You are here

  • Home
  • Online First
  • Cognitive trajectory in the first year after first-ever ischaemic stroke in young adults: the ODYSSEY study

Article Text

Article menu
Cognition
Original research
Cognitive trajectory in the first year after first-ever ischaemic stroke in young adults: the ODYSSEY study
  1. Mijntje M I Schellekens1,2,
  2. Ravi C S Springer1,
  3. Esther M Boot1,2,
  4. Jamie I Verhoeven1,2,
  5. Merel S Ekker1,2,
  6. Mayte E van Alebeek3,
  7. Paul J A M Brouwers4,
  8. Renate M Arntz4,
  9. Gert W van Dijk5,
  10. Rob A R Gons6,
  11. Inge W M van Uden6,
  12. Tom den Heijer7,
  13. Julia H van Tuijl8,
  14. Karlijn F de Laat9,
  15. Anouk G W van Norden10,
  16. Sarah E Vermeer11,
  17. Marian S G van Zagten12,
  18. Robert J Van Oostenbrugge13,14,
  19. Marieke J H Wermer15,16,
  20. Paul J Nederkoorn17,
  21. Frank G van Rooij18,
  22. Ido R van den Wijngaard19,
  23. Paul L M de Kort8,
  24. Frank-Erik De Leeuw1,2,
  25. Roy P C Kessels2,20,21,
  26. Anil M Tuladhar1,2
  1. 1Neurology, Radboudumc, Nijmegen, The Netherlands
  2. 2Radboud University Donders Institute for Brain Cognition and Behaviour, Nijmegen, The Netherlands
  3. 3Neurology, Gelre Hospitals, Zutphen, The Netherlands
  4. 4Neurology, Medisch Spectrum Twente, Enschede, The Netherlands
  5. 5Neurology, Canisius-Wilhelmina Hospital, Nijmegen, The Netherlands
  6. 6Neurology, Catharina Hospital, Eindhoven, The Netherlands
  7. 7Neurology, Franciscus Gasthuis & Vlietland, Rotterdam, The Netherlands
  8. 8Neurology, Elisabeth-TweeSteden Hospital, Tilburg, The Netherlands
  9. 9Neurology, Haga Hospital, Den Haag, The Netherlands
  10. 10Neurology, Amphia Hospital, Breda, The Netherlands
  11. 11Neurology, Rijnstate Hospital, Arnhem, The Netherlands
  12. 12Neurology, Jeroen Bosch Hospital, 's-Hertogenbosch, The Netherlands
  13. 13Neurology, Maastricht University Medical Centre, Maastricht, The Netherlands
  14. 14University Maastricht School for Mental Health and Neuroscience, Maastricht, The Netherlands
  15. 15Neurology, Leiden University Medical Centre, Leiden, The Netherlands
  16. 16Neurology, University Medical Centre Groningen, Groningen, The Netherlands
  17. 17Neurology, Amsterdam University Medical Centre, location AMC, Amsterdam, The Netherlands
  18. 18Neurology, Medical Centre Leeuwarden, Leeuwarden, The Netherlands
  19. 19Department of Neurology, Medical Centre Haaglanden, Den Haag, The Netherlands
  20. 20Vincent Van Gogh Instituut for Psychiatry, Venray, The Netherlands
  21. 21Department of Medical Psychology and Radboudumc Alzheimer Center, Radboudumc, Nijmegen, The Netherlands
  1. Correspondence to Anil M Tuladhar, Neurology, Radboudumc, Nijmegen 6500 HB, The Netherlands; Anil.Tuladhar{at}radboudumc.nl

Abstract

Background Limited data exists on cognitive recovery in young stroke patients. We aimed to investigate the longitudinal course of cognitive performance during the first year after stroke at young age and identify predictors for cognitive recovery.

Methods We conducted a multicentre prospective cohort study between 2013 and 2021, enrolling patients aged 18–49 years with first-ever ischaemic stroke. Cognitive assessments were performed within 6 months and after 1 year following the index event, covering seven cognitive domains. Composite Z-scores using normative data determined cognitive impairment (Z-score<−1.5). A Reliable Change Index (RCI) assessed cognitive recovery (RCI>1.96) or decline (RCI<−1.96).

Results 393 patients (median age 44.3 years, IQR 38.4–47.2) completed cognitive assessments with a median time interval of 403 days (IQR 364–474) between assessments. Based on RCI, a similar proportion of patients showed improvement and decline in each cognitive domain, while the majority exhibited no cognitive change. Among cognitively impaired patients at baseline, improvements were observed in processing speed (23.1%), visuoconstruction (40.1%) and executive functioning (20.0%). Younger age was associated with better cognitive recovery in visuoconstruction, and larger lesion volume was related to cognitive recovery in processing speed. No other predictors for cognitive recovery were identified.

Conclusions Cognitive impairment remains prevalent in young stroke even 1 year after the event. Most patients showed no cognitive change, however, recovery may have occurred in the early weeks after stroke, which was not assessed in our study. Among initially cognitively impaired patients, cognitive recovery is observed in processing speed, visuoconstruction and executive functioning. It is still not possible to predict cognitive recovery in individual patients.

  • COGNITION
  • STROKE

Data availability statement

Data are available upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jnnp-2023-332104

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Almost half of the young adults with stroke experience cognitive impairment afterwards. However, the longitudinal course of cognitive function in these individuals and factors associated with these changes are currently unknown.

    WHAT THIS STUDY ADDS

    • Comparable recovery and decline rates were found for the examined cognitive domains, but in most patients no cognitive change after the subacute phase was observed. We found no uniform predictors for cognitive recovery.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • Clinicians may use these results to inform their patients about their post-stroke prognosis, however caution is advised in discussing the expectation of cognitive recovery after the subacute phase.

    Introduction

    At least 1.5 million young adults (18–50 years) are affected by stroke every year, with an increasing global incidence.1 Post-stroke outcome does not only depend on motor function, but also on cognitive performance after stroke. A meta-analysis revealed that nearly half of young adults experience cognitive impairment after stroke, both in the (sub)acute and chronic phase (often after excluding patients with aphasia).2 These findings align with our recent study, which reported that up to 37% of young stroke patients experience impairment in five cognitive domains in the subacute phase (on average 3 months after ischaemic stroke): episodic memory; processing speed; visuoconstruction; executive functioning and attention and working memory.3 However, previous studies on cognitive performance at different post-stroke intervals were predominantly cross-sectional rather than longitudinal.2 Surprisingly, limited research has been done on the incidence and risk factors associated with cognitive recovery or decline after stroke at a young age. Information about post-stroke cognitive impairment and to identify those who will recover is essential for young stroke patients, as they will have to cope with the consequences for the rest of their lives, which can have an effect on family planning or a career.4 Several studies have examined potential predictors for cognitive recovery after ischaemic stroke in older patients. These predictors included younger age, female sex, higher level of education, stroke location and severity, vascular risk factors, emotional status, lesion volume and white matter hyperintensity severity, although results varied across studies.5–10 It remains unclear whether similar patterns and risk factors for cognitive performance apply to the younger stroke population. We therefore aimed to investigate the longitudinal trajectory of cognitive performance across multiple cognitive domains during the first year (ie, subacute and chronic phase) after first-ever ischaemic stroke, as well as explore factors associated with cognitive recovery in a cohort of young stroke patients.

    Patients and methods

    Patients and study design

    This study is part of the ‘Observational Dutch Young Symptomatic StrokE studY’, a multicentre prospective cohort study examining risk factors and prognosis of stroke at young age.3 11 The present study included patients aged 18–49 years with a first-ever ischaemic stroke with radiological evidence of cerebral ischaemia. Patients were included between May 2013 and February 2021. Exclusion criteria were a history of stroke, retinal infarction and cerebral venous sinus thrombosis. Detailed information regarding data collection has been provided elsewhere.11

    Cognitive assessment at baseline and follow-up

    Patients underwent an extensive neuropsychological assessment at baseline (within 6 months after stroke) and follow-up (1 year after stroke). The seven most relevant cognitive domains were assessed using multiple tests: episodic memory (3-trial version of the Rey Auditory Verbal Learning Test), processing speed (the written version of the Symbol-Digit Modalities Test, the abbreviated Stroop Color Word Test, parts I and II), visuoconstruction (Rey-Osterrieth Complex Figure-copy trial), executive functioning (fluency test, Stroop interference score, Brixton Spatial Anticipation Test), visual neglect (Star Cancellation of the Behavioral Inattention Test), language deficits (Short Token test), attention and working memory (Digit Span subtest from the Wechsler Adult Intelligence Scale-Fourth Edition). Global cognitive functioning was examined with the Mini-Mental State Examination.

    Further details regarding the collection of cognitive data can be found elsewhere.3

    Normative data from the Advanced Neuropsychological Diagnostics Infrastructure (ANDI), which includes data of 26 000 healthy individuals across all age groups were employed for most tests. This allowed fine-grained adjustment based on age, sex and education level.12 Prior to inputting the data of the abbreviated Stroop Color Word Test into ANDI, we multiplied the raw scores by two, to account for using half of the card. For the written version of the Symbol-Digit Modalities Test, we used the normative data from the test’s manual (n=1307),13 adjusted for age and education level, as this test is not available in the ANDI data set. For the Star Cancellation Test, which by definition is a negatively skewed outcome variable, we used a cut-off value (<44) instead of Z-scores, to indicate visuospatial neglect.14

    Based on the regression-based normative data from ANDI, adjusted for age and education level, raw test scores were converted to Z-scores per test for each participant. To correct for outliers, we adjusted Z-scores >3 or <−3 to 3 and −3, respectively.3 We used simple imputation with the median (education category five, ie, middle school/secondary vocational training) for the missing data on education level in two patients. Subsequently, the composite Z-score for each cognitive domain was computed by averaging the Z-scores of cognitive tests that are reflective of the same cognitive domain. If one test within a particular domain was missing, the domain score was based on the remaining tests within that domain. For each test, cognitive impairment was defined as a Z-score of <−1.5 (ie, reflecting performance more than 1.5 SD below the age-adjusted and education-adjusted normative mean). Cognitive impairment on a domain was defined as a composite Z-score of <−1.5 and below average performance was defined as a composite Z-score between −1.0 and −1.5.3 15

    To calculate mean test scores and proportions within the cognitive performance categories based on composite Z-score, only patients who fulfilled the domain score in both assessments were included.

    Reliable Change Index

    The raw test scores were converted to a Reliable Change Index (RCI) with the Chelune formula, which accounts for measurement errors and practice effects.16 More detailed information on the Chelune formula can be found in online supplemental file. Data from control groups was found in the literature.5 17–22 For the Star Cancellation Test, an RCI could not be computed, due to the non-parametric nature of the test’s outcome. After calculating the RCI for each individual cognitive test, we averaged RCIs of cognitive tests that assessed the same cognitive domain into domain RCI scores. If one test of a specific domain was missing, the domain RCI was based on the remaining tests within that domain. Next, we calculated the total RCI by averaging the RCIs per domain, reflecting an individual’s change across all cognitive domains. Reliably cognitive recovery was defined as an RCI greater than 1.96, while reliable decline was defined as an RCI less than −1.96. An RCI between −1.96 and 1.96 indicated an unchanged cognitive performance.16

    Supplemental material

    Other measurements

    Level of education was scored with a Dutch scoring system comprising seven categories23 that align with the UNESCO international classification of education levels.24 We assessed symptoms of depression and fatigue using the Mini International Neuropsychiatric Interview25 and the subscale Subjective Fatigue of the revised Checklist Individual Strength (CIS-20R),26 respectively.

    We used the Barthel Index27 and modified Rankin Scale (mRS)28 to assess functional outcome at the time of the baseline cognitive assessment. We defined good functional outcome as an mRS score of 0–1 and a Barthel Index of≥85.

    Additionally, we evaluated the aetiology of stroke (based on modified Trial of ORG 10172 in Acute Stroke Treatment)29 30 and severity at admission and discharge (National Institutes of Health Stroke Scale (NIHSS)),31 if necessary retrospectively, using a validated approach,32 33 because this scale was not consistently applied in all medical files.

    Lesion locations were visually identified, and semi-automatic segmentation was performed on diffusion-weighted imaging (DWI) (MRI<14 days after stroke) or FLAIR images (MRI>14 days after stroke) using ITK-SNAP. All lesion segmentations were visually inspected, and lesion volumes were calculated. We determined whether there was recurrent stroke before or between the cognitive assessments based on patient records.

    Statistical analysis

    We compared the baseline characteristics of patients who completed both the baseline and follow-up cognitive assessment with those who completed only a baseline cognitive assessment, using the independent t-test, Mann-Whitney U test or Pearson’s χ2 test (or Fisher’s exact test when an expected cell count was less than five) when appropriate.

    To investigate differences in proportions of cognitively impaired patients between baseline and follow-up, we used the McNemar test. A difference in proportion of recovered and declined patients per cognitive domain based on RCI was examined with Pearson’s χ2 test. We determined the association of post-stroke fatigue (based on CIS-20R) at baseline and the RCI per cognitive domain using linear regression.

    Potential predictive factors for recovery and cognitive performance at follow-up, selected based on the literature, including sex, age, NIHSS at admission and discharge, education level, lesion location, lesion volume, acute treatment, discharge destination, post-stroke fatigue, post-stroke depression,5–10 were tested. Additionally, the interval between stroke and first assessment was specifically assessed as a predictive factor for recovery and cognitive performance at baseline was specifically assessed as a predictive factor for cognitive performance at follow-up. Potential predictors for recovery were tested univariately in cognitively impaired patients at baseline using independent t-test, Mann-Whitney U test or Pearson’s χ2 test (or Fisher’s exact test when an expected cell count was less than five). In univariate tests, there was, if any, only one potential predictor significant for each domain. Therefore, multivariable analyses were not applicable.

    To analyse the potential predictors for cognitive performance at follow-up, we used linear regression. First, we assessed strong correlations among numeric potential predictors, setting a threshold at 0.7, and found that none exceeded this limit. Subsequently, we tested potential variables through univariable analysis, and if they were found to be significant, we included them in our multivariable linear regression. If the variable was categorical, we selected a reference category.

    To investigate the influence of recurrent stroke on the results, we performed post-hoc analyses, in which we conducted all above-described analyses after excluding patients with a recurrent stroke between baseline and follow-up assessment.

    All statistical analyses were performed using RStudio 2022.02.01.

    Results

    In total, 1322 patients had an ischaemic stroke at young age, of whom 598 completed a baseline neuropsychological assessment and 393 participated in both, a baseline and follow-up neuropsychological assessment (figure 1). Baseline characteristics of the study population are presented in table 1. Median age of patients at stroke onset was 44.3 years (IQR 38.4–47.2), 49.6% (n=195) were women. Median NIHSS at admission was 2 (IQR 1–5). Median time from index event to the first cognitive assessment was 80 days (IQR 54–114) and median time between the first and second cognitive assessment was 403 days (IQR 364–474). Baseline characteristics of patients with a baseline cognitive assessment only (n=205) are presented in online supplemental table 1. Participants who dropped out after the baseline assessment had a lower education level (p=0.037), a higher NIHSS at admission (p=0.010), more often had an unfavourable Barthel Index (p=0.027) and a longer time to baseline neuropsychological assessment (p=0.026) compared with the participants who completed both assessments.

    View this table:
    Table 1

    Baseline characteristics

    Figure 1
    Figure 1

    Flowchart of the study population.

    Individual neuropsychological test scores and the percentage of patients with cognitive impairment for each test are presented in table 2.

    View this table:
    Table 2

    Raw neuropsychological test scores and percentage of patients with cognitive impairment on a test

    Below-average performance and cognitive impairment

    Proportions of patients with below average performance and cognitive impairment based on composite Z-scores per domain are shown in figure 2. A smaller proportion of patients was classified as cognitively impaired at follow-up compared with baseline on episodic memory (11.7% vs 19.1%; p=0.001), processing speed (23.0% vs 27.6%; p=0.021) and language deficits (9.1% vs 17.0%; p<0.001).

    Figure 2
    Figure 2

    Domain-specific below average performance and cognitive impaired at baseline and follow-up. The proportion of patients (%) with young ischaemic stroke with domain-specific below average performance (composite Z-score between −1.5 and −1.0) or a cognitive impairment (composite Z-score<−1.5) at baseline and follow-up. For visual neglect a raw score <44 indicates impairment. Only patients with a domain score at baseline and follow-up were included. Missing values: episodic memory 10 (2.5%); processing speed 2 (0.5%); visuoconstruction 19 (4.8%); executive functioning 1 (0.3%); language deficits 29 (7.4%); attention and working memory 17 (4.3%).

    Change in performance category

    For all cognitive domains, most patients neither improved nor declined in performance category between baseline and follow-up (figure 3). We found the highest proportion of patients that improved in performance category for episodic memory (21.9%), processing speed (16.6%), visuoconstruction (26.8%) and language deficits (17.6%). In the other domains, less than 10% of patients improved. Focusing on patients who were cognitively impaired at baseline, most patients improved to below average or normal performance at follow-up in episodic memory (65.8%), executive functioning (56.0%) and language deficits (67.7%). In other cognitive domains, improvement in one-third to half of patients was observed.

    Figure 3
    Figure 3

    Change in performance categories between baseline and follow-up. Patients with improved, constant or declined performance based on composite Z-score between baseline and follow-up assessment for each cognitive domain. Only patients with a domain score at baseline and follow-up were included. Missing values: episodic memory 10 (2.5%); processing speed 2 (0.5%); visuoconstruction 19 (4.8%); executive functioning 1 (0.3%); language deficits 29 (7.4%); attention and working memory 17 (4.3%).

    Cognitive recovery per cognitive domain

    The total RCI and RCI per cognitive domain per patient are presented in figure 4. Overall, there was no cognitive change in most patients (98.7%) based on total RCI. Among the cognitive domains, the highest percentage of patients showed improvement in processing speed (7.2%) and visuoconstruction (18.1%). However, equal proportions of patients also experienced decline in these cognitive domains, with 8.2% declining in processing speed and 11.5% in visuoconstruction. For all domains, we observed equal proportions of patients with cognitive recovery and decline based on RCI. In patients who were cognitively impaired at baseline, recovery was observed in 25/108 (23.1%) in processing speed, 51/127 (40.1%) in visuoconstruction and 5/25 (20.0%) in executive functioning, while in other domains improvement was less than 4%.

    Figure 4
    Figure 4

    Reliable Change Index overall and per cognitive domain. RCI>1.96 indicates cognitive recovery, RCI<−1.96 indicates cognitive decline. Not displayed values (out of range): RCI total 1 (10.8), RCI processing speed 3 (−17.8, −11.0, 30.6). Missing values: episodic memory 10 (2.5%); processing speed 2 (0.5%); visuoconstruction 19 (4.8%); executive functioning 1 (0.3%); language deficits 29 (7.4%); attention and working memory 17 (4.3%).

    Higher scores of post-stroke fatigue at baseline were associated with a lower RCI in executive functioning (β=−0.012; p=0.004). However, the effect size was small with an R2adjusted of 0.021. There were no significant associations between post-stroke fatigue and RCI of the other cognitive domains.

    Predictive factors for cognitive recovery in cognitively impaired patients at baseline

    Patients who were cognitively impaired in processing speed at baseline and who recovered (based on RCI) had a larger lesion volume (26.8 mL (SD 37.7) vs 10.1 mL (SD 17.8), p=0.006) compared with patients who did not recover. In visuoconstruction, patients who were cognitively impaired at baseline and who recovered, were younger than patients who did not recover, with a median age of 40.2 years versus 44.6 years (p=0.045), respectively. No other predictive factors were found to be significant in any of the tested cognitive domains.

    Predictive factors for cognitive performance at follow-up

    Domain specific predictive factors for cognitive performance at follow-up are presented in online supplemental table 2. In all domains, the Z-score at baseline of a domain was significantly associated with the Z-score of that particular domain at follow-up. The effect sizes were medium to large (β varied between 0.39 and 0.83). In episodic memory, a high education level (compared with low) and in visuoconstruction middle and high education level (both compared with low) were significantly associated with lower Z-scores at follow-up (β varied between −0.19 and −0.29).

    Recurrent stroke

    After excluding patients with a recurrent stroke before the follow-up assessment (n=23), younger age was no longer a predictive factor for recovery of visuoconstruction in cognitively impaired patients at baseline. The results of all other analyses remained similar.

    Discussion

    In this prospective cohort study of young ischaemic stroke patients, we found that in most patients their cognitive performance did not change based on RCI in the first year after their index stroke. However, in cognitively impaired patients at baseline, we observed recovery based on RCI in one to two-fifths of the patients in processing speed, visuoconstruction and executive functioning. One to two-thirds of patients who were cognitively impaired at baseline were no longer categorised as cognitively impaired at follow-up in all domains (ie, they were classified as below average or normal). We could not identify predictive factors for cognitive recovery in patients who were cognitively impaired at baseline.

    We found that even 1 year after relatively mild strokes (median NIHSS 2) multi-domain cognitive impairment is still prevalent in young stroke patients. This is in line with previous cross-sectional young stroke studies with longer follow-up periods (up to several years).2 34 35 Studies on cognitive recovery in young stroke patients are scarce. A study in a relatively small sample of young stroke patients (n=87), with a median NIHSS of two, found lower prevalence rates of cognitive impairment after 3 months, compared with the acute phase (within 3 weeks), in processing speed, attention, executive functioning, but not in fluency.36 We found lower prevalence rates in the chronic phase compared with the subacute phase in episodic memory, processing speed and language deficits, but not in the other tested domains. This could be explained by differences in cognitive domains studied and tests used for assessing different domains. Besides that, it is possible that certain cognitive domains recover faster, resulting in differences in prevalence rates at different phases after stroke. A relatively small sample study (n=38) in stroke patients, aged 18–65, showed significant improvement in visuospatial function in the subacute phase (7 months) compared with the acute phase (1 week) and in working memory in the chronic phase (10 years), compared with the acute and subacute phase.37 Another study (n=153) in stroke patients, aged 18–65, showed that the majority of recovery in all seven tested cognitive domains took place within the first 6 months after stroke, but little recovery occurred after 6 months, as 90% of the patients remained either cognitively unimpaired or cognitively impaired between 6 months and 2 years.38 While these studies, with a smaller sample size, looked at group-level changes and did not use the RCI, most recovery seemed to occur in the first months after stroke. Therefore, it could be that we found relatively little recovery because our baseline assessment took place on average 3 months after stroke, and not in the acute phase. Also, recovery detected through RCI is a fairly strict method, which may have led to fewer patients with recovery than studies that used other methods to measure change in cognition. The latter could also explain why we were unable to find strong predictive factors for cognitive recovery. Since relatively few patients recovered, our study sample may not have had enough power to identify predictors for recovery. However, using the RCI to detect recovery is more reliable than using delta scores for change, since it reduces the chance of overestimating a true recovery.16 Post-stroke fatigue may play a role in the cognitive recovery after stroke. However, we only identified a significant association between post-stroke fatigue and cognitive recovery in executive functioning, but this does not explain the cognitive recovery in this domain as the effect size was very small. Although we did not find a uniform predictor, we found that in visuoconstruction younger patients who were cognitively impaired at baseline were more likely to recover. In processing speed patients with a larger stroke lesion who were cognitively impaired at baseline were more likely to recover. The latter finding could be explained by the fact that lesion volumes were mainly derived from DWI images, and a DWI lesion does not necessarily result in permanent brain damage. Second, the role of the exact location of the lesion and other structural brain changes, such as brain atrophy and pre-existent vascular damages, might influence cognitive recovery.39 The best predictor for cognitive performance at follow-up in our study was the cognitive performance at baseline. This finding is not surprising, considering the limited cognitive change over time. The finding that higher education levels were associated with lower cognitive performance in episodic memory and visuoconstruction at follow-up, as compared with lower educational levels, could be attributed to the prior adjustment of Z-scores for education differences. This unexpected finding, however, needs to be replicated and further investigated in more detail in future studies. We found a similar proportion of patients who declined compared with those who have recovered. This decline was not expected and cannot be explained by the presence of recurrent strokes. Other factors such as secondary neurodegeneration, ongoing chronic neuroinflammation, or ongoing damage to neural networks may be involved in determining the clinical status 1 year after the initial stroke index. However, future studies are needed to validate these findings and to understand the underlying pathophysiological mechanisms of this decline.

    Strong elements of this study are the large sample size of patients at young age with first-ever, radiologically confirmed, ischaemic stroke and its prospective and longitudinal multicentre character, in which general and university hospitals participated. Second, we used extensive neurocognitive testing instead of short cognitive screening tests, with limited missing data. Third, we used RCI analyses to assess cognitive recovery, which allowed us to adjust for learning effects. Finally, we used regression-based normative data based on large groups of healthy controls, resulting in clinically relevant classifications of test performances.

    However, several limitations need to be addressed. First, cognitive data of patients who were unable or refused to participate were lacking (for example because of severe stroke). This resulted in a population with relatively mild strokes (median NIHSS 2). This might affect the generalisability of our results to the whole young stroke population. Since patients without cognitive assessment had higher NIHSS scores,3 we expect that this bias, if any, would most likely lead to underestimation of the actual deficits. Second, cognitive data of patients who were unable or refused to participate in the follow-up assessment were lacking. These patients had more severe stroke, as indicated by higher NIHSS scores at admission, compared with patients with complete assessment. Third, due to logistic reasons not all neuropsychological tests were performed at the same time after stroke. The baseline assessment was assessed up to 6 months after stroke (with a median of 80 days after the index stroke), which might have affected the results, as recovery may have occurred in the first weeks after stroke. Furthermore, the wide temporal range of the baseline assessments may potentially influence the outcomes, as cognition may not remain stable during the first months after stoke. Fourth, we did not collect data on other interventions patients received, such as cognitive rehabilitation, which could have mediated cognitive recovery. Fifth, there could be a ceiling effect of multiple tests we used, making them less sensitive.

    Clinicians could use the results of this study to inform patients and their caregivers about what to expect regarding their cognitive performance after stroke. Future research should address the role of change in cognitive performance on functional outcomes, such as resuming work. In addition, it would be interesting to identify reliable predictors for the development and recovery of cognitive impairment. Strategic infarct locations, lower white matter integrity, or brain reserve might play a role in this process. Future studies should focus on these topics, since this can clarify the mechanisms of post-stroke cognitive recovery or decline, which also might be important in the neurorehabilitation process.

    In conclusion, cognitive impairment after mostly mild stroke in young adults is still common 1 year after stroke and in most patients, there was no cognitive change after the first months after stroke. Cognitive recovery could have taken place in the acute phase after stroke, which we were unable to investigate because of the lack of neuropsychological data for this timeframe in our study. These findings are relevant for young stroke patients, as they have to cope with these consequences for the rest of their lives. Predicting cognitive recovery for the individual patient remains difficult.

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involves human participants and was approved by the Medical Review Ethics Committee region Arnhem-Nijmegen approved the study (NL41531.091.12). We obtained written informed consent from all participants. If the patient was unable to provide informed consent, consent was provided by the patient’s legal representative.

    References

    1. ↵
      1. Ekker MS,
      2. Verhoeven JI,
      3. Vaartjes I, et al
      . Stroke incidence in young adults according to age, subtype, sex, and time trends. Neurology 2019;92:e2444–54. doi:10.1212/WNL.0000000000007533
      OpenUrlCrossRefPubMed
    2. ↵
      1. Weterings RP,
      2. Kessels RP,
      3. de Leeuw F-E, et al
      . Cognitive impairment after a stroke in young adults: A systematic review and meta-analysis. Int J Stroke 2023;18:888–97. doi:10.1177/17474930231159267
      OpenUrl
    3. ↵
      1. Schellekens MMI,
      2. Boot EM,
      3. Verhoeven JI, et al
      . Subacute cognitive impairment after first-ever transient ischemic attack or ischemic stroke in young adults: the ODYSSEY study. European Stroke Journal 2023;8:283–93. doi:10.1177/23969873221132032
      OpenUrl
    4. ↵
      1. Rutten-Jacobs LCA,
      2. Arntz RM,
      3. Maaijwee NAM, et al
      . Long-term mortality after stroke among adults aged 18 to 50 years. JAMA 2013;309:1136–44. doi:10.1001/jama.2013.842
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Nys GMS,
      2. Van Zandvoort MJE,
      3. De Kort PLM, et al
      . Domain-specific cognitive recovery after first-ever stroke: a follow-up study of 111 cases. J Int Neuropsychol Soc 2005;11:795–806. doi:10.1017/s1355617705050952
      OpenUrlCrossRefPubMed
    6. ↵
      1. Rasquin SMC,
      2. Lodder J,
      3. Verhey FRJ
      . Predictors of reversible mild cognitive impairment after stroke: a 2-year follow-up study. J Neurol Sci 2005;229–230:21–5. doi:10.1016/j.jns.2004.11.015
      OpenUrl
    7. ↵
      1. Khan M,
      2. Heiser H,
      3. Bernicchi N, et al
      . Leukoaraiosis predicts short-term cognitive but not motor recovery in ischemic stroke patients during rehabilitation. J Stroke Cerebrovasc Dis 2019;28:1597–603. doi:10.1016/j.jstrokecerebrovasdis.2019.02.037
      OpenUrlCrossRefPubMed
    8. ↵
      1. Patel M,
      2. Coshall C,
      3. Rudd AG, et al
      . Natural history of cognitive impairment after stroke and factors associated with its recovery. Clin Rehabil 2003;17:158–66. doi:10.1191/0269215503cr596oa
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Aben HP,
      2. De Munter L,
      3. Reijmer YD, et al
      . Prediction of cognitive recovery after stroke: the value of diffusion-weighted imaging-based measures of brain Connectivity. Stroke 2021;52:1983–92. doi:10.1161/STROKEAHA.120.032033
      OpenUrl
    10. ↵
      1. Shin M,
      2. Sohn MK,
      3. Lee J, et al
      . Effect of cognitive Reserve on risk of cognitive impairment and recovery after stroke: the KOSCO study. Stroke 2020;51:99–107. doi:10.1161/STROKEAHA.119.026829
      OpenUrlCrossRefPubMed
    11. ↵
      1. Arntz RM,
      2. van Alebeek ME,
      3. Synhaeve NE, et al
      . Observational Dutch young symptomatic stroke studY (ODYSSEY): study rationale and protocol of a Multicentre prospective cohort study. BMC Neurol 2014;14:55. doi:10.1186/1471-2377-14-55
    12. ↵
      1. de Vent NR,
      2. Agelink van Rentergem JA,
      3. Schmand BA, et al
      . Advanced neuropsychological diagnostics infrastructure (ANDI): A normative database created from control Datasets. Front Psychol 2016;7:1601. doi:10.3389/fpsyg.2016.01601
    13. ↵
      1. Smith A
      . SDMT Symbol Digit Modalities Test Amsterdam. The Netherlands: Hogrefe Uitgevers B.V, 2010.
    14. ↵
      1. Evans S,
      2. Tsao JCI,
      3. Lu Q, et al
      . Parent-child pain relationships from a Psychosocial perspective: A review of the literature. J Pain Manag 2008;1:237–46.
      OpenUrlPubMed
    15. ↵
      1. Sachdev PS,
      2. Brodaty H,
      3. Valenzuela MJ, et al
      . The neuropsychological profile of vascular cognitive impairment in stroke and TIA patients. Neurology 2004;62:912–9. doi:10.1212/01.wnl.0000115108.65264.4b
      OpenUrlCrossRefPubMed
    16. ↵
      1. Bruggemans E
      . Het Meten Van Cognitieve Veranderingen over de Tijd. In: KR HM, M G, B S, eds. Neuropsychologische diagnostiek: 2010:130-45. De klinische praktijk Amsterdam: Boom, 2010: 130–45.
    17. ↵
      1. Van der Elst W,
      2. Van Boxtel MPJ,
      3. Van Breukelen GJP, et al
      . Detecting the significance of changes in performance on the Stroop color-word test, Rey’s verbal learning test, and the letter digit substitution test: the regression-based change approach. J Int Neuropsychol Soc 2008;14:71–80. doi:10.1017/S1355617708080028
      OpenUrlPubMed
    18. ↵
      1. Levine AJ,
      2. Miller EN,
      3. Becker JT, et al
      . Normative data for determining significance of test-retest differences on eight common neuropsychological instruments. Clin Neuropsychol 2004;18:373–84. doi:10.1080/1385404049052420
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. van den Berg E,
      2. Nys GMS,
      3. Brands AMA, et al
      . The Brixton spatial anticipation test as a test for executive function: validity in patient groups and norms for older adults. J Int Neuropsychol Soc 2009;15:695–703. doi:10.1017/S1355617709990269
      OpenUrlCrossRefPubMed
    20. ↵
      1. Hensel A,
      2. Angermeyer MC,
      3. Riedel-Heller SG
      . Measuring cognitive change in older adults: reliable change indices for the mini-mental state examination. J Neurol Neurosurg Psychiatry 2007;78:1298–303. doi:10.1136/jnnp.2006.109074
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Van der Elst W,
      2. Van Boxtel MPJ,
      3. Van Breukelen GJP, et al
      . Normative data for the animal, profession and letter M naming verbal fluency tests for Dutch speaking participants and the effects of age, education, and sex. J Int Neuropsychol Soc 2006;12:80–9. doi:10.1017/S1355617706060115
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Wechsler D
      . Wechsler Adult Intelligence Scale – Fourth Edition - Nederlandse Bewerking: Technische Handleiding. Amsterdam: Pearson, 2012.
    23. ↵
      1. Verhage F
      . Intelligence and Age in Adults. Assen: Van Gorcum, 1964.
    24. ↵
      1. UNESCO
      . International Standard Classification of Education (ISCED 2011). Montreal, QC, Canada: UNESCO Institute for Statistics, 2011.
    25. ↵
      1. Sheehan DV,
      2. Lecrubier Y,
      3. Sheehan KH, et al
      . The mini-International neuropsychiatric interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry 1998;59 Suppl 20:22–33.
      OpenUrlPubMed
    26. ↵
      1. Vercoulen JH,
      2. Swanink CM,
      3. Fennis JF, et al
      . Dimensional assessment of chronic fatigue syndrome. J Psychosom Res 1994;38:383–92. doi:10.1016/0022-3999(94)90099-x
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. MAHONEY FI,
      2. BARTHEL DW
      . Functional evaluation: the Barthel index. Md State Med J 1965;14:61–5.
      OpenUrlCrossRefPubMed
    28. ↵
      1. van Swieten JC,
      2. Koudstaal PJ,
      3. Visser MC, et al
      . Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19:604–7. doi:10.1161/01.str.19.5.604
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Hennerici MG,
      2. Investigators PS
      . Rationale and design of the prevention of cerebrovascular and cardiovascular events of ischemic origin with Terutroban in patients with a history of ischemic stroke or transient ischemic attack (PERFORM) study. Cerebrovasc Dis 2009;27 Suppl 3:28–32. doi:10.1159/000209263
      OpenUrl
    30. ↵
      1. Ekker MS,
      2. Verhoeven JI,
      3. Schellekens MMI, et al
      . Risk factors and causes of ischemic stroke in 1322 young adults. Stroke 2023;54:439–47. doi:10.1161/STROKEAHA.122.040524
      OpenUrl
    31. ↵
      1. Brott T,
      2. Adams HP,
      3. Olinger CP, et al
      . Measurements of acute cerebral infarction: a clinical examination scale. Stroke 1989;20:864–70. doi:10.1161/01.str.20.7.864
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Kasner SE,
      2. Chalela JA,
      3. Luciano JM, et al
      . Reliability and validity of estimating the NIH stroke scale score from medical records. Stroke 1999;30:1534–7. doi:10.1161/01.str.30.8.1534
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Williams LS,
      2. Yilmaz EY,
      3. Lopez-Yunez AM
      . Retrospective assessment of initial stroke severity with the NIH stroke scale. Stroke 2000;31:858–62. doi:10.1161/01.str.31.4.858
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Schaapsmeerders P,
      2. Maaijwee NAM,
      3. van Dijk EJ, et al
      . Long-term cognitive impairment after first-ever ischemic stroke in young adults. Stroke 2013;44:1621–8. doi:10.1161/STROKEAHA.111.000792
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. de Bruijn MAAM,
      2. Synhaeve NE,
      3. van Rijsbergen MWA, et al
      . Long-term cognitive outcome of ischaemic stroke in young adults. Cerebrovasc Dis 2014;37:376–81. doi:10.1159/000362592
      OpenUrlCrossRefPubMed
    36. ↵
      1. Pinter D,
      2. Enzinger C,
      3. Gattringer T, et al
      . Prevalence and short-term changes of cognitive dysfunction in young ischaemic stroke patients. Eur J Neurol 2019;26:727–32. doi:10.1111/ene.13879
      OpenUrl
    37. ↵
      1. Elgh E,
      2. Hu X
      . Dynamic trajectory of long-term cognitive improvement up to 10 years in young community-dwelling stroke survivors: A cohort study. Front Neurol 2019;10:97. doi:10.3389/fneur.2019.00097
    38. ↵
      1. Turunen KEA,
      2. Laari SPK,
      3. Kauranen TV, et al
      . Domain-specific cognitive recovery after first-ever stroke: A 2-year follow-up. J Int Neuropsychol Soc 2018;24:117–27. doi:10.1017/S1355617717000728
      OpenUrlCrossRef
    39. ↵
      1. Hobden G,
      2. Moore MJ,
      3. Colbourne E, et al
      . Association of neuroimaging markers on clinical CT scans with domain-specific cognitive impairment in the early and later Poststroke stages. Neurology 2023;101:e1687–96. doi:10.1212/WNL.0000000000207756
      OpenUrl

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors MMIS researched literature, did data analysis, wrote the first draft of the manuscript,

      is responsible for the overall content as guarantor. RMA, MEvA, F-EDL, RK were involved in protocol development. F-EDL gained ethical approval. All authors were involved in patient recruitment. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests AMT is a junior staff member of the Dutch Heart Foundation (grant number 2016T044). F-EDL is a clinical established investigator of the Dutch Heart Foundation (2014 T060). MJHW has received a VIDI grant (9171337) of the ZonMw/NWO and the clinical established investigator Dutch Heart Foundation grant (2016T86).

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.