Age better discriminates early dysphagia than cough function, respiratory muscle strength, or trunk control after mild stroke

Abstract

This retrospective study aimed to identify factors that best discriminate the presence of early dysphagia after mild stroke. We performed a retrospective analysis of prospectively collected routine-care data from consecutive first-ever subacute stroke patients. We included 140 first-ever subacute stroke patients greater than 18 years or older, within 3 weeks of stroke onset, with hemiparesis and a Mini-Mental State Examination score greater than 24, recruited from an inpatient rehabilitation unit. Dysphagia was assessed using the Volume-Viscosity Swallow Test and videofluoroscopy, with swallowing function evaluated by the Penetration-Aspiration Scale. The outcome measures included maximal inspiratory and expiratory muscle pressures (MIP and MEP) for respiratory muscle strength, peak expiratory cough flow (PECF) for cough function, and the Trunk Control Test for trunk control. We found respiratory muscle weakness and impaired cough function (MIP 55.5 ± 23.4; MEP 84.5 ± 33.9 cmH₂O; PECF 244.9 ± 87.5 L/min); however, no deficits in trunk control were observed. Dysphagia was present in 30 patients (26.3%), who were older and had lower PECF (215.0 ± 85.8 vs. 255.6 ± 90.3 L/min; P = 0.02). In multivariable analysis, only age predicted dysphagia [odds ratio 1.09, 95% confidence interval (CI): 1.03–1.16; P = 0.03]; receiver operating characteristic area under the curve was 0.717 (95% CI: 0.613–0.821; P < 0.001) with a cutoff of 64.5 years. In early poststroke patients, dysphagia was primarily associated with older age, whereas respiratory muscle strength, cough function, and trunk control – although frequently impaired – were not independent predictors. An age threshold of ~65 years best discriminated between patients with and without dysphagia.

Introduction

Poststroke dysphagia is an independent risk factor for poor functional recovery [1,2]. Therefore, recognition and treatment of dysphagia after stroke are of great significance to the prognosis of stroke patients. Early after admission to the acute hospital, stroke patients should receive a formal dysphagia screening, including a bedside water swallowing test or a multi-consistency test, followed by instrumental assessment when indicated [3]. Beyond that, several studies have also identified clinical factors associated with dysphagia, including advanced age, greater stroke severity, and comorbidities [4,5].

Reduced cough strength has also been linked to swallowing difficulties; indeed, compared with non-dysphagic patients, those with dysphagia show lower values for cough flow [6]. Impairments in cough function in stroke patients may, at least in part, be explained by respiratory muscle weakness [7]. An effective cough needs powerful and coordinated contraction of expiratory (abdominal and intercostal) muscles to generate high intrathoracic pressure and high peak cough flow; consequently, cough effectiveness is closely related to respiratory muscle strength.

Respiratory muscles also contribute significantly to trunk control [8]. Following a stroke, most patients experience impairments in trunk control, which are associated with reduced strength of the core and expiratory muscles, mobility limitations, and respiratory dysfunctions [9–11]. Despite the known interdependence between cough function, respiratory muscle strength, and trunk control in stroke, their combined role in the early detection of dysphagia remains poorly understood.

Cough, respiratory muscles, and trunk control are three clinical features that are closely linked and could play a crucial role in the early detection of dysphagia. In this vein, the respiratory muscles have a central role; they contribute to cough effectiveness and play a role in stabilizing the trunk [12]. Therefore, the study of respiratory muscles in the context of stroke is of particular interest, since these patients’ present alterations in muscle structure leading to decreases in respiratory muscle strength, ventilatory disturbances, decreased cough ability, and decreased trunk control [13,14]. Therefore, understanding the role of respiratory muscle strength in dysphagia screening may clarify whether assessing respiratory muscle strength should be routinely considered. Additionally, the voluntary cough assessment would also add valuable information during the dysphagia screening, helping to identify stroke patients at increased risk of aspiration and serving as a useful adjunct to standard dysphagia screening tests.

The relationship between cough function, respiratory muscle strength, and trunk control in relation to dysphagia needs to be explored through further research to understand its clinical applicability. Therefore, this study aims to identify factors that best discriminate against the presence of early dysphagia after mild stroke.

Methods

Design

This observational study retrospectively analyzed prospectively collected routine-care data from consecutive subacute stroke patients admitted to an inpatient neurorehabilitation facility at a tertiary hospital in Barcelona, Spain (2015–2018). All assessments were part of standardized clinical care, with no additional research procedures. Baseline assessments included the evaluations of dysphagia, cough function, respiratory muscle strength, and trunk control. This study complied with the Declaration of Helsinki and with the Spanish data protection laws (Organic Law 3/2018, December 5). Upon admission in our unit, patients provided written consent for routine clinical care, as approved by the Clinical Ethics Committee of the Institut Hospital del Mar d’Investigacions Mèdiques. This document relates solely to routine clinical management. This study has been reported following the STROBE Statement [15]. AI tools were used for language editing and clarity. The authors reviewed and approved the final content and take full responsibility for the manuscript.

Participants

The inclusion criteria were: first-ever ischemic stroke, age greater than 18 years old, within 3 weeks of stroke onset, with hemiparesis in the upper or lower limb, and a Mini-Mental State Examination (MMSE) score greater than 24 and mild-to-moderate impairment: National Institute of Health Stroke Scale (NIHSS) score between 4 and 13. Patients with a history of cardiopulmonary disease, neurological condition other than stroke, significant alcohol abuse (80 g of ethanol/day) in the months preceding the stroke, use of medications that could affect muscle function, and inability to perform voluntary muscle assessments were excluded.

Study variables

Baseline patient data were collected, and stroke characteristics were described as follows: ischemic stroke etiology (atherosclerotic, cardioembolic, small vessel, undetermined, atherothrombotic, or unusual); stroke severity according to NIHSS; and stroke location classified as total anterior circulation infarct, partial anterior circulation infarct, posterior circulation infarct, or lacunar infarct. Lesions were further classified by brain hemisphere as: (right, left, and bilateral) and anatomical location (supratentorial/infratentorial). The degree of independence in activities of daily living was measured using the Barthel Index [16]. Acute disability after stroke was assessed using the Modified Ranking Scale [17]. Cognitive status was evaluated with the 30-question MMSE [18].

Dysphagia was assessed using the Volume-Viscosity Swallow Test (V-VST) [19]. The test was administered by a trained clinician at the bedside using boluses of different volumes (5, 10, and 20 ml) and viscosities (thin liquid, nectar-like, and pudding-like), prepared with a commercial thickener. The assessment started with the nectar-like viscosity and progressed to thin liquid and pudding-like consistencies when tolerated. Clinical signs of impaired safety (cough, voice changes/wet voice, and oxygen desaturation ≥3% measured by pulse oximetry) and impaired efficacy (oral leakage, oral residue, multiple swallows, and suspected pharyngeal residue) were recorded. Patients who showed impairments in safety and efficacy underwent videofluoroscopy. The videofluoroscopy assessments were performed through videofluoroscopy: the lateral profile sequence of fluoroscopy was captured by continuous recording using digital storage of images at a rate of 25 frames/s. Visipaque, a water-soluble contrast, and Thickener Resource, a ThickenUp, were used to obtain three distinct viscosities (liquid, nectar, and pudding). Water-soluble contrast is available in bottles of 50 ml (16-mg iodine) or 320 mg/ml. Liquid viscosity was obtained by mixing 1 : 1 mineral water and water-soluble contrast at room temperature (100 ml/16-mg iodine). Nectar and pudding viscosities were achieved by adding 3.5 and 8 g of thickener, respectively, to the liquid solution in a mixing pan to obtain nectar and pudding viscosity. The solutions were prepared 10 min before the assessment. Boluses of each viscosity and a soft solid, in increasing volumes of 5, 10, and 20 ml, were administered to the following patients using a syringe. Swallowing function was evaluated using the Penetration-Aspiration Scale (PAS) [20]. The 8-point scale was used to characterize the depth and response to airway invasion during videofluoroscopy. Indeed, patients with a PAS score greater than or equal to 3 were classified as having dysphagia [21].

Respiratory muscle strength was assessed by measuring the maximal inspiratory pressure (MIP) and the maximal expiratory pressure (MEP) using a pressure transducer Micro RPM (Micro Medical Care Fusion, Kent, UK). For MIP, patients were asked to perform a maximum inspiration effort from the residual volume. For MEP, patients were asked to perform a maximum expiration from the total lung capacity (TLC). A flanged mouthpiece with a small orifice was used to ensure an optimal mouth seal and to minimize the involvement of facial and mouth muscles. Each measurement was repeated at least three times, and the highest value was recorded [22].

Cough function was assessed throughout voluntary peak expiratory cough flow (PECF) using an oronasal mask connected to a portable analogue peak flow meter (Mini-Wright Peak flow meter). Participants were seated and instructed to perform five volitional maximal cough maneuvers, after inspiration to TLC. The highest value of the five reproducible maneuvers (<10% variability between values) was used for analysis. No signs of fatigue were observed in the participants, and no participant requested discontinuation of the study [23].

Trunk control was assessed using the Trunk Control Test (TCT), which evaluates four distinct axial movements: rolling from the supine position to the weak side (T1), rolling from the supine position to the strong side (T2), sitting up from a lying-down position (T3), and sitting in a balanced position on the edge of the bed with feet off the ground for 30 s (T4). Scoring criteria: unable to perform the movement without assistance (0 points); able to perform the movement but in an abnormal manner (12 points); and able to complete the movement normally (25 points). The total TCT score was the sum of the scores for the four movements, ranging from 0 to 100 [24].

Statistical analysis

Categorical variables were presented as counts and percentages, and continuous variables as means and SD. T-tests were used to compare the main study variables between patients with and without dysphagia. Univariate and multivariable logistic regression analyses were performed to explore the possible association between dysphagia diagnosis and other relevant clinical variables. The candidate predictor variables were included in the model using a manual selection approach based on clinical relevance, and multicollinearity among predictors was assessed using the variance inflation factor. Indeed, receiver operating characteristic (ROC) curve analysis was performed to evaluate the ability of age at inclusion to discriminate between patients with and without dysphagia. Discrimination was quantified using the area under the curve (AUC) with 95% confidence intervals. The optimal cutoff was selected by maximizing Youden’s index.

Results

Of the 140 patients recruited (72.8% men), 26 were excluded: 19 were unable to perform the study maneuvers, 3 had cognitive impairment, 2 had severe facial paralysis, 1 had a history of poliomyelitis, and 1 was discharged before completing the assessments. Consequently, 114 patients were included in the descriptive analysis.

Demographic and clinical characteristics are presented in Table 1. Patients with dysphagia were older when compared with patients without dysphagia [70.0 (SD 9.2) years vs. 61.4 (SD 11.4) years; P = 0.00]. The mean of the PAS score was higher in patients with dysphagia [6.0 (SD 1.63) vs. 1.4 (SD 0.6); P = 0.00]. No significant differences were observed between groups for the remaining clinical variables.

Table 1.

Baseline demographic and clinical characteristics of the sample

	Total sample N = 114	Dysphagia N = 30	Non-dysphagia N = 84	P value
Age, years, mean	63.7 (SD 11.7)	70.0 (SD 9.2)	61.4 (SD 11.7)	0.00
Sex, n (%)  Man  Woman	83 (72.8%) 31 (27.2%)	23 (74.2%) 8 (25.8%)	59 (71.1%) 24 (28.0%)	0.469
Smoking history, n (%)  Current smoker  Past smoker  Never smoker	37 (34.9%) 16 (15.1%) 53 (50.0%)	6 (20.7%) 3 (10.3%) 20 (69%)	31 (40.3%) 13 (16.9%) 33(42.9%)	0.056
COPD, n (%)	2 (2.1%)	1 (1.3%)	1 (5.6%)	0.341
Hypertension, n (%)	74 (66.1%)	21 (28.4%)	53 (71.6%)	0.384
Diabetes, n (%)	36 (31.9%)	8 (22.2%)	28 (77.8%)	0.318
Dyslipidemia, n (%)	47 (14.6%)	13 (27.7%)	34 (72.3%)	0.494
Stroke onset (days), mean	11.1 (SD 9.8)	9.1 (SD 5.8)	11.89 (SD 10.9)	0.190
Stroke hemisphere, n (%)  Right  Left  Bilateral	53 (48.2%) 55 (50.0%) 2 (1.8%)	13 (46.4%) 14 (50%) 1(3.6%)	40 (48.8%) 41(50%) 1 (1.2%)	0.719
Stroke etiology, n (%)  Cardioembolic  Small vessel  Undertermined  Atherotrombotic  Atherosclerotic  Unusual	9 (13.6%) 18 (27.3%) 29 (43.9%) 7 (10.6%) 1 (1.5%) 2 (3.0%)	2 (9.5%) 8 (38.1%) 88 (38.1) 2(9.5%) 0 1 (4.8%)	7 (15.6%) 10 (22.2%) 21 (46.7%) 5 (11.1%) 1 (2.2%) 1 (2.2%)	0.739
OXFORD clasification, n (%)  TACI  PACI  POCI  LACI	12(20.7%) 15(25.9%) 11(19.0%) 20 (34.5%)	5 (21.7%) 3 (13.0%) 7 (30.4%) 8 (34.8%)	7 (20.0%) 12 (34.3%) 4 (11.4%) 12 (34.3%)	0.166
Anatomical location, n (%)  Supratentorial  Infratetorial	90 (81,8%) 20 (18.2%)	22 (73.3%) 8 (26.7%)	68 (85%) 12 (15%)	0.129
MMSE (mini-mental)	25.6 (SD 4.1)	25.3 (SD 3.9)	26.9 (SD 2.7)	0.055
NIHSS on admission	5.2 (SD 3.1)	5.06 (SD 2.6)	5.26 (SD 3.4)	0.771
Modified Ranking Scale Score	3.22 (SD 0.8)	3.19 (SD 0.8)	3.23 (SD 0.8)	0.844
Barthel Index on Admission	54.1(SD 17.4)	51.60 (SD 17.7)	54.94 (SD 17.3)	0.377
Penetration-Aspiration scale	3.2 (SD 2.5)	6.0 (SD 1.63)	1.4 (SD 0.6)	0.00

Data are shown as mean (SD) or n (%).

The bold value represent a statistically significant results.

NIHSS, National Institute of Health Stroke Scale; PACI, partial anterior circulation infarct; PAS, Penetration-Aspiration scale; POCI, posterior circulation infarct; TACI, total anterior circulation infarct.

Of the 114 patients, 36 showed no impairments in safety and efficacy parameters on the V-VST. The remaining 78 patients, who presented abnormalities on V-VST, underwent videofluoroscopy, and 30 patients (38.5%) were diagnosed with dysphagia. PAS score distribution was (n, %): 1 : 29 (37.2); 2 : 14 (17.9); 3 : 5 (6.4); 4 : 7 (9.0); 5 : 5 (6.4); 6 : 1 (1.3); 7 : 10 (12.8); and 8 : 7 (9.0). Overall, 55.0% of patients demonstrated normal swallowing (PAS ≤ 3), while the remaining patients showed impaired swallowing function, with PAS scores greater than or equal to 4.

Our sample of stroke patients exhibited respiratory muscle weakness and impaired cough function (Table 2), with very low values for MIP and MEP [55.5 (SD 23.4) cmH₂O; 84.5 (SD 33.9) cmH₂O], respectively, and for PECF 244.9 (SD 87.5). Patients with dysphagia showed significantly lower peak cough flow compared with patients without dysphagia [215.0 (SD 85.8) vs. 255.6 (SD 90.3) L/min; P = 0.02]. In contrast, no significant between-group differences were observed for respiratory muscle strength or trunk control.

Table 2.

Description of cough function, respiratory muscle strength, and trunk control and the differences in these variables between participants with and without dysphagia

	Total sample N = 114	Dysphagia N = 30	No-dysphagia N = 84	Differences	95% CI	P value
Maximal inspiratory pressure (cmH2O), mean	55.5 (SD 23.4)	52.90 (SD 16.1)	56.51 (SD 25.5)	3.61 (SD 7.2)	−16.97 to 11.71	0.376
Maximal expiratory pressure (cmH2O), mean	84.5 (SD 33.9)	86.46 (SD 30.5)	83.83 (SD 35.1)	2.63 (SD 4.9)	−6.27 to 13.50	0.717
Peak expiratory cough flow (l/min), mean	244.9 (SD 87.5)	215.00 (SD 85.9)	255.59 (SD 86.1)	40.59 (SD 18.3)	4.32 to 7.86	0.029
Trunk control test, mean	78.8 (SD 24.8)	77.40 (SD 24.8)	79.43 (SD 25.0)	2.03 (SD 5.3)	−8.65 to 12.72	0.707

The bold value represent a statistically significant results.

CI, confidence interval.

Exploratory analyses examining the association between dysphagia diagnosis and relevant clinical variables were conducted in 99 patients with complete data (Table 3). Age was significantly associated with dysphagia diagnosis [odds ratio (OR) = 1.09, 95% confidence interval (CI): 1.03–1.16; P = 0.03]. No other variables in the model showed a statistically significant association with dysphagia.

Table 3.

Multivariable model without PEmax

Dysphagia
	Odd ratio	CI	P value
Intercept	0.006	0.000–1.151	0.063
PECF	0.994	0.988–1.001	0.081
MIP	1.007	0.983–1.030	0.576
TCT_TOTAL	1.004	0.980–1.030	0.723
Sex (female)	0.871	0.257–2.791	0.819
Age	1.090	1.033–1.160	0.003
NIHSS	1.005	0.836–1.192	0.952
Barthel Index	0.984	0.950–1.019	0.375

The bold value represent a statistically significant results.

CI, confidence interval; TCT, Trunk Control Test.

The ROC analysis showed that age discriminated dysphagia with an AUC of 0.717 (95% CI: 0.613–0.821; P < 0.001). The optimal cutoff by the Youden index was greater than or equal to 64.5 years, yielding a sensitivity of 83.9% and a specificity of 56.0% (Fig. 1).

Fig. 1.

Receiver operating characteristic (ROC) curve for age to discriminate dysphagia.

Discussion

Our cohort with mild ischemic stroke showed reduced cough function and respiratory muscle strength vs. normative values. Patients with dysphagia had poorer cough performance, but multivariable analyses found no independent associations between dysphagia and respiratory muscle strength, cough function, or trunk control. Age was the only discriminative factor, with an optimal threshold of ~65 years.

Regarding cough function, participants with dysphagia consistently exhibited lower PECF values across age groups, and this difference appeared more pronounced in older participants. These findings are consistent with previous studies reporting impaired cough function in stroke patients with dysphagia [25–27]. However, published PECF values show variability across studies [25–27]. Kulnik et al. [25] observed higher PECF values in patients with dysphagia (≈383 L/min), despite comparable age and time since stroke to our cohort. This difference may be attributed to clinical and methodological factors, including greater stroke severity, dysphagia identified through bedside screening, and cough assessment performed using a pneumotachograph. In contrast, substantially lower PECF values were reported by Sohn et al. and Han et al. (≈116 L/min). Although both studies assessed cough function using a peak flow meter, their cohorts presented marked cognitive impairment (MMSE ≈ 17) and severe disability (mRS ≈ 5) compared with our sample, factors that may have limited volitional performance and contributed to lower cough flows. Overall, differences in cough assessment methods, dysphagia criteria, and cognitive status likely contribute to between-study variability in PECF values. Although pneumotachograph remains the gold standard, portable peak flow meters offer a practical and reproducible alternative [28]. Given the potential influence of cognition on volitional cough, cognitive status should be considered alongside PECF when interpreting results and supporting clinical decision-making.

Our results for respiratory muscle strength align with previous studies, showing that respiratory muscle weakness is common after stroke [27,29,30]. However, our patients presented higher MEP values compared with those reported previously, and this pattern was also observed in patients with dysphagia, where a higher impairment of MEP is anticipated [27]. This discrepancy may be explained by differences in patient characteristics, as our rehabilitation unit primarily admits patients with mild stroke, preserved cognition, and a favorable functional prognosis [31]. In addition, compared with Han et al. [27], our cohort showed lower swallowing impairment (mean PAS 6 vs. 7) and better functional status (Barthel Index 51.5 ± 17.7 vs. 31.9 ± 31.5). It is also important to note that Han et al. defined dysphagia using a PAS greater than or equal to 2, whereas the literature more commonly considers PAS greater than or equal to 3 as indicative of airway invasion; therefore, this lower threshold may have increased the number of patients classified as dysphagic. Taken together, the milder functional and swallowing impairment in our sample likely contributed to the higher values for respiratory muscle strength. This interpretation is supported by previous research showing that declines in respiratory muscle strength are associated with greater functional impairment after stroke [32].

Prior literature reports correlations between expiratory muscle function and trunk control and highlights the importance of respiratory muscles in generating effective cough [33,34]. Therefore, we hypothesized that respiratory muscle strength, cough function, and trunk control would be associated with dysphagia, as greater impairment in these parameters would be expected among patients with swallowing dysfunction.

In line with this rationale, Han et al. [27] proposed cutoff values for voluntary PECF (≤151 L/min) and respiratory muscle pressures (MIP < 20 cmH₂O), and MEP (<38 cmH₂O) associated with an increased dysphagia risk, and reported higher odds of dysphagia for PECF below this threshold (OR 4.19). In contrast, in our cohort, none of the assessed parameters showed an independent association with dysphagia. Although this finding was unexpected, we believe that the discrepancy may partly be explained by differences in patient characteristics (Han et al.). Reference [27] included patients with greater overall impairment, a higher proportion of dysphagia, and hemorrhagic strokes, whereas our sample comprised predominantly mild ischemic strokes with generally lower dysphagia severity.

In our cohort, age was significantly different between patients with and without dysphagia. This finding is consistent with two recent meta-analyses [35,36] demonstrating that increasing age is a significant risk factor for poststroke dysphagia (OR = 2.13; 95% CI: 1.53–2.97; P < 0.001). Additionally, previous studies have reported a significant association between age and impaired swallowing function [37]. Accordingly, current stroke guidelines recommend [38] routine dysphagia screening in all patients, with particular attention to older adults due to their elevated risk profile.

Pereira et al. [39] identified age greater than 70 years as a threshold associated with a higher risk of dysphagia, and this cutoff has been frequently reported in the literature. However, this threshold appears to have been established arbitrarily. In contrast, our study performed a specific analysis to determine the optimal age threshold for predicting dysphagia, identifying 64.5 years as the most accurate cutoff.

Strengths of this study include the identification of a novel age cutoff value for predicting dysphagia, and the inclusion of clinical features that are physiologically relevant to dysphagia and can complement routine screening. Limitations include: patient’s characteristics which are mainly mild ischemic strokes with good cognitive status. And a small number of patients with dysphagia, most of whom had lower dysphagia severity.

Overall, our findings support a multifactorial view of poststroke dysphagia and reinforce the need for comprehensive swallowing assessment. They also suggest that commonly used age thresholds for heightened dysphagia risk may warrant reconsideration, although larger studies are needed to confirm the 65-year threshold before any change in clinical recommendations.

Acknowledgements

M.M.-S., A.G.-S., and E.M. have given substantial contributions to the conception or the design of the manuscript. M.M.-S. and D.B. contributions to acquisition, analysis, and interpretation of the data. All authors have participated in drafting the manuscript. M.M.-S., D.B., and E.M. revised it critically. All authors read and approved the final version of the manuscript. All authors contributed equally to the manuscript and read and approved the final version of the manuscript.

Both participants and their family members were provided with clear, understandable oral and written information about the study procedures. Before admission to our unit, all patients must sign an informed consent form, which has been approved by the Clinical Ethics Committee of the Institut Hospital del Mar d´Investigacions Mèdiques.

Data could be available upon request if considered by the corresponding author.

Conflicts of interest

There are no conflicts of interest.