A Review of Risk Factors for Cognitive Impairment in Stroke Survivors

Table 1

Studies on cognitive impairment among stroke patients.

Study	Objective of study	Sample	Methodology	Setting	Prevalence	Outcome
						Demographic	Clinical	Psychological	Physical
Sachdev et al. [27]	(a) Identify determinants for CI; (b) Identify clinical features associated with development of VaMCI and VaD	169	Cross-sectional study	Hospital based	58% CI (VaD = 21.3%; VaMCI = 36.7%); Tools used to assess CI: (a) MMSE; (b) Neuropsychological assessments; (c) Consensus diagnosis by experts	(i) Older age (S); 60 years as cut-off; (ii) Lower education (NS); (iii) Female (NS)	Medical factors: (a) Lower NART-IQ (S) (premorbid intellectual ability); (b) Higher IQCODE (S) (preexisting CI); (c) Higher ESS (S) (stroke severity); (d) Previous stroke (NS); (e) Previous TIA (NS); Vascular factors: (a) Homocysteine (NS); (b) Higher CVRF (NS); Neurological factors: (a) Infarct volume (S); (b) Laterality of stroke (NS); (c) Cerebral hemisphere (NS); (d) Lacunar infarct (NS); (e) Brain atrophy (NS); (f) Number of infarcts (NS); (g) Higher WMH score (NS)	Depression (NS); Tools used: (a) GDS; (b) HDRS	NR; Tools used: (a) SOFAS; (b) ADL; (c) I-ADL
De Ronchi et al. [34]	Detect the impact of stroke on the occurrence of dementia and CIND in different age, sex, and education	7930	Cross-sectional study	Population based	11.6% CI (CIND = 5.1%; M: 4.1%; F: 5.7%) (dementia = 6.5%; M: 4.9%; F: 7.5%); Tools used to assess CI: (a) MMSE and Global Deterioration Scale; (b) DSM-III-R and experts opinion	(a) Age and education modify effect of stroke on dementia (the risk was twofold stronger in older (75+ years old) and young people (61–75 years old) who had low education (0–3 years of schooling) with stroke as compared with higher education (4+ years of schooling) with stroke); (b) Sex did not modify the effect of stroke on dementia	History of stroke increased the risk ratio for dementia and CIND	NA	NA
Saxena et al. [35]	Determine the prevalence of depressive symptoms and cognitive impairment in stroke patients at 3 phases of rehabilitation processes	200	Observational study	Hospital based	Admission = 54.5% CI; Discharge = 33.7%; Follow-up (6 months) = 40.4%; Tool used to assess CI: (a) AMT	(a) Older age (S) ≥76 years; (b) Divorced (S)	Medical factors: (a) Poststroke dysphasia (S); (b) Poststroke urinary incontinence (S); (c) Cognitive impairment on admission (S); Vascular factor: (a) Ischemic heart disease (S)	Depressive symptoms on admission (S); Tool used: (a) GDS	Severe physical functioning (S); Tool used: (a) BI
Cederfeldt et al. [36]	Examine the differences in performance of P-ADL in relation to cognitive impairment at pre- and postdischarge	45	Longitudinal study	Hospital based & community based	Acute phase = 29% CI; 6 months = 20% CI; 12 months = 10% CI; Tool used to assess CI: (a) MMSE; (b) CIMP-QUEST; (c) Neuropsychological battery	NA	NA	NA	(a) Intact cognition (before & after): improved P-ADL; (b) Impaired cognition (before & after): nonimproved P-ADL; Tool used: (a) BI
Zhou et al. [37]	Identify the frequency and risk factors of cognitive impairment after stroke	434	Cross-sectional study	Hospital based	Poststroke CI = 37.1%; Stroke-related CI = 32.2%; First-ever stroke CI = 29.6%; Tool used to assess CI: (a) MMSE; (b) IQCODE	Personal factors: (a) Older age (S) (8.5 years older than cognitive intact group); (b) Lower education (S) (≤6 years); Life style factor: (a) Everyday alcohol drinking (S)	Medical factors: (a) History of stroke (S); (b) Dysphasia (S); Neurological factor: (a) Left carotid infarction (S)	NA	NA
Hurford et al. [38]	Examine domain-specific patterns of cognitive change after ischemic stroke	209	Cross-sectional study	Hospital based	Changes in the prevalence of cognitive impairment in each cognitive domain from less than 1 month to over 3 months:
					(a) Speed and attention: 72.4% to 37.9%
					(b) Frontal executive function: 34.4% to 16.2%
					(c) Nominal skills: 30.2% to 8.1%
					(d) Perceptual skills: 29.5% to 8.1%
					(e) Visual memory: approximately 18% to 10%
					(f) Verbal memory: approximately 28% to 18%
					Tool used to assess CI:
					(a) Neuropsychological tests battery
					No determinants on demographic, clinical, psychological, and functional data were reported
Claesson et al. [39]	Explore the impact of cognitive impairment on ADL performances, utilisation, and costs of health care	149	Cross-sectional study	Hospital based	CI = 72% (dementia: 28%); Tool used to assess CI: (a) Rating of cognitive symptoms; (b) DSM III-R	NA	NA	NA	Cognitively impaired: high dependency in I-ADL (i.e., continence, indoor mobility, toilet management, transfer, dressing/undressing, grooming, cooking, bath/shower, housework, and outdoor mobility); Tool used: (a) BI; (b) SI
Liman et al. [40]	(a) Determine the frequency and predictors of CI; (b) Examine the prevalence and factors associated with cognitive recovery and cognitive stability	630	Longitudinal study	Population based	(a) CI at 3 months = 14.8%; (b) CI at 1 year = 13.3%; (c) CI at 3 years = 11.8%; Tool used to assess CI: (a) MMSE	(a) Predictor at 3 months after stroke: Older age (S); (b) Predictor for cognitive stability up to 3 years: Younger age (S)	(a) Predictor at 3 months after stroke: Diabetes mellitus (S)	NA	(a) Predictor at 3 months after stroke: stroke severity on day 7/BI (S); (b) Predictor for cognitive stability up to 3 years: less stroke severity/high BI (S); Tool used: (a) BI
Stephens et al. [41]	Determine the relationship between attention, executive performance, and memory impairments with ADL impairments	339	Cross-sectional study	Hospital based	MCI = 19%; CIND = 26%; AACD = 44%; Tool used to assess CI: (a) CAMCOG; (b) CRT task performance; (c) DSM-III-R; (d) Experts’ opinion	NA	NA	Depression is positively correlated with 18 BADLS items; Tool used: (a) GDS	(a) CRT = disabilities in basic self-care; (b) Executive function = disabilities in intermediate self-care; (c) MMSE = disabilities in complex self-management; (d) Memory impairment: not associated with any disabilities in ADL; Tool used: (a) BADLS
Tang et al. [42]	Determine the relationship between CMBs and CIND reversion	328	Longitudinal study	Hospital based	Impairment at baseline: (a) Visuomotor speed = 60.6%; (b) Executive function = 49.6%; (c) Visual memory = 29.9%; (d) Verbal memory = 29.1%; (e) Visual construction = 20.5%; (f) Language = 13.4%; (g) Attention = 6.3%; Tools used to assess CI: (a) Modified VDB; (b) FAB; (c) MMSE; (d) DSM-IV	Determinant of reversion of CIND: (a) Younger age (S); (b) Higher education (NS); Determinant of reversion of language domains: (a) Younger age (NS)	Determinants of reversion of CIND: (I) Medical factors: (a) Less visual constructional impairment (NS); (b) Less verbal memory impairment (NS); (c) More likely having visual memory impairment (NS); (d) More likely having executive function impairment (NS); (II) Vascular factors: (a) Absence of CMBs (S); (b) Absence of lobar CMBs (NS); (c) Hypertension (NS); (III) Neurological factors: (a) Small volume of WMHs (NS); Determinants of reversion of language domains: (a) Absence of CMBs (NS); (b) Small volume of WMHs (NS)	NA	NA
Akbari et al. [43]	(a) Investigate whether test performance in neurological and cognitive areas is able to predict daily task performance; (b) Examine whether the potential of tool measures can predict functional outcomes	27	Cross-sectional study	Population based	NR; Tool used to assess cognitive status: (a) LOTCA	(a) Stroke severity (i.e., motor impairment) correlates with dependency in ADL performance; (b) Self-care activities correlate with general cognitive performance (total score of LOTCA); (c) Visual perception and visuomotor organization correlated with BI components; (d) Determinants on demographic, clinical, psychological, and physical data were not reported; (e) LOTCA is not suitable to predict dependency in BADLS performance after stroke; Tool used: (a) Neurological impairment = NIHSS; (b) ADL performance = BI
Narasimhalu et al. [44]	Determine neuroimaging measures (i.e., infarcts, WMH, and neurodegeneration) associated with subjective cognitive impairment (SCI) in cognitively intact patients with lacunar stroke	145	Cross-sectional study	Hospital based	SCI = 30.9%; NCI = 69.1%; Tools used to assess CI: (a) MMSE; (b) MoCA; (c) FAB	(a) Age (NS); (b) Education (NS)	Medical factors: (a) MMSE (S); Vascular factors: (a) Diastolic BP (NS); Neurological factors: (a) Basal ganglia infarct (S); (b) Temporal lobe infarct (NS)	Depression (NS); Tool used: (a) PHQ-9	NR; Tool used: (a) Modified Rankin score
Khedr et al. [45]	(a) Determine the relative frequency of first-ever PSD; (b) Determine the risk factors of PSD; (c) Determine the relationship between total Hcy level and PSD	81	Cross-sectional study	Hospital based	PSD = 21%; (PSD in ischemic stroke: 76.5%, PSD in hemorrhagic stroke: 23.5%); Tools used to assess CI: (a) Neuropsychological tests battery; (b) MMSE; (c) CASI; (d) WMS-R; (e) DSM-IV	Personal factors: (a) Older age (S); (b) Lower education (S); (c) Family history of dementia (S); Life style factor: (a) Smoking (S)	Neurological factors: (a) Large size of infarction (S); (b) Lacunar infarct (S); (c) Dominant hemispheric lesion (S); Vascular factors: (a) Hypertension (S); (b) High Hcy level (S); (c) Atrial fibrillation (NS); (d) Ischemic heart disease (NS); (e) Carotid stenosis (NS)	Depression (NS); Tool used: (a) HDRS	Motor and functional disability (S); Tools used: (a) SSS; (b) BI
Makin et al. [33]	Determine the factors associated with the progression of cognitive impairment after stroke	193	Longitudinal study	Hospital based	3 months: (a) Cognitive intact = 72%; (b) CIND = 9.3%; (c) Dementia = 18.6%; Tool used to assess CI: (a) SPMSQ; (b) IQCODE; (c) DSM-IV; (d) Experts’ opinion	(a) Older age (S); (b) Polydrug used (S)	Medical factor: (a) Previous CI (IQCODE) (S); Vascular factors: (a) Low diastolic pressure on admission (S); (b) Hypotension during admission (S)^#	Depressive symptoms: not associated with cognitive evolution (progress or no progress); Tool used: (a) CDR; (b) CES-D	NR; Tool used: (a) BI
Čengić et al. [46]	Analyze and compare motor and cognitive impairment in stroke patients at acute, subacute, and chronic phases	50	Cross-sectional study	Hospital based	CI at acute, subacute, and chronic phases = 12%; Tools used to assess CI: (a) Modified MMSE; (b) SKT	Personal factors: (a) Older age (NS) (patient’s age under 75 years); (b) Lower education (S) (partially or fully completed elementary school); (c) Heredity (S); Life style and habits: (a) Smoking (S); (b) Alcohol drinking (S); (c) Obesity (S)	Vascular factors: (a) Hypertension (S); (b) Hyperlipoproteinemia (S); (c) Diabetes (S); (d) Heart disease (S)	Stress (S)	(a) Physically inactive (S); (b) Acute and subacute phases: better motor recovery and better cognitive status; Tool used: (a) ESS
Douiri et al. [32]	Evaluate the prevalence of cognitive impairment after first-ever stroke up to 15 years	4212	Longitudinal study	Population based	(a) CI at 3 months = 24%; (b) CI at 5 years = 22%; (c) CI at 10 years = 18%; (d) CI at 14 years = 21%; Tool used to assess CI: (a) MMSE; (b) AMT	(a) Older age (S); (b) Ethnicity (black peoples) (S); (c) Socioeconomic status (manual worker) (S)	Neurological factors: (a) Lacunar infarct (S); (b) Small vessel occlusion (S)	NA	NR; Tool used: (a) BI
Knopman et al. [47]	Determine the association of history of stroke with the diagnosis of MCI or cognitive impairment	2050	Cross-sectional study with case-control	Population based	MCI = 10.9%; Tool used to assess CI: (a) Neuropsychological tests battery; (b) DSM-IV	(I) Association of stroke with MCI:
						(a) History of stroke was associated with a higher risk of MCI (adjusted for age, sex, and education)
						(b) Association between history of stroke and MCI subtypes (aMCI and naMCI) did not change when diabetes, coronary heart disease, APOE genotype, and hypertension were added to the model
						(c) History of stroke was associated with both aMCI and naMCI, while APOE ɛ4 genotype was associated with aMCI only
						(II) Association of stroke with cognitive domains:
						(a) History of stroke was significantly associated with lower cognitive function in other domains (language, executive, and visuospatial) except memory
						(b) The magnitude of the association was strongest for the executive function domain in unadjusted analyses
						(c) Association was elevated about 2-fold for language and visuospatial domains after being adjusted for age, sex, and education
						(d) Association of stroke with language, executive, and visuospatial domains did not change when diabetes, coronary heart disease, APOE genotype, and hypertension were added to the model
						(e) APOE ɛ4 genotype was only associated with poor performance in the memory domain
Sundar and Adwani [48]	(a) Assess cognitive dysfunction at 3 months after ischemic stroke; (b) Assess frontal executive function using MMSE; (c) Evaluate the degree and type of cognitive dysfunction in ischemic stroke subgroups	164	Cross-sectional study	Hospital based	Cognitive dysfunction = 31.7%; (17.07% had been impaired on frontal executive functions only); Tool used to assess CI: (a) Modified Folstein’s MMSE; (b) FAB	(a) Memory was significantly and commonly affected in multi-infarct strokes as compared to single infarcts; (b) Frontal executive dysfunction was not significantly different in single versus multiple infarcts and cortical versus subcortical infarcts; (c) Number of infarcts did not appear to influence cognitive dysfunction at 3 months of stroke; (d) Determinants on demographic, clinical, psychological, and physical data were not reported
Jokinen et al. [49]	Explore the severity and location of WMHs as predictor of neuropsychological test performance	323	Cross-sectional study	Hospital based	Dementia = 14.6%; Tool used to assess CI: (a) Neuropsychological test battery; (b) DSM-III-R	Predictors of neuropsychological deficits: (a) Age (S); (b) Education (S)	Predictors of neuropsychological deficits: (a) Total infarct volume (S); (b) Cerebral WMHs (S)^#; (c) Cortical atrophy (S)^#; Predictors of WMHs: (a) Poor executive functions (S)^#; (b) Poor speed of mental processing (S)^#; (c) Poor visual memory (S); (d) Poor delayed recall of objects (S); (e) Poor visuospatial task (S); (f) Poor short-term memory (NS); (g) Poor story recall (NS); (h) Poor verbal conceptualization (NS)	NA	NA
Cao et al. [50]	(a) Identify the neuropsychological impairments; (b) Identify the clinical characteristics related to cognitively impaired patients	40	Cross-sectional with case-control study	Hospital based	(a) Dementia = 12.5%; (b) CI = 7.5%; (c) Partial CI = 20%; Tool used to assess CI: (a) MMSE; (b) Neuropsychological tests battery	(a) Lower education level is positively correlated with cognitive performance (global/partial impairment)
						(b) Token test, RPM, and AVLT delay and similarities were more often significantly failed tests by patients than control. However, these tests did not correlate with the number and site of lesions, ultrasound pattern, and neurological conditions
						(c) No correlation between size, number and side of lesions within demented patients, globally or partially impaired patients, and etiological diagnosis of stroke
						(d) Dementia and CI were associated with a lower BI score
						Tool used:
						(a) Daily activity abilities: BI
						(b) Depression scale: SDS
Mizrahi et al. [51]	Evaluate the relationship between diabetes and overall cognitive status in patients with ischemic stroke	707	Retrospective study	Hospital based	CI: NR; Tool used to assess CI: (a) MMSE	(a) Older age (S); (b) Gender (female) (S)	Medical factor: (a) Previous stroke (S); Vascular factors: (a) Dementia (S); (b) Diabetes mellitus (S)	NA	NA
Mizrahi et al. [52]	Evaluate the relationship between atrial fibrillation (AF) and overall cognitive status in patients with ischemic stroke	707	Retrospective study	Hospital based	CI: NR; Tool used to assess CI: (a) MMSE	(a) Older age (S); (b) Gender (female) (S)^#	Medical factor: (a) Previous stroke (S); Vascular factors: (a) Diabetes mellitus (S); (b) Dementia (S); (d) Atrial fibrillation (S)^#	NA	NA
Tang et al. [42]	Examine the frequency and clinical determinants of poststroke cognitive impairment in Chinese stroke patients without dementia	179	Cross-sectional study	Hospital based	CI: 21.8% after 3-month stroke; Tool used to assess CI: (a) MMSE	(a) Lower education (S); (b) Gender (female) (S)^#	Medical factors: (a) NIHSS dysarthria score (S); (b) Urinary incontinence (S); Vascular factor: (a) Atrial fibrillation (S)^#	NA	NA
Sachdev et al. [53]	(a) Investigate neuropsychological features of the VaMCI and its progression over 3 years among stroke patients without dementia; (b) Investigate risk factors for VaMCI conversion to dementia; (c) Examine relationship of MRI measures with conversion	104 patients; 84 controls	Longitudinal study	Hospital based	Dementia over 3 years: (a) VaMCI subject: 24.4%; (b) NCI subject: 8.5%; Tool used to assess CI: (a) Consensus diagnosis by experts based on cognitive domains, functional status, and vascular etiologies	(i) Clear determinants of progression did not emerge
						(ii) Neuropsychological impairment at baseline tended to predict greater decline
						(iii) Global cognitive and functional impairment at baseline may be of importance in predicting dementia
						(iv) Converters and nonconverters of VaMCI to VaD did not differ by age, sex, education, burden of vascular risk factors, or structural changes in brain
						(v) VaMCI group had more vascular risk factors and more white matter hyperintensities at baseline than the NCI and control groups
						(vi) Neuropsychological factor: greater decline of logical memory in VaMCI group
						(vii) MRI measures: stroke patients had larger volumes of total, deep, periventricular WMHs and smaller amygdala volume (VaMCI group)
						(a) Tools used in neuropsychological assessments: WMS-R, WAIS-R, Boston Naming Test, TMT, SDMT, Western Aphasia Battery, and NART
						(b) Tools used in medical and psychiatric assessments: SOFAS, ADL, I-ADL, ESS, GHQ, GDS, HDRS, and Neuropsychiatric Inventory
Tu et al. [17]	(a) Explore the prevalence and effects of vascular cognitive impairment (VCI) among ischemic stroke patients; (b) Provide a basis for prevention and treatment strategies	689	Cross-sectional study with control group	Community based	VCI: 41.8%; (a) VaCIND: 32.1%; (b) VaD: 9.7%; Tool used to assess CI: (a) MMSE and MoCA; (b) Criteria in NINDS and AIREN; (c) Expert consensus	Personal factors: (a) Older age (S); (b) Low level of education (S); (c) Professional worker (S); (d) Living alone (S)^#; Behaviour and life style: (a) High alcohol intake (S); (b) Lack of hobbies (S)^#; (c) Longer sleep (S); (d) Irregular health check-up (S); Dietary habits: (a) Not eating fruit/vegetables (S); (b) Not drinking milk (S); (c) Not drinking tea (S)	Medical factors: (a) Family history of stroke (S); (b) Aconuresis (S); Vascular factors: (a) Hyperlipidemia (S)^#; (b) TIA (S); Neurological factors: (a) High level of paraventricular WML (S); (b) Macroangiopathy disease (S); (c) Brain atrophy (S)^#	NA	NA
Zhang et al. [54]	Examine the incidence, neuropsychological characteristics, and risk factors of cognitive impairment 3 months after stroke in China	577	Cross-sectional study	Hospital based	PSCI: 30.7%; (a) Visual impairment = 22.4%; (b) Executive impairment = 11.6%; (c) Memory impairment = 10.4%; (d) Attention impairment = 3.1%; Tools used to assess CI: (a) MMSE and MoCA; (b) FOM, RVR, BD, and DS in WAIS-R; (c) Consensus criteria and experts’ opinion	Personal factors: (a) Older age (S) ≥65 years old; (b) Low education level (S)^#(<7 years)	Medical factor: (a) Stroke severity: 3-month poststroke (S); Vascular factor: (a) Obesity (S)^# (waist circumference = men ≥ 102 cm; women ≥ 98 cm); Neurological factor: (a) Cortex lesion (S)	Depressive symptom (S); Tool used to assess CI: (a) CES-D	NA
Narasimhalu   et al. [44]	(a) Compare cognitive performance and quality of life (QOL) in stroke survivors and controls; (b) Examine predictors for cognitive function and QOL	109	Cross-sectional with case-control study	Hospital based	CI = 17.4%; Tool used to assess CI: (a) Modified MMSE	(a) Older age (NS); (b) Sex (NS); (c) Lower education (NS)	Medical factors: (a) Previous psychiatric illness (NS); (b) Stroke duration (NS); (c) Paresis (NS)	Psychiatric morbidity (S); Tool used to assess CI: (a) GHQ-30	NA

Significant determinant in multivariate analysis; ^#strong predictor; ⁺basic self-care (i.e., transferring, dressing, hygiene, teeth cleaning, going to toilet, bathing, eating, mobility, and drink preparation); intermediate self-care (i.e., shopping, house/gardening, transport, and games/hobbies); complex self-management (i.e., finances, oriented to time, use telephone, and communication).
S = significant determinant; NS = nonsignificant determinant; NR = not reported; M = male; F = female; NA = not available; P-ADL = personal activities of daily living; I-ADL = instrumental activities of daily living; WMHs = white matter hyperintensities; WML = white matter lesion; ADL = activities of daily living; BI = Barthel Index; CI = cognitive impairment; MCI = mild cognitive impairment; VaMCI = vascular mild cognitive impairment; SCI = subjective cognitive impairment; NCI = no cognitive impairment; PSCI = poststroke cognitive impairment; VaD = vascular dementia; ESS = European stroke scale; NART = national adult reading test; NART-IQ = national adult reading test-intelligence quotient; IQCODE = informant questionnaire for cognitive decline in the elderly; CVRF = cardiovascular risk factor; TIA = transient ischemic attack; CIND = cognitive impairment no dementia; VaCIND = vascular cognitive impairment no dementia; AACD = age-associated cognitive decline; BADLS = Bristol activities of daily living scale; CRT = choice reaction time; CMBs = cerebral microbleeds; BP = blood pressure; FAB = frontal assessment battery; SSS = Scandinavian stroke scale; MMSE = mini mental state examination; MoCA = Montreal cognitive assessment; PSD = poststroke dementia; Hcy = plasma homocysteine; SPMSQ = short portable mental status questionnaire; aMCI = amnestic mild cognitive impairment; naMCI = nonamnestic mild cognitive impairment; RPM = Raven’s progressive matrices; AVLT = auditory-verbal learning test; BSRT = Babcock story recall test; DST-B = digit span test backward; GDS = geriatric depression scale; HDRS = Hamilton depression rating scale; AMT = abbreviated mental test; CIMP-QUEST = cognitive impairment questionnaire; CDR = clinical dementia rating scale; SI = Sunnaas index of ADL; CAMCOG = Cambridge assessment of mental disorder in the elderly; VDB = vascular dementia battery; FAB = frontal assessment battery; LOTCA = Loewenstein occupational therapy cognitive assessment; PHQ-9 = patient health questionnaire; CASI = cognitive abilities screening instruments; WMS-R = Wechsler memory scale-revised; CES-D = Center for Epidemiologic Studies depression scale; SKT = short cognitive performance test for assessing memory and attention; SDS = self-rating depression scale; NIHSS = National Institutes of Health Stroke Scale; TMT = trail making test; SDMT = symbol digit modalities test; SOFAS = social and occupational functioning scale; GHQ = general health questionnaire; NINDS = National Institute of Neurological Disorders and Stroke; AIREN = Association Internationale pour la Recherche et L’Enseignement en Neurosciences; FOM = Fuld object memory test; RVR = rapid verbal retrieval; WAIS-R = Wechsler adult intelligence scale-revised; BD = block design; DS = digit span; DSM = diagnostic and statistical manual; MRI = magnetic resonance imaging.