Abstract
BACKGROUND: The nasal cannula is widely regarded as a safe and effective means of administering low- and high-flow oxygen to patients irrespective of their age. However, variability in delivered oxygen concentration (FDO2FDO2) via nasal cannula has the potential to pose health risks. The present study aimed to evaluate predictive equations for FDO2 over a large parameter space, including variation in breathing, oxygen flow, and upper-airway geometry representative of both young children and adults.
METHODS: Realistic nasal airway geometries were previously collected from medical scans of adults, infants, and neonates. Nasal airway replicas based on these geometries were used to measure the FDO2 for low-flow oxygen delivery during simulated spontaneous breathing. The present study extends previously published data sets to include higher oxygen flows. The extended data sets included nasal cannula oxygen flows that ranged from 6 to 65 L/min for the adult replicas, and from 0.5 to 6 L/min for the infant replicas. For both age groups, FDO2 was measured over a range of breathing frequencies, inspiratory to expiratory time ratios, and tidal volumes. Measured FDO2 values were compared with values predicted by using a previously derived flow-weighted equation.
RESULTS: For both age groups, FDO2 was observed to increase nonlinearly with the ratio between oxygen flow supplied to the nasal cannula and the average inhalation flow. The previously derived flow-weighted equation over-predicted FDO2 at higher oxygen flows. A new empirical equation, therefore, was proposed to predict FDO2 for either age group as a function of nasal cannula flow, tidal volume, and inspiratory time. Predicted FDO2 values matched measured values, with average relative errors of 2.4% for infants and 4.3% for adults.
CONCLUSIONS: A new predictive equation for FDO2 was obtained that accurately matched measured data in both adult and infant airway replicas for low- and high-flow regimens.
Introduction
Supplemental oxygen therapy is commonly administered to patients with acute or chronic respiratory failure across a broad range of conditions, patient categories, and health-care settings.1-5 Despite this widespread use and the countless benefits of oxygen therapy as a standard of care, risks associated with uncontrolled oxygen administration are often underappreciated. Evidence is accumulating to support the need to titrate oxygen delivery so as to maintain oxygenation within target ranges. For example, in exacerbations of COPD, titrated oxygen treatment delivered by nasal cannula has been shown to more effectively reduce mortality, hypercapnia, and respiratory acidosis than 100% oxygen administered at a flow of 8 to 10 L/min through a non-rebreather face mask.6
Current guidelines also recommend that oxygen therapy be titrated against pulse oximetry in the management of asthma exacerbations for both adults and children7,8 because controlled or titrated oxygen therapy is associated with lower mortality and better outcomes than is high concentration (100%) oxygen therapy.9-12 For long-term oxygen therapy, oxygen flows are titrated against pulse oximetry at an initial visit after eligibility for oxygen therapy has been confirmed, although practices vary with regard to re-evaluation of oxygen flows at follow-up visits.13,14 Oxygen use during neonatal resuscitation remains a subject of debate and controversy,15 with the most recent recommendation being to commence resuscitation with air before possibly following up with oxygen therapy.16 In addition, results of observational studies have suggested that giving oxygen to preterm infants in high concentrations may have more negative effects than positive benefits.17-20
For open patient interfaces, for example, nasal cannulas, oxygen administration is complicated by uncontrolled variability in the delivered oxygen concentration (FDO2). Indeed, results of studies show that, for adults in both resting and rapid breathing conditions, nasal cannula oxygen administration results in variable absolute amounts of delivered oxygen and different values of FDO2, even when the oxygen flow is fixed.21-25 Variability in FDO2 can lead to harmful effects.31-33
To account for this variability, recent studies have proposed predictive equations as a noninvasive means of estimating FDO2 as a function of oxygen flow and breathing parameters.21,22,34-37 In vitro, or benchtop, test apparatuses that incorporate realistic airway replicas and simulated breathing have proven helpful in evaluating oxygen delivery by providing a highly controlled testing environment for conducting repeatable experiments.21-23,38-40 Based on airway geometries segmented from medical images, these airway replicas are more closely representative of human upper airways than are highly simplified geometries previously used for evaluating inhaled concentrations of medical gases.41
Most recently, Sabz et al30 conducted an in vitro investigation to evaluate variation in FDO2 across the full range of experimental data.30
In the present work, we aimed to unify our findings with previous work conducted on adult22 and infant30 airway replicas by extending the range of oxygen flow considered for both age groups. We hypothesize that a single correlation can be developed to accurately predict variation in FDO2 with breathing pattern in both adult and infant airway replicas for low- and high-flow regimens.
QUICK LOOK
Current Knowledge
The nasal cannula is used to administer low- and high-flow oxygen to patients of all ages. However, considerable variability in delivered oxygen concentration (FDO2 with breathing parameters and oxygen supply flow has previously been described by using a simple equation that assumes ideal gas mixing in the upper airway, recent studies have highlighted inaccuracies in this description.
What This Paper Contributes to Our Knowledge
A wide range of parameters that influence FDO2 was obtained that accurately matched measured data in both adult and infant airway replicas for low- and high-flow regimens.
Methods
Nasal Airway Replicas
Infants.
Previously, Tavernini et al42 developed 7 neonatal and infant replicas based on computed tomographies of children between the ages of 5 d and 3 months. Replicas were constructed by using photopolymers and included the face, nasal vestibule and valve, the turbinate and nasopharynx, and the trachea.42 These replicas were used by Sabz et al30 to investigate FDO2 closest to the average values reported by Sabz et al.30
Adults.
In addition, 15 adult nasal airway replicas were developed previously by Chen et al21 based on computed tomography and magnetic resonance images of subjects ranging from 27 to 72 years old. A single replica (female; airway replica volume, 35,900 mm3; airway surface area, 23,500 mm2) from these 15 replicas was used by Katz et al22 to evaluate FDO2 representative of high-flow oxygen therapy, as described below.
Experimental Design
The experimental setup is illustrated in Figure 1. Tidal breathing was simulated by using a lung simulator (ASL 5000, IngMar Medical, Pittsburgh, Pennsylvania), with the chamber volume recorded at a frequency of 512 Hz. The lung simulator was operated in its flow-pump mode to impose user-specified breathing patterns. Airway replicas were continuously supplied oxygen through nasal cannulas (1611-7-50 or 1601-7-50 for infant replicas, Salter Labs [Arvin, California]; model 1103 for adult replicas, Hudson RCI [Teleflex Medical, Research Triangle Park, North Carolina]) from an oxygen cylinder connected to a mass flow controller (MCMC-Series Mass Flow Controller, Alicat [Tucson, Arizona]). A laser diode analyzer (GA-200, iWorx [Dover, New Hampshire]), was used to sample the gas at the exit of the replica (representing the trachea) at a sampling flow of 200 mL/min. The oxygen concentration was measured at a frequency of 34.88 Hz. The oxygen concentration, measured by the analyzer, and the volume of the chamber of the test lung, recorded by the lung simulator, were streamed via RS-232 serial port and TCP/IP, respectively, to a desktop computer, where they were accessed by Virtual Instrument Software Architecture and synchronized by using LabVIEW (National Instruments, Austin, Texas).
A schematic of the experimental setup.
For adult replicas, 22-mm internal diameter breathing circuit tubing connected the replica outlet to the lung simulator. The internal volume of this tubing was 135 cm3, which approximates the volume of the conducting airways from the trachea to the terminal bronchioles for an average adult with a functional residual capacity of 3 L.43 For the infant replicas, the total volume of the connecting tube was reduced to 13.8 cm3 by minimizing the internal volume of connectors on either side of the gas sampling port. This volume is representative of the volume of the intrathoracic conducting airways of a 6-kg infant.44
Simulated inhalation and exhalation flows over time were calculated from the chamber volume data by using the forward difference method for numerical differentiation. At the start of each experiment, oxygen was administered through the nasal cannula before starting the lung simulator to achieve a characteristic drop in oxygen concentration as soon as the simulated breath began. It was then possible to synchronize the oxygen flow and concentration data by matching the inspiratory flow's start time with this oxygen concentration decrease. Synchronized flow and oxygen concentration waveforms agreed with oxygen waveform features described in the literature,21 according to which the lowest oxygen fraction corresponds to the highest inhalation flow due to room air dilution. Moreover, this synchronization technique minimized the difference between the amounts of oxygen inhaled and exhaled. Because there was no oxygen uptake mechanism in the experimental setting, the difference is zero for an ideal synchronization at steady state.
For each breathing cycle, the times that inspiration started and ended were identified as the times of zero flow. The volume of oxygen inhaled past the trachea during each breath was then determined by integrating the product of oxygen concentration and flow by using the trapezoid rule. The inspiratory flow was integrated over the inspiratory period to calculate the FDO2 was averaged over 5 consecutive breaths after a steady state in expiratory oxygen concentration was observed (after at least 50 breaths). Experiments were repeated on 3 separate days, with the nasal cannula repositioned for each set of measurements.
Selection of Breathing Parameters
Infants.
In newborns and infants with respiratory diseases, including acute bronchiolitis, lower respiratory illness, asthma, and severe chronic lung disease, various studies were conducted to determine typical ranges of breathing parameters.45-47 Based on these studies, Sabz et al30 used the representative breathing parameters shown in Table 1 during their investigation of low-flow oxygen delivery in infants. Breathing parameters for the present study were selected from this table to investigate high-flow oxygen delivery in infant replicas.
Breathing Parameters Used With Infant Replicas
Adults.
Chen et al21 defined 3 breathing patterns representative of adult patients with COPD, as reported during rest and light exercise by Chatila et al48 and during sleep by Hudgel et al.49 These are summarized in Table 2. Katz et al22 used the same breathing parameters in their study, which included FDO2 up to 65 L/min, as described below.
Breathing Parameters Used With Adult Replicas
Selection of Nasal Cannula Flows
For both the infant and adult nasal airway replicas, separate ranges of nasal cannula flow were defined to represent low- and high-flow oxygen delivery.
Low-Flow Oxygen Delivery.
Previous measurements conducted by Sabz et al30 evaluated low-flow oxygen delivery in 7 infant replicas for FDO2 < 6 L/min, measured values previously reported by Katz et al22 were used.
High-Flow Oxygen Delivery.
High-flow oxygen delivery, as defined in the context of this study, refers to the administration of oxygen at flows > 2 L/min for infants50-53 and > 15 L/min for adults.54,55 New experiments were conducted in the present study for both infant and adult airway replicas to evaluate FDO2 for infants and adults undergoing low- and high-flow oxygen delivery are summarized in Table 4.
Breathing Parameters and Nasal Cannula Flow for High-Flow Oxygen Delivery
Sources of Data Used for Analysis of FDO2 via Nasal Cannula
Results
Comparison of Measured FDO2 with Previous Studies
As shown in Figure 2, values of FDO2 measured in the present study for the adult replica demonstrated close agreement with the average values for 15 adult replicas studied by Chen et al21 and with the values measured by Katz et al22 for the same individual adult replica used in this study.
A comparison of values of delivered oxygen concentration (FDO2 = nasal cannula flow.
Predicting FDO2 Values By Using a Flow-Weighted Calculation
Measured values of FDO2 was investigated, with the results presented in the next section.
Measured values of delivered oxygen concentration (FDO2 = nasal cannula flow.
A Universal Equation for Predicting FDO2 Delivered via Nasal Cannula
Measured FDO2 for both age groups as well as for both the low- and high-flow ranges.
Measured values of delivered oxygen concentration (FDO2 = inspiratory flow.
A variety of functions were examined to find the best fit, including third-order polynomial, asymptotic regression, logistic growth, and Weibull growth functions. In the case of nonlinear regression, the standard error of the regression was used as a reference point to determine the most suitable function. Although standard error of the regression values were similar between the best-fit equation formed when using the asymptotic regression, logistic growth, and the Weibull growth functions, ultimately, the equation formed by using the logistic growth function was selected owing to its closest agreement with the physically realistic value of FDO2 = 0. This equation is provided below:
(5)
Equation (5) improved the accuracy of predicted FDO2 in comparison with predictions made by using Equation (4), reducing the average relative error from 21.5% to 2.4% and 12.6% to 4.3% for infants and adults, respectively.
Equation (5) is plotted in Figure 4 alongside data obtained for the adult and infant airway replicas, demonstrating a good fit to experimental data for both age categories. FDO2 and average flow is the following:
(6)
Therefore, when FDO2 can be expected to approach a value of 1, as is seen in Figure 4.
Bland-Altman plots are provided in Figure 5 to further compare the accuracy of FDO2 on their average value (seen in Figure 5A) was reduced for both adults and infants (seen as the data distributed around a horizontal line in Figure 5B).
Bland-Altman plots based on measured and calculated values by using A: Equation (4), and B: Equation (5), for infants and adults in low- and high-flow oxygen delivery.
Discussion
This study evaluated FDO2 across the full range of flows studied.
An accurate estimation of the FDO2, to patients receiving oxygen via nasal cannula.57,58 To achieve these goals, especially to support clinical decision-making, accurate real-time monitoring of breathing parameters will be critical.
The present study relied on FDO2 measurements in previous studies.21,30
Moreover, simulated breathing experiments did not include gas exchange, so gases exhaled back through replicas during tidal breathing remained oxygenated and devoid of carbon dioxide. Previously, it was shown that exhaled gases are unlikely to be re-inhaled from room air near the nares59; however, for the experimental methods used herein, gas contained within nasal airway replicas at the end of expiration will be re-inhaled past the oxygen sampling point (Fig. 1). This introduces the potential for FDO2 and values reported in vivo for adults (sampled from the trachea through a catheter) during oxygen delivery via nasal cannula.60
The present study explored limited variation in respiratory parameters, VT, breathing frequency, and I:E, during high-flow oxygen delivery. It is acknowledged that clinical scenarios can involve a broader range of respiratory parameter combinations, and our findings should be interpreted within the context of this limitation. In addition, our study introduced variability in FDO2 under controlled conditions.
Finally, the present study demonstrated an accurate estimation of FDO2 for older children. Less clear is the applicability of Equation (5) in the presence of upper-airway obstruction, especially nasal or nasopharyngeal obstruction, which could potentially alter mixing between oxygen and entrained room air.
Conclusions
FDO2 was obtained based on logistic growth function that accurately matched measured data in both adult and infant airway replicas for low- and high-flow regimens.
Footnotes
- Correspondence: Andrew R Martin PhD, 10–265 Donadeo Innovation Centre for Engineering, University of Alberta, Edmonton, AB T6G 1H9, Canada. E-mail: andrew.martin{at}ualberta.ca
Financial support was provided by the Natural Sciences and Engineering Research Council of Canada.
A version of this paper was presented by Dr Martin at the annual conference of the American Thoracic Society, held in San Francisco, California, May 15–18, 2022.
The authors have disclosed no conflicts of interest.
- Copyright © 2024 by Daedalus Enterprises