Long‐term cognitive outcomes after decompressive hemicraniectomy for right‐hemisphere large middle cerebral artery ischemic stroke

Abstract

Background and Purpose

Decompressive hemicraniectomy (DH) improves survival and functional outcome in large middle cerebral artery (MCA) infarcts. However, long‐term cognitive outcomes after DH remain underexplored. In a cohort of patients with large right‐hemisphere MCA infarction undergoing DH, we assessed the rates of long‐term cognitive impairment over 3‐year follow‐up.

Methods

We prospectively evaluated consecutive patients included in the Lille Decompressive Surgery Database (May 2005–April 2022) undergoing DH according to existing guidelines for large hemisphere MCA infarction. We included patients with right‐sided stroke and screened with the Mini‐Mental State Examination (MMSE) in at least one of the prespecified follow‐ups (3‐month, 1‐year, 3‐year). Cognitive impairment was defined as an MMSE score < 24. We included only right‐hemisphere strokes to avoid testing biases related to severe aphasia. We compared clinical and neuroimaging data in patients with and without cognitive impairment.

Results

Three hundred four patients underwent DH during the study period. Among 3‐month survivors, 95 had a right‐hemisphere stroke and underwent at least one cognitive screening (median age = 51 years, 56.8% men). Forty‐four patients (46.3%) exhibited cognitive impairment at least once during the 3‐year follow‐up. Baseline characteristics did not significantly differ between patients with and without cognitive impairment. Regarding long‐term temporal trends, cognitive impairment was observed in 23 of 76 (30.3%), 25 of 80 (31.3%), and 19 of 66 (28.8%) patients at 3‐month, 1‐year, and 3‐year follow‐up, respectively, and it was associated with higher rates of functional disability (all p < 0.05).

Conclusions

The persistently high rates of cognitive impairment after DH highlight the importance of cognitive monitoring to improve the long‐term management of survivors.

INTRODUCTION

Large middle cerebral artery (MCA) infarcts represent a critical subset of ischemic strokes characterized by extensive hemispheric brain damage, often leading to a rapid progression of brain edema. Large MCA infarcts—if untreated—result in death in up to 80% of cases, often in the first week after onset [1, 2]. The European guidelines recommend decompressive hemicraniectomy (DH) in patients aged ≤60 years with large MCA infarcts who can be treated within 48 h of stroke onset [3]. Surgical decompression improves poststroke outcomes by decreasing intracranial hypertension and preventing brain herniation [4]. Although aggregate data from several clinical trials carried out in the past 2 decades showed that DH decreases the chance of death and reduces the degree of disability, approximately 40% of survivors experience severe long‐term disability [1, 5, 6]. Although cognitive impairment might contribute to poststroke functional decline [7, 8], there is a scarcity of data regarding the long‐term cognitive trajectories of patients undergoing DH for large MCA ischemic strokes. A retrospective study from 2011 with limited sample size found that 18 of the 20 included DH survivors, followed‐up at least 1 year after stroke, showed some degree of cognitive impairment [9]. In a substudy on 20 patients (13 of them undergoing surgical decompression) with large hemispheric infarction enrolled in the HAMLET trial [10], cognitive testing with Mini‐Mental State Examination (MMSE) was found to be feasible in 12, with half of them exhibiting impaired MMSE scores [11]. Neurological long‐term follow‐up and regular cognitive screening of DH survivors is challenging, due to severe functional impairment and high rates of aphasia, especially in left‐hemisphere strokes.

In a cohort of consecutive patients with large right‐hemisphere MCA ischemic stroke treated with DH, we aimed to assess the rates of long‐term cognitive impairment, clinical–radiological associated factors, and temporal trends over 3‐year follow‐up.

METHODS

Patient selection

We retrospectively analyzed a prospective cohort of consecutive patients undergoing DH for large MCA infarcts from May 2005 to April 2022, included in the Lille Decompressive Surgery Database. The Lille University Hospital serves as the only tertiary care center for a population of 4 million residents.

As part of our in‐house protocol, adults older than 18 years (an upper age limit of 60 years was upheld until the publication of the DESTINY II [Decompressive Surgery for the Treatment of Malignant Infarction of the Middle Cerebral Artery] trial in 2014) [12], with ischemic stroke in MCA territory, with a National Institutes of Health Stroke Scale (NIHSS) score of 15 or higher, within the initial 48 h of stroke onset, and who had an infarct volume larger than 145 mL on diffusion‐weighted images (or an infarct involving at least two thirds of the MCA territory on computed tomography [CT] scan in patients with contraindications for magnetic resonance imaging [MRI]) underwent DH. Patients with life expectancy of less than a few years due to unrelated diseases, and those with signs of irreversible brainstem dysfunction do not undergo DH.

Baseline clinical and radiological data

We gathered information on demographics, vascular risk factors, and previous history of ischemic/hemorrhagic stroke, as detailed previously [13]. Severity of stroke symptoms was evaluated using the NIHSS score. Prestroke disability was defined as a modified Rankin Scale (mRS) score of ≥2 [14]. We documented whether intravenous thrombolysis with recombinant tissue plasminogen activator, mechanical thrombectomy, or both were administered before DH for all included patients. Patients without contraindications underwent brain MRI upon admission using a 1.5‐T device (Achieva; Philips Medical Systems, Best, the Netherlands), including a diffusion‐weighted imaging sequence with specific parameters (repetition time = 3900 ms, echo time = 87 ms, flip angle = 90°, b‐values = 0 and 1000 s/mm², field of view = 230 × 230 mm, matrix = 392 × 216, slice thickness = 5 mm, number of slices = 30, and no interslice gap) [5]. Infarct volume was calculated on b‐1000 diffusion‐weighted images using a semiautomated segmentation tool based on intensity variation and edge detection, utilizing stroke‐dedicated software (Olea Sphere V.3.0; Olea Medical, La Ciotat, France); if MRI was not performed, the infarct volume was calculated using the ABC/2 method on CT scan [15, 16]. The Heidelberg bleeding classification was employed to define hemorrhagic transformation in ischemic strokes before DH, further differentiating hemorrhagic infarction and parenchymal hematoma [17]. We also assessed the presence of additional infarcts in ipsilateral anterior cerebral artery, ipsilateral posterior cerebral artery, posterior fossa, or contralateral territories. All radiological variables were analyzed on the last MRI/CT performed before DH by certified neuroradiologists blinded to clinical data.

Surgical technique

We adopted the DH technique employed in randomized controlled trials, employing a duraplasty after the realization of a large bone flap [10, 18, 19]. The detailed surgical procedure used in our cohort was thoroughly described in a previous study [5].

Follow‐up

According to our in‐house protocol, patients were followed up at 3 months, 1 year, and 3 years. At each follow‐up, we evaluated the functional status with mRS and neurological deficit with NIHSS, and we performed a cognitive screening using MMSE. MMSE was administered by an experienced stroke specialist to all patients who attended the in‐site follow‐up appointment. Being one of the most widely used screening tests for cognitive functions, MMSE is composed of 11 questions, encompassing orientation, attention, memory, visuospatial, and language skills, with scores ranging from 0 to 30 [20]. When patients were not able to attend the outpatient clinic, they were interviewed by telephone to record vital status, and the mRS was assessed with the help of their caregivers, attending physician, or both.

Cognitive outcome definition

We set a cutoff of MMSE < 24 to define cognitive impairment [21, 22, 23]. For the purpose of this study, we analyzed data from patients alive 3 months after DH who underwent at least one cognitive screening using MMSE over the 3‐year follow‐up. Because only a minority of patients with left‐side strokes (15%) were able to undergo cognitive screening, we only included patients with right‐sided strokes, to avoid biases attributable to severe aphasia [24]. For the same reason, we also excluded patients undergoing DH for a right‐hemisphere MCA infarction who also had contralateral territory ischemic involvement.

Statistical analysis

Quantitative variables are presented as median (interquartile range [IQR]), and categorical variables are expressed as frequency (percentage). The primary outcome measure was the rate of cognitive impairment in the study population, defined as an MMSE score < 24 at any of the three follow‐up timepoints, across all participants. To assess the selection bias related to the absence of cognitive screening during follow‐up, baseline clinical and radiological variables were compared between eligible patients with at least one MMSE assessment (included patients, n = 100) versus those without (n = 20) using chi‐squared test or Fisher exact test for categorical variables and Mann–Whitney U‐test for quantitative variables. Among the included patients (n = 100), baseline clinical and radiological variables were also compared according to presence or absence of cognitive impairment during follow‐up using chi‐squared test or Fisher exact test for categorical variables and Mann–Whitney U‐test for quantitative variables. To represent the rate of cognitive impairment over time, we displayed the MMSE scores (categorized as <24, =24, and >24) and the temporal trends of cognitive impairment over the 3‐year follow‐up using a Sankey diagram in patients who underwent MMSE assessment at all three timepoints (SankeyMATIC, https://sankeymatic.com/). The 95% confidence interval (CI) for proportions was obtained using the binomial exact calculation.

At each follow‐up timepoint, we compared the mRS (treated as ordinal variable and as binary variable, defining poor functional outcome as a score ≥ 4) and NIHSS scores according to presence or absence of cognitive impairment using chi‐squared test or Mann–Whitney U‐tests. No statistical comparisons were conducted for categorical variables with fewer than five events. The magnitude of the between‐group differences was assessed by calculating the standardized differences with theirs 95% CIs (Cohen d), calculated on rank‐transformed data for quantitative variables; standardized differences were interpreted as small, medium, and large differences for absolute values of 0.20, 0.50, and 0.80, respectively [25].

We performed all statistical analyses using SPSS software (version 22; IBM, Armonk, NY, USA). Statistical testing was performed with a two‐tailed α level of 0.05. Data were analyzed using the SAS software package, release 9.4 (SAS Institute, Cary, NC, USA).

Ethical approval

The study protocol was regarded as observational by the internal review board of the Lille University Hospital, which granted ethical approval for this study. Patients or their relatives or primary caregiver gave informed consent for follow‐up. The database was declared to the ad hoc commission protecting personal data.

RESULTS

Three hundred four patients were treated with DH for large MCA ischemic stroke consecutively between May 2005 and April 2022. Among them, 56 patients died before the 3‐month follow‐up. Of 248 three‐month survivors, 136 patients had left‐hemisphere strokes or contralateral territory involvement and were therefore not included in this study. Among the 112 patients with right‐hemisphere stroke without contralateral territory involvement, 17 did not undergo MMSE testing at any of the three prespecified timepoints (eight patients did not attend the outpatient follow‐up, six had severe impairment or behavioral manifestations preventing the administration of the questionnaire, three had missing data; see Table S1). Ninety‐five patients received at least one cognitive screening during the 3‐year follow‐up time and were therefore included in the study (Figure 1). Figure S1 displays the number of eligible patients (i.e., survivors who had reached the 3‐month outpatient evaluation) for each timepoint and who underwent cognitive screening during the long‐term follow‐up.

FIGURE 1.

Flowchart of patient inclusion.

Median age of participants was 51 years (IQR = 42–56); 54 (56.8%) were male. Compared to patients with cognitive testing, those who did not undergo cognitive screening were older, more often had diabetes, and had more severe stroke symptoms (i.e., NIHSS score) at onset and before DH (see Table S1).

Median MMSE score was 26 (IQR = 22–28) at 3‐month follow‐up, 26 (IQR = 23–29) at 1‐year follow‐up, and 27 (IQR = 23–29) at 3‐year follow‐up. Forty‐four patients (46.3%, 95% CI = 36.2–56.4) had an MMSE score < 24 at at least one of the three timepoints and were defined as cognitively impaired. Seventy‐six, 80, and 66 patients underwent cognitive screening at 3‐month, 1‐year, and 3‐year follow‐up, respectively. Twenty‐three patients (30.3%, 95% CI = 19.9–40.6) showed cognitive impairment at 3‐month, 25 (31.3%, 95% CI = 21.1–41.5) at 1‐year, and 19 (28.8%, 95% CI = 17.8–39.7) at 3‐year follow‐up (Figure 2).

FIGURE 2.

Rates of cognitive impairment during long‐term follow‐up.

Fifty‐one patients underwent cognitive screening at all three timepoints; the Sankey diagram in Figure 3 shows the number of patients with cognitive impairment at each follow‐up, along with the trajectories of patients over long‐term follow‐up.

FIGURE 3.

Trajectories of patients with cognitive impairment across the three timepoints. This graph shows the 51 patients who underwent cognitive screening at all three timepoints. MMSE, Mini‐Mental State Examination. Made with SankeyMATIC.

We did not observe any statistically significant difference in baseline clinical and radiological characteristics between patients with and without cognitive impairment, despite a slight trend for older age (median 53 vs. 48 years) and numerically lower rates of deep MCA territory involvement (86.4% vs. 96.1%) in patients with cognitive impairment (Table 1).

TABLE 1.

Baseline characteristics of patients with and without cognitive impairment at follow‐up.

	Cognitive impairment, n = 44	No cognitive impairment, n = 51	p	Standardized difference (95% CI)
Demographics and medical history
Age, years	53 (44–57)	48 (41–54)	0.082	0.36 (−0.05 to 0.77)
Male sex	23 (52.3)	31 (60.8)	0.40	−0.17 (−0.57 to 0.23)
Prestroke comorbidities
Hypertension	13 (29.5)	19 (37.3)	0.43	−0.16 (−0.56 to 0.24)
Diabetes	5 (11.4)	3 (5.9)	0.47	0.20 (−0.23 to 0.62)
Dyslipidemia	13 (29.5)	15 (29.4)	0.99	0.003 (−0.41 to 0.41)
Smoking	28 (63.6)	30 (58.8)	0.63	0.10 (−0.31 to 0.51)
Excessive alcohol intake	10 (22.7)	8 (15.7)	0.38	0.18 (−0.24 to 0.60)
Previous stroke or TIA	1 (2.3)	7 (13.7)	0.065	−0.43 (−0.80 to −0.07)
Atrial fibrillation	2 (4.5)	3 (5.9)	1.00	−0.06 (−0.46 to 0.34)
Prestroke disability	2 (4.4)	0 (0.0)	‐	NA
Stroke‐related variables
NIHSS score at hospital presentation [1]	18 (15–19)	17 (15–19)	0.23	0.25 (−0.15 to 0.66)
NIHSS score before hemicraniectomy	18 (17–20)	18 (16–20)	0.30	0.21 (−0.20 to 0.62)
Intravenous thrombolysis	19 (43.2)	22 (43.1)	1.00	0.001 (−0.41 to 0.41)
Mechanical thrombectomy	13 (29.5)	22 (43.1)	0.32	−0.20 (−0.60 to 0.19)
Neuroimaging
Infarct size, mL [1]	199 (160–258)	205 (163–247)	0.89	−0.03 (−0.44 to 0.38)
Superficial MCA territory involvement [2]	44 (100)	49 (100)	‐	NA
Deep MCA territory involvement	38 (86.4)	49 (96.1)	0.14	−0.35 (0.72 to 0.03)
Homolateral ACA territory involvement [1]	13 (29.5)	12 (24.0)	0.54	0.13 (−0.29 to 0.54)
At least one homolateral territory involvement [1] a	14 (31.8)	13 (26.0)	0.53	0.13 (−0.29 to 0.54)
Hemorrhagic transformation	17 (38.6)	23 (45.1)	0.52	−0.13 (−0.53 to 0.27)
Hemorrhagic infarction [2]	13 (30.2)	10 (20.0)	0.25	0.24 (−0.18 to 0.66)
Parenchymal hematoma [2]	5 (11.6)	12 (24.0)	0.12	−0.33 (−0.71 to 0.06)

Note: Standardized differences (Cohen d, calculated on rank‐transformed values for quantitative variables) for cognitive versus no cognitive impairment. Categorial variables are expressed as n (%); continuous variables are expressed as median (interquartile range). Presence of cognitive impairment was defined as Mini‐Mental Status Examination score < 24 at any of the three follow‐up times (3 months, 1 year, 3 years). Numbers in brackets represent missing cases.

Abbreviations: ACA, anterior cerebral artery; CI, confidence interval; MCA, middle cerebral artery; NA, not applicable; NIHSS, National Institutes of Health Stroke Scale; TIA, transient ischemic attack.

^a Other than MCA.

The results of the analyses assessing the associations between cognitive status (i.e., patients with and without cognitive impairment) and neurological symptom severity and functional status at each follow‐up are shown in Table 2.

TABLE 2.

Functional outcomes of patients with and without cognitive impairment at each follow‐up timepoint.

Measure	3‐month follow‐up
	Cognitive impairment, n = 23	No cognitive impairment, n = 53	p	Standardized difference (95% CI)
Modified Rankin Scale, median (IQR)	4 (4–4)	4 (3–4)	0.009	0.70 (0.19 to 1.20)
Modified Rankin Scale = 4 or 5, n (%)	19 (82.6)	30 (56.6)	0.030	0.59 (0.06 to 1.12)
NIHSS score, median (IQR)	14 (11–15)	10 (7–13)	0.003	0.80 (0.29 to 1.31)

	1‐year follow‐up
	Cognitive impairment, n = 25	No cognitive impairment, n = 55	p	Standardized difference (95% CI)
Modified Rankin Scale, median (IQR)	4 (3–4)	3 (3–4)	0.045	0.50 (0.02 to 0.99)
Modified Rankin Scale = 4 or 5, n (%)	14 (56.0)	18 (32.7)	0.049	0.48 (−0.02 to 0.98)
NIHSS score, median (IQR)	12 (6–14)	9 (6–12)	0.18	0.32 (−0.17 to 0.81)

	3‐year follow‐up
	Cognitive impairment, n = 19	No cognitive impairment, n = 47	p	Standardized difference (95% CI)
Modified Rankin Scale, median (IQR)	4 (3–4)	3 (3–4)	0.009	0.77 (0.21 to 1.32)
Modified Rankin Scale ≥ 4, n (%)	10 (52.6)	12 (25.5)	0.035	0.58 (0.01 to 1.15)
NIHSS score, median (IQR)	10 (8–13)	8 (5–11)	0.013	0.72 (0.16 to 1.27)

Note: Standardized differences (Cohen d, calculated on rank‐transformed values for modified Rankin Scale and NIHSS scores) for cognitive versus no cognitive impairment.

Abbreviations: CI, confidence interval; IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale.

DISCUSSION

Among 95 consecutive patients with large right‐hemisphere MCA infarcts treated with DH, 46% exhibited cognitive impairment (i.e., MMSE < 24) at least once during the 3 years following the index stroke. We further assessed the rate of cognitive impairment at 3‐month, 1‐year, and 3‐year timepoints, revealing a stable rate of 1 in 3 patients across the 3‐years follow‐up. Patients with and without cognitive impairment had comparable baseline clinical and neuroimaging characteristics, but cognitive impairment was associated with neurological severity and worse functional outcome during follow‐up.

The current study reports the first systematic evaluation of cognitive function in a large prospective cohort of consecutive patients treated with DH for large MCA infarcts. Cognitive impairment assessment in severe ischemic strokes poses challenges, particularly in cases of left‐hemisphere infarcts, where cognitive screening is limited by aphasia and often infeasible [26, 27, 28]. This challenge is exacerbated by the reliance on verbally oriented tests in clinical settings, contributing to the well‐established link between lower cognitive scores and left‐hemisphere lesions, an association that may be—at least partly—attributed to the confounding factor of language impairment rather than dementia [24]. Given these considerations, in our study involving individuals with extensive anterior‐circulation ischemic strokes, we decided to focus solely on right‐hemisphere infarcts to mitigate potential biases.

In the absence of previous relevant studies focusing on post‐DH cognitive impairment, our results are not easily comparable to previous evidence. One study involving a sample as small as 20 patients undergoing DH for stroke on the nondominant hemisphere showed that almost all tested individuals had some degree of cognitive impairment [9]. In a small study derived from the HAMLET trial involving 20 patients with large hemispheric strokes, half of patients who were able to undergo testing had impaired MMSE scores [11]. The rate of cognitive impairment in our cohort, as assessed by MMSE, was substantial, with 46% of patients exhibiting impairment at least once during the 3‐year follow‐up, and approximately 1 in 3 patients showing impaired MMSE scores at each timepoint. This proportion seems in line with the estimated rate of 34% for incident dementia 1 year following severe (NIHSS > 10) stroke, as reported in a broad population study [29]. However, it is equally noteworthy that more than 1 in 2 patients demonstrated preserved MMSE cognitive screening despite experiencing extensive hemispheric lesions; although our findings suggest some level of cognitive resilience in a portion of patients, it is important to note that the use of MMSE as a screening tool likely underestimated the prevalence of cognitive impairment [30]. Therefore, more comprehensive cognitive assessments are recommended for a clearer understanding of long‐term outcomes.

Regarding long‐term temporal trends, we observed that approximately 1 in 3 patients experienced cognitive impairment at each follow‐up point, spanning up to 3 years after the index event. In our study, among the 51 of 95 patients who underwent cognitive screening at all three timepoints, there was only minimal transition between cognitive states during the extended follow‐up period (Figure 3). Although this slight variability could be attributed to fluctuating symptoms, the rate of cognitive impairment remained consistent across the three timepoints (Figure 2). Although cognitive outcomes have frequently been overlooked in prior studies on DH for large MCA ischemic strokes, our findings highlight the importance of a comprehensive evaluation—inclusive of cognitive testing—during long‐term follow‐up [1]. Identifying cognitive symptoms could aid in establishing a personalized care and rehabilitation plan, thereby improving the quality of life for both patients and caregivers.

Despite a slight trend toward older age and lower rates of deep MCA territory involvement in patients with cognitive impairment, we did not observe statistically significant differences in baseline clinical and radiological characteristics. The homogeneity of our patient population in terms of age, previous medical history, acute stroke characteristics, and treatment standards (a consequence of the recommendations for performing DH) makes it challenging to identify robust predictors for cognitive impairment [3, 31]. To date, current knowledge does not allow the identification of patients at high risk for cognitive decline during acute ischemic stroke phase; further research may yield additional insights into the complex nature of cognitive outcomes following DH for large MCA ischemic stroke.

At each follow‐up, cognitive impairment was associated with neurological severity and worse functional outcome. Given that all enrolled patients shared similar characteristics regarding stroke severity, we might hypothesize that cognitive function plays a pivotal role in shaping the functional status and long‐term outcomes of survivors [8]. This hypothesis suggests that, upon discharge from rehabilitation, cognitive rehabilitation should be encouraged for individuals who have experienced large hemispheric strokes.

The strengths of our study lie in the systematic cognitive screening in a substantial number of patients treated with DH, the standardization of surgical treatment aligning with indications from major trials, a long follow‐up, and comprehensive clinical and radiological patient characterization. However, our study has limitations. It is a retrospective analysis of prospectively collected data conducted at a single center, potentially raising questions about the generalizability of our results. The lack of information on prestroke cognitive status might have, to some extent, influenced the interpretation of the results; nevertheless, in accordance with existing guidelines for performing DH, the patients included were typically young and very unlikely to have prestroke cognitive impairment [3, 6, 10]. The inclusion of patients for this study spanned approximately 20 years; although new effective reperfusion treatments for the acute stroke phase have been approved during this period, the indications for performing DH as well as the surgical techniques did not change substantially. We acknowledge that cognitive impairment was defined based on a cutoff value on MMSE, a screening tool rather than a comprehensive battery. Consequently, it is likely that higher rates of cognitive impairment might have been found if comprehensive testing had been administered. We recognize that the exclusion of a subset of patients who were not assessed at any timepoint may have led to an underestimation of the prevalence of cognitive impairment. However, MMSE is easily employed in clinical practice and serves to identify patients who may benefit from a more in‐depth evaluation. The consistent rate of cognitive impairment across multiple timepoints reassures regarding the validity of our findings. The reliance of MMSE on verbal skills has limited the feasibility of cognitive screening in patients with left‐hemisphere stroke, where severe language impairment is extremely common; therefore, our results are applicable mainly to right‐hemisphere strokes. As cognitive screening in aphasic patients with stroke remains an unresolved challenge, the use of less verbally reliant batteries might enhance the feasibility of cognitive screening in this population [32, 33]. We acknowledge that also in patients with large right hemisphere stroke the risk of bias remains, mainly due to symptoms such as neglect, which might impair the cognitive assessment. We also acknowledge that the exclusion of a minority of patients with right‐hemisphere stroke who did not undergo cognitive screening may have limited the interpretability of our results. Due to our relatively small cohort size, we did not perform multivariate analysis, and therefore results were not adjusted on several baseline characteristics when assessing the association between cognitive impairment and functional status. Although MMSE was used in this study, future research could benefit from employing more comprehensive tools like the Montreal Cognitive Assessment and the Oxford Cognitive Screen to better capture cognitive impairment across a broader range of domains.

In conclusion, cognitive screening is feasible and valuable for patients undergoing DH for right‐hemisphere large MCA ischemic strokes. The notable rate of cognitive impairment—with consistent temporal trends over 3‐year follow‐up—highlights the importance of cognitive assessment to enhance long‐term patient management.

AUTHOR CONTRIBUTIONS

Giuseppe Scopelliti: Conceptualization; investigation; data curation; methodology; visualization; writing – original draft. Hilde Henon: Conceptualization; investigation; writing – original draft; methodology; visualization; writing – review and editing; data curation; supervision; project administration. Olivier Masheka‐Cishesa: Investigation; writing – original draft; writing – review and editing; visualization; data curation. Julien Labreuche: Formal analysis; writing – review and editing; conceptualization. Gregory Kuchcinski: Writing – review and editing; data curation. Rabih Aboukais: Writing – review and editing; data curation. Charlotte Cordonnier: Conceptualization; investigation; methodology; writing – review and editing; supervision; data curation; project administration. Barbara Casolla: Conceptualization; investigation; writing – review and editing; visualization; methodology; data curation; supervision; project administration.

FUNDING INFORMATION

None.

CONFLICT OF INTEREST STATEMENT

C.C. has received fees for advisory boards (Bayer, Biogen, Novartis) and speaker fees (Amgen). B.C. has received speaker fees (Amgen, Sanofi‐Aventis France, Acticor Biotech).

Supporting information

Data S1.

Scopelliti G, Henon H, Masheka‐Cishesa O, et al. Long‐term cognitive outcomes after decompressive hemicraniectomy for right‐hemisphere large middle cerebral artery ischemic stroke. Eur J Neurol. 2025;32:e16492. doi: 10.1111/ene.16492

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.