Cross-System Integration of Respiration and Deglutition: Function, Treatment, and Future Directions

Abstract

Swallowing occurs preferentially in the expiratory phase of the quiet breathing cycle and at mid-to-low tidal volume. This coordinative pattern imparts important biomechanical advantages to swallowing and airway protection and facilitate laryngeal elevation, laryngeal vestibular and vocal fold closure, and cricopharyngeal sphincter opening. This preferred coordinative relationship between breathing and swallowing is impaired in a variety of patient populations, including head and neck cancer survivors with dysphagia. We developed a training protocol to re-establish more optimal phasing of swallowing with breathing in these patients with striking outcomes, including reduced swallowing physiological impairments and improved airway protection. This motivated us to continue to refine and expand this training protocol and develop new assistive technologies for swallowing monitoring outside of the lab. In this review, we highlight the origins of our optimal respiratory-swallowing coordination hypothesis, describe the biomechanical advantages it provides, carefully describe our training protocol and findings, and chart a course for the next phase of this work. Our overall goal is to harness technology combined with carefully constructed learning paradigms to improve the lives of patients with impaired respiratory-swallowing coordination consequent to a variety of pathologies including head and neck cancer and degenerative neurological conditions such as Parkinson’s disease.

Introduction

Much theoretical attention and experimental attention has been given to understanding physiologic impairments that occur within the oropharyngeal swallowing system. These impairments are identified in patients with dysphagia and are directly targeted using evidence-based swallowing interventions. Until very recently, cross-system impairments, including those between respiration and swallowing, were not identified in traditional swallowing assessments. This is despite the fact that these impairments can negatively impact swallowing function and airway protection in vulnerable populations [1–13]. Work in our combined labs for over 30 years has been directed toward elucidating normal coordinative relationships between breathing and swallowing using animal [14, 15], infant [16–18], and adult human models [1, 15, 19–27]. We have advanced several key hypotheses about the neural control, biomechanical and airway protective mechanisms underlying normal and impaired respiratory-swallowing coordination. We have translated this foundational knowledge in the development of a respiratory-swallow training therapy (RST) that establishes a more optimal respiratory-swallow phase relationship and facilitates improved airway protection and swallowing efficiency. We will provide the background to this innovative therapeutic protocol, explain its fundamentals, and provide a summary of data on its application from previous and ongoing clinical trials.

Biomechanical Advantages of Initiating Swallowing During Expiration at Mid-to-low Lung Volumes

Physiologic activities of muscles and structures involved in resting, spontaneous, quiet breathing have been the focus of basic and clinical studies for almost a century [26–31]. It is well-established that muscles of the tongue, pharynx, larynx, and cervical esophagus are all active to varying degrees during quiet breathing and facilitate airway patency [28–33]. Differential posturing and movements of oral, pharyngeal, and laryngeal structures have also been documented [28–33]. Fluoroscopic and endoscopic observations show that during inspiration, the tongue moves forward, pharynx widens, vocal folds abduct, and the larynx descends due to contraction of the diaphragm, placing counterproductive traction on the upper aerodigestive tract. See Figs. 1a, 2 [28, 31, 32, 34]. The direction of inspiratory movements is in opposition to the movements required for safe and efficient swallowing [25, 26]. See Figs. 1b, 2 [30,33,34,36]. In contrast, the expiratory phase of respiration results in posterior positioning of the tongue, [29, 30, 33] narrowing of the pharynx [32], superior positioning of the larynx [31], and inward posturing of the arytenoid-true vocal fold complex [28, 30]. This posturing creates a biomechanically efficient set point for swallowing movements necessary to safely close the airway and transfer the bolus [28–30] (Fig. 1c [25, 28–33]).

Fig. 1.

(1) Tongue Base (2) Pharynx (3) Larynx positioning during inspiration, expiration, and height of swallow

Fig. 2.

Glottic configuration during quiet breathing

We have demonstrated in healthy adults that swallowing is typically initiated at mid-to-low quiet breathing lung volumes, between mid-expiratory and resting expiratory level (REL) (Fig. 3) [15, 28].

Fig. 3.

Respiratory phase and lung volume [15, 28]

This respiratory phase and volume relationship provides key advantages in swallowing mechanics and airway protection, including facilitating laryngeal elevation, cricopharyngeal sphincter opening, laryngeal vestibular closure, tongue base retraction, and pharyngeal clearance [32, 35]. We have subsequently coined the term ‘optimal respiratory-swallowing coordination’ as the phase-volume target for swallowing initiation [26, 28, 33, 36]. Based on these physiological principals, our theoretical position of optimal respiratory-swallowing coordination and our findings of non-optimal coordination in patients with swallowing disorders [1, 8, 9], we developed a Respiratory-Swallowing Training (RST) intervention.

The Respiratory-Swallow Training Protocol

The RST protocol [9] is divided into 3 learning modules—Identification, Acquisition, and Mastery, based on a hierarchy of motor skill acquisition. The identification module (Goal 1 and 2) typically takes 1 session. The acquisition and mastery modules (Goal 3–6) take approximately 4–6 sessions. Each RST session lasts approximately one hour, and typically occurs once a week. Patients complete 10 trials of each subgoal. If a subgoal is met, patients move onto the next subgoal. Based on our experience, patients can meet the RST goals with ease. However, if a subgoal is not met in the first 10 trials, patients complete an additional 10 trials, for up to 4 sets of 10 trials. If the subgoal is not met after 4 sets of 10 trials, the therapy session for that day is terminated. Patients are offered the opportunity to repeat each session one more time (2 times total). If the goal is not met during the second session, treatment sessions are discontinued. See Fig. 4.

Fig. 4.

Identification, acquisition, and mastery modules for initial RST protocol

Identification Module (Goals 1 and 2).

The purpose of the identification module is to train patients to first identify respiratory-swallow events. In Goal 1, patients are trained to identify the inspiratory and expiratory phases (Goal 1a), respiratory cessation period accommodating the swallows (Goal 1b), and optimal and nonoptimal respiratory-swallow coordination (phase and lung volume) (Goal 1c) from 10 printed still illustrations of respiratory-swallow signals with 80% accuracy. They learn that upward deflections of the illustration of quiet breathing represents inspiration and downward represents expiration. Swallowing events are represented by a pause in the breathing trace. Patients are instructed to identify the target mid-to-low quiet breathing, expiratory lung volume range as the latter half of the expiratory phase of the simulation. In Goal 2, patients are trained to identify the same events from 10 pre-recorded videos of respiratory-swallow signals with 80% accuracy.

Acquisition Module (Goals 3 and 4).

The purpose of the acquisition module is to train patients to initiate swallowing the optimal respiratory phase and volume range. Patients are fitted with Inductobands (Inductotrace System; Ambulatory Monitoring, LLC, Ardsley, New York) and a nasal cannula to record their respiratory movements and these movements are displayed on a computer monitor for their viewing. In Goal 3, patients are trained to initiate swallowing in optimal respiratory-swallowing coordination (during expiration at mid-to-low lung volumes) with visualization of their own respiratory-swallow signals displayed in real-time and with verbal coaching from clinicians. Patients are trained during natural, self-administered liquid swallows of standardized and graduated consistencies of thin (i.e., water), nectar, and honey thick liquids (Thick & Easy^® Clear Drinks, Hormel Health Labs, LLC). First, clinicians train patients to initiate swallows during the expiratory phase of respiration by providing them immediate feedback on their performance (Goal 3a) and then during the expiratory phase of respiration at mid-to-low lung volumes (Goal 3b) with self-administered 5 mL thin liquid swallows. The same protocol is followed for 5 mL nectar (Goals 3c and 3d) and 5 mL honey thick (Goals 3e and 3f) liquid swallows. Training continues until the patient achieves 80% accuracy in 10 trials with each liquid consistency (thin, nectar, honey). In Goal 4, patients are trained to initiate swallowing during expiration at mid-to-low lung volumes with visualization of their own respiratory-swallow signals displayed in real-time, but without verbal coaching from clinicians. To meet Goal 4 and advance to the Mastery Module, patients must initiate swallowing during expiration with 80% accuracy, without verbal coaching, in 10 trials with each liquid consistency. Patients are trained only on liquid consistencies with PAS scores ≤ 6 identified on the baseline MBSS to optimize participant safety during training. If a patient is only able to train with one consistency, they complete three sets of 10 trials with that consistency for each goal to ensure equivalent practice for each patient.

Mastery Module (Goals 5 and 6).

The purpose of the mastery module is for patients to develop proficiency in the initiation of swallowing during expiration at mid-to-low lung volumes across all liquid consistencies and without visualization of participant’s own respiratory-swallow signals and without verbal coaching. Patients are still fitted with Inductobands and a nasal cannula for continuous recording, but not visualization, of respiratory movements and nasal airflow during the training sessions. In Goals 5 and 6, the real-time respiratory-swallow signals are hidden from the participant, but clinicians can view these signals in real-time to judge accuracy. To meet Session Goals 5 and 6, patients must initiate swallowing during expiration at mid-to-low lung volumes, without any visualization of signals or performance and with 90% accuracy in 10 trials of each liquid consistency. Similar to the acquisition module, patients are trained only on liquid consistencies with PAS scores ≤ 6 identified on the baseline MBSS to optimize participant safety during training This will not be completely clear to readers this previous sentence. Needs some context and/or additional explanation. If a patient is only able to train with one consistency, they complete three sets of 10 trials with that consistency for each goal to ensure equivalent practice for each patient.

Impact of Respiratory-Swallow Training

The impact, feasibility and safety of RST were examined in a Phase I clinical trial with a cohort of 28 survivors of oropharyngeal head and neck cancer [9]. These participants, many of whom were long past cancer treatment (range: 1–13 yrs., mean: 3.4 yrs.), had refractory dysphagia, and they initiated swallowing during the inspiratory phase of respiration observed on a baseline modified barium swallow study that included graduated volumes and consistencies of standardized liquid barium contrast (Varibar^® barium sulfate 40% w/v, Bracco Diagnostics, Inc., Monroe Township, NJ). The videofluroscopic image was synchronized with the respiratory-swallow signals acquired using Inductobands and nasal cannula and allowed us to simultaneously examine respiratory-swallow coordination and swallowing function. All 28 participants successfully completed the RST without any adverse effects and demonstrated significant increases in the frequency of swallows initiated during expiration compared to baseline (p < 0.0001). Further, training safety was confirmed in that airway invasion (Penetration-Aspiration Scale [PAS] scores [37]) did not worsen, rather improved (p < 0.0001) [9]. Three key physiologic components of swallowing also showed significant improvements measured by the Modified Barium Swallow Impairment Profile (MBSImP^™[38, 39]) including, laryngeal vestibular closure (p = 0.0004), tongue base retraction (p < 0.0001), and pharyngeal residue (p = 0.01) [1]. Morphometric analyses [40] collaborated beneficial mechanical changes measured with MBSImP scores. Significant improvements were also seen in dysphagia-related quality of life (MD Anderson Dysphagia Inventory (MDADI) p = 0.003) [41], with 30% of the sample reaching clinical significance [42].

A recent study also investigated the impact of RST on the timing of swallowing movements. A frame-by-frame analysis of Modified Barium Swallow Studies (MBSS) was completed pre- and post-treatment. MBSS bolus presentations included two trials of 5 mL, 15 mL and self-selected cup sip of standard Varibar contrasts (thin, nectar, honey) presented in sequential order. Bolus trials were grouped and averaged. Analyses of MBSS from 8 participants with chronic dysphagia (> 1 year) after oropharyngeal head and neck treatment showed that the duration of laryngeal vestibular closure (p = 0.016), laryngeal elevation (p = 0.0075), tongue base retraction (p = 0.026), hyoid elevation (p = 0.047), and pharyngoesophageal segment opening (p = 0.037) increased following Respiratory-Swallow Training. The results of this study also showed that RST resulted in increased duration of critical swallowing events that contribute to swallowing safety (i.e., laryngeal vestibular closure, elevation, and hyoid excursion) and efficiency (i.e., tongue base retraction, PES opening). These findings support previously reported clinical trial data showing improvements in MBSImP physiologic swallowing components TBR and LVC as well as improved airway safety (PAS [37] scores) as a result of RST (Table 1). Acknowledging the lack of randomization and a sham-controlled treatment in this trial, these encouraging, but preliminary findings, combined with the reaction of our research and clinical colleagues expressing their interest in this work, have motivated our team to conduct randomized-controlled trials investigating the efficacy of RST in larger cohorts of head and neck cancer survivors.

Table 1.

Pre- and Post-RST Scores

Outcome measure	Pre-intervention	Post-intervention	P-value
Optimal respiratory-swallow phase	43.0	86.0	< 0.0001
PAS ≥ 3	78.0	38.3	< 0.0001
MBSImP Component Scores
Laryngeal vestibular closure	78.2	55.5	0.0004
Tongue base retraction ≥ 2	96.0	88.6	< 0.0001
Pharyngeal residue ≥ 2	96.1	88.4	0.01
MDADI (mean ± SD)	60.4 ± 14.8	65.3 ± 16.7	0.003

Neural Mechanisms Underlying Improved Respiratory-Swallow Coordination

Although a detailed exploration of neural mechanisms is beyond the scope and focus of the paper, it is potentially useful to know our current understanding of neural mechanisms underlying improved respiratory-swallow coordination [43, 44]. Briefly, careful consideration of the respiratory [45, 46] and swallowing [47, 48] literature leads us to believe that cortical sites involved in swallowing initiation and expiratory effort might be necessary to learn to initiate swallowing near the mid-expiratory phase of respiration. We have further speculated [43, 49] that RST results in bi-system, cortico-cortical, long-term potentiation that enhances swallowing function, also potentially enhancing subcortical circuit control of swallowing. As the task is learned, the motor enhancement is embedded within cortico-basal ganglionic and cerebellar circuits that are shared between the swallowing and respiratory control systems [43, 44].

Using more traditional models of motor control, it could be hypothesized that RST fits within a classical, motor skill learning paradigm. An alternative but complementary theoretical explanation is that the bi-system cortical changes may be due to perceptual, somatosensory learning [50]. Recent evidence in limb motor systems suggests that a large portion of the change in motor areas of the brain that occurs in learning skilled movements is a by-product of perceptual learning [51]. A main component of RST is the perceptual identification of appropriate respiratory phase and volume for swallowing initiation. We guide our patients to determine and refine the somatosensory goals of appropriate coordination and to establish the best motor commands to meet these goals. The secondary, and equally important, feature of RST is repetition, or skilled practice through structured training sessions with appropriately timed feedback of results. This refinement and repetition of task, combined with the physiological benefits that optimal respiratory phase and volume create in the swallowing system, results in a positive perceptual and motor learning environment that resets and establishes more appropriate respiratory-swallow coordination.

Ongoing Clinical Trials

Randomized-Controlled Crossover Trial in Survivors of Head and Neck Cancer Who Are in the Chronic Phases of Their Recovery

The promising findings in patients with oropharyngeal head and neck cancer have motivated us to test the efficacy of RST in a more diverse cohort of head and neck cancer survivors, including survivors of oral, oropharyngeal, pharyngeal, and laryngeal cancers, increasing the generalizability of RST to a broader group of head and neck cancer survivors. The purpose of our ongoing randomized-controlled crossover trial is to determine the benefits of RST on eating, drinking, health, and quality of life of head and neck cancer survivors who are in the chronic phases of their recovery (> 6 months post-cancer treatment completion) and who suffer from refractory dysphagia. Patients are randomized into 2 groups: experimental and control. Patients in the experimental group begin RST described in Section III. right after a baseline MBSS synchronized with respiratory-swallow recordings. Patients in the control group wait one month after the baseline MBSS with no active treatment and then receive a second baseline MBSS and crossover to the experimental group.

Similar to the methods used in our preliminary study, patients in this ongoing clinical trial are fitted with Inductotrace Bands and a nasal cannula for continuous visualization and recording of respiratory movements and nasal airflow during the training sessions. The key difference from our preliminary trial is that real-time swallowing initiation feedback is provided by custom-made, interactive software developed in our lab [52]. The program uses rib cage and abdomen movements provided by the Inductotrace System and estimated lung volume using the calibration technique developed by Banzett et al. [53] from baseline examinations of quiet breathing. Automated detection provides visual signals to the patients and clinicians and indicates to the participants the target respiratory phase and lung volume to initiate swallowing. More specifically, participants are trained to initiate their swallows at or after the referent line in the lung volume signal indicating 50% of the expiratory phase of their quiet breathing. The respiratory pause to accommodate swallowing is used for swallowing detection.

The automated software detects if the swallow is initiated in optimal phase and lung volume [52]. Respiratory phase is categorized as inspiration or expiration based on the direction of the slope in the respiratory signal. If the direction of the slope at the onset of the swallow is downward (expiration), the visual message displayed on the screen reads, “CORRECT”. If the slope direction is upward at the initiation of the swallow, the feedback message reads, “INCORRECT”. If the program does not detect a swallow (no respiratory pause), the message reads, “MISSED” at the end of the swallow trial. Lung volume is represented by a visual cue (a thick horizontal line) bisecting the respiratory signal at 50% of the participant’s calibrated tidal volume during exhalation. If the swallow (respiratory pause) is initiated above the 50% line during the expiratory phase, the performance feedback given to the patient is “LUNGS TOO FULL”. If swallow initiation occurs within the targeted mid-to-low volume expiratory range (at or below the 50% line of the slope), the performance feedback given on the screen is, “CORRECT” (See Fig. 5).

Fig. 5.

Automated software detecting swallow event

Our preliminary findings show that patients learn to initiate a swallow in the optimal phase and lung volume range. The impact of RST on the primary and secondary outcomes, including swallowing physiology, health, and quality of life, will be assessed at the conclusion of the trial (Fig. 6).

Fig. 6.

Identification, acquisition, and mastery modules for telehealth RST study with sensor

Randomized, Sham-Controlled Clinical Trial in Survivors of Head and Neck Cancer Who Are in the Subacute Phases of Their Recovery

The logical extension of this work is to examine the safety and efficacy of RST in oropharyngeal head and neck cancer survivors who are in earlier subacute phases of treatment recovery, a time when there is greater potential for functional gains in swallowing [10].

The purpose of this randomized, sham-controlled clinical trial is to determine if RST results in increased frequency of swallows initiated during expiration and improved swallowing safety and efficiency in oropharyngeal cancer survivors who are in the subacute phases (3–6 months post-cancer treatment completion) of their recovery, when swallowing impairments and acute side effects of treatment toxicities have stabilized [54, 55]. In this randomized, phase II, sham-controlled clinical trial, 88 survivors with impaired respiratory-swallow phase patterning and airway safety and who are being treated for primary oropharyngeal head and neck cancer are randomized 1:1 into two group arms using a stratified, permuted-block randomization. Because of the known toxicity of radiation therapy on swallowing, strata are based on radiation therapy versus no radiation therapy to ensure balance across the two arms. Participants in the experimental group (n = 44) are trained to initiate swallowing during expiration and participants in the sham group (n = 44) receive sham training. The purpose of the sham training group is to ensure the highest level of rigor in our experimental design. During our research and clinical interactions with patients, we observed that when patients are trained to initiate their swallows during expiration, they typically initiate swallows at mid-to-low lung volumes followed by a brief post-swallow expiration. These observations motivated us to refine our treatment protocol and train patients to initiate swallows during expiration, without directly training lung volume range. The structure of the sham training is similar to the experimental training, but without the key therapeutic element of the treatment under investigation (swallow initiation during expiration). We hypothesize that patients in the experimental group will improve airway protection and efficiency when they learn the target skill (i.e., initiation of swallowing during expiration) compared to the patients in the sham group.

Both the experimental and sham training are delivered using a multi-use, wearable sensor, and customized sensor application (iOS) for recording respiration synchronized with swallowing combined with a telehealth platform to provide remote synchronous therapy to oropharyngeal head and neck cancer survivors. Until now, the identification and synchronous recording of respiratory and swallowing signals required cumbersome, non-portable and expensive equipment. Our initial validation studies [56, 57] have shown that a single wearable sensor has the capability of detecting and differentiating breathing and swallowing events, thereby eliminating the need for this burdensome laboratory recording equipment.

We also aim to examine the impact of RST on the frequency of swallows initiated during expiration outside of the clinic in wakeful, naturalistic swallowing environments, including eating and drinking. A challenge to all behavioral interventions is the ability to determine if the desired therapy outcome extends to natural every-day behavior. The wearable sensor used for treatment is also used for periodic ambulatory monitoring of quiet breathing and swallowing in natural contexts throughout the trial. To our knowledge, this is the first time a wearable sensor has been used to detect the phase of respiration (inspiration vs. expiration) during which swallowing is initiated in natural contexts for extended periods of time. The outcome of this study aim will provide an unprecedented viewpoint into the generalizability of swallow initiation during expiration.

Conclusion

Our long-standing research programs have provided foundational knowledge of the cross-system control of breathing and swallowing in normal individuals and persons with swallowing disorders. This work led to the development and testing of an optimal respiratory-swallowing coordination hypothesis and a training protocol to improve coordination in vulnerable individuals. Preliminary application of this RST resulted in improvements in airway protection and swallowing performance in individuals with refractory dysphagia consequent to head and neck cancer. We continue to refine our training protocols and develop innovative technologies to test the impact of RST in clinical trials with more powerful experimental designs and expanded patient populations. Our medium-term goals are to extend our RST approach to other clinical populations including patients with neurologic disease and pulmonary diseases such as Chronic Obstructive Pulmonary Disease (COPD) and to further our understanding of fundamental mechanisms of coordination. This work will bring rapid advancements in the development of training protocols and technologies to improve respiratory-swallowing coordination and the quality of life of patients suffering from the devastating consequences of swallowing impairments.