Characteristics and outcomes of patients receiving high flow nasal cannula therapy prior to mechanical ventilation in COVID-19 respiratory failure: A prospective observational study.

BACKGROUND: Treatment strategies for acute hypoxic respiratory failure secondary to coronavirus 2019 disease (COVID-19) had significant variation early in the pandemic. We sought to determine if patients treated with high flow nasal cannula (HFNC) prior to mechanical ventilation had differing outcomes compared to those treated only with conventional oxygen.
METHODS: This was a prospective, observational study of patients with COVID-19 admitted to a tertiary care medical center with a diagnosis of acute hypoxic respiratory failure. Adult patients with a positive polymerase chain reaction test for COVID-19 who required mechanical ventilation were included.
RESULTS: A total of 91 patients met the inclusion criteria for our study. The mean age was 68.4 years (standard deviation [SD] ± 12) and 58% were male. The mean initial partial pressure of arterial oxygen to fraction of inspired oxygen (PaO2/FiO2) ratio was 152 (SD ± 65) and was not significantly different between the HFNC group and the conventional oxygen strategy group (P = 0.95). The observed mortality rate was 30% in the HFNC group versus 52% in the conventional oxygen strategy group (P = 0.05). The multivariate odds ratio of mortality for patients on HFNC was 0.375 compared to a conventional oxygen strategy (95% confidence interval 0.122-1.151, P = 0.09).
CONCLUSIONS: While HFNC appears to be safe as the initial treatment strategy for COVID-19 associated respiratory failure, HFNC did not result in a statistically significant difference in mortality compared to a conventional oxygen strategy. Further studies are needed to confirm these findings. Copyright:

INTRODUCTION

Variation in the initial treatment strategies for coronavirus 2019 disease (COVID-19)-induced acute hypoxic respiratory failure evolved due to early reports of high mortality with mechanical ventilation,[1] projections of ventilator shortages,[2] and concerns regarding health-care worker safety related to aerosolization. In addition, patients with COVID-19 may show little distress or report minimal dyspnea despite significant hypoxia, which may delay intubation and affect patient outcomes. In a multicenter randomized trial of patients with non-COVID-related hypoxemic respiratory failure, treatment with high flow nasal cannula (HFNC) resulted in lower mortality and more ventilator-free days, but did not reduce intubation rates.[3] The outcomes of mechanically ventilated patients with COVID-19 initially treated with HFNC versus standard oxygen therapy are unknown.

The aim of this investigation was to determine if patients with COVID-19 initially treated with HFNC who went on to require mechanical ventilation had improved outcomes compared to patients who received conventional oxygen therapy. We hypothesized that the application of HFNC in a population of patients with COVID-19 respiratory failure would decrease overall mortality. In addition, we sought to determine if specific characteristics of patients who were placed on high flow nasal cannula would be predictive of mortality.

METHODS

We conducted a prospective, observational study of patients with COVID-19 admitted to a tertiary care medical center between March 2020 and May 2020. The study was approved by our institutional review board (approval number 20203054) and the need for informed consent was waived. Our study was performed according to STROBE guidelines. We included consecutive adult patients (age >18 years) with confirmed COVID-19 infection by polymerase chain reaction testing of a nasopharyngeal or tracheal aspirate sample (RealTime SARS-COV-2 assay, Abbott Laboratories, Des Plaines, Illinois, USA) presenting with acute hypoxic respiratory failure requiring mechanical ventilation. We excluded patients treated with noninvasive ventilation prior to intubation. Patient demographics, clinical variables, and outcomes data were recorded and entered into an online database (MS Excel). All outcomes data were censored at day 28.

Therapy with HFNC or mechanical ventilation for acute hypoxic respiratory failure was initiated based on clinician discretion and availability of HFNC equipment and negative pressure rooms. Patients treated with HFNC received therapy with heated and humidified oxygen with titrated liter flow rate from 20 to 60 liters per minute based on the patient's respiratory rate and work of breathing using the PM5200 Air-Oxygen blender (Precision Medical, Northampton, Pennsylvania, USA). The inspired fraction of oxygen (FiO2) was titrated from 21% to 100% to maintain an oxygen saturation >92%. Patients receiving conventional oxygen therapy received oxygen through a non-rebreather mask or a reservoir nasal cannula with oxygen flows of 10–15 liters per minute. Mechanical ventilation was initiated at the discretion of the treating providers. The management of ventilated patients was according to evidence-based guidelines[4] for acute respiratory distress syndrome, including low tidal volume ventilation[5] and prone positioning,[6] if necessary.

The continuous variables were tested using the Student's t-test if the values were normally distributed or Wilcoxon Rank-Sum test if the values were non-normally distributed. Levene's F test was used to assess for homoscedasticity. The categorical variables were tested using Fisher's exact test if the variables were dichotomous or Chi-Square test if the variable included more than two categories. A multivariate logistic regression model was used to determine the association of mortality with HFNC and conventional oxygen use. We did not include missing data points in our analysis. The statistical analysis for this study was generated using Statistical Analysis Software (SAS), version 9.4 for Windows (SAS Institute, Cary, North Carolina, USA).

RESULTS

A total of 92 patients with COVID-19 were included in the analysis. The baseline demographics are presented in Table 1. The mean age for the cohort was 68.4 years (standard deviation [SD] ± 12) and 58% were male. For all patients, the mean SOFA score was 7.3 (SD ± 2.8) and the mean initial partial pressure of arterial oxygen to fraction of inspired oxygen (PaO2/FiO2) ratio was 152 (SD ± 65). Thirty (32.6%) patients were treated with HFNC prior to mechanical ventilation. Patients who did not receive HFNC had more coronary artery disease and heart failure at baseline when compared to patients in the HFNC group. HFNC was provided for a median of 21 (interquartile range [IQR] 9–36) hours prior to mechanical ventilation compared to patients treated with conventional oxygen therapy who had a median time of 4 (IQR 1–11) hours prior to intubation. There was no difference in the initial PaO2/FiO2 ratio between the two groups. The mean SOFA score was significantly lower in the HFNC group compared to the conventional oxygen strategy group (6.6 vs. 7.7, P = 0.05).

Table 1

Characteristics of patients treated with and without high flow nasal cannula prior to intubation

Variable	HFNC (n=30)	No HFNC (n=62)	P
Age (years), mean±SD	68.6±12.5	68.3±11.9	0.91b
Race, n (%)
Caucasian	23 (76.7)	51 (82.3)	0.77c
African American	4 (13.3)	5 (8.1)
Hispanic	2 (6.7)	2 (3.2)
Asian	1 (3.3)	3 (4.8)
Native American	0	1 (1.6)
Gender (male), n (%)	21 (70)	37 (59.7)	0.37d
BMI, mean±SD	32.2±8.1	31.4±9.8	0.70b
Comorbidities, n (%)
CKD	6 (20)	14 (22.6)	1.00d
Hypertension	16 (53.3)	45 (72.6)	0.10d
Diabetes mellitus	9 (30)	27 (43.6)	0.26d
CAD	1 (3.3)	13 (21.0)	0.03d
CHF	0	8 (12.9)	0.05d
Cirrhosis	0	1 (1.6)	1.00d
Cancer	4 (13.3)	10 (16.1)	1.00d
COPD	6 (20)	6 (9.7)	0.20d
Asthma	1 (3.3)	5 (8.1)	0.66d
Dementia	2 (6.7)	11 (17.7)	0.21d
SOFA, mean±SD	6.6±2.2	7.7±3	0.05
Admission source, n (%)
Home	15 (50)	19 (30.7)	0.20c
Nursing home	6 (20)	17 (27.4)
Outside hospital transfer	9 (30)	26 (41.9)
Initial PaO2/FiO2, mean±SDa	152±62	153±67	0.95b
Hours on>6L O2 prior to intubation, median (IQRe)	20.5 (9-36)	4 (1-11)	<0.0001
Prone ventilation, n (%)	23 (76.7)	31 (50)	0.02d
Inhaled pulmonary vasodilator, n (%)	2 (6.7)	9 (14.5)	0.49d
Laboratory studies (IQRe)
Ferritin, median	1859.5 (1235-2486)	1735.5 (884-3892)	0.10f
D-dimer, median	2000 (1069-2000)	2000 (760-2000)	0.67f
LDH, median	522 (419-661)	569 (425-748)	0.40f
IL-6, median	16 (9-39)	50 (7-111)	0.28f
COVID treatments, n (%)
Convalescent plasma	1 (3.3)	3 (4.8)	1.00d
Hydroxychloroquine	17 (56.7)	39 (62.9)	0.65d
Steroids	24 (80)	36 (58.1)	0.06d
Tocilizumab	7 (23.3)	12 (19.4)	0.78d
Statin	25 (83.3)	50 (80.7)	1.00d
Azithromycin	7 (23.3)	29 (46.8)	0.04d

bStudent’s t-test, cFisher’s exact test, dIQR, eSigned rank Wilcoxon test, fChi-square test. BMI: Body mass index, CKD: Chronic kidney disease, CAD: Coronary artery disease, CHF: Congestive heart failure, COPD: Chronic obstructive pulmonary disease, LDH: Lactate dehydrogenase, IL-6: Interleukin 6, IQR: Interquartile range, SD: Standard deviation, SOFA: Sequential organ failure assessment, COVID: Coronavirus disease, HFNC: High flow nasal cannula

The observed mortality rate was 30% in the HFNC group versus 52% in the conventional oxygen strategy group (P = 0.05) [Table 2]. Although not significant, multivariate logistic regression analysis suggested a trend toward lower mortality with treatment with HFNC (odds ratio 0.38, 95% confidence interval 0.12–1.15, P = 0.09) after controlling for other predictors of mortality [Table 3]. There was a trend toward a higher mean intensive care unit (ICU) length of stay and more days on a ventilator in the HFNC; however, these differences were not statistically significant.

Table 2

Outcomes of patients with and without high flow nasal cannula prior to intubation

Variable	HFNC (n=30)	No HFNC (n=61)	P
ICU LOS, mean±SD	18±8.2	14.5±9.5	0.08b
ICU free days, median (IQRc)	0.5 (0-11)	0 (0-9)	0.18d
Ventilator days, mean±SD	17.2±8.4	13.6±9.5	0.08b
Ventilator free days, median (IQRd)	1.5 (0-14)	0 (0-10)	0.41d
Successfully extubated, n (%)	15 (50)	22 (35.5)	0.26e
Disposition, n (%)
Discharged home	2 (6.7)	5 (8.1)	0.15f
Skilled nursing facility	5 (16.7)	5 (8.1)
Died	9 (30)	33 (53.2)
Still in hospital	14 (46.7)	53.2 (30.6)
Mortality, n (%)	9 (30)	33 (53.2)	0.05

bStudent’s t-test, cIQR, dSigned rank Wilcoxon test, eFisher’s exact test, fChi-square test. LOS: Length of stay, IQR: Interquartile range, SD: Standard deviation, ICU: Intensive care unit, HFNC: High flow nasal cannula

Table 3

Multivariate logistic regression to predict mortality

Covariates	OR (95% CI)	Correlation P
HFNC	0.375 (0.122-1.151)	0.0866
Prone	0.888 (0.336-2.342)	0.8096
Azithromycin	0.547 (0.195-1.539)	0.2533
SOFA	1.420 (1.138-1.773)	0.0020
Hours on>6L O2 prior to intubation	1.004 (0.988-1.021)	0.6099

OR: Odd ratio, CI: Confidence interval, HFNC: High flow nasal cannula, SOFA: Sequential organ failure assessment

DISCUSSION

In our analysis, we found that patients who were treated with HFNC as their initial treatment strategy for COVID-19 associated respiratory failure who subsequently required mechanical ventilation had a lower observed mortality rate compared to patients who underwent conventional oxygen therapy prior to intubation despite similar initial PaO2/FiO2 ratios. However, after controlling for other predictors, our multivariate logistic regression model did not find a difference in mortality, which may be due to our small sample size. We did identify a trend toward longer ICU length of stay in the HFNC group, although this finding may be related to the higher survival rate.

Initial treatment strategies for hypoxic respiratory failure secondary to COVID-19 varied among clinicians early in the pandemic.[7] Reports of increased mortality among ventilated patients may have caused some to adopt a noninvasive management strategy in an attempt to avoid or delay intubation, resulting in a relatively high rate of HFNC utilization. In contrast, other centers almost entirely avoided the use of HFNC due to concerns about health-care worker exposure related to aerosol generation. As a result, the reported use of HFNC for critically ill patients with COVID-19 vary widely, ranging from as low as 4.8% to as high as 63%.[89] A previous multicenter, randomized trial of patients with acute hypoxic respiratory failure due to non-COVID related illness (primarily pneumonia) treated with HFNC found a lower mortality rate in the ICU and at day 90.[3] However, there is not significant data regarding the outcomes of patients with COVID-19 treated with HFNC.[10] The Surviving Sepsis Campaign provided a weak recommendation in favor of the use of HFNC over conventional oxygen therapy in patients with acute hypoxic respiratory failure due to COVID-19.[11] This recommendation was based primarily on literature showing improved outcomes in trials conducted in patients without COVID-19. Our data suggest that the use of HFNC is a safe intervention in hypoxic respiratory failure due to COVID-19 when compared to the use of conventional oxygen therapy.

Retrospective studies have also demonstrated that HFNC does not appear to confer an increased risk of contracting COVID-19 over conventional oxygen therapy for health-care providers treating patients with COVID-19 respiratory failure. HFNC presented a contamination risk similar to that of conventional oxygen in studies evaluating bacterial environmental contamination.[12] Further evidence in patients with SARS demonstrates that HFNC did not present an increased risk of developing disease.[13] All patients receiving HFNC in our study were treated in negative pressure rooms.

Potential limitations should be considered when interpreting the results of this investigation. Our data were collected at a single center, which may limit generalizability, due to the small sample size. Given our study was observational, there is potential for selection bias as the choice of intervention was at the discretion of the treating provider. Although there was no difference in the mean PaO2/FiO2 ratio between the two groups, the higher mean SOFA score in the conventional oxygen strategy group raises the possibility that patients with more significant multisystem illness may have been evaluated with a lower threshold to utilize mechanical ventilation sooner. Patients treated with HFNC could only receive therapy in a negative pressure room and thus the decision to utilize conventional oxygen therapy may have been biased by room availability in a small number of circumstances. In addition, patients with high oxygen requirements may have been brought to the ICU in anticipation of further worsening, which may have increased ICU length of stay in the HFNC group. Finally, differences in baseline characteristics between the two groups may have contributed to the observed mortality difference.

CONCLUSIONS

While HFNC appears to be safe as the initial treatment strategy for COVID-19 associated respiratory failure, HFNC did not result in a statistically significant difference in mortality compared to a conventional oxygen strategy. Larger, multicenter studies are needed to confirm our findings.

Research Quality and ethics statement

This study was approved by the Institutional Review Board / Ethics Committee at Lahey Hospital and Medical Center (Approval # 20203054). The authors followed the applicable EQUATOR Network (http://www.equator-network.org/) guidelines, specifically the STROBE Guidelines, during the conduct of this research project.

Financial support and sponsorship

This research was conducted without funding or sponsorship.

Conflicts of interest

The authors report no conflicts of interest.