Printer Friendly

Evaluation of a humidified nasal high-flow oxygen system, using oxygraphy, capnography and measurement of upper airway pressures.

Several devices have recently entered clinical use allowing the delivery of humidified high-flow oxygen via the nasopharynx. Oxygen therapy of this type has several potential advantages over conventional low flow devices and mask oxygen, including minimal variation of the inspired oxygen fraction, humidification and warming, improved comfort (1-3), the ability to eat, talk and cough easily and the delivery of a positive airway pressure with high gas flow rates (T. David, Canterbury University, New Zealand, personal communication).

Clinically it has been found that exercise capacity is improved in patients with chronic obstructive pulmonary disease while receiving humidified nasal oxygen at 20 1/minute (4).

One example, the Optiflow[TM] system, has been developed in New Zealand by Fisher and Paykel Healthcare (East Tamaki, Auckland). The manufacturer states that the system can deliver up to 60 litres per minute of an oxygen/air mixture, at optimal humidity (44 mg water vapour/Lat 37[degrees]C) via a soft silicon nasal interface.

This study was undertaken to define the performance of this system by measuring expired oxygen fractions and upper airway pressure in volunteers in an experimental setting.


The study was approved by the Northern Y Ethics Committee, New Zealand.


Healthy volunteers over the age of 18 years were invited to participate. Exclusion criteria were: pregnancy, allergy to local anaesthetic agents, upper respiratory tract disease or anatomical deformity and cardiac or respiratory disease precluding heavy exercise. Written, informed consent was obtained from each participant.


The system delivers an air/oxygen mixture from a blender (Bird 3800 Microblender) through a variable area flow meter (Key Instruments) to the Optiflow system. This consists of a heated humidifier (Fisher and Paykel Healthcare MR880), heated tubing (Fisher and Paykel Healthcare RT241) and a nasal interface (Fisher and Paykel Healthcare RT034). The nasal interface is made of soft silicon, with a wider bore than traditional nasal cannulae, and is designed specifically to deliver humidified gas.

Figure 1 shows the gas delivery and measurement circuit used for this study. A fixed oxygen fraction of 0.6 was used throughout the study. The delivery system was calibrated for flow using a TSI 4040 flow meter (TSI Flow Meters Ltd, Co Laois, Ireland) before passing through the humidifier. The system was also calibrated for delivered oxygen fraction using a direct connection to the gas analysis port of a Datex AS/3 anaesthetic monitor after each change in flow rate.

A 10 Fr PVC catheter was inserted through the nose to the hypopharynx of each participant under topical local anaesthesia (lignocaine spray 10%). Correct placement was confirmed by ensuring the tip was past the oropharynx on visual inspection and by ensuring a stable capnograph trace after connection to the gas analysis port of the Datex AS/3 monitor.

Oxygraphy and capnography of inhaled and exhaled airway gases and airway pressure measurements were obtained by direct connection to the end of the hypopharyngeal catheter. Gas analysis was performed by the Datex AS/3 monitor. This uses sidestream, breath by breath, infrared capnography and paramagnetic oxygraphy. It samples gases at a rate of 200 m1/minute (+/-20 m1/minute). Pharyngeal pressure was measured using a Honeywell precision pressure transducer (PPT - 0001 DWW2VA-B, Honeywell International Ltd) with a laptop computer interface.


Peak inspiratory flow rate (PIFR) was measured at rest and with exercise using the Spiroson-AS ultrasonic flow sensor with PC interface (ndd Medizintechnik AG). This has a flow range of up to 16 l/second and an accuracy of +/- 3% or 20 ml/second (manufacturer's product information sheet).

Each participant was monitored with electrocardiogram, using the Datex AS/3 monitor.


Hypopharyngeal pressures and fractions of inspiratory [O.sub.2] (Fi[O.sub.2]), end-tidal [O.sub.2] (FE[O.sub.2]), and end-tidal C[O.sub.2] ([F.sub.E]C[O.sub.2]) were measured from the hypopharyngeal catheter. Each participant, while wearing the nasal interface at rest, was asked to breathe through their nose with their mouth closed. A fixed oxygen fraction of 0.6 at flow rates of 10, 20, 30, 40 and 50 1/minute was delivered in a random order. Recordings were made after a period of stabilisation, ensuring at least four breaths of stable capnography and oxygraphy data. Measurements were then repeated at one minute to ensure a steady state. This process was then repeated for each participant breathing through their mouth for each delivered flow rate.

Respiratory rate and heart rate were monitored throughout. Randomisation of delivered flow rates for each subject was performed using a computer generated randomisation table. Participants were blinded to the flow settings.

Participants were then asked to exercise on a stationary bicycle. Participants increased their exercise work rate to achieve a PIFR of greater than 100 1/minute. Heart rate was measured at this point and a stable work rate was achieved by maintaining the heart rate at a constant level. Fi[O.sub.2], [F.sub.E][O.sub.2], [F.sub.E]C[O.sub.2], respiratory rate and heart rate were recorded while wearing the nasal interface and breathing through the nose for each delivered flow rate in descending order from 50 to 10 1/minute.


Fi[O.sub.2] was calculated from the measured [F.sub.E]C[O.sub.2] and [F.sub.E][O.sub.2] by rearranging the alveolar gas equation (5) assuming that: [F.sub.E][O.sub.2] is equal to alveolar oxygen fraction ([F.sub.A][O.sub.2]), [F.sub.E]C[O.sub.2] is equivalent to alveolar carbon dioxide fraction ([F.sub.A]C[O.sub.2]) and the respiratory quotient (RQ) for each participant =0.8 (Figure 2).


Statistical analysis of this hierarchically structured data (different flow rates within each subject) was with linear mixed models. In particular, each subject was assumed to have his or her own relationship between response and flow rate, such that each regression coefficient is a random sample from some population of possible coefficients: the random coefficient mixed model (6). Plots were used to display the raw data and best fitting models. The usual residual and other diagnostics were performed. Statistical significance was assumed at P <0.05.


Ten subjects (eight men and two women) completed the study (Table 1) Mean PIFR at rest was 27.9 (SD 9.23) 1/minute and with exercise 119.9 (SD 20.04) 1/minute. All but one of the participants (who achieved greater than 90 1/minute) were able to generate a PIFR of greater than 100 1/minute.

Table 2 gives the mean values for [F.sub.E][O.sub.2], [F.sub.E]C[O.sub.2] and calculated Fi[O.sub.2] for all three breathing patterns versus delivered flow rates.

Figures 3, 4, 5 and 7 show individual, raw data, means per flow rate and fitted curves plotted against delivered gas flow rate.

For calculated Fi[O.sub.2] (Figure 3) there is a significant upward trend with increasing flow rate (P <0.001) for all breathing patterns. Calculated Fi[O.sub.2] is highest, for all delivered gas flow rates, when breathing at rest through the nose, and lowest when exercising to a PIFR of >90 1/minute. There is a significant difference between the calculated Fi[O.sub.2] for each of the three breathing patterns (P <0.001).

For respiratory rate (Figure 4), there are significant differences between the three breathing patterns (P <0.001). Respiratory rate is highest with exercise and lowest when breathing at rest with the mouth closed. There is also a downward trend of respiratory rate plotted against delivered gas flow rate for all three breathing patterns; this is not statistically significant (P=0.15).




For [F.sub.E]C[O.sub.2] (Figure 5), there are again significant differences between the three breathing patterns (P <0.001), highest with exercise and lowest when breathing at rest with the mouth open. There is no change seen with increasing flow rate.

Figure 6 is a plot of upper airway pressure versus time for one participant breathing 10 and 40 1/minute through the nose and 40 1/minute with the mouth open. Peak pressures are noticeably higher at the higher delivered gas flow rate and with breathing with the mouth closed rather than with the mouth open.


For mean upper airway pressure (Figure 7) both slope (P <0.001) and means of individual data (P <0.001) with mouth open and closed showed differences; with mouth closed in particular showing increasing pressure with increasing delivered gas flow rate. Here there is almost a linear response in mean upper airway pressures with increasing delivered gas flow rates above 30 1/minute; at 30 1/minute the mean upper airway pressure approaches 3 cm[H.sub.2]0, at 40 1/minute it is approximately 4 cm[H.sub.2]0 and at 50 1/minute the mean upper airway pressure is approximately 5 cm[H.sub.2]0.


This study evaluates a high-flow nasal oxygen delivery system in healthy volunteers. We found that the calculated Fi[O.sub.2] approached the prescribed Fi[O.sub.2] when the delivered gas flow rates were greater than participant PIFR. Furthermore, when the subjects breathed with their mouth closed, the system delivered a clinically relevant mean positive airway pressure which was directly proportional to delivered gas flow rates.

The method of pharyngeal oxygraphy and capnography is well described and has been found to be reliable (7-14). Two other papers have also described this method to measure upper airway pressures (15,16) and have had similar findings to this study. Tracheal sampling via a cricothyroid puncture has been used in the assessment of Fi[O.sub.2] (17) but this is invasive and at least three deaths have been attributed to this technique (18). Hypopharyngeal sampling compares favourably to tracheal sampling (9), is less invasive and is better tolerated by the awake subject.


The Fi[O.sub.2] was calculated by rearrangement of the alveolar gas equation (Figure 2) (7,8). Fi[O.sub.2] was measured directly, but concerns regarding gas streaming in the upper airway causing incomplete mixing in the hypopharynx have previously been raised and this has been reason for criticism of a recent paper (19,20). The Riley version of the equation was used in this study to allow for the difference in inspired and expired volumes (5). Several assumptions need to be made to use this method, which results in a derivative of the [F.sub.E][O.sub.2], which is, physiologically, the most relevant indicator of alveolar oxygen fraction ([F.sub.A][O.sub.2]).

We used a single oxygen fraction of 0.6 in this study. This was chosen to be high enough to show significant differences with each flow rate while minimising potential adverse effects from high concentration oxygen, in particular, absorption atelectasis.

We observed a constant, small but statistically significant difference for calculated Fi[O.sub.2] between nose breathing and mouth breathing at rest (Figure 3). There are three possible reasons for this observation. The first is an increase in air entrainment caused by the participant's inspiratory flow when the mouth is open. This was also described by Lomholt in 19682. If this were the only cause, we would have expected this effect to decrease with increasing delivered flow rates causing the two curves to converge; this wasn't seen in this study.

A second possible cause is sampling error; the Datex AS3 samples from the pharyngeal catheter at a rate of approximately 200 m1/minute. It is possible that this causes entrainment of air into the sampling line when the mouth is open. This phenomenon could also explain the similar difference seen in [F.sub.E]C[O.sub.2] between nose and mouth breathing (Figure 5).

Third, a venturi effect might explain the difference between mouth and nose breathing. When the mouth is open, increased delivered flow rates are associated with increasing air entrainment. The system with the mouth closed is sealed enough to prevent this effect.

It is possible that all of these effects are present and have a net effect which results in a relatively constant difference in expired oxygen between mouth and nose breathing, as delivered flow rate increases.

We used exercise to simulate respiratory distress. Previous studies have simulated increased inspiratory flow rates using a model (21) or have asked participants to increase their respiratory rate voluntarily using a metronome (9,11). The high PIFR with exercise attained in this study was associated with lower calculated Fi[O.sub.2] than those measured at rest (Figure 3). Furthermore the curves are more linear with exercise. It is likely that the gas flow rates used in this study are not high enough to minimise air entrainment when the participant is breathing with a PIFR in excess of 90 1/minute. It is not known how clinically relevant these findings are as we were unable to find any published data describing the breathing pattern, particularly peak inspiratory flow rates, in patients with acute respiratory disease.

Positive end-expiratory pressure is an important treatment for a number of acute hypoxic diseases. The high flows delivered by the studied system act as a resistance to exhalation and deliver a clinically significant mean airway pressure when the mouth is closed. The positive mean airway pressures in this study were similar to end-expiratory oral pressures measured in volunteers using the same system and wearing the same nasal interface in an unpublished study from Canterbury University, New Zealand (T. David, personal communication). The positive airway pressure delivered by such devices may explain part of their efficacy in treating hypoxemia.

A third proposed benefit of this type of system is the ability to deliver warmed humidified gas to the nasopharynx and upper airway. Mucosal function is impaired above and below the optimum level of temperature and humidity; i.e. core temperature and 100% relative humidity (22). The humidification of delivered oxygen has been found to be beneficial for patients in other studies (23,24). We chose to focus on the delivery of Fi[O.sub.2] and positive airway pressure and we await further humidity data with anticipation.

Respiratory rate was not a primary measurement in this study. However, the data suggest that there could be a decrease in respiratory rate with an increase in delivered gas flow rate for all breathing patterns (Figure 4). Another study comparing 15 l/minute of delivered oxygen through a Hudson mask and nasal cannula found that respiratory rate was lower when using the nasal cannula (25). This might be followed by a decrease in minute volume but [F.sub.E]C[O.sub.2] measurements in our study do not support this in healthy volunteers. It is likely that tidal volume is increased in proportion to the decrease in respiratory rate. It is also likely that the high flows delivered by this system washout anatomical dead space, contributing to the effect seen on [F.sub.E]C[O.sub.2]. Further work on this is needed in patients with respiratory failure.


The studied oxygen delivery system delivers a prescribed Fi[O.sub.2] at delivered gas flow rates greater than peak inspiratory flow rate whether or not the wearer is breathing through their mouth or nose. However the system could not be considered a fixed oxygen delivery system as described by Leigh (26) as accuracy of the system depends on the wearer's breathing pattern.

With the mouth closed, the system delivers a clinically relevant positive airway pressure proportional to the delivered gas flow rate.


We would like to thank Kevin Butler (Charge Biomedical Engineer, Middlemore Hospital), Craig White (Product Development Manager, Fisher and Paykel Healthcare Limited), Steven Korner (Product Development Engineer, Fisher and Paykel Healthcare Limited), the participants and staff of the Critical Care Complex, Middlemore Hospital.

Source of support

Fisher and Paykel Healthcare Limited provided equipment and technical support for this study. They also provide a research grant to Counties Manukau District Health Board, Auckland, New Zealand.


(1.) Kory RC, Bergmann JC, Sweet RD, Smith JR. Comparative evaluation of oxygen therapy techniques. JAMA 1962; 179:123-128.

(2.) Lomholt N. Continuous controlled humidification of inspired air. Lancet 1968; 2:1214-1216.

(3.) Roca O, Riera J, Torres F, Masclans JR. High-flow oxygen therapy in acute respiratory failure. Respir Care 2010; 55:408-413.

(4.) Chatila W, Nugent T, Vance G, Gaughan J, Criner GJ. The effects of high-flow vs low-flow oxygen on exercise in advanced obstructive airways disease. Chest 2004; 126:1108-1115.

(5.) Nunn J. Nunn's Applied Respiratory Physiology, 4th ed. Oxford: Butterworth-Heinemann Ltd 1993; p. 658.

(6.) Littell RC, Milliken GA, Stroup WW, Wolfinger RD, Schabenberger O. Random Coefficient Models. In: SAS for Mixed Models, 2nd ed. SAS Publishing 2006. p. 321-348.

(7.) Waldau T, Larsen VH, Bonde J. Evaluation of five oxygen delivery devices in spontaneously breathing subjects by oxygraphy. Anaesthesia 1998; 53:256-263.

(8.) Slessarev M, Somogyi R, Preiss D, Vesely A, Sasano H, Fisher JA. Efficiency of oxygen administration: sequential gas delivery versus "flow into a cone" methods. Crit Care Med 2006; 34:829-834.

(9.) Schacter EN, Littner MR, Luddy P, Beck GJ. Monitoring of oxygen delivery systems in clinical practice. Crit Care Med 1980; 8:405-409.

(10.) Larsen VH, Waldau T, Oberg B. Oxygraphy in spontaneously breathing subjects. Acta Anaesthesiol Scand Suppl 1995; 107:81-85.

(11.) Dunlevy C, Tyl S. The effect of oral versus nasal breathing on oxygen concentrations received from nasal cannulas. Respiratory Care 1992; 37:357-360.

(12.) Waldau T, Oberg B, Larsen VH. Reliability of C[O.sub.2] measurements from the airway by a pharyngeal catheter in unintubated, spontaneously breathing subjects. Acta Anaesthesiol Scand 1995; 39:637-642.

(13.) Lenz G, Heipertz W, Epple E. Capnometry for continuous postoperative monitoring of nonintubated, spontaneously breathing patients. J Clin Monit 1991; 7:245-248.

(14.) Oberg B, Waldau T, Larsen VH. The effect of nasal oxygen flow and catheter position on the accuracy of end-tidal carbon dioxide measurements by a pharyngeal catheter in unintubated, spontaneously breathing subjects. Anaesthesia 1995; 50:695698.

(15.) Groves N, Tobin A. High flow nasal oxygen generates positive airway pressure in adult volunteers. Aust Crit Care 2007; 20:126-131.

(16.) Parke R, McGuinness S, Eccleston M. Nasal high-flow therapy delivers low level positive airway pressure. Br J Anaesth 2009; 103:886-890.

(17.) Robertson GS. Cricothyroid puncture in the assessment of equipment for postoperative oxygen therapy. Lancet 1969; 1:801-803.

(18.) Unger K, Moser K. Fatal complcation of transtracheal aspiration. Archives of Internal Medicine 1973; 132:437-439.

(19.) Wettstein RB, Shelledy DC, Peters JI. Delivered oxygen concentrations using low-flow and high-flow nasal cannulas. Respir Care 2005; 50:604-609.

(20.) Casaburi R. Assessing the dose of supplemental oxygen: let us compare methodologies. Respir Care 2005; 50:594-595.

(21.) Foust GN, Potter WA, Wilons MD, Golden EB. Shortcomings of using two jet nebulizers in tandem with an aerosol face mask for optimal oxygen therapy. Chest 1991; 99:1346-1351.

(22.) Williams R, Rankin N, Smith T, Galler D, Seakins P. Relationship between the humidity and temperature of inspired gas and the function of the airway mucosa. Crit Care Med 1996; 24:1920-1929.

(23.) Rea H, McAuley S, Jayaram L, Garrett J, Hockey H, Storey L, et al. The clinical utility of long-term humidification therapy in chronic airway disease. Respir Med 2010; 104:525-533.

(24.) Hasani A, Chapman TH, McCool D, Smith RE, Dilworth JP, Agnew JE. Domiciliary humidification improves lung mucociliary clearance in patients with bronchiectasis. Chron Respir Dis 2008; 5:81-86.

(25.) Lund J, Holm-Knudsen RJ, Nielsen J, Foge Jensen PB. Nasal cannula versus hudson facemask in oxygen therapy. Journal of the Danish Medical Association 1996; 158:4077-4079.

(26.) Leigh JM. Variation in performance of oxygen therapy devices. Anaesthesia 1970; 25:210-222.

J. E. RITCHIE *, A. B. WILLIAMS ([dagger]), C. GERARD ([double dagger]), H. HOCKEY ([section])

Critical Care Complex, Middlemore Hospital, Auckland, New Zealand

* M.B., Ch.B. (Auckland), F.A.N.Z.C.A., F.C.I.C.M., Intensive Care Specialist.

([dagger]) M.B., Ch.B., F.A.N.Z.C.A., F.C.I.C.M., Specialist.

([double dagger]) B.Sc. (Hons), Clinical Research Scientist, Fisher and Paykel Healthcare Limited.

([section]) M.Sc., Statistician, Biometrics Matters Limited, Hamilton.

Address for correspondence: Dr J. Ritchie, email: jo.ritchie@middlmeore.

Accepted for publication on June 28, 2011.
Table 1
Participant demographics

Demographic Mean Range

Age, y 30.1 23-43
Height, m 1.79 1.52-1.89
Weight, kg 78.4 52-95

Table 2
Mean values for [F.sub.E][O.sub.2], [F.sub.E]C[O.sub.2] and calculated
Fi[O.sub.2], for all three breathing patterns versus delivered flow

Delivered Nose breathing at rest
flow rate
 [F.sub.E] [F.sub.E]
 [O.sub.2] C[O.sub.2] Fi[O.sub.2]

50 l/min 0.511 0.052 0.568
40 l/min 0.488 0.059 0.550
30 l/min 0.458 0.055 0.519
20 l/min 0.404 0.054 0.465
10 l/min 0.309 0.053 0.369

Delivered Mouth breathing at rest
flow rate
 [F.sub.E] [F.sub.E]
 [O.sub.2] C[O.sub.2] Fi[O.sub.2]

50 l/min 0.484 0.046 0.535
40 l/min 0.448 0.048 0.502
30 l/min 0.431 0.049 0.486
20 l/min 0.361 0.048 0.416
10 l/min 0.284 0.051 0.342

Delivered Nose breathing with exercise
flow rate
 [F.sub.E] [F.sub.E]
 [O.sub.2] C[O.sub.2] Fi[O.sub.2]

50 l/min 0.328 0.071 0.408
40 l/min 0.289 0.070 0.370
30 l/min 0.242 0.068 0.321
20 l/min 0.235 0.066 0.313
10 l/min 0.188 0.064 0.263

[F.sub.E][O.sub.2]=fraction of end-tidal [O.sub.2],
[F.sub.E]C[O.sub.2] =fraction of end-tidal C[O.sub.2],
Fi[O.sub.2]=fraction of inspiratory [O.sub.2].

Figure 2: Riley version of the alveolar gas equation (Equation 1)3
and rearrangement for calculation of Fi[O.sub.2] (Equation 2).

Equation 1 [P.sub.a][O.sub.2] = [P.sub.i][O.sub.2] -
[P.sub.A]C[O.sub.2]/RQ [1 - Fi[O.sub.2](1 - RQ)]

Equation 2 Fi[O.sub.2] = [F.sub.E][O.sub.2] x RQ +
[F.sub.E]C[O.sub.2]/[RQ + [F.sub.E]C[O.sub.2](1 - RQ)]
COPYRIGHT 2011 Australian Society of Anaesthetists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ritchie, J.E.; Williams, A.B.; Gerard, C.; Hockey, H.
Publication:Anaesthesia and Intensive Care
Article Type:Report
Geographic Code:8NEWZ
Date:Nov 1, 2011
Previous Article:A comparison of the forces applied to a manikin during laryngoscopy with the GlideScope[R] and Macintosh laryngoscopes.
Next Article:Benefits and risks associated with the R100 high frequency oscillatory ventilator for patients with severe hypoxaemic respiratory failure.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters