Treatment of carbon monoxide poisoning: high-flow nasal cannula versus non-rebreather face mask

Abstract

Objective

In this study, the aim was to compare the rate of decrease in carboxyhemoglobin (COHb) values at consecutive time points and calculate the half-life of COHb (COHb t½) in patients admitted to the emergency department (ED) with carbon monoxide (CO) poisoning and treated with either high flow nasal cannula oxygen therapy (HFNC) or non-rebreather face mask (NRFM).

Methods

This retrospective, cohort study with historical controls was performed over a 2-year period and included adult patients with CO poisoning, whose COHb values were checked. The HFNC group consisted of patients admitted to the ED with CO poisoning when HFNC was available in the hospital, while the NRFM group consisted of patients who presented to the ED with CO poisoning before the availability of HFNC. The primary outcome of the study was to detect the COHb t½.

Results

A total of 71 patients were enrolled in the study. While 35 patients received oxygen with NRFM, 36 patients received HFNC. The mean COHb t½ in the HFNC group was 41.1 min (95% CI: 31.0–58.4) and 64.0 min (95% CI: 43.5–114.4) in the NRFM group. We did not find a significant difference in the COHb t½ between the HFNC group and NRFM group (p = 0.099). COHb levels between treatment arms at serial time points showed a statistically significant difference at 60 min (p = 0.048). We compared the decay constant and half-life of COHb between groups according to gender. In both genders, COHb t½ was significantly different between groups, and COHb t½ was lower in the patients treated with HFNC.

Conclusion

HFNC was effective in reducing the half-life of COHb values in patients with carbon monoxide poisoning. Prospective studies to be conducted in larger groups are needed to fully understand the effect of HFNC on carbon monoxide poisoning.

Keywords:

• Carbon monoxide poisoning
• high-flow nasal cannula oxygen therapy
• non-rebreather face mask

Disclosure statement

No potential conflict of interest was reported by the author(s).