Respiratory Medicine

Non-rebreather mask and low-flow nasal cannula vs high-flow nasal cannula in severe COVID-19 pneumonia in the emergency department

a b s t r a c t

Background: To assess the effectiveness of non-rebreather mask combined with low-flow nasal cannula (NRB + NC) compared to high-flow nasal cannula in improving oxygenation in patients with COVID-19-related hypoxemic respiratory failure (HRF).

Methods: This retrospective study was conducted in emergency departments of two tertiary hospitals from June 1 to August 31, 2021. Consecutive patients aged >18 years admitted for COVID-19-related HRF (World Health Organization criteria: confirmed COVID-19 pneumonia with respiratory rate > 30 breaths/min, severe respiratory distress, or peripheral oxygen saturation < 90% on room air) requiring NRB + NC or HFNC were screened for en- rollment. Primary outcome was improvement of partial pressure arterial oxygen (PaO2) at two hours. Secondary outcomes were Intubation rate, ventilator-free days, hospital length of stay, and 28-day mortality. Data were ana- lyzed using linear regression with inverse probability of treatment weighting (IPTW) based on propensity score. Results: Among the 110 patients recruited, 52 (47.3%) were treated with NRB + NC, and 58 (52.7%) with HFNC. There were significant improvements in patients’ PaO2, PaO2/FIO2 ratio, and respiratory rate two hours after the initiation of NRB + NC and HFNC. Comparing the two groups, after IPTW adjustment, there were no statistically significant differences in PaO2 improvement (adjusted mean ratio [MR] 2.81; 95% CI -5.82 to 11.43; p = .524), in- tubation rate (adjusted OR 1.76; 95% CI 0.44 to 6.92; p = .423), ventilator-free days (adjusted MR 0.00; 95% CI -8.84 to 8.85; p = .999), hospital length of stay (adjusted MR 3.04; 95% CI -2.62 to 8.69; p = .293), and 28-day mortality (adjusted OR 0.68; 95% CI 0.15 to 2.98; p = .608).

Conclusion: HFNC may be beneficial in COVID-19 HRF. NRB + NC is a viable alternative, especially in resource- limited settings, given similar improvement in oxygenation at two hours, and no significant differences in long- term outcomes. The effectiveness of NRB + NC needs to be investigated by a powered randomized controlled trial.

(C) 2022

  1. Introduction

More than 460 million people have been infected with corona- virus disease 2019 , killing 6 million during the global pandemic [1]. Early in the pandemic, many hospitals received

* Corresponding author.

E-mail addresses: [email protected] (M.K. Mohd Kamil), [email protected] (K.P. Yuen Yoong), [email protected]

(A.M. Noor Azhar), [email protected] (A. Bustam), [email protected] (M.H. Md Yusuf), [email protected] (A. Zambri), [email protected] (A.Z. Ahmad Zahedi).

COVID-19 patients with hypoxemic respiratory failure (HRF) requir- ing supplemental oxygen and ventilator support [2,3]. In a study of 5700 patients with COVID-19 hospitalized in the United States be- tween March and April 2020, 27.8% received supplemental oxygen, 14.2% admitted in the intensive care unit (ICU), and 12.2% required endotracheal intubation and mechanical ventilation [4]. ICUs have been overwhelmed and high mortality was observed in COVID-19 patients requiring invasive ventilation [4]. The initial respiratory treatment modalities for these patients are widely debated, and dif- ferent strategies including non-invasive oxygen therapy have yielded variable outcomes [5,6].

0735-6757/(C) 2022

Non-invasive respiratory support such as High-flow nasal cannula is an attractive strategy for avoiding invasive ventilation. HFNC is simple to use and has physiological benefits [7-10]. In previ- ous studies, HFNC reduced the need for endotracheal intubation in HRF due to chronic obstructive pulmonary disease, cardiogenic pul- monary oedema, community-acquired pneumonia, and COVID-19 [7,10-13]. However, there were no differences in mortality rates, ICU admission, or length of stay [7]. Many Low- and middle-income countries are struggling with scarcity of Healthcare resources, partic- ularly oxygen supply amid a devastating COVID-19 surge [14]. HFNC devices are dependent on the wall oxygen system and consume 5 to 10 times the amount of oxygen that a mechanical ventilator does [15]. However, the oxygen pipes and vaporizers in many older hospi- tals are unable to accommodate the higher flow needs due to building structural limitations. conventional oxygen therapy (COT) can be delivered using portable Oxygen cylinders to lessen the demand for wall oxygen [15].

Current trials do not provide definitive evidence to recommend the use of HFNC in COVID-19 with HRF [16-18]. In resource-limited settings, particularly in the emergency department (ED), COT may offer a viable option for treating severe COVID-19 infection [8,19]. Problems arise in patients with severe HRF who require higher con- centrations of oxygen that are not met by Non-rebreather mask alone. Previous reports in India have demonstrated an im- provement in oxygenation with the usage of combined NRB and low-flow nasal cannula (NRB + NC) therapy for COVID-19-related HRF [20]. A retrospective study of 54 ICU patients found that NRB + NC yielded comparable results to HFNC in terms of mortality, intuba- tion rate, and length of ICU and hospital stay [21]. However, this study recruited a small sample size in an ICU setting. It is uncertain if the re- sults of this study would be applicable in the ED.

This retrospective, bicenter, observational study was conducted in COVID-19-related HRF patients to compare Arterial oxygenation, intu- bation rate, invasive ventilation-free days, hospital length of stay, and 28-day mortality between NRB + NC and HFNC treatment groups.

  1. Methods
    1. Study design and setting

This retrospective, observational study was conducted in the ED of two tertiary hospitals in Malaysia. The Medical Research Ethics and the National Medical Research Register approved the study protocol (MREC ID: 2021819-10491, NMRR ID: 21-02094-TYT) and waived the

requirement for informed consent. This study was conducted in accor- dance with Declaration of Helsinki.

    1. Selection of participants

All consecutive adult patients admitted in the ED between June 1 to August 31, 2021 for severe COVID-19 pneumonia were screened for en- rollment. Severe COVID-19 pneumonia was based on the World Health Organization (WHO) criteria: clinical signs of pneumonia (fever, cough, dyspnea) plus one of the following; respiratory rate > 30 breaths/min, severe respiratory distress, or peripheral oxygen saturation (SpO2)

< 90% on room air [22]. Other inclusion criteria were patients aged 18 years and older requiring NRB + NC or HFNC. COVID-19 infection was confirmed via reverse transcriptase-polymerase chain reaction (RT-PCR) assay.

Patients were excluded if emergent invasive ventilation was re- quired upon presentation or within two hours of initiation of oxygen therapy. Other exclusion criteria were cross-treatment between NRB

+ NC and HFNC in ED, acute exacerbation of chronic pulmonary dis- eases, moderate to severe heart failure (New York Heart Association class >=3 or left ventricular ejection fraction <40%), end-stage renal dis- ease, and pregnancy.

    1. Measurements

Data extracted from patients’ medical records were age, gender, co- morbidities, vaccination status, vital signs, chest radiograph findings, ar- terial blood gas (ABG), lactate, C-reactive protein, ferritin, D-dimer, and treatment provided. The decision to initiate oxygen therapy was at the discretion of the attending physician and based on the availability of the HFNC device and the oxygen capacity in the ED. In the NRB + NC group, patients received oxygen via 15 L/min of non-rebreather mask and 5 L/min of nasal cannula with fraction of inspired oxygen (FIO2) determined as 1.0. In the HFNC group, patients received oxygen via AIRVO 2 Optiflow (Fisher Paykel, New Zealand) with the flow rate and FIO2 determined by the treating physician. The need for invasive ventilation was determined two hours after the initiation of oxygen therapy based on clinical parameters and ratio of oxygen saturation (SpO2/FIO2) to respiratory rate (ROX) as per standard institutional protocol. Partial pressure of arterial oxygen (PaO2) to FIO2 ratio (PFR), Sequential Organ Failure Assessment , ROX, and APACHE II scores were calculated from the recorded parameters.

Treatment included Prone positioning and concomitant medical therapies such as steroids and other immunomodulatory agents, antibi- otics, antiviral agents, and vasopressors as determined by the treating physicians. Patients were followed-up for 28 days.

Data were collected retrospectively from the respective hospital’s medical records. The International Classification of Disease (ICD-10- CM) code U07.1 was used to extract patient records for “Confirmed COVID-19 virus identified”. Data abstraction was performed by two ex- perienced medical doctors who had undergone training on the study protocol to minimize inter-rater variability. The data abstractors were not blinded to the Study objectives and hypotheses. Patients’ clinical and biochemical characteristics were recorded on the data-abstraction forms. Three certified emergency physicians (KPYY, AMNA, and AB) val- idated that all participants included in the study fulfilled the definition for severe COVID-19 pneumonia according to the WHO criteria.

    1. Outcomes

The primary outcome of the study was the improvement of PaO2 after receiving NRB + NC or HFNC at two hours. Secondary outcomes were intubation rate, ventilator-free days, length of hospital stay, 28- day mortality, improvement in SpO2, PFR, and respiratory rate, and number of days to intubation or to de-escalation of oxygen therapy. Ventilator-free days was defined as the number of days the patients were liberated from mechanical ventilation. If a patient died within the 28 days, the ventilator-free days was documented as 0.

The sample size was calculated using G*power version with an effect size of 0.5 and ?-error probability of 0.05. The effect size of

0.5 was used since there were no previous comparable studies. For a study power of 0.80, the total sample size required was 102 including a dropout rate of 10%.

    1. Statistical analysis

The results were analyzed using Statistical Package for the Social Sci- ences (SPSS) version 26 (IBM Crop, Armonk, NY) and R Project for Sta- tistical Computing (version 4.0.4). Descriptive statistics were expressed as frequencies (percentages), mean (standard deviation), or median (interquartile range). Categorical variables were analyzed using Pearson’s chi-square test or Fisher’s exact test when appropriate. Test of normality was determined with the Shapiro-Wilk test for continuous variables. Normally distributed continuous data were analyzed using Student’s t-test and reported as mean differences and mean ratio (MR), whereas non-parametric data was analyzed with Mann- Whitney U test and reported as odds ratio (OR). The Paired Samples t-test and Wilcoxon Signed-Rank Test were used to calculate the

differences in physiological variables at baseline and two hours of administration of oxygen therapy.

In the sensitivity analysis, heterogeneity in the NRB + NC and HFNC groups were adjusted with propensity score using inverse probability of treatment weighting (IPTW) due to the small cohort in both groups. The data between NRB + NC and HFNC groups were heterogeneous in vac- cination status, heart rate, respiratory rate, SpO2, pH, partial pressure of carbon dioxide (PaCO2), bicarbonate (HCO), serum lactate, SOFA, and APACHE II scores, and thus were selected as covariates for the IPTW. The IPTW method creates a pseudo-population where the weighted data can mitigate the covariate bias. Average treatment effects were calcu- lated using 1/propensity score for NRB + NC group and inverse of (1-propensity score) for HFNC group [23]. Generalized linear model and regression analyses were used to analyze the adjusted outcomes. The survival analysis and cumulative intubation rates were plotted as IPTW-adjusted Kaplan-Meier curves using Cox proportional hazards regression model. The findings were considered statistically significant if the p < .05.


  1. Results
    1. Characteristics of study subjects

Out of the 531 severe COVID-19 pneumonia patients admitted to the 2 participating EDs between June 1 and August 31, 2021, 110 met the inclusion criteria and were included in the analysis (Fig. 1). Fifty-two patients (47.3%) were in the NRB + NC group, and 58 (52.7%) were in the HFNC group. No patients were lost to follow-up and no missing data was reported. There was no crossover of patients between the two treatment arms. In the HFNC group, the mean initial flow rate and FIO2 were 58.3 L/min (95% CI 57.3 to 59.3 L/min) and 0.58 (95% CI 0.57 to 0.59). Mean ROX indices at two hours for NRB + NC and HFNC were 2.73 (95% CI 2.11 to 4.08) and 6.21 (95% CI 4.04 to 8.44), respec-

tively. Two patients from each group had a do-not-intubate (DNI) order. Baseline characteristics of the patients are summarized in Table 1. The physiologic variables at baseline and two hours of applica- tion of oxygen therapy are shown in Table 2 and Fig. 2.

Image of Fig. 1

Fig. 1. Selection of study participants.

Baseline characteristics of patients, according to study group.


No. (percentage, %)


Non-rebreather mask and low-flow

High-flow nasal

nasal cannula (n = 52)

cannula (n = 58)

Age, median (IQR), years

48 (36-60)

53 (41-63)





33 (63.5)

38 (65.5)


19 (36.5)

20 (34.5)

Relevant comorbidities Type 2 diabetes

19 (36.5)

16 (27.6)



29 (55.8)

26 (44.8)



28 (53.8)

9 (15.5)



1 (1.9)

4 (6.9)


Ischaemic Heart Disease

1 (1.9)

4 (6.9)


Congestive cardiac failure

1 (1.9)



Chronic Kidney Disease

1 (1.9)



Cerebrovascular disease


2 (3.4)


bronchial asthma


4 (6.9)


History of cancer


Breast cancer

1 (1.9)

1 (1.7)

Vaccination status



28 (53.8)

51 (87.9)

dPartially vaccinated

22 (42.3)

6 (10.3)


7 (13.4)

1 (1.7)


15 (28.9)

5 (8.6)

eCompletely vaccinated

2 (3.8)

1 (1.7)


1 (1.9)



1 (1.9)

1 (1.7)

Blood pressure, median (IQR) mmHg Systolic

127 (88-139)

129 (111-135)



74 (55-82)

77 (68-82)


Mean arterial pressure

92 (65-101)

94 (83-98)


Heart rate, mean (SD) beats/min

106 (12)

98 (17)


Respiratory rate, median (IQR) breaths/min

38 (37-39)

30 (28-36)


Temperature, median (IQR), ?C

38.0 (37.8-39.0)



Oxygen saturation on non-rebreather 15 L/min, SpO2 median (IQR) SpO2

82 (80-84)

92 (87-95)


Glasgow Coma Scale (GCS)




arterial blood gases on non-rebreather 15 L/min, median (IQR)


7.30 (7.26-7.36)

7.48 (7.45-7.51)


PaO2, mm Hg

65.7 (61.0-72.4)

64.5 (54.7-72.8)


PaCO2, mm Hg

31.7 (28.9-36.6)

28.9 (26.1-31.8)


HCO, mmol/L

16.9 (15.4-18.7)

23.8 (21.6-25.5)


Base deficit, mmol/L

2.7 (1.7-3.6)

1.7 (0.6-3.7)


PaO2 / FIO2 ratio

66 (61-72)

69 (58-83)


Serum lactate, median (IQR) mmol/L

3.00 (2.63-3.48)

1.49 (1.17-2.11)


C-reactive protein, median (IQR) mg/L

206.5 (167.0-312.5)

130.1 (95.7-165.1)


Serum ferritin, median (IQR), ug/L

467 (339-736)

1075 (637-2495)


Positive D-dimer (qualitative)

16 (30.8)

11 (19.0)


bSOFA score, median (IQR)

3 (2-6)

2 (2-3)


cAPACHE II score, median (IQR)

17 (13-22)

11 (8-13)


A-a gradient, median (IQR)

608 (598-613)

608 (594-616)


Bilateral infiltrates on chest x-ray

52 (100)

58 (100)

Concomitant medications


52 (100)

58 (100)


52 (100)

58 (100)


51 (98.1)

47 (81.0)



8 (15.4)

12 (20.7)



3 (5.8)

15 (25.9)



2 (3.8)

7 (12.1)


Prone position

46 (88.5)

12 (20.7)


Vasopressor / inotropic support


Single (Noradrenaline)

20 (38.5)

5 (8.6)


Abbreviations: FIO2, fraction of inspired oxygen; IQR, interquartile range; PaCO2, partial pressure of carbon dioxide; PaO2, partial pressure of arterial oxygen; SD, standard deviation; SOFA, Sequential Organ Failure Assessment; SpO2, peripheral arterial oxygenation.

a For qualitative variables, p-values refer to the Chi-square test or Fisher exact test, whereas for quantitative variables, p-values tabulated from the student’s t-test or Mann-Whitney U test.

b SOFA score was calculated from 6 variables at enrollment. Scores range from 0 to 24, with higher scores indicating more severe disease.

c APACHE II score was calculated from 12 variables at enrollment. Scores range from 0 to 71, with higher scores indicating more severe disease and higher mortality.

d Patients who received a COVID-19 vaccine injection or received two injections but <14 days from date of ED admission.

e Patients who received two injections with at least 14 days from date of admission.

    1. Main results
      1. Primary outcome

The median improvement of PaO2 at two hours after application of oxygen therapy was 13.0 mmHg (IQR 7,18) in the

NRB + NC group and 4.0 mmHg (IQR -6.4, 17.0) in the HFNC group (p = .009). The mean difference and MR between the NRB + NC and HFNC groups were 6.0 mmHg (95% CI 0 to 12.0) and 5.61 mmHg (95% CI -0.59 to 11.81), respectively

(Table 3).

Table 2

Physiologic variables at baseline and two hours of application of oxygen therapy, according to study group.


Two hours



Two hours


Oxygen saturation, median (IQR), SpO2 (%)a

82 (80-84)

90 (88-94)


92 (87-95)

93 (89-95)


Respiratory rate, median (IQR) breaths/mina

38 (37-39)

35 (32-36)


30 (28-36)

25 (23-30)


Heart rate, mean (SD) beats/minb

106 (12)

98 (9)


98 (17)

89 (14)


Systolic blood pressure, median (IQR), mm Hga

127 (88-139)



129 (111-135)

127 (115-138)


Diastolic blood pressure, median (IQR), mm Hga

74 (55-82)

75 (68-82)


77 (68-82)

71 (60-78)


Mean arterial pressure, median (IQR), mm Hga

92 (65-101)

91 (86-98)


94 (83-98)

88 (82-98)


arterial pH, median (IQR)a

7.30 (7.26-7.36)

7.32 (7.28-7.35)


7.48 (7.45-7.51)

7.45 (7.41-7.48)


PaO2, median (IQR), mm Hga

65.7 (61.0-72.4)

80.2 (72.2-84.0)


64.5 (54.7-72.8)

69.1 (57.6-80.2)


PaCO2, median (IQR) mm Hga

31.7 (28.9-36.6)

34.0 (30.3-38.0)


28.9 (26.1-31.8)

32.6 (29.0-35.8)


HCO, median (IQR), mmol/La

16.9 (15.4-18.7)

17.8 (15.3-20.0)


23.8 (21.6-25.5)

23.6 (21.4-26.1)


PaO2 / FIO2 ratio (PFR), median (IQR)a

66 (61-72)

80 (72-84)


69 (58-83)

118 (105-141)


Serum lactate, median (IQR) mmol/La

3.00 (2.63-3.48)

2.35 (1.70-3.08)


1.49 (1.17-2.11)

1.20 (1.10-1.58)


Variable Non-rebreather mask and low-flow nasal cannula High-flow nasal cannula


Abbreviations: HCO, bicarbonate; IQR, interquartile range; PaO , partial pressure of arterial oxygen; PaCO , partial pressure of carbon dioxide; SD, standard deviation; SpO , peripheral

3 2 2 2

arterial oxygenation.

a Wilcoxon Signed-Rank Test.

b Paired-Samples t-Test.

      1. Secondary outcomes

The mean improvement of SpO2 at two hours in the NRB + NC group vs HFNC group were 8% vs 1% (mean difference 6%; 95% CI 4 to 8; p <= 0.001). The median improvement of PFR at two hours in the NRB + NC group vs HFNC group were 13 vs 49 (mean difference - 36; 95% CI -47 to -25; p <= 0.001). The reduction of respiratory rate in the NRB

+ NC group vs HFNC group were 3 vs 5 breaths/min (mean difference

– 2; 95% CI -4 to 0; p = .038).

The rates of endotracheal intubation within 28 days in the NRB + NC and HFNC groups were 53.8% vs 37.9% (OR 1.91; 95% CI 0.89 to 4.09; p =

.094). The median days to intubation in the NRB + NC and HFNC groups were 0 vs 3 days (mean difference – 4; 95% CI -7 to -1; p <= 0.001). The median ventilator-free days in the NRB + NC group compared to the HFNC group were 23 vs 28 days (mean difference - 2; 95% CI -6 to 3;

p = .145).

The median length of hospital stay in the NRB + NC and HFNC groups were 16 vs 13 days (mean difference 0; 95% CI -3 to 3; p =

.561). The 28-day mortality rates in the NRBC+NC vs HFNC groups were 28.8% vs 27.6% (OR 1.06; 95% CI 0.46 to 2.44; p = .883)

(Table 3).

      1. Sensitivity analyses

Fig. 3 shows the covariates before and after IPTW adjustment. After the adjustment, there were no differences between NRB + NC and HFNC groups in improvement of PaO2 (adjusted MR 2.81; 95% CI -5.82 to 11.43; p = .524), requirement for intubation (adjusted OR 1.76; 95% CI 0.44 to 6.92; p = .423), ventilator-free days (adjusted MR 0.00; 95% CI -8.84 to 8.85; p = .999), hospital length of stay (adjusted MR 3.04; 95% CI -2.62 to 8.69; p = .293), and 28-day mortality (adjusted OR 0.68; 95% CI 0.15 to 2.98; p = .608).

Image of Fig. 2

Fig. 2. Physiologic variables at baseline and two hours of application of oxygen therapy.

















Image of Fig. 3

Fig. 3. Standardized mean differences of covariates for IPTW adjustment.

ATE weighted

Adjusted odds ratio / mean ratio (95% CI)

Odds ratio / mean ratio (95% CI)


1.11 (1.03 to 1.19)

5.61 (-0.59 to 11.81)



1.09 (0.93 to 1.27)

2.81 (-5.82 to 11.43)

-1.38 (-3.13 to 0.38)

6.18 (4.21 to 8.14)

0.64 (0.59 to 0.69)

-36.08 (-46.79 to -25.38)

7.31 (5.73 to 8.89)

-1.85 (-3.64 to 0.06)

1.91 (0.89 to 4.09)

– 0.41 (-0.63 to – 0.25)

-1.51 (-6.12 to 3.09)

-0.37 (-0.58 to -0.23)











2.40 (-1.27 to 6.08)

4.80 (-0.04 to 9.63)

0.64 (0.54 to 0.75)

-42.09 (-56.98 to -27.20)

3.50 (0.98 to 6.03)

-1.35 (-2.78 to 0.09)

1.76 (0.44 to 6.92)

0.60 (0.32 to 1.12)

0.00 (-8.84 to 8.85)

0.54 (0.29 to 1.00)

1.06 (0.46 to 2.44)

-0.32 (-3.23 to 2.59)



0.68 (0.15 to 2.98)

3.04 (-2.62 to 8.69)

Abbreviations: ATE, Average treatment effects; FIO22, fraction of inspired oxygen; IQR, interquartile range; PaO2, partial pressure of arterial oxygen; SpO2, peripheral arterial oxygenation.

a Linear regression analysis.

However, there were differences in the outcomes observed between NRB + NC and HFNC groups in PFR improvement (13 vs 49; adjusted MR -42.09; 95% CI -56.98 to -27.20; p <= 0.001), respiratory rate at

two hour (35 vs 25 breaths/min; adjusted MR 3.50; 95% CI 0.98 to 6.03; p = .006), and duration to de-escalate to lower respiratory sup- port, such as simple face mask or NC (4 vs 5 days; adjusted MR 0.54; 95% CI 0.29 to 1.00; p = .049).


Absolute / mean difference

6.0 (0 to 12.0)

6.0 (0 to 12.0)

-1 (-3 to 0) 6 (4 to 8)

-46 (-56 to -37)

-36 (-47 to -25)

7 (5 to 9)

-2 (-4 to 0)

-4 (-7 to -1)

-2 (-6 to 3)

-2 (-5 to 0)

0 (-3 to 3)

  1. Discussion

In this retrospective study of COVID-19 patients with HRF present- ing to the ED, both NRB + NC and HFNC groups demonstrated sig- nificant improvements in PaO2 at two hours. The NRB + NC group had a greater two-hour median difference in PaO2 improvement compared to the HFNC group (13 vs 4 mmHg; p-value = .009), but this difference was not statistically significant after IPTW adjustment (adjusted MR 2.81; p-value = .524). In contrast, SpO2 improvement was significantly different between the two groups. This could be explained by the NRB + NC group having lower SpO2 at baseline, attributable to a greater proportion of severely ill patients compared to the HFNC group (APACHE II score 17 vs 11; p-value = .001). More- over, more patients in the NRB + NC group required proning com- pared to the HFNC group (46 vs 12; p-value = .001) to achieve a similar targeted SpO2.

No (%)

Non-rebreather mask and low-flow nasal cannula

High-flow nasal cannula

Primary Outcome

Partial pressure of arterial oxygen at 2 h, PaO2 median (IQR) mm Hga Improvement in partial pressure of arterial oxygen at 2 h,

PaO2 median (IQR) mm Hga

Secondary Outcomes

Oxygen saturation at 2 h, SpO2 median (IQR) %a

Improvement in oxygen saturation at 2 h, SpO2 mean (SD), %a PaO2 / FIO2 at 2 h, median (IQR)a

Improvement in PaO2 / FIO2 at 2 h, median (IQR)a Respiratory rate at 2 h, median (IQR) breaths/mina

Reduction in respiratory rate at 2 h, median (IQR) breaths/mina Intubation within 28 days from enrollmentb

Days to intubation, median (IQR)c

Invasive ventilation-free days within 28d, median (IQR)a

Days to de-esclation to lower respiratory support (e.g: nasal cannula, simple face mask), median (IQR)b

Mortality within 28db

Hospital length of stay, median (IQR) daysa

80.2 (72.0 to 84.0)

13.0 (7.0 to 18.0)

69.1 (57.6 to 80.2)

4.0 (-6.4 to 17.0)

90 (88 to 94)

8 (4)

80 (72 to 84)

13 (7 to 18)

35 (32 to 36)

3 (2 to 5)

28 (53.8)

0 (0 to 1)

23 (0 to 28)

4 (3 to 4)

93 (89 to 95)

1 (6)

118 (105 to 141)

49 (30 to 68)

25 (23 to 30)

5 (1 to 9)

22 (37.9)

3 (1 to 6)

28 (0 to 28)

5 (3 to 7)

15 (28.8)

16 (8 to 19)

16 (27.6)

13 (10 to 19)

The findings of this study corroborate previous trials that observed significant improvements in PFR and reduction in respiratory rate after initiating HFNC [13]. A systematic review found that HFNC sig- nificantly improves PFR and respiratory rate [13]. HFNC reduces oxygen dilution, allows a more reliable FIO2, eliminates physiological dead space, and generates positive end expiratory pressure (PEEP) [24]. Nonetheless, improvement of PFR and respiratory rate were also seen in the NRB + NC group. The use of a low-flow NC in combination with NRB reduces air mixing and increases FIO2 closer to 1.0, which cannot be achieved by NRB alone [13,20,21].

Table 3

Primary and secondary outcomes, according to study group.


Early observational studies demonstrated high mortality rates in COVID-19 patients with HRF requiring invasive ventilation [4,6]. The usage of non-invasive respiratory support avoids Ventilator-induced lung injury and the complications related to prolonged sedation and Neuromuscular paralysis [25]. Evidence for HFNC as an effective treatment for HRF is drawn from studies on populations other than COVID-19 pneumonia [7,11-13].

Logistic regression analysis.

c Poisson regression analysis.

Despite the presumed physiological benefits of HFNC use over COT, there is no evidence that it reduces mortality. A meta-analysis by Rochwerg et al. of 9 trials involving 2093 patients with HRF of various


aetiologies, found no difference in mortality in patients treated with HFNC compared to COT (relative risk 0.94; 95% CI 0.67 to 1.31) [7]. Sim- ilarly, other recent trials did not demonstrate the Mortality benefits of HFNC over COT [16-18]. This suggests that neither HFNC nor COT have a direct effect on the disease process and impacting mortality, but both have similar efficacy in terms of improving oxygenation and preventing intubation in a greater proportion of patients. Early relief of respiratory effort could theoretically reduce patients’ self-inflicted lung injury and result in improved clinical outcomes [17]. However, COVID-19 is a unique disease from other causes of HRF, with complex manifestations of dysregulated immuno-inflammatory response, thrombotic, parenchymal, and endotheliopathy derangements [26]. Various factors and treatments may influence long-term outcomes of patients.

In COVID-19-related HRF, current trials have shown inconsistent evidence in the role of HFNC in avoiding intubation [16-18]. In the HifLo-Covid trial, the use of HFNC reduced the rate of endotracheal intubation compared with COT (34.3% vs 51.0%) [17]. However, the RECOVERY-RS trial contradicted this finding [18]. The RECOVERY-RS trial randomized 1273 COVID-19 patients to continuous positive airway pressure (CPAP), HFNC, or COT, and found reduced rate of endotracheal intubation in CPAP compared to COT (36.3% vs 44.4%), but no significant difference between HFNC and COT (44.3% vs 45.1%) [18].

In this study, patients in the NRB + NC group were more severely ill at baseline with more dyspnea, hypoxia, and acidosis. They also had a greater Vasopressor requirement than patients in the HFNC group. Con- sequently, the cumulative intubation rates over 28 days showed that patients in the NRB + NC group were more likely to require early intu- bation, with the majority requiring it within a day (Log rank p = .02) (Fig. 4). This correlated with the low ROX index after two hours of initi- ating NRB + NC therapy. Although the patients treated with NRB + NC appeared to have a lower survival rate than those treated with HFNC, this was not statistically significant (Log rank p = .503) (Fig. 5). The curves imply that a considerable number of patients in the NRB + NC died early in the course of the disease within 8 days, before reaching a plateau at day 10 and nearly converges with the HFNC curve at day 24. The HFNC group did not show a statistically significant difference in intubation rate compared to the NRB + NC group. Comparatively, the rates of intubation in recent COVID-19-related HRF trials were 34% and 44% [17,18]. In this study, patients in the HFNC group had more

Image of Fig. 4

Fig. 4. Cumulative incidence of intubation over 28 days using IPTW-adjusted Kaplan- Meier plots.

Image of Fig. 5

Fig. 5. Cumulative Incidence of Survival over 28 days using IPTW-adjusted Kaplan-Meier plots.

ventilator-free days compared to those in the NRB + NC group. Despite the patients in the NRB + NC group being more critically ill at baseline, the difference in ventilator-free days was not statistically significant compared to the HFNC group.

It was observed in this study that patients on HFNC took longer to de-escalate to COT. This was likely due to a greater inclination among physicians to maintain HFNC use until the required FIO2 was within

0.30. This corresponds with experts’ recommendation that FIO2 should be <0.4 before weaning off HFNC [27,28]. However, the ideal weaning strategy of HFNC is not yet established. The current SLOWH trial is investigating different HFNC weaning protocols among patients with various causes of HRF [29].

Successful avoidance of intubation could optimize resource alloca- tion in the ED, especially in the context of the COVID-19 pandemic. In settings where access to HFNC is available, using HFNC is reasonable, but NRB + NC may be a viable alternative. This study was conducted in a developing country with significant resource limitations, especially in regards to the availability of oxygen canisters and HFNC devices that were further exacerbated by the massive caseload of COVID-19 pan- demic. HFNC may consume more oxygen than NRB + NC, and invasive ventilation is cumulatively more resource-intensive than non-invasive ventilation or COT. Therefore, this necessitated the use of the readily available NRB + NC which was deemed an efficient alternative to the more resource-intensive HFNC.

This study has several limitations. Firstly, this was a retrospective observational study with a possibility of selection bias. Secondly, despite sensitivity analysis through IPTW, optimal matching could not be achieved due to the heterogeneity of the patients’ baseline characteris- tics in the two treatment groups. Potentially significant covariates which were not matched were vasopressor use and prone position. However, there was no difference in the baseline blood pressure and there is insufficient evidence regarding the benefits of awake prone po- sitioning in the management of nonintubated COVID-19 patients with HRF [30]. Thirdly, given the use of oxygen therapy with varying FIO2, PFR may be a more accurate reflection of improvement in oxygenation, especially in the setting of a right-to-left shunt due to ventilation- perfusion mismatch. Lastly, because of its retrospective nature, there may have been some variability in the exact timing of ABG sampling, although local protocol requiring ROX score to be obtained two hours after therapy is initiated may reduce this error if adhered to. Further multi-centered randomized controlled trials are required to assess the efficacy and safety of NRB + NC in severe COVID-19 infection.

  1. Conclusions

In summary, HFNC may be beneficial in COVID-19 HRF. NRB + NC is a viable alternative, especially in resource-limited settings, given similar improvement in oxygenation at two hours, and no significant differ- ences in long-term outcomes. The effectiveness of NRB + NC needs to be investigated by a powered randomized controlled trial.

CRediT authorship contribution statement Muhammad Khidir Mohd Kamil: Writing – review & editing, Writ-

ing – original draft, Supervision, Methodology, Investigation, Formal

analysis, Conceptualization. Khadijah Poh Yuen Yoong: Writing – review & editing, Supervision, Methodology, Conceptualization. Abdul Muhaimin Noor Azhar: Writing – review & editing, Supervision, Software, Methodology, Formal analysis, Conceptualization. Aida Bustam: Writing – review & editing, Supervision, Software, Methodol- ogy, Conceptualization. Ahmad Hariz Abdullah: Data curation. Mohd Hafyzuddin Md Yusuf: Writing – review & editing, Supervision, Meth- odology, Conceptualization. Aliyah Zambri: Writing – review & editing. Ahmad Zulkarnain Ahmad Zahedi: Supervision, Methodology, Conceptualization. Hidayah Shafie: Data curation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper.


We would like to express our appreciation to the Director General of Ministry of Health (MOH), Malaysia for the publication of this journal.


  1. World Health Organization. WHO COVID-19 dashboard. World Health Organization;

2022. Available: [accessed March 19, 2022].

  1. Ministry of Health (MOH). COVIDNOW in Malaysia. Ministry of Health; 2022. Available: [accessed March 19, 2022].
  2. COVID-19 cases in categories 4, 5 on the rise – Dr Adham. , BERNAMA. 2021. Avail- able: [accessed March 19, 2022].
  3. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting characteristics, comorbidi- ties, and outcomes among 5700 patients hospitalized with COVID-19 in the New York city area. JAMA. 2020;323(20).
  4. Franco C, Facciolongo N, Tonelli R, et al. Feasibility and clinical impact of out-of-ICU noninvasive respiratory support in patients with COVID-19-related pneumonia. Eur Respir J. 2020;56(5):2002130.
  5. Grasselli G, Zangrillo A, Zanella A, et al. COVID-19 Lombardy ICU network. Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to ICUs of the Lombardy region, Italy. JAMA. 2020;323(16):1574-81. https://doi. org/10.1001/jama.2020.5394.
  6. Rochwerg B, Granton D, Wang DX, et al. High flow nasal cannula compared with conventional oxygen therapy for acute hypoxemic respiratory failure: a systematic review and meta-analysis. Intensive Care Med. 2019 May;45(5):563-72. https://
  7. Brewster DJ, Chrimes N, Do TB, et al. Consensus statement: safe airway society prin- ciples of airway management and tracheal intubation specific to the COVID-19 adult patient group. Med J Aust. 2020 Jun;212(10):472-81. 50598.
  8. Alhazzani W, Moller MH, Arabi YM, et al. Surviving sepsis campaign: guidelines on the management of critically ill adults with coronavirus disease 2019 (COVID-19). Intensive Care Med. 2020 Mar 28:1-34. 06022-5.
  9. Bonnet N, Martin O, Boubaya M, et al. High flow nasal oxygen therapy to avoid inva- sive mechanical ventilation in SARS-CoV-2 pneumonia: a retrospective study. Ann Intensive Care. 2021 Feb 27;11(1):37.
  10. Ni YN, Luo J, Yu H, et al. Can high-flow nasal cannula reduce the rate of reintubation in adult patients after extubation? A meta-analysis. BMC Pulm Med. 2017 Nov 17;17


  1. Kang Hanyujie, Zhao Zhiling, Tong Zhaohui. Effect of High-flow nasal cannula oxygen therapy in immunocompromised subjects with acute respiratory failure. Respir Care. 2020;65(3):369-76.
  2. Lee CC, Mankodi D, Shaharyar S, et al. High flow nasal cannula versus conventional oxygen therapy and non-invasive ventilation in adults with acute hypoxemic respi- ratory failure: a systematic review. Respir Med. 2016 Dec;121:100-8. https://doi. org/10.1016/j.rmed.2016.11.004.
  3. World Health Organization (WHO). COVID-19 oxygen emergency impacting more than half a million people in low- and middle-income countries every day, as demand surges. World Health Organization; 2021. Available: https://www.who. int/news/item/25-02-2021-covid-19-oxygen-emergency-impacting-more-than- half-a-million-people-in-low-and-middle-income-countries-every-day-as-demand- surges [accessed March 19, 2022].
  4. Toner Eric. Potential solutions to the COVID-19 oxygen crisis in the United States. John Hopkins Bloomberg School of Public Health; 2021. Available: https://www. oxygen-memo.pdf [accessed March 19, 2022].
  5. Grieco DL, Menga LS, Cesarano M, et al. Effect of helmet noninvasive ventilation vs high-flow nasal oxygen on days free of respiratory support in patients with COVID-19 and moderate to severe hypoxemic respiratory failure: the HENIVOT ran- domized clinical trial. JAMA. 2021 May 4;325(17):1731-43. jama.2021.4682.
  6. Ospina-Tascon GA, Calderon-Tapia LE, Garcia AF, et al. Effect of High-flow oxygen therapy vs conventional oxygen therapy on invasive mechanical ventilation and clinical recovery in patients with severe COVID-19: a randomized clinical trial. JAMA. 2021 Dec 7;326(21):2161-71.
  7. Perkins GD, Ji C, Connolly BA, et al. Effect of noninvasive respiratory strategies on in- tubation or mortality among patients with acute hypoxemic respiratory failure and COVID-19. The RECOVERY-RS randomized clinical trial. JAMA. 2022;327(6):546-58.
  8. Montrief T, Ramzy M, Long B, et al. COVID-19 respiratory support in the emergency department setting. Am J Emerg Med. 2020;38(10):2160-8. 1016/j.ajem.2020.08.001.
  9. Kumar A, Sinha C, Kumar A, et al. Low flow nasal oxygen supplementation in addi- tion to non-rebreathing mask: an alternative to high flow nasal cannula oxygenation for acute hypoxemic COVID-19 patients in resource limited settings. Trends Anaesth Crit Care. 2021 Jun;38:24-5.
  10. Kabak M, Cil B. Feasibility of non-rebreather masks and nasal cannula as a substitute for high flow nasal oxygen in patients with severe COVID-19 infection. Acta Medica Mediterr. 2021;37(2):949-54.
  11. World Health Organization (WHO). Living guidance for clinical management of COVID-19. World Health Organization; 2021. Available: publications-detail-redirect/WHO-2019-nCoV-clinical-2021-2 [accessed March 19, 2022].
  12. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treat- ment effects in observational studies. Stat Med. 2015 Dec 10;34(28):3661-79.
  13. Ritchie JE, Williams AB, Gerard C, Hockey H. Evaluation of a humidified nasal high- flow oxygen system, using oxygraphy, capnography and measurement of upper air- way pressures. Anaesth Intensive Care. 2011 Nov;39(6):1103-10. 1177/0310057X1103900620.
  14. Tobin MJ, Laghi F, Jubran A. Caution about early intubation and mechanical ventila- tion in COVID-19. Ann Intensive Care. 2020 Jun 9;10(1):78. s13613-020-00692-6.
  15. Osuchowski MF, Winkler MS, Skirecki T, et al. The COVID-19 puzzle: deciphering pathophysiology and phenotypes of a new disease entity. Lancet Respir Med. 2021;9(6):622-42.
  16. Blez D, Soulier A, Bonnet F, et al. Monitoring of high-flow nasal cannula for SARS- CoV-2 severe pneumonia: less is more, better look at respiratory rate. Intensive Care Med. 2020 Nov;46(11):2094-5.
  17. Winck JC, Scala R. Non-invasive respiratory support paths in hospitalized patients with COVID-19: proposal of an algorithm. Pulmonology. 2021 Jul-Aug;27(4): 305-12.
  18. Kim MC, Lee YJ, Park JS, et al. Simultaneous reduction of flow and fraction of inspired oxygen (FIO2) versus reduction of flow first or FIO2 first in patients ready to be weaned from high-flow nasal cannula oxygen therapy: study protocol for a random- ized controlled trial (SLOWH trial). Trials. 2020 Jan 14;21(1):81. 1186/s13063-019-4019-7.
  19. Alhazzani W, Evans L, Alshamsi F, et al. Surviving sepsis campaign guidelines on the management of adults with coronavirus disease 2019 (COVID-19) in the ICU: first update. Crit Care Med. 2021 Mar 1;49(3):e219-34. 0000000000004899.