A Retrospective Analysis of COVID-19 ARDS Patients Successfully Extubated to High Flow Nasal Cannula Oxygen Therapy

Background Coronavirus disease- 2019 (COVID-19) related to acute respiratory distress syndrome (ARDS) caused by the highly infectious SARS-CoV-2 novel coronavirus is a major cause of death during the pandemic period. Here we aim to present the retrospective data analysis of extubation success to High Flow Nasal Cannula Oxygen Therapy (HFNO) in COVID-19 ARDS patients. Methods The data of 22 laboratory-conrmed COVID-19 ARDS patients who were extubated to HFNO therapy at an intensive care unit (ICU) were analyzed. Respiratory variables as well as demographic characteristics were collected on admission. The mechanical ventilation volumes and pressures together with blood gas measurements were recorded during the intubation period. HFNO ow rate, FiO 2 , and oxygenation variables were collected 5 consecutive days after extubation. The reintubation rate within the 5 days following planned extubation, duration of ICU stay, and mortality were recorded.


Introduction
High Flow Nasal Oxygen therapy (HFNO) is one of the newer methods of oxygenation commonly used in critical care during acute hypoxemic respiratory failure that can deliver heated and humidi ed gas up to 100% oxygen at a maximum ow of 60 L minute − 1 via nasal cannula. It has also been reported that HFNO can generate ow-dependent, low-level positive airway pressure, reduce airway resistance, and ush nasopharyngeal dead space [1].
In COVID-19 patients with acute respiratory failure, HFNO reduced the intubation rate compared to noninvasive ventilation (NIV) when used as initial support [2]. HFNO was shown to be superior to conventional oxygen treatment (COT) in reducing treatment failure when used as a primary support strategy and in reducing rates of extubation failure and reintubation when used after extubation [2].
However, there is an important concern that the high gas ow used might cause aerosol dispersion leading to the transmission of the virus into the environment. In vitro studies demonstrated that generation and dispersion of bio-aerosols via HFNO show a similar risk to standard oxygen masks [3].
Thinking about the advantages of HFNO in reducing the risk of intubation and the need for mechanical ventilation, it is not wise to discard this technique for the support of acute respiratory failure patients due to COVID-19. HFNO was widely used in our institution during the pandemic period both before and after invasive mechanical ventilation. In this retrospective data analysis, we aimed to report extubation success to HFNO and report the outcome data of COVID-19 ARDS patients.

Material And Methods
This retrospective study was conducted at Istanbul University Hospital after ethics approval was obtained from the hospital's local ethics committee (Approval number: 04/06/2020-89807). Twenty-two ARDS patients (18 years of age or older) with laboratory-con rmed COVID-19 infection who were admitted to the hospital's four ICUs between 18 March 2020 and 30 May 2020 were included in this analysis. The written informed consent from individual patients was waived due to the nature of the retrospective chart review. The directive for follow-up of COVID-19 patients was documented on Supplement 1. COVID-19 disease was de ned as a positive result on a reverse-transcriptase-polymerase-chain-reaction (RT-PCR) assay of a nasopharyngeal swab collected by the local hospital health authority. SARS-CoV-2 pneumonia was diagnosed according to World Health Organization guidance and those patients who required respiratory support or had oxygen saturation below 90% with a standard oxygen mask were admitted to the ICU. ARDS was de ned according to Berlin de nition [4]. Data was collected from available electronic medical records and patient les by the attendings responsible for the research facilities of the Department of Critical Care Medicine of the Istanbul Medical Faculty.
The demographic and clinical data including age, sex, admission disease severity scores [Acute Physiology and Chronic Health Evaluation II (APACHE II) and Sequential Organ Failure Assessment (SOFA) score], underlying comorbidities (chronic cardiac disease, hypertension, chronic pulmonary disease, diabetes mellitus, chronic renal failure, chronic liver disease, malignancies, cerebrovascular disease, autoimmune disease, and immunosuppressive state), duration between symptom initiation and ICU admission, and intubation period were recorded. PaO 2 /FiO 2 (arterial oxygen partial pressure/fractional inspired oxygen) ratio before intubation, days in mechanical ventilation were recorded. Blood gas analysis and respiratory parameters including inspiratory support pressure, positive end-expiratory pressure (PEEP), respiratory frequency, tidal volume (Vt), and frequency, as well as PaO 2 /FiO 2 ratio right before extubation, were added to the data chart.
Mechanical ventilation volumes, pressures, and blood gas analysis results were recorded during the intubation period.
The weaning protocol included daily screening for weaning readiness according to the respiratory and clinical criteria. Patients were extubated when they ful ll the criteria of extubation (low PEEP level (5-8 cm H 2 O), with no electrolyte disturbance, provided hemodynamic stability, interrupted sedation and followed up in spontaneous breathing in pressure support mode, good state of consciousness, received su cient Vt (at least 5 ml kg − 1 ), and had su cient cough re ex were evaluated in terms of sputum amount, character and viscosity, aspiration frequency of more than 2 hours, pain control were achieved, breath rate is < 30 / min, oxygen saturation (SpO 2 ) > 90%, PaO 2 > 60 mmHg, rapid shallow breathing index < 105). Patients were continuously treated by HFNO alone with a ow and FiO 2 adjusted to obtain adequate oxygenation, with a SpO 2 as measured by pulse oximeter of at least 92%. To provide su cient humidi cation, the temperature of the heated humidi er was set at 37°C. When there were no signs of respiratory failure (respiratory rate > 35 minute − 1 more than ve minutes, hypoxemia that SpO 2 < 90%, tachycardia that heart rate (HR) > 140 minute − 1 or 20% increase, bradycardia that 20% reduction in HR, hypertension that systolic blood pressure > 180 mmHg, hypotension that systolic blood pressure < 90mmHg, acidosis that pH < 7.32 and > 10 mmHg increase in PaCO 2 (arterial carbon dioxide partial pressure), consciousness changes that agitation, sweating or anxiety symptoms, cyanosis, ndings of increased breathing effort that accessory muscle use, stress symptoms on the face, increased breathlessness) after extubation, treatment was stopped and switched to standard oxygen.
HFNO ow rate, FiO 2 , and oxygenation variables including PaO 2 , PaCO 2 , SaO 2 (arterial oxygen saturation) were collected 5 consecutive days after extubation. The reintubation rate within the 48 hours and 5 days following extubation, ICU length of stay, and mortality were recorded. Data collection was stopped in those patients who were either switch to COT or invasive mechanical ventilation.

Statistical analysis
Data analysis was performed with SPSS (IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp.) One sample Kolmogorov-Smirnov test was performed to evaluate the distribution of the continuous variables. The categorical variables were presented as number and percentage values. The mean, standard deviation, minimum and maximum values of continuous variables were presented. Spearman correlation analysis was performed to evaluate the association between clinical features and pre-extubation mechanic ventilation volumes, pressures, and blood gas parameters. A p-value less than .05 was considered statistically signi cant.

Results
The demographic and clinical characteristics of the patients are summarized in Table 1

Discussion
The main nding of the present retrospective data is that high-risk ARDS COVID-19 patients can be successfully extubated to HFNO. Among the non-invasive modalities, high ow oxygen therapy offers many physiological bene ts which include improved oxygenation, decreased anatomical dead space, decreased metabolic demand of breathing, decreased production of carbon dioxide. Most importantly this technique serves up to superior comfort and improved work of breathing. In a small group of patients, the delivery of heated and humidi ed oxygen with high-ow nasal cannula was shown to be superior to high-ow oxygen via a non-rebreathing mask. Breathing frequency and inspiratory effort were reduced with HFNO compared with the non-rebreathing mask. HFNO therapy decreases post-extubation neuroventilatory drive and work of breathing in patients with chronic obstructive pulmonary disease [5]. We did not measure electrical diaphragmatic activity, but we think that HFNO treatment reduces the possibility of reintubation due to high ventilatory impulse and respiratory work in patients with extubated COVID-19 ARDS.
Many other studies showed that HFNO was superior to COT in reducing treatment failure when used as a primary support strategy and in reducing rates of extubation failure [6]. In a few trials, HFNO reduced the intubation rate compared to NIV when used as initial support but demonstrated no bene cial effects after extubation. Post extubation respiratory failure and reintubation rates were compared between HFNO and non-invasive ventilation in a group of high-risk patients. In this multicentric randomized clinical trial, HFNO offered many clinical advantages and proved that it is not inferior to NIV for preventing reintubation and post-extubation respiratory failure. A higher reintubation rate was reported (19%) with NIV most probably due to switching to COT after 24 hours [7,8]. Other data suggest that more prolonged high-ow conditioned oxygen therapy could improve outcomes in critically ill patients after extubation [9]. In a general population of critically ill patients randomized to receive either high-ow conditioned oxygen therapy or COT for 48 hours, Maggiore and colleagues found persistent improvement in oxygenation and comfort parameters and achieved a lower reintubation rate (3.8%) [10]. In another multicentric randomized trial including high-risk extubation failure patients, the reintubation rate was 18.2% with 48 hours HFNO treatment [11]. The reintubation rate was 10% (2/20) in our retrospective data which was similar to previous trials. We continued HFNO treatment for 48 hours after planned extubation after which they switch to standard oxygen therapy. Considering the high risk of COVID-19 ARDS patients for extubation, prolonging the HFNO therapy for at least 5 days could improve extubation success.
Several reports discussed if endotracheal intubation could be prevented by HFNO treatment in COVID-19 patients who presented with acute respiratory failure with moderate ARDS. Twelve randomized controlled trials provided low-certainty evidence that HFNO may reduce invasive ventilation in non-COVID-19 patients [2]. Results provided no support for differences in mortality or length of ICU stay. HFNO seems to have been seldomly used during the current COVID-19 pandemic in the western world. This is most probably due to the fear of risk of aerosolization and viral dispersion which might lead to infection transmission. However, the World Health Organization and other scienti c societies list HFNO among the possible options for ventilator support [12]. Four studies evaluating droplet dispersion and three evaluating aerosol generations and dispersion provided very low certainty evidence. Two simulation studies and a crossover study showed mixed ndings regarding the effect of HFNO on droplet dispersion. Although two simulation studies reported no associated increase in aerosol dispersion, one reported that higher ow rates were associated with increased regions of aerosol density [13][14][15][16][17][18][19][20][21]. However, in-vitro and clinical studies demonstrated that placing a simple surgical protection mask on patients signi cantly reduces dispersion distance [22]. Smoke simulation studies also demonstrated that dispersion with a ow rate of 60 L minute − 1 is similar to the one observed with a simple oxygen mask at 15 L minute − 1 [13,23]. We followed the same rule that all patients wore a facial mask during HFNO treatment and the mean ow rates were lower than 50 L minute − 1 in 5 days' follow-up after extubation which we believed that sustained minimum dispersion.
Our data evaluation has several limitations rst being its retrospective nature. Second, we did not have a control group so that we were not able to compare the data with other respiratory support systems. We haven't used any xed protocol in terms of time period after extubation. However, patients were switched to a standard oxygen mask when they ful ll the necessary clinical and respiratory criteria. Third, the number of patients might not be enough to come to any strong conclusion however we think that the rate of extubation success in our data of high risk of COVID-19 patients worth considering.

Conclusions
Among high-risk ARDS COVID-19 patients who have undergone extubation, HFNO therapy should be considered for preventing reintubation and post-extubation respiratory failure.