Prone Positioning During High Flow Oxygen Therapy in Acute Hypoxemic Respiratory Failure

Background Nasal high flow oxygen (NHF) is a new and promising tool for oxygen therapy in critically ill patients: NHF allows accurate delivery of the set FiO2, anatomical dead space clearance due to a washout effect in the upper airways and provides a small, variable amount of positive end end-expiratory pressure. Different studies have investigated its safety and efficacy in several clinical setting and, recently, a randomized controlled trial showed that NHF, as compared to NIV, may reduce the intubation rate in severely hypoxemic patients with de novo acute respiratory failure (AHRF).

Prone positioning (PP) is a simple and cost-effective strategy that since 1980s has been used in mechanically ventilated patients with acute respiratory failure to treat oxygenation impairment. In the acute respiratory distress syndrome (ARDS) PP reduces intrapulmonary shunt (Qs/Qt) and enhances lung recruitment, modifying both lung ventilation (VA) and lung perfusion (Q) distribution, finally generating an improvement in VA/Q matching and reversing oxygenation impairment. In addition, the reduction of transpulmonary gradient (mean pulmonary arterial pressure-pulmonary artery occlusion pressure) due to a higher pulmonary arterial occlusion pressure allows pulmonary vascular recruitment, possibly lowering dead space fraction. Indeed, prone positioning sessions have been shown to reduce right ventricular afterload, increase cardiac index in subjects with preload reserve and reverse acute cor pulmonale in severe ARDS patients. Finally, a large randomized study detected a relevant survival benefit by prone positioning in patients with moderate to severe acute respiratory distress syndrome (ARDS) undergoing invasive mechanical ventilation and managed with the ARDS network peep-FiO2 table strategy.

Theoretically, PP benefits may concern also spontaneous breathing patients, in whom it could possibly contribute to the success of the noninvasive strategy. Data concerning PP application in such context are limited to small case series and a retrospective study involving patients undergoing NIV or oxygen therapy and suggesting good patients tolerance, significant improvement in oxygenation during the procedure and no increase in nurse workload. In particular, in the retrospective study by Pesenti and co. 43 procedures in 15 patients were analysed: three patients (20%) received mild sedation without dosing adjustments over the course of the study, median [IqR] duration of prone positioning was 3[2-4] hours and only in two cases (5%) pronation had to be interrupted after 30 minutes due to patient's discomfort, while no other respiratory or technical complications, such as displacement of indwelling catheters, facial oedema, pressure sores, pressure neuropathies, compression of nerves, and retinal vessels or vomiting, have been documented. Given its high tolerability, simplicity to use and positive effect on patients' comfort when compared both to NIV and to low flow oxygen delivered through a face mask, NHF may be the optimal strategy for the respiratory management of spontaneous breathing patients during prone positioning.

Interestingly, a recent electrical impedance tomography study on 20 healthy volunteers during NHF showed a more homogeneous dorsal-to-ventral ventilation pattern in PP, as compared to the supine position: notably, a dorsal-to-ventral more uniform air distribution may benefit hypoxemic patients, given that PaO2 response to oxygen supplementation can be fostered by more homogeneous regional lung inflation and better oxygenation has been shown to be a predictor of NHF treatment success.

The investigators designed a pilot feasibility study to assess the safety and efficacy of prone positioning in patients wit AHRF undergoing high flow oxygen therapy via nasal cannula.

Methods:

Setting: 21-bed medical-surgical ICU, Fondazione Policlinico Universitario "Agostino Gemelli", Rome, Italy..

Patients: adult hypoxemic (PaO2/FiO2 < 200 mmHg) non-hypercapnic patients with acute respiratory failure will be assessed for the enrolment. To validate the presence of the oxygenation criteria, each eligible patient, in the absence of exclusion criteria, will receive 15 minutes of heated and humidified 50% oxygen at a rate of 50 l/min via a non-rebreathing face mask. An ABG will be then collected and PaO2/FiO2 ratio computed: given the high flows used, actual FiO2 will be approximated to the set one. Patients with a P/F ratio<80 mmHg and/or a RR>40 bpm will be excluded for safety reasons.

Protocol Eligible patients will undergo NHF for 1 hour in the supine semi recumbent position (baseline, BL step). Oxygen therapy will be commenced at 60 l/min and flows will be decreased in case of intolerance. Humidification chamber (MR860, Fisher and Paykel healthcare) will be set at 37 °C or 31 °C according to patient's comfort. FiO2 will be titrated to obtain and SpO2 >92 % and <98 %.

Afterwards, the patient will be placed in the prone position for 2 hours (Prone step). During pronation FiO2 will be increased up to 80% and then gradually decreased to the baseline value within the first 10 minutes of prone positioning. Any further modifications in the NHF settings will be discouraged over the entire course of the study; nonetheless, if needed to achieve the SpO2 target, an increase in FiO2 will be allowed and recorded.

As suggested by Pesenti and co., in conscious and cooperative patients pronation will be performed by 2 operators and the attending physician, while in patients with impaired mobility up to 5 operators will be necessary. Evaluation of gastric residual volume through a nasogastric tube will reduce the risk of aspiration and application of appropriate skin protections will avoid pressure sores. In addition, careful application of appropriate cushions will enhance patient tolerance to the manoeuvre.

For safety reasons, enteral feeding, whether prescribed, will be interrupted 1 hour before prone positioning and re-established after the study ending.

After a 2-hour PP period, the patient will be rotated and will undergo 1 hour of NHF in the semi recumbent supine position (Supine step).

As already stated, FiO2 will be increased up to 80% for the procedure and then gradually decreased to the baseline value within the following 10 minutes.

In case of intolerance to PP, all clinical data will be collected and the patients will be promptly rotated in supine position. The reasons for prematurely unproning the patient will be accurately recorded. In case of sudden worsening of the oxygenation impairment or haemodynamics, 100% FiO2 will be set and the patient will be promptly positioned in the supine semi-recumbent position.

Measurements Patient's demographics will be collected at study entry: initials, age, sex, height, weight, BMI, cause of hospital and ICU admission, SAPS II, Apache, SOFA, date and time of ICU admission, date and time of enrolment, comorbidities, NYHA category prior to respiratory failure, body temperature, chest x-ray (jpeg images), chest CT scan (if available).

At study enrolment an electrical impedance tomography (EIT) belt will be placed around the thorax between the 5th or 6th parasternal intercostal space and connected to a dedicated device to record electrical impedance signals.

During the study, each patient will undergo a standard ICU monitoring: ECG, Invasive blood pressure, SpO2, respiratory rate, diuresis.

At the end of the BL period, and then on a hourly basis the following data will be collected:

• Respiratory rate, SpO2, pH, PaCO2, PaO2, SaO2, PaO2/FiO2;
• Heart Rate, arterial blood pressure;
• Dyspnoea, as defined by the Borg dyspnoea scale;
• Discomfort, as defined by a visual analogic scale (VAS) adapted to rate the procedural pain of ICU patients;
• End expiratory lung impedance (EELI), tidal volume distribution and global and regional lung dynamic strain (Change in lung impedence due to tidal volume/EELI). At the end of each step, ten-minute EIT signals will be recorded and offline reviewed using a dedicated software. Image acquisition rate will be 30 Hz. Lungs will be divided into four regions (ventral, mid-ventral, mid-dorsal and dorsal): the percentage of impedance variation related to tidal volume and the percentage EELI in the four regions as compared to the absolute values will be calculated.
• PTP of the esophageal pressure
• Esophageal pressure inspiratory swings
• Respiratory system, lung and chest wall mechanics, according to the method proposed by Mauri et al. (Am J Resp Crit Care Med 2017)
• Pendelluft phenomenon

As a safety tool the following adverse events will be assessed in each of the three study periods:

• The number of displacements of the nasal cannula
• The number of oxygen desaturations (SpO2 <90%)
• Episodes of haemodynamic instability (Systolic arterial pressure<80 mmHg or FC>120 BPM)
• Displacement of venous line (peripheral or central), if documented
• Displacement of arterial line, if documented
• Displacement of urinary catheter, if documented.

At the end of the study, the nurse in charge will be asked to anonymously rate the additional workload due to PP (in minutes) and to judge the feasibility and safety of the procedure using two analog scales ranging from 0 (completely unfeasible/unsafe) to 10 (totally feasible/safe).

Sample size Given the pilot, physiological and exploratory design of the study, a formal sample power analysis was not performed. It is planned to enrol 15 patients over a period of 12 months.