Early nasal high-flow versus Venturi mask oxygen therapy after lung resection: a randomized trial

Abstract

Background: Data on high-flow nasal oxygen after thoracic surgery are limited and confined to the comparison with low-flow oxygen. Different from low-flow oxygen, Venturi masks provide higher gas flow at a predetermined fraction of inspired oxygen (FiO2). We conducted a randomized trial to determine whether preemptive high-flow nasal oxygen reduces the incidence of postoperative hypoxemia after lung resection, as compared to Venturi mask oxygen therapy.

Methods: In this single-center, randomized trial conducted in a teaching hospital in Italy, consecutive adult patients undergoing thoracotomic lung resection, who were not on long-term oxygen therapy, were randomly assigned to receive high-flow nasal or Venturi mask oxygen after extubation continuously for two postoperative days. The primary outcome was the incidence of postoperative hypoxemia (i.e., ratio of the partial pressure of arterial oxygen to FiO2 (PaO2/FiO2) lower than 300 mmHg) within four postoperative days.

Results: Between September 2015 and April 2018, 96 patients were enrolled; 95 patients were analyzed (47 in high-flow group and 48 in Venturi mask group). In both groups, 38 patients (81% in the high-flow group and 79% in the Venturi mask group) developed postoperative hypoxemia, with an unadjusted odds ratio (OR) for the high-flow group of 1.11 [95% confidence interval (CI) 0.41-3] (p = 0.84). No inter-group differences were found in the degree of dyspnea nor in the proportion of patients needing oxygen therapy after treatment discontinuation (OR 1.34 [95% CI 0.60-3]), experiencing pulmonary complications (OR 1.29 [95% CI 0.51-3.25]) or requiring ventilatory support (OR 0.67 [95% CI 0.11-4.18]). Post hoc analyses revealed that PaO2/FiO2 during the study was not different between groups (p = 0.92), but patients receiving high-flow nasal oxygen had lower arterial pressure of carbon dioxide, with a mean inter-group difference of 2 mmHg [95% CI 0.5-3.4] (p = 0.009), and were burdened by a lower risk of postoperative hypercapnia (adjusted OR 0.18 [95% CI 0.06-0.54], p = 0.002).

Conclusions: When compared to Venturi mask after thoracotomic lung resection, preemptive high-flow nasal oxygen did not reduce the incidence of postoperative hypoxemia nor improved other analyzed outcomes. Further adequately powered investigations in this setting are warranted to establish whether high-flow nasal oxygen may yield clinical benefit on carbon dioxide clearance.

Trial registration: ClinicalTrials.gov, NCT02544477 . Registered 9 September 2015.

Keywords: High-flow oxygen therapy; Hypoxemia; Noninvasive ventilation; Postoperative pulmonary complications; Thoracotomic lobectomy.

Conflict of interest statement

Ethics approval and consent to participate

The study was approved by local Ethics Committee and written informed consent to data analysis was obtained by all studied patients

Consent for publication

Not applicable

Competing interests

MA has received payments for board participation from Maquet, Air Liquide, and Chiesi. MA and DLG disclose a research grant received by General Electrics Health Care. DLG has received payments for travel expenses by Maquet, Getinge, and Air Liquide and is supported by grants from SIAARTI/Merck Sharp & Dohme and ESICM (2017 Bernhard Dräger Award for Advanced Treatment of Acute Respiratory Failure). All other authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Fig. 1

Consort flow diagram

Fig. 2

Kaplan–Meier plots of the cumulative incidence of postoperative hypoxemia

Fig. 3

Postoperative PaO2/FiO2 ratio and PaCO2 in the two study groups. Results are expressed as means and standard deviation. No differences were detected in the PaO2/FiO2 ratio (ANOVA p = 0.92). Patients in the HFNC group showed lower PaCO2 over the entire course of the study (ANOVA p = 0.015), with a mean difference between study treatments of 1.7 mmHg [95% CI 0.3–3]. This difference was particularly evident while the assigned treatments were administered, with a mean difference between groups of 2 mmHg [95% CI 0.5–3.4] (ANOVA p = 0.009). *Indicates p < 0.05 for the comparison between HFNC and Venturi mask at the single timepoint

Fig. 4

Kaplan–Meier plots of the cumulative incidence of postoperative hypercapnia in the two study groups. The inter-group difference remained significant after adjustment for age, history of clinically documented pulmonary infections in the month preceding surgery, and preoperative PaCO2, with a hazard ratio for HFNC of 0.33 [0.14–0.74] (p = 0.007). Please note that this analysis was not prespecified and should be considered exploratory in nature.