Asymmetrical nasal high flow ventilation improves clearance of CO2 from the anatomical dead space and increases positive airway pressure

Abstract

Positive airway pressure that dynamically changes with breathing, and clearance of anatomical dead space are the key mechanisms of noninvasive respiratory support with nasal high flow (NHF). Pressure mainly depends on flow rate and nare occlusion. The hypothesis is that an increase in asymmetrical occlusion of the nares leads to an improvement in dead-space clearance resulting in a reduction in re-breathing. Clearance was investigated with volumetric capnography in an adult upper-airway model, which was ventilated by a lung simulator with entrained carbon dioxide (CO2) at respiratory rates (RR) of 15-45 min-1 and at 18 min-1 with chronic obstructive pulmonary disease (COPD) breathing patterns. Clearance was assessed at NHF of 20-60 L/min with a symmetrical interface (SI) and an asymmetrical interface (AI). CO2 kinetics visualized by infrared spectroscopy and mathematical modeling were used to study the mechanisms of clearance. At a higher RR (35 min-1) and NHF of 60 L/min, clearance in the upper airway was significantly higher with the AI when compared with the SI (29.64 ± 9.96%, P < 0.001), as opposed to at a lower RR (15 min-1) (1.40 ± 6.25%, P > 0.05), (means ± SD). With COPD breathing, clearance by NHF was reduced but significantly improved with the AI by 45.93% relative to the SI at NHF 20 L/min (P < 0.0001). The maximum pressure achieved with the AI was 6.6 cmH2O and NHF was 60 L/min at the end of expiration. Pressure differences between nasal cavities led to the reverse flow observed in the optical model. Asymmetrical NHF increases dead-space clearance by reverse flow through the choanae and accelerates purging of expired gas via the less occluded nare.NEW & NOTEWORTHY The asymmetrical interface generated reverse flow in the nasal cavities and across the choana, which led to unidirectional purging of expired gas from the upper airways. This accelerated the clearance of anatomical dead space and reduced re-breathing while increased resistance to flow resulted in higher positive end-expiratory pressure (PEEP). These findings are relevant to patients with elevated respiratory rates or with expiratory flow limitations where dead-space clearance by NHF can be substantially reduced.

Keywords: asymmetrical; dead space; nasal cannula; nasal high flow; respiratory support.

Conflict of interest statement

S.T., M.R., A.G., and L.G.T.v.d.H. are employees of Fisher & Paykel Healthcare. S.T. disclosures a US Patent No. 10569043: Asymmetrical nasal delivery elements and fittings for nasal interfaces.

Figures

Graphical abstract

Figure 1.

A: the relationship between total nare area, percent occlusion, and total leak area by the large (black OPT 946) and medium (gray OPT944) symmetrical interface (SI) and asymmetrical large (dark blue, OPT966) and medium (light blue OPT964) interfaces (AI) used in the study. Reduction of the nare area may lead to a complete occlusion of both nares with the SI, which is less likely with the different-sized prongs of the AI. The dotted line represents the averaged adult nare area used in the model. B: a schematic view of the prong’s cross-sectional area (internal circles) and the nares (external circles) with the leak area as a gap between the internal and external circles; these are depicted in proportion.

Figure 2.

Effect of an increased occlusion by symmetrical cannula interface (SI) and asymmetrical cannula interface (AI) on dead-space clearance in the anatomically correct three-dimensional upper airway model during nasal high flow (NHF) rates of 20, 40, and 60 L/min and an increasing respiratory rate (RR). A: breathing patterns used in the lung simulator at RRs of 15, 25, 35, and 45 min−1. Gray vertical bars demonstrate time of the last and first 100 mL during expiration and inspiration. This corresponds to the time when the majority of clearance occurs. An increase in RR leads to reduced clearance time. B: clearance achieved by increased total nare occlusion by an SI (left) and AI (right). A standard medium-sized SI was used as control, as shown in the central Fig. C: this illustrates the difference in clearance due to increased nare occlusion with SI (left) and AI (right) relative to the SI control. Asymmetrical occlusion increases dead-space clearance. Symmetrical occlusion leads to reduced clearance at higher RRs where the difference in clearance between the two interfaces is the greatest. At an RR of 15 min−1 and NHF of 60 L/min, the difference is not significant (see P value table in graph).

Figure 3.

The relationship between the dead-space clearance and positive end-expiratory pressure (PEEP) measured in the upper airway model during nasal high flow (NHF) rates of 20, 40, and 60 L/min with the symmetrical cannula interface (SI; A) and asymmetrical cannula interface (AI; B) during simulated breathing at an respiratory rate (RR) of 45 min−1 (Ti:Te 1:1) and 15 min−1 (Ti:Te 1:2) for different levels of occlusion (larger and smaller cannulae). The SI at an RR of 45 min−1 demonstrates a more linear relationship between the clearance and PEEP when compared with the AI (R2 = 0.79 vs. R2 = 0.65 respectively).

Figure 4.

Effect of simulated breathing patterns at an respiratory rate (RR) of 18 min−1 on the dead-space clearance with the symmetrical cannula interface (SI) and asymmetrical cannula interface (AI) with nasal high flow (NHF) rates of 20, 40, and 60 L/min. A: breathing patterns with (chronic obstructive pulmonary disease, COPD 1 and 2) and without (control) expiratory flow limitations used in the experiment. COPD 2 also generates “intrinsic positive end-expiratory pressure (PEEP)” and has higher expiratory flow toward the end of expiration, but with the same Ti:Te ratio as in the control. Gray vertical bars demonstrate the time of the last and first 100 mL during expiration and inspiration where most of dead-space clearance occurs. B: in COPD 1 and 2 the clearance achieved with the SI is reduced; the clearance by AI was significantly less affected by the reduction in clearance time. The dead-space clearance with the SI at NHF of 60 L/min and the AI at 20 L/min were not significantly different. The table shows the differences in the average percentage of clearance between the AI and SI at the different flow rates used within each group (means ± SD). The breathing patterns for both COPD groups showed a reduced clearance time regardless of the Ti:Te ratio, resulting in a reduction in dead-space clearance, which was significantly improved by NHF with the AI.

Figure 5.

CO2 kinetics in an optical upper airway model during one breathing cycle, at an respiratory rate (RR) of 15 min−1 and nasal high flow (NHF) of 40 L/min. A: the breathing pattern used in the subsequent graphs. CO2 percentage was measured with infrared spectroscopy in rectangular regions of interests (ROI 1 and ROI 2). B: time graph along with the corresponding screenshots of both nasal cavities during a single breath. The vertical dotted line represents the time when expiration changes to inspiration. The dashed line on the graph represents the top cavity and the solid line the bottom cavity. The cavity with the larger asymmetrical prong had markedly less CO2 throughout respiration. C: using the area under the curve of the time graphs in B, the amount of CO2 in the nasal cavity during expiration (black) and inspiration (gray).

Figure 6.

Breathing by an operator through the optical model via the symmetrical interface (SI) and the asymmetrical interface (AI) at nasal high flow (NHF) rate of 20 L/min. A: voluntary increase of respiratory rate (RR) over a minute. B: the first three breaths from A show the percentage of CO2 (measured by infrared spectroscopy) in each nasal cavity (solid line -> right nasal cavity, dotted line -> left nasal cavity). In the AI, the left prong has the larger diameter and the difference in T50% can be seen by the left shift of the dotted line. C: multivariable plots demonstrate that a reduction in re-breathing (measured in the trachea) at a low RR with the AI was related to a difference in time to clear 50% of the CO2 (ΔT50%) in the respective nasal cavities.

Figure 7.

A: differential pressure measured in nasal cavities of the three-dimensional model during nasal high flow (NHF) rates of 20, 40, and 60 L/min via symmetrical interface (SI) and asymmetrical interface (AI) with a respiratory rate (RR) range from 15 to 45 min−1 across all breaths. LOWESS lines are fitted to each dataset. The pressure in cmH2O at flow rates of −0.1 and 0.1 L/min for the SI and the AI respectively are: 20 L/min 0.016 ± 0.005 vs. 0.066 ± 0.017, P < 0.001; 40 L/min 0.023 ± 0.011 vs. 0.120 ± 0.021, P < 0.001; and 60 L/min 0.007 ± 0.027 vs. 0.201 ± 0.019, P < 0.001. B: differential pressure ports in the model.

Figure 8.

Mathematical modeling of flow in the upper airways with an asymmetrical interface (AI). A: a schematic view of upper airways with the AI demonstrates resistance in the larger prong (RLP), the nare occluded with the larger prong (RNL), smaller prong (RSP), the nare occluded with the smaller prong (RSN), nasal high flow (NHF), and tidal breathing flow (expiration and inspiration). B: the Wheatstone bridge circuit consisted of five resistors: resistance in the smaller prong was higher than in the larger prong and resistance in the nare occluded by the smaller prong was lower than resistance of the nare occluded by the larger prong, and the resistance in the nasopharynx was very small. During inspiration, the lowest resistance in the cannula was in the larger prong. During expiration, the lowest resistance was in the less occluded nare with the smaller prong, which created asymmetry and diverted the expired gas via the less occluded nare. C: NHF at 20, 40, and 60 L/min generated flow-dependent differential pressure between the nasal cavities during the breathing cycle. D: reverse flow between the nasal cavities peaked at the end of expiration and led to unidirectional purging of expired gas via the nare occluded by the smaller prong.

Figure 9.

Flow within the nasal cavities and asymmetrical cannula prongs with nasal high flow (NHF) at 60 L/min during a single breath. The shaded area represents time for clearance as the last and the first 100 mL of expiration and inspiration, respectively. The vertical dotted line denotes the change from expiration to inspiration. A: flow of a single breath (respiratory rate 18 min−1, Ti:Te 1:2). B: flow out of the nares during a single breath (solid line = nare with the larger prong, dotted line = nare with the smaller prong). C: flow in the nasal prongs during a single breath (solid line = nare with the larger prong, dotted line = nare with the smaller prong). D: reverse flow through the choanae; this became maximal at the end of expiration and dropped during inspiration. E: schematic view of flows.

Figure 10.

Schematic representation of the flow direction in cannulae and the upper airways during inspiration (top) and expiration (bottom) in a symmetrical interface (SI) (left) and an asymmetrical interface (AI) (right). Blue arrows indicate nasal high flow (NHF), which is equally split between the prongs in the SI. In the AI, NHF is biased toward the larger prong due to its lower resistance and the streamline of gas velocity within the cannula. Expired gas flow is indicated by red arrows. During expiration, the SI leads to equal mixing and purging via both nares. In the AI, the nare occluded by the smaller prong creates a lower resistance path for the expired gas to be cleared from the nasal cavity. The biased flow from the larger prong is also directed to the contralateral nasal cavity via the choanae, forming the reverse flow that peaks at the end of expiration.