Abstract
Stroke, characterized as a neurological deficit of cerebrovascular cause, is very common in older adults. Increasing evidence suggests stroke contributes to the risk and severity of cognitive impairment. People with cognitive impairment following stroke often face with quality-of-life issues and require ongoing support, which have a profound effect on caregivers and society. The high morbidity of post-stroke cognitive impairment (PSCI) demands effective management strategies, in which preventive strategies are more appealing, especially those targeting towards modifiable risk factors. In this review article, we attempt to summarize existing evidence and knowledge gaps on PSCI: elaborating on the heterogeneity in current definitions, reporting the inconsistent findings in PSCI prevalence in the literature, exploring established or less established predictors, outlining prevention and treatment strategies potentially effective or currently being tested, and proposing promising directions for future research.
INTRODUCTION
Stroke, one of the leading causes of death and adult disability, is highly associated with increased risks of cognitive impairment and dementia. Comorbidity is very common in stroke and has strong prognostic implications for both function and mortality in all age groups [1]. Among stroke survivors, cognitive impairment is a major cause underlying disability/dependence. As the tremendous burden of stroke continues to rise, post-stroke cognitive impairment (PSCI) has become a mounting public healthcare challenge. In this review article, we aim to summarize recent development in the definition, epidemiology, risk factors, prevention, and treatment of PSCI, and to provide insights in both clinical and public health sectors.
DEFINITION AND CLASSIFICATION
PSCI is a broad concept that covers the full spectrum from mild cognitive impairment to dementia and includes cases with solely vascular pathology or mixed pathologies. In order to differentiate from prestroke dementia, in most studies, PSCI was defined as cognitive decline developed following a clinical cerebrovascular event. PSCI or post-stroke dementia (PSD) was recommended to be diagnosed at least 3 to 6 months after the index cerebrovascular event, in which period the brain would recover at least partly from the event. The concept of PSCI is relatively explicit. However, the identification is often challenging due to the indeterminacy on recovery and the latency of symptoms presentation. In 2017, the diagnosis and potential mechanisms of early-onset PSD and delayed-onset PSD were reviewed by Mok et al. [2, 3]. Panel 1 presents PSCI-related terminology.
EPIDEMIOLOGY
The epidemiological data of PSCI have been widely reported. However, the prevalence of PSCI varies widely across studies depending on time of assessment, setting of study, population characteristics, and the diverse cognitive tests and cut-offs. We summarized the data of distribution of PSCI in Fig. 1 and the Supplementary Material contains the related references.
The distribution of the post-stroke cognitive impairment. PSCI, post-stroke cognitive impairment; PSD, post-stroke dementia; VCI, vascular impairment; NCD, neurocognitive disorder.
Pooled data analyses indicated a prevalence of 38% (95% confidence intervals, 32% to 43%) of PSCI [4] and a prevalence 18.4% (95% confidence intervals, 7.4% to 38.7%) [5] of PSD in the first year after stroke. The Stroke and Cognition Consortium (STROKOG) harmonized data from 13 studies based in 8 countries found that 44% of stroke participants admitted to hospital were impaired in global cognition [6]. According to several longitudinal studies, about one in five patients with an initial or previous history of stroke developed early-onset PSD, and 4.4% to 23.9% patients developed delayed-onset PSD [7–9]. The cognition of stroke survivors mostly improves within the first 12 months, but gradually deteriorates afterwards. In addition, post-stroke patients had the highest risk of dementia after 1 year, and the increased risk remains persistent even after a decade [10, 11].
Characteristics and mechanisms vary with different stroke subtypes, which can also influence the prevalence of PSCI. Following ischemic stroke, intracerebral hemorrhage (ICH) is the second most common subtype of stroke, but it accounts for app-roximately 40% of stroke-related fatality [12]. A study demonstrated that the prevalence of cognitive impairment ranged between 9–29% and 14–88% for pre-ICH and post-ICH patients, and approximately 19.0% ICH patients had developed early dementia within 6 months after recovery. For those who did not develop early dementia, the estimated annual dementia incidence was 5.8% (median follow-up, 47.4 months) [13, 14].
Nearly 10% of all ischemic strokes occur in young adults (under 50 years of age) [15]. Approximately 1 year after stroke, up to 60% of young survivors show below-average performance in the cognitive function [16]. Generally, the course of cognitive recovery in younger patients possibly continues beyond 1 year after stroke [17]. Fortunately, the chance of returning to work in younger patients continues to enhance over years after the stroke. However, studies with longer follow-up and adjusting for age reported that after a mean follow-up of 11 years, 50% of young patients may still have poor cognitive performance [18].
RISK AND PROTECTIVE FACTORS FOR PSCI
Demographic factors
Few prospective studies have examined the risks that are potentially associated with the development of PSCI. Increasing age, low educational status, ethnicity, place of birth, and pre-stroke cognitive status were reported that related with PSCI [19–21]. Figure 2 presents the major risk factors. Taken together, it is possible that socioeconomic status played a crucial role in aggravating the severity of cognitive declines.
The major risk factors for post-stroke cognitive impairment. The interaction between specific environmental factors and genetic variants could lead to the development of stroke, which subsequently increase the risk of post-stroke cognitive impairment.
Stroke characteristics
The effects of stroke-related factors, such as a previous stroke history, stroke lesion characteristics, volume of brain infarction, and clinical stroke severity, on risk of PSCI were evident. For patients with major stroke events, the incidence of dementia was 50 times higher compared to the general population [22]. For transient ischemic attack (TIA) patients, significant delay atrophy in global gray matter and cognitive worsening were also observed [23]. The damages in different strategic brain regions, including basal ganglia, thalamus, corpus callosum, internal capsule, cingulate cortex, enlarged perivascular space in the basal ganglia, and the left superior fronto-occipital fasciculus and part of corpus callosum have been reported in association with global PSCI. In the recent large-scale multicohort lesion-symptom mapping study, Weaver and colleagues with data from 2950 patients showed that infarcts in the left frontotemporal lobes, right parietal lobe, and left thalamus were most strongly predictive of PSCI [24]. Compared to cortical and cerebellar infarcts, subcortical infarcts have also been considered to generate greater contribution regarding the incidence of dementia [25]. To validate the relationship of structure-function, the Groupe de Réflexion pour l’Évaluation Cognitive Vasculaire (GRECogVASC) Study Group found that lesions in the strategic areas were the main determinant of post-stroke cognition, accounting for 22.5% of the variance in the Global Cognitive Score. Interestingly, the total stroke volume and total medial temporal lobe atrophy only accounted for a small proportion of the variability in cognitive performance, in which they were independent but weak determinants [26].
Vascular risk factors
Vascular and related factors, including atrial fibrillation, the number and stability of carotid plaques, diabetes, chronic kidney disease, alcohol use, hypertension, and smoking have been reported to be associated with PSCI. An additive effect of Type 2 diabetes mellitus and chronic kidney disease on cognitive decline doubles the risk for PSCI relative to the presence of one disorder [27]. Also, glycemic variability and hyperglycemia in acute ischemic stroke were identified as novel predictors for PSCI [28].
Genetic susceptibility
The genetic data from PSCI are relatively scarce, and most are derived from Alzheimer’s disease. So far, only one gene (rs12007229) on the X chromosome was reported to be related to PSCI [29]. Recently, a longitudinal population-based cohort study has shown that homozygosity for the APOE ɛ4 allele was associated with both pre- and post-event dementia in patients with TIA and stroke [30].
Risk score models
Only a few prospective studies attempted to develop predictive models of PSCI by incorporating risk factors [31–33]. The SIGNAL2 (Stenosis, Infarct type, Global Cortical Atrophy, Number of years of education, Age, Leuokoariosis/White Matter Hyperintensity, Lacune count) risk score, comprising clinical and neuroimaging variables, seemed to be effective in identifying stroke survivors who are at risk for PSCI [34]. The CHANGE (Chronic lacunes, Hyperintensities, Age, Non-lacunar cortical infarcts, Global atrophy, and Education) risk score model has been considered as a reliable tool for screening the risk of PSCI in both subacute and chronic ischemic stroke survivors [35]. For patients with nonvalvular atrial fibrillation-induced cardioembolic stroke, the CHADS2 (Congestive Heart Failure, Hypertension, Age ≥75years, Diabetes mellitus, Stroke) and R2-CHADS2 (CHADS2 + creatinine clearance < 60 mL/min) may also be effective for predicting PSCI [36]. Table 1 shows each of the different stroke-specific models including their component variables, predictive, and validation accuracy [37, 38].
Risk score models for predicting post-stoke cognitive impairment
Risk score models’ properties, including sample size, sensitivity, specificity, AUC, are driven from validation data. aThis model has not been validated; AUC, Area under the curve.
ETIOLOGICAL MECHANISMS
PSCI has complex etiologies incorporating large artery disease, cerebral small vessel disease, and non-vascular pathology. Experimental studies have revealed a new functional and pathogenic synergy between neurons, glia, and vascular cells which provide a framework to reflect how vascular brain lesions affect cognition and what the nature relationship is of vascular pathology with neurodegeneration [39, 40].
Dysfunction of neurovascular unit (NVU)
The NVU, accomplished by a group of cells, including neurons, glia, vascular cells, pericytes, and extracellular matrix [41], plays an extremely important role on maintaining the homeostasis of the cerebral microenvironment. These cells work collectively to detect energy demands for neural activity and trigger responses to regulate regional changes in cerebral blood flow. Ischemic stroke caused by blood supply disruption induces NVU dysfunction, resulting in sudden or delayed neurological deficits. It is widely accepted that a series of ischemic injury cascades, consisting energy failure due to disruption of blood flow, calcium overloading, oxidative stress, blood-brain barrier dysfunction [42], injury-related inflammation [43, 44], can induce NVU dysfunction, and then subsequently cause sudden or delayed neurological deficits. The brain tissue may have complex reparative responses, which are however curtailed in ageing brains.
Neuroinflammation responses
Considerable evidence has strongly supported the role of inflammation in the pathophysiology of stroke. Stroke-induced compromised blood-brain barrier coincides with hypoperfusion and contributes to amyloid accumulation in brain parenchyma, which triggers neuroinflammatory responses [45]. The followed attraction of both local and systemic immune cells in the lesion sites could then lead to irreversible astrocyte injuries and the disruption of gliovascular interactions in the frontal white matter, resulting in the development of PSCI [46]. Meanwhile, the complement system activated by ischemic stroke, promoting tissue repair but also triggering inflammatory responses from long-lasting systemic activation, reduces post-stroke neuroplasticity and leads to poor cognitive outcomes [47]. Additionally, mouse models have shown the B-lymphocyte-mediated inflammation in an infarct lesion, with antibody diffusion into the surrounding neuropil, can directly cause delayed-onset PSCI [43]. Pro-inflammatory T helper cell subpopulations also promote neuroinflammation. Anti-inflammatory subpopulations, such as regulatory T cells, are critically involved in maintaining immune homeostasis.
Glymphatic pathway impairment
The glymphatic pathway is a brain interstitial met-abolic solute clearance system that supports the recirculation of cerebrospinal fluid (CSF) through the brain parenchyma. In mice, the microinfarcts can cause a rapid decrease in brain-wide glymphatic CSF influx and accumulation of tracers within the tissue. The results suggested the small, disperse ischemic lesions can disrupt glymphatic function and trap interstitial solutes within the brain parenchyma, increasing the risk of amyloid plaque formation [48]. Another study used the middle cerebral artery occlusion and bilateral common carotid artery occlusion rats to mimic a clinical setting where chronic cerebral hypoperfusion (CCH) is superimposed after acute territorial infarct, who found that neuroinflammation and amyloid pathology were enhanced in the ipsilateral cortex, thalamus, and hippocampus. These experimental findings suggested that CCH combined with territorial infarcts could significantly aggravate cognitive impairment. Notably, the glymphatic pathway-related aquaporin-4 (AQP4) distribution changes from perivascular to parenchymal pattern indicated that the aberrant dislocation of AQP4 water channel, along with neuroinflammation, might contribute to the post-stroke amyloid deposit. Therefore, CCH may affect the development of PSD by affecting both functional and structural changes of glymphatic pathways related to amyloid clearance [49].
Relation of infarcts with Alzheimer’s disease pathology
Cerebral infarctions and neurodegenerative disease commonly occur in older people. The most common causes of PSD are vascular dementia (VaD), AD, and mixed dementia, all of which need autopsy to make definitive diagnosis. Although numerous studies have demonstrated the co-existence of stroke and AD-like pathology, vascular causes of dementia and their contribution to neurodegenerative processes have not been well understood. Recently, vascular biomarkers have been recommended to be incorporated into the National Institute on Aging and Alzheimer’s Association (NIA-AA) Research Framework, to clearly define AD in the context of multifactorial pathophysiology and to broaden the perspective for therapeutic strategies[50]. Approximately 30% of patients with dementia after stroke or TIA have shown AD-like Pittsburg compound B binding, which is almost four times greater compared to patients without dementia. As a result, the concurrent AD pathology might also be involved in the onset of PSCI [51, 52]. The stroke-induced Aβ and tau deposits could also co-localized with increased levels of beta-secretase 1, along with its substrate-neuregulin 1 type III, and the components of a myelin repair pathway. This demonstrated that AD-related pathology could be a chronic sequela of ischemic stroke [53].
BIOMARKERS
Although scientists and clinicians have been wor-king hard to explore for news biomarker to diagnose or predict PSCI, most biomarkers have shown relatively low specificity. Plasma Aβ42, bedtime cortisol, hair cortisol concentrations, 8-hydroxydeoxyqua-nosine, matrix metalloproteinase-9, serum gluta-mine, kynurenine, lysophosphatidylcholine (18 : 2), S100β, high-sensitivity cardiac troponin T, homocysteine, low-density lipoprotein, D-amino acid oxidase, and uric acid have been suggested to correlate with PSCI. Recent work has shown that serum quinolinic acid (QUIN) over kynurenic acid (KYNA) ratios not only were strongly correlated with degradation of long-term memory, neuronal death, microglia/macrophage infiltration and white matter rarefaction in mice, but also, at the acute phase of stroke, were significantly different between the patients who will or will not present a cognitive dysfunction [54]. In addition, proinflammatory cytokines (e.g., interleukin-1β in CSF, interleukin-6 in serum, interleukin-12) and anti-inflammatory cytokines (e.g., interleukin-10 in CSF) have significantly elevated levels in patients with PSCI [55]. Using a high-dimensional elastic net analysis of the mass cytometry dataset to profile the stroke patients’ peripheral immune response over the course of a year, a study has successfully identified stroke-related phenotypical and functional immunological changes. Specifically, in peripheral blood, the characterization of the acute (innate) immunological magnitude has been observed, in which the signaling responses for signal transducer and activator of transcription 3 were increased. Such responses have been suggested with its strong correlation with long-term cognitive trajectories [56]. However, since most of the study did not either include baseline cardiovascular risk factors demographic or long-term follow-up data, it is challenging to draw definite conclusions.
Neuroimaging has a vital role in the assessment of PSCI by providing key information on anatomical substrates [57, 58]. Cerebral atrophy and medial temporal lobe atrophy [59] are considered as the most consistent predictors for cognitive impairment. Moreover, electroencephalography (EEG)-based ap-proaches may offer complementary information for functional MRI [60] and post-stroke recovery monitoring [61]. A number of studies demonstrated that quantitative measures of specific EEG feature indices, including slower EEG alpha generation, synchronization, the delta/alpha ratio, as well as delta/theta ratio, can reflect functional cortical connectivity that is sensitive to PSCI.
ASSESSMENT METHODS OF COGNITIVE PERFORMANCE
Challenges of characterizing cognitive deficits in stroke survivors
More than half of stroke survivors with excellent functional recovery by modified Rankin Scale continue to have cognitive impairment, indicating the need for greater attention to multiple domains in recovery and a more holistic approach to outcome assessment in stroke patients [62]. We herein review the assessment methods of post-stroke cognitive function and propose a timeframe for post-stroke cognitive function assessments, emphasizing the need for differing assessment methods at differing time points after stroke (Fig. 3).
Cognitive assessment throughout the stroke pathway. The schematic illustrates the potential approaches to cognitive assessment at various stages after stroke. Note that the mentioned tests are given as examples rather than recommendations. In the early period after stroke, the administration of detailed cognitive assessments may not be an immediate priority. Instead, simpler assessments focusing on the pre-stroke cognition, cognitive screening, and stroke-related impairments are applied. The GRECogVASC cognitive risk score can help identify stroke survivors who are eligible for a comprehensive cognitive assessment. Additionally, timely follow-up with patients in clinic or using telemedicine is also necessary. MoCA, Montreal Cognitive Assessment; GRECogVASC, Groupe de Réflexion pour l’Évaluation Cognitive Vasculaire; NIHSS, National Institutes Health Stroke Scale; MMSEadj, adjusted Mini-Mental State Examination; T-MoCA, Telephone Montreal Cognitive Assessment; TICS, Telephone Interview of Cognitive Status; IQCODE, Informant Questionnaire for Cognitive Decline; OCS, The Oxford Cognitive Screen; NINDS-CSN, Neurological Disease and Stroke-Canadian Stroke Network protocols.
Cognitive processing is heterogeneous in the stroke population. The related cognitive assessment for linguistic and motor function typically involves hand-written tasks, which can be significant barriers for many stroke survivors [63]. Patients with significant cognitive impairments (68%) or communication issue (62%) were mostly excluded in the evaluation of validity of cognitive assessments, which implied that specific stroke subgroups are poorly represented in such studies [64]. Assessment of cognitive impairment following stroke is complicated by additional stroke-related disfunction. Accordingly, the selection of assessment should take the availability and feasibility into account (Table 2).
Properties of selected tools to detect cognitive impairment after stroke
Test properties are from meta-analyses. aAccuracy of NINDS-CSN 5-min protocol for assessment of PSD at 3 months after stroke; bAccuracy of IQCODE for assessment of PSD in the longer term after stroke; cOCS was validated to be a stroke-specific short cognitive screening tool; MMSE, Mini-Mental State Examination; MoCA, The Montreal Cognitive Assessment; NINDS-CSN 5-min, The National Institute of Neurological Disease and Stroke-Canadian Stroke Network 5-minute neuropsychology protocol; PSD, post stroke dementia; IGCODE, The Informant Questionnaire on Cognitive Decline in the Elderly; ACE-R, Addenbrooke’s Cognitive Examination–Revised; OCS, Oxford Cognitive Screen.
Assessing pre-stroke problems
Although retrospective assessment is likely to be biased, retrospectively portraying the pre-stroke cognition state is important. Various tools exist to retrospectively assess the cognitive function, but there is no validated assessment method available for detection of pre-stroke dementia. Informant Questionnaire for Cognitive Decline in the Elderly (IQCODE) in stroke may have prognostic value, but the availability of informants required in the assessment is not guaranteed due to the recall bias, which means the assessment should be performed close to the event, to ascertain the accuracy of test results for pre-stroke cognitive state [65].
Cognitive assessment in post-stroke aphasia and spatial neglect
Post-stroke complications (e.g., impaired sensorimotor and language functions, mood disorders) and the timing of testing may influence patients’ performance on cognitive assessment scores. It is difficult to identify preserved and impaired functions in patient with aphasia. Oxford Cognitive Screen is a short and domain-specific cognitive screening tool for stroke survivors, which is user-friendly for patients with aphasia, apraxia, and neglect [66].
Assessment in the stroke unit
In the early phase after stroke, the cognitive assessment may not be an immediate priority. More detailed assessment is usually better conducted once a patient is in a stabilized condition. A shorter bedside evaluation with a sensitive screening instrument is often an appropriate choice. Since no test is significantly superior to another, choices of test should be informed by purpose of testing, feasibility, acceptability, and opportunity cost [67]. Compared to Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA) has a higher sensitivity and specificity for initial cognitive function screening [67, 68]. The National Institute of Neurological Disease and Stroke-Canadian Stroke Network (NINDS-CSN) protocols, including the 60-, 30-, and 5-min protocols, are all valid in discriminating stroke patients from normal controls [69, 70]. The NINDS-CSN 5-minute protocol consisting of only verbally conducted tests, could be administered in patients with dominant hand or visual field defect following stroke [71].
Assessment in outpatient clinic
Conducting assessment in cognitive performance 3 or 6 months after stroke appears to be appropriate. A predictive model for PSCI based on clinical and radiological datasets, including NIHSS score, multiple strokes, adjusted MMSE (MMSEadj) score and the deep white matter Fazekas score, could help clinicians to identify patients in need of comprehensive neuropsychological testing [72]. The use of telemedical assessments has also gained popularity for its accessibility and convenience, especially for stroke and disabled patients that require urgent screening and medical intervention [73]. In this context, both the Telephone Montreal Cognitive Assessment (T-MoCA) and Telephone Interview of Cognitive Status have shown validity for initial cognition evaluation and benefits for making subsequent clinical or research assessments.
PREVENTION AND TREATMENT STRATEGIES
Recognize the uniqueness of stroke in the development of PSCI
Nearly 90% of stroke [74] and 30% of dementia [75] are preventable. The decrease in stroke incidence has shown a strong correlation with a concomitant decrease in all types of dementia [76, 77]. As above-mentioned, symptomatic stroke as an independent, substantial, and potentially modifiable risk factor for all-cause dementia [78, 79], plays a key causal role in the development of PSCI. With the multifactorial etiology of PSCI, multidomain interventions targe-ting at several potentially modifiable risk factors (primary prevention), including vascular, lifestyle-related and psychosocial factors, should be imple-mented simultaneously for effective prevention/treatment [80].
Though no consistent evidence has found that intensive treatment of vascular risk factors is beneficial for the prevention of PSCI, the holistic approach and interdisciplinary studies to establish evidence-based interventions management of these risk factors in primary and secondary prevention are extremely warranted (Fig. 4), to narrow the gap between what we know and what is done [81, 82].
The cognitive trajectories and prevention of post-stroke cognitive impairment. The blue-solid line indicates the cognitive trajectories of individuals with stroke associated risk factors but without stroke occurrence. The yellow-solid line indicates the cognitive trajectories of individuals with stroke but without evident cognitive impairment in cognitive assessment three to six months after the index stroke. The red-solid line indicates the cognitive trajectories of individuals with stroke and cognitive impairment in cognitive assessment three to six months after the index stroke. The primordial and primary should be respectively initiated when the risk factors are identified. As for the secondary and tertiary prevention, the initiation should begin once index stroke occurs, or the assessments reveal evident cognitive impairment in three to six months after the index stroke.
Secondary prevention
Secondary prevention, in the context of PSCI, implies the disruption of disease progression before cognitive impairment becomes clinically recognizable in patients with stroke. Stroke or TIA survivors are at high risk of developing cognitive decline, recurrent stroke, and severe complications. Delivering acute stroke treatments and preventing stroke recurrence to decrease neuropathological damage could be beneficial for the prevention of PSCI [2].
General treatment of acute stroke
Treatment of acute stroke might not only immediately alleviate disability but also reduce the long-term risk of PSCI [22]. Selected patients treated with intravenous recombinant tissue plasminogen activator after stroke onset has shown improvement on survival and cognitive tests at 90 days [83, 84]. In patients with isolated posterior cerebral artery occlusion, intravenous thrombolysis plus conventional treatment showed potential better efficacy over conventional treatment alone on cognitive outcomes but may not as good as the effect of endovascular treatment [85]. Patients treated with endovascular therapy 6 to 16 hours after stroke have better Quality of Life outcome than patients treated with medical therapy alone, including superior cognition, better mobility, and more social participation [86]. Likewise, the REVASCAT (Endovascular Revascularization With Solitaire Device Versus Best Medical Therapy in Anterior Circulation Stroke Within 8 Hours) trial reported that thrombectomy could improve cognitive function [87] and health-related quality of life [88] in patients with acute stroke due to anterior-circulation large vessel occlusion. As compared with conventional treatment, better MoCA and MMSE performance were also observed in patients treated with thrombectomy in 3-month follow-up [89].
Preventing recurrent ischemic stroke
The risk of PSD is at least twice times higher after recurrent stroke compared to first ever stroke[20]. In a 12-year follow-up study, patients with PSD also had a shorter mean time to recurrent stroke (7.13 years) than patients without dementia (9.41 years) [90]. Thus, the prevention of recurrent ischemic stroke is particularly important to reduce the burden of cognitive impairment after stroke. Cilostazol may potentially improve cognitive function in patients with non-cardioembolic ischemic stroke during convalescent rehabilitation [91]. Although rate cases of cognitive impairment have been reported regarding the administration of statins, a recent meta-analysis concluded that post-stroke statin use was associated with decreased risk of cognitive impairment [92].
Delaying cognitive decline after stroke
Physical activity may help prevent cognitive decline in stroke survivors, but it is important to build a framework to guide selection of the time, the optimal exercise types, and intensities of physical activity for such purpose. Targeted exercise training has also been found with its effectiveness in promoting cognitive improvement through neuroplasticity processes adjustment [93, 94]. However, the AVERT (A Very Early Rehabilitation Trial) indicated that exposure to very early and more frequent mobilization after stroke (initiated within 24 hours of stroke onset and lasting for 14 days or until discharge from the acute stroke unit) was associated with reduced functional capacity for recovery without affecting cognition [95].
Tertiary prevention and interventions under investigation
The presence of numerous comorbidities complicates the delivery of home health care and magnifies the difficulties experienced by people with long-term conditions. In this section, we summarize the therapeutic options to improve cognitive function after stroke, including pharmacological and non-pharmacological approaches.
Pharmacological treatment
Over the past two decades, significant efforts have been made by the pharmaceutical industry in the discovery of novel drug agents, but to this date, ground-breaking therapy for cognitive impairment and dementia is still far from reach [96]. The most recent meta-analysis of 7 studies reported that acetylcholinesterase inhibitors (donepezil and rivastigmine) maintain a stable pattern of improved cognitive function compared to the placebo group through 24 weeks treatment in PSCI and VaD patients [97]. Recent meta-analyses for the treatment of VaD have demonstrated the potential efficacy of memantine [98], ginkgo biloba extract [99, 100], cerebrolysin [101], and traditional Chinese herbal medicine [102]. Besides, double-blind RCT have shown that DL-3-n-butylphthalide [103] was effective for improving cognitive and global functioning in patients with VCI, and actovegin [104] was effective for improving cognitive outcomes in patients with PSCI. It is acknowledged that these results warrant confirmation in additional robustly designed studies.
Non-pharmacological treatment
Non-pharmacological interventions with disease-modifying effects mostly involve psychological and physical rehabilitation strategies. An umbrella review retrieved 18 systematic reviews (129 RCTs; 7,985 participants) of interventional trials, and found that interventions including physical activity or cognitive rehabilitation showed a positive effect on cognitive function, while due to limited methodological quality the certainty of interventions was rated low [105]. Specifically, a pooled patient-level data analysis including 322 patients reported multidomain interventions to target lifestyle and vascular risk factors in stroke patients with cognition compared with standard care can improve the scores in trail making test A (TMT-A, assessing attention), but not in TMT-B (executive functions) one year after stroke [106]. A recent meta-analysis combing seven non-randomized controlled studies findings showed that the psychological interventions after stroke had small but statistically significant overall effects on cognition [107]. As a cognitive rehabilitation strategy, enriched environment has been reported with its potency to facilitate physical and social abilities in ischemic stroke-induced cognitive impairment [108]. Remote limb ischemic conditioning has been found to improve blood supply to the brain and cognitive function, in patients with first-ever non-cardioembolic ischemic stroke [109]. Whereas no effects were found in cardiac rehabilitation has shown on cognitive function [110]. Acupuncture could be used as additional therapies to improve the clinical outcomes of patients with PSCI [111]. A sequential combination of aerobic exercise and cognitive training also has preferable effects on cognitive outcomes of PSCI patients [112]. Furthermore, evidence suggests that high-frequency transcranial magnetic stimulation exerts significant neuroprotective and pro-cognitive effects by enhancing neurogenesis and activating BDNF/TrkB signaling pathway [113]. Transcranial magnetic stimulation and transcranial direct current stimulation are prone to be promising areas in the rehabilitation of cognitive impairment [114].
CONCLUSION AND RESEARCH PRIORITIES
Cognitive impairment affects a huge number of stroke survivors, has an enormous impact on patients, caregivers, and imposes a major burden to the society and economy. Greater efforts are urgently needed to improve its management. However, the heterogeneity in many aspects of the condition, the inadequate understanding of the pathophysiology, and a collective failure to identify effective treatments are all challenges.
Treatments to slow the progression of cognitive impairment to dementia and improve the quality of life in patients with PSCI are desperately needed. With a greater understanding of the multiple pathogenic processes of the post-ischemic cascade, a number of promising biological agents are being investigated [115]. On the one hand, concomitant vascular and neurodegenerative pathologies may double the risk of developing dementia; on the other hand, it indicates that a significant proportion of dementia could be mitigated, delayed, or prevented [116]. A feasible strategy to reduce stroke incidence and, once a stroke occurs, to treat acute stroke and prevent recurrence could be helpful.
In view of the above-mentioned facts, adjustments are also required for the ongoing and further study. More RCTs and observational studies in stroke patients were recommended to use cognition as a primary or secondary endpoint, rather than focus only on physical functions. Besides, the design of RCTs need to be adapted to the intended patient groups, with longer-term interventions, and larger sample size to reveal the potential protective effects against PSCI in at-risk patients [117]. Given that single-agent interventions might not be enough to affect cognition, simultaneously targeting several risk factors might be promising for prevention and treatment of PSCI.
Although PSCI could potentially be prevented and cognitive decline might be reversible, the current diagnosis and treatment strategies are far from being satisfactory in real-world practice. Future research on the cognitive effects of stroke should involve more extensive collaborations between physicians in vascular and cognitive neurology, cardiovascular medicine and neurorehabilitation, to optimize prevention, assessment, and treatment of PSCI [21].
Footnotes
ACKNOWLEDGMENTS
This study was supported by grants from the National Natural Science Foundation of China (82071201), Shanghai Municipal Science and Technology Major Project (No.2018SHZDZX01) and ZHANGJIANG LAB, Tianqiao and Chrissy Chen Institute, and the State Key Laboratory of Neurobiology and Frontiers Center for Brain Science of Ministry of Education, Fudan University.
