Overall
To assess whether Pharyngeal Electrical Stimulation (PES) is safe and effective at improving post-stroke dysphagia (PSD).

• Does 6 days of PES accelerate return to oral intake of food and drink as assessed using the dysphagia severity rating scale and blinded to treatment?

• Does PES improve swallowing and reduce pneumonia, antibiotic exposure, hospital length of stay and disability?
• Does PES increase quality-of-life?
• Is PES cost effective as compared to usual care?
• What subgroups predict response to PES?

The null hypothesis is that PES does not alter DSRS at day 14 post recruitment in participants with PSD. The primary outcome, DSRS (dysphagia severity rating scale) will be compared between PES and no PES using multiple linear regression with adjustment for stratification and minimisation variables.

Assuming alpha 5%, power 90%; DSRS difference 1.2, standard deviation 5.0; losses 3%, crossovers 3%; sample rounded up, a sample size of N=800 is needed (PES n=400, control n=400) (assumptions based on pilot trials and STEPS trial).

800 hospitalised adults (age ≥18) with recent (2-31 days) ischaemic or haemorrhagic anterior or posterior circulation stroke (as diagnosed clinico-radiologically) at a stroke centre, and Clinical dysphagia defined as a functional oral intake scale (FOIS) score of 1 (nothing by mouth, feeding by NGT/PEG),2 (tube dependent with minimal attempts of food or liquids) or 3 (tube dependent with consistent oral intake of food or liquids).  NIHSS item 1a score score of 1 (nothing by mouth, feeding by NGT/PEG),2 (tube dependent of 0, 1 or 2 (where the patient requires repeated stimulation to arouse). Baseline DSRS supervision score of either 3 (requiring therapeutic feeding by SALT team; on oral trials) or 4 (No oral feeding).

Non-stroke dysphagia, e.g., due to traumatic brain haemorrhage, subarachnoid haemorrhage, brain tumour, Parkinson’s disease, multiple sclerosis, severe dementia, head or neck cancer.  NIHSS item 1a score of 2 (where the patient only responds to pain) or NIHSS item 1a score of 3.  Pre-stroke dysphagia or dependency (modified Rankin scale, mRS 4/5).  Investigator believes dysphagia will be short-term, e.g. there are signs of impending recovery in swallowing (i.e. enrolment is reserved for patients whom the investigator considers are unlikely to resolve spontaneously over the next few days).  Patient expected to be repatriated to a separate organisation.  Patient expected to be rehabilitated at a separate organisation.  Patient not likely to be in the treating hospital for at least 14 days. Ongoing or anticipated ventilation/intubation/tracheostomy. Ongoing treatment of dysphagia with other forms of electrical / magnetic stimulation (e.g. NMES, TCDS, rTMS or Ampcare), or devices (e.g. EMST, IQORO, IOPI, biofeedback that uses EMG electrodes, or chin tuck against resistance using a ball/chin depressor) – this applies for the duration of the trial Malignant middle cerebral artery syndrome.  Pregnant.  Pacemaker, cochlear implant or implantable cardioverter-defibrillator.  Need for >35% of oxygen. Two or more NGT pulled out within the last week unless nasal bridle in place.  Investigator feels patient will not tolerate PES catheter.  Palliative care.  Presence of a pharyngeal pouch. Participant is risk-feeding at time of screening. 

PES on top of guideline-based standard-of-care. PES will be administered on days 1-6 using a commercial catheter with integral feeding tube. PES involves six daily 10 minute treatments at 5 Hz; threshold and tolerability currents will be assessed and the treatment current set at threshold + 0.75 x (tolerability - threshold) with current generated by a base-station. Dosing levels will be monitored, and sites informed if the stimulation current is too low, i.e. <20 mA; sites will be re-trained on the importance of delivering adequate current, if necessary. The catheter will be replaced if removed or pulled out prior to 6 treatments being administered. Treatment will be administered by PES-trained research coordinators, nurses or SLTs who are not involved in outcome data collection.

No PES catheter/stimulation on top of best guideline-based standard-of-care dysphagia management. A standard NGT will be used for feeding as necessary.

Primary at day 14−1 +3 (day 13-17)
Dysphagia assessed using DSRS, based on bedside clinical assessment/management conducted at days 14±1 +3. Outcome assessment will be assessed by DSRS/FOIS-trained research coordinators, nurses or SLTs who are not involved in treatment.

Secondary at days 1-9
PES threshold, tolerability and stimulation currents; number of catheters used.

Secondary at day 14−1 +3 (day 13-17)
DSRS >3, FOIS, EAT-10 and feeding status score (FSS); NGT/PEG in situ; pneumonia; antibiotic use; weight; EQ5D-5L, EQ-VAS (source recorded - patient or proxy), cognition (MoCA, TICS, MMSE, semantic verbal fluency, phonemic verbal fluency, clinical diagnosis of dementia, cognition from informant, (IQCODE), dependency and disability (mRS, BI, NIHSS), frailty (CFI), mood / depression (Zung), global (Stroke Impact Scale)

Note 1: We have included several measures of dysphagia and feeding: DSRS, FOIS, EAT-10 and FSS to triangulate the presence and magnitude of dysphagia and effects on feeding. These will be analysed together using the Wei-Lachin test.

Secondary at discharge/death by hospital assessor:
Length of stay; antibiotic use; swallowing therapy contact time; time to removal of NGT/PEG; EQ5D-5L, EQ-VAS (source recorded - patient or proxy); admitted to ICU; discharged with PEG; disposition (home, residential home, nursing home, hospital, death).

Secondary at day 90±7 by central telephone assessor blinded to baseline, treatment and in-hospital data (or by post)*:
DSRS, FOIS, EAT-10, FSS; home time; dependency (modified Rankin Scale ¹), disability (Barthel Index), quality of life (EQ-5D5L/EQ-VAS.), cognition (TICS ² , MoCA, TICS, MMSE, semantic verbal fluency, phonemic verbal fluency, clinical diagnosis of dementia, IQCODE), mood (Zung ³);^1,4-8 disposition, frailty (CFI), global (Stroke Impact Scale), Health Economic Resources
Note 1: These outcomes are all sensitive to therapeutic change.

Secondary at day 180±7 by central telephone assessor blinded to baseline, treatment and in-hospital data (or by post)*:
DSRS, FOIS, EAT-10, FSS; home time; dependency (modified Rankin Scale ¹), disability (Barthel Index), quality of life (EQ-5D5L/EQ-VAS.), cognition (TICS ² , MoCA, TICS, MMSE, semantic verbal fluency, phonemic verbal fluency, clinical diagnosis of dementia, IQCODE), mood (Zung ³);^1,4-8 disposition, frailty (CFI), global (Stroke Impact Scale)

Secondary at day 365±7 by central telephone assessor blinded to baseline, treatment and in-hospital data (or by post)*:
DSRS, FOIS, EAT-10, FSS; home time; dependency (modified Rankin Scale ¹), disability (Barthel Index), quality of life (EQ-5D5L/EQ-VAS.), cognition (TICS ² , MoCA, TICS, MMSE, semantic verbal fluency, phonemic verbal fluency, clinical diagnosis of dementia, IQCODE), mood (Zung ³);^1,4-8 disposition, frailty (CFI), global (Stroke Impact Scale), all cause mortality.
*If participant is an inpatient at day 90, 180 or 365, the follow up can be completed by the staff at site face to face.

Safety
PES has an excellent safety record in previous trials. Participants with PSD, who usually have severe stroke, will have multiple adverse events and SAEs. Hence, we will limit recording to: SAEs over 0-9 days, procedure/device-related (S)AEs over days 0-14; fatal SAEs over days 10-90 and all-cause mortality to day 365.

Costs
Health care resource use at discharge, day 90, day 180 and day 365.

The primary outcome, DSRS, will be compared between PES and no PES using multiple linear regression with adjustment for minimisation variables. The null hypothesis is that PES does not alter DSRS at day 14 in participants with PSD.

Assuming alpha 5%, power 90%; DSRS difference 1.2, standard deviation 5.0; losses 3%, crossovers 3%; sample rounded up, a sample size of N=800 is needed (PES n=400, control n=400).