Cardiorespiratory fitness in stroke patients with different activities of daily living levels: a cross-sectional study

Abstract

Understanding the relationship between cardiorespiratory function (CRF) and activities of daily living (ADL) in stroke patients is essential for improving rehabilitation outcomes. A total of 153 participants were enrolled in this study. Cardiopulmonary exercise was tested at admission. A multivariable linear regression was performed to identify variables associated with peak oxygen uptake (VO_2peak). Participants with low ADL exhibited poorer responses to exercise than those with high ADL levels. After adjusting for confounders, the multiple linear regression analysis showed that albumin-to-globulin ratio (AGR) and left ventricular posterior wall thickness (LVPW) were significantly associated with VO_2peak in all patients. In the low ADL subgroup, the positive association between AGR and VO_2peak was consistent. Conversely, the negative association between LVPW and VO_2peak was uncertain. Otherwise, no significant association were found between AGR, LVPW, and VO_2peak in the high ADL subgroup. This study provides new insights into the relationship between CRF and ADL in stroke patients, with a focus on ARG and LVPW. Future studies with larger sample sizes are needed to further explore the role of AGR and LVPW in improving the CRF in stroke patients.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-97293-9.

Introduction

Cardiorespiratory fitness (CRF) is an important indicator to reflect the ability of the circulatory and respiratory systems to supply oxygen to working muscles¹. High CRF provides protective benefits for the vascular, immune, and nervous systems, and reduces the risk of disease and all-cause mortality, making it a valuable biomarker for aiding clinical decision-making^2–4. Stroke is a leading cause of disability worldwide, frequently leading to a combination of motor impairment, physical inactivity, and systemic complications⁵. CRF of patients with stroke is 25 − 45% lower than that of age-matched healthy individuals⁶. Impaired CRF not only limits daily activities but also hinders the effectiveness of rehabilitation efforts and increases the risk of recurrent strokes and other related health conditions.

Activities of daily living (ADL) impairment is a common consequence of stroke, with approximately 35% of stroke survivors remaining dependent on assistance for daily activities during the first year after stroke⁷. The impact of ADL on CRF is particularly pronounced in clinical populations with physical inactivity and deconditioning⁸. While previous studies have investigated CRF in stroke patients, the relationship between ADL levels and CRF in this population remains insufficiently explored, particularly among those with severe functional limitations^9,10.

Peak oxygen uptake (VO_2peak) is often considered the most practical and effective indicator of CRF in populations with exercise inexperience or lack of motivation¹¹. Cardiac and extracardiac factors, such as age, body composition, cardiac output, left and right ventricular systolic function, complications, and hemoglobin level, are always associated with the VO_2peak level of individuals^12–14. However, no studies have explored how these associated factors differ between patients with varying ADL levels. Identifying the specific factors that influence VO_2peak in stroke patients with different functional limitations is important for designing targeted and effective interventions to enhance CRF.

This study aimed to investigate the relationship between CRF and ADL levels, as well as to explore the factors associated with VO_2peak in stroke patients with different ADL levels. This study is the first to address the gap in understanding how ADL levels influence CRF, with a particular focus on the factors associated with VO_2peak. Our study provides preliminary evidence that can help healthcare providers develop more effective and personalized plans to improve CRF in stroke patients with varying ADL levels.

Methods

Design, setting, and participants

This was a single-center, cross-sectional study. We enrolled hospitalized patients diagnosed with stroke between January 2019 and December 2023. The inclusion criteria were as follows: (1) age ≥ 18 years, (2) being diagnosed with ischemic or hemorrhagic stroke considering the 10th edition of the International Classification of Stroke, (3) being in a medically stable state, and (4) having the ability to fully understand the exercise testing instructions and completed the exercise test. The exclusion criteria were as follows: (1) heart failure and other unstable cardiovascular diseases that required medical treatment or surgery, (2) acute pulmonary embolism, hematological malignancy, or systemic infection, (3) kidney and renal dysfunction, (4) current use of beta-blocker medications, and (5) mental illness, aphasia, or severe cognitive disability. This study was approved by the Ethics Committee of Shenzhen Second People’s Hospital (no.20180926006). All the procedures were conducted according to the Declaration of Helsinki. Participants signed an informed consent form prior to taking part in the study.

Sample size

The sample size required for the multiple regression analysis was calculated using G*Power 3.1 software. Based on findings from a previous study, an estimated effect size of 0.17 was used for the calculation¹⁵. With a type I error rate (α) of 0.05 and a statistical power of 80%, the minimum sample size was determined to be 101 participants in this study.

Data collection

The clinical information of participants included age (< 50 years, 50 − 59 years, and ≥ 60 years), gender, body mass index (BMI) (< 24 kg/m² vs. ≥ 24 kg/m²), stroke type (ischemic or hemorrhagic), duration of stroke (≤ 1 month vs. > 1month), hemiplegic side, ADL level, and comorbidities (hypertension and diabetes mellitus).

The ADL of the participants were evaluated using the Longshi scale. The Longshi scale, a pictorial assessment tool, is convenient for professionals and non-professionals to assess the ADL level of individuals (Supplementary Fig. 1)^16,17. Previous studies have demonstrated the reliability and validity of the Longshi scale^18,19. Based on the ability to transfer out of the bed or to go outside, participants were categorized into bedridden, domestic, and community groups. Each of these groups was evaluated based on three items. The bedridden group was evaluated based on bladder and bowel management, feeding, and entertainment activities. The domestic group was assessed based on their abilities to use the toilet, self-clean, and do housework. The community group was evaluated based on their ability to engage in community activity, shop, and participate in social activities. Each item was scored from 1 to 3, and the total score of the Longshi scale ranged from 3 to 9. Considering the subgroup and total scores, the ADL of participants can be categorized into six levels: 1, totally dependent; 2, severely dependent; 3, moderately dependent; 4, mildly dependent; 5, slightly dependent; and 6, totally independent (Fig. 1). According to the purpose of this study, ADL level was divided into two groups: low ADL (≤ 3 grade) and high ADL (4 − 6 grade) levels.

Fig. 1.

Flow chart for the evaluation of the Longshi Scale.

The laboratory and echocardiographic data of the participants were obtained at admission. Laboratory data included hemoglobin level, albumin level, globulin level, albumin-to-globulin ratio (AGR), serum creatinine level, and urea level. Echocardiographic data included left ventricular ejection fraction (LVEF), left ventricular dimension in systole (LVS), left ventricular dimension in diastole (LVD), left ventricular posterior wall thickness (LVPW), and left ventricular outflow tract (LVOT).

Cardiopulmonary exercise test (CPET), standard and validated assessment tools, is an integrative and comprehensive assessment for CRF, which has been shown to provide reliable measures of VO_2peak²⁰. All assessments were performed by trained professionals following the same standardized procedures. The assessors were blinded to the ADL levels and other specific characteristics to ensure consistency across the data collection process. It was performed on a bicycle ergometer (Ergoline GmbH, ergoselect 200, Germany) as described in a previous study²¹. Participants were asked to wear a mask and breathe through a calibrated volume sensor connected to a metabolic cart. Breath-by-breath measurements of oxygen consumption, carbon dioxide production, and respiratory exchange ratio were measured. The bicycle ergometer and gas analysis system were fully calibrated before the test according to the manufacturer’s instructions. Throughout the test, the 12-lead electrocardiography was recorded continuously, blood pressure was measured at 2-minute intervals, and oxygen saturation was measured at 1-minute intervals. The protocol of CPET began with a rest stage sitting, where participants sat on the bicycle ergometer for 3 min to establish a steady state. This was followed by a 3-minute warm-up stage, during which the participants pedaled without resistance to prepare for the exercise. In the incremental exercise stage, the participants were required to pedal with an increasing workload of 4 − 10 watts per minute, maintaining a cycling speed of 55 − 65 revolutions per minute, and completing the exercise within 8–12 min. Incremental resistance was adjusted based on the participants’ fitness level and clinical condition. The exercise stage was terminated when the participants reached their maximum effort, showed signs of significant distress, exhibited any abnormal physiological responses, or could not maintain the pedaling speed despite strong verbal encouragement. The recovery stage was conducted for 3 min with a resistance of 10 watts. The respiratory exchange ratio (RER)>1.1 was used to assess the “maximal” performance in the exercise test²². The CPET parameters included oxygen uptake at anaerobic threshold, VO_2peak, percentage of predicted maximum oxygen uptake, peak heart rate (HR_peak), percentage of predicted maximum heart rate, metabolic equivalents of task (MET), ventilatory equivalent ratio for carbon dioxide, oxygen uptake efficiency slope (OUES).

Statistical analysis

Data were analyzed using Statistical Package for the Social Sciences V.22(IBM, Armonk, NY, USA). The Kolmogorov-Smirnov test was used to determine the normality of the data distribution. Continuous variables are presented as mean (standard deviation) or median (interquartile range). Categorical variables are expressed as frequencies and percentages. Differences between low and high ADL groups were analyzed using independent t-tests or non-parametric Mann − Whitney U tests. The interaction effect between groups was analyzed using a general linear model. The correlation between VO_2peak and both laboratory and echocardiographic data was tested using Spearman’s rank correlation coefficient. Then, multivariate linear regression models were performed in all stroke patients. The primary outcome measure was VO_2peak, representing the maximum rate of oxygen consumption during CPET. The independent variables included laboratory and echocardiographic variables, which were selected based on their statistically significant association with VO_2peak. Several potential confounders were considered in the analysis, including age, gender, BMI, ADL level, duration after stroke, hypertension and diabetes mellites. These variables were selected based on previous studies and clinical relevance^11,23,24. The model 1 was performed without adjusting for any confounders. The model 2 was performed after adjusting for the age, gender, ADL levels, and duration of stroke. The model 3 was performed after adjusting for age, gender, ADL levels, duration of stroke, BMI, hypertension and diabetes mellitus. Multicollinearity among the independent variables in the models was tested. The value of the variance inflation factors was less than 10, indicating that these variables had no multicollinearity. Furthermore, separate linear regression models were conducted to evaluate whether the association between laboratory and echocardiographic factors and VO_2peak differed between high and low ADL levels. Participants with missing data of interest were excluded from the analysis. A P-value < 0.05 was considered statistically significant.

Results

A total of 153 participants were enrolled in the study. Participants’ demographic information, stroke situation, laboratory data, and echocardiography variables are presented in Tables 1 and 2. Most patients were younger than 60 years old (n = 111, 72.5%), male (n = 129, 84.3%), diagnosed with ischemic stroke (n = 108, 70.6%), and had a stroke duration longer than 1 month (n = 93, 60.8%). 41.2% (n = 63) of the patients exhibited low ADL levels. The prevalence of hypertension and diabetes mellitus among the participants was 79.7% (n = 112) and 31.4% (n = 48), respectively.

Table 1.

Baseline demographic and clinical characteristic of participants.

Variable	Total
Age(years), n (%)
<50	52(34.0)
50–59	59(38.6)
≥ 60	42(27.5)
Gender, n (%)
Male	129(84.3)
Female	24(15.7)
BMI(Kg/m 2 ), n (%)
<24	70(45.8)
≥24	83(54.2)
Type of stroke, n (%)
Ischemic	108(70.6)
Hemorrhage	45(29.4)
Duration of stroke, n (%)
≤ 1 month	60(39.2)
> 1 month	93(60.8)
Hemiplegic side, n (%)
Left	71(46.4)
Right	82(53.6)
Hypertension, n (%)	112(79.7)
Diabetes mellitus, n (%)	48(31.4)
ADL level, n (%)
Low	63(41.2)
High	90(58.8)

BMI, body mass index; ADL, activity of daily living.

Table 2.

Laboratory and echocardiography variables of participants.

Variables	Total
Laboratory variables
Hemoglobin(g/L), Mean (SD)	134.31(15.99)
Albumin(g/L), Mean (SD)	41.87(3.11)
Globulin(g/L), Mean (SD)	25.63(3.58)
AGR, Mean (SD)	1.67(0.28)
Serum creatinine (umol/L), Median (IQR)	69.00(25.05)
Urea(mmol/L), Median (IQR)	5.40(3.36)
Echocardiography variables
LVEF(%),Median (IQR)	67.00(7.00)
LVS(mm), Median (IQR)	29.00(4.00)
LVD(mm), Median (IQR)	46.00(6.00)
LVPW(mm), Median (IQR)	10.00(2.00)
LVOT, Median (IQR)	23.00(4.00)

ADL, activity of daily living; AGR, the ratio of albumin to globulin; LVEF, left ventricular ejection fraction; LVS, left ventricular systolic dimension; LVD, left ventricular diastolic dimension; LVPW, left ventricular posterior wall thickness; LVOT, left ventricular outflow tract.

The results of CPET are listed in Table 3. The mean VO_2peak achieved was 16.54 mL/kg/min, and the mean HR_peak was 134 beats per minute (bpm). Participants with low ADL levels showed a poorer response to exercise compared to those with high ADL levels in terms of VO_2peak (14.90 ± 3.07 mL/kg/min vs. 17.69 ± 3.75 mL/kg/min, P < 0.001), HR_peak (128.53 ± 20.24 bpm vs. 137.92 ± 24.34 bpm, P = 0.014), MET (4.29 ± 0. 89 vs. 5.03 ± 1.06, P < 0.001), and OUES (1219.24 ± 303.45 ml/logL vs. 1447.46 ± 311.53 ml/logL, P < 0.001).

Table 3.

The parameters of cardiopulmonary exercise test.

Variable	Total	Low ADL level	High ADL level	P
VO2AT(mL/ kg / min), Mean (SD)	10.67(2.45)	9.94(2.27)	11.17(2.45)	0.002
VO2peak (mL/min/kg), Mean (SD)	16.54(3.74)	14.90(3.07)	17.69(3.75)	< 0.001
%VO2max, Median (IQR)	58.06(18.62)	54.33(20.32)	59.53(17.64)	0.048
O2/HR (ml), Median (IQR)	8.51(1.96)	8.06(1.98)	8.83(1.88)	0.016
HRpeak,(bpm) Mean (SD)	134.09(23.08)	128.53(20.24)	137.92(24.34)	0.014
% HRmax, Mean (SD)	81.51(13.18)	79.17(11.74)	83.15(13.93)	0.066
MET, Mean (SD)	4.72(1.06)	4.29(0.89)	5.03(1.06)	< 0.001
VE/VCO2 (L/L), Median (IQR)	31.02(5.92)	31.37(6.42)	30.92(4.81)	0.236
OUES (ml/logL), Mean (SD)	1353.49(327.24)	1219.24(303.45)	1447.46(311.53)	< 0.001

VO_2AT, oxygen uptake at anaerobic threshold; VO_2peak, peak oxygen uptake; %pp VO2, percentage of maximum oxygen uptake predicted; HR_peak, peak heart rate; % maximum HR, percentage of maximum predicted heart rate; METs, metabolic equivalents of task; VE/VCO2, ventilatory equivalent ratio for carbon dioxide; OUES, oxygen uptake efficiency slope; and bpm, beats per minute.

The interaction analysis of ADL level and other factors in determining VO_2peak is presented in Table 4. No significant interaction effect was found between ADL level and age (P = 0.395), gender (P = 0.699), BMI (P = 0.693), duration of stroke (P = 0.534), hypertension(P = 0.172), or diabetes mellitus (P = 0.059).

Table 4.

Two-way ANOVA analysis for the interaction of ADL levels and other factors in determining VO_2peak.

Subgroups	VO2peak (mL/ kg / min)		β(95%CI)		P	P for Interaction
	Low ADL level	High ADL level
All	14.90(3.07)	17.69(3.75)	2.79(1.66, 3.93)		< 0.001
Age(years)						0.395
<50	14.28(2.95)	18.18(4.16)	3.90 (1.59, 6.22)		0.001
50–59	15.83(3.21)	18.15(3.20)	2.32(0.64, 4.00)		0.008
≥ 60	14.14(2.76)	16.22(3.55)	2.08(0.08, 4.08)		0.042
Gender						0.699
Male	15.09(3.15)	17.88(3.89)	2.79(1.49, 4.08)		< 0.001
Female	14.16(2.74)	16.35(2.30)	2.18(0.02, 4.35)		0.048
BMI(kg/m 2 )						0.693
< 24	15.29(2.86)	17.85(3.59)	2.56(0.97, 4.15)		0.002
≥ 24	14.54(3.25)	17.56(3.91)	3.02(1.39, 4.65)		< 0.001
Duration of stroke						0.534
≤ 1 month	14.84(2.47)	18.12(3.93)	3.28(1.61, 4.96)		< 0.001
> 1 month	14.96(3.63)	17.50(3.68)	2.54(0.94, 4.14)		0.002
Hypertension						0.172
No	16.33(2.64)	17.54(3.78)	1.20(-1.37, 3.76)		0.344
Yes	14.56(3.09)	17.73(3.77)	3.17(1.90, 4.44)		< 0.001
Diabetes mellitus						0.059
No	15.49(3.13)	17.60(3.84)	2.10(0.72, 3.50)		0.003
Yes	13.42(2.40)	17.88(3.64)	4.46(2.52, 6.40)		< 0.001

VO_2peak, peak oxygen uptake; ADL, activity of daily living; and BMI, body mass index.

The correlations between VO_2peak and laboratory and echocardiographic variables are presented in Supplementary Figs. 2 and 3, AGR (P = 0.003) and LVPW (P = 0.009) were found to be associated with VO_2peak in patients with stroke. Multiple linear regression analysis was further performed to identify the factors associated with VO_2peak. After adjusting for confounders of age, gender, ADL level, duration of stroke, BMI, hypertension and diabetes mellitus, AGR (β = 2.63, 95%CI: 0.64 − 4.61, P = 0.010), and LVPW (β=-0.47, 95%CI:-0.88−-0.06, P = 0.024) were significantly associated with VO_2peak in all patients with stroke (Table 5). Furthermore, multiple linear regression analysis was performed to identify factors associated with VO_2peak in the two ADL groups separately. In the low ADL group, AGR was positively associated with VO_2peak across all three regression models. Conversely, LVPW exhibited a negative association with VO_2peak in both in model 1 and model 2, but no significant association was observed in the model 3. Otherwise, no significant association were found between AGR, LVPW, and VO2peak in the high ADL group (Table 6).

Table 5.

Factors associated with VO_2peak in all stroke patients.

Variables	Model 1		Model 2		Model 3
	β(95%CI)	P	β(95%CI)	P	β(95%CI)	P
AGR	3.51(1.49–5.52)	0.001	2.62(0.66, 4.59)	0.009	2.63(0.64, 4.61)	0.010
LVPW	-0.62(-1.02, -0.22)	0.002	-0.51(-0.89, -0.13)	0.009	-0.47(-0.88, -0.06)	0.024

Multiple linear regression analysis was performed. Model 1 was performed without adjusting for any confounders. Model 2 was performed after adjusting for age, gender, ADL level, and duration of stroke. Model 3 was performed after adjusting for age, gender, body mass index, ADL level, duration after stroke, hypertension and diabetes mellitus. AGR, the ratio of albumin to globulin; LVPW, left ventricular posterior wall thickness.

Table 6.

Factors associated with VO_2peak in stroke patients with different ADL levels.

Variables	Model 1		Model 2		Model 3
	β(95%CI)	P	β(95%CI)	P	β(95%CI)	P
Low ADL levels
AGR	2.92(0.37, 5.47)	0.026	2.96(0.15, 5.77)	0.040	3.04(0.21, 5.86)	0.036
LVPW	-0.51(-1.00, -0.02)	0.040	-0.51(-1.02, -0.00)	0.048	-0.39(-0.94, 0.17)	0.172
High ADL levels
AGR	2.75(-0.12, 5.62)	0.060	2.62(-0.21, 5.44)	0.069	2.73(-0.16, 5.62)	0.064
LVPW	-0.46(-1.04, 0.12)	0.117	-0.53(-1.11, 0.05)	0.071	-0.50(-1.11, 0.11)	0.105

Multiple linear regression analysis was performed. Model 1 was performed without adjusting for any confounders. Model 2 was performed after adjusting for age, gender and duration of stroke. Model 3 was performed after adjusting for age, gender, duration of stroke, body mass index, hypertension and diabetes mellitus. ADL, activity of daily living; AGR, the ratio of albumin to globulin; LVPW, left ventricular posterior wall thickness.

Discussion

The present study uniquely explores the relationship between CRF and ADL in stroke patients, and found that patients with low ADL levels exhibit lower CRF. Our study fills a critical gap by focusing on CRF in stroke patients with significant ADL limitations, a subgroup often underrepresented. Notably, the present study is among the first to identify AGR and LVPW as significant factors associated with VO_2peak in stroke patients, particularly emphasizing those with low ADL levels. It provides a novel perspective on how physiological and functional factors impact CRF. These findings highlight the need for tailored interventions to improve CRF for stroke patients and lay the groundwork for future research in this area.

Our study found that the CRF was lower in stroke patients with low ADL levels than in those with high ADL levels. Importantly, this relationship remained consistent and independent of other potential confounding factors, demonstrating the stability of the association between functional independence and aerobic capacity in stroke patients. A previous study has reported that the level of N-terminal pro-B-type natriuretic peptide was higher in bedridden stroke patients than in non-bedridden stroke patients, suggesting that lower ADL level in stroke patients is associated with worse cardiac function²⁵. Individuals with low ADL levels often experience severe physical and mental disabilities, which may further accelerate the decline in exercise capacity^26–28. These results emphasize the important role of ADL as a key determinant of CRF among stroke patients, independent of other clinical or demographic variables.

Several factors influence VO_2peak in individuals. However, our study found no correlation between VO_2peak and hemoglobin, which contrasts with a previous study suggesting that the hemoglobin levels, including hemoglobin mass and hemoglobin concentration, are positively correlated with VO_2peak²⁹. It has been reported that increased blood viscosity and greater resistance in skeletal muscle, which are common in patients with stroke, may disrupt the positive relationship between VO_2peak and hemoglobin²⁹. This may explain why no significant association between VO_2peak and hemoglobin was observed in our study. Additional studies are required to further explore the relationship and underlying mechanisms linking VO_2peak and hemoglobin in this population.

AGR, considering the characteristics of both albumin and globulin, is a key indicator of the body’s nutritional and inflammatory status³⁰. Our study found a positive association between AGR and VO_2peak in patients with stroke. Higher AGR levels indicate better muscle mass and lower inflammation, both of which are associated with improved muscle strength and efficiency, as well as the ability to utilize oxygen effectively during physical activity^31,32. This, in turn, may contribute to better exercise performance and higher VO_2max. In the low ADL group, AGR showed a consistent and robust association with VO_2peak, demonstrating that AGR may be a critical independent predictor of aerobic capacity in patients with lower functional abilities. ADL limitations are closely linked to low albumin levels and reduced AGR^33–35. Nutritional deficiencies and increased inflammation may exacerbate muscle weakness and impaired oxygen utilization, leading to low VO_2peak. Our study offers preliminary evidence for the relationship between AGR and VO₂_peak, suggesting that AGR may serve as a useful biomarker of CRF in stroke patients, particularly those with significant ADL limitations.

Our study suggests that LVPW is negatively correlated with VO_2peak in patients with stroke. LVPW is an important indicator of left ventricular hypertrophy, which is often observed in individuals with functional decline³⁶. Increased LVPW can reduce the diastolic filling capacity and impair the systolic function of the heart, thereby reducing optimal blood flow and oxygen delivery to the muscles, which may contribute to a decrease in VO_2max^36,37. In the low ADL group, the association between LVPW and VO_2peak was inconsistent after adjusting for different confounders. This variability may result from the association between LVPW and VO_2peak being mediated or influenced by other factors, such as comorbidities^38,39. Additionally, the small sample size in the low ADL group had limitations of statistical power to detect a significant association after adjusting more confounders. Thus, larger studies are needed to explore the relationship between LVPW and VO_2peak in stroke patients with significant functional limitations.

Interesting, neither ARG nor LVPW showed significant correlations with VO_2peak in stroke patients with high ADL levels. This suggests that nutritional status, inflammation, and cardiac function may not be limiting factors in the patients with high ADL levels. The reason for this could be that high functional independence allow patients to rely more on muscle strength, neurological recovery and neuromuscular efficiency, reducing the impact of physiologic and cardiac related limitations^40,41. It indicates that the determinants of aerobic capacity may differ between patients with varying levels of ADL, highlighting the importance of tailored interventions to improve CRF based on functional status.

This study had several limitations. First, due to the small sample sizes in subgroups of ADL, the results should be interpreted with caution. Further studies with larger sample sizes are necessary to confirm these findings. Second, despite adjusting for several confounders, other confounding variables, such as stroke severity, medication use, and psychological factors, may influence both ADL levels and CRF. Further studies are required to explore the effects of more factors on the relationship between ADL levels and CRF in patients with stroke. Third, the generalizability of our findings may be limited due to the specific sample used in this study. The participants were recruited from a single center, which may not fully represent the diversity of the broader stroke patient population. Finally, as this was a cross-sectional study, it offered only an initial analysis of the findings. Additional studies are required to establish definitive evidence.

Conclusion

This study found that stroke patients in low ADL group exhibit significantly lower CRF. AGR and LVPW were associated with VO_2peak in stroke patients. Specifically, AGR demonstrated a consistent association with VO_2peak in low ADL group, while the relationship between LVPW and VO_2peak remains uncertain. These results highlight the importance of addressing both physiological and functional factors in post-stroke rehabilitation. Future studies with larger sample sizes are needed to further explore the role of AGR and LVPW in improving the CRF and rehabilitation outcomes in stroke patients.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Yulong Wang conceived the study, Mingchao Zhou, Jianjun Long, and Fubing Zha conducted and analyzed the data, Dongxia Li drafted the manuscript.

Funding

The study was supported by the Sanming Project of Medicine in Shenzhen (No.SZSM202111010); National Natural Science Foundation of China(No.82205065).

Data availability

The datasets during the current study are not available publicly, but are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical considerations

This study was approved by the Ethics Committee of Shenzhen Second People’s Hospital (no.20180926006). All the procedures were performed considering the Declaration of Helsinki.