Association Between Fatigue and Cognitive Impairment at 6 Months in Patients With Ischemic Stroke Treated With Acute Revascularization Therapy

Background: Fatigue is a frequent symptom after stroke. We aimed to determine the association between fatigue and cognitive performance in patients with ischemic stroke who received acute revascularization therapy (IV thrombolysis and/or mechanical thrombectomy). Methods: Seventy patients were prospectively included in the stroke unit of the University Hospital of Dijon, France. A follow-up was performed at 6 months with clinical examination, fatigue assessment by the Fatigue Severity Scale (FSS), and a comprehensive neuropsychological evaluation. Patients with fatigue (FSS score >4) were compared with patients without fatigue. Neuropsychological factors associated with fatigue at 6 months were analyzed using multivariable logistic regression models. Results: Fatigue was reported by 34.3% of patients. Patients with fatigue were older, had more frequent residual handicap, depressive symptoms, and impaired quality of life. They had more frequently low score (<26) on the MoCA scale (79.2 vs. 47.8%, OR = 4.15; 95% CI: 1.32–13, p = 0.015), memory impairment (60 vs. 30.6%, OR = 3.41; 95% CI: 1.09–10.7, p = 0.035), and executive dysfunction (65 vs. 30.8%, OR = 4.18; 95% CI: 1.33–13.1, p = 0.014). In multivariable logistic regression analysis, only memory impairment was independently associated with fatigue (OR = 5.70; 95% CI: 1.09–29.6, p = 0.039). Further analyses restricted to non-depressed patients (n = 58, 84.1%) showed in multivariable models that a score < 26 on MoCA scale (OR 5.12; 95% CI: 1.00–26.2, p = 0.05), and a memory impairment (OR = 6.17; 95% CI: 1.06–35.9, p = 0.043) were associated with fatigue. There was also a non-significant trend toward an association between divided attention deficit and fatigue (OR = 6.79; 95% CI: 0.80–57.6, p = 0.079). Conclusion: The association between fatigue and subtle cognitive impairment including memory or attention deficits could be of interest in elaborating future interventional studies to evaluate the impact of therapeutic strategies, including cognitive rehabilitation, on fatigue.


INTRODUCTION
Ischemic stroke has beneficiated from major therapeutic advances over the past 20 years, leading to improvements in patients' survival and motor outcome. Consequently, several previously neglected post-stroke complications have emerged as important concerns affecting quality of life. Fatigue is a major complaint of patients after ischemic stroke. Fatigue is a subjective, multidimensional experience, with perceptualmotor, emotional and cognitive components, making it difficult to define and measure. Objective fatigue, defined as an observable and measurable decrease in performance that occurs during the repetition of a mental or physical activity, can be distinguished from subjective fatigue, defined as a feeling of exhaustion, weakness, and aversion to effort (1). For Staub and Bogousslavsky, it is a reversible decrease or loss of ability, associated with an intense feeling of physical and/or mental overwork, even in the absence of special effort, and a pronounced feeling of exhaustion that disrupts or makes it impossible to perform routine activities (2).
The prevalence of fatigue after ischemic stroke ranges between 29 and 77% in the literature (3). Several reasons can explain these inconsistent estimates such as the use of various scales for assessment, various time points after stroke for evaluation, differences in population settings, and characteristics of patients including residual motor disability, chronic pain, sleep disorders, and depression, that may influence post-stroke fatigue (4). Fatigue can persist long after the onset of the cerebral infarction and has a negative impact in terms of survival, recovery, and be associated with a profound deterioration in several aspects of daily life (family, social, and professional) (5).
Although it is the subject of a growing literature, the exact mechanisms underlying post-stroke fatigue have not yet been fully elucidated. Some pathophysiological studies carried out on various neurological diseases suggest that primary fatigue could be linked to damage in neural networks, particularly reticular formation and attention network (6)(7)(8)(9). Surprisingly, while cognitive impairment is a major complication after stroke (10), there is few data on its association with fatigue. Indeed, a very limited number of studies focusing on poststroke fatigue incorporated cognitive assessment of patients, and assessments were most often based on global tests or limited to a few areas of cognition, thus making conclusions uncertain (11).
Therefore, the aim of this study was to investigate the prevalence of fatigue 6 months after ischemic stroke and its association with cognitive performances in patients who received acute revascularization therapy.

Study Population
This observational prospective cohort study was conducted at the Stroke Unit of the University Hospital of Dijon, France (Study Registration Number 2017-A01906-47; Clinical Trial Number NCT03288090). This study received approval from the national ethics committee (CPP SUD-EST 4). All patients gave their oral consent to participate according to the French legislation.
Patients ≥18 years old who suffered hemispheric acute ischemic stroke defined according to the WHO criteria and confirmed by imaging (CT-scan or MRI) who received a revascularization therapy (intravenous thrombolysis and/or mechanical thrombectomy) according to current recommendations were included between August 2017 and May 2018. Exclusions criteria were major co-morbidities prior to ischemic stroke that may be responsible for significant fatigue, previous dementia or aphasia, persistence of severe aphasia, or severe disability (modified Rankin scale score > 3) at 6 months preventing a reliable cognitive assessment. Because of the exploratory nature of this study in the absence of previous reliable data in ischemic stroke patients treated with acute revascularization therapy, no sample size calculation was performed.

Data Collected, Patients' Follow-Up, and Outcomes Measured
At inclusion following data were collected: demographics, vascular risk factors including hypertension (high blood pressure recorded in a patient's medical history or patients under antihypertensive treatment), diabetes mellitus (reported in the medical record or patients taking insulin or oral hypoglycemic agents), hypercholesterolemia (reported in the medical history or patients treated with lipid-lowering therapy), current smoking, history of coronary heart disease, atrial fibrillation, excessive alcohol consumption (defined as alcohol intake ≥3 units a day in men and ≥2 in women). Stroke severity at onset was evaluated using the National Institutes of Health Stroke Scale (NIHSS) score.
Patients were evaluated face-to-face by a senior neurologist at 6 months. Clinical examination included evaluation of the NIHSS score, and functional handicap using the modified Rankin Score (mRS). Fatigue was evaluated by the Fatigue Severity Scale (FSS) (12). This scale includes 9 items rated from 1 to 7, assessing physical fatigue, the impact of fatigue on the psychosocial environment and fatigue in general (3 items in each category). Each item is a quote for which patients are asked to judge the application for themselves of "strongly disagree" (1 point) to "strongly agree" (7 points). The total score therefore varies from 9 to 63. The average of the total score is calculated by dividing the total score by the number of items to which the patient has responded. This average defines the FSS score, which ranges from 1 to 7. Patients with a score >4 were considered as having fatigue in accordance with the literature (13). A vertical visual numerical scale for fatigue assessment was also performed in addition to the FSS scale (rated from 0 to 100 on a 10 cm scale). Anxiety and depression were rated using the Hospital Anxiety Depression scale (HAD). A score >7 for either anxiety or depression was retained as significant (14). Global pain and quality of life were rated on a vertical numerical scale (rated from 0 to 100 on a 10 cm scale). A sleep assessment was also performed using the Pittsburgh Sleep Quality Index (PSQI). A total score >5 was considered to reflect sleep disturbances.
A comprehensive neuropsychological assessment was performed by a neuropsychologist according to the GRECoGVASC battery (15,16), a French-language adaptation and standardization of the National Institute of Neurological Disorders and Stroke-Canadian Stroke Network comprehensive battery (17). Several cognitive domains were explored: visuospatial and visuoconstructive skills assessed by the Rey Figure, (21), and the Verbal Fluences (categorical and lexical) (22). To assess divided attention, the computerized double task divided attention test of the Test of Attentional Performance (TAP) was administered to patients (23). Working memory was also tested using the Corsi cubes test and the serial span test taken from the Wechsler Memory Scale (MEM III) (24). Finally, global cognitive function was evaluating by the Montreal Cognitive Assessment scale (MoCA) (25).
The results of the neuropsychological evaluation were dichotomized into two groups (normal or abnormal). The standards chosen were defined according to the GREFEX (French Focus Group on Executive Functions Assessment) standards for the evaluation of executive functions (22), and the GREMEM (French Focus Group on Memory Assessment) standards for the evaluation of memory functions (26). These norms are stratified on age and socio-cultural level. Concerning the interpretation of the tests from the MEM III and WAIS-IV scales, the pathological thresholds were defined on the basis of standard scores, stratified on age. Working memory and/or information processing speed were considered abnormal if a standard score was ≤5 for one of the scores in each event. The pathological threshold for the divided attention test was ≤10th percentile for the number of omissions, as used in clinical practice. Patients were considered to have executive functions impairment if they had an abnormal test in at least one of the following domains: mental flexibility, fluences, working memory, or processing speed.

Statistical Analysis
Categorical variables were presented as frequencies, and quantitative variables were summarized as mean ± standard deviation (SD), or median (interquartile interval, IQR) values. Two groups of patients were considered: patients who had significant fatigue at 6 months (FSS scale >4) vs. patients without fatigue. Characteristics of patients were compared between groups using the Chi-2 test or exact Fisher test for categorical variables, and the Wilcoxon-Mann-Whitney test for quantitative  variables. Logistic regression models were used to evaluate factors associated with fatigue at 6 months. In multivariate analyses, confounding variables with a p < 0.10 were introduced into the final models. A sub-group analysis was performed in non-depressed patients. Statistical analysis was performed with STATA@13 software (StataCorp LP, College Station, Texas, USA).

Prevalence of Fatigue and Patients' Characteristics
Of the 215 patients admitted for acute ischemic stroke who received a revascularization therapy between August 2017 and  May 2018, 106 were initially included. Among these patients, 70 were finally assessed at 6 months (59% men, mean age: 65.7 ± 14.5 years). Reasons for exclusion of patients were withdrawal of consent during follow-up (n = 18), severe residual handicap not compatible with a neuropsychological evaluation (n = 11), stroke recurrence (n = 4), death (n = 2), or loss of follow-up (n = 1) (Figure 1). All patients responded to all items of the FSS. There was a good correlation between FSS score and score on fatigue vertical numerical scale (Spearman coefficient: 0.66, p < 0.001). Twentyfour patients (34.3%) reported significant fatigue (defined as FSS score >4). Characteristics of patients according to fatigue are shown in Table 1 In multivariable analysis, factors associated with significant fatigue at 6 months were a history of hypercholesterolemia, the presence of depressive symptoms on the HAD scale, mRS score ≥2 and, lower score on quality of life scale ( Table 2).

Neuropsychological Evaluation and Fatigue at 6 Months
There was no difference in the proportion of patients who were administered neuropsychological tests according to fatigue at 6 Frontiers in Neurology | www.frontiersin.org   Figure 3 shows the proportion of non-depressed

DISCUSSION
This study demonstrated that fatigue is a common symptom 6 months after ischemic stroke in patients who received acute revascularization therapy. Indeed, nearly a third of patients had significant fatigue. The frequency of fatigue observed in our population is nevertheless rather low when we consider the wide distribution of the prevalence of fatigue in epidemiological studies in which up to 77% of patients may suffer this symptom (4). Since fatigue is correlated with residual motor disability, the observed difference could be explained by the fact that our population was composed of patients with a low residual disability (72.9% of patients had a NIHSS = 0 score and 61.4% had a mRs score ≤1). A study conducted in patients with minor ischemic stroke (initial NIHSS score ≤6 and 6-month mRs score ≤1) found a prevalence of fatigue of 30%, which is consistent with our results (27). To account for the relationship between fatigue and residual handicap, it could be argue that physical activity is more important in patients with a good motor recovery thus improving fatigue through different mechanisms including aerobic training that could strengthen muscle capacity and allow better physical performance by reducing the feeling of fatigue. Physical exercise could also contribute to higher self-esteem and more social interaction (28).
Depression appears to be closely associated with fatigue. In our study 29% of patients with fatigue had depressive symptoms, which is consistent with previous studies (29). Fatigue has long been considered a symptom of post-stroke depression. Although fatigue and depression are closely related, it is now recognized that fatigue can exist independently of depression. Indeed, fatigue after ischemic stroke can occur in the absence of depression, which has also been highlighted in other neurological diseases including Parkinson's disease (30), and multiple sclerosis (31).
The question remains open regarding the causal relationships between depression and fatigue: depression may in some cases be more a consequence than a causal factor of fatigue (32). Although patients with fatigue were older than those without fatigue, we did not find a significant statistical correlation between age and FSS score (Spearman coefficient: 0.11, p = 0.39). In the literature, the association between post-stroke fatigue and age has been inconsistently reported (4).
The presence of cognitive impairment appears to be associated with fatigue. Considering that depression may influence cognitive evaluations, we studied a subgroup of non-depressed patients. As a result, a low score on MoCA scale was associated with fatigue in non-depressed patients. Several studies did not report association between fatigue and cognitive impairment (33). This may be explained by the fact that these studies used the Mini-Mental State Examination (MMSE) score that have a lower sensitivity than the MoCA in assessing post-stroke cognitive impairment (34). Indeed, unlike the MMSE score, the MoCA scale allows an assessment of executive and attentional functions, which are frequently affected in vascular cognitive disorders.
When analyzing specifically cognitive domains, memory impairment was significantly associated with fatigue in both overall and non-depressed patients. One previous study found more frequent memory problems in patients with post-stroke fatigue compared with those without fatigue (35). Conversely, two studies did not conclude to such an association, but neuropsychological tests used for memory assessment differed from our study (27,36). In addition, only minor strokes for one study and both minor strokes and transient ischemic attacks for the other were included. We also found a trend toward an association between fatigue and divided attention impairment in non-depressed patients, although statistical significance was not reached probably due to a lack of power. To the best of our knowledge, one study found an association between sustained attention deficit and fatigue 1 year after ischemic stroke, as well as with executive disorders in a group of non-depressed patients (27). Another study found attention deficit but in a more heterogeneous population that also involved patients with subarachnoid hemorrhage (37). Several studies have shown a correlation between impaired processing speed and fatigue after stroke (35)(36)(37). However, in one study, the deficit was only detected at 3 and did not persist after 6 months, the other two studies did not specify when the neuropsychological assessment was performed during follow-up.
Cognitive impairment has been shown to contribute to fatigue after subarachnoid hemorrhage (38) or traumatic brain injury (39). Our study would reinforce the idea that this may also be the case for ischemic stroke. If cognitive disorders and fatigue appear to be related, it is difficult to specify causal links. Ischemic stroke is an interesting condition for understanding major cognitive functions due to possible focal lesion. Some so-called strategic lesions have helped to establish neuroanatomical correlates of cognitive functions. Regarding fatigue, a pilot study conducted in patients with ischemic stroke who had no significant sequelae or mood disorders found that brainstem and thalamic lesions were more frequent in patients with fatigue (40). The authors suggested that so-called primary fatigue could be the result of minimal attention deficits secondary to an interruption of neural networks, particularly the reticular formation. This hypothesis was not confirmed by a subsequent study conducted in a larger number of patients, in which no association between fatigue and the side or site of the lesion was found (27). Since fatigue is a multidimensional and complex phenomenon, any injury can contribute to increasing the risk of fatigue due to its multiple expression.
Several limitations must be acknowledged. Despite a large screening, the number of patients finally included was limited, thus leading to a lack of power for some multivariable analyses. In a few cases, it was not possible to perform several neuropsychological tests, but there was no difference between patients with and those without fatigue. These limitations underline the fact that neuropsychological assessment after stroke, as well as evaluation of other functional outcomes including fatigue, is difficult and time consuming in routine clinical practice, which could explain the paucity of data in the literature. Finally, it was not possible to assess with the FSS which specific dimension of fatigue the patients complained with. Hence, it would be interesting to evaluate whether cognitive impairment could be more frequently observed in patients with mental fatigue rather than physical fatigue.
To conclude, fatigue after ischemic stroke is a frequent symptom even in patients treated with acute revascularization therapies and good functional recovery. Although it is difficult to establish a causal link, the association between fatigue and subtle cognitive impairment including memory or attention deficits could be of interest in developing further interventional studies so as to define therapeutic strategies to improve these symptoms.

DATA AVAILABILITY
The datasets generated for this study are available on request to the corresponding author.