The swallowing process needs coordinated movements of muscles of the mouth, pharynx, larynx, and esophagus [1]. These movements are all controlled by the central nervous system, particularly the cerebral cortex, medulla oblongata, and cranial nerve nuclei [2,3]. Dysphagia is common in stroke patients. The average incidence of dysphagia in stroke patients is approximately 40% [4].

Stroke patients with breathing difficulty may also have swallowing difficulty because of brief apnea during swallowing. The incidence of aspiration pneumonia is frequently associated with dysphagia [5-11]. Moreover, patients with decreased diaphragmatic movements may have high risk of aspiration pneumonia due to their insufficient capacity to induce effective coughing reflex [12-14].

Given the importance of respiratory complications and dysphagia in patients’ quality of life, we decided to assess the correlation between severity of post-stroke dysphagia and breathing difficulty. The National Institutes of Health Stroke Scale (NIHSS) is the most commonly used scale to evaluate stroke patients [15]. Modified Rankin Scale (mRS) is commonly used to measure the degree of disability or dependence of daily life in stroke patients. These scales are associated with patient’s prognosis and quality of life [16,17]. Therefore, we examined difference in functional assessment battery according to swallowing function.

The objective of this study was to investigate differences in bilateral diaphragm movement for patients with tracheal aspiration according to post-stroke residue severity and determine correlations of Penetration Aspiration Scale (PAS), residue scale, and bilateral diaphragm movement.

This was a prospective cross-sectional study to assess inpatients diagnosed with hemorrhagic or ischemic stroke who were referred to the rehabilitation department of a university hospital. All eligible patients were recruited from March 2017 to February 2018 at the time of evaluation. Inclusion criteria were: first-time subacute (<3 months) stroke, age ≥19 years at stroke onset, and corresponding lesion on computed tomography or magnetic resonance imaging showing a single hemisphere lesion. Patients with histories of systemic illness or medication usage leading to neuropathy, loss of consciousness or lack of cooperative behavior, with sensory or global aphasia, having Mini-Mental State Examination (MMSE) score of less than 20 points, or having chest deformity were excluded from this study. All subjects provided written informed consent before inclusion in this study. Approval of this study was obtained from the Institutional Review Board of Pusan National University Hospital (No. H-1809-019-071).

Initially, 60 patients diagnosed with stroke and requested for dysphagia evaluation were enrolled. Finally, 47 hemiplegic stroke patients (28 men and 19 women; age, 66.4±12.8 years) satisfied our inclusion criteria. They were enrolled in the study. The period from onset to enrollment was 12.77±12.07 days. Four persons were within two standard deviations from normal diaphragm movement. There were 45 cases of ischemic stroke and 2 cases of hemorrhagic stroke. Of these 47 cases, 21 had right hemispheric stroke and 26 had left hemispheric stroke. Affected levels were as follows: cortical in 8, subcortical in 14, brainstem in 4, and multiple level in 21 cases (Table 1). Peripheral nerve lesion was not identified on phrenic nerve conduction.

Patients with tracheal aspiration (PAS≥6) had significantly (p=0.035) lower ipsilateral diaphragm movement during spontaneous breathing. Post-swallow residue severity showed statistically significant (p=0.028) difference in patients with ipsilateral diaphragm movement during forced breathing (Table 2). Post hoc testing with the Mann–Whitney U test also showed significant difference between groups 0 and 3 (p=0.007), groups 1 and 3 (p=0.006), and groups 2 and 3 (p=0.030).

In linear regression analysis, PAS showed weak correlations with ipsilateral spontaneous diaphragm movement (r=0.397, p=0.006), ipsilateral forced diaphragm movement (r=0.384, p=0.008), and contralateral forced diaphragm movement (r=0.323, p=0.027). Weak correlation was also observed between post swallow residue severity and ipsilateral diaphragm movement during spontaneous breathing (r=0.343, p=0.018) and forced breathing (r=0.331, p=0.023) (Table 3).

After dividing patients with or without tracheal aspiration based on PAS grade, no significant difference in mRS (p=1.0) or NIHSS (p=0.848) was observed between the two groups. There was no significant difference in mRS (p=0.773) or NIHSS (p=0.576) among the four groups divided based on the severity of residue either (Table 4).

In this study, we demonstrated that the extent of diaphragmatic movements in the ipsilateral side during forced or spontaneous breathing and in the contralateral side during forced breathing via fluoroscopic examination was correlated with patients’ swallowing function. Several studies have reported that stroke can affect breathing, swallowing, and motor impairment [23]. However, to the best of our knowledge, no study has assessed swallowing functions with bilateral diaphragmatic movements through fluoroscopy in stroke patients. In this study, the diaphragmatic movement test by fluoroscopy was performed simultaneously with VFSS for a short period of time. This is advantageous to observe the movement of the diaphragm objectively rather by ultrasonography. Our results revealed that forced breathing and swallowing were related to each other. Diaphragmatic movement during deep breathing may be considered similar to forced vital capacity. Forced vital capacity refers to air volume that can be exhaled when deep exhalation occurs. Compared with spontaneous breathing, forced breathing was associated with swallowing function more. The difference in diaphragmatic movements during spontaneous breathing was very small which might limit patient comparisons. Such difference might be due to large difference in forced breathing.

Stroke may disrupt breathing in many ways. It may influence respiration rhythm, cause problems in the respiratory pathway from the brain, and affect diaphragmatic movement. In the respiratory process, the diaphragm is the most important organ. Diaphragmatic movement is regulated by several pathways. Two major descending pathways are the corticospinal pathway from the cortex to respiratory motor neurons known to be responsible for voluntary breathing [24] and the bulbospinal pathway from the medulla to the respiratory motor neurons that is associated with the control of automatic breathing [25]. Moreover, a putative connection between the motor cortex and pontomedullary respiratory centers exists [26]. Because of these various control pathways, bilateral diaphragmatic movements in stroke patients may vary. In addition, the correlation between breathing and swallowing processes is clearly demonstrated by shared muscle components and anatomical structures. Breathing shares many muscles involved in swallowing [27,28]. There is approximately 1 second of apnea during swallowing [29]. Patients with poor pulmonary function have difficulty holding their breath which in turn affects their swallowing function. Moreover, during swallowing, non-spontaneous respiration should be controlled. This is done by the swallowing center. If damage occurs, this process becomes problematic [30]. Therefore, several previous studies have attempted to confirm the association of swallowing function with respiratory function.

Fregosi and Fuller [27] have measured the diaphragmatic movement in stroke patients with or without swallowing difficulty and healthy control group by M-mode ultrasonography while patients perform quiet breathing, deep breathing, and voluntary coughing. The phrenic nerve has a corticospinal projection. Hence, when comparing relationships between swallowing function and respiration, ipsilateral and contralateral sides of the diaphragm should be evaluated and their values showed be compared with swallowing function rather than only numerically comparing one side of diaphragmatic movement. However, in their study, only the diaphragm of the hemi-side was evaluated. Thus, we compared diaphragmatic movements on both sides. Fregosi and Fuller [27] used ultrasonography to evaluate diaphragmatic movement. The most popular method to evaluate diaphragmatic movement is ultrasonography. However, the quality of ultrasonography evaluation depends on technical skill of the operator. The gold standard for evaluating swallowing dysfunction is fluoroscopy. Hence, after VFSS, fluoroscopic diaphragm examination was done in the same laboratory.

In one study, nasal cannulae were used to determine differences in respiratory air flow during swallowing between normal and stroke patients. However, there was no statistically significant difference in apnea time. In addition, there is no evidence linking swallowing respiratory characteristics to aspiration [31]. However, mRS or NIHSS, the most basic functional assessment of stroke patients, was not significantly associated with dysphagia severity. The present study suggests that mRS or NIHSS cannot assess patient’s impairment in a multidisciplinary manner. This is probably because neither tool has a domain for dysphagia evaluation. Because these scales do not predict the reduction in quality of life (QOL) due to dysphagia, a common domain to assess activities of daily living, QOL depression due to dysphagia needs to be assessed using America Speech-Language Hearing Association National Outcome Measurement System swallowing scale.

This study has some limitations. First, because patient coordination is important to perform the diaphragmatic movement test and VFSS, all patients with an MMSE of 20 or less were excluded from enrollment. Hence, the number of subjects enrolled was very small. A similar study with a larger sample size will be needed to generalize results of this study. Second, most stroke patients included in our study were ischemic stroke patients and the distribution of patients was biased to one side. In addition, the proportion of the affected site of stroke in patients was not evenly distributed. In the future, it will be necessary not only to increase the number of patients, but also to make the distribution of patient groups more uniform so that the affected side of stroke does not become a bias in the study. Third, diaphragm movement of most patients was deviated from the normal range. This is due to the fact that no paper has published the normal range by measuring diaphragm movement using fluoroscopy. Thus, we compared diaphragm movement with the normal value using ultrasonography. In addition, because there were too many patients out of the normal range, there was a limit in additional analysis to result in statistically significant results except for these patients.

In conclusion, this study evaluated swallowing function and diaphragmatic movement through fluoroscopic examination and confirmed the relationship between swallowing function and bilateral diaphragmatic movement. Deep breathing exercise in addition to dysphagia therapy to enhance diaphragmatic movement could be administered to patients with swallowing problems and patients who cannot breathe deeply. Further longitudinal studies are needed to assess the effect of breathing exercise on post-stroke dysphagia.