The relationship between diaphragmatic dysfunction and sleep, emotional state, quality of life, and functional status in patients with chronic ischemic stroke

Abstract

Objectives

To examine whether hemiplegic diaphragm thickening fraction (DTF, %) is associated with functional status, walking capacity, sleep quality, mood, and stroke-specific quality of life (SS-QoL) in chronic ischemic stroke, and to determine a clinically useful DTF threshold.

Materials and Methods

This single-center cross-sectional study included 68 patients with chronic ischemic stroke (≥6 months). Hemiplegic DTF was quantified by ultrasonography. Outcome measures were the Barthel index (BI), Functional Ambulation Category, 6-minute walk test (6-MWT), SS-QoL, Pittsburgh sleep quality index (PSQI), Hospital Anxiety and Depression Scale-Anxiety (HADS-A) and Hospital Anxiety and Depression Scale-Depression (HADS-D) scales, and Fugl-Meyer upper extremity (FMA-UE). Statistical analyses included non-parametric comparisons (DTF ≤20% vs. >20%), Spearman correlations, and multivariable regression adjusted for age, sex, body mass index, and stroke duration. Receiver operating characteristic and logistic regression analyses examined PSQI >5.

Results

Hemiplegic diaphragm dysfunction (DTF ≤20%) was present in 36.8% of participants. Higher DTF was independently associated with better BI, 6-MWT, and SS-QoL scores, and with lower PSQI, HADS-A, and HADS-D scores (all p<0.05), but not with FMA-UE. DTF showed good discrimination for poor sleep [area under the curve: 0.779, 95% confidence interval (CI): 0.649-0.892], with an optimal threshold around 21% (sensitivity 0.77, specificity 0.83). Each 1% increase in DTF reduced the odds of poor sleep by approximately 7% (odds ratio ≈ 0.93; 95% CI: 0.89-0.98).

Conclusion

In chronic ischemic stroke, greater hemiplegic DTF is independently associated with better function, mobility, and quality of life, and fewer sleep and mood problems. A DTF around 21% may serve as a practical threshold for risk stratification and rehabilitation targeting.

Stroke is a leading cause of disability worldwide, resulting in motor deficits and psychosocial impairments that negatively affect quality of life in survivors.^{[1, 2, 3]} In addition to limb motor deficits, respiratory muscle dysfunction—particularly involving the diaphragm—has gained increasing attention in stroke populations.

The diaphragm is the primary muscle of respiration. Any lesion that disrupts neural stimulation, muscle fibers, or the mechanical connection to the thoracic cage may lead to diaphragmatic dysfunction. This condition may manifest as exertional dyspnea and reduced exercise tolerance.^{[2, 4]} Recent studies have demonstrated that the diaphragm muscle also plays a significant role in swallowing, postural stability, balance, regulation of emotional states (emotional autonomy), and modulation of pain threshold.^[5] The diaphragm is prone to atrophy in central nervous system disorders and often remains asymptomatic.^{[6, 7]}

Studies conducted in stroke patients have shown that post-stroke hemiplegic diaphragmatic mobility and thickening fraction are significantly reduced during deep inspiration.^[6] In healthy adults, diaphragm thickness varies according to sex, side, body position, and respiratory phase. In a large ultrasonographic reference study, end-expiratory diaphragm thickness was approximately 2.1±0.4 mm in men and 1.9±0.4 mm in women on the right side, and 2.0±0.4 mm in men and 1.7±0.3 mm in women on the left side. e-IT during quiet breathing was approximately 2.8±0.6 mm in men and 2.5±0.6 mm in women on the right side, and 2.6±0.5 mm in men and 2.3±0.5 mm in women on the left side. During maximal inspiration, diaphragm thickness increases further, reflecting preserved contractile reserve. Therefore, abnormally low end-expiratory thickness (e-ET) may indicate diaphragm atrophy or reduced muscle mass, whereas reduced e-IT or impaired thickening suggests insufficient contractile activation. These abnormalities may decrease ventilatory reserve, impair cough and airway clearance, increase exertional dyspnea, and contribute to sleep-disordered breathing, especially in patients with neurological impairment. In this context, diaphragm thickening fraction (DTF), which reflects the relative increase in diaphragm thickness during inspiration, is considered a functional marker of diaphragmatic contractility. Higher DTF values indicate better diaphragmatic function and are generally associated with improved respiratory performance and functional status, whereas reduced DTF suggests impaired contractile capacity.^[8] A thickening fraction of less than 20% is considered indicative of diaphragmatic dysfunction, a diagnosis that can be reliably established using diaphragmatic ultrasonography. In a cohort of patients with acute ischemic stroke, the incidence of diaphragmatic dysfunction has been reported as 52%.^[9] Ultrasound studies have also demonstrated that reduced diaphragmatic muscle thickness and mobility are positively correlated with motor function of the hemiplegic limb and balance.^[6] Randomized controlled trials have confirmed that inspiratory muscle training enhances respiratory capacity and walking ability by increasing the DTF.^[10]

Diaphragmatic dysfunction may increase the risk of respiratory disturbances during rapid eye movement (REM) sleep by reducing upper airway patency. Positional obstructive sleep apnea has been detected in the majority of patients with unilateral hemidiaphragmatic weakness.^[11] It has also been suggested that diaphragmatic dysfunction may alter both respiratory rhythm and emotional state by influencing the phrenic and vagus nerves through baroreceptor pathways.^[12]

Evidence indicates that one-third of stroke survivors develop major depression, which is associated with increased early mortality. Post-stroke depression typically peaks within the first six months, leading to cognitive decline, delaying rehabilitation, and negatively affecting quality of life. Furthermore, depression has also been linked to an increased risk of recurrent stroke.^{[13, 14]}

There is a bidirectional relationship between stroke and sleep disorders. Respiratory-related sleep disorders, in particular, are among the major risk factors for stroke. In chronic stroke, conditions such as sleep apnea (obstructive/central) and sleep-related hypoventilation have been frequently reported, contributing significantly to impaired quality of life in patients.^{[15, 16]} These disorders may lead to abnormal airway pressures. Post-stroke depression is also closely associated with sleep disturbances.^[16]

In light of this evidence, the multidimensional interactions among diaphragmatic function, respiration, sleep, emotional state, and functionality are gaining increasing importance.

However, the clinical impact of diaphragmatic dysfunction in stroke remains uncertain.

Moreover, studies investigating the long-term outcomes of diaphragmatic dysfunction in chronic patients are quite limited. Our review revealed no comprehensive study investigating the relationship between hemiplegic diaphragmatic function and sleep quality, depression, quality of life, and functionality in patients with chronic stroke.

This study aims to investigate the relationship between hemiplegic diaphragmatic function parameters (ultrasound-based thickness and thickening fraction) and sleep quality, anxiety/depression levels, quality of life, and functional independence in individuals with chronic ischemic stroke.

MATERIALS AND METHODS

Study design

This cross-sectional observational study was conducted between December 2023 and April 2025 at the department of physical medicine and rehabilitation of a university training and research hospital, in accordance with the Declaration of Helsinki. The study was approved by the University of Health Sciences Türkiye, Bursa Yüksek İhtisas Training and Research Hospital Ethics Committee (decision code: 2011-KAEK-25, 2023/11-18, date: 29.11.2023).

Patients

A total of 80 stroke survivors were invited to participate in the study.

Inclusion criteria were as follows: age between 20 and 80 years, stroke duration longer than 6 months, radiologically confirmed diagnosis of ischemic stroke, Montreal Cognitive Assessment (MoCA) score ≥26 for cognitive function, presence of unilateral hemiplegia, and a Functional Ambulation Category (FAC) of at least 1.

Exclusion criteria included: acute or chronic respiratory diseases, restrictive heart disease, history of spinal or abdominal surgery, neuromuscular disorders, restrictive orthopedic limitations, hearing-visual-comprehension impairments, MoCA score <26, facial paralysis, and alcohol use. A total of 68 patients who met the criteria constituted the study population. All participants were informed about the study and provided written informed consent.

The study design is summarized in a flowchart (Figure 1).

Figure 1.

Flowchart.

MoCA: Montreal Cognitive Assessment; FAC: Functional Ambulation Category; DTF: Diaphragmatic thickening fraction.

Assessment parameters

Demographic and clinical data of the patients were obtained from hospital records. Neurological and functional assessments were performed using standard methods.

Ultrasound measurement of diaphragm thickness

Diaphragmatic ultrasonography is a reliable method for detecting diaphragmatic dysfunction.^[17] In this study, all measurements were performed by a physiatrist experienced in ultrasonography, who was blinded to the patients’ clinical data. A high-resolution B-mode ultrasound device (Logic S7 Expert, Korea) was used for the measurements.

Participants were positioned in a semi-supine posture at approximately 45° during the measurement. For imaging, a linear transducer with a frequency range of 3-11 MHz (11L-D, H40432LN) was used. The probe was placed over the “zone of apposition,” where the costal portion of the diaphragm meets the abdominal structures, near the mid-axillary line, usually at the level of the 9^th-10^th intercostal space. On ultrasound, the diaphragm was identified as a hypoechoic band located between two hyperechoic lines corresponding to the pleura and peritoneum.^{[6, 17]} Diaphragm thickness was determined by measuring thickness at the end of maximal inspiration and maximal expiration (Figure 2). The DTF was calculated using the following formula:DTF = (thickness at end-inspiration - thickness at end-expiration) / thickness at end-expiration × 100.

Figure 2.

Representative ultrasonographic images of diaphragm thickness measurement. Diaphragm thickness was measured between the pleural (X) and peritoneal (Y) hyperechoic lines at end-expiration (A) and end-inspiration (B).

During inspiration, the diaphragm contracts and thickens as muscle fibers shorten, whereas during expiration it relaxes and returns to its resting thickness. Therefore, the DTF reflects the contractile activity of the diaphragm and is considered a surrogate marker of diaphragmatic function. Higher DTF values indicate greater contractility, while lower values suggest impaired diaphragmatic performance. In the literature, a DTF value of less than 20% is considered an indicator of diaphragmatic dysfunction.^[9]

Assessment scales

Brunnstrom stages of motor recovery (BMR): A six-stage clinical classification system used to evaluate post-stroke motor recovery by separately scoring the upper extremity, hand, and lower extremity. Higher stages (5-6) indicate advanced motor recovery.^[18]

Functional ambulation categories: A commonly used clinical mobility scale that assesses walking ability and level of independence, scoring from 0 to 5. Higher scores reflect better walking ability and independence.^[19]

Barthel index of activities of daily living (BI-ADL): Assesses the degree of independence in ten basic daily living activities (e.g., feeding, dressing, bathing, mobility). The total score ranges from 0 to 100, with higher scores indicating greater independence.^[20]

Fugl-Meyer assessment of upper extremity motor function (FMA-UE): Evaluates upper extremity motor function in terms of reflexes, movement synergies, coordination, and speed. The total score is given out of 66.^[21]

Pittsburgh sleep quality index (PSQI): Assesses sleep quality over the past month on a scale ranging from 0 to 21. A total score >5 indicates poor sleep quality.^[22]

Hospital Anxiety and Depression Scale (HADS): Evaluates symptoms of anxiety (HADS-A) and depression (HADS-D) using separate subscales, each scored between 0 and 21.^[23]

6-minute walk test: Measures the total distance an individual can walk in six minutes. Higher values indicate better functional capacity.^[24]

Stroke-specific quality of life scale (SS-QoL): A self-reported measure consisting of 12 domains and 49 items, scored between 49 and 245.^[25]

Statistical analysis

Data were analyzed using SPSS version 25.0 (IBM Corp., Armonk, NY, USA) and Python (Pandas, Statsmodels, and SciPy). Categorical variables are reported as counts and percentages. Because the distribution of continuous variables deviated from normality, they are summarized as median (minimum-maximum). Normality was assessed with the Kolmogorov-Smirnov test.

Categorical variables were compared using the chi-square test or Fisher’s exact test, and continuous variables were compared using the Mann-Whitney U test. Diaphragmatic parameters on the hemiplegic and non-hemiplegic sides were analyzed with non-parametric methods. Patients were stratified into two groups according to hemiplegic DTF ≤20% vs. >20%, and subgroup analyses were performed. Associations between variables were assessed with Spearman’s rank correlation coefficient (ρ). Linear regression models were fitted for the BI-ADL, FMA-UE, SS-QoL, FAC, PSQI total score, HADS-A, HADS-D, and the 6-minute walk distance. A two-stage approach was adopted:

Model 1 (univariable): DTF as the sole independent variable.

Model 2 (multivariable): DTF, age, sex, body mass index (BMI, kg/m²) and time since stroke (months); all covariates were entered simultaneously (enter method) based on clinical rationale.

Effect sizes were expressed as unstandardized coefficients (β) with 95% confidence intervals (CI) per 1% increase in DTF. Model fit was summarized by R² (model 1) and adjusted R² (model 2). Assumptions were checked via residual-versus-fitted and component-plus-residual plots; heteroskedasticity was tested with Breusch-Pagan (HC3 robust SEs applied when needed), and multicollinearity was screened (variance inflation factor <5). Poor sleep (PSQI >5) was modeled using logistic regression with DTF (continuous) adjusted for age, sex, BMI, and time since stroke; results are reported as odds ratios (OR) with 95% CIs per 1% DTF. Discrimination was assessed by receiver operating characteristic (ROC) with area under the curve (AUC) and 95% CIs from 1,000-sample bootstrap; the optimal cutoff was chosen by the Youden J index,^[26] and sensitivity and specificity at that threshold were reported with binomial CIs.

Sensitivity analyses: Primary results were reported from model 2, while model 1 estimates were additionally provided to demonstrate consistency in direction and magnitude. The choice of statistical methods was guided by the distributional characteristics of the data, and model assumptions were systematically evaluated to ensure the validity of the analyses. Correlation strength and model estimates were interpreted in accordance with established methodological conventions. Statistical significance was defined as p<0.05.

Sample size

An a priori calculation was conducted in G*Power 3.1, assuming a moderate effect (r=0.40), α=0.05, and power (1-β)=0.80. Under these assumptions, at least 47 participants were required for correlation analyses and 46 for regression analyses. The final cohort of 68 patients exceeded these thresholds, indicating adequate statistical power for the planned analyses.

RESULTS

Of the 80 patients invited, 68 met the eligibility criteria and were included in the analysis. Median age was 66.5 years (range, 23-80) and median BMI was 26.76 kg/m² (20.03-35.27); 55.9% were male. The median time since stroke was 24 months (6-72). Hypertension was the most frequent comorbidity (80.9%). Medians for clinical/demographic variables and scale scores are summarized in Table 1.

TABLE 1. Patient demographics, clinical profile, and assessment characteristics (n=68).

Agea	66.5 (23-80)
Gender, n (%)	-
BMI (kg/m2)	26.76 (20.03-35.27)
Smoking n (%)
Non-smoker	36 (52.94%)
Smoker	22 (32.35%)
Ex-smoker	10 (14.71%)
Dominant hand, right	65 (95.59%)
Diabetes mellitus, n (%) Hypertension n (%) Hyperlipidemia n (%) Coronary artery disease n (%)	28 (41.18%) 55 (80.88%) 10 (14.71%) 17 (25%)
Stroke duration (months)a	24 (6-72)
Stroke side, right, n (%)	39 (57.35%)
BMR-UEa	4 (2-6)
BMR-Ha	4 (2-6)
BMR-LEa	5 (3-6)
FACa	4 (1-5)
BI-ADLa	85 (15-100)
FMA-UEa	49.5 (0-66)
PSQI-comp 1a	1 (0-3)
PSQI-comp 2a	1 (0-3)
PSQI-comp 3a	1 (0-3)
PSQI-comp 4a	1 (0-3)
PSQI-comp 5a	1 (0-3)
PSQI-comp 6a	0 (0-3)
PSQI-comp 7a	0 (0-3)
PSQI totala	4 (0-18)
PSQI >5	26 (38.24%)
HADS-Aa	6 (0-21)
HADS-Da	5 (0-19)
6-MWDa	200 (30-450)
SS-QoLa	130 (59-221)

^a: Median (minimum-maximum); BMI: Body mass index; BMR-UE: Brunnstrom stages of motor recovery-upper extremity; BMR-H: Brunnstrom stages of motor recovery-hand; BMR-LE: Brunnstrom stages of motor recovery-lower extremity; FAC: Functional Ambulation Categories; BI-ADL: Barthel index of activities of daily living; FMA-UE: Fugl-Meyer assessment of upper extremity motor function; PSQI-comp: Pittsburgh Sleep Quality Index component; HADS-A: Hospital Anxiety and Depression Scale-Anxiety; HADS-D: Hospital Anxiety and Depression Scale-Depression; 6-MWD: six-minute walk distance; SS-QoL: Stroke-specific quality of life scale.

Hemiplegic vs. non-hemiplegic

All diaphragmatic indices—end-inspiratory thickness (e-IT), e-ET, and DTF—were lower on the hemiplegic side (p<0.001; Table 2).

TABLE 2. Diaphragm thickness and thickening fraction in hemiplegic versus non-hemiplegic sides in stroke patients (n=68).

-	Hemiplegic side	Non-hemiplegic side	p
e-IT	21.0 (7-43)	31.0 (15-47)	<0.001
e-ET	16.0 (6-28)	20.0 (10-32)	<0.001
DTF	28.0 (7-69)	50.0 (17.0-78.0)	<0.001

e-IT: End-inspiratory diaphragmatic thickness; e-ET: End-expiratory diaphragmatic thickness; DTF: Diaphragmatic thickening fraction.

DTF subgroups (≤20% vs. >20%)

A hemiplegic DTF ≤20% was observed in 36.8% of patients (n=25). This subgroup showed poorer function and walking capacity, with significantly lower BI-ADL, SS-QoL, and 6-minute walk distance (all p<0.001). FAC scores were also lower (p=0.001). All PSQI subscales—except sleep latency—were higher in the DTF ≤20% group.

Poor sleep (PSQI >5) was observed in 76.0% (19/25) of the DTF ≤20% group and 16.3% (7/43) of the DTF >20% group (p<0.001; Table 3).

TABLE 3. Clinical, demographic, and assessment comparisons between patients with hemiplegic DTF ≤20% and >20%.

-	h DTF ≤20% n=25 (36.76%)	h DTF >20% n=43 (63.24%)	p
Agea	68.0 (40.0-79.0)	63.0 (23.0-80.0)	0.249
Gender, n (%)b Male Female	8 (32.0%) 17 (68.0%)	30 (69.8%) 13 (30.2%)	0.006
BMI (kg/m2)a	26.6 (22.5-35.2)	27.1 (20.0-35.3)	0.8636
Smoking n (%)c Non-smoker Smoker Ex-smoker	16 (64.0%) 9 (36.0%) 0 (0.0%)	20 (46.5%) 13 (30.2%) 10 (23.3%)	0.032
Dominant hand, right, n (%)b	24 (96.0%)	41 (95.3%)	1.000
Diabetes mellitus, n (%)b	12 (48.0%)	16 (37.2%)	0.448
Hypertension, n (%)b	22 (88.0%)	33 (76.7%)	0.345
Hyperlipidemia, n (%)b	6 (24.0%)	4 (9.3%)	0.154
Coronary artery disease, n (%)b	8 (32.0%)	9 (20.9%)	0.387
Stroke duration (months)a	15.0 (6.0-60.0)	24.0 (6.0-72.0)	0.042*
BMR-UEa	4.0 (2.0-6.0)	4.0 (2.0-6.0)	0.532
BMR-Ha	4.0 (2.0-6.0)	4.0 (2.0-6.0)	0.886
BMR-LEa	4.0 (3.0-6.0)	5.0 (3.0-6.0)	0.241
FACa	4.0 (1.0-5.0)	4.0 (2.0-5.0)	0.001
BI-ADLa	65.0 (15.0-100.0)	90.0 (40.0-100.0)	<0.001
FMA-UEa	51.0 (0.0-66.0)	49.0 (10.0-66.0)	0.702
PSQI-comp 1a	1.0 (0-3)	0.0 (0-1)	<0.001
PSQI-comp 2a	2.0 (0-3)	1.0 (0-3)	0.071
PSQI-comp 3a	2.0 (0-3)	1.0 (0-3)	<0.001
PSQI-comp 4a	1.0 (0-3)	0.0 (0-3)	<0.001
PSQI-comp 5a	1.0 (0-3)	0.0 (0-1)	0.003
PSQI-comp6a	1.0 (0-3)	0.0 (0-2)	<0.001
PSQI-comp7a	1.0 (0-3)	0.0 (0-2)	<0.001
PSQI totala	9.0 (0.0-18.0)	3.0 (0.0-9.0)	<0.001
PSQI >5b	19 (76%)	7 (16.28%)	<0.001
HADS-Aa	10.0 (0.0-21.0)	4.0 (0.0-19.0)	0.002
HADS-Da	9.0 (0.0-19.0)	4.0 (0.0-16.0)	0.002
6-MWD (meter)a	115.0 (30.0-400.0)	210.0 (120.0-450.0)	<0.001
SS-QoLa	98.0 (59.0-205.0)	144.0 (76.0-221.0)	<0.001

^a: Mann-Whitney U; ^b: Chi-square test; ^c: Fisher’s exact test; h DTF: Hemiplegic-side diaphragmatic thickening fraction; BMR-UE: Brunnstrom stages of motor recovery-upper extremity; BMR-H: Brunnstrom stages of motor recovery-hand; BMR-LE: Brunnstrom stages of motor recovery-lower extremity; FAC: Functional Ambulation Categories; BI-ADL: Barthel index of activities of daily living; FMA-UE: Fugl-Meyer assessment of upper extremity motor function; PSQI: Pittsburgh Sleep Quality Index; HADS-A: Hospital Anxiety and Depression Scale-Anxiety; HADS-D: Hospital Anxiety and Depression Scale-Depression; 6-MWD: six-minute walk distance; SS-QoL: stroke-specific quality of life scale.

Correlations

Hemiplegic DTF correlated positively—strongly with SS-QoL, moderately with 6-minute walk distance and the BI-ADL, and weakly with FAC; and negatively—moderately with total PSQI and weakly with HADS-A, HADS-D, and time since stroke. In addition, the relationship between hemiplegic DTF and Brunnstrom motor recovery stages was examined. Hemiplegic DTF showed weak positive correlations with BMR-UE, BMR-H, and BMR-LE scores; however, these associations did not reach statistical significance (Table 4).

TABLE 4. Correlations between hemiplegic diaphragmatic thickening fraction and assessment parameters.

-	h DTF
-	r	p
Age	-0.150	0.221
BMI (kg/m2)	-0.073	0.555
Stroke duration	0.177	0.149
BMR-UE	0.153	0.214
BMR-H	0.106	0.391
BMR-LE	0.168	0.172
FAC	0.383	0.001
BI-ADL	0.495	<0.001
FMA-UE	0.099	0.420
PSQI total	-0.509	<0.001
HADS-A	-0.325	0.007
HADS-D	-0.342	0.004
6 MWD	0.559	<0.001
SS-QoL	0.601	<0.001

BMI: Body mass index; DTF: Diaphragmatic thickening fraction; BMR-UE: Brunnstrom stages of motor recovery-upper extremity; BMR-H: Brunnstrom stages of motor recovery-hand; BMR-LE: Brunnstrom stages of motor recovery-lower extremity; FAC: Functional Ambulation Categories; BI-ADL: Barthel index of activities of daily living; FMA-UE: Fugl-Meyer assessment of upper extremity motor function; PSQI: Pittsburgh Sleep Quality Index; HADS-A: Hospital Anxiety and Depression Scale-Anxiety; HADS-D: Hospital Anxiety and Depression Scale-Depression; 6-MWD: Six-minute walk distance; SS-QoL: Stroke-specific quality of life scale.

ROC analysis (DTF → PSQI >5)

DTF demonstrated moderate-to-good discrimination of poor sleep: AUC: 0.779 (95% CI, 0.649-0.892). The Youden J-derived optimal cutoff was ≈21%, yielding sensitivity 0.77 and specificity 0.83 (n=68).

Linear regression (DTF → continuous outcomes)

We adopted a two-model strategy: a univariable model including only DTF and a multivariable model adjusting for age, sex, BMI, and months since stroke. Inferences are based on the multivariable model; the univariable estimates were directionally and quantitatively consistent.

Functional and clinical status: In adjusted models, each 1-percentage-point increase in DTF was independently associated with higher BI-ADL scores, a longer 6-minute walk distance, and better SS-QoL (Table 5, model 2). The FAC showed a similarly positive association.

TABLE 5. Effects of diaphragmatic thickening fraction on continuous outcomes: results of model 1 and model 2 presented in a single table.

Outcome	Model 1 β (95% CI)	p	R2	Model 2 β (95% CI)	p	Adjusted R2
FAC	0.029 (0.012-0.046)	0.001	0.149	0.027 (0.008-0.046)	0.006	0.106
BI-ADL	0.620 (0.334-0.907)	<0.001	0.149	0.657 (0.344-0.971)	<0.001	0.206
FMA-UE	0.151 (-0.179–0.482)	0.364	0.013	0.200 (-0.140-0.540)	0.250	0.160
PSQI total	-0.110 (-0.164– -0.056)	<0.001	0.198	-0.114 (-0.173– -0.055)	<0.001	0.208
HADS-A	-0.100 (-0.180– -0.019)	0.016	0.084	-0.113 (-0.202– -0.024)	0.013	0.061
HADS-D	-0.109 (-0.182– -0.037)	0.004	0.122	-0.097 (-0.175– -0.018)	0.017	0.113
6-MWD	3.210 (1.907-4.512)	<0.001	0.268	2.769 (1.368-4.169)	<0.001	0.279
SS-QoL	1.461 (0.920-2.001)	<0.001	0.306	1.521 (0.934-2.107)	<0.001	0.304

Model 1: Univariate linear regression (DTF only).

Model 2: Multivariate linear regression; covariates = age + sex + BMI (kg/m²) + stroke duration (months).

Effect unit: β (per 1% increase in DTF).

DTF: Diaphragmatic thickening fraction; BMI: Body mass index; CI: Confidence interval; BI-ADL: Barthel index of activities of daily living; FMA-UE: Fugl-Meyer assessment of upper extremity motor function; PSQI: Pittsburgh Sleep Quality Index; HADS-A: Hospital Anxiety and Depression Scale-Anxiety; HADS-D: Hospital Anxiety and Depression Scale-Depression; 6-MWD: Six-minute walk distance; SS-QoL: Stroke-specific quality of life scale; β: Regression coefficient; R²: Coefficient of determination.

Sleep and mood: Higher DTF was independently associated with lower PSQI total scores. DTF also showed inverse associations with HADS-A and HADS-D.

Motor recovery: In the adjusted model (model 2), DTF showed no significant association with FMA-UE. Effect estimates with 95% CIs are presented in Table 5 (model 2); univariable summaries are also provided in Table 5.

Logistic regression (PSQI >5)

In the multivariable model adjusting for age, sex, BMI, and months since stroke, each 1-percentage point increase in DTF was associated with ~7% lower odds of poor sleep (OR ≈ 0.93; 95% CI 0.89-0.98; p≈0.007). No other covariate reached statistical significance.

DISCUSSION

This study is the first to examine the relationships between diaphragmatic dysfunction and sleep quality, anxiety/depression levels, and quality of life in individuals with chronic ischemic stroke. In our cohort, hemiplegic diaphragmatic dysfunction (DTF ≤20%) was identified in 36.8% of participants. A higher hemiplegic DTF was associated with better functional status, greater walking capacity, and higher SS-QoL, together with lower severity of sleep disturbance, anxiety, and depression. Patients with diaphragmatic dysfunction showed poorer functional performance and quality of life, and a markedly higher prevalence of poor sleep (PSQI >5).

Reported prevalence of diaphragmatic dysfunction after ischemic stroke varies by phase. Within the first 48 hours, DTF <20% has been described in ~52% of patients;^[9] and ~47% in the early/chronic period (first 6 months).^[6] Prior studies indicate higher rates in the acute and early convalescent stages with partial recovery over time.^{[9, 27]} Our chronic-phase estimate (36.8%) is consistent with the view that dysfunction persists in a subset of patients. In our cohort, dysfunction was more frequent among those with shorter time since stroke; the median time since stroke in the DTF ≤20% group was 15 months, which may partly account for the lower overall prevalence observed. We also found more pronounced dysfunction in women. One study identified female sex as an independent predictor of diaphragmatic dysfunction,^[9] and healthy-population data show that women have a structurally thinner diaphragm.^{[8, 28]} Following stroke, a smaller baseline thickness may hasten depletion of diaphragmatic functional reserve, which could have contributed to the sex pattern and prevalence in our sample.

The diaphragm is not only the principal inspiratory muscle but also a key contributor to trunk stabilization, working in coordinated synergy with the abdominal musculature and the limbs.^{[5, 29]} Imaging studies indicate that greater diaphragmatic thickness and a higher thickening fraction contribute to better postural control. Conversely, reduced thickness and limited excursion on the hemiplegic side are associated with poorer motor and balance performance, as reflected by the Fugl-Meyer assessment.^[6] In our multivariable models, each 1-percentage-point increase in hemiplegic DTF was independently associated with higher Barthel scores, a longer 6-minute walk distance, and better SS-QoL; a similar positive trend was observed for the FAC. The associations between DTF and clinical outcomes remained significant after adjustment for BMI, suggesting that these relationships were independent of body composition. Higher DTF is associated with better daily activities, walking, and functional outcomes. This relationship can be explained by a biomechanical mechanism in which increased diaphragmatic contractility enhances core stability.

No significant association was observed between hemiplegic diaphragmatic function and the FMA-UE. FMA-UE scores were similar between patients with DTF ≤20% and those with DTF >20%. Although a study conducted in the convalescent phase reported small-to-moderate positive correlations between diaphragmatic thickness/excursion and FMA scores,^[6] that work evaluated earlier post-stroke periods and used thickness/mobility metrics rather than a DTF-threshold approach. These differences in disease stage and measurement strategy likely account for the discrepant findings.

Brunnstrom motor recovery stages were also analyzed to further evaluate the relationship between diaphragmatic function and motor recovery. Although hemiplegic DTF showed weak positive correlations with upper extremity, hand, and lower extremity Brunnstrom stages, these associations were not statistically significant. This finding suggests that DTF may be more closely related to functional independence, ambulation, walking endurance, and quality of life than to categorical motor recovery stages. The ordinal and relatively broad structure of the Brunnstrom classification may also be less sensitive to subtle differences in motor performance compared with continuous functional measures such as the BI, 6-minute walk distance, or SS-QoL scores.

Prior work in other conditions suggests that unilateral diaphragmatic dysfunction can promote disordered breathing during REM sleep by altering upper-airway geometry and pleuro-abdominal pressure dynamics.^{[4, 11]} In stroke, sleep disturbances are both a risk factor and a consequence, with sleep-disordered breathing prominently reported and linked to poorer quality of life.^{[15, 16]} In line with this framework, we found higher rates of poor sleep among patients with diaphragmatic dysfunction in the chronic phase, and the hemiplegic DTF correlated inversely with sleep quality. On ROC analysis, DTF showed moderate-to-good discrimination for PSQI >5; the Youden index indicated an approximate threshold of 21%, yielding sensitivity 0.77 and specificity 0.83. The PSQI >5 cut-off has been validated as identifying “poor sleep” with high sensitivity/specificity.^[22] In an adjusted logistic model, each 1-percentage-point increase in DTF was associated with ~7% lower odds of poor sleep. Collectively, these observations are congruent with the proposed pathophysiology whereby unilateral diaphragmatic impairment predisposes to REM-related ventilatory instability via upper-airway and pleuro-abdominal mechanisms.^{[4, 11]}

Our data showed small yet statistically significant inverse associations between hemiplegic DTF and both anxiety and depression. Moreover, patients with diaphragmatic dysfunction had higher anxiety and depression scores than those without dysfunction. Mechanistically, impaired diaphragmatic activity may alter phrenic and vagal signaling via baroreceptor pathways, with downstream effects on respiratory rhythm and affective regulation.^[12] Our findings are consistent with this proposition. Additional support for a respiration-mood axis comes from chronic respiratory disorders such as COPD, in which anxiety and depression are common and the diaphragm exhibits structural-functional alterations.^[30] In stroke populations, the strong links between sleep quality and affective symptoms are well documented.^[31] In the present cohort, the negative association between DTF and PSQI suggests that sleep disturbance could partly mediate the relationship between diaphragmatic dysfunction and mood. Future work should test whether combining respiratory muscle training with sleep-focused interventions can improve emotional outcomes.

An independent, positive association between DTF and the 6-minute walk distance supports the contribution of respiratory muscle function to submaximal exercise performance. Post-stroke respiratory muscle training has improved diaphragmatic parameters and exercise capacity in some trials, yet quality-of-life gains have been inconsistent; this heterogeneity likely reflects differences in intervention type/dose, recovery phase, and coexisting deficits.^{[32, 33, 34]} In our cohort, the independent positive link between DTF and SS-QoL indicates that, in the chronic phase, better diaphragmatic function may translate into greater activity and participation.

The main strength of this study lies in being the first to comprehensively examine the relationship between diaphragmatic dysfunction and sleep quality, anxiety/depression levels, and quality of life in individuals with chronic ischemic stroke. Conducting the study in a single center with a relatively small sample size constitutes one of its limitations. Another limitation of this study is the lack of pulmonary function test or spirometry data, which could have provided additional information on respiratory capacity and its relationship with diaphragmatic function. In addition, habitual physical activity was not assessed in this study. It is recommended that future protocols incorporate objective activity assessments and M-mode ultrasonographic evaluations of diaphragmatic excursion.

In chronic stroke, a higher hemiplegic DTF appears to be associated with better functional status and health-related quality of life, as well as with lower levels of sleep disturbance, anxiety, and depression. A pragmatic cut-point of DTF ≤20% may aid risk stratification and the targeting of individualized rehabilitation goals. Diaphragm-focused assessment and interventions could therefore support multidomain recovery after stroke.