key: cord-0980673-q56bhpe6 authors: Stebbins, Julie; Saigal, Raveen; Hooper, Robbie; Shortland, Adam title: The effect of entraining oxygen at different locations in a noninvasive ventilator date: 2020-09-09 journal: Br J Anaesth DOI: 10.1016/j.bja.2020.09.006 sha: 1edd82f565f85015156d5402a04867628ba8ddfa doc_id: 980673 cord_uid: q56bhpe6 nan The effect of entraining oxygen at different locations The effect of entraining oxygen at different locations The effect of entraining oxygen at different locations The effect of entraining oxygen at different locations in a noninvasive ventilator in a noninvasive ventilator in a noninvasive ventilator in a noninvasive ventilator Editor -Guidance on using noninvasive ventilation produced by NHS England, 1 as well as the Association for Respiratory Technology and Physiology (ARTP) COVID Group 2 suggests that oxygen can be entrained into the breathing system at the patient end, directly into the heat and moisture exchange (HME) filter or through an oxygen entrainer. This is contrary to manufacturer guidance (for the Breas Vivo 2, the system in use at the Nightingale hospital), which recommends entraining the oxygen into the dedicated port at the back of the machine. The aim of this study was to determine whether entraining oxygen at the patient end or machine end of the breathing system caused a difference in delivered fractional oxygen (FiO 2 ) or pressure to the patient. This was done using continuous positive airway pressure (CPAP) and bi-level positive airway pressure (BiPAP) modes to also assess if this was dependent on ventilation mode. The following experiments took place in the Nightingale Hospital, London in an unused ward adjacent to the patient ward. The named experimenters were assisting clinical staff in a technical support role. Experiments were conducted largely at night when the ward activity was at its quietest. Consequently, the experimenters wore full personal protective equipment (PPE). They had limited access to measurement instrumentation that they may have used in a more standard setting. The Vivo 2 (Breas, Sweden) Noninvasive Ventilator was set up as if it were being used on a patient, including a heat and moisture exchange (HME) filter. In addition, a second filter (HEPA filter) was placed in series with the usual filter. The extra filter was added so that FiO 2 could be measured via a sampling line using a Penlon 465 anaesthetic machine (Penlon Ltd, Oxfordshire, UK). A test lung (Drager Ltd, Lubeck, Germany) was attached in place of a patient. Initially, oxygen was entrained through the dedicated port on the Vivo 2. Using the CPAP mode, pressure was set sequentially to 5, 10 and 15 cm H 2 O. For each CPAP pressure setting, oxygen flow rate was incrementally increased from 0 to 15 L min -1 (via a flow regulator attached to the piped oxygen supply) and FiO 2 was recorded. The whole process was then repeated with oxygen entrained directly into the HME filter. The experiment was then repeated with one experimenter depressing the test lung once every four seconds to simulate patient breathing. The Vivo 2 was then set to BiPAP mode, and the above procedure repeated. (IPAP of 12-14 cm H 2 O, EPAP of 8 cm H 2 O, respiratory rate of 16 breaths per min, Tinsp set to 1.5 s, and volume target set to 480 ml). Two different FiO 2 recordings were made as the readings fluctuated with each "breath". A high reading and a low reading were taken once these stabilised (after about 2 min). Finally, to investigate the effect on pressure, the patient breathing system was attached to the Penlon anaesthetic machine. The Noninvasive Ventilator was used to provide expiratory air in a CPAP mode. CPAP pressure was gradually increased and the Penlon machine was used to measure the pressure in the breathing system in real time. Using the CPAP mode and with oxygen entrained in the back of the machine, FiO 2 linearly increased with flow rate. At 5 cm H 2 O, FiO 2 plateaued at a flow rate of 4 L min -1 . The plateau occurred later with increasing CPAP pressure. When the oxygen was entrained directly into the HME filter at the patient end, FiO 2 increased to a maximum level (~95%) as soon as flow rate was initiated (0.5 L min -1 ) and plateaued at this point (Fig. 1) . The same result was achieved for all CPAP pressure levels. When the test lung was compressed to simulate breathing, the results fluctuated significantly with each simulated "breath", but the same pattern was observed with a much earlier plateau in FiO 2 when oxygen was entrained at the patient end compared to the Noninvasive Ventilator port. When the Noninvasive Ventilator was set to BiPAP mode, the pattern was similar to CPAP. When 2 L min -1 of oxygen was entrained at the filter, FiO 2 plateaued at its maximum level. There was a linear response with increasing flow rate when oxygen was entrained at the back of the machine, with a plateau at around 10 L min -1 . When measuring the effect on delivered pressure, entraining oxygen at the filter had minimal impact (1-2 cm H 2 O) compared to the machine end, even at flow rates of 15 L min -1 . These results indicate that entraining oxygen via the dedicated port on the Noninvasive Ventilator facilitates more controlled titration of delivered FiO 2 . This allows modification of FiO 2 depending on patient needs and thus potentially improves patient management. These results are in line with a previous study conducted on patients with chronic obstructive pulmonary disease 3 . The linear range was larger for CPAP pressures of 10 and 15 cm H 2 O. We believe that this is because to increase pressure (when oxygen flow rate is constant) more air is suctioned by the Noninvasive Ventilator. Therefore the relative percentage of oxygen is lower. This is important for clinicians to note as FiO 2 should be adjusted accordingly. These results are in line with findings from the ARTP COVID group, 2 however it should be noted that the FiO 2 values we achieved were significantly higher at comparable oxygen flow rates than the ones reported by ARTP. When the test lung was compressed to simulate breathing, FiO 2 fluctuated significantly if oxygen was delivered directly to the filter. In contrast, FiO 2 remained stable and predictable if oxygen was delivered to the port in the Noninvasive Ventilator. This suggests that FiO 2 is highly dependent on patient specific respiratory parameters when oxygen is delivered to the filter. The clinician has little control over this and cannot guarantee the FiO 2 being delivered. For this reason, delivering the FiO 2 via the port in the Noninvasive Ventilator seems preferable. There was minimal change in pressure to the patient when delivering oxygen directly to the filter compared to the Noninvasive Ventilator port. This suggests that there is an internal safety value within the Vivo2 Noninvasive Ventilator which ensures that pressures are limited within the closed breathing system. A limitation to this study is that we were unable to measure the results with the breathing system attached to a patient to comply with infection prevention measures. The actual FiO 2 delivered to the patient may vary due to the leakage of flow around the mask. These results suggest that oxygen should preferentially be entrained via the dedicated port in the Noninvasive Ventilator, as this allows greater control of delivered FiO 2 . It is difficult to deliver an FiO 2 < 90% when entraining via the filter, which may be too high for some patients. In addition, patient specific breathing mechanics make predicting FiO 2 uncertain when delivered via the filter. Our results given here can be used to estimate FiO 2 with a given CPAP, oxygen flow rate and mode of ventilation. However, different Noninvasive Ventilator machines and patient settings may result in different relationships, so we suggest creation of a look-up table for each set up to allow clinicians to set an estimated FiO 2 using a given flow rate and CPAP. This will only work however, when oxygen is supplied via the Noninvasive Ventilator port. Guidance for the role and use of non-invasive respiratory support in adult patients with COVID-19 (confirmed or suspected) The effect of entrainment site and inspiratory pressure on the delivery of oxygen therapy during non-invasive mechanical ventilation (NIMV) in acute COPD patients The authors declare no conflicts of interest.