The relationship between post-stroke fatigue and trunk impairment in the subacute phase: a cross-sectional study

Abstract

[Purpose] This study aimed to investigate the correlation between the severity of post-stroke fatigue and trunk impairment, and to identify the associations between specific fatigue characteristics and trunk function. [Participants and Methods] This study included 20 patients with hemiplegia following a first-time stroke who were admitted to a convalescent rehabilitation ward. Fatigue was measured using the Multidimensional Fatigue Inventory, and trunk function was assessed using the Trunk Impairment Scale. Correlation analysis was conducted between the Multidimensional Fatigue Inventory and Trunk Impairment Scale scores. Regression analysis was performed with fatigue characteristics as dependent variables and trunk functions as independent variables. [Results] The prevalence of post-stroke fatigue was 40% (8 patients) at admission. The Multidimensional Fatigue Inventory scores showed a significant negative correlation with Trunk Impairment Scale scores. Regression analyses of the subscales indicated that static sitting balance was significantly associated with reduced activity; dynamic sitting balance was associated with physical fatigue and reduced activity; and coordination was associated with general, physical, and mental fatigue. [Conclusion] Post-stroke fatigue was significantly correlated with trunk impairment, and fatigue characteristics were associated with specific trunk functions. These findings suggest that assessing trunk function may help manage post-stroke fatigue in the subacute phase.

INTRODUCTION

Post-stroke fatigue (PSF) is one of the most common complications of stroke¹⁾, with an estimated prevalence of approximately 50%^{2, 3)}. PSF has been defined as “a feeling of exhaustion, weariness, or lack of energy that can be overwhelming, and which can involve physical, emotional, cognitive, and perceptual contributors, is not relieved by rest, and affects a person’s daily life”⁴⁾. PSF is a significant predictor of dependence in activities of daily living (ADL) and institutionalization⁵⁾. Additionally, it is associated with reduced quality of life and survival rates^{6, 7)}, and it is considered a barrier to stroke rehabilitation⁸⁾.

According to the PSF onset concept model proposed by Wu et al.⁹⁾, PSF occurs within three months after stroke onset and may persist for more than one year. Therefore, early prevention and management of PSF are important^{10, 11)}; however, the physical, mental, and psychological factors associated with PSF are multifaceted^{1, 9, 12)}, and evidence for its prevention and treatment remains insufficient¹³⁾. Research on PSF has been reported less frequently in East Asia than in Europe and the United States³⁾, and cultural differences influence fatigue through psychosocial factors²⁾. Therefore, examining PSF across diverse cultural backgrounds is crucial for a deeper understanding of its global characteristics.

PSF is generally assessed subjectively and objectively using self-report questionnaires and performance tests¹²⁾. In a systematic review, Usman et al. reported that performance measures such as 6-minute walk distance, 10-meter walk speed, lower-limb motor function, 5-times sit-to-stand time, and balance function were associated with PSF¹⁴⁾. However, in patients with subacute stroke for whom standing or walking is difficult, traditional performance tests make it challenging to objectively assess PSF in severe cases.

Trunk function in stroke patients is a fundamental motor skill essential for performing many functional tasks¹⁵⁾. Trunk function refers to the ability to execute selective trunk movements to maintain an upright posture, control weight shifts, and preserve the base of support during static and dynamic postural adjustments¹⁶⁾. Previous studies have reported that trunk function is significantly associated with mobility and balance^{17, 18)} and that it predicts functional outcomes¹⁹⁾. The Trunk Impairment Scale (TIS)²⁰⁾, which assesses trunk function, can be applied to patients in a seated position, even in severe cases where standing and walking are difficult. Moreover, because the TIS does not exhibit a ceiling effect in subacute to chronic stroke patients¹⁸⁾, it is also suitable for evaluating trunk function in mild cases in which patients are able to walk.

Based on these findings, we hypothesized that trunk function—which is important for mobility and ADL^{14,15,16,17,18,19)}, is related to PSF; however, no reports have clarified the association between the Multidimensional Fatigue Inventory (MFI)²¹⁾ and the TIS. Many studies have reported an association between the Fatigue Severity Scale (FSS)²²⁾ and performance tests¹⁴⁾.

The novelty of this study lies in its multidimensional evaluation of fatigue, as PSF is believed to affect not only general fatigue but also mental fatigue and motivation²³⁾, and in its focus on the association with trunk function, which is essential for mobility. Trunk function can be assessed in a seated position, making it applicable even in severe cases in which standing or walking is difficult. Therefore, it is useful for objectively evaluating fatigue in cases where PSF is severe and motor function is impaired¹⁴⁾. Management and prevention of PSF in subacute stroke patients are important for reducing the likelihood that PSF will persist into the chronic phase.

The primary objective of this study was to investigate the correlation between the MFI and TIS in subacute stroke patients with hemiplegia. The secondary objective was to identify the associations between MFI fatigue characteristics and the TIS subscales.

PARTICIPANTS AND METHODS

This was a cross-sectional study conducted at a single institution during hospitalization. The study took place in the convalescent rehabilitation ward (50 beds) of the Akita Prefectural Center for Rehabilitation and Psychiatric Medicine between May 2024 and February 2025, targeting stroke patients admitted during that period.

The inclusion criteria were: (1) diagnosis of cerebral hemorrhage (I61) or cerebral infarction (I63) according to the International Classification of Diseases, 10th Revision, confirmed by computed tomography and/or magnetic resonance imaging; (2) age ≥18 years with hemiplegia due to a first-time stroke; and (3) admission within three months of stroke onset. The exclusion criteria were: (1) severe aphasia; (2) dementia, or a score ≤20 on the revised version of Hasegawa’s Dementia Scale (HDS-R)²⁴⁾; (3) coexisting orthopedic disorders or ataxia affecting postural control; (4) history of mental disorders or current use of antidepressant or antipsychotic medications; (5) unstable physical condition; (6) blindness; (7) transfer to other hospitals; and (8) inability to provide informed consent or refusal to participate.

This study was conducted in accordance with the Declaration of Helsinki and the Ethical Guidelines for Medical and Biological Research Involving Human Subjects in Japan. All participants received both oral and written explanations of the study procedures before enrollment, and each provided written informed consent. The study protocol was approved by the Ethics Review Committee of the Akita Prefectural Center for Rehabilitation and Psychiatric Medicine (Approval Number: 1030).

Demographic and clinical information for the participants included age, sex, height, weight, body mass index (BMI), stroke type, lesioned side, paralyzed side, number of days from stroke onset to admission, and HDS-R score. Functional outcomes included the Japanese version of Functional Independence Measure (FIM), version 3.0^25,26,27,28); the Stroke Impairment Assessment Set–motor item (SIAS-motor)²⁹⁾; the TIS; the Berg Balance Scale (BBS)³⁰⁾; and the Functional Ambulation Category (FAC)³¹⁾.

Psychological outcomes were assessed using the Japanese version of MFI³²⁾ and the Japanese version of the Hospital Anxiety and Depression Scale (HADS)^{33, 34)}. All assessments were conducted within two weeks of hospitalization, and data were collected from electronic medical records.

The FIM was used to assess ADL. The motor item scores range from 13 to 91, the cognitive item scores range from 5 to 35, and the total scores range from 18 to 126. Higher scores indicate greater independence in ADL. The FIM was administered by the physical therapists, occupational therapists, speech therapists, and nurses responsible for the participants.

The SIAS-motor was used to assess the severity of motor impairment. It consists of five tests (knee–mouth, finger function, hip flexion, knee extension, and foot pad). Each item is scored from 0 to 5, yielding a total score range of 0 to 25, with higher scores indicating less severe motor impairment.

The TIS was used to evaluate trunk function. It consists of three subscales: static sitting balance, dynamic sitting balance, and coordination. The score range for each subscale is 0–7, 0–10, and 0–6, respectively, yielding a total score of 0–23, with higher scores indicating better trunk function. The static sitting balance subscale evaluates whether the patient can maintain a sitting position with both feet on the floor. This is followed by passive leg crossing performed by the therapist and then active leg crossing performed by the patient. The dynamic sitting balance subscale evaluates selective lateral flexion ability initiated from the shoulders and pelvic girdle. The coordination subscale measures selective rotational movements of the upper and lower trunk within a 6-second time limit.

The BBS was used to assess balance ability. It consists of 14 items reflecting daily activities. Each item is scored on a scale from 0 to 4, yielding a total score of 0–56, with higher scores indicating better balance ability.

The FAC was used to assess walking ability. It is a 6-point scale ranging from 0 to 5, where 0 indicates inability to walk or walking that requires assistance from two people, and 5 indicates independent walking in all environments, including flat surfaces and stairs. Higher scores indicate better walking ability.

The MFI was used to assess fatigue. It consists of five dimensions: general fatigue, physical fatigue, reduced activity, reduced motivation, and mental fatigue. General fatigue measures overall fatigue; physical fatigue reflects physical sensations related to fatigue; reduced activity assesses daily activity levels; reduced motivation evaluates the level of motivation for daily activities; and mental fatigue measures cognitive symptoms related to concentration. Each dimension is scored on a scale of 4–20, yielding a total score of 20–100, with higher scores indicating greater fatigue. Following previous studies, the cutoff value for PSF was set at general fatigue ≥12^{2, 35, 36)}.

The HADS was used to assess symptoms of anxiety and depression. It includes two subscales—anxiety and depression—each consisting of seven items. The score range is 0–3, with a total score of 0–42, where higher scores indicate greater levels of anxiety or depression.

The SIAS-motor, TIS, BBS, and FAC were measured by the first author in the physical therapy room. The MFI and HADS were administered as self-report questionnaires completed by the participants in their hospital rooms. If participants were unable to read or write clearly, the first author read the questions aloud and recorded their responses. All assessments (SIAS-motor, TIS, BBS, FAC, MFI, and HADS) were conducted on the same day.

Statistical analysis was performed as follows. The Shapiro–Wilk test was used to determine whether numerical data followed a normal distribution. Numerical data were presented as mean ± standard deviation or as median and interquartile range, while categorical data were presented as frequency (%). Correlation analysis between functional outcomes (FIM-motor, SIAS-motor, TIS, BBS, and FAC) and MFI total scores was conducted using Spearman’s rank correlation coefficient. Because significant correlations were observed among functional outcomes in the correlation matrix, partial correlation analysis was conducted between functional and psychosocial outcomes to confirm spurious correlations. Simple linear regression analysis was performed to examine the association between each TIS subscale and the MFI, with MFI fatigue characteristics as dependent variables and TIS subscales as independent variables.

The sample size was calculated using G*Power 3.1.9.7 (Heinrich-Heine-University, Düsseldorf, Germany; freeware) based on the correlation coefficient. With the significance level (α) set at 0.05, statistical power (1–β) at 0.80, and an effect size of −0.62³⁷⁾ derived from a previous study, the minimum sample size was determined to be 18. Assuming that 10% of the participants had missing or incomplete data, the target sample size was set at 20 participants. All statistical analyses were performed using R version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria; freeware), with the significance level set at 5%.

RESULTS

A flow diagram of the 20 stroke patients with hemiplegia who were analyzed is presented in Fig. 1. The demographic and clinical characteristics of the patients at admission are summarized in Table 1. The prevalence of PSF was 40% (8 patients) with MFI general fatigue scores ≥12.

Fig. 1.

Flow diagram of the study.

Table 1. Demographic and clinical characteristics of the patients at admission.

Variables	Score ranges	All patients (n=20)
Age (years)		61.1 ± 10.2
Sex (male/female)		13 (65)/7 (35)
Height (cm)		162.9 ± 9.7
Weight (kg)		65.4 ± 14.4
BMI (kg/m2)		24.6 ± 4.9
Stroke type (hemorrhage/infarction)		11 (55)/9 (45)
Lesioned side (right/left)		11 (55)/9 (45)
Paralyzed side (right/left)		9 (45)/11 (55)
Number of days from stroke onset to admission (days)		36.0 (29.5–41.5)
HDS-R (score)	0–30	28.5 (26.8–29.0)
FIM-total (score)	18–126	110 (93–114)
FIM-motor (score)	13–91	78.5 (62.0–86.3)
FIM-cognitive (score)	5–35	30.5 (27.0–32.3)
SIAS-motor total (score)	0–25	19.5 (12.0–22.5)
TIS-total (score)	0–23	18.5 (13.3–20.0)
TIS-static sitting balance (score)	0–7	7 (7–7)
TIS-dynamic sitting balance (score)	0–10	8.0 (4.8–9.0)
TIS-coordination (score)	0–6	3.5 (1.8–4.0)
BBS (score)	0–56	51 (26–55)
FAC (score)	0–5	4 (3–5)
MFI-total (score)	20–100	53.1 ± 15.5
MFI-general fatigue (score)	4–20	10.3 ± 4.4
MFI-physical fatigue (score)	4–20	11.3 ± 4.4
MFI-reduced activity (score)	4–20	11.6 ± 4.3
MFI-reduced motivation (score)	4–20	9.2 ± 2.5
MFI-mental fatigue (score)	4–20	10.8 ± 3.2
HADS-total (score)	0–42	11.5 ± 7.5
HADS-anxiety (score)	0–21	5.0 ± 3.7
HADS-depression (score)	0–21	6.5 ± 4.2

Values are presented as mean ± standard deviation, median (25th–75th percentiles) or number (%).

BMI: body mass index; HDS-R: revised version of Hasegawa’s dementia scale; FIM: functional independence measure; SIAS: stroke impairment assessment set; TIS: trunk impairment scale; BBS: Berg balance scale; FAC: functional ambulation categories; MFI: multidimensional fatigue inventory; HADS: hospital anxiety and depression scale.

Because we hypothesized that functional outcomes would be significantly correlated, we examined the correlation matrix of FIM-motor, SIAS-motor total, TIS-total, BBS, and FAC. The results showed significant correlations between SIAS-motor total and TIS-total (ρ=0.55, p=0.012) and among all other outcomes (ρ=0.68–0.91, p<0.001) (Table 2). In the simple correlation analysis between functional outcomes and MFI total scores, only TIS-total showed a significant negative correlation (ρ=−0.50, p=0.024). When functional outcomes other than TIS-total were used as control variables in partial correlation analysis, only TIS-total remained significantly negatively correlated with MFI total (ρ=−0.59, p=0.016) (Table 3).

Table 2. Correlations between functional outcomes.

	FIM-motor	SIAS-motor total	TIS-total	BBS
FIM-motor	–	–	–	–
SIAS-motor total	0.74***	–	–	–
TIS-total	0.70***	0.55*	–	–
BBS	0.91***	0.78***	0.74***	–
FAC	0.89***	0.76***	0.68***	0.89***

*p<0.05; **p<0.01; ***p<0.001; values are Spearman’s rank coefficients.

FIM: functional independence measure; SIAS: stroke impairment assessment set; TIS: trunk impairment scale; BBS: Berg balance scale; FAC: functional ambulation categories.

Table 3. Simple and partial correlation analyses between functional outcomes and MFI-total scores.

	Simple correlation coefficients	Partial correlation coefficients
FIM-motor	−0.20	0.06
SIAS-motor total	−0.38	−0.47
TIS-total	−0.50*	−0.59*
BBS	−0.21	0.26
FAC	−0.19	0.13

*p<0.05; **p<0.01; ***p<0.001; Values are Spearman’s rank coefficients.

MFI: multidimensional fatigue inventory; FIM: functional independence measure; SIAS: stroke impairment assessment set; TIS: trunk impairment scale; BBS: Berg balance scale; FAC: functional ambulation categories.

Results of the simple linear regression analysis of TIS subscales on MFI fatigue characteristics showed that static sitting balance was associated with reduced activity (b=−0.69, p<0.001); dynamic sitting balance was associated with physical fatigue (b=−0.51, p=0.023) and reduced activity (b=−0.56, p=0.009); and coordination was associated with general fatigue (b=−0.46, p=0.040), physical fatigue (b=−0.54, p=0.015), and mental fatigue (b=−0.49, p=0.030) (Table 4).

Table 4. Results of the simple linear regression analysis of TIS subscales on MFI fatigue characteristics.

	Static sitting balance			Dynamic sitting balance			Coordination
	b	95% CI	R2	b	95% CI	R2	b	95% CI	R2
General fatigue	−0.10	−0.59 to 0.39	0.01	−0.24	−0.72 to 0.24	0.06	−0.46*	−0.90 to −0.02	0.21
Physical fatigue	−0.33	−0.80 to 0.14	0.11	−0.51*	−0.93 to −0.08	0.26	−0.54*	−0.95 to −0.12	0.29
Reduced activity	−0.69***	−1.05 to −0.34	0.48	−0.56**	−0.97 to −0.15	0.32	−0.35	−0.81 to 0.12	0.12
Reduced motivation	−0.21	−0.69 to 0.28	0.04	−0.20	−0.69 to 0.28	0.04	−0.25	−0.73 to 0.23	0.06
Mental fatigue	−0.30	−0.77 to 0.17	0.09	−0.33	−0.80 to 0.14	0.11	−0.49*	−0.92 to −0.05	0.24

*p<0.05; **p<0.01; ***p<0.001; b: standardized regression coefficient; CI: confidence interval; R²: coefficient of determination; TIS: trunk impairment scale; MFI: multidimensional fatigue inventory.

DISCUSSION

This study is the first to demonstrate a significant association between fatigue and trunk impairment in patients with subacute hemiplegic stroke. The MFI and TIS total scores showed a significant negative correlation not only in simple correlation analysis but also in partial correlation analysis after controlling for FIM-motor, SIAS-motor, BBS, and FAC.

Boz et al.³⁸⁾ reported significant associations between the TIS and forced expiratory volume in one second (FEV₁) as well as forced vital capacity in stroke patients. Oyake et al.³⁹⁾ demonstrated that the severity of PSF is associated with delayed increases in oxygen consumption and cardiac output at the onset of exercise in subacute stroke patients. In summary, the TIS is associated with respiratory function, and impaired cardiopulmonary function is associated with PSF. Furthermore, trunk function in stroke patients is important for controlling distal muscle movement and for performing basic movements such as standing and walking^15,16,17,18), and it is also related to cardiopulmonary function. This interrelationship may explain why the TIS and MFI were found to be correlated.

In contrast to reports from previous meta-analyses and studies¹⁴⁾, no significant correlation was found with functional outcomes other than the TIS. Trunk function plays a central role in maintaining posture and movement efficiency^15,16,17,18), and its impairment may increase energy expenditure and effort during daily activities, leading to greater fatigue. While other functional outcomes, such as the FIM, SIAS-motor, BBS, and FAC, primarily reflect independence levels and mobility and are susceptible to environmental and assistance factors, they may have weakened the direct association with subjective fatigue. Furthermore, many previous studies^{2, 3)} used the FSS, a unidimensional fatigue scale, and focused on patients in the chronic phase (≥6 months after onset). In this study, the MFI was used to assess not only general fatigue but also psychological aspects, including mental fatigue and motivation, of patients in the subacute phase (median 36 days after onset) (Table 1). At this phase, when compensatory strategies and behavioral adaptations were insufficient, trunk impairment may have been associated with fatigue. Therefore, the use of different fatigue scales and differences in disease phase may also have influenced the results of this correlation analysis.

Interestingly, the simple linear regression analysis of TIS subscales on MFI fatigue characteristics showed that different aspects of trunk function were associated with distinct fatigue characteristics. The TIS subscales have a hierarchical structure based on the difficulty of trunk functions⁴⁰⁾. The static sitting balance subscale, which has the lowest level of difficulty, requires the ability to maintain an upright posture. Verheyden et al.¹⁹⁾ reported that static sitting balance in the TIS was an independent predictor of Barthel Index scores six months after stroke. We considered that a decline in static sitting balance affects inactivity in daily life and, through this effect, is associated with reduced activity in the MFI. The dynamic sitting balance subscale requires selective lateral flexion ability of the scapular and pelvic girdles. van Nes et al.⁴¹⁾ reported a significant association between sitting lateral balance and the BBS in subacute stroke patients. Dynamic sitting balance is more challenging than static sitting balance, and we considered that the increased physical load on body functions accompanying functional improvement may explain its association with both reduced activity and physical fatigue in the MFI. The coordination subscale requires selective rotational movements of the scapular and pelvic girdles and the ability to complete the movement within 6 seconds. Muci et al.⁴²⁾ reported an association between walking speed and fatigue during a dual-task test in ambulatory stroke patients. We considered that the coordination subscale was associated with physical fatigue, mental fatigue, and general fatigue scores on the MFI by influencing physical fatigue through rotational movements, mental fatigue through the cognitive load imposed by time constraints, and general fatigue through the overall demands of task performance. Therefore, even mild trunk impairment may be associated with physical and mental fatigue, and recovery of sitting balance in patients with subacute stroke hemiplegia may involve a trade-off with fatigue characteristics. Mutai et al.³⁶⁾ conducted a study using the MFI to evaluate PSF in Japanese patients two weeks after stroke onset and reported that physical fatigue, reduced activity, and mental fatigue were higher than other fatigue characteristics. The results of the present study were similar, suggesting that PSF should be evaluated not only in terms of general fatigue but also in terms of physical fatigue, reduced activity, and mental fatigue.

The clinical application of these results is that PSF can be objectively assessed by combining the MFI, which evaluates multidimensional fatigue, with the TIS, which evaluates trunk function. This approach is applicable not only to patients with mild symptoms but also to those with severe symptoms who have difficulty in standing or walking. This study enriches research on PSF from East Asian and contributes to the development of rehabilitation programs for patients with stroke.

This study has several limitations. First, PSF was assessed using a self-report questionnaire, which, as in many previous studies, excluded patients with severe aphasia. Kao and Chan⁴³⁾ reported that patients with aphasia are at higher risk of developing pathological fatigue and depression than patients without aphasia. Because 46% of those who met the exclusion criteria in this study had severe aphasia (Fig. 1), the generalizability of the findings is limited. Second, although the sample size met the requirements for correlation coefficient tests, the analysis was limited to univariate analysis of the MFI and the TIS. Factors associated with PSF include biological, physical, and psychological factors^{1, 4, 9, 12)}. Biological factors include age, sex, lesion location, and inflammatory biomarkers; physical factors include physical disability, pain, and sleep disorders; psychological factors include anxiety and depression. Therefore, conducting multivariate analyses considering the potential influence of these factors in a larger sample may yield different results. Finally, because this was a cross-sectional study conducted at admission, the causal relationship between fatigue and trunk function remains unclear. Future studies should clarify whether trunk impairment at admission influences fatigue at discharge through longitudinal designs.

In conclusion, this study demonstrated that multidimensional fatigue was significantly correlated with trunk impairment and that specific fatigue characteristics were associated with different components of trunk function, including static sitting balance, dynamic sitting balance, and coordination. These findings suggest that assessing trunk function may be helpful for the management of post-stroke fatigue in the subacute phase.

Funding

This study received no external funding.

Conflict of interest

The authors declare no conflicts of interest.