Effect of Different Interfaces on FIO2 and CO Rebreathing During Noninvasive Ventilation.

BACKGROUND

Improving [Formula: see text] and reducing CO rebreathing ([Formula: see text]) are the key means to improve the therapeutic efficacy of noninvasive ventilation (NIV). This study aimed to investigate the impact of interface design on [Formula: see text] and [Formula: see text] during NIV.

METHODS

A simulated lung model was established to analyze 17 different interfaces. CO was injected into the outlet of the simulated lung, and the noninvasive ventilator was connected to the simulated lung to simulate the application of NIV in patients with COPD with hypercapnia. [Formula: see text] and [Formula: see text] were calculated by mathematical integration of synchronously collected data pertaining to real-time pressure, flow, oxygen concentration, and CO concentration in the breathing circuit. Comparisons were performed between different types (nasal vs oronasal) and models of interfaces as well as between interfaces with different leak positions. Correlation of [Formula: see text] and [Formula: see text] with inner volume and leakage, respectively, and the correlation between [Formula: see text] and [Formula: see text] were analyzed.

RESULTS

[Formula: see text] levels were significantly different with a nasal or an oronasal mask (0.45 ± 0.05% vs 0.41 ± 0.08%, respectively; < .001). [Formula: see text] levels associated with different models of interfaces varied significantly (all < .001); [Formula: see text] did not differ significantly among the different interfaces ( = .19). Leak position significantly affected [Formula: see text] and [Formula: see text] (all < .001). Both inner volume and leakage significantly correlated with [Formula: see text] (r = -0.23, < .001; r = -0.08, = .02). There was a significant correlation between [Formula: see text] and [Formula: see text] (r = 0.43, < .01); the general linear equation was = 0.17 + 0.02 (r = 0.43, R = 0.19).

CONCLUSIONS

The design of the interface had a significant impact on [Formula: see text] and [Formula: see text] during NIV. [Formula: see text] and [Formula: see text] showed a significant positive correlation, although the effect size of correlation was moderate.