PDF: PHAST-TRAC
(2018)Objective
To determine whether pharyngeal electrical stimulation (PES) increases readiness for decannulation in patients with stroke who required tracheotomy due to severe neurogenic dysphagia
Study Summary
• Trial stopped early for efficacy after 69 of planned 140 patients by IDSMB recommendation
• Of 17 PES responders, 14 (82%) had tracheal tube removed and 3 (18%) had cuff deflated
• Open-label phase: 37 (57%) of 65 patients who received at least one PES cycle were ready for decannulation
• Second PES cycle in initial non-responders: 4/15 (27%) retreated patients became ready for decannulation
• Sham non-responders given open-label PES: 16/30 (53%) ready for decannulation
• No recannulations within 48 hours or up to 30 days/hospital discharge
• No deaths or SAEs judged related to PES; mortality 20% PES vs 9% sham (OR 2.58; p=0.31, NS)
Intervention
Pharyngeal electrical stimulation (PES) via Phagenyx catheter at 5 Hz for 10 min daily on 3 consecutive days, with individually optimized intensity (mean 33.6 mA), vs sham stimulation using patient simulator box
Inclusion Criteria
Age >18, supratentorial stroke (ischaemic or haemorrhagic), mechanically ventilated ≥48h, weaned from ventilator but retained tracheotomy, sedation-free ≥3 days, RASS ≥-1, could not be decannulated due to severe dysphagia on FEES assessment. Exclusions: infratentorial stroke, pre-existing dysphagia, cardiac pacemaker/ICD, life expectancy <3 months
Study Design
Arms: Array
Patients per Arm: PES 35, Sham 34 (69 total)
Outcome
• Open-label decannulation (all PES-treated): 37/65 (57%)
• Retreatment responders: 4/15 (27%)
• Sham crossover to PES: 16/30 (53%) decannulated
• DSRS at day 90: PES 4.6 vs sham 5.7 (MD -1.10; p=0.44)
• FOIS at day 90: PES 4.6 vs sham 3.9 (MD 0.745; p=0.29)
• mRS at day 90: PES 4.1 vs sham 4.3 (MD -0.203; p=0.44)
• Zero recannulations during follow-up
• SAEs: PES 29% vs sham 23% (OR 1.30; p=0.79); 7 deaths PES vs 3 sham (p=0.31, NS)
Bottom Line
PES significantly increased readiness for decannulation in tracheotomized stroke patients with severe dysphagia (49% vs 9%; OR 7.00; p=0.0008), leading to early trial termination for efficacy. A second PES cycle rescued additional non-responders. Treatment was earlier onset (<28 days) and shorter ventilation (<15 days) predicted better response. No recannulations occurred, confirming the durability of the treatment effect.
Major Points
- PHAST-TRAC was a prospective, single-blind, randomized controlled trial across 9 sites (7 acute hospitals, 2 rehab facilities) in Germany, Austria, and Italy, enrolling 69 tracheotomized stroke patients with severe dysphagia who could not be decannulated (May 2015 - July 2017).
- The primary endpoint — readiness for decannulation assessed by FEES-based algorithm 24-72h after treatment — was met in 17/35 (49%) PES patients vs 3/34 (9%) sham patients (OR 7.00; 95% CI 2.41-19.88; p=0.0008). The trial was stopped early for efficacy by the IDSMB after 69 of 140 planned patients.
- Among the 17 PES responders judged ready for decannulation, 14 (82%) had complete tracheal tube removal and 3 (18%) had permanent cuff deflation. No patient required recannulation at 48h or up to 30 days, confirming durable treatment effect.
- In the open-label phase, 30 sham non-responders received PES: 16 (53%) became ready for decannulation. Of 15 PES non-responders who received a second (retreatment) cycle, 4 (27%) became ready. Overall, 37/65 (57%) patients who received at least one PES course achieved decannulation readiness.
- Prespecified subgroup analyses revealed significant treatment-by-subgroup interactions for onset-to-randomization time (<28 days: OR 56.37; ≥28 days: OR 1.58; interaction p=0.03) and ventilation duration (<15 days: OR 108.20; ≥15 days: OR 0.82; interaction p=0.0004), indicating PES is most effective when given early.
- Clinical dysphagia scores (DSRS, FOIS) and stroke severity (NIHSS, mRS) did not differ significantly between groups at any timepoint, though the study was not powered for these secondary endpoints.
- Baseline characteristics were well-balanced: mean age 64 years, 64% male, 71% ischaemic stroke, mean NIHSS 17.5, median onset-to-randomization 28 days, median ventilation 15 days. All patients had maximum dysphagia severity (DSRS 12/12, FOIS 1/7).
- Safety: 10 PES patients (29%) and 8 sham patients (23%) had at least one SAE (OR 1.30; p=0.79). Seven deaths in PES group vs 3 in sham (OR 2.58; p=0.31), none judged related to PES. The mortality difference was not statistically significant and reflected the severely ill patient population.
Study Design
- Study Type
- Prospective, single-blind, sham-controlled, randomized trial with sequential (triangular) design
- Randomization
- Yes
- Blinding
- Single-blind (outcome assessors blinded; treating investigators unmasked)
- Sample Size
- 69
- Follow-up
- 90 days
- Centers
- 9
- Countries
- Germany, Austria, Italy
Primary Outcome
Definition: Readiness for decannulation (FEES-based algorithm) 24-72h after treatment
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 3/34 (9%) | 17/35 (49%) | 7 (2.41-19.88) | 0.0008 |
Limitations & Criticisms
- Trial stopped early at 69 of 140 planned patients — risk of overestimating treatment effect.
- Single-blind design — treating investigators were unmasked, introducing potential performance bias.
- Primary endpoint (readiness for decannulation) is a surrogate; clinical outcomes (DSRS, FOIS, mRS) were not significantly different between groups.
- Mortality numerically higher in PES group (20% vs 9%) though not statistically significant — raises safety concern in severely ill population.
- Small sample size limits power for secondary endpoints and subgroup analyses.
- Manufacturer-funded (Phagenesis Ltd) — potential bias.
- Very selected population (tracheotomized stroke patients with maximum dysphagia severity) — results may not generalize to less severe dysphagia.
Citation
Lancet Neurol 2018;17:849-859