia ² (Figure 1). VCI itself is a heterogeneous syndrome which can be caused by multiple pathologies. A useful division is into post-stroke VCI, representing acute cognitive decline occurring soon after a stroke, and chronic cerebrovascular disease, in particular CSVD, which can cause insidious cognitive decline in the absence of clinical stroke. Recognizing this heterogeneity, and the underlying pathogenesis, is important when planning treatment approaches.

Figure 1.

The major types of vascular cognitive impairment.

The major types of VCI

Post-stroke cognitive decline

Post-stroke VCI and VaD have been estimated to occur after 38% and 18% of all strokes, respectively.^2–5 The term post-stroke cognitive impairment (PSCI) is used to cover any degree of cognitive impairment resulting from stroke. Risk factors include age and cardiovascular risk factors such as hypertension, hyperlipidemia, diabetes, atrial fibrillation, physical inactivity, smoking, and obesity ² Whether the stroke event induces cognitive impairment depends on a complex interplay between the features of the stroke lesion itself and the brain resilience of the individual. ² Cognitive impairment is more likely after large lesions or lesions located at “strategic” sites. ⁶ Brain resilience depends on pre-existing brain diseases (such as previous strokes, small vessel disease (SVD), and other neurodegenerative pathologies), as well as cognitive reserve, for which one marker is the number of years of education. Even a small lesion located in a non-strategic region, or a transient ischemic attack, may trigger post-stroke cognitive impairment in someone with low brain resilience. Furthermore, after the cognitive decline associated with an acute stroke has settled, the long-term rate of cognitive decline over subsequent years is faster when compared to that in stroke-free individuals, and such a decline occurs even in the absence of recurrent stroke. ⁷

While stroke is associated with PSCI, an interesting paper in this issue of IJS by Huang et al. ⁸ looked at dementia risk before stroke diagnosis. The study included 24,056 individuals with stroke and 74,136 matched controls from the UK Biobank. The risk of dementia was higher in individuals with stroke compared to non-stroke controls in each time window before stroke diagnosis, with odds ratios ranging from 1.43 in 5–10 years before diagnosis to 5.11 in the year immediately before the stroke. These results demonstrate that participants who developed stroke had a higher risk of dementia beginning a decade before stroke onset. They emphasize the bidirectional relationship between stroke and dementia and that if we wish to reduce the risk of post-stroke dementia, we need to also pay close attention to primary prevention of stroke risk factors.

It is important to recognize that the pathology underlying post-stroke is varied. It includes both ischemic stroke and intracerebral hemorrhage, and within these broad categories, there are many different pathophysiology processes causing the stroke. VCI has also been shown to be common after other cerebrovascular disease, including cerebral venous thrombosis. ⁹

SVD and VCI

SVD leads to about 25% of ischemic strokes, most hemorrhagic strokes in adults more than 65 years, and is the most common pathology underlying VCI and VaD. ¹⁰ Other clinical symptoms include gait disorders and neurobehavioral disorders including apathy. It can result in cognitive impairment, even in the absence of clinical stroke. The core neuroimaging markers of SVD include recent small subcortical infarcts, lacunes, white-matter hyperintensities (WMHs), enlarged perivascular spaces (PVSs), and cerebral microbleeds (CMBs). ¹¹ Diffusion tensor imaging (DTI) studies have shown diffuse white-matter abnormalities, not only in areas of WMH but also in normal-appearing white matter. Even in symptom-free individuals, WMH and lacunar infarcts, which become common in older individuals, are predictors of future stroke and dementia risk. ¹²

A characteristic pattern of cognitive impairment is seen with prominent early involvement of information processing speed and executive dysfunction. ¹³ In general, cognitive impairment is associated with increasing extent or severity of magnetic resonance imaging (MRI) markers of tissue damage, and the strongest associations have generally been reported with both DTI abnormalities within the white matter and the number of lacunes. DTI tractography techniques can be used to construct brain networks, and studies using this technique have shown disruption of brain networks dependent on white-matter tract integrity in SVD. ¹⁴ The degree of network disruption was both correlated with cognitive decline and predicted future dementia risk,^15,16 while it also mediated the effect of WMH and lacunes on cognition. ¹⁴ This suggests that disruption of complex brain networks plays a central role in the genesis of VCI and VaD caused by SVD.

The importance of network disruption in PSCI is further emphasized in a paper in this issue of IJS by Zhou et al. ¹⁷ In a cohort of 376 mild-to-moderate first acute ischemic stroke patients without pre-stroke dementia, they investigated network-level disconnection patterns underlying PSCI. First, they used connectome-based lesion-symptom mapping (CLSM) to characterize how infarcts and WMH result in network disconnection. They then investigated the relationship between this network disruption and cognitive deficits. They found that both spatially heterogeneous infarcts and WMH converge on functionally/structurally coherent disconnection patterns. To systematically quantify the impact on brain networks, they calculated a separate disconnection score for each lesion, and the derived lesion impact scores were significantly correlated with cognitive deficits and demonstrated greater prediction for 3-month cognitive outcomes than traditional volumetric measures. WMH-induced disconnections significantly mediated the relationship between infarcts and PSCI deficits. This exciting paper emphasizes the key role of network disruption in PSCI, and how SVD can reduce brain resilience after stroke, increasing the chance that the stroke will result in PSCI. It also highlights directions in which therapeutic approaches guided by this finding could be developed.

The huge global burden of SVD, and its contribution to VCI, is highlighted in a systematic review and meta-analysis of large MRI studies, also published in this issue of the IJS. Eighty-five studies (spanning 17 global regions, and including a total of 1,562,765 individuals, with a mean age of 64.4 years) were ¹⁸ included. The pooled prevalence of moderate-to-severe WMH was 18.9%, and the pooled prevalence of lacunes, CMBs, and moderate-to-severe enlarged PVS were 11.2%, 10.3%, and 22.6%, respectively. A higher lacune prevalence (13.3% vs. 7.3%) was found in Asia compared with Europe. A complementary study, also in this issue, looked specifically at incident lacunes. Thirty-one studies comprising 12,646 participants and 56,073 person-years were included. ¹⁹ The pooled overall incidence was 3.27 per 100 person-years ranging from 1.50 to 8.03 across populations. Rates were the highest in SVD patients (8.03), intermediate in stroke and memory-clinic patients, and lower in community-based, hypertensive, and non-specific artery disease cohorts. Predictors of new lacunes included male sex, baseline lacunes, hypertension, and diabetes. Together, these studies highlight the enormous global burden of SVD and, due to its key importance in VCI, emphasize that a key to tackling the epidemic of VaD is better prevention and treatment of SVD.

The vascular contribution to Alzheimer’s disease

Even though it creates an enormous global burden of disease, cognitive impairment caused directly by vascular disease, either from acute stroke or chronic cerebrovascular disease, does not fully capture the full contribution of cerebrovascular disease to dementia. The complex relationship between cerebrovascular disease with Alzheimer’s disease (AD) and other neurodegenerative dementias is covered in the World Stroke Organization (WSO) scientific statement on the vascular contribution to dementia, also published in this issue of IJS. ²⁰ This comprehensive statement, written by an international panel of experts, provides a state-of-the-art description covering the full breadth of how vascular disease contributes to dementia.

The contribution of vascular pathology to AD has been known for many years, although underappreciated and under-researched. ²¹ Early data emphasizing this association came from the Nun Study, in which an order of nuns was studied clinically during life and donated their brains at death. Of those with Alzheimer’s pathology on post-mortem, only 50–60%% had clinical dementia during life, but this increased to 70–80% in the presence of any stroke, and over 90% in the presence of SVD. ²² This strong modifying effect of cerebrovascular disease on whether Alzheimer’s pathology results in clinical dementia has been confirmed in other studies. It is particularly relevant in the general population. While many research studies in AD have been performed in highly selected younger populations, community-based studies with post-mortem neuropathology have shown that in the vast majority of dementia cases, a single pathology is rare, and most cases have two or more pathologies, of which vascular pathology is a frequent component.^23,24

As discussed in detail in the WSO scientific statement, this close relationship between AD pathology and vascular disease raises the question of whether vascular disease causes or accelerates cognitive decline. ²⁰ Is the association explained simply by vascular disease reducing the brain’s cognitive reserve, and therefore making AD pathology more likely to cause cognitive impairment, or could the two biological processes interact? A number of methods by which these two processes could interact have been suggested. A key component of sporadic AD is thought to be the failure to clear beta-amyloid peptides. ²⁵ The glymphatic system, first described in 2012, plays a key role in brain waste clearance, and it has been suggested that disruption of this clearance mechanism could impair beta-amyloid clearance. ²⁶ This hypothesis is supported by rodent studies, in which glymphatic function has been measured by injections of tracers into the cerebrospinal fluid (CSF).^26,27 Performing such studies in large human cohorts is not possible, but recently some indirect MRI-based techniques to assess CSF dynamics have become available, including diffusion tensor imaging along perivascular space (DTI-ALPS), choroid plexus volume, and enlarged PVS. ²⁸ A recent study in 40K community-based subjects found DTI-ALPS, blood oxygen level-dependent CSF (BOLD-CSF) coupling, and choroid plexus (CP) volume all predicted all-cause dementia over a median 5.3 years of follow-up. ²⁹

Glymphatic transport occurs along PVS surrounding penetrating cerebral vessels, is driven primarily by vascular pulsatility, and is mediated by aquaporin-4 (AQP4) water channels expressed on astrocytic endfeet. ³⁰ Given that glymphatic transport occurs predominantly along perivascular pathways, it is anatomically and physiologically intertwined with the cerebral small vessels. Therefore, structural alterations of small vessels caused by SVD, such as vessel wall thickening, amyloid deposition, and luminal narrowing, could directly disrupt glymphatic flow. ²⁸ Increasing evidence suggests that glymphatic dysfunction is a key feature of SVD, ²⁸ and it has indeed been suggested that SVD represents a “glymphadenopathy.” ³¹

Current prevention of VCI and VaD

Current cardiovascular risk-factor management and specific treatments such as anticoagulants for atrial fibrillation have a major impact on reducing stroke incidence, and one would expect them to therefore reduce the risk of PSCI, although data supporting this hypothesis is limited, largely due to a lack of suitably powered studies with sufficient duration of follow-up. ² However, a more controversial question is whether lifestyle measures, including risk-factor modification, reduce the risk of all-cause dementia. The 2024 Lancet commission reported that 14 modifiable risk factors for dementia account for nearly half of all cases: less education, hearing loss, high low-density lipoprotein cholesterol, uncorrected vision loss, traumatic brain injury, high blood pressure (BP), physical inactivity, diabetes, obesity, smoking, excessive alcohol consumption, air pollution, depression, and social isolation. ³² Many of these risk factors are well-recognized stroke risk factors, and therefore a significant proportion of this modifiable risk may be mediated by cerebrovascular disease. This suggests that risk-factor modification could reduce the risk of not only VaD but also AD.

Assessing this requires intervention studies with prolonged follow-up, of which there have been few, but evidence is emerging, particularly for hypertension. The SPRINT trial randomized 9361 persons with a systolic BP of 130 mm Hg or higher and an increased cardiovascular risk, but without diabetes, to a systolic BP target of less than 120 mm Hg (intensive treatment) or a target of less than 140 mm Hg (standard treatment). It reported a 25% reduction in the primary composite outcome of myocardial infarction, other acute coronary syndromes, stroke, heart failure, or death from cardiovascular causes. ³³ Innovatively, the study included a cognitive substudy, SPRINT-MIND, in which cognitive assessments were performed during follow-up; 8563 (91.5%) completed at least 1 follow-up cognitive assessment. The median treatment intervention period was 3.34 years. During a total median follow-up of 5.11 years, intensive BP control reduced the risk of mild cognitive impairment (14.6 vs. 18.3 cases per 1000 person-years), a significant 19% reduction, and also of the combined rate of mild cognitive impairment or probable dementia. ³⁴ Interestingly, in an MRI substudy, intensive BP lowering was also associated with slower progression of WMH, suggesting that at least part of this reduction in VCI may be mediated by reduced SVD progression. ³⁵ Further support for a protective effect of BP control on dementia is provided by a cluster randomized trial which recruited 33,995 individuals aged ⩾40 years with uncontrolled hypertension in rural China. ³⁶ In total, 163 villages were randomly assigned to a non-physician community healthcare provider-led intervention and 163 villages to usual care. In the intervention group, antihypertensive medications were titrated according to a simple stepped-care protocol to achieve a systolic BP goal of <130 mm Hg and a diastolic BP goal of <80 mm Hg. Over 48 months, the net reduction in systolic BP and diastolic BP was 22.0 mm Hg and 9.3 mm Hg, respectively, in the intervention group compared to usual care. The primary outcome of all-cause dementia was 15% significantly lower in the intervention group than in the usual care group. It is important to highlight that the endpoint in both these studies was all-cause dementia, emphasizing that the benefit of BP reduction is beyond reducing VaD alone, although the results do not tell us whether this effect is due to a direct effect on AD pathology or to an impact on the vascular component of mixed dementia. Nevertheless, it has important public health implications.

Data on modification of other cardiovascular risk factors are less convincing. A meta-analysis of recent randomized controlled trials (RCTs) suggested that anti-diabetic agents were associated with less cognitive decline in diabetic patients. ³⁷ There is considerable current interest in glucagon-like peptide-1 receptor antagonists (GLP1-RAs), and a meta-analysis of 82,140 subjects from 11 RCTs found a 16% relative reduction in stroke compared with placebo. ³⁸ GLP1-RAs have also been associated with reduced cognitive decline, ³⁹ and this led to their testing in RCTs in AD, but the EVOKE trials were negative. ⁴⁰ Whether they may be effective in VaD remains to be tested, although some data suggest they may be particularly protective in SVD, the major pathology underlying VaD. ⁴¹

While a recent meta-analysis of observational studies reported statin use was associated with reduced dementia risk, ⁴² the majority of RCTs evaluating cognitive outcomes after statin therapy have failed to show a cognitive benefit, ² including the recent STAREE trial in 9971 individuals aged 70 or older followed for a median of 5.9 years. ⁴³

Current treatment of established VCI and VaD

The treatments discussed above may reduce cognitive impairment by preventing brain damage, but are there effective treatments for those with established VCI and VaD? While both cholinesterase inhibitors and memantine are used as symptomatic treatments in AD, and it has been suggested they may have benefit in VCI, perhaps due to denervation resulting in loss of cholinergic innervation, there is no convincing evidence that they benefit individuals with pure VaD, although they may have limited benefit in mixed dementia containing both an AD and a vascular component.^2,44 There are no other pharmacological interventions that have definitively shown to improve symptoms, or delay progression, of PSCI or VCI, and high-quality adequately powered clinical trials in this area are badly needed. ⁴⁴

An alternative non-pharmacological approach is lifestyle interventions. The FINGER study showed that a multidomain lifestyle (vascular risk monitoring, Mediterranean-like diet, physical exercise, cognitive training) intervention could improve or maintain cognitive functioning in at-risk elderly people without dementia. ⁴⁵ Other multidomain trials have given mixed results. ² One reason for this may be that the evolution from subclinical to clinical stages of VCI can take years to decades, and therefore large studies of long duration will be required to show an effect of interventions on incident dementia. ² However, it is a widely applicable potential intervention including in low- and middle-income countries (LMICs), as exemplified by the recent LatAm-FINGERS trial. In this single-blind, multicenter, randomized clinical trial conducted in 11 Latin American countries (Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, Mexico, Peru, and Uruguay), individuals aged 60–77 years with high risk of dementia (cardiovascular risk factors, aging, and dementia risk score ⩾6), and suboptimal cognitive performance were randomly assigned to receive either a 2-year systematic lifestyle intervention (SLI group) or a flexible lifestyle intervention (FLI group). ⁴⁶ The SLI provided structured multidomain lifestyle interventions with supervised support and monitoring; FLI offered health advice. In total, 1065 participants were included. Global cognitive composite scores increased more in the SLI group, with a mean annual change of 0.31 SD (95% confidence interval (CI) = 0.28–0.34) per year compared with 0.20 SD (0.17–0.23) per year in the FLI group (p < 0.0001).

The current status of multidomain interventions is reviewed in this issue of IJS by Yap et al. ⁴⁷ Synthesizing data from all the available trials, they highlight a number of important points. First, exercise seems a key component of such interventions. Second, in populations with documented cerebrovascular injury, particularly recent and subacute post-stroke cohorts, multidomain interventions without structured exercise components improved systemic vascular markers without corresponding cognitive benefit over typical trial durations. This suggests that once structural vascular injury is established, modification of systemic risk alone may be insufficient to produce measurable short- to medium-term cognitive change. Third, in the primary prevention setting, studies designed to modify vascular or metabolic risk factors did not consistently demonstrate cognitive benefit, even when systemic markers improved. In contrast, trials explicitly targeting cognition, particularly those incorporating structured and supervised exercise, were more likely to demonstrate modest improvements in composite or domain-specific measures. ⁴⁷

Understanding disease mechanisms to develop better treatments for VCI

Therefore, with the possible exception of cardiovascular and lifestyle measures in primary prevention, our current treatment options for established VCI and VaD are very limited. One major reason has been a lack of potential therapeutic targets. However, recent advances in the understanding of the pathophysiology of processes leading to VCI, particularly SVD, are suggesting a number of exciting possible new treatment approaches (Figure 2 and Table 1).

Figure 2.

Novel treatment approaches in cerebral small vessel disease and vascular cognitive impairment.

Table 1.

Ongoing trials in SVD and VCI.

Mechanism	Drug	Trial name	Phase of trial	Sample size	Clinical study registry number	Notes
Vasodilators	Isosorbide mononitrate	LACI-3	3	1300	ISRCTN44436843	Tested in a 2 × 2 factorial design with cilostazol
	Cilostazol	LACI-3	3	1300	ISRCTN44436843	Tested in a 2 × 2 factorial design with isosorbide mononitrate
	Cilostazol		3	632	NCT07180472
	Isosorbide mononitrate	IMPACT	3	3156	NCT07252544	Tested in a 2 × 2 factorial design with butylphthalide
	Sildenafil	SILENCE-China	3	1640
	Sildenafil	SILENCE-UK	3	1360
GLP-1 agonists	exenatide	GAPP-SVD	2	110	NCT05356104
Anti-oxidant/free radical scavenger	Edaravone	240	2	240		Ye et al. 48
	Edaravone	TASTE-SVD	3	600	NCT06674460
Endothelial function	XBD173		2	106	NCT06643013	XBD173 targets the 18 kDa translocator protein (TSPO) to improve brain endothelial cell function.
Remote ischemic conditioning	Not applicable	Various	2	20-100	NCT07317557 NCT05967728 NCT05225948	Trials of feasibility and with surrogate marker outcomes. Not powered for clinical outcomes.

CSVD is not a single disease but can be caused by diverse pathological processes. The two most common are arteriolosclerosis caused by aging, hypertension, and other conventional vascular risk factors, and cerebral amyloid angiopathy (CAA) caused by vascular deposition of β-amyloid. Other less common causes include monogenic forms of SVD, the most common of which is cerebral autosomal dominant arteriopathy with subcortical ischemic strokes and leukoencephalopathy (CADASIL). Age-related hypertensive non-amyloid SVD may also be heterogeneous. Diffuse arteriosclerosis affecting the small vessels plays an important role, but lacunar infarcts can also result from atherosclerosis at or near the origin of the perforating arteries.^10,49

The conventional hypothesis has been that in SVD, hypoperfusion results in cerebral ischemia and tissue damage such as WMH, particularly in the regions furthest from the origin of the perforating arteries. While this is likely to be important, more recent data suggest other processes may also contribute. These include endothelial activation and dysfunction, increased blood–brain barrier (BBB) permeability, inflammation, and glymphatic dysfunction. It has been proposed that early endothelial dysfunction leads to the disruption of the BBB with leakage of fluid and toxic plasma proteins into the vascular media and surrounding tissues, with secondary effects on vascular reactivity, pericyte function, oligodendrocyte proliferation, and perivascular fluid drainage pathways. ⁵⁰ A major disease target is believed to be the extracellular matrix and associated proteins, together referred to as the matrisome, disruption of which would be expected to result in increased BBB permeability. ⁵¹ A better understanding of the molecular pathways underlying these processes may lead to new targets for drug therapy. ⁵² In this issue of IJS, we have reviews on a number of these areas, which examine how the individual processes are disrupted in SVD and look toward new treatments targeting this dysfunction.

Webb et al. ⁵³ review the role of impaired cerebral autoregulation in SVD. Autoregulation is the homeostatic process whereby the body maintains a stable cerebral blood flow (CBF) in the face of variations in BP to limit the effects of hypoperfusion and hyperperfusion and depends on both endothelial and smooth muscle cell function. Autoregulation therefore represents cerebrovascular responsiveness to BP changes and is the principal mechanism protecting the brain against hypoperfusion or hyperperfusion. ⁵³ The review highlights how a number of studies have demonstrated impaired autoregulation in SVD and that this is positively associated with the extent of SVD markers on MRI, although these studies are all cross-sectional, and there are little data showing whether impaired autoregulation predicts future progression of SVD. Although not directly targeting impaired autoregulation, one way of tackling the resultant hypoperfusion is cerebral vasodilators. The vasodilators cilostazol and isosorbide were associated with improved vasodilatory function and possibly improved cognition in SVD in the LACI-1 and LACI-2 trials.^54,55 The LACI-3 trial is examining their effect on clinical outcomes in a phase 3 setting in 1300 participants. ⁵⁶ Phosphodiesterase type 5 (PDE5) inhibitors, which reduce the breakdown of cGMP in vascular smooth muscle cells following nitric oxide release, also show promise in SVD. Considerable experimental and genetic data suggest they could be effective in SVD. ⁵⁷ Treatment with the PDE5 inhibitor sildenafil, in the OxHARP trial in 65 individuals, improved cerebrovascular reactivity and absolute CBF, ⁵⁸ while 3 months of tadalafil in the ETLAS-2 trial improved cerebral perfusion. ⁵⁹ The phase 3 SILENCE family of trials, currently funded in China and the United Kingdom, is examining the efficacy of sildenafil in CSVD.

Disordered inflammatory responses have also been implicated. Both systemic inflammation and central nervous system (CNS) inflammation have been demonstrated in SVD. Current thinking in this area is reviewed by Brown et al. ⁶⁰ in this issue of IJS. CNS inflammation has been demonstrated in SVD both in post-mortem brains and in vivo using positron emission tomography with radioligands targeted against the translocator surface proteins in microglia, but data showing such changes are causally related to disease progression are lacking. Peripheral inflammation can be measured in the blood using targeted assays and, more recently, large-scale proteomic panels. Proteins involved in coagulation, endothelial cell activation, and immune cell adhesion have been associated with SVD, as well as specific cytokines, although not all findings have been replicated, and there are limited data examining whether individual proteins predict future disease progression. A key question is whether such changes are causal and, therefore, whether targeting them could delay disease progression, or whether they are merely secondary to tissue damage. Insights into causality can be gained from longitudinal studies and from genetic data using Mendelian randomization approaches, but ultimately support has to come from intervention studies. A pathway linking hypoperfusion and hypoxia to BBB disruption and inflammation has been proposed based on studies in a rodent model of white-matter ischemia, and minocycline administration was associated with a significant reduction in white-matter damage and improved behavioral outcomes. ⁶¹ Minocycline is known to have anti-inflammatory properties within the brain, reducing the activation of microglia and may stabilize the BBB. However, a phase 2 trial examining the effectiveness of minocycline on microglial activation measured using positron emission tomography (PET), and BBB permeability measured using MRI in humans, showed no treatment effect. ⁶² Brown et al. ⁶⁰ conclude that a better understanding of the molecular basis of inflammation and immune dysfunction in SVD is needed to allow more precise therapeutic targeting.

Increased BBB permeability has been demonstrated in multiple studies in patients with different phenotypes of SVD including lacunar stroke, MRI markers of SVD, and VCI. ⁶³ It is usually measured using dynamic contrast-enhanced MRI, with prolonged imaging after administration of an MRI contrast agent such as gadolinium. BBB disruption can trigger neuroinflammatory cascades which can itself contribute to BBB breakdown, creating a detrimental feedback loop. However, whether BBB disruption predicts future cognitive decline is uncertain, with limited longitudinal studies giving conflicting results.^64,65 Approaches have been suggested that may stabilize the BBB, including matrix metalloproteinase (MMP) inhibitors. MMPs degrade the extracellular matrix around small vessels. However, as yet there are no human treatment trials of agents to stabilize the BBB.

A number of other approaches are being evaluated in ongoing clinical trials. Edaravone is a free radical scavenger and neuroprotectant, which exhibits significant anti-inflammatory and immunomodulatory properties. ⁴⁸ In rodent models of CSVD, edaravone has been shown to reduce oxidative stress, suppress inflammation, and modulate the N-methyl-D-aspartate receptor signaling pathway, improving learning ability. ⁶⁶ A sublingual preparation is being examined in a phase 2 trial in 240 patients with SVD-related cognitive impairment, randomized between active treatment and placebo. The primary outcome is the Montreal Cognitive Assessment score change from baseline to week 48. ⁴⁸ GLP-1 agonists are being evaluated in a phase 2 RCT (ClinicalTrials.gov ID NCT05356104) to assess tolerability and clinical, imaging, and hemodynamic parameters in 110 patients with SVD of Age-Related White Matter Changes Scale of 2 or 3, followed for 78 weeks. Remote ischemic conditioning, which involves brief cycles of limb ischemia and reperfusion, may be a neuroprotective adjuvant that provides neuroprotective effect and reduces ischemic injury. ⁶⁷ A number of studies are evaluating it in SVD.

Finally, the suggestion that glymphatic dysfunction plays a key role in SVD and VCI raises the possibility that strategies to improve CSF dynamics could reduce SVD progression. Impaired CSF dynamics estimated using DTI-ALPS predicted future dementia risk in 120 patients with lacunar strokes and confluent white-matter hyperintensities followed up for 5 years, consistent with, but not proving, a causal association. ⁶⁸ Potential approaches which could improve glymphatic function are discussed in detail in the article by Hong in this issue of IJS. ²⁸ They include controlling cardiovascular risk and improving sleep, during which glymphatic flow is believed to be most active including with the use of orexin receptor antagonists, pharmacological modulation of the noradrenergic system, and neuromodulation. A number of these are currently being evaluated in AD, based on the hypothesis that impaired glymphatic function reduces clearance of amyloid beta and tau proteins.

The advances above have largely been in sporadic SVD, but CAA also causes VCI, as well as being the primary cause of lobar intracerebral hemorrhage. It affects cortical and leptomeningeal vessels and is characterized pathologically by progressive deposition of amyloid-β in the cerebrovascular wall. BBB disruption has also been suggested to play a role in CAA. In this issue of IJS, Voigt and colleagues used contrast-enhanced MRI to assess global parenchymal gadolinium extravasation across the BBB. Perhaps surprisingly, BBB permeability was lower in patients with CAA compared with controls. ⁶⁹ They suggest this observation can be explained in terms of limited capillary blood perfusion and/or the possibility that vascular amyloid-β deposition impairs molecular transport across the BBB. In contrast, focal leptomeningeal enhancement was higher in CAA, reflecting vessel wall infiltration. The role of inflammation in CAA is becoming particularly interesting. For many years, we have known that in rare patients with CAA, an unusual inflammatory presentation can occur with a distinct clinical presentation, typically with a subacute encephalopathy and characteristic neuroimaging features, in particular asymmetrical confluent WMH. ⁷⁰ It is usually treated with corticosteroids and other immunosuppressive agents. However, recent evidence has suggested lesser degrees of inflammation such as sulcal hyperintensities, gyral swelling, and leptomeningeal enhancement may occur in typical, “non-inflammatory” CAA. It has been suggested that, while the directions of causality between the various forms of inflammation and symptoms remain to be elucidated, treating coexisting inflammation in sporadic CAA could reduce symptoms and subsequent hemorrhagic events—an exciting hypothesis that could be tested in clinical trials. ⁷⁰

Conclusion

In summary, VCI is a common and increasingly important consequence of cerebrovascular disease, particularly SVD. Current management remains centered on vascular risk-factor modification, with BP control supported by the strongest evidence. However, growing understanding of SVD pathophysiology has identified several promising therapeutic targets, including cerebral hypoperfusion and impaired autoregulation, endothelial and BBB dysfunction, neuroinflammation, and impaired glymphatic clearance. Emerging pharmacological and non-pharmacological interventions targeting these mechanisms offer potential new approaches, but adequately powered clinical trials with meaningful cognitive outcomes are required before they can be incorporated into routine clinical practice.