With reported incidence of 12.2 million people worldwide in 2019, stroke is the world’s second-leading cause of death and the third-leading cause of death and disability combined in the world (Feigin et al., 2021). Among the common disabilities, with a prevalence of 45-65% (Khedr et al., 2021; Martino et al., 2005), oropharyngeal dysphagia has major consequences on quality of life of patients and caregivers, and is associated with substantial costs for the healthcare system. However, little is known about the course of its recovery, and relationships between initial presentation and symptom resolution. Based on increased understanding of the contribution from both the brainstem and supratentorial structures for swallowing initiation and modulation, recent literature has identified distinct patterns of pathophysiological deficits that may involve not only a strength but also a skill component, responsible for impaired motor planning and execution (Ng et al., 2021). These preliminary findings suggest that variability in recovery of swallowing function following stroke may be subserved by pathophysiological deficits specific to lesion site. Moreover, they challenged the historical assumption of weakness being the main deficit underlying dysphagia, and the main treatment target. Following on from previous research from Ng et al. (2021), this research program addresses the gaps in the literature in respect to the relationship between acute presentation and long-term evolution of dysphagia post stroke (DPS), and further investigates the effects of a swallowing skill-based training on swallowing and quality of life in patients with chronic DPS.

Study 1 was a prospective, observational study designed in two parts. The first aimed to document the presentation and evolution of dysphagia in the first six months following stroke. The second aimed to characterise swallowing impairments in acute stroke patients according to the pathophysiological mechanisms: submental muscle weakness (strength deficit) or lack of precision in muscle contraction (skill deficit). Further, this study evaluated how dysphagic characterisation predicts long-term outcomes. Sixty consecutive adult participants (mean age 73.3  14.1 [29 – 101],

gender ratio [M:F] 1.4:1) were recruited in a tertiary acute hospital at a mean time of 2.8 days  1.5 days after a first-ever acute stroke, following a positive dysphagia screening. Participants were followed longitudinally for six months, with data collected at five timepoints. Sociodemographic and stroke typology data, occurrence of pneumonia, readmission rate and mortality, as well as nutritional data using FOIS and the IDDSI-FDS, were collected during patient interviews or chart reviews. The average length of stay (LoS) in the acute hospital was 7.6  5.6 days. On discharge, 45% of participants were consuming modified diet with IDDSI-FDS 5.9  2.8, and 20% had enteral nutrition. The average length of tube feeding was 16 days. Seventy percent were admitted to a rehabilitation hospital with a mean LoS of 29.1  21.8 days. During the follow up period, 20% of all participants were readmitted, and the rate of pneumonia and mortality was 23% and 8%, respectively. At six months, place of residence was home for 82% and hospital-level care for 13%, while 2% were patients in the rehabilitation hospital. One participant remained fed enterally via gastrostomy (2%) while 11% had a FOIS <7 and IDDSI-FDS<8, reflecting some level of modified intake. Six percent (n=3) of participants were still undergoing dysphagia management by a speech- language therapist, one participant for active rehabilitation, two for texture management. Multivariate analyses were conducted to identify prognostic factors for poor clinical outcomes at discharge and at six months. Patients with more severe stroke (NIHSS on admission) were more likely to be discharged to rehabilitation hospital. No other variables of interest were identified to predict the occurrence of pneumonia, mortality, readmission, residence, or diet level.

Twenty-two participants of the sample consented to be further evaluated to characterise dysphagia. A cranial nerve assessment, objective tests of functional oral intake (TOMASS and TWST), and a self-assessment of swallowing-related quality of life (SWAL-QOL) were completed. Muscle strength and contraction accuracy components of the suprahyoid muscles were assessed using surface electromyography (sEMG) coupled with a biofeedback device (BiSSkiT) during saliva swallowing and jaw opening tasks. Each SWAL-QOL parameter was significantly influenced by stroke laterality and swallowing temporal errors. After six months, TOMASS scores were within the range of

normal for 32% of participants and 15% for the TWST. Mixed effect models revealed that TOMASS number of swallows increased significantly between three and six months (+0.61, p=0.03). Using multiple linear regression analysis, significant influence of stroke site on TOMASS number of swallows (p=0.02) and TWST swallowing capacity (p=0.04) were found. A significant influence of swallowing temporal (p=0.04) and amplitude (p=0.002) errors were also found on TWST ingestion time. The results of mixed effect models analysis revealed a significant effect of time point on the sEMG variables. Specifically, effortful hit rate increased significantly between 10 days and one month (p=0.04), swallowing temporal error and hit rate from admission to six months (p<0.01), and jaw opening hit rate between three and six months and from admission to six months (p=0.04). Right strokes were significant predictors of decreased swallowing precision (hit rate) (p=0.03). Subcortical involvement was a significant predictor of decreased jaw opening precision (temporal errors) compared to cortical strokes (p=0.005).

The second part of this research program (Study 2) was an exploratory, interventional study investigating the effectiveness of a swallowing skill-based training on swallowing and quality of life for individuals with persistent dysphagia in the chronic stage of recovery, six months post stroke. Designed as a case series on an A-B-A design, the protocol included two baseline measures across two weeks, followed by a two-week treatment period, and two follow-up periods (two weeks and two months apart, respectively). Participants who completed the first study were eligible if they presented persistent dysphagia six months after their stroke. Of a possible 13 participants, four consented to participate and were recruited. During the two-week treatment period, one-hour daily sessions involved a swallowing skill-based training protocol, using sEMG paired with biofeedback. Clinical and swallowing outcome measures included patient’s reported dysphagia (EAT-10) and quality of life (SWAL-QOL). The TOMASS and TWST provided clinical measures of ingestive swallowing of solids and fluids, respectively, and a cranial nerve examination was completed, along with physiological swallowing measures of strength and skill derived from the sEMG device coupled with the BiSSkiT software. Additional measures were collected to inform on swallowing

biomechanics using videofluoroscopy (hyoid excursion, pharyngeal residue ratio [PRR], penetration- aspiration scale [PAS], duration of laryngeal vestibule closure [LVC] and upper oesophageal segment [UES] opening) and durational parameters using sEMG (premotor time [PMT], preswallow time [PST], total duration of contraction [TDC]). Descriptive statistics showed increased SWAL-QOL scores following treatment for all participants. Although they all demonstrated some level of functional improvement based on clinical outcomes (TOMASS, TWST, sEMG temporal measures of PMT, PST and TDC), the results were inconclusive in the absence of distinct patterns of change. In terms of strength and skill measures, swallowing temporal precision improved in half of the participants, and swallowing amplitude in all of them; jaw-opening temporal precision improved for one third and amplitude in one half of participants. With respect to VFSS, only one participant was able to undergo the assessment, and results were reported descriptively.

This study offers a preliminary and novel step in exploring the nature and evolution of strength and skill deficits underlying dysphagia. It suggests that distinct pathophysiological deficits of swallowing strength and skill may be influenced by stroke characteristics. It also provides preliminary evidence that swallowing temporal error may influence functional water swallowing and quality of life, suggesting that pathophysiological mechanisms of strength and skill may influence functional recovery of swallowing following a stroke. These findings support the need for development of differential diagnostic tools to better identify patients at risk of developing chronic dysphagia based on their pathophysiological deficits. They also provide reason enough to further explore these pathophysiological mechanisms with studies on larger samples to develop targeted rehabilitation approaches. This should greatly improve the effectiveness of our interventions by producing long- term effects on swallowing and thus on the quality of life of individuals following a stroke.