Guidelines International Journal of Stroke 2022, Vol. 17(3) 260­–268 International stroke genetics consortium © 2021 World Stroke Organization https://doi.org/10.1177/17474930211007288 recommendations for studies of genetics Article reuse guidelines: sagepub.com/journals-permissions of stroke outcome and recovery DOI: 10.1177/17474930211007288 journals.sagepub.com/home/wso Arne G Lindgren1,2 , Robynne G Braun3, Jennifer Juhl Majersik4, Philip Clatworthy5, Shraddha Mainali6, Colin P Derdeyn7, Jane Maguire8, Christina Jern9,10, Jonathan Rosand11, John W Cole12,13, Jin-Moo Lee14, Pooja Khatri15, Paul Nyquist16 , Stéphanie Debette17,18, Loo Keat Wei19, Tatjana Rundek20, Dana Leifer21, Vincent Thijs22 , Robin Lemmens23,24, Laura Heitsch14, Kameshwar Prasad25, Jordi Jimenez Conde26,27, Martin Dichgans28, Natalia S Rost11, Steven C Cramer29,30, Julie Bernhardt22 , Bradford B Worrall31 and Israel Fernandez- Cadenas32; International Stroke Genetics Consortium Abstract Numerous biological mechanisms contribute to outcome after stroke, including brain injury, inflammation, and repair mechanisms. Clinical genetic studies have the potential to discover biological mechanisms affecting stroke recov- ery in humans and identify intervention targets. Large sample sizes are needed to detect commonly occurring genetic variations related to stroke brain injury and recovery. However, this usually requires combining data from multiple studies where consistent terminology, methodology, and data collection timelines are essential. Our group of expert stroke and rehabilitation clinicians and researchers with knowledge in genetics of stroke recovery here present recommendations for harmonizing phenotype data with focus on measures suitable for multicenter genetic studies of 1 Department of Clinical Sciences Lund, Neurology, Lund University, Lund, Sweden 18 2 Department of Neurology, Skåne University Hospital, Lund, Sweden Neurology Department, Bordeaux University Hospital, Bordeaux, France 19 3 Department of Neurology, University of Maryland, Baltimore, MD, USA Department of Biological Science, Faculty of Science, Universiti Tunku 4 Department of Neurology, University of Utah, Salt Lake City, UT, USA Abdul Rahman, Perak, Malaysia 20 5 Department of Neurology, North Bristol NHS Trust, Bristol, UK Department of Neurology, University of Miami Miller School of 6 Department of Neurology, The Ohio State University, Columbus, OH, USA Medicine, Miami, FL, USA 21 7 Department of Radiology, University of Iowa, Iowa City, IA, USA Department of Neurology, Weill Cornell Medicine, New York, NY, USA 22 8 Faculty of Health, University of Technology Sydney, Ultimo, NSW, Stroke Theme, Florey Institute of Neuroscience and Mental Health, Australia Melbourne, Vic, Australia 23 9 Department of Laboratory Medicine, Institute of Biomedicine, Department of Neuroscience, University of Leuven, Leuven, Belgium 24 University of Gothenburg, Gothenburg, Sweden Department of Neurology, University Hospitals Leuven, Leuven, 10 Department of Clinical Genetics and Genomics, Sahlgrenska University Belgium 25 Hospital, Gothenburg, Sweden Rajendra Institute of Medical Sciences, Ranchi, Jharkhand, India 26 11 Department of Neurology, Massachusetts General Hospital, Boston, Neurology Department, Neurovascular Research Group, Institut MA, USA Hospital del Mar d’Investigació Mèdica, Barcelona, Spain 27 12 Neurology Service, Baltimore Veterans Affairs Medical Center, Neurology, Universitat Autònoma de Barcelona, Barcelona, Spain 28 Baltimore, MD, USA Institute for Stroke and Dementia Research, University Hospital, LMU, 13 Department of Neurology, University of Maryland School of Medicine, Munich, Germany 29 Baltimore, MD, USA Department of Neurology, UCLA, Los Angeles, CA, USA 30 14 Department of Neurology, Washington University School of Medicine, California Rehabilitation Institute, Los Angeles, CA, USA 31 St. Louis, MO, USA Department of Neurology, University of Virginia, Charlottesville, VA, 15 Department of Neurology and Rehabilitation Sciences, University of USA 32 Cincinnati, Cincinnati, OH, USA Stroke Pharmacogenomics and Genetics Group, Sant Pau Biomedical 16 Neurology, Anesthesiology/Critical Care Medicine, Neurosurgery, and Research Institute, Barcelona, Spain General Internal Medicine, Johns Hopkins School of Medicine, Baltimore, Corresponding author: MD, USA Arne G Lindgren, Department of Neurology, Skåne University Hospital, 17 Bordeaux Population Health, Inserm U1219, University of Bordeaux, SE-22185 Lund, Sweden. Bordeaux, France Email: arne.lindgren@med.lu.se International Journal of Stroke, 17(3) International Journal of Stroke, 0(0) Lindgren 2 et al. 261 International Journal of Stroke 0(0) ischemic stroke brain injury and recovery. Our recommendations have been endorsed by the International Stroke Genetics Consortium. Keywords Data collection, genetics, ischemic stroke, outcome, phenotype, recovery, standardization Received: 9 March 2021; accepted: 12 March 2021 timing and metric choices across studies, the Stroke Introduction Recovery and Rehabilitation Roundtable taskforce in Genetic studies can potentially discover biological 2017 recommended core outcomes for trials and stan- mechanisms affecting stroke recovery with treatment dardized measurement time points to reduce implications. However, they need large sample sizes heterogeneity.11 only achievable by combining data from multiple stu- Here, we focus specifically on design of prospective dies, where harmonized terminology, methodology, genetic studies of ischemic stroke (IS) recovery, aiming and data collection timelines are essential. to ascertain the underlying genetic influences on stroke The terms stroke outcome and stroke recovery recovery biology. Our recommendations complement differ in meaning. Stroke outcome describes the degree existing advise for standardizing phenotype data12 of function at specific time points; stroke recovery and biological sample collection13 for stroke risk and encompasses the degree of improvement (or deterior- recovery studies11,14 by providing recommendations for ation) over time and better captures dynamic biological pre-specified harmonized data sets suitable for large, processes. Stroke recovery evaluation requires high-quality, multi-center collaborations in prospective initial stroke severity data, without which only stroke stroke genetic recovery studies. We propose measures outcome is measurable. It is also important to distin- comprehensive enough to provide both stroke- and guish restitution (‘‘true’’) recovery from behavioral domain-specific data, but simple enough to allow col- compensation. For example, ‘‘true’’ motor recovery lection of large sample sizes across numerous and suggests restoration of pre-stroke movement patterns1 diverse enrollment sites. This will allow opportunities whereas ‘‘compensation,’’ implies new (possibly dys- to discover genetic factors influencing hitherto functional) movement patterns for accomplishing func- unknown biological pathways affecting the dynamics tional tasks.2 of IS recovery. We do not here consider intracerebral The dynamics of stroke recovery depend on intrinsic hemorrhage (ICH) given ICH recovery mechanisms and extrinsic factors.3 Each patient’s recovery pattern differ from IS. uniquely reflects the combined influences of lesion size and location, biological mechanisms of brain repair, comorbidities, pre-morbid health status, and post- Methods stroke factors including acute recanalization, rehabili- Methods for reaching a consensus on these recommen- tation, psychosocial factors, and environmental influ- dations are described in the Supplement. ences. Consequently, the degree of stroke recovery varies considerably between individuals, and even skilled clinicians have difficulty making accurate recov- Results ery predictions.4 The need for improved predictive models of Overview of phenotypic variables stroke recovery has become a major research focus5,6 We prioritized phenotypic variables into categories: (1) and recent studies suggest that genetic variations influ- minimum variables—mandatory for all studies, (2) pre- ence recovery after stroke.7–9 Despite multiple studies, ferred variables— recommended but may be precluded findings remain heterogeneous, due to differences in by practical limitations, and (3) optional variables— populations, recovery metrics, assessment time interesting for some multi-center projects. Table 1 points, and study designs. Most studies using global shows the minimum (mandatory) variable set. assessments incorporate the modified Rankin Scale Supplemental Table 1 lists a detailed comprehensive (mRS)10 while some use more detailed modality-specific set. Supplemental Table 2 suggests variable formats to functions, for example, upper extremity (UE) motor facilitate compilation of joint data. Regularly updated function, language or cognitive function,3,11 or versions will be kept at the Global Alliance for patient-reported outcome measures (PROMs). Few stu- International Stroke Genetics Consortium (ISGC) dies use repeated measures, leading to knowledge Acute and Long-term Outcome studies (https://genes- gaps on stroke recovery time course. To standardize troke.wixsite.com/alliesinstroke). International Journal of Stroke, 0(0) International Journal of Stroke, 17(3) Table 1. Recommended minimum variable sets for genetic studies of ischemic stroke recovery. 262 Lindgren Evaluation time Clinical Stroke clinical Stroke imaging Treatment Functional et al. Stroke onset- Pre-stroke/demographicsa � Main stroke typeb � Initial CT/MRI examin- � Thrombolysis � Initial stroke severity: NIHSSc two days � Pre-stroke mRS � TOAST/CCS subtype ation performed � Thrombectomy within 6 hd after hospital � Charlson Comorbidity IndexEP � SurvivalLP � Time to initial CT/MRI presentation (when possible) � Age at time of stroke onset � Time from stroke scanLO or just before recanalization � Sex to deathLP � CT at 24 hLO therapy � Race � Time to CT 24 h scanLO � Time from stroke onset to Risk factors at stroke onseta � Hemorrhagic trans- initial NIHSSLP formation International Journal of Stroke, 17(3) � Hypertension � NIHSSc 24 h after recanaliza- � Atrial fibrillation on 24 h CT scanLO tion therapy/24 h after base- � Coronary heart disease line NIHSS, if no � Diabetes mellitus recanalization therapyLO � Smoking � Time from stroke onset to 24 h � Hypercholesterolemia NIHSSLO � Previous stroke Seven days/ As above, if not already done � Survival � NIHSSc at seven days or at discharge � Time from stroke discharge, if earlierEP to deathLP � Time from stroke onset to NIHSS at seven days/ dischargeEP Three months � Time of evaluationEO � NIHSSc,EO � SurvivalEO � mRSEO � Recurrent strokeEO 12 months, yearly � Time of evaluationEO � NIHSSc,EO thereafter � SurvivalEO � mRSEO � Recurrent strokeEO Listed variables are recommended as minimum for Early Phase Studies (with focus on 0–48 h and seven days/hospital discharge) and Later Phase Studies (with focus on three months and beyond), unless otherwise specified. A comprehensive outline of all suggested minimum, preferred, and optional variables are shown in Supplemental Table 1. Time of evaluation after stroke onset (hours for up to 72 h; days thereafter) should be registered. mRS: modified Rankin Scale; IS: ischemic stroke; ICH: intracerebral haemorrhage; h: hours; NIHSS: National Institutes of Health Stroke Scale; CT: computed tomography; MRI: magnetic resonance imaging; TOAST: Trial of ORG 10172 in Acute Stroke Treatment; CCS: Causative Classification of Stroke. a Can often be collected somewhat later. b Only IS in this manuscript. c Including individual subitems. d For Later Phase studies: NIHSS within 6 h ¼ preferred, NIHSS within 24 h ¼ minimum. EP Early Phase studies, preferred. LP Later Phase studies, preferred. EO Early Phase studies, optional. LO International Journal of Stroke, 0(0) 3 International Journal of Stroke 17(3) Later Phase studies, optional. Lindgren 4 et al. 263 International Journal of Stroke 0(0) Figure 1. Framework showing time points post stroke related to current known biology of stroke recovery. Time post stroke should always be included in data acquisition. Adapted to represent ischaemic stroke only, from: Bernhardt J et al, Int J Stroke 2017, Vol. 12(5) 444–450, copyright  2017 by World Stroke Organization. Reprinted by permission of SAGE Publications, Ltd. stroke preceding the index stroke versus other condi- tions. We also recommend the Charlson Comorbidity Timing of recovery assessment Index,20 with information about pre-existing key med- Stroke recovery starts immediately at symptom onset ical conditions. For further details, see Table 1 and and continues for years thereafter (Figure 1). Blood Supplemental Table 1. biomarkers, and other biomarker evaluations, for All studies should provide demographics: age at example, magnetic resonance imaging (MRI), often stroke onset, sex, race/ethnicity, residential area type vary across time points. To provide simplification, we (urban/rural), educational status, living situation recommend the time course for assessment of evolution (housing type), and available social support (living and recovery into three phases post-stroke (Day 0 is alone/with someone).21 day of stroke onset): 0 to 24–48 h, seven days, and approximately Day 90 after stroke onset and when pos- Baseline clinical and imaging information. Baseline charac- sible one year and later. When appropriate, studies may teristics of current IS should include initial National choose additional precisely defined time periods. Institutes of Health Stroke Scale (NIHSS) total and Studies of hyperacute recovery and therapy should individual component scores, and Trial of ORG perform evaluations within 6 h (when possible) or at 10172 in Acute Stroke Treatment22 and/or Causative least within 24 h after stroke onset and before revascu- Classification of Stroke subtype.23 Specific ‘‘other larization therapy, followed by a new evaluation at 24 h determined’’ stroke etiologies (e.g., cervical artery dis- post stroke15 or 24 h after recanalization therapy (see section) could be detailed. Laboratory parameters and below). Glasgow Coma Scale may be recorded. Seven days post-stroke is often recommended for We recommend baseline registration of head com- evaluation.1 However, because many patients leave puted tomography/magnetic resonance (CT/MRI), hospital earlier, we recommend evaluation either at and CT/MRI angiography and perfusion, because, for seven days or discharge, whichever occurs earlier. example, dynamic blood flow changes may be related to IS studies often conclude evaluations at three genetic influences.24 months.16,17 However, improvement may occur at 6–12 months and possibly beyond.18 Recovery is not Stroke treatment and neuroimaging at 0–48 h and seven days/ linear, and time frames may vary by different domains, hospital discharge. Treatment with thrombolysis and for example, cognitive versus motor function.19 To thrombectomy should be noted. Final expanded evaluate three-month recovery independently of early thrombolysis in cerebral infarction (TICI) score25 acute phases, sometimes influenced by treatments, for indicating degree of revascularization achieved example, revascularization, we recommend measuring should be mandated in large vessel occlusion recovery as functional change between Day 7 (or dis- stroke studies. Additional treatments possibly affect- charge), and three months. If possible, additional evalu- ing recovery include carotid endarterectomy/stenting ations at one and three years are strongly recommended and pharmacologic interventions for blood pressure, to evaluate longer term recovery. dyslipidemia, or atrial fibrillation. Follow-up imaging at 24 h after recanalization ther- apy with CT/MRI is valuable to evaluate location and Recommended phenotypic variables extent of the acute ischemic lesion(s). When possible, Pre-stroke variables, demographic data. Pre-stroke func- MRI with FLAIR, DWI, MRI angiography, and tional status affects stroke outcome and should be mea- GRE/T2*/SWI is recommended within 24 h (or within sured as mRS, ideally specifying whether due to a seven days at the latest) after stroke onset. MRI International Journal of Stroke, 0(0) International Journal of Stroke, 17(3) 264 Lindgren et al. International Journal of Stroke 17(3)5 performed later might also have value. Imaging meas- linked with quality of life and participation level, and ures of leukoaraiosis, microhemorrhages, prior infarcts, testing does not require much time. On Day 7, we rec- and arterial stenoses could be considered. Injury extent ommend recording the ability to walk 10 m independ- to specific neural structures, such as corticospinal tract, ently (yes/no), and for those able, a 10 -m walk test. may be useful for some hypotheses. This may be repeated at later time points (see below). Neuroimaging biomarkers can serve as endopheno- Early complications such as infections and recurrent types. For examples, please see Supplement. stroke may also influence recovery and should be considered. Clinical measures at 0–48 h and seven days/hospital discharge. In the first days after stroke, neurological def- Considerations and treatment information up to three months icits can be highly unstable, with patients rapidly and beyond. Stroke recurrence, with a 30%–40% cumu- improving, or deteriorating. Serial NIHSS scores,26 lative risk among first stroke survivors, is a common often standard of care in acute stroke, capture these cause of worsening disability and requires tracking.29,30 changes. A change in NIHSS between baseline (<6 h Secondary prevention measures, complications (e.g., from onset) and 24 h (NIHSS6–24 h) is related to 90- depression, infections, seizures, fractures after falls), day outcome independent of baseline NIHSS27 with a level of physical activity, and socioeconomic factors genome wide association study (GWAS) of NIHSS6– may substantially affect outcome and recovery. At the 24 h having revealed genes potentially involved in ische- designing stage, studies should define which of these mic brain injury (data not shown). We therefore recom- variables to collect as confounding factors, exclusion mend NIHSS (including subitems) at baseline <6 h or criteria, or endpoint/dependent variables. at least within 24 h after stroke onset, and short-term Rehabilitation treatment is heterogeneous across follow-up at 24 h after stroke onset/after recanalization centers and difficult to uniformly register. We suggest therapy, noting number of hours since stroke onset. registering how often the treatment is administered per Recovery during the initial days after stroke onset is week or month and duration of rehabilitation in days. difficult to measure, and we recommend evaluations The starting day after stroke onset and treatment dose including NIHSS (with subitems) either at seven days (minutes per day) may be recorded. or discharge from hospital, whichever occurs earlier. Treatment with antidepressants and other psycho- The Shoulder Abduction Finger Extension score tropic medication31 should be noted as should other during the first three days after stroke predicts upper rehabilitation adjuncts, whether pharmacologic or limb motor outcome.28 This complements the NIHSS device-based (e.g., transcranial magnetic stimulation). and is useful as an early marker, easier to assess than more complex motor assessments such as the Fugl- Evaluation at three months and beyond. Factors influencing Meyer (FM) or Action Research Arm Test (ARAT). long-term recovery (improvement/deterioration) may Gait performance measured as walking speed pre- differ from those important in earlier time periods. As dicts walking recovery and falls risk. Gait is also mentioned above, we recommend evaluation at Day 7, Figure 2. Suggested domain-specific screening by using NIHSS. Detected deficits are subsequently assessed with more detailed evaluations. NIHSS: National Institutes of Health Stroke Scale. International Journal of Stroke, 17(3) International Journal of Stroke, 0(0) Lindgren 6 et al. 265 International Journal of Stroke 0(0) or discharge (if earlier) as a new baseline for long-term language differences between populations. We favor recovery at three months. the Western Aphasia Battery-Revised version bedside Stroke variably affects different functional screening test, which takes 10–15 min and is well- domains.32 We recommend that specific domains are accepted by researchers.40 Language evaluation items considered separately and only in more detail where in cognitive tests are also a possibility. appropriate. For example, if a motor deficit is detected on the NIHSS, more in-depth motor testing can be per- Neglect. NIHSS item 11 provides a screening tool for formed (Figure 2 and Supplemental Table 1). In this neglect and hemi-inattention. Among the many avail- way, the NIHSS subitems provide screening for deficits able bedside assessments, the Star cancellation test is requiring more detailed evaluation, saving time, and recommended. resources. Mood. The Hospital Anxiety and Depression Scale41 had most consensus in our group for utility across dif- Evaluation of specific recovery domains ferent time points. Alternatives have been recom- Motor function. Motor deficits are seen in >80% of IS33 mended by others.42 and can be screened by NIHSS items 5 and 6. A more detailed assessment of motor deficit changes over time Other specific domains. Post-stroke visual field loss, eye is of great importance to evaluate recovery. The FM- movement abnormalities, dysphagia, balance disorders, UE motor scale34 is well known and recommended but fatigue, frailty, and urinary incontinence are all import- requires trained personnel.35 The FM lower extremity ant aspects of post-stroke recovery. We agreed that no motor scale may be considered,34 but limited reprodu- specific recommendations can be made for these cibility, a high concordance with UE weakness, and domains at this time but provide some suggestions in overlapping recovery mechanisms may limit its value. the Supplement. UE motor function is best captured with ARAT, but this requires equipment.11 Global assessment. The three-month mRS is used in most Gait velocity (see above) is also useful for long-term stroke trials and should be performed in studies of motor function evaluation. stroke recovery genetics to facilitate comparison across cohorts. Evaluation of mRS at other time Sensory function. The FM Sensory Examination or the points (e.g., 6 months, 12 months, and yearly there- Nottingham Sensory Scale could be considered. after) may be useful. The mRS offers advantages of ease of administration, good inter-observer reproduci- Cognitive function. Combining the four NIHSS items, bility, certification, and available phone-based evalu- Orientation (1b), Executive function (1c), Language ation.10,43. The mRS score has been analyzed both as (9), and Inattention (11), has similar value as the a continuous and an ordinal variable,44,45 but dichoto- Mini-Mental State Examination in detecting severe mization may lose information and statistical power. cognitive impairment.36 A more elaborate cognitive Other functional scales, such as Barthel Index and evaluation with the Montreal Cognitive Assessment the Nottingham extended activities of daily living scale Scale37 is recommended when possible. When even (ADL), have limitations such as ceiling effects or rarer more detailed or longitudinal understanding of specific usage. cognitive domains is needed, an in-depth neuropsycho- logical assessment may encompass multiple cognitive Patient-reported outcome measures. Outcome and recovery domains, especially verbal episodic memory, executive evaluations important to clinicians are not always con- function, and processing speed. Pre-stroke cognitive gruent with those of patients. When possible, PROMs assessment with tools such as the informant question- should be included in recovery studies to support the naire on cognitive decline in the elderly (IQCODE)38 is validity of other measures and reflect meaningful stroke important, as pre-stroke cognitive impairment is fre- outcomes and recovery. PROMs can assess disability, quent and associated with post-stroke dementia.39 mood, cognitive function, pain, mobility, and fatigue. The genetics of post-stroke cognitive impairment are The Patient-Reported Outcome Measurement not covered here but addressed in separate working Information System, 36-Item Short Form Survey, groups of the ISGC (www.strokegenetics.org) and the EuroQuality of life 5 dimension questionnaire (EQ- Cohorts for Heart and Aging Research in Genomic 5D), and Stroke Impact Scale are frequently used Epidemiology consortium. PROMs.46 Speech function. NIHSS item 9 provides a screening tool Combining dynamic changes from different domains. Genetic for aphasia. Aphasia evaluations are hampered by correlates of recovery mechanisms may influence International Journal of Stroke, 0(0) International Journal of Stroke, 17(3) 266 Lindgren et al. International Journal of Stroke 17(3)7 several functional domains. Combined measurements domains can be complementary, as can combined across domains can be obtained by quantification of measures such as the predict recovery potential 2 algo- the domain with greatest impairment in individual sub- rithm evaluating clinical function, MRI, and surrogate jects (defined as the system with the worst baseline sub- parameters to predict three-month UE motor func- score from the baseline NIHSS), and computing the tion.28 Other clinical evaluations to predict recovery percentage of the maximum possible score for this such as sitting balance for independent walking and domain followed by comparing these measures on ability to comprehend and repeat spoken language are Day 7 and Day 90. Recovery is calculated as the uncommonly standardized and systematically investi- remaining deficit (% recovery ¼ 100 � (1�(ScoreMax gated and may currently have less value for genetic � Scored90)/(ScoreMax � Scored7))) for each subject.47 studies of stroke recovery. Increasing importance is being placed on PROMs to ensure that recovery mea- Neuroimaging. Neuroimaging after stroke can detect sured using tools based on neurological impairment is new infarcts, hemorrhages, and small vessel disease meaningful from the patient’s perspective, although the including white matter changes and brain atrophy. role of PROMs in stroke genetics research has not been For these purposes, MRI including FLAIR and established. GRE/T2*/SWI sequences could be considered at three Training, certification, and recertification are essen- months, one year, and later. tial to reduce error and inter-rater variance. A plan for Several other forms of neuroimaging and training, certification, and recertification for each associated methods have been examined in relation to behavioral scale should be a part of every stroke recov- genetic variation, for examples—please see ery study or trial. Supplement. Statistical considerations are important. Many scales are ordinal and non-linear. An improvement in the NIHSS of 10 points, for instance, may signify different Discussion degrees of improvement when a patient improves from We here recommend a specific set of phenotype out- 20 to 10 versus from 10 to 0. Additional concerns come variables, time frames, and covariates for pro- regarding repeated measurements include regression spective genetic studies of recovery after IS. To detect to the mean and management of missing data. changes in the patient-specific evolution of symptoms Analyses must consider collinearity when employing the same variables should, when possible, be measured the same variable to calculate both the independent at the different time points. and the dependent variables to avoid misinterpretation Our suggested time points for evaluations and the of paired observations when comparing baseline scores assessments categorized as minimum, preferred, or with follow-up results.48 Analyses combining different optional can be useful tools for individual studies, com- domains may be considered for detecting genetic influ- parative studies, and multi-center studies on stroke ence on general stroke recovery. recovery genetics. Of the large number of potential evaluation tools available for assessment of IS recov- ery, we suggested tools that should be simple and Conclusions accessible while detailed enough to capture dynamic The rapid progress of genetic research methodologies changes in the designated domains. provides excellent opportunities to discover new factors Physical follow-up examinations after the acute influencing stroke recovery. However, to obtain optimal phase of stroke are labour intensive. Patient tele- efficiency, harmonized and well-accepted phenotyping phone interviews may be an alternative. Live exam- instruments across studies are required. We suggest inations permit detailed determination of many selected evaluations of stroke recovery to measure neurological features but come at a higher price important recovery domains. Harmonization of these such as cost and travel. Phone and video-based evaluations between studies will allow performance of examinations are less expensive, but more limited large prospective studies of genetic influence on recovery in the data that can be reliably measured. Given dynamics in the early and later phases after stroke. the focus of the current recommendations, we advise live examinations for studies focusing on recovery at 90 days and beyond to be performed Declaration of conflicting interests whenever resources permit. The author(s) declared the following potential conflicts of We stress the use of NIHSS, including its subscores, interest with respect to the research, authorship, and/or pub- for screening because it is already widely utilized in lication of this article: Apart from what is mentioned under clinical routine, clinical trials, and recovery studies. Disclosures in the Supplement, there is no other potential More elaborate evaluations focusing on specific conflicts of interest. International Journal of Stroke, 17(3) International Journal of Stroke, 0(0) Lindgren 8 et al. 267 International Journal of Stroke 0(0) Funding 8. Söderholm M, Pedersen A, Lorentzen E, et al. The author(s) disclosed receipt of the following financial sup- Genome-wide association meta-analysis of functional port for the research, authorship, and/or publication of this outcome after ischemic stroke. Neurology 2019; 92: article: This work was supported by National Institute of e1271–e1283. Child Health and Human Development (K12HD093427); 9. Mola-Caminal M, Carrera C, Soriano-Tarraga C, et al. Lund University; National Institute of Neurological PATJ low frequency variants are associated with worse Disorders and Stroke (K23NS099487-01, R01-NS082285, ischemic stroke functional outcome. Circ Res 2019; 124: R01-NS086905, U19-NS115388 and 5U10NS086606); 114–120. National Institutes of Health (KL2TR003016, R01- 10. Shinohara Y, Minematsu K, Amano T and Ohashi Y. NS085419, R01-NS100178, R01-NS105150, R01-NS114045, Modified Rankin Scale with expanded guidance scheme R21-NS106480, U24-NS107230, U24-NS107237 and U24- and interview questionnaire: Interrater agreement and NS107222); Maestro Project and Ibiostroke project funded reproducibility of assessment. Cerebrovasc Dis 2006; 21: by Eranet-Neuron, ISCIII and FEDER; CaNVAS project 271–278. funded by National Institutes of Health; Department of 11. Kwakkel G, Lannin NA, Borschmann K, et al. Biotechnology, Ministry of Science and Technology, India; Standardized measurement of sensorimotor recovery in Skåne University Hospital; Region Skåne; The Swedish stroke trials: Consensus-based core recommendations Heart and Lung Foundation; and The Swedish Research from the stroke recovery and rehabilitation roundtable. Council (2018–02543 and 2019–01757); The Swedish Int J Stroke 2017; 12: 451–461. Government (under the ‘‘Avtal om Läkarutbildning och 12. Majersik JJ, Cole JW, Golledge J, et al. Medicinsk Forskning, ALF’’); Sparbanksstiftelsen Färs och Recommendations from the international stroke genetics Frosta; Freemason’s Lodge of Instruction Eos, Lund; FWO consortium, Part 1: Standardized phenotypic data collec- Flanders (1841918N); US Department of Veteran Affairs, the tion. Stroke 2015; 46: 279–284. American Heart Association (15GPSG23770000, 13. Battey TW, Valant V, Kassis SB, et al. Recommendations 17IBDG3300328). from the international stroke genetics consortium, Part 2: Biological sample collection and storage. Stroke 2015; 46: 285–290. ORCID iDs 14. Boyd LA, Hayward KS, Ward NS, et al. Biomarkers of Arne G Lindgren https://orcid.org/0000-0003-1942-7330 stroke recovery: Consensus-based core recommendations Paul Nyquist https://orcid.org/0000-0001-6078-3543 from the Stroke Recovery and Rehabilitation Vincent Thijs https://orcid.org/0000-0002-6614-8417 Roundtable. Int J Stroke 2017; 12: 480–493. Julie Bernhardt https://orcid.org/0000-0002-2787-8484 15. Torres-Aguila NP, Carrera C, Muino E, et al. Clinical variables and genetic risk factors associated with the Supplementary material acute outcome of ischemic stroke: A systematic review. J Stroke 2019; 21: 276–289. Supplementary material is available for this article online. 16. Jorgensen HS, Nakayama H, Raaschou HO, Vive-Larsen J, Stoier M and Olsen TS. Outcome and time course of References recovery in stroke. Part II: Time course of recovery. The 1. Bernhardt J, Hayward KS, Kwakkel G, et al. Agreed Copenhagen stroke study. Arch Phys Med Rehabil 1995; definitions and a shared vision for new standards in 76: 406–412. stroke recovery research: The Stroke Recovery and 17. Duncan PW, Goldstein LB, Matchar D, Divine GW and Rehabilitation Roundtable taskforce. Int J Stroke 2017; Feussner J. Measurement of motor recovery after stroke. 12: 444–450. Outcome assessment and sample size requirements. 2. Levin MF, Kleim JA and Wolf SL. What do motor Stroke 1992; 23: 1084–1089. ‘‘recovery’’ and ‘‘compensation’’ mean in patients follow- 18. Cortes JC, Goldsmith J, Harran MD, et al. A short and ing stroke? Neurorehabil Neural Repair 2009; 23: 313–319. distinct time window for recovery of arm motor control 3. Lindgren A and Maguire J. Stroke recovery genetics. early after stroke revealed with a global measure of tra- Stroke 2016; 47: 2427–2434. jectory kinematics. Neurorehabil Neural Repair 2017; 31: 4. Nijland RH, van Wegen EE, Harmeling-van der Wel BC 552–560. and Kwakkel G. Accuracy of physical therapists’ early 19. Ballester BR, Maier M, Duff A, et al. A critical time predictions of upper-limb function in hospital stroke window for recovery extends beyond one-year post- units: the EPOS study. Phys Ther 2013; 93: 460–469. stroke. J Neurophysiol 2019; 122: 350–357. 5. Stinear CM. Prediction of motor recovery after stroke: 20. Charlson ME, Pompei P, Ales KL and MacKenzie CR. A advances in biomarkers. Lancet Neurol 2017; 16: new method of classifying prognostic comorbidity in lon- 826–836. gitudinal studies: Development and validation. J Chronic 6. Carrera C, Cullell N, Torres-Aguila N, et al. Validation Dis 1987; 40: 373–383. of a clinical-genetics score to predict hemorrhagic trans- 21. Counsell C, Dennis M and McDowall M. Predicting formations after rtPA. Neurology 2019; 93: e851–e863. functional outcome in acute stroke: comparison of a 7. Pfeiffer D, Chen B, Schlicht K, et al. Genetic imbalance is simple six variable model with other predictive systems associated with functional outcome after ischemic stroke. and informal clinical prediction. J Neurol Neurosurg Stroke 2019; 50: 298–304. Psychiatry 2004; 75: 401–405. International Journal of Stroke, 0(0) International Journal of Stroke, 17(3) 268 Lindgren et al. International Journal of Stroke 17(3)9 22. Adams HP Jr, Bendixen BH, Kappelle LJ, et al. implications for clinical trials. Neurorehabil Neural Classification of subtype of acute ischemic stroke. Repair 2013; 27: 732–741. Definitions for use in a multicenter clinical trial. 36. Cumming TB, Blomstrand C, Bernhardt J and Linden T. TOAST. Trial of ORG 10172 in Acute Stroke The NIH stroke scale can establish cognitive function Treatment. Stroke 1993; 24: 35–41. after stroke. Cerebrovasc Dis 2010; 30: 7–14. 23. Ay H, Benner T, Arsava EM, et al. A computerized algo- 37. Nasreddine ZS, Phillips NA, Bedirian V, et al. The rithm for etiologic classification of ischemic stroke: The Montreal cognitive assessment, MoCA: A brief screening Causative Classification of Stroke system. Stroke 2007; tool for mild cognitive impairment. J Am Geriatr Soc 38: 2979–2984. 2005; 53: 695–699. 24. Lucitti JL, Sealock R, Buckley BK, et al. Variants of Rab 38. Jorm AF. The informant questionnaire on cognitive GTPase-Effector Binding Protein-2 cause variation in the decline in the elderly (IQCODE): a review. Int collateral circulation and severity of stroke. Stroke 2016; Psychogeriatr 2004; 16: 275–293. 47: 3022–3031. 39. Pendlebury ST and Rothwell PM. Prevalence, incidence, 25. Liebeskind DS, Bracard S, Guillemin F, et al. eTICI and factors associated with pre-stroke and post-stroke reperfusion: defining success in endovascular stroke ther- dementia: a systematic review and meta-analysis. Lancet apy. J Neurointerv Surg 2019; 11: 433–438. Neurol 2009; 8: 1006–1018. 26. Brott T, Adams HP Jr, Olinger CP, et al. Measurements 40. Wallace SJ, Worrall L, Rose T, et al. A core outcome set of acute cerebral infarction: A clinical examination scale. for aphasia treatment research: The ROMA consensus Stroke 1989; 20: 864–870. statement. Int J Stroke 2019; 14: 180–185. 27. Heitsch L, Ibanez L, Carrera C, et al. Early neurological 41. Zigmond AS and Snaith RP. The Hospital Anxiety and change after ischemic stroke is associated with 90-day Depression Scale. Acta Psychiatr Scand 1983; 67: outcome. Stroke 2021; 52: 132–141. 361–370. 28. Stinear CM, Byblow WD, Ackerley SJ, Smith M–C, 42. Towfighi A, Ovbiagele B, El Husseini N, et al. Borges VM and Barber PA. REP2: A biomarker-based Poststroke depression: A scientific statement for algorithm for predicting upper limb function after stroke. healthcare professionals from the American Heart Ann Clin Transl Neurol 2017; 4: 811–820. Association/American Stroke Association. Stroke 2017; 29. Burn J, Dennis M, Bamford J, Sandercock P, Wade D 48: e30–e43. and Warlow C. Long-term risk of recurrent stroke after a 43. Weimar C, Kurth T, Kraywinkel K, et al. Assessment of first-ever stroke. The Oxfordshire community stroke pro- functioning and disability after ischemic stroke. Stroke ject. Stroke 1994; 25: 333–337. 2002; 33: 2053–2059. 30. Hardie K, Hankey GJ, Jamrozik K, Broadhurst RJ and 44. Nunn A, Bath PM and Gray LJ. Analysis of the modified Anderson C. Ten-year risk of first recurrent stroke and rankin scale in randomised controlled trials of acute disability after first-ever stroke in the Perth community ischaemic stroke: A systematic review. Stroke Res Treat stroke study. Stroke 2004; 35: 731–735. 2016; 2016: 9482876. 31. Goldstein LB. Common drugs may influence motor 45. Rhemtulla M, Brosseau-Liard PE and Savalei V. When recovery after stroke. The Sygen in acute stroke study can categorical variables be treated as continuous? A investigators. Neurology 1995; 45: 865–871. 32. Cramer SC, Koroshetz WJ and Finklestein SP. The case comparison of robust continuous and categorical SEM for modality-specific outcome measures in clinical trials estimation methods under suboptimal conditions. of stroke recovery-promoting agents. Stroke 2007; 38: Psychol Methods 2012; 17: 354–373. 1393–1395. 46. Salinas J, Sprinkhuizen SM, Ackerson T, et al. An inter- 33. Rathore SS, Hinn AR, Cooper LS, Tyroler HA and national standard set of patient-centered outcome meas- Rosamond WD. Characterization of incident stroke ures after stroke. Stroke 2016; 47: 180–186. signs and symptoms: Findings from the atherosclerosis 47. Liew SL, Zavaliangos-Petropulu A, Jahanshad N, et al. risk in communities study. Stroke 2002; 33: 2718–2721. The ENIGMA stroke recovery working group: Big data 34. Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S and neuroimaging to study brain-behavior relationships after Steglind S. The post-stroke hemiplegic patient. 1. A stroke. Hum Brain Mapp 2020. DOI:10.1002/hbm.25015. method for evaluation of physical performance. Scand J 48. Hope TMH, Friston K, Price CJ, Leff AP, Rotshtein P Rehabil Med 1975; 7: 13–31. and Bowman H. Recovery after stroke: not so propor- 35. See J, Dodakian L, Chou C, et al. A standardized tional after all? Brain 2019; 142: 15–22. approach to the Fugl-Meyer assessment and its International Journal of Stroke, 17(3) International Journal of Stroke, 0(0)