Effect of high-flow nasal cannula versus non-invasive ventilation in preventing re-intubation in high-risk chronic obstructive pulmonary disease patients: A randomised controlled trial

ABSTRACT

Background:

Currently, a high-flow nasal cannula (HFNC) has been shown to improve extubation outcomes. However, there is a lack of evidence on the utilisation of HFNC in high-risk chronic obstructive pulmonary disease (COPD) patients. This study aimed to compare the effectiveness of HFNC versus non-invasive ventilation (NIV) in preventing re-intubation following planned extubation in high-risk COPD patients.

Patients and Methods:

In this prospective, randomised, controlled trial, 230 mechanically ventilated COPD patients at high risk for re-intubation who fulfilled the criteria for planned extubation were enrolled. Post-extubation blood gases and vital signs at 1, 24, and 48 hours were recorded. The primary outcome was the re-intubation rate within 72 hours. Secondary outcomes included post-extubation respiratory failure, respiratory infection, intensive care unit and hospital length of stay, and mortality rate at 60 days.

Results:

230 patients after planned extubation were randomly allocated to receive either HFNC (n = 120) or NIV (n = 110). Re-intubation within 72 hours was significantly lower in the high-flow group: 8 patients (6.6%) versus 23 patients (20.9%) in the NIV group {absolute difference, 14.3% [95% confidence interval (CI), 10.9–16.3]; P = 0.001}. The frequency of post-extubation respiratory failure was less in patients assigned to HFNC than in those allocated NIV (25% vs. 35.4%) [absolute difference, 10.4% (95% CI, 2.4–14.3); P = 0.001]. There was no significant difference between the two groups regarding reasons for respiratory failure after extubation. It was observed that the 60-day mortality rate was lower in patients who received HFNC than in those assigned to NIV (5% vs. 13.6%) [absolute difference, 8.6 (95% CI, 4.3 to 9.10); P = 0.001].

Conclusion:

The use of HFNC after extubation appears to be superior to NIV in reducing the risk of re-intubation within 72 hours and 60-day mortality in high-risk COPD patients.

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is considered the fourth leading cause of death worldwide as it negatively impacts patients’ health, hospitalisation, and mortality rates. Especially for COPD patients who received invasive mechanical ventilation, hospital mortality rated 28 to 30%. Invasive ventilation may indeed be necessary to rescue COPD patients with severe hypercapnic respiratory failure. The global strategy for the diagnosis, management, and prevention of COPD reported that[1] using non-invasive ventilation (NIV) is recommended as a weaning-facilitating strategy in predominantly COPD with acute exacerbation and decreases hospital mortality and the incidence of ventilator-associated pneumonia without increasing the risk of re-intubation or weaning failure. It has been shown to improve gas exchange, reduce the work of breathing, and improve survival.[2,3]

The frequency of post-extubation respiratory failure was increased up to 48% in mechanically ventilated patients with chronic respiratory disorders and hypercapnia.[3] Thus, preventing post-extubation respiratory failure and re-intubation rate plays a key role in successful liberation from mechanical ventilation in COPD patients. Numerous studies have noted that NIV as a weaning strategy reduced the incidence of post-extubation respiratory failure and improved the patients’ outcomes.[4,5,6] However, NIV intolerance was reported in more than 25% of patients; it can cause an increased risk of treatment failure and re-intubation.[7,8]

The recent technological improvements have enabled high-flow oxygen therapy to improve conventional oxygen therapy performance via a nasal cannula,[9] capable of delivering humidified and heated oxygen at a flow rate of up to 80 L/min. In addition, there is a low level of continuous positive airway pressure with an increased end-expiratory lung volume and reduced work of breathing, partly through intrinsic positive end-expiration pressure compensation and CO₂ washout from the anatomical dead space. Moreover, the delivery of warmed humidified inspired air up to physiologic conditions results in alleviating inflammation of the tracheobronchial mucosa and better spontaneous respiratory secretion management.[10] Based on these potential advantages, HFNC has been described in several clinical trials in the treatment of both stable and exacerbated COPD patients.[10,11,12]

HFNC therapy has promised clinical benefits in critically ill patients after extubation.[13,14] A randomised trial by Maggiore et al.[11] investigated the effect of high-flow therapy after extubation in critically ill patients on post-extubation respiratory failure and the re-intubation rate, suggesting that HFNC reduces the rate of re-intubation.

Therefore, we hypothesise that the application of HFNC therapy immediately after planned extubation would reduce the need for re-intubation in high-risk COPD patients.

MATERIALS AND METHODS

Study design and patients

This prospective randomised controlled study was carried out at Assiut University Hospital from November 2020 to May 2022. The study was conducted in accordance with the Declaration of Helsinki. All the patients provided informed consent.

All COPD patients according to Global Initiative for Obstructive Lung Disease (GOLD) criteria,[15] intubated for 48 h or more, who tolerated a spontaneous breathing trial (SBT) through a T-piece after recovery but are considered high risk for re-intubation were eligible for this study [see Figure 1].

Figure 1.

Flow chart

High risk for re-intubation has been defined as the presence of at least one of the following:[16,17] an age above 65 years old, moderate or severe chronic obstructive pulmonary disease, multiple attempts at weaning before SBT success, prolonged mechanical ventilation for ≥7 days, heart failure, acute physiology and chronic health evaluation (APACHE II score ≥12), body mass index ≥30 kg/m², or excessive airway secretions.

Exclusion criteria were as follows: 1) patients who were tracheostomies, who were accidentally extubated or had self-extubation, 2) contraindications to NIV, 3) and inability to give informed consent.

Patient demographics including age, gender, years of COPD diagnosis, and APACHE II score within the first 24 hours after admission were recorded.

Weaning protocol

As per the protocol in our study, daily screening for weaning readiness was performed according to the following criteria:[18] recovery from the precipitating illness, respiratory criteria with adequate oxygenation (PaO₂ >60 mmHg with FIO₂ ≤0.4, positive end-expiratory pressure <8 cmH₂O, and arterial pH >7.35), and clinical criteria [haemodynamic parameters were stable or the blood pressure was supported by minimal vasopressors (not more than 5 μg/kg/minute of dopamine or 0.5 μg/kg/minute of noradrenaline with mean arterial pressure ≥65 mmHg, heart rate <140/min, temperature <38°C, haemoglobin >8 g/dL, no need for sedatives, and appropriate spontaneous cough)].

During screening, the rapid shallow breathing index (RSBI) was assessed: the ratio of respiratory frequency to tidal volume (fR/VT); if fR/VT was >105 min^-1.L^-1, a spontaneous breathing trial was attempted for 30 min with a T-tube on an inspired oxygen fraction (FiO₂) of 0.3 to 0.4. Extubation was then ordered by the attending physician. Spontaneous breathing trial (SBT) failure criteria were used in the study[18] (see Appendix 1 in Supplement Material).

At extubation, the following variables were recorded: arterial blood gases (ABGs) [pH, partial pressure of arterial carbon dioxide (PaCO₂), partial pressure of arterial oxygen/fraction of inspired oxygen (PaO₂/FiO₂)], and monitored vital signs [heart rate (HR), mean arterial pressure (MAP), respiratory rate (RR)] at 1 hour, 24 hours, and 48 hours after extubation.

Patients were observed for 72 hours after extubation. During this time, extubation-related complications, reasons for re-intubation, and time to re-intubation in hours were recorded. All patients were followed up to 60 days from hospital admission. Lengths of stay in the intensive care unit (ICU) and hospital and mortality at 60 days were also recorded.

Randomisation

Immediately after extubation, eligible patients were randomised to receive either HFNC or NIV respiratory support throughout the hospitalisation period. Randomisation was stratified by a laboratory scientist not involved in the study using the technique of shuffled sealed envelopes containing equal numbers of each treatment arm. Both groups were treated by the same medical management, nursing, and respiratory therapy staff. The HFNC or NIV therapy started immediately after extubation.

Interventions

In the HFNC group, a high-flow device (Optiflow; Fisher & Paykel Healthcare, Auckland, New Zealand) was utilized. The humidifier temperature was set at 37°C via large-bore bi-nasal prongs, and inspired oxygen (FiO₂) was adjusted to maintain oxygen saturation by pulse oximetry (SpO₂) ≥90%. The flow was initially set at 10 L/min and titrated upward in 5 L/min steps until patients experienced discomfort.

NIV was delivered with an oro-nasal mask (Fisher & Paykel Health-care) connected to an ICU ventilator with a dedicated NIV mode (Evita XL, Evita 4, or Evita 2 dura, Dräger, Lübeck, Germany) equipped with a heated humidifier. Patients were ventilated by NIV with a pressure support level targeting a tidal volume of 6–8 mL/kg; the inspiratory positive airway pressure (IPAP) was initiated at 10–12 cmH₂O, and the expiratory positive airway pressure (EPAP) started at 4–5 cmH₂O. FIO₂ was adjusted to maintain SpO₂ ≥90%.

We continuously monitored patients’ vital signs and arterial blood gases. We did not allow meals during the first 24 h after extubation to avoid aspiration. Cough and expectoration were assisted by respiratory therapists.

Outcomes

The primary outcome was re-intubation within 72 hours after extubation.

The criteria for immediate re-intubation were[19,20] cardiac arrest or obvious haemodynamic instability, refractory hypoxemia (PaO₂ <50 mmHg with sufficient oxygen therapy), significant hypercapnia with pH ≤7.20, loss of consciousness or gasping for air, psychomotor agitation, or severe dyspnoea (respiratory frequency >40/min) (See Appendix 2 in Supplementary Material).

The secondary outcomes were post-extubation respiratory failure[21] (See Appendix 3 in Supplementary Material), ICU and hospital lengths of stay, and mortality at 60 days.

Additional secondary outcomes included associated complications: respiratory infection [(ventilator-associated pneumonia[22] or ventilator-associated tracheobronchitis,[23] sepsis, multi-organ failure[24] (see Appendix 4 in Supplementary Material)].

Postextubation respiratory failure definition: the presence of any of the following criteria for at least 30 min post-extubation and any of the following parameters requiring re-intubation: 1) respiratory acidosis and hypercapnia (arterial pH <7.35 along with an increase in carbon dioxide arterial tension (P_{a, CO2}) of >20% from the time of extubation), 2) hypoxemia, defined as SpO₂ by pulse oximetry <90% or PaO₂ <60 mmHg at FiO₂ >0.5, 3) a decreased level of consciousness and agitation rendering the patient unable to tolerate NIV or HFNC, 4) clinical signs suggestive of respiratory muscle fatigue and/or increased work of breathing, such as the use of respiratory accessory muscles, paradoxical abdominal motion, or intercostal indrawing, and 5) inability to clear secretions.[21]

We addressed causes of post-extubation respiratory failure, with published definitions: respiratory acidosis, hypoxia, upper-airway obstruction, aspiration or excess respiratory secretions, and encephalopathy.[25]

Statistical analysis and sample size calculation

Based on a previous study,[16] the absolute reduction in re-intubation rate was estimated at 8% from a basal rate of 18%. Thus, given a power of 80% with an α error of 5%, to detect the re-intubation rate using HFNC in COPD patients, a sample size of 110 patients in each group of the study was considered adequate for a two-sided test.

Data were collected at multiple time points. The comprehensive two-factor repeated measurement analysis between HFNC and NIV at different time points was performed. Kolmogorov–Smirnov statistic tests were used to test the normality of distribution for considered variables. Normally distributed data were expressed as means ± standard deviation, and the skewed distributed data were reported as medians with inter-quartile (25^th–75^th) percentiles. The inter-group comparison used Wilcoxon's rank sum test. Differences in categorical variables were assessed with the χ2 test or Fisher's exact probabilities test. Repeated measures analysis of variance (ANOVA) was performed for the data collected at multiple time points, and LSD-t test was compared between two groups; t test (d-value) was used for the time pairwise comparison, in which the significance level was adjusted using the Bonferroni correction method.

Data analysis was conducted with SPSS (SPSS 17.0 for windows; SPSS; Chicago, IL).

We ascertained cumulative 60-day mortality probability with Kaplan–Meier curves and used the log-rank test to compare groups.

RESULTS

Over the study period, 230 mechanically ventilated COPD patients were admitted to the ICU for acute exacerbation and underwent planned extubation. Overall, 120 (52%) patients randomised to receive HFNC and 110 (47.8%) patients randomised for NIV were available for analysis. Table 1 shows the patient’s characteristics. None of the variables depicting patients’ characteristics at baseline significantly differed between the two groups.

Table 1.

Baseline characteristics

	HFNC group (n=120)	NIV group (n=110)	P
Age, yrs	68.4±6.8	67.9±6.9	0.07
COPD history, yrs	21.2±10.2	19.4±11.5	0.48
Male gender (No. %)	76 (63.3%)	64 (58.1%)	0.34
Day 1 APACHE II score	13.9±3.2	14.2±2.5	0.19
Length of mechanical ventilation before extubation, h	118±34	124±30	0.671
Respiratory mechanics indexes after intubation
Tidal volume, mL	448.7±53.7	451.2±51.2	0.343
Minute volume, L/min	8.5±2.0	9.1±2.5	0.292
Crs, ml/cmH2O	67.1±21.6	67.5±22.9	0.86
Raw, cmH2O/L/s	15.5±3.8	14.9±4.1	0.38
PEEP, cmH2O	5.7±2.1	5.1±1.7	0.59
PS, cmH2O	10.4±2.4	10.9±3.2	0.54
Indices during spontaneous breathing trial
fR/VT, min/L	74.6±11.6	71.5±13.7	0.16
NIV settings
IPAP, cmH2O	-----	11.4±2.0	-----
EPAP, cmH2O	------	4.6±0.5	----
FIO2	------	0.5±0.1	-----
HFNC settings
Flow, L/min	52.4±5.4	------	------
FIO2	0.4±0.2	------	------

Data are presented as mean±standard deviation or No. (%). APACHE II=Acute Physiology and Chronic Health Evaluation II score (range, 0-71, with higher scores indicating higher severity of illness). Crs: respiratory system compliance; Raw: airway resistance; PEEP: positive end expiratory pressure; PS: support pressure; EPAP: expiratory positive airway pressure; FIO₂: fraction of inspiration oxygen; fR/VT: ratio of respiratory frequency to tidal volume; HFNC: high-flow nasal cannula; IPAP: inspiratory positive airway pressure; NIV: noninvasive ventilation

Before extubation, pressure support ventilation mode was utilized, and pressure support levels were similar in the HFNC and NIV groups (10.4 ± 2.4 vs. 10.9 ± 3.2 cmH₂O, P = 0.54). The severity of illness showed no significant differences [Table 1].

Respiratory rates and heart rates

Patients in both the HFNC and NIV groups had faster mean respiratory rates 1 hour after extubation. After 24 hours, the NIV group had higher respiratory rates, compared with the HFNC group (24.4 ± 3.9 vs. 22.9 ± 4.3 breaths/minute, P < 0.05). Both groups had returned to baseline by 48 hours after extubation. As for HR and MAP, no significant differences were noted between groups.

For the comparison of ABGs between groups

At 1 hour after extubation, mean PaCO₂ values were significantly increased in both groups as compared with baseline (P < 0.05). However, mean PaCO₂ in HFNC (58.4 ± 9.2 mmHg) was significantly lower than PaCO₂ with NIV (62.4 ± 11.1 mmHg, P < 0.01) at 48 hours after extubation. There were no statistically significant differences in PaO₂/FiO₂ and pH values in either group at 24 hours or 48 hours after extubation [Table 2].

Table 2.

Vital signs and arterial blood gas analysis at different time points

	Baseline	1st h post-extubation	24 h post-extubation	48 h post-extubation	F	P
HFNC (n=120)
HR (b/min)	78.7±16.9a,b,c,d	83.8±15.6a,b,c,d	84.7±18.6a,b,c,d	86.5±12.8*a,b,c,d	143.93	0.454
MAP (mmHg)	96.2±9.0a,b,c,d	88.7±9.3a,b,d	87.0±8.8a,b,d	86.1±8.1a,c,d	12.32	0.129
RR (bpm)	20.8±3.6a,b,c,d	23.6±4.9a,b,c,d	22.9±4.3a,d	20.9±3.1a,d	25.13	0.001
pH	7.45±0.03	7.45±0.03	7.44±0.05c,d	7.43±0.05c,d	0.23	0.002
PaCO2 (mmHg)	54.8±7.2a,d	56.2±8.4a,d	56.4±5.6a,d	58.4±9.2#a,d	26.058	0.001
PaO2/FiO2 (mmHg)	233.2±46.0a,b,c,d	240.5±38.3*a,b,d	240.4±39.3a,b,d	245.3±38.9a,b,d	243.20	0.000
NIV (n=110)
HR (b/min)	80.5±13.3a,b,c,d	84.8±14.2a,b,c,d	86.5±14.6a,b,c	86.9±13.2a,b,c	151.90	0.183
MAP (mmHg)	96.5±12.3a,b,c,d	90.6±9.3a,b,c,d	88.6±10.2a,b,c,d	87.7±9.2a,b,c	42.87	0.197
RR (bpm)	21.3±4.2a,b,c,d	25.2±5.9a,c,d	24.4±3.9#a,c,d	21.9±2.9a,b,c,d	48.14	0.000
pH	7.46±0.03	7.45±0.03	7.45±0.05	7.44±0.03	0.56	0.324
PaCO2 (mmHg)	53.4±8.3b,c,d	55.4±0.8.7b,c,d	58.8±10.2a,b,c,d	62.4±11.1a,b,c,d	32.65	0.002
PaO2/FiO2 (mmHg)	230.3±44.9a,b,c,d	237.9±34.4a,d	240.1±36.4a,d	244.4±40.2a,c,d	262.70	0.002

Data are shown as means±standard deviation. HFNC=High-flow nasal cannula oxygen therapy; NIV=non-invasive ventilation;. Baseline=just prior to the spontaneous breathing trial. HR=heart rate; MAP=mean arterial pressure; RR=respiratory rate; PaCO₂=partial pressure of arterial carbon dioxide; PaO₂=partial pressure of arterial oxygen; FiO₂=fraction of inspired oxygen. ^#Compared with NIV at the same time point, P<0.05. a Compared with the baseline value in the same group, P<0.05 after Bonferroni correction;. b, c, d Comparison of multiple time points within the group with 1 h, 24 h, and 48 h post-extubation. Significance values have been adjusted by the Bonferroni correction for multiple comparisons

Table 3 summarises the outcome variables in the study. All patients were followed up for 72 hours after extubation. Re-intubation within 72 hours was lower in the high-flow group: 8 patients (6.6%) versus 23 patients (20.9%) in the NIV group {absolute difference, 14.3% [95% confidence interval (CI), 10.9–16.3]; P = 0.001}.

Table 3.

Outcome variables, length of stay, and mortality

Outcome variables	HFNC (n=120)	NIV (n=110)	Difference Between Groups (95% CI)	P
Primary outcome
Re-intubation, No. (%)	8 (6.6%)	23 (20.9%)	14.3 (10.9 to 16.3)	0.001
Time to re-intubation, median (IQR), h	19 (12 to 28)	15 (9 to 31)	-4 (-54 to 46)	0.34
Reasons of re-intubation
Cardiorespiratory arrest	0	2 (1.8%)		0.02
Inability to clear secretions	2 (1.6%)	12 (10.9%)
Haemodynamic impairment	3 (2.5%)	4 (3.6%)
Hypoxemia	2 (1.6%)	4 (3.6%)
Upper airway obstruction	1 (0.8%)	1 (0.9%)
Complication
Post-extubation respiratory failure, No. (%)	30 (25%)	39 (35.4%)	10.4 (2.4 to 14.3)	0.001
Respiratory infection, No. (%)	9 (7.5%)	11 (10%)	2.5 (-0.6 to 6.2)	0.07
Ventilator-associated tracheobronchitis	5 (4.1%)	5 (4.5%)	0.4 (-1.0 to 4.4)	0.220
Ventilator-associated pneumonia	4 (3.3%)	6 (5.4%)	2.1 (-1.3 to 3.9)	0.310
Sepsis, No. (%)	1 (0.8%)	2 (1.8%)	1 (-2.4 to 1.5)	0.32
Multiorgan failure, No. (%)	0	1 (0.9%)	1 (-2.1 to 1.5)	0.21
Main causes of Post-extubation respiratory failure, No. (%)				0.321
Respiratory acidosis	12 (10%)	15 (13.6%)
Hypoxia	10 (8.3%)	11 (10%)
Aspiration or excess respiratory secretions	4 (3.3%)	8 (7.2%)
Unbearable dyspnoea	2 (1.6%)	2 (1.8%)
Upper airway obstruction	1 (0.8%)	2 (1.8%)
Encephalopathy	1 (0.8%)	1 (0.9%)
Length of stay
Intensive-care unit stay, median (IQR) (days)	6 (2 to 8)	6 (2 to 9)	0 (−10 to 24)	0.23
Hospital stay, median (IQR) (days)	11 (6 to 15)	12 (6 to 16)	1 (−28 to 32)	0.43
Mortality
Intensive-care unit mortality	5 (4.1%)	10 (9%)	4.9 (1.30 to 6.7)	0.12
Hospital mortality	5 (4.1%)	11 (10%)	5.9 (2.7 to 8.3)	0.001
Mortality at 60 days	6 (5%)	15 (13.6%)	8.6 (4.3 to 9.10)	0.001

Respiratory acidosis: pH less than 7.30; pH less than with PaCO₂ greater than 45 mmHg. Hypoxia: SpO₂ less than 90% or PaO₂ less than 60 mmHg at FIO₂ greater than 0.4. Haemodynamic impairment: A heart rate less than 50/min with loss of alertness or severe haemodynamic instability (systolic blood pressure <90 mmHg for >30 min) unresponsive to fluids and vasoactive drugs

The time to re-intubation was not statistically significant between the two groups, {19 hours [inter-quartile range (IQR), 12–28] in the high-flow group vs. 15 hours (IQR, 9–31) in the NIV group; P = 0.34}.

Incidence of post-extubation respiratory failure was less common in those patients allocated to high flow: 30 patients (25%) versus 39 patients (35.4%) in those who received NIV [absolute difference, 10.4% (95% CI, 2.4–14.3); P = 0.001]. There was no significant difference between the two groups regarding reasons for respiratory failure after extubation, except for excessive secretion, which reported a higher rate in the NIV group (P = 0.321) [Table 3].

Moreover, there was no statistically significant difference in the median ICU length of stay in the high-flow group as compared with the NIV group.

Mortality in the ICU and the hospital was significantly lower among patients assigned to HFNC as compared with the NIV group. However, the 60-day mortality was 5% in the HFNC group versus 13.6% in those who received NIV [absolute difference, 8.6 (95% CI, 4.3 to 9.10); P = 0.001]. Additionally, the Kaplan–Meier curve showed a statistical difference in 60-day mortality rates between the two groups (log-rank test 0.421, P = 0.001, see Figure 2).

Figure 2.

Kaplan–Meier curve analysis for cumulative incidence of hospital mortality

DISCUSSION

The main finding of this study was that the re-intubation rate within 72 hours in COPD patients at high risk for extubation failure was significantly lower in the high-flow group when compared with those who received NIV (6.6 vs. 20.9%). The present study revealed that the frequency of respiratory failure after planned extubation was substantially lower in those assigned to the HFNC group. Thus, these findings confirm the benefits of early use of HFNC after extubation in high-risk COPD patients during a spontaneous breathing trial.

HFNC has been increasingly used for hypoxemic patients. Compared with traditionally low-flow oxygen therapy, HFNC was associated with better comfort, fewer desaturation episodes, and a lower re-intubation rate.[13] Using HFNC in critically ill patients with persistent respiratory failure also showed a sustained improvement in clinical and physiologic parameters.[26] However, few data are available about re-intubation rates in selected populations with high risk for extubation failure as it can present significant challenges. The reported re-intubation rates in high-risk groups ranged from 22% to 24%, mainly depending on the selected criteria to represent a high risk for re-intubation.[13,25,26]

These results are in line with a large multi-centre randomised clinical trial in three ICUs, testing if HFNC is non-inferior to NIV for preventing post-extubation respiratory failure and re-intubation in patients at high risk of re-intubation. The authors reported that 22.8% in the high-flow group versus 19.1% in the NIV group were re-intubated; 26.9% in the high-flow group versus 39.8% in the NIV group experienced post-extubation respiratory failure. Therefore, high-flow oxygen therapy was not inferior to NIV for preventing re-intubation and post-extubation respiratory failure.[27]

Also, Tan et al.[28] studied 86 COPD patients with hypercapnic respiratory failure, who were randomised to HFNC or NIV at extubation. The treatment failure rate (defined as resumption of invasive ventilation) in the HFNC group was 22.7% and 28.6% in the NIV group-risk difference of - 5.8% (95% CI, - 23.8-12.4%, P = 0.535), which was significantly lower than the non-inferior margin of 9%. An analysis of causes of treatment failure showed that intolerance in the HFNC group was significantly lower than that in the NIV group.

In mechanically ventilated patients at high risk of extubation failure, Thille et al.[29] reported that the use of high-flow nasal oxygen with NIV immediately after extubation significantly decreased the risk of re-intubation compared with high-flow nasal oxygen alone.

In a randomised controlled trial by Fernandez and colleagues comparing high-flow oxygen with the conventional oxygen in non-hypercapnic patients at high risk of extubation failure, the authors reported that post-extubation respiratory failure developed in 16 (20%) in patients who received HFNC and in 21 (27%) with conventional oxygen. Thus, HFNC is considered over conventional oxygen therapy in preventing the occurrence of respiratory failure in non-hypercapnic patients at high risk of extubation failure,[30] suggesting that the physiologic effects of HFNC might perfectly match the patient need during the post-extubation period, as it could effectively interrupt the post-extubation vicious circle of edema, excessive effort, lung injury, and muscle fatigue facilitating the recovery of lung function.[31]

A study by Lee et al.[32] compared the effectiveness of the HFNC therapy to that of NIV in severe acute exacerbation of chronic obstructive pulmonary disease (AECOPD) with hypercapnic acute respiratory failure (ARF) concerning re-intubation rate and mortality. The authors reported no difference in the intubation rate and 30-day mortality between HFNC and NIV groups. Similarly, Jing et al.[33] reported no differences between HFNC and NIV in re-intubation, ICU length of stay, and 28 days mortality.

Although our study was based on short-term evaluation, the most common reason for re-intubation was inability to clear secretions among patients who received NIV. HFNC is an oxygen therapy that supplies heated, humidified high-flow mixed oxygen to the upper respiratory tract through a large-calibre nasal cannula. This effect of HFNC accounts for improved patients’ comfort and secretion clearance.[14]

In this study, we found that the mean PaCO₂ and respiratory rate in the HFNC group were lower than those in the NIV group after 48 h of therapy. There were no statistically significant differences in PaO₂/FiO₂ in both groups at different time points.

In a prospective study by Braunlich et al.,[34] who compared HFNC to NIV in 54 patients with hypercapnic COPD patients, they found that HFNC significantly decreased PaCO₂. Similarly, PaCO₂ reduction was also observed in some small sample size studies in COPD exacerbation. Pilcher et al.[35] measured the transcutaneous carbon dioxide tension (PtCO₂) after 30 min of HFNC use in 24 exacerbated COPD patients; HFNC significantly reduced the PtCO₂ after 30 min compared to COT [mean difference (MD) -1.4 mmHg; 95% CI: −2.2 to −0.6; P = 0.001]. On the other hand, several studies did not observe any difference in PaCO₂.[19,20] Jing et al.[33] randomised 42 exacerbated COPD patients after extubation to receive HFNC or NIV. The authors reported that PaCO₂ was not different between HFNC and NIV at 3, 24, and 48 h (P = 0.780 for all comparisons). In 14 exacerbated COPD patients receiving HFNC and COT after extubation, Di Mussi et al.[36] found that PaCO₂ was not different. Oxygenation was also assessed in several studies that did not report differences in oxygenation between HFNC and COT[35] or between HFNC and NIV.[32,33]

The present study revealed a significant reduction in the respiratory rate in HFNC as compared with NIV. Similarly, Pilcher et al.[35] reported a trend towards a significant reduction of the respiratory rate at 30 min during HFNC, as compared to conventional oxygen therapy, while others did not report differences in respiratory rate between HFNC and NIV.[32,33]

Therefore, based on our study results and previous studies on the physiological benefits of HFNC supportive therapy, HFNC has emerged as a safe, useful therapy in the management of COPD patients. Warm and adequately humidified gas delivered at high flow has apparent beneficial physiological effects on the airway, which can improve mucus properties, attenuate inspiratory resistance, and increase expiratory resistance.[37] Moreover, it causes a washout of CO₂, which in turn allows a higher minute ventilation fraction to facilitate gas exchange.[38]

The present study revealed no significant differences in the ICU or hospital total lengths of stay between the two groups. The 60-day mortality in the HFNC group was 6%, which was significantly lower than 13% in the NIV group. In contrast, for patients with COPD weaned off invasive ventilation, Lee et al.[32] did not find any difference in the re-intubation between HFNC and NIV regarding re-intubation rate, ICU length of stay, and 28 days mortality.

A multi-centre, randomised, controlled trial was performed to compare mortality at day 28 on patients who experienced post-extubation respiratory failure within 7 days following extubation. Patients were treated with NIV alternating with high-flow nasal oxygen or high-flow nasal oxygen alone. They found that the mortality rate on day 28 was 18% (15/84) using NIV alternating with high-flow nasal oxygen and 29% (18/62) with high-flow nasal oxygen alone.[39]

There are several limitations to the present study. First, the data were collected from patients in a single clinical centre. A large and multi-centre randomised control trial is required. The majority of our study patients did not have pulmonary function tests in the past or within 6 months; therefore, we could not determine their pulmonary function status or classifications before intubation. Future studies are needed to determine the relationship between COPD patients’ pulmonary function and HFNC success in weaning. Another limitation was the nature of the study interventions; blinding was not possible. This might be a source of bias.

CONCLUSION

In conclusion, the use of high-flow nasal cannula oxygen compared with NIV therapy reduced the risk of re-intubation within 72 hours and 60-day mortality in high-risk COPD patients.

Thus, routine implementation of this strategy after extubation is advisable. HFNC appears to be an effective means of respiratory support for COPD patients after planned extubation.

Ethical approval and consent to participate

The research received ethical approval from the Ethics Committee of the Faculty of Medicine. The data were confidential. All procedures in the current study were performed according to the ethical standards of the institutional research committee.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

SUPPLEMENTARY MATERIAL

Appendix 1. Criteria for Spontaneous Breathing Trial Failure.

Appendix 2. Criteria for immediate reintubation.

Appendix 3. Definition of Postextubation Respiratory Failure.

Appendix 4. Definition of Ventilator-Associated Pneumonia and Tracheobronchitis.

STUDY PROTOCOL.

Appendix 1:

Criteria for Spontaneous Breathing Trial Failure.

Criteria for spontaneous breathing trial failure were agitation, anxiety, depressed mental status, diaphoresis, cyanosis, evidence of increasing respiratory effort, increased accessory muscle activity, facial signs of distress, dyspnea, PaO2 lower than 60 mmHg or SpO2 lower than 90% on inspired fraction of oxygen higher than 0.5, PaCO2 higher than 50 mmHg or increased more than 8 mmHg from baseline value, arterial pH lower than 7.32 or decreased more than 0.07 from baseline value, respiratory rate higher than 35 breaths per minute or increased more than 50% from baseline value, heart rate higher than 140 beats per minute or increased more than 20% from baseline value, systolic arterial pressure higher than 180 mmHg or increased more than 20% from baseline value, systolic arterial pressure lower than 90 mmHg, or cardiac arrhythmias.

Appendix 2: Primary outcome:

Reintubation within 72 hours after extubation.

Intubation rate was defined that the patients underwent endotracheal intubation with mechanical ventilation due to continuous hypoxia and hypercapnia despite HFNC or NIV therapy.

Defined criteria for immediate reintubation:

Presence any of the following major clinical events: Respiratory or cardiac arrest or obvious hemodynamic instability (unresponsive to fluids and vasoactive drugs), refractory hypoxemia (PaO₂ <50 mmHg with sufficient oxygen therapy), significant hypercapnia with pH =7.20, loss of consciousness or gasping for air, psychomotor agitation or severe dyspnea (respiratory frequency >40/min).

Appendix 3:

Secondary outcomes:

1. Postextubation respiratory failure classified according to the postextubation respiratory failure definition, and according to a clinical diagnosis of extubation failure if applicable.

Postextubation respiratory failure definition: presence any of the following criteria for at least 30 min post extubation any of the following parameters requiring reintubation: 1) respiratory acidosis and hypercapnia (arterial pH < 7.35 along with an increase in carbon dioxide arterial tension (Pa, CO2) of > 20% from the time of extubation), 2) hypoxaemia: defined as SpO2 by pulse oximetry < 90% or PaO2 < 60 mmHg at FiO2 > 0.5, 3) decreased level of consciousness, agitation rendering the patient unable to tolerate NIV or HFNC, 4) clinical signs suggestive of respiratory muscle fatigue and/or increased work of breathing, such as the use of respiratory accessory muscles, paradoxical abdominal motion, or intercostal indrawing, and 5) inability to clear secretions.

We assigned causes to respiratory failure after extubation, with adapted published definitions: respiratory acidosis, hypoxia, upper-airway obstruction; aspiration or excess respiratory secretions; and encephalopathy.

Appendix 4:

2. Respiratory infection (ventilator-associated pneumonia or ventilator-associated tracheobronchitis).

• Ventilator-associated pneumonia (VAP) was defined as fever (temperature > 38°C) or altered leukocyte count (>12,000/mL or < 4,000/mL) plus new onset of purulent endotracheal secretions or change in sputum, with new and progressive or persistent infiltrate or consolidation or cavitation and a significant pathogen culture (>105 cfu/mL in semiquantitative endotracheal aspirate, >104 cfu/mL in bronchoalveolar lavage fluid, or > 103 cfu/mL in protected brush specimens).

• Ventilator-associated tracheobronchitis (VAT) was defined by the same criteria but without new infiltrates.

• Sepsis or multiorgan failure:

• Sepsis was defined according to Surviving Sepsis Campaign: International Guidelines for Management of Severe Sepsis and Septic Shock, 2012.

• Multiorgan failure was defined according to Surviving Sepsis Campaign: International Guidelines for Management of Severe Sepsis and Septic Shock, 2012.

4. ICU and hospital length of stay.

5. Mortality at 60 days.