A high-flow nasal cannula system with relatively low flow effectively washes out CO2 from the anatomical dead space in a sophisticated respiratory model made by a 3D printer

Abstract

Background: Although clinical studies of the high-flow nasal cannula (HFNC) and its effect on positive end-expiratory pressure (PEEP) have been done, the washout effect has not been well evaluated. Therefore, we made an experimental respiratory model to evaluate the respiratory physiological effect of HFNC.

Methods: An airway model was made by a 3D printer using the craniocervical 3D-CT data of a healthy 32-year-old male. CO₂ was infused into four respiratory lung models (normal-lung, open- and closed-mouth models; restrictive- and obstructive-lung, open-mouth models) to maintain the partial pressure of end-tidal CO₂ (P_ETCO₂) at 40 mmHg. HFNC flow was changed from 10 to 60 L/min. Capnograms were recorded at the upper pharynx, oral cavity, subglottic, and inlet sites of each lung model.

Results: With the normal-lung, open-mouth model, 10 L/min of HFNC flow decreased the subglottic P_ETCO₂ to 30 mmHg. Increasing the HFNC flow did not further decrease the subglottic P_ETCO₂. With the normal-lung, closed-mouth model, HFNC flow of 40 L/min was required to decrease the P_ETCO₂ at all sites. Subglottic P_ETCO₂ reached 30 mmHg with an HFNC flow of 60 L/min. In the obstructive-lung, open-mouth model, P_ETCO₂ at all sites had the same trend as in the normal-lung, open-mouth model. In the restrictive-lung, open-mouth model, 20 L/min of HFNC flow decreased the subglottic P_ETCO₂ to 25 mmHg, and it did not decrease further. As HFNC flow was increased, PEEP up to 7 cmH₂O was gradually generated in the open-mouth models and up to 17 cmH₂O in the normal-lung, closed-mouth model.

Conclusions: The washout effect of the HFNC was effective with relatively low flow in the open-mouth models. The closed-mouth model needed more flow to generate a washout effect. Therefore, HFNC flow should be considered based on the need for the washout effect or PEEP.

Keywords: High-flow nasal cannula; PEEP; Rebreathing; Ventilation; Washout effect; Work of breathing.

Fig. 1

The airway model. The actual airway model was made by a 3D printer using 3D-CT data. Sampling ports were made to record capnograms at each site in the model

Fig. 2

The respiratory model. Experimental system used in the study. The airway model (Fig. 1) was connected to the lung model. The physiological dead space was adjusted to 3 mL/kg. The respiratory patterns of the lung model could be changed, and the model had pressure sensors to measure the internal pressure. Sampling tubes were connected to the capnogram at each site mentioned

Fig. 3

Capnograms and airway flows recorded. Capnograms were recorded with an HFNC flow of 10–60 L/min in each lung model. Airway flows were recorded without HFNC by the flow sensor in the lung model. There were no differences between the capnograms recorded at the subglottic and the inlet sites of the lung model, indicating the flow generated by HFNC does not reach further than the subglottic area. In open-mouth models, an HFNC flow of 10–20 L/min washed out the CO₂ of the upper pharynx and the oral cavity. The closed-mouth model needed more HFNC flow to wash out the CO₂ of the upper pharynx and the oral cavity

Fig. 4

Relationship between P_ETCO₂, PEEP, and inspiratory effort to maintain Vt. a P_ETCO₂ measured at the subglottic site in each respiratory model. P_ETCO₂ in the open-mouth models reaches a minimum value with a relatively low flow. The closed-mouth model required more flow to establish the washout effect. b The Pmus needed to maintain the initial Vt without HFNC was counted as 100%. As HFNC flow was raised, the generated PEEP and Pmus required to maintain the initial value increased