Ischemic Stroke Incidence and Severity and Poststroke Cognitive Decline and Incident Dementia

Silvia Koton; Alden L Gross; Hugo J Aparicio; Alexa S Beiser; Emily M Briceño; Josef Coresh; Mitchell S V Elkind; Bruno J Giordani; Rebecca F Gottesman; Rodney A Hayward; Virginia J Howard; Michelle C Johansen; Ronald M Lazar; Mellanie V Springer; Robert J Stanton; Jeremy Sussman; Hang Wang; Rachael T Whitney; Wen Ye; Deborah A Levine

doi:10.1001/jamanetworkopen.2026.8900

. 2026 Apr 16;9(4):e268900. doi: 10.1001/jamanetworkopen.2026.8900

Ischemic Stroke Incidence and Severity and Poststroke Cognitive Decline and Incident Dementia

Silvia Koton ^1,^2,^3,^✉, Alden L Gross ³, Hugo J Aparicio ⁴, Alexa S Beiser ^4,⁵, Emily M Briceño ⁶, Josef Coresh ¹, Mitchell S V Elkind ⁷, Bruno J Giordani ⁸, Rebecca F Gottesman ⁹, Rodney A Hayward ^10,¹¹, Virginia J Howard ¹², Michelle C Johansen ¹³, Ronald M Lazar ¹⁴, Mellanie V Springer ¹⁵, Robert J Stanton ¹⁶, Jeremy Sussman ^10,¹¹, Hang Wang ³, Rachael T Whitney ¹⁰, Wen Ye ¹⁷, Deborah A Levine ^10,¹⁴

¹Departments of Population Health and Medicine, New York University Grossman School of Medicine, New York

²Department of Nursing, Gray Faculty of Medical and Health Sciences, Tel Aviv University, Tel Aviv, Israel

³Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland

⁴Department of Neurology, Boston University Chobanian and Avedisian School of Medicine, Boston, Massachusetts

⁵Department of Biostatistics, Boston University School of Public Health, Boston University School of Medicine, Boston, Massachusetts

⁶Department of Physical Medicine and Rehabilitation, University of Michigan, Ann Arbor

⁷American Heart Association, Dallas, Texas

⁸Department of Psychiatry, University of Michigan, Ann Arbor

⁹Stroke Branch, National Institute of Neurological Disorders and Stroke Intramural Research Program, Bethesda, Maryland

¹⁰Department of Internal Medicine, University of Michigan, Ann Arbor

¹¹VA Ann Arbor Healthcare System, Ann Arbor, Michigan

¹²Department of Epidemiology, School of Public Health, University of Alabama at Birmingham

¹³Department of Neurology, The Johns Hopkins University School of Medicine, Baltimore, Maryland

¹⁴Department of Neurology, University of Alabama at Birmingham

¹⁵Department of Neurology, University of Michigan, Ann Arbor

¹⁶Department of Neurology and Rehabilitation Medicine, University of Cincinnati, Cincinnati, Ohio

¹⁷Department of Biostatistics, University of Michigan, Ann Arbor

Accepted for Publication: March 3, 2026.

Published: April 16, 2026. doi:10.1001/jamanetworkopen.2026.8900

Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License, which does not permit alteration or commercial use, including those for text and data mining, AI training, and similar technologies. © 2026 Koton S et al. JAMA Network Open.

^✉

Corresponding Author: Silvia Koton, PhD, MOccH, RN, Optimal Aging Institute, NYU Langone Health, 227 E 30th St, 7th Floor, New York, NY 10016 (silvia.koton@nyulangone.org).

Author Contributions: Prof Koton and Dr Gross had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Koton, Gross, Coresh, Elkind, Giordani, Hayward, Sussman, Ye.

Acquisition, analysis, or interpretation of data: Koton, Gross, Aparicio, Beiser, Briceño, Giordani, Gottesman, Howard, Lazar, Johansen, Springer, Stanton, Wang, Whitney, Ye, Levine.

Drafting of the manuscript: Koton, Gross, Coresh, Giordani.

Critical review of the manuscript for important intellectual content: Koton, Gross, Aparicio, Beiser, Briceño, Elkind, Giordani, Gottesman, Hayward, Howard, Johansen, Lazar, Springer, Stanton, Sussman, Wang, Whitney, Ye, Levine.

Statistical analysis: Koton, Gross, Hayward, Wang, Whitney, Ye.

Obtained funding: Aparicio, Coresh, Howard, Levine.

Administrative, technical, or material support: Beiser, Briceño, Howard, Lazar, Sussman, Whitney, Levine.

Supervision: Koton, Gross, Giordani, Stanton, Sussman, Levine.

Conflict of Interest Disclosures: Dr Briceño reported receiving grants from the National Institutes of Health (NIH) during the conduct of the study and grants and personal fees from NIH outside the submitted work; she is also an associate editor for Neuropsychology and editorial board member for the Journal of Clinical and Experimental Neuropsychology. Dr Howard reported receiving grants paid to her institution from the NIH National Institute of Neurological Disorders and Stroke (NINDS) and National Institute on Aging (NIA) during the conduct of the study. Dr Johansen reported receiving grants from NINDS outside the submitted work. Dr Lazar reported receiving grants from NIH and McKnight Brain Research Foundation and personal fees from Eisai during the conduct of the study. Drs Springer, Stanton, and Sussman reported receiving grants from NIH during the conduct of the study. Dr Levine reported receiving grants from Northwestern University and consulting fees from Tufts University outside the submitted work. No other disclosures were reported.

Funding/Support: This study was funded by grant RF1AG068410 from the NIA, NIH, US Department of Health and Human Services (DHHS). The Atherosclerosis Risk in Communities (ARIC) study is performed as a collaborative study supported by contracts 75N92022D00001, 75N92022D00002, 75N92022D00003, 75N92022D00004, and 75N92022D00005 from the National Heart, Lung, and Blood Institute (NHLBI). The ARIC Neurocognitive Study is supported by grants U01HL096812, U01HL096814, U01HL096899, U01HL096902, and U01HL096917 from NIH (NHLBI, NINDS, NIA, and National Institute on Deafness and Other Communication Disorders). The Reasons for Geographic and Racial Differences in Stroke (REGARDS) project is supported by cooperative agreement U01NS041588 cofunded by NINDS and NIA, NIH, DHHS. The Framingham Heart Study is supported by grant 75N92019D00031 from NHLBI, with additional support from grants HHSN268201500001, R01HL107385, and N01HC251952 from NHLBI and U19AG023122 from NIA, NIH.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of NINDS or NIA.

Data Sharing Statement: See Supplement 2.

Additional Contributions: The authors thank the staff and participants of the ARIC Study for their important contributions. The authors thank the other investigators, the staff, and the participants of the REGARDS study for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at https://www.uab.edu/soph/regardsstudy/.

^✉

Corresponding author.

PMCID: PMC13087813 PMID: 41989779

Key Points

Question

Are stroke incidence and severity associated with long-term cognitive decline and dementia risk?

Findings

In this cohort study 42 342 participants, higher stroke severity was associated with progressively steeper cognitive decline. Compared with participants without stroke, the risk of dementia was significantly increased in those with stroke, with nearly twice the risk for minor stroke, more than 3 times the risk for mild to moderate stroke, and 5 times the risk for moderate to severe stroke.

Meaning

These findings underscore the importance of identifying mechanisms for the observed associations.

This cohort study examines the association between ischemic stroke incidence and severity and poststroke cognitive decline and dementia risk among participants from 3 US prospective cohort studies.

Abstract

Importance

The association between stroke severity and dementia is well established. However, reports on trajectories of cognitive decline comparing stroke survivors with individuals without stroke in large cohorts are insufficient.

Objectives

To examine associations of ischemic stroke incidence and severity with cognitive decline and dementia risk and to explore whether vascular risk factors modify these associations.

Design, Setting, and Participants

This cohort study pooled longitudinal data on cognitive function of participants aged 45 years or older and without stroke and dementia at baseline from 3 US prospective cohorts: the Atherosclerosis Risk in Communities study (1987-2019), Framingham Offspring Study (1971-2019), and Reasons for Geographic and Racial Differences in Stroke study (2003-2019). First definite ischemic strokes were reported in each cohort using consistent protocols, with severity defined using the National Institutes of Health Stroke Scale (NIHSS). The data analysis was completed February 27, 2026.

Exposure

Incident ischemic stroke categorized as minor (NIHSS 0-5), mild to moderate (NIHSS 6-10), or moderate to severe (NIHSS ≥11).

Main Outcomes and Measures

The primary outcomes were decline in global cognition and incident dementia. Secondary outcomes were changes in memory and executive function. Multivariable linear mixed-effects models were used to test the association of stroke incidence and severity with cognitive decline.

Results

A total of 42 342 participants from the pooled cohorts were included (mean [SD] age, 61.3 [9.8] years; 55.0% female). Longitudinal cognitive testing data were available for a median of 11.1 years (range, 0-29.7 years) with 397 344 person-years of observation for dementia incidence. Stroke severity data were available for 1055 of 1505 first-ever ischemic stroke survivors (70.1%). Compared with participants with no stroke, adjusted hazard ratios for incident dementia were 1.93 (95% CI, 1.52-2.45) for NIHSS 0 to 5, 3.26 (95% CI, 1.93-5.53) for NIHSS 6 to 10, and 5.06 (95% CI, 2.71-9.45) for NIHSS 11 or higher. Over the follow-up, higher stroke severity was associated with progressively steeper cognitive declines across all domains, with more prevalent dose-response associations for global cognition (ranging from a mean −0.18 [95% CI, −0.19 to −0.18] points per year for no stroke to −0.58 [95% CI, −0.73 to −0.42] points per year for moderate to severe stroke) and memory (ranging from a mean −0.15 [95% CI, −0.16 to −0.14] points per year for no stroke to −0.36 [95% CI, −0.51 to −0.21] points per year for moderate to severe stroke) than for executive function (ranging from a mean −0.33 [95% CI, −0.34 to −0.32] points per year for no stroke to −0.52 [95% CI, −0.66 to −0.39] points per year for moderate to severe stroke).

Conclusions and Relevance

This large cohort study of participants from 3 prospective cohorts found that greater stroke severity was associated with substantially elevated dementia risk and accelerated decline in global cognition, memory, and executive function. These findings underscore the critical importance of stroke prevention, particularly severe stroke, and identifying mechanisms that may link stroke to cognitive decline.

Introduction

Cognitive impairment and dementia are common following stroke.¹ Prevalence rates of cognitive impairment after stroke vary widely, depending on the timing of assessment and definitions used,² with estimates ranging from 20% to more than 80% in stroke survivors.^{3,4,5,6,7,8,9} Additionally, approximately 25% of patients develop incident dementia within the first year after stroke.¹⁰ These high rates, combined with increased longevity, rising stroke incidence among older adults, and declining mortality due to therapeutic advances,¹¹ emphasize the importance of poststroke cognitive decline for health in older adults.

This study adds to the well-documented cognitive impairment after stroke^12,13 by leveraging data from 3 US prospective cohorts (Atherosclerosis Risk in Communities [ARIC],¹⁴ Framingham Offspring Study [FOS],¹⁵ and Reasons for Geographic and Racial Differences in Stroke [REGARDS]¹⁶) with adjudicated strokes; information on stroke severity; and harmonized, longitudinal cognitive assessments before and after stroke, as well as long-term longitudinal cognitive data for individuals without stroke. Although the association between stroke severity and dementia is well established, most previous studies lacked long-term longitudinal cognitive data on cohort participants with stroke as well as those without stroke. Stroke severity is a major risk estimator of poststroke cognitive outcomes¹⁷; however, age, sex,¹⁸ comorbidities,^19,20 vascular risk factors,²¹ and stroke etiology²² are also associated with poststroke cognitive impairment and dementia.^{2,17,21,23,24} Complementary evidence from population-based and cardiovascular cohorts without stroke has shown that vascular risk factors, including blood pressure, body mass index (BMI), and blood glucose, are associated with brain changes, cognitive decline, and dementia.^{20,23,25,26,27,28} Prior stroke studies have indicated that these vascular risk factors may mediate or modify the association between stroke severity and subsequent cognitive impairment.²⁹ More severe stroke may worsen vascular and metabolic status, leading to ongoing small vessel disease, white matter damage, and neurodegeneration; these processes can act as mediators through which initial stroke severity contributes to later cognitive impairment and dementia.^25,26

Accordingly, we assessed the association of stroke incidence and severity with longitudinal cognitive decline and incident dementia and examined the potential influence of poststroke vascular risk factors on these associations. We hypothesized that a graded increase in risk across severity categories from no stroke to minor stroke (National Institutes of Health Stroke Scale [NIHSS] score 0-5), mild to moderate stroke (NIHSS score 6-10), and moderate to severe stroke (NIHSS score ≥11) would inform which patients may benefit from intensified monitoring and risk factor management after the acute phase.

Methods

This cohort study was approved by the University of Michigan Institutional Review Board. Participating institutions’ ethics review boards approved the cohort studies. Participants provided written informed consent before inclusion in the cohort studies. The study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.

Setting and Participants

Data on stroke severity were collected from ARIC (1987-2019), FOS (1971-2019), and REGARDS (2003-2019). Individuals younger than 45 years at baseline, lacking cognitive evaluations before and after stroke, and having prestroke dementia or prevalent stroke were excluded. The final analytic sample is shown in the eFigure in Supplement 1. Only events adjudicated in each cohort as ischemic stroke and defined as definite ischemic stroke^16,30,31,32 were included. First-ever definite ischemic strokes were identified in the harmonized pooled data.

Exposure Variables

Stroke incidence was defined as the first definite ischemic stroke, adjudicated within each cohort,³³ occurring after baseline cognitive testing and before the final follow-up visit. Stroke was evaluated as a time-varying exposure, with participants with stroke contributing nonstroke and poststroke time. Stroke severity was assessed using the NIHSS, a validated measure considered the criterion standard for clinical trials.³⁴ Each participating cohort used similar protocols to determine the severity of stroke by NIHSS on admission,^35,36 a measure recognized as providing reliable estimates of functional outcomes and mortality after stroke.³⁷ Deidentified hospital medical records were reviewed by trained physicians, and data on all items in the NIHSS were evaluated according to an algorithm for retrospective collection of the NIHSS score,^35,36 which has been shown to be valid across the entire stroke severity spectrum.^38,39 To address potential nonlinear scaling of stroke severity and for clinical utility, we classified stroke severity according to the NIHSS score as follows: no stroke or minor stroke (NIHSS 0-5), mild to moderate stroke (NIHSS 6-10), and moderate to severe stroke (NIHSS ≥11), as previously classified in the ARIC study.^35,36

In FOS, the NIHSS was adopted for determination of stroke severity in the late 1980s; before then, the FOS assessed stroke severity using a measurement scale reflecting the degree of neurologic deficits causing disability.⁴⁰ We developed a crosswalk between the FOS severity scale and the NIHSS based on a theoretical mapping of the categories (eMethods in Supplement 1). In sensitivity analyses, we leveraged a subset of strokes with both grading approaches to impute NIHSS scores based on stroke severity and on different sets of adjustment variables, as reported in the Results.

Outcome Variables

Cognitive Function

Dementia and global cognition (primary outcomes) and memory and executive function (secondary outcomes) were assessed by each cohort. For the pooled cohort analysis, data on cognitive function were statistically harmonized across cohorts as previously described^41,42 and treated as scaled continuous variables (mean [SD] at baseline, 50 [10]).^33,41,43,44 For the study of risk of dementia, diagnostic classifications were based on cohort-specific criteria. The participating cohorts implemented state-of-the-art testing for cognitive function and diagnosis of dementia; therefore, cognitive tests were not adapted to accommodate sensory, motor, or language impairment from stroke, and inability to complete test components due to speech or other deficits resulted in missing data. However, with respect to participants who did not attend follow-up, the cohorts implemented various procedures to ascertain dementia. In ARIC, cognitive function was assessed through in-person administration of a 3-test battery from ARIC 1990-1992 and the use of an expanded 10-test battery and informant interviews from ARIC 2011-2013. Based on these tests, a computer algorithm produced a preliminary diagnosis that was verified by expert review of all the information collected through cognitive tests, detailed neurologic history, and interviews. When in-person neurocognitive assessment was not available, it was conducted through structured telephone cognitive testing and informant interviews and/or medical record surveillance, including continuous surveillance of hospitalization and death certificate codes.^45,46 According to these procedures, in addition to the expert classification (ARIC Neurocognitive Study level 1 dementia ascertainment), dementia was also classified using level 2 and level 3 dementia ascertainment.³⁶ Details on the ARIC levels of dementia ascertainment are presented in the eMethods in Supplement 1.

In the FOS cohort, dementia surveillance incorporated serial cognitive screening and neuropsychological testing, with case identification supplemented by interim triggers (annual health status updates, family and physician reports), caregiver telephone interviews, and adjudication using outside medical records. For participants who had died, postmortem and nursing records up to death were used to detect cognitive decline since the last examination.^47,48

The REGARDS cognitive outcome ascertainment is telephone based, using a 3-test battery during follow-up and a prespecified algorithm to classify incident cognitive impairment. The REGARDS study does not have physician-adjudicated data on incident dementia; therefore, a Six-Item Screener (SIS) cutoff score of less than 4 (on a scale of 0-6)⁴⁹ in 2 consecutive tests (or 1 SIS score <4 points for 3 participants with only 1 score available), consistent with dementia, was used. The SIS cutoff of less than 4 points used in this study showed 89% sensitivity and 88% specificity for a diagnosis of dementia in a community-based study.⁴⁹ Participants who had insufficient telephone cognitive testing to apply the algorithm could not be classified for incident impairment and were therefore considered to have a missing outcome status.^16,50 In the present study, dementia after stroke was defined if diagnosed at least 1 year after incident stroke to exclude cases of acute temporary changes in cognitive function after stroke.³⁶

Covariates

Baseline age, sex, race and ethnicity, education, cohort, BMI, blood glucose, systolic blood pressure (SBP), and diastolic blood pressure (DBP) were included in all models. Apolipoprotein E (APOE) ε4 status was included as a covariate in sensitivity analyses because genetic data were not available for all participants. Race and ethnicity were self-identified in each cohort (ARIC, non-Hispanic White and other race and ethnicity [African American or Black, other, or unknown]; FOS, non-Hispanic White and African American or Black; REGARDS, non-Hispanic White and African American or Black), and for this analysis, we classified these groups as non-Hispanic White or other race and ethnicity. We adjusted for race and ethnicity because of their known association with both stroke and dementia risk, with implications for prevention strategies.

Statistical Analysis

The data analysis was completed on February 27, 2026. Variables are described using means and SDs for continuous data and counts and percentages for categorical data. Statistical significance was defined as a 2-sided α of .05.

Multivariable linear mixed-effects models were used to test associations of stroke incidence and severity with cognitive decline, adjusting for cohort, age, and age-by-time interaction (model 1) and further adjusting for BMI, SBP, DBP, and blood glucose (model 2). Follow-up time was expressed as years since study baseline. To estimate the difference in rate of change in cognitive function, we used linear mixed-effects models with random effects for participants and time. Indicators for stroke incidence and severity (no stroke [reference group], minor stroke [NIHSS 0-5], mild to moderate stroke [NIHSS 6-10], and moderate to severe stroke [NIHSS ≥11]) were included as variables, and their interactions with time in study were used to estimate the difference in rate of change between each NIHSS category and the no-stroke reference group. Models included fixed effects for age, sex, education, race and ethnicity, and cohort. An equation describing the general model setup is shown in the eMethods in Supplement 1. To test whether levels of SBP, DBP, blood glucose, and BMI altered the association between stroke severity and cognitive decline, in separate models (model 3 sets) we entered 3-way interactions among NIHSS, time, and each vascular risk factor. In exploratory analyses, we further investigated whether the associations between stroke and cognitive changes were moderated by underlying vascular risk factors.

Cox proportional hazards models were used to estimate dementia risk, adjusting for baseline covariates (models 1, 2, and 3). Follow-up began at baseline or age 55 years, whichever occurred later, and ended with dementia diagnosis (or SIS scores consistent with dementia in REGARDS), death, loss to follow-up, or administrative censoring. We chose age 55 years to begin follow-up because only 3 cases of dementia occurred before that age. Sensitivity analyses included imputation of early FOS stroke severity using theoretical and empirical methods; adjustment for APOE ε4 in available participants; and cohort-specific models separately evaluated for ARIC, FOS, and REGARDS. The data analyses were performed using Stata, version 17.0 (StataCorp LLC).

Results

Participant Characteristics

A total of 42 342 participants were included in the pooled analysis (40 837 [96.4%] without stroke and 1505 [3.6%] with a first-ever incident stroke). At baseline, participants’ mean (SD) age was 61.3 (9.8) years; 55.0% were female and 45.0% male; and 52.4% self-identified as non-Hispanic White and 47.5% as another race or ethnicity (Table 1). Participants with missing data on cognitive testing (1944 [4.3%]) tended to be younger (mean [SD], 59 [9.4] years vs 61.3 [9.8] years; P < .001), to be male (1043 [53.7%] vs 19054 [46.3%]; P < .001), and to have less education (some college or more, 601 [34.1%] vs 24 011 [56.7%]; P < .001) than participants with cognitive testing. Across the studies, participants who dropped out with less than 5 years of follow-up (9378 [22.1%]) tended to be older (odds ratio [OR], 1.03; 95% CI, 1.03-1.03), to be male (OR, 0.90; 95% CI, 0.86-0.94), to have lower education (OR, 0.74; 95% CI, 0.72-0.76), and to have lower general cognition at baseline (OR, 0.95; 95% CI, 0.95-0.96). Characteristics by stroke status and cohort are presented in Table 1 and eTable 1 in Supplement 1.

Table 1. Descriptive Characteristics of the Pooled Sample.

Characteristic	Participants, No. (%)
Characteristic	Total (N = 42 342)	No stroke (n = 40 837)	Stroke (n = 1505)
Age at first cognitive visit, mean (SD), y	61.3 (9.8)	61.2 (9.8)	64.3 (9.8)
Sex
Female	23 281 (55.0)	22 534 (55.2)	747 (49.6)
Male	19 061 (45.0)	18 303 (44.8)	758 (50.4)
Race and ethnicity
Non-Hispanic White	22 184 (52.4)	21 403 (52.4)	781 (51.9)
Other^a	20 158 (47.6)	19 434 (47.6)	724 (48.1)
Education
Less than high school	6334 (15.0)	6097 (15.0)	237 (15.8)
High school	11 921 (28.2)	11 453 (28.1)	468 (31.2)
Some college	9065 (21.5)	8725 (21.4)	340 (22.7)
Completed college or more	14 946 (35.4)	14 492 (35.6)	454 (30.3)
Participating cohort
ARIC	12 376 (29.2)	12 092 (29.6)	284 (18.9)
FOS	801 (1.9)	731 (1.8)	70 (4.7)
REGARDS	29 165 (68.9)	28 014 (68.6)	1151 (76.5)
No. of visits, mean (SD)	9.1 (6.9)	9.0 (6.9)	10.9 (6.5)
Follow-up, median (IQR), y	11.1 (5.5-14.1)	10.3 (5.1-14.1)	12.1 (7.1-14.1)
Stroke severity category
No stroke	40 837 (97.5)	40 837 (100)	0
NIHSS 0-5 (minor)	879 (2.1)	0	879 (83.3)^b
NIHSS 6-10 (mild to moderate)	125 (0.3)	0	125 (11.9)^b
NIHSS ≥11 (moderate to severe)	51 (0.1)	0	51 (4.8)^b
Missing (among strokes), No.	450	NA	450
Incident dementia^c	1870 (4.4)	1739 (4.3)	131 (8.7)
Baseline cardiovascular variables
BMI, mean (SD)	28.8 (6.0)	28.8 (6.0)	29.0 (5.7)
Hypertension	25 742 (61.5)	24 586 (61.0)	1156 (77.2)
SBP, mean (SD), mm Hg	131.4 (17.8)	131.3 (17.8)	135.0 (18.5)
DBP, mean (SD), mm Hg	81.6 (12.6)	81.6 (12.6)	81.5 (12.7)
Diabetes history	7686 (18.9)	7266 (18.5)	420 (29.0)
Blood glucose, mean (SD), mg/dL^d	105.5 (37.2)	105.2 (36.5)	113.2 (51.4)
APOE ε4
Any ε4 allele	6817 (33.0)	6515 (33.0)	302 (31.5)
No ε4 allele	13 874 (67.1)	13 218 (67.0)	656 (68.5)
Missing, No.	21 651	21 104	547
Baseline cognitive scores, mean (SD)
Global cognition	54.5 (6.5)	54.5 (6.6)	54.3 (6.3)
Executive function	51.4 (10.0)	51.5 (10.0)	50.0 (9.9)
Memory	54.2 (4.9)	54.2 (4.9)	53.9 (5.0)

Open in a new tab

Abbreviations: APOE, apolipoprotein E; ARIC, Atherosclerosis Risk in Communities; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); DBP, diastolic blood pressure; FOS, Framingham Offspring Study; NIHSS, National Institutes of Health Stroke Scale; REGARDS, Reasons for Geographic and Racial Differences in Stroke; SBP, systolic blood pressure.

^{^a}

In ARIC, the other race and ethnicity group mostly included African American or Black individuals (n = 2901 [23%]), as well as other or unknown race (n = 26 [0.2%]). In FOS, 798 participants were non-Hispanic White, and 3 participants (0.37%) were African American or Black. In REGARDS, the other race and ethnicity group included African American or Black individuals (n = 17 228 [59%]).

^{^b}

Percentages are based on stroke events with nonmissing NIHSS scores.

^{^c}

For REGARDS, Six-Item Screener scores were consistent with dementia.

^{^d}

To convert to millimoles per liter, multiply by 0.0555.

Overall, 1505 first-ever definite ischemic strokes were identified in the harmonized pooled data, among which data on severity was determined for 1055 (70.1%) events. Participants missing NIHSS scores (74 from ARIC, 1 from FOS, and 375 from REGARDS) were younger (OR, 0.97 per year; 95% CI, 0.96-0.99 per year; P = .003) and had more study visits (OR, 1.02 per visit; 95% CI, 1.00-1.05 per visit; P = .03) but were similar in sex, race and ethnicity, and education. Overall, participants’ characteristics were similar in the different cohorts. With regard to stroke severity data, 879 strokes (83.3%) were categorized as minor (NIHSS 0-5), 125 (11.9%) as mild to moderate (NIHSS 6-10), and 51 (4.8%) as moderate to severe (NIHSS ≥11). Follow-up with cognitive testing across the cohorts lasted a median of 11.1 years (range, 0-29.7 years; 397 344 person-years of observation for dementia incidence), with a median (IQR) number of visits with cognitive testing of 11 (5-17) among stroke survivors. The median (IQR) visits with cognitive testing were 4 (2-9) before stroke and 4 (2-8) after stroke. In contrast with prior reports,¹² baseline global cognition scores did not differ by future stroke status (B, −0.20; 95% CI, −0.54 to 0.14), a finding that may reflect inclusion of a relatively healthy, cognitively intact cohort with limited variability in baseline cognition. Follow-up for dementia lasted a median of 9.4 years (range, 0.1-29.6 years) overall, given surveillance methods of the cohorts that did not require in-person visits. Overall, 1870 cases of incident dementia (for REGARDS, SIS scores consistent with dementia) were reported (Table 1).

Stroke Severity and Global Cognition

The rate of change in global cognition among participants without stroke was −0.18 points per year, which given the scaling of the cognitive score (SD, 10), corresponds to 0.18 SDs per decade (Table 2). Higher stroke severity was associated with a progressively steeper annual rate of decline in global cognition during follow-up, ranging from a mean −0.18 points per year (95% CI, −0.19 to −0.18 points per year) among participants with no stroke to −0.36 points per year (95% CI, −0.39 to −0.32 points per year) among participants with minor stroke, −0.47 points per year (95% CI, −0.58 to −0.36 points per year) among participants with mild to moderate stroke, and −0.58 points per year (95% CI, −0.73 to −0.42 points per year) among participants with moderate to severe stroke. Interactions among DBP, time, and stroke severity for minor (P = .048) and moderate to severe (P = .04) stroke suggested attenuation of cognitive decline at higher DBP (eTable 2 in Supplement 1).

Table 2. Association of Stroke Severity With General Cognitive Change in the Pooled Sample (N = 42 342).

Coefficient	Model 1^a			Model 2^b
Coefficient	β (SE)	P value	Age-standardized β^c	β (SE)	P value	Age-standardized β^c
Main effect sizes
Stroke severity category (reference, no stroke)
NIHSS 0-5 (minor)	−0.24 (0.16)	.13	1.28	−0.13 (0.16)	.42	0.69
NIHSS 6-10 (mild to moderate)	−0.66 (0.45)	.14	3.50	−0.40 (0.45)	.38	2.10
NIHSS ≥11 (moderate to severe)	0.05 (0.73)	.94	−0.28	0.21 (0.72)	.77	−1.09
Years since baseline	−0.18 (0)	<.001	0.96	−0.18 (0)	<.001	0.96
Interactions between time and stroke severity (reference, no stroke)
Time × NIHSS 0-5	−0.18 (0.02)	<.001	0.93	−0.17 (0.02)	<.001	0.91
Time × NIHSS 6-10	−0.27 (0.06)	<.001	1.42	−0.28 (0.06)	<.001	1.50
Time × NIHSS ≥11	−0.39 (0.08)	<.001	2.07	−0.39 (0.08)	<.001	2.07

Open in a new tab

Abbreviation: NIHSS, National Institutes of Health Stroke Scale.

^{^a}

Adjusted for cohort membership, baseline age, sex, race and ethnicity, education, and age-by-time interaction.

^{^b}

Model 1 additionally adjusted for baseline BMI, systolic blood pressure, diastolic blood pressure, and blood glucose.

^{^c}

Age-standardized β refers to the β coefficient divided by the corresponding coefficient for age in the model.

Stroke Severity and Memory

Memory decline over follow-up that included prestroke trajectories was progressively steeper with greater stroke severity. Higher stroke severity was associated with a progressively steeper annual rate of decline in memory during follow-up, ranging from a mean of −0.15 points per year (95% CI, −0.16 to −0.14 points per year) among participants with no stroke to −0.20 points per year (95% CI, −0.24 to −0.17 points per year) among participants with minor stroke, −0.31 points per year (95% CI, −0.42 to −0.20 points per year) among participants with mild to moderate stroke, and −0.36 points per year (95% CI, −0.51 to −0.21 points per year) among participants with moderate to severe stroke (Table 3). No statistically significant interactions were observed among time; stroke severity; and BMI, SBP, DBP, or blood glucose (eTable 3 in Supplement 1).

Table 3. Association of Stroke Severity With Memory Change in the Pooled Sample (n = 41 703).

Coefficient	Model 1^a			Model 2^b
Coefficient	β (SE)	P value	Age-standardized β^c	β (SE)	P value	Age-standardized β^c
Main effect sizes
Stroke severity (reference, no stroke)
NIHSS 0-5 (minor)	−0.07 (0.12)	.54	0.68	−0.02 (0.12)	.87	0.19
NIHSS 6-10 (mild to moderate)	−0.52 (0.31)	.10	5.07	−0.29 (0.32)	.37	2.78
NIHSS ≥11 (moderate to severe)	−0.20 (0.50)	.68	1.98	−0.14 (0.50)	.77	1.40
Years since baseline	−0.15 (0)	<.001	1.44	−0.15 (0)	<.001	1.46
Interactions between time and stroke severity (reference, no stroke)
Time × NIHSS 0-5	−0.05 (0.02)	.01	0.52	−0.05 (0.02)	.01	0.52
Time × NIHSS 6-10	−0.14 (0.06)	.01	1.35	−0.16 (0.06)	.01	1.51
Time × NIHSS ≥11	−0.21 (0.08)	.01	2.06	−0.21 (0.08)	.01	2.06

Open in a new tab

Abbreviation: NIHSS, National Institutes of Health Stroke Scale.

^{^a}

Adjusted for cohort membership, baseline age, sex, race and ethnicity, education, and age-by-time interaction.

^{^b}

Model 1 additionally adjusted for baseline BMI, systolic blood pressure, diastolic blood pressure, and blood glucose.

^{^c}

Age-standardized β refers to the β coefficient divided by the corresponding coefficient for age in the model.

Stroke Severity and Executive Function

Executive function, over follow-up that included prestroke trajectories, declined faster than global cognition or memory (B = −0.33 points per year) (Table 4). Higher stroke severity was associated with a progressively steeper annual rate of decline in executive functioning during follow-up, ranging from a mean −0.33 points per year (95% CI, −0.34 to −0.32 points per year) among participants with no stroke to −0.49 points per year (95% CI, −0.54 to −0.45 points per year) among participants with minor stroke, −0.52 points per year (95% CI, −0.64 to −0.41 points per year) among participants with mild to moderate stroke, and −0.52 points per year (95% CI, −0.66 to −0.39 points per year) among participants with moderate to severe stroke.

Table 4. Association of Stroke Severity With Executive Functioning Change in the Pooled Sample (n = 36 839).

Coefficient	Model 1^a			Model 2^b
Coefficient	β (SE)	P value	Age-standardized β^c	β (SE)	P value	Age-standardized β^c
Main effect sizes
Stroke severity (reference, no stroke)
NIHSS 0-5 (minor)	−1.34 (0.32)	<.001	3.82	−1.02 (0.32)	<.001	2.90
NIHSS 6-10 (mild to moderate)	−2.89 (0.84)	<.001	8.21	−2.41 (0.84)	<.001	6.82
NIHSS ≥11 (moderate to severe)	−1.31 (1.23)	.29	3.72	−0.95 (1.21)	.43	2.70
Years since baseline	0.33 (0)	<.001	0.95	−0.33 (0)	<.001	0.94
Interactions between time and stroke severity (reference, no stroke)
Time × NIHSS 0-5	−0.16 (0.02)	<.001	0.46	−0.16 (0.02)	<.001	0.45
Time × NIHSS 6-10	−0.18 (0.06)	<.001	0.52	−0.19 (0.06)	<.001	0.54
Time × NIHSS ≥11	−0.19 (0.07)	.01	0.54	−0.19 (0.07)	.01	0.54

Open in a new tab

Abbreviation: NIHSS, National Institutes of Health Stroke Scale.

^{^a}

Adjusted for cohort membership, baseline age, sex, race and ethnicity, education, and age-by-time interaction.

^{^b}

Model 1 additionally adjusted for baseline BMI, systolic blood pressure, diastolic blood pressure, and blood glucose.

^{^c}

Age-standardized β refers to the β coefficient divided by the corresponding coefficient for age in the model.

Statistically significant interactions were observed among time, SBP, and stroke severity for minor stroke (P = .003), suggesting that the association of minor stroke severity with executive function decline was attenuated with higher SBP (eTable 4 in Supplement 1). There also were 3-way interactions among time, DBP, and stroke severity for minor (P = .003) and mild to moderate stroke (P = .005), suggesting that the associations of minor and mild to moderate severity with executive function decline were attenuated with higher DBP (eTable 4 in Supplement 1).

Stroke Severity and Dementia Risk

Stroke severity was associated with dementia risk. Compared with participants without stroke, adjusted hazards ratios (HRs) for minor stroke, mild to moderate stroke, and moderate to severe stroke were 1.93 (95% CI, 1.52-2.45), 3.26 (95% CI, 1.93-5.53), and 5.06 (95% CI, 2.71-9.45), respectively. Details on incident dementia by NIHSS scores and HRs for associations of stroke severity levels with incident dementia are presented in Table 5. Because REGARDS used an algorithm for dementia instead of clinical adjudication, we reestimated models among 13 177 participants not from REGARDS, and results were comparable. Compared with the group without stroke, minor stroke (HR, 2.11; 95% CI, 1.46-3.06), mild to moderate stroke (HR, 4.77; 95% CI, 2.46-9.23), and moderate to severe stroke (HR, 7.55; 95% CI, 3.58-15.94) were associated with an elevated hazard of dementia. In an exploratory analysis of the dementia outcome, we observed no interactions among time; stroke severity level; and each of BMI, SBP, DBP, and glucose level.

Table 5. Association of NIHSS With Hazard of Dementia in the Pooled Sample (n = 43 008)^a.

Coefficient	No. (%)		Follow-up, person-years	HR (95% CI)
Coefficient	Incident dementia	No incident dementia	Follow-up, person-years	Model 1^b	Model 2^c
Stroke severity^d
NIHSS 0-5 (minor)	73 (0.08)	806 (0.92)	4913.77	1.97 (1.55-2.49)	1.93 (1.52-2.45)
NIHSS 6-10 (mild to moderate)	14 (0.11)	111 (0.89)	642.39	3.37 (1.99-5.71)	3.26 (1.93-5.53)
NIHSS ≥11 (moderate to severe)	10 (0.20)	41 (0.80)	257.61	5.40 (2.89-10.06)	5.06 (2.71-9.45)

Open in a new tab

Abbreviations: HR, hazard ratio; NIHSS, National Institutes of Health Stroke Scale; REGARDS, Reasons for Geographic and Racial Differences in Stroke.

^{^a}

For REGARDS, Six-Item Screener cutoffs consistent with dementia were used.

^{^b}

Adjusted for cohort membership, baseline age, sex, race and ethnicity, education, and age-by-time interaction.

^{^c}

Model 1 additionally adjusted for baseline BMI, systolic blood pressure, diastolic blood pressure, and blood glucose.

^{^d}

No stroke was the reference group. Among participants without stroke, during 390 193 person-years of follow-up, there were 1739 participants (<0.1%) with incident dementia and 40 564 participants (1.0%) without dementia.

Sensitivity Analyses

Sensitivity analyses were conducted using alternative mappings for FOS stroke severity based on multivariate imputation by chained equations, first based on the NIHSS only (eTables 5-8 in Supplement 1) and then adjusting additionally for age and sex (eTables 9-12 in Supplement 1). In addition, we assessed the influence of APOE ε4 on the studied associations (eTables 13-16 in Supplement 1) and produced cohort-specific models (eTables 17-28 in Supplement 1). Findings were consistent with those in the primary analyses.

Discussion

This pooled cohort study analyzed individual data from 42 342 participants across 3 large, prospective cohorts in the US with up to 30 years of follow-up. The study found that higher stroke severity is associated with progressively steeper cognitive decline across global cognition, memory, and executive function domains along with significantly higher risks of dementia.

The finding showing probable steeper cognitive decline after more severe stroke compared with less severe stroke is biologically plausible. Major strokes cause greater initial structural and network damage, thereby reducing cognitive reserve and leaving the brain less able to compensate for subsequent pathology.^51,52 More severe stroke is also associated with a higher risk of recurrent vascular events and progression of small vessel disease, which can further accelerate cognitive deterioration.^2,21,24 In addition, individuals with major stroke often experience more pronounced physical disability and reduced social and cognitive engagement, both of which have been associated with faster cognitive decline and dementia.^26,53 Different relative rates of decline across domains may reflect that each domain depends on partly distinct neural systems with different vulnerability to vascular and neurodegenerative injury. Processing speed and executive function, which rely on fronto-subcortical circuits and white matter, are particularly sensitive to small vessel disease and diffuse ischemia and, therefore, often show earlier and steeper declines than episodic memory or language.^54,55 In contrast, memory functions supported by medial temporal lobe structures may be relatively preserved until more advanced or mixed pathology, and differences in cognitive reserve can further modulate the apparent rate of change across domains.^26,52

The rate of decline of 0.18 SDs per decade observed in participants without stroke is consistent with declines of 0.2 to 0.5 SDs observed in most other cohorts of community-living older adults aged 65 years or older,^{56,57,58,59,60} a finding that supports the generalizability of our population to others. Compared with individuals without stroke, those with mild to moderate stroke experienced cognitive decline equivalent to as if they were 1.8 years older at baseline, while those with moderate to severe stroke declined as if they were 2.6 years older at baseline. Furthermore, mild to moderate and moderate to severe stroke severity correlated with 3-fold and 5-fold increased risks of dementia, respectively, compared with no stroke. Our findings on cognitive aging and incidence of dementia associated with stroke are lower than previously reported.^10,36 Such differences may be associated with differences in age and health characteristics of the study population, survival bias, differences in rates of minor vs severe stroke, differences in tests used for the evaluation of cognitive function, the timing of the evaluation, and attrition rates.

Stroke severity plays a pivotal role in poststroke cognitive outcomes. Severe strokes have been linked to higher dementia risk and persistent cognitive impairment. The Oxford Vascular Study found an approximately 50-fold increase in dementia incidence reported within 1 year of a major stroke compared with the general population.¹⁰ Severity-related cognitive decline affects multiple domains due to the multifocal nature of stroke-related injuries.⁶¹ However, cognitive impairment is also observed in patients with low NIHSS scores and even transient ischemic attack.^{10,11,62,63,64} Silent strokes may also lead to future cognitive dysfunction.⁵⁴ Of note, 1 study involving patients with stroke with a mean admission NIHSS score of 2.7 indicated that NIHSS scores remained associated with cognitive function, even though some degree of improvements across domains were evident 6 months after stroke.⁶⁵

Mechanisms underlying poststroke cognitive changes and dementia are multifactorial. They involve vascular damage, neurodegenerative processes, and chronic inflammation.⁶⁶ Risk factors contributing to cognitive impairment after stroke include age; stroke severity; recurrence; location; and comorbidities, such as diabetes and poststroke seizures.^10,11 Chronic brain inflammation, potentially exacerbated by prolonged immune responses and inefficient myelin debris clearance, might also play a role in poststroke dementia.⁶⁷ Cognitive outcomes can vary among patients with similar stroke severity due to interactions among prestroke cognitive status, vascular risk factors, white matter hyperintensities, systemic immune responses, stroke management, and presence of subclinical infarcts.^24,54,67

Prospective cohort studies provide valuable data for analyzing the association between stroke and cognitive outcomes, as they account for heterogeneous conditions and long-term effects. In 2 previous articles using data from 2 cohorts participating in the Stroke-Cog pooled cohort, links between stroke and cognitive outcomes have been reported. For example, the ARIC study found stroke to be an independent risk factor for dementia, with risks elevated further by stroke severity and recurrence.³⁶ In the REGARDS study, accelerated cognitive decline in global cognition and executive function after stroke highlighted domain-specific differences in the patterns of cognitive decline.¹² We have now assessed these associations in a larger cohort including individuals with and without stroke and with the use of the most recent data available.

Previous research examining blood pressure and glucose effects on poststroke cognition has revealed complex associations,³³ which may explain the different influences of risk factors on the associations of stroke severity and cognitive function observed in our exploratory analyses. Our exploratory findings suggest different influences of SBP and DBP on the associations of cognitive function with different stroke severities. These differences may be associated with different vascular processes reflected by SBP and DBP. While elevated SBP has been shown to be associated with large artery stiffness, higher pulse pressure, and damage of fragile cerebral microvessels, and therefore the promotion of white matter damage,^26,54 DBP has been more closely associated with the perfusion pressure needed to maintain adequate diastolic cerebral blood flow. Low DBP may compromise perfusion, whereas very high DBP may exacerbate small vessel stress.⁶⁸ These, and potentially other, possible mechanisms that explain the observed complex associations warrant future research. In addition, based on our main analysis findings, future research studying associations in subgroups by age, race and ethnicity, and specific risk factors (eg, blood pressure levels) is recommended.

Strengths and Limitations

This study had several strengths. First, we analyzed individual participant data pooled from 3 prospective cohort studies with decades-long follow-up and large numbers of adjudicated incident strokes treated as a time-varying exposure. The availability of cognitive assessments before and after stroke enabled evaluation of trajectories over time and identification of delayed cognitive changes distinct from transient deficits at the time of stroke. Change-point and segmented trajectory approaches in cohort data have further shown the utility of isolating acute vs chronic components.¹² Analyses of large and diverse populations, both geographically and clinically, enhanced the generalizability of findings. Second, detailed demographic and clinical data allowed comprehensive adjustment for risk factors influencing stroke, dementia, and their interactions with time and stroke severity. Third, harmonization of longitudinal cognitive assessments facilitated examination of cognitive trajectories across domains, mitigating variability in methodology among cohorts. Exploratory analyses also suggested the potential influence of vascular risk factors, such as blood pressure and glucose, on the associations between stroke severity and cognitive function.

Despite the study’s strengths, some limitations must be noted. First, the findings are generalizable only to community-dwelling stroke survivors without dementia prior to the incident ischemic stroke. Second, brain imaging data, neuroinflammatory or neurodegenerative biomarkers, and information on stroke location and recurrent strokes were unavailable, limiting detailed mechanistic analyses. In addition, data on peak adult cognitive ability, known as an important factor associated with both stroke risk and poststroke cognitive impairment, were not available. However, all the statistical models included educational attainment, which may correlate with both early-life cognitive ability and later-life cognitive trajectories. Third, we excluded early diagnoses of dementia in the first year after stroke, which may represent cognitive changes that reverse. By excluding early dementia, we were not including these early, common, poststroke changes. Selection bias, in a conservative direction, may arise from the underrepresentation of patients with severe stroke, as these individuals may have fewer poststroke follow-up visits. Missing data and attrition during longitudinal follow-up may have led to conservative biases. Another sample size–related limitation is that despite the pooled cohorts we were able to leverage, more than 80% events were categorized as minor stroke, while more severe stroke categories had sparse numbers; despite this, we observed clear dose-response associations in cognitive outcomes. Finally, while comprehensive neuropsychological test data were included, physician-adjudicated dementia diagnoses were unavailable in the REGARDS cohort, and some study visits relied on brief cognitive assessments, potentially reducing sensitivity for detecting mild cognitive changes.^12,43

Conclusions

This large cohort study of participants in 3 prospective cohorts found robust evidence of the graded association of stroke severity with cognitive decline and dementia risk. The findings highlight the importance of prevention strategies, vigilance in cognitive monitoring, and more aggressive treatment of risk factors aimed at mitigating dementia development across severity levels.

Supplement 1.

eFigure. Consolidated Standards of Reporting Trials Flow Diagram

eMethods.

eTable 1. Descriptive Characteristics of the Sample, by Cohort

eTable 2. Associations of NIH Stroke Severity With General Cognitive Change, Model 3

eTable 3. Associations of NIH Stroke Severity With Memory Change, Model 3

eTable 4. Associations of NIH Stroke Severity With Executive Functioning Change, Model 3

eTable 5. Associations of NIH stroke Severity With General Cognitive Change Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, With No Adjustment

eTable 6. Associations of NIH Stroke Severity With Memory Change Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, With No Adjustment

eTable 7. Associations of NIH Stroke Severity With Executive Functioning Change Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, With No Adjustment

eTable 8. Associations of NIH Stroke Severity, Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, With No Adjustment, With Hazard of Dementia

eTable 9. Associations of NIH Stroke Severity With General Cognitive Change Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, Adjusted for Age and Sex

eTable 10. Associations of NIH Stroke Severity With Memory Change Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, Adjusted for Age and Sex

eTable 11. Associations of NIH Stroke Severity With Executive Functioning Change Using Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, Adjusted for Age and Sex

eTable 12. Associations of NIH Stroke Severity, Based on Alternative Mapping of FOS Stroke Severity Based on MICE Imputation, Adjusted for Age and Sex, With Hazard of Dementia

eTable 13. Associations of NIH Stroke Severity With General Cognitive Change, Adjusted for APOE ε4

eTable 14. Associations of NIH Stroke Severity With Memory Change, Adjusted for APOE ε4

eTable 15. Associations of NIH Stroke Severity With Executive Functioning Change, Adjusted for APOE ε4

eTable 16. Associations of NIH Stroke Severity With Hazard of Dementia, Adjusted for APOE ε4

eTable 17. Associations of NIH Stroke Severity With General Cognitive Change: Results From ARIC

eTable 18. Associations of NIH Stroke Severity With Memory Change: Results From ARIC

eTable 19. Associations of NIH Stroke Severity With Executive Functioning Change: Results From ARIC

eTable 20. Associations of NIH Stroke Severity With Hazard of Dementia: Results From ARIC

eTable 21. Associations of NIH Stroke Severity With General Cognitive Change: Results From FOS

eTable 22. Associations of NIH Stroke Severity With Memory Change: Results From FOS

eTable 23. Associations of NIH stroke Severity With Executive Functioning Change: Results From FOS

eTable 24. Associations of NIH Stroke Severity With Hazard of Dementia: Results From FOS

eTable 25. Associations of NIH Stroke Severity With General Cognitive Change: Results From REGARDS

eTable 26. Associations of NIH Stroke Severity With Memory Change: Results From REGARDS

eTable 27. Associations of NIH Stroke Severity With Executive Functioning Change: Results From REGARDS

eTable 28. Associations of NIH Stroke Severity With Hazard of Dementia: Results From REGARDS

jamanetwopen-e268900-s001.pdf^{(2MB, pdf)}

Supplement 2.

Data Sharing Statement

jamanetwopen-e268900-s002.pdf^{(12.6KB, pdf)}

References

1.Rost NS, Bottle A, Lee JM, et al. ; Global Comparators Stroke GOAL Collaborators . Stroke severity is a crucial predictor of outcome: an international prospective validation study. J Am Heart Assoc. 2016;5(1):e002433. doi: 10.1161/JAHA.115.002433 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Pendlebury ST, Rothwell PM. Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: a systematic review and meta-analysis. Lancet Neurol. 2009;8(11):1006-1018. doi: 10.1016/S1474-4422(09)70236-4 [DOI] [PubMed] [Google Scholar]
3.Douiri A, Rudd AG, Wolfe CD. Prevalence of poststroke cognitive impairment: South London Stroke Register 1995-2010. Stroke. 2013;44(1):138-145. doi: 10.1161/STROKEAHA.112.670844 [DOI] [PubMed] [Google Scholar]
4.Ihle-Hansen H, Thommessen B, Wyller TB, et al. Incidence and subtypes of MCI and dementia 1 year after first-ever stroke in patients without pre-existing cognitive impairment. Dement Geriatr Cogn Disord. 2011;32(6):401-407. doi: 10.1159/000335361 [DOI] [PubMed] [Google Scholar]
5.Ivan CS, Seshadri S, Beiser A, et al. Dementia after stroke: the Framingham study. Stroke. 2004;35(6):1264-1268. doi: 10.1161/01.STR.0000127810.92616.78 [DOI] [PubMed] [Google Scholar]
6.Jacquin A, Binquet C, Rouaud O, et al. Post-stroke cognitive impairment: high prevalence and determining factors in a cohort of mild stroke. J Alzheimers Dis. 2014;40(4):1029-1038. doi: 10.3233/JAD-131580 [DOI] [PubMed] [Google Scholar]
7.Mahon S, Parmar P, Barker-Collo S, et al. Determinants, prevalence, and trajectory of long-term post-stroke cognitive impairment: results from a 4-year follow-up of the ARCOS-IV study. Neuroepidemiology. 2017;49(3-4):129-134. doi: 10.1159/000484606 [DOI] [PubMed] [Google Scholar]
8.Rist PM, Chalmers J, Arima H, et al. Baseline cognitive function, recurrent stroke, and risk of dementia in patients with stroke. Stroke. 2013;44(7):1790-1795. doi: 10.1161/STROKEAHA.111.680728 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Sachdev PS, Brodaty H, Valenzuela MJ, et al. Clinical determinants of dementia and mild cognitive impairment following ischaemic stroke: the Sydney Stroke study. Dement Geriatr Cogn Disord. 2006;21(5-6):275-283. doi: 10.1159/000091434 [DOI] [PubMed] [Google Scholar]
10.Pendlebury ST, Rothwell PM; Oxford Vascular Study . Incidence and prevalence of dementia associated with transient ischaemic attack and stroke: analysis of the population-based Oxford Vascular study. Lancet Neurol. 2019;18(3):248-258. doi: 10.1016/S1474-4422(18)30442-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Gottesman RF, Hillis AE. Predictors and assessment of cognitive dysfunction resulting from ischaemic stroke. Lancet Neurol. 2010;9(9):895-905. doi: 10.1016/S1474-4422(10)70164-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Levine DA, Galecki AT, Langa KM, et al. Trajectory of cognitive decline after incident stroke. JAMA. 2015;314(1):41-51. doi: 10.1001/jama.2015.6968 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Zheng F, Yan L, Zhong B, Yang Z, Xie W. Progression of cognitive decline before and after incident stroke. Neurology. 2019;93(1):e20-e28. doi: 10.1212/WNL.0000000000007716 [DOI] [PubMed] [Google Scholar]
14.Wright JD, Folsom AR, Coresh J, et al. The ARIC (Atherosclerosis Risk in Communities) study: JACC focus seminar 3/8. J Am Coll Cardiol. 2021;77(23):2939-2959. doi: 10.1016/j.jacc.2021.04.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Feinleib M, Kannel WB, Garrison RJ, McNamara PM, Castelli WP. The Framingham Offspring Study. design and preliminary data. Prev Med. 1975;4(4):518-525. doi: 10.1016/0091-7435(75)90037-7 [DOI] [PubMed] [Google Scholar]
16.Howard VJ, Cushman M, Pulley L, et al. The Reasons for Geographic and Racial Differences in Stroke study: objectives and design. Neuroepidemiology. 2005;25(3):135-143. doi: 10.1159/000086678 [DOI] [PubMed] [Google Scholar]
17.Rost NS, Meschia JF, Gottesman R, Wruck L, Helmer K, Greenberg SM; DISCOVERY Investigators . Cognitive impairment and dementia after stroke: design and rationale for the DISCOVERY study. Stroke. 2021;52(8):e499-e516. doi: 10.1161/STROKEAHA.120.031611 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Viticchi G, Falsetti L, Plutino A, Bartolini M, Buratti L, Silvestrini M. Sex influence in ischemic stroke severity and outcome among metabolically unhealthy overweight patients. J Neurol Sci. 2020;416:116955. doi: 10.1016/j.jns.2020.116955 [DOI] [PubMed] [Google Scholar]
19.Appelros P, Nydevik I, Seiger A, Terént A. Predictors of severe stroke: influence of preexisting dementia and cardiac disorders. Stroke. 2002;33(10):2357-2362. doi: 10.1161/01.STR.0000030318.99727.FA [DOI] [PubMed] [Google Scholar]
20.Xie W, Zheng F, Yan L, Zhong B. Cognitive decline before and after incident coronary events. J Am Coll Cardiol. 2019;73(24):3041-3050. doi: 10.1016/j.jacc.2019.04.019 [DOI] [PubMed] [Google Scholar]
21.Portegies ML, Wolters FJ, Hofman A, Ikram MK, Koudstaal PJ, Ikram MA. Prestroke vascular pathology and the risk of recurrent stroke and poststroke dementia. Stroke. 2016;47(8):2119-2122. doi: 10.1161/STROKEAHA.116.014094 [DOI] [PubMed] [Google Scholar]
22.Johansen MC, Chen J, Schneider ALC, et al. Association between ischemic stroke subtype and stroke severity: the Atherosclerosis Risk in Communities study. Neurology. 2023;101(9):e913-e921. doi: 10.1212/WNL.0000000000207535 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Cox SR, Lyall DM, Ritchie SJ, et al. Associations between vascular risk factors and brain MRI indices in UK Biobank. Eur Heart J. 2019;40(28):2290-2300. doi: 10.1093/eurheartj/ehz100 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Mijajlović MD, Pavlović A, Brainin M, et al. Post-stroke dementia - a comprehensive review. BMC Med. 2017;15(1):11. doi: 10.1186/s12916-017-0779-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Biessels GJ, Despa F. Cognitive decline and dementia in diabetes mellitus: mechanisms and clinical implications. Nat Rev Endocrinol. 2018;14(10):591-604. doi: 10.1038/s41574-018-0048-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Gorelick PB, Scuteri A, Black SE, et al. ; American Heart Association Stroke Council, Council on Epidemiology and Prevention, Council on Cardiovascular Nursing, Council on Cardiovascular Radiology and Intervention, and Council on Cardiovascular Surgery and Anesthesia . Vascular contributions to cognitive impairment and dementia: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2011;42(9):2672-2713. doi: 10.1161/STR.0b013e3182299496 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Launer LJ, Miller ME, Williamson JD, et al. ; ACCORD MIND Investigators . Effects of intensive glucose lowering on brain structure and function in people with type 2 diabetes (ACCORD MIND): a randomised open-label substudy. Lancet Neurol. 2011;10(11):969-977. doi: 10.1016/S1474-4422(11)70188-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Lüscher TF. The heart and the brain: cardiovascular risk factors, atrial fibrillation, and dementia. Eur Heart J. 2019;40(28):2271-2275. doi: 10.1093/eurheartj/ehz508 [DOI] [PubMed] [Google Scholar]
29.Hua J, Zhou Y, Chen L, Tang X, Diao S, Fang Q. How do cardiovascular risk factors correlate with post-stroke cognitive function: directly or indirectly through stroke severity? Front Neurol. 2022;13:917295. doi: 10.3389/fneur.2022.917295 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Rosamond WD, Folsom AR, Chambless LE, et al. Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke. 1999;30(4):736-743. doi: 10.1161/01.STR.30.4.736 [DOI] [PubMed] [Google Scholar]
31.Seshadri S, Beiser A, Kelly-Hayes M, et al. The lifetime risk of stroke: estimates from the Framingham study. Stroke. 2006;37(2):345-350. doi: 10.1161/01.STR.0000199613.38911.b2 [DOI] [PubMed] [Google Scholar]
32.Wolf PA, D’Agostino RB, Belanger AJ, Kannel WB. Probability of stroke: a risk profile from the Framingham study. Stroke. 1991;22(3):312-318. doi: 10.1161/01.STR.22.3.312 [DOI] [PubMed] [Google Scholar]
33.Levine DA, Chen B, Galecki AT, et al. Associations between vascular risk factor levels and cognitive decline among stroke survivors. JAMA Netw Open. 2023;6(5):e2313879. doi: 10.1001/jamanetworkopen.2023.13879 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Lyden P. Using the National Institutes of Health Stroke Scale: a cautionary tale. Stroke. 2017;48(2):513-519. doi: 10.1161/STROKEAHA.116.015434 [DOI] [PubMed] [Google Scholar]
35.Koton S, Patole S, Carlson JM, et al. Methods for stroke severity assessment by chart review in the Atherosclerosis Risk in Communities study. Sci Rep. 2022;12(1):12338. doi: 10.1038/s41598-022-16522-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Koton S, Pike JR, Johansen M, et al. Association of ischemic stroke incidence, severity, and recurrence with dementia in the Atherosclerosis Risk in Communities cohort study. JAMA Neurol. 2022;79(3):271-280. doi: 10.1001/jamaneurol.2021.5080 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Adams HP Jr, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Neurology. 1999;53(1):126-131. doi: 10.1212/WNL.53.1.126 [DOI] [PubMed] [Google Scholar]
38.Lindsell CJ, Alwell K, Moomaw CJ, et al. Validity of a retrospective National Institutes of Health Stroke Scale scoring methodology in patients with severe stroke. J Stroke Cerebrovasc Dis. 2005;14(6):281-283. doi: 10.1016/j.jstrokecerebrovasdis.2005.08.004 [DOI] [PubMed] [Google Scholar]
39.Williams LS, Yilmaz EY, Lopez-Yunez AM. Retrospective assessment of initial stroke severity with the NIH Stroke Scale. Stroke. 2000;31(4):858-862. doi: 10.1161/01.STR.31.4.858 [DOI] [PubMed] [Google Scholar]
40.Lin HJ, Wolf PA, Kelly-Hayes M, et al. Stroke severity in atrial fibrillation. the Framingham study. Stroke. 1996;27(10):1760-1764. doi: 10.1161/01.STR.27.10.1760 [DOI] [PubMed] [Google Scholar]
41.Briceño EM, Gross AL, Giordani BJ, et al. Pre-statistical considerations for harmonization of cognitive instruments: harmonization of ARIC, CARDIA, CHS, FHS, MESA, and NOMAS. J Alzheimers Dis. 2021;83(4):1803-1813. doi: 10.3233/JAD-210459 [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Gross AL, Li C, Briceño EM, et al. Harmonisation of later-life cognitive function across national contexts: results from the Harmonized Cognitive Assessment Protocols. Lancet Healthy Longev. 2023;4(10):e573-e583. doi: 10.1016/S2666-7568(23)00170-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Levine DA, Whitney RT, Ye W, et al. Associations between stroke type, ischemic stroke subtypes, and poststroke cognitive trajectories. Stroke. 2025;56(4):898-907. doi: 10.1161/STROKEAHA.124.047640 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Springer MV, Whitney RT, Ye W, et al. Education levels and poststroke cognitive trajectories. JAMA Netw Open. 2025;8(3):e252002. doi: 10.1001/jamanetworkopen.2025.2002 [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Gottesman RF, Albert MS, Alonso A, et al. Associations between midlife vascular risk factors and 25-year incident dementia in the Atherosclerosis Risk in Communities (ARIC) cohort. JAMA Neurol. 2017;74(10):1246-1254. doi: 10.1001/jamaneurol.2017.1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Knopman DS, Gottesman RF, Sharrett AR, et al. Midlife vascular risk factors and midlife cognitive status in relation to prevalence of mild cognitive impairment and dementia in later life: the Atherosclerosis Risk in Communities study. Alzheimers Dement. 2018;14(11):1406-1415. doi: 10.1016/j.jalz.2018.03.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Chêne G, Beiser A, Au R, et al. Gender and incidence of dementia in the Framingham Heart Study from mid-adult life. Alzheimers Dement. 2015;11(3):310-320. doi: 10.1016/j.jalz.2013.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Satizabal CL, Beiser AS, Chouraki V, Chêne G, Dufouil C, Seshadri S. Incidence of dementia over three decades in the Framingham Heart Study. N Engl J Med. 2016;374(6):523-532. doi: 10.1056/NEJMoa1504327 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Callahan CM, Unverzagt FW, Hui SL, Perkins AJ, Hendrie HC. Six-item screener to identify cognitive impairment among potential subjects for clinical research. Med Care. 2002;40(9):771-781. doi: 10.1097/00005650-200209000-00007 [DOI] [PubMed] [Google Scholar]
50.Gillett SR, Thacker EL, Letter AJ, et al. Correlates of incident cognitive impairment in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. Clin Neuropsychol. 2015;29(4):466-486. doi: 10.1080/13854046.2015.1042524 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Cramer SC. Repairing the human brain after stroke. II. restorative therapies. Ann Neurol. 2008;63(5):549-560. doi: 10.1002/ana.21412 [DOI] [PubMed] [Google Scholar]
52.Stern Y. Cognitive reserve in ageing and Alzheimer’s disease. Lancet Neurol. 2012;11(11):1006-1012. doi: 10.1016/S1474-4422(12)70191-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Fratiglioni L, Paillard-Borg S, Winblad B. An active and socially integrated lifestyle in late life might protect against dementia. Lancet Neurol. 2004;3(6):343-353. doi: 10.1016/S1474-4422(04)00767-7 [DOI] [PubMed] [Google Scholar]
54.Debette S, Markus HS. The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: systematic review and meta-analysis. BMJ. 2010;341:c3666. doi: 10.1136/bmj.c3666 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Jokinen H, Kalska H, Ylikoski R, et al. ; LADIS group . Longitudinal cognitive decline in subcortical ischemic vascular disease–the LADIS study. Cerebrovasc Dis. 2009;27(4):384-391. doi: 10.1159/000207442 [DOI] [PubMed] [Google Scholar]
56.Gross AL, Nichols E, Renteria MA, et al. Risk and protective factors against cognitive decline in older adults from a nationally representative sample in India: results from the LASI-DAD. medRxiv. 2025;doi: 10.1101/2025.07.07.25331027 [DOI]
57.Hayden KM, Reed BR, Manly JJ, et al. Cognitive decline in the elderly: an analysis of population heterogeneity. Age Ageing. 2011;40(6):684-689. doi: 10.1093/ageing/afr101 [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Kunicki ZJ, Nichols E, De Vito AN, et al. The pace of memory decline in older adults without a neurocognitive disorder: results from 22 years of follow-up in a nationally representative sample. Neuroepidemiology. Published online September 4, 2025. doi: 10.1159/000548005 [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Palta P, Sharrett AR, Deal JA, et al. Leisure-time physical activity sustained since midlife and preservation of cognitive function: the Atherosclerosis Risk in Communities study. Alzheimers Dement. 2019;15(2):273-281. doi: 10.1016/j.jalz.2018.08.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Royall DR, Palmer R, Chiodo LK, Polk MJ. Normal rates of cognitive change in successful aging: the Freedom House study. J Int Neuropsychol Soc. 2005;11(7):899-909. doi: 10.1017/S135561770505109X [DOI] [PubMed] [Google Scholar]
61.Hochstenbach JB, den Otter R, Mulder TW. Cognitive recovery after stroke: a 2-year follow-up. Arch Phys Med Rehabil. 2003;84(10):1499-1504. doi: 10.1016/S0003-9993(03)00370-8 [DOI] [PubMed] [Google Scholar]
62.Jokinen H, Melkas S, Ylikoski R, et al. Post-stroke cognitive impairment is common even after successful clinical recovery. Eur J Neurol. 2015;22(9):1288-1294. doi: 10.1111/ene.12743 [DOI] [PubMed] [Google Scholar]
63.Makin SD, Turpin S, Dennis MS, Wardlaw JM. Cognitive impairment after lacunar stroke: systematic review and meta-analysis of incidence, prevalence and comparison with other stroke subtypes. J Neurol Neurosurg Psychiatry. 2013;84(8):893-900. doi: 10.1136/jnnp-2012-303645 [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Del Bene VA, Howard G, Gropen TI, et al. Cognitive decline after first-time transient ischemic attack. JAMA Neurol. 2025;82(4):323-332. doi: 10.1001/jamaneurol.2024.5082 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Girgenti SG, Brunson AO, Marsh EB. Baseline function and rehabilitation are as important as stroke severity as long-term predictors of cognitive performance post-stroke. Am J Phys Med Rehabil. 2023;102(2 suppl 1):S43-S50. doi: 10.1097/PHM.0000000000002125 [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Khan R, Devlin P, Urayama A, Ritzel RM. Models and mechanisms of post-stroke dementia and cognitive impairment. Front Stroke. 2025;4:1563924. doi: 10.3389/fstro.2025.1563924 [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Doyle KP, Buckwalter MS. Immunological mechanisms in poststroke dementia. Curr Opin Neurol. 2020;33(1):30-36. doi: 10.1097/WCO.0000000000000783 [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Iadecola C, Davisson RL. Hypertension and cerebrovascular dysfunction. Cell Metab. 2008;7(6):476-484. doi: 10.1016/j.cmet.2008.03.010 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials