ObjectiveTo compare the therapeutic effects of invasive-high-flow oxygen therapy (HFNC) and invasive-non-invasive ventilation (NIV) sequential strategies on severe respiratory failure caused by chronic obstructive pulmonary disease (COPD), and explore the feasibility of HFNC after extubation from invasive ventilation for COPD patients with severe respiratory failure.MethodsFrom October 2017 to October 2019, COPD patients with type Ⅱ respiratory failure who received invasive ventilation were randomly assigned to a HFNC group and a NIV group at 1: 1 in intensive care unit (ICU), when pulmonary infection control window appeared after treatments. The patients in the HFNC group received HFNC, while the patients in the NIV group received NIV after extubation. The primary endpoint was treatment failure rate. The secondary endpoints were blood gas analysis and vital signs at 1 hour, 24 hours, and 48 hours after extubation, total respiratory support time after extubation, daily airway care interventions, comfort scores, and incidence of nasal and facial skin lesions, ICU length of stay, total length of stay and 28-day mortality after extubation.ResultsOne hundred and twelve patients were randomly assigned to the HFNC group and the NIV group. After secondary exclusion, 53 patients and 52 patients in the HFNC group and the NIV group were included in the analysis respectively. The treatment failure rate in the HFNC group was 22.6%, which was lower than the 28.8% in the NIV group. The risk difference of the failure rate between the two groups was –6.2% (95%CI –22.47 - 10.43, P=0.509), which was significantly lower than the non-inferior effect of 9%. Analysis of the causes of treatment failure showed that treatment intolerance in the HFNC group was significantly lower than that in the NIV group, with a risk difference of –38.4% (95%CI –62.5 - –3.6, P=0.043). One hour after extubation, the respiratory rate of both groups increased higher than the baseline level before extubation (P<0.05). 24 hours after extubation, the respiratory rate in the HFNC group decreased to the baseline level, but the respiratory rate in the NIV group was still higher than the baseline level, and the respiratory rate in the HFNC group was lower than that in the NIV group [(19.1±3.8) vs. (21.7±4.5) times per minute, P<0.05]. 48 hours after extubation, the respiratory rates in the two groups were not significantly different from their baseline levels. The average daily airway care intervention in the NIV group was 9 (5 - 12) times, which was significantly higher than the 5 (4 - 7) times in the HFNC group (P=0.006). The comfort score of the HFNC group was significantly higher than that of the NIV group (8.6±3.2 vs. 5.7±2.8, P= 0.022), while the incidence of nasal and facial skin lesions in the HFNC group was significantly lower than that in the NIV group (0 vs. 9.6%, P=0.027). There was no significant difference in dyspnea score, length of stay and 28-day mortality between the two groups.ConclusionsThe efficacy of invasive-HFNC sequential treatment on COPD with severe respiratory failure is not inferior to that of invasive-NIV sequential strategy. The two groups have similar treatment failure rates, and HFNC has better comfort and treatment tolerance.