--- tags: Glia-ARM, Testing, ISO, FDA, UHN --- # 20201210 Final Report of Device Testing and Clinical Trial Data from Glia Aerosol-Reducing Mask (ARM) v1.0 for Non-invasive Ventilation [![hackmd-github-sync-badge](https://hackmd.io/RWzsJT1jTlS7SDs0I3eBFQ/badge)](https://hackmd.io/RWzsJT1jTlS7SDs0I3eBFQ) 1. ISO 17510:2015 Annexes B-F 2. Achievable FiO2 3. Clinical trial Report date: 2020.12.10 Azad Mashari, Maggie Xiao, Vahid Anwari, Chelsea Darling, Ben Thomson [TOC] ## 1. ISO 17510:2015 Report Testing performed 29 September - 30 October 2020, by Vahid Anwari and Azad Mashari at Toronto General Hospital and Thornhill Laboratories, under supervison of Azad Mashari MD FRCPC, Staff Anesthesiologist, Toronto General Hospital, University Network. ### General Set-up #### Head Phantom A head phantom was custom built by silicon casting from 3D printed molds (figure 1.1). The phantom was equipped with two sensor ports: 1. A reversed standard 22 mm airway taper at the mouth allowing an in-line pressure/flow sensor to be inserted from the back. 2. A pressure or gas sampling port at the nose, consisting of the barrel of a 3 ml syringe connected to pressure tubing with a luer connector. ![1.1 Head Phantom - front](https://i.imgur.com/WORJbPU.jpg) ***Fig 1.1**. Test phantom - front view showing pressure-flow (mouth) and pressure/gas sampling (nose) ports*. ![1.2. Head Phantom - back](https://i.imgur.com/Et9KvC5.jpg) ***Fig 1.2**. Test phantom - rear view Airway tubing is connected to FluxMed sensor. Sample tubes for the FluxMed and the pressure/gas sensor pass through the bottom of the neck*. #### Viral filters System was tested with two different viral filters compatible with the the ARM. These models were selected due to their availability and their low flow resistance values. Flo-Guard has the lowest resistance in an N95 filter. Hydro-Guard Mini has a resistance value typical of most models on the market. * Intersurgical Hydro-Guard Mini * [Manufacturer Page](https://fda.report/MAUDE/MDR/7742777) * Filtration Efficiency: >99.999% * Resistance at 30L/min: 1.3 cm H2O * Resistance at 60L/min: 2.9 cm H2O * Compressible volume: 63 ml * Minimum tidal volume: >200ml * Weight: 30 g * Connectors: 22F/15M-22M/15F * [FDA Report](https://fda.report/MAUDE/MDR/7742777) * Intersurgical Flo-Guard. * [Manufacturer Page](https://https://www.intersurgical.com/products/airway-management/floguard-low-resistance-breathing-filter-for-cpap-and-bilevel) * Filtration efficiency: >99.99% * Resistance at 30L/min: 0.4 cm H2O * Resistance at 60L/min: 0.8 cm H2O * Resistance at 120L/min: 2.0 cm H2O * Resistance at 240L/min: 5.1 cm H2O * Compressible volume: 80ml * Weight: 27.8g * [FDA Report](https://fda.report/PMN/K121026) #### Anti-asphyxia valve (required in some configurations) * Resmed Leak valve 24991: [FDA](https://fda.report/GUDID/00619498249916) #### Flow and pressure sensor: * FluxMed GrH: [Manufacturer Page](https://https://mbmed.com/fluxmed-grh/) ### Annex B: Exhaust Flow September 29, 2020, Toronto General Hospital, Hyperbaric Medicine Unit. #### Objective Define relationship between pressure provided to patient and exhaust flow #### Apparatus & Set-up * ARM Mask: AV2000 M, 3M Lot #031120 FT 936 * ARM Gasket and Adapter V1.0. ABS. Black. * Filters * Intersurgical Hydro-Guard Mini * Intersurgical Flo-Guard * Anti-asphyxia valve: Resmed Leak valve * Flow and Pressure Sensor: FluxMed GrH (s/n 2802020022) ![B1. Annex B testing set-up](https://i.imgur.com/60qBCWi.jpg) ***Fig B1**. Exhaust flow testing set-up with Hydro-mini filter. FluxMed sensor is connected at patient connection port. Wall oxygen (via a sequence of adapters) is connected to FluxMed. Flow readings are recoded from FluxMed sensor. Connection sequence: Seal - Mask - Gasket - Adaptor - Filter (Flo-Guard | Hydro-Guard Mini) - Asphyxiation valve ( No | Yes ) - FluxMed sensor - Flow source* #### Procedure 1. Sensor was positioned at the Patient Connection Port 2. To exclude flow source, the patient interface was sealed with plastic wrap (between the head mannequin and the mask) 3. Flow rate was adjusted to achieve the minimum pressure (5 cm H2O) in the RATED pressure range (0-40 cm H2O), pressure was measured at this point. 4. Step 3 was repeated at the following pressures: 10, 15, 20, 25, 30, 35, 40 cm H2O. 5. Flow rates were averaged over 10 seconds at sampling frequency of 8 Hz #### Results ##### Series 1: Dual-limb ventilator configuration – Flo-Guard Filter: Intersurgical Flo-Guard (Lot #32053172). Asphyxiation Valve: None |Trial|Target Pressure<br>(cm H2O)|Actual Pressure<br>(cm H20)|Required Flow<br>(L/min)| |-|-|-|-| |1.1|5|5.0|0.0| |1.2|10|10.1|0.1| |1.3|15|15.1|0.0| |1.4|20|20.0|0.1| |1.5|25|25.0|0.2| |1.6|30|30.0|0.0| |1.7|35|35.0|0.0| |1.8|40|40.0|0.2| ##### Series 2: Single-limb ventilator configuration - Flo-Guard Filter: Flo-Guard (Lot #32053172). Asphyxiation valve: Resmed |Trial|Target Pressure<br>(cm H2O)|Actual Pressure<br>(cm H2O)|Required Flow<br>(L/min)| |-|-|-|-| |2.1|5|5.3|25.0| |2.2|10|10.0|34.5| |2.3|15|14.9|41.4| |2.4|20|20.0|47.3| |2.5|25|24.8|52.3| |2.6|30|30.0|56.7| |2.7|35|35.6|62.3| |2.8|40|40.3|67.4| ##### Series 3: Dual-limb ventilator configuration – Hydro-Guard Mini Filter: Hydro-Guard Mini Lot# 32053312 Asphyxiation valve: None |Trial|Target Pressure<br>(cm H2O)|Actual Pressure<br>(cm H2O)|Required Flow<br>(L/min)| |-|-|-|-| |3.1|5|5.5|1.1| |3.2|10|10.0|1.0| |3.3|15|15.0|0.8| |3.4|20|20.0|1.0| |3.5|25|25.0|1.1| |3.6|30|30.0|0.9| |3.7|35|35.0|1.1| |3.8|40|40.0|0.9| ##### Series 4: Single-limb ventilator configuration – Hydro-Guard Mini Filter: Hydro-Guard Mini; Lot# 32053312 Asphyxiation valve: Resmed (Lot #) Date: Tues 29 Sept,2020, 12:30 pm |Trial|Target Pressure<br>(cm H2O)|Actual Pressure<br>(cm H2O)|Required Flow<br>(L/min)| |-|-|-|-| |4.1|5|5.3|23.6| |4.2|10|9.8|32.3| |4.3|15|15.2|40.3| |4.4|20|20.0|46.2| |4.5|25|25.3|51.5| |4.6|30|30.0|55.5| |4.7|35|35.0|59.7| |4.8|40|40.2|65.0| ### Annex C: Resistance to flow October 2, 2020, 10:30 AM #### Objective Measure pressure drop between source and patient airway #### Apparatus & Set-up * Flo-Guard Lot# 32053312 * Hydro-Guard Mini Lot# 32053312 * Sensor 1: FluxMed GrH (s/n 2802020022) – Patient-Connection Port * Sensor 2: FluxMed GrH (s/n 2802019032) – Airway ![C1. Annex C testing set-up](https://i.imgur.com/V9YtC8F.jpg) ***Fig C1**. Flow resistance set-up*. #### Procedure 1. Set-up as shown in Figure C1. Patient interface was open to the atmosphere to achieve the specified flowrates. 1. Flow/Pressure sensors were placed at the patient connection port (visible in image) and inside the mannequin in the mouth (line coming from the bottom of the mannequin visible) 1. At the target flow rate of 50 l/min, pressure measurements were averaged at a 10 second interval (8 Hz sampling rate) and the pressure delta recorded. 1. The test was repeated with saturated filters. Filters were soaked in a water bath for 10 minutes. removed, excess water drained and tested as above. #### Results |Filter|State|Patient Connection Port Flow<br>(L/min)|Patient Connection Port Pressure<br>(cm H2O)|Airway Flow<br>(L/min)|Airway Pressure<br>(cm H2O)|**Pressure difference**<br>**(cmH2O)**| |-|-|-|-|-|-|-| |Flo-Guard|Dry|50.1|0.7|41.2|0.8|**0.1**| |Flo-Guard|Dry|100.1|1.5|79.0|1.5|**0.0**| |Flo-Guard|Wet|50.0|0.9|40.4|0.7|**0.2**| |Flo-Guard|Wet|100.3|1.8|40.1|2.1|**0.3**| |Hydro-GuardMini|Dry|50.4|2.1|41.5|0.72|**1.4**| |Hydro-GuardMini|Dry|100.3|4.5|79.5|1.45|**3.1**| |Hydro-GuardMini|Wet|50.1|2.3|41.7|0.7|**1.6**| |Hydro-GuardMini|Wet|100.3|5.1|80.1|1.6|**3.5**| ### Annex D: Anti-Asphyxia Valve Opening and Closing Pressures Tuesday 29 Sept 2020. Toronto General Hospital #### Objective Determine the pressures at which the anti-asphyxia valve opens and closes to atmosphere #### Apparatus & Set-up ![D1. Annex D test set-up](https://i.imgur.com/z9kxO13.jpg) ***Fig D1**. Anti-asphyxia valve opening and closing pressure measurement set-up*. #### Procedure ##### Opening Pressure Set up is shown in figure D1, the FluxMed sensor was attached between anti-asphyxia valve and flow source. Pressure was increased until the anti-asphyxia valve closed to atmosphere, the closing pressure was recorded. Pressure was then decreased until the valve opened, the opening pressure was recorded. The test was repeated for both the Flo-Guard and Hydro-Guard filters. ##### Closing Pressure Starting with no flow and the anti-asphyxia valve open, flow was gradually increased until valve closure. Pressure at closure was documented for each filter. #### Results |Filter|Opening Pressure<br>(cm H2O)|Opening Flow<br>(L/min)|Closing Pressure<br>(cm H2O)|Closing Flow<br>(L/min)| |-|-|-|-|-| |Flo-Guard|2.2|13.4|36.8|103.0| |Hydro-Guard Mini|4.8|21.1|44.7|98.9| ### Annex E Determination of the inspiratory and expiratory resistance under single fault condition Friday 2 October 2020, 12:00 pm Toronto General Hospital #### Objective To determine the inspiratory and expiratory resistance at a flowrate of 50 l/min in single fault condition. #### Apparatus & Set-up * Flo-Guard * Hydro-Guard Mini * Anti-aphyxia valve: Resmed * Fluxmed sensor in mannequin airway (mouth) * Wall suction (Inspiratory flow source) * Wall oxygen (Expiratory flow source) ![E1a. Annex E Expiratory Resistance Set-up](https://i.imgur.com/xMm7wHm.jpg) ***Fig E1a**. Expiratory resistance set-up*. ![E1b. Annex E Expiratory Resistance Set-up Schematic](https://i.imgur.com/hpA76ZR.png) ***Fig E1b**. Expiratory resistance set-up schematic*. ![E2a. Annex E Inspiratory Resistance Set-up](https://i.imgur.com/BtqN0GT.jpg) ***Fig E2a**. Inspiratory resistance set-up*. ![E2b. Annex E Inspiratory Resistance Set-up Schematic](https://i.imgur.com/x2qDRmb.png) ***Fig E2b** Inspiratory resistance set-up schematic* #### Procedure A filter and anti-asphyxia valve were connected to the ARM device. The free end of the anti-asphyxia valve was sealed with plastic film to simulate obstruction of the ventilator flow source. 1. Expiratory resistance (Fig E1a, E1b): an oxygen line was connected to the phantom airway and adjusted to obtain an out-flow of approximately 50 L/min, measured by the FluxMed sensor in the phantom airway. The achieved pressure and flow were recorded. 1. Inspiratory resistance (Fig E2a, E2b): a suction line was connected to the phantom airway and adjusted to obtain a flow of approximately 50L/min, measured by the FluxMed sensor in the phantom airway. The achieved pressure and flow were recorded. #### Results |Filter|Inspiratory Flow<br>(L/min)|Inspiratory Pressure<br>(cm H2O)|Expiratory Flow<br>(L/min)|Expiratory Pressure (cm H2O)| |-|-|-|-|-| |Flo-Guard|-49.8|-1.3|50.2|0.6| |Hydro-Guard Mini|-49.3|-4.4|50.1|1.8| ### Annex F: CO<sub>2</sub> Rebreathing October 30, 2020 9:00-16:30 Azad Mashari, Vahid Anwari, Ben Thomson, Drew Miller (Thornhill) Thornhill Medical Labs, Mars Building, 661 University Ave. 2nd Floor Toronto, Ontario #### Objective To document safety of the device against CO2 rebreathing under normal function and relevant single fault conditions. Three supported configurations were tested: 1. Mask and filter only 2. BVM configuration with PEEP valve 3. Single-limb NIV configuration (with anti-asphyxiation valve) #### Set-up ![F1. Annex F CO2 Rebreathing Set-up Schematic](https://i.imgur.com/l7PUeow.png) ***Fig F1**. Set-up for Annex F testing, CO<sub>2</sub> rebreathing. No mixing chamber ("~10L MIX Ch.") was included. Components described further in text.* 1. Breathing simulator: ASL5000 * Mode: Volume Pump * VT 500ml * RR 15 * I:E 1:2 * FRC (URC) 1.5 L. 2. Gas analyzer: Oxisoft EIN0016 SN:O2-03504; Software 0.8.4 * Calibrated for CO2 and O2 * Sampling flow 280 mL/min. Set to match CO2 flow titrated at baseline * Data collected at 10Hz. 3. Pressure & flow sensor: Fluxmed GrH SN:2802020022 * Caliberated * Recording at 16Hz 4. CPAP ventilator: Respironics REMstar auto Intl. REF 551P SN: P04341300C18D. * Mode: CPAP * Ramp: Off * Flex type: None * Tubing Type 22mm 1.9m (fixed) 5. CO2 supply: 4.02% O2 with balance (95.98%) CO2. Connected by 50 foot, 1/4 inch ID tube to flux med. Flow meter inserted in-line (SFM 4300) #### Results values averaged over 15 simulated breaths (1 minute). Sampled at 10Hz. #### 0. Baseline (No device on manniqu) End-tidal CO<sub>2</sub> after titration [mean(SD)]: 5.04(0.01)% CO2 flow (1 minute average at 30Hz): 279(2) ml/min #### 1. Mask & filter only |Trial|Condition|ETCO2 max [mean(SD)]%|% Increase| |---|---|---|---| |0.0|Baseline: No Mask|5.04(0.01)|0 |0.1|Mask only|5.86 (0.02)|16.3| |1.1|Flo-guard dry|7.81 (0.02)|55.0| |1.2|Flo-guard wet|7.89 (0.02)|56.5| #### 2. BVM (Ambu with PEEP valve) No single fault test applicable since no flow is generated and there is no breathing tube to be blocked. |Trial|PEEP Set|PEEP Actual|ETCO2 max [mean(SD)]%|% Increase| |---|---|---|---|---| |2.1|0|0|8.26 (0.2)|**63.9**| |2.2|20|20.8|8.11 (0.02)|**60.9**| Acceptable increase is < 20%. **BVM configuration does not meet requirements**. #### 3. CPAP with anti-asphyxia valve and oxygen connector (capped off) |Trial|CPAP Pressure (cm H2O)|ETCO2 max [mean(SD)]%|% Increase| |---|---|---|---| |3.1|5|5.70 (0.2)|13.1| |3.2|10|4.31 (0.2)|**-14.5**| |3.3|20|2.3 (0.2)|**-54.4**| Note with CPAP support > 5 cm H<sub>2</sub>O, ETCO<sub>2</sub> reading dropped below baseline. ##### Single Fault CPAP tests |Trial|Fault|ETCO2 max [mean(SD)]%|% Increase| |---|---|---|---| |3.4|Patient connection port occluded|8.35 (0.03)|**65.7**| |3.5|CPAP connected with no flow|8.35 (0.01)|**65.7**| Acceptable increase in single fault mode is <60%. **Mask does not meet single fault requirements** for unmonitored use. ## 2. Achievable FiO2 Testing ### Objective To evaluate achievable FiO<sub>2</sub> with home CPAP configurations with supplimental oxygen. ### Set-up * Set-up identical to CO<sub>2</sub> rebreathing trials (figure F1). * Configurations and settings for Breathing simulator unchanged (Vt: 500ml; RR: 15; I:E 1:2). * O<sub>2</sub> sensor in pharynx (30 ml deadspace to mouth). * Trough values extracted for each breath. Average of 15 breaths (1 minute) after equilibraration (typically 3 minutes) ### Results | Oxygen Flow (L/min) | CPAP (cm H<sub>2</sub>O) | Minimum FiO<sub>2</sub>% | |---------------------|---------------|---------------| | 4 | 5 | 32 | | 8 | 5 | 47.4 | | 15 | 5 | 85.3 | | 4 | 10 | 31.2 | | 8 | 10 | 38.2 | | 15 | 10 | 55 | ![O1. Minimal FiO<sub>s</sub>](https://i.imgur.com/8mbxue3.png) ***Fig O1**. Minimum FiO<sub>2</sub> in Phantom Pharynx vs. Oxygen Flow at CPAP pressures of 5 and 10 cm H<sub>2</sub>O.* ## 3. Clinical Trial ### Objective To evaluate the effective leakage of the ARM compared to current standard of care NIV masks in patients undergoing emergent CPAP or BiPAP treatment. ### Methods Opportunistic randomized controlled trial. After [HSREB approval at Western University](https://github.com/GliaX/Aerosol-Reducing-Mask/blob/master/ethics_application/2020-04-27%20Ethics%20Approval.pdf) participants determined to require non-invasive ventilation by attending MDs in the emergency departments of affiliated teaching hospitals were recruited and randomized to the ARM or control standard mask ([Fisher & Paykel RT043M non-vented mask with anti-asphyxiation valve](https://www.fphcare.com/en-ca/hospital/adult-respiratory/noninvasive-ventilation/rt043/)). Patients requiring NIV for an actue indications were assessed for inclusion. Exclusion criteria were: Age <18 years; respiratory failure due to non-pulmonary pathology; impaired level of consciousness (Glasgow coma scale <10); upper gastrointestinal bleeding; Chest trauma; Agitatation or violence. Patients were ventialted using a Resmed (San Diego, US) Stellar 150 Non-invasive ventilator. The device's internal leak logging system was used to document mask leak. Trial was continued until either NIV was no longer deemed medically indicated by the attending team to a maximum of 24 hours. Arterial or venous blood gases collected during routine care of the patient were monitored to ensure adequacy of ventilation. Primary end-point was proportion of time under treatment with mask leak < 0.5 L/min (registering as 0 L/min by machine). Secondary end-points included ulceration score (National Advisory Panel on Pressure Ulcer Scale), mask fit (Likert-like 5 point scale, 1 = poor; 5 = excellent); and patient tolerance of the device (1 = poor; 5 = excellent). Additional collected data included: demographics (age, sex); morphologic parameters (height, weight, presence of dentures, facial hair); metabolic data (arterial or venous blood gas); vital signs (temperature, pulse, respirations, blood pressure, oxygen saturation); Ventilation parameters (mode, IPAP, EPAP) and fit and comfort (mask fit evaluated by RT, and patient tolerance of mask). ### Results The study was stopped after randomization of 16 patient (6 intervention, 10 control) due to results suggesting failure of the ARM to reduce leaks compared to the control device (Fig T1 & T2). There were no major adverse events during the trial related to the device. ![T1. Mask Leak - Control](https://i.imgur.com/PwNWAIe.png) ***Fig T1**. Distribution of mask leak flows for each patient with the control mask (N=10). 7/10 patients had leak of 0.5 L/min or less for at least half of the treatment duration* ![T2. Mask Leak - ARM](https://i.imgur.com/LMyFhO7.png) ***Fig T2**. Distribution of mask leak flows or each patient with the ARM (N=6). Only 2 of 6 patients had leaks of 0.5 L/min or less for more than half of the treatment duration*. ## Conclusion Despite Interim Order approval by Health Canada for clinical use during the pandemic, bench testing and clinical trial data suggest that the device as currently designed is likely ineffective at reducing aerosolization and poses risks of significant CO<sub>2</sub> rebreathing in the event of a ventilator or oxygen supply failure if used with limited supervision, as might be expected in remote nursing stations or in large hospitals during pandemic surge situations with strained staffing. We therefore recommend suspension of any pending device sales or distributions and revision of the design. To date no devices have been distributed. As such no recall of devices is required.