A noninvasive high frequency oscillation ventilator: Achieved by utilizing a blower

and a valve
YueYang Yuan, JianGuo Sun, Baicun Wang, Pei Feng, and ChongChang Yang

Citation: Rev. Sci. Instrum. 87, 025113 (2016); doi: 10.1063/1.4942048

View online: http://dx.doi.org/10.1063/1.4942048
View Table of Contents: http://aip.scitation.org/toc/rsi/87/2
Published by the American Institute of Physics
 REVIEW OF SCIENTIFIC INSTRUMENTS 87, 025113 (2016)

A noninvasive high frequency oscillation ventilator: Achieved

by utilizing a blower and a valve
YueYang Yuan,1,2 JianGuo Sun,2 Baicun Wang,3 Pei Feng,1 and ChongChang Yang1
1
College of Mechanical Engineering, Donghua University, Shanghai 210620, China
2
Curative Medical Technical Co., Ltd., Suzhou 215163, China
3
College of Chemical and Biological Engineering, Zhejiang University, 38 Zheda Road,
Hangzhou 310027, China
(Received 26 July 2015; accepted 3 February 2016; published online 24 February 2016)
After the High Frequency Oscillatory Ventilation (HFOV) has been applied in the invasive ventilator,
the new technique of noninvasive High Frequency Oscillatory Ventilation (nHFOV) which does not
require opening the patient’s airway has attracted much attention from the field. This paper proposes
the design of an experimental positive pressure-controlled nHFOV ventilator which utilizes a blower
and a special valve and has three ventilation modes: spontaneous controlled ventilation combining
HFOV, time-cycled ventilation combining HFOV (T-HFOV), and continuous positive airway pressure
ventilation combining HFOV. Experiments on respiratory model are conducted and demonstrated the
feasibility of using nHFOV through the control of fan and valve. The experimental ventilator is able
to produce an air flow with small tidal volume (VT) and a large minute ventilation volume (MV)
using regular breath tubes and nasal mask (e.g., under T-HFOV mode, with a maximum tidal volume
of 100 ml, the minute ventilation volume reached 14 400 ml). In the process of transmission, there
is only a minor loss of oscillation pressure. (Under experimental condition and with an oscillation
frequency of 2-10 Hz, peak pressure loss was around 0%-50% when it reaches the mask.) C 2016 AIP
Publishing LLC. [http://dx.doi.org/10.1063/1.4942048]

I. INTRODUCTION to achieve the nHFOV is still a challenge for engineers and

With conventional mechanical ventilations, a number of
patients cannot be adequately ventilated even with intensified,
higher breathing rates and ventilation pressures. But due to
the small tidal volume and low airway pressure as well as
the capacity of supporting patient’s life, the high frequency
ventilation (HFV) has been proven to be a safe and effective
way by experienced practitioner who applied it1–3 not only
with the invasive ventilation but also with the noninvasive
ventilation. The ventilation methods of HFV mainly include
High Frequency Positive Pressure Ventilation (HFPPV), High
Frequency Jet Ventilation (HFJV), and High Frequency
Oscillatory Ventilation (HFOV), etc. The HFOV has gained
widespread attention in latest decades. And the noninvasive
high frequency oscillatory ventilation (nHFOV) which does
not require opening of a patient’s airway has become a
promising ventilation technique. The noninvasive ventilator
with achieving the HFOV methods is still in its early stage of
development.
Furthermore, in the existing invasive HFOV ventilator, the
oscillation flows and pressures are produced using a special
reciprocating piston4,5 or a high-power large loudspeaker
membrane.6 With these methods, the component of oscillator
is more complex in structure and controlling algorithms.
Additionally, in most cases, opening the patient’s airway
would be necessary when ventilating with the invasive
HFOV ventilator. The endotracheal tube (ET-tube) is always
indispensable for bypassing the patient upper airway. As a
result, the risks (such as trachea injures and surgical infections)
are unavoidable with the invasive ventilation. Therefore, how
physicians.
In addition, the positive pressure ventilation7 has been
synonymous with the growth of critical care medicine since it
came into its own in 1950s. And the noninvasive positive
pressure ventilation refers to positive pressure ventilation,
which is delivered through a noninvasive interface (nasal
mask, face mask, or nasal prong), rather than an invasive
interface (ET-tube, tracheotomy). It has become common
due to its benefits in therapy for patients with ARDS (acute
respiratory distress syndrome) and acute lung injury8–10 is
increasingly being recognized. Prompted by the interest of
the nHFOV technique, a positive pressure-controlled nHFOV
ventilator is proposed in this report.
II. METHODS
A. Design and implementation of the ventilator
To achieve the nHFOV, the detailed materials about
the designing for an experimental nHFOV ventilator are
introduced in Subsection II B of hardware and software.
As shown in Fig. 1, along the gas passageway (indicated
by the dash arrow), the gas is pumped into the ventilator from
the gas import by the high-speed blower (33 000 rpm at non-
loaded). First, the filtered gas is monitored by the first flow
sensor. Then it is pumped into the blower and exported from
the high-speed three-way pressure balance valve.11 The gas
flow from this valve is monitored by the second flow sensor and
delivered to the mask (the end of gas passageway). In the mask,
the gas flow brings a positive pressure due to the resistance

0034-6748/2016/87(2)/025113/6/$30.00 87, 025113-1 © 2016 AIP Publishing LLC

025113-2 Yuan et al.
FIG. 1. Pneumatic and electric circuit block diagram. Dashed arrow lines denote gas flow. Solid arrow lines denote electrical signal.
of patient airway.12 The pressure is monitored by the pressure
sensor. A leakage hole is deliberately placed in the mask to
eliminate CO2 (carbon dioxide). During the inspiration time,
part of flow is inhaled by the patient into airway and the rest of
the flow leaks out from the mask. During the expiration time,
gases from both the ventilator and from patient are taken out
from the mask and dissipate into the atmosphere.
The mechanism of valve for regular gas flow is explained
in Fig. 2. The valve output flow (QVOUT ) is in proportion to the
valve input flow (QVIN ). It is regulated by the valve leakage
flow (QVCON ). The leakage flow rate is in positive proportion to
the gap size (gap). And the gap size is in positive proportion
to the electric current in the valve solenoid (IV ). A smaller
current in the valve solenoid means a smaller gap size, which
means a smaller leakage flow and a larger output flow. Their
relationship can be expressed in Equation (1). K1 and K2 stand
two proportional coefficients,


 QVOUT = QVIN − QVCON

 QVLEAK = K1 × gap

. (1)

 Gap = K2 × IV


 tion Integration Differentiation) control algorithm is designed
The electric circuit system is mainly composed of a micro-
computer, sensor signal process circuits, the blower driver
circuit, the valve driver circuit, operation panel, and LCD
(Liquid Circuit Display). Through the operation panel, the user
may set the ventilation pressures, IE ratio (the time ratio of
inspiration and expiration), BPM (breathing rate per min), etc.
FIG. 2. Valve and its flow control method. Q VIN , Q VOUT , and Q VCON stand input flow, output flow, and leakage flow.
Rev. Sci. Instrum. 87, 025113 (2016)
The corresponding messages for confirming the settings
are displayed on LCD. After the ventilator is started, the
blower is driven to carry a constant gas flow into the valve. By
regulating the electric current of solenoid, the valve transports
a gas flow to the patient end in proportion. These electrical
analogy signals from two flow sensors and one pressure sensor
are processed by their signal processes circuits and fed back
to the microcontroller.
Based on the hardware, the embedded software for
this pressure-controlled ventilator is programmed into the
microcontroller. In this software, three basic ventilation modes
are prepared: the spontaneous controlled mode combining
HFOV (S-HFOV mode), time-cycled mode combining HFOV
(T-HFOV mode), and the continuous positive airway pressure
ventilation mode13 combining HFOV (CPAP-HFOV mode).
Fig. 3 describes the embedded software.
When the ventilator is powered on, its system restores
the parameters back to the last operation condition. Those
parameters can also be set by buttons in the panel and then
will be displayed on LCD. After the ventilation is started, the
corresponding output can be carried out. The PID14 (Propor-
to control the ventilator to work in preset requirements. The
blower outputs a constant gas flow to the valve, and then the
valve transports the flow to produce a positive pressure at
the patient end. Finally the system will calculate the volumes
based on the pressure and flow rate with Equation (2) and
display them on LCD,
025113-3 Yuan et al. Rev. Sci. Instrum. 87, 025113 (2016)

FIG. 4. Experimental setup.

the manometer (8252-2psi, AZ INSTRUMENT) for pressures

and with the flow meter (TSI4000, TSI) for flow rates. The
FIG. 3. Block flowchart of ventilator embedded software. normal sampling rate was designed to be 200sps (sampling
per s). A “chest” is simulated by a sealed plastic bucket. Its
leakage is less than 0.5 l/min measured at the internal pressure
QLEAK = F(PA)

 of 40 cm H2O. The bucket is connected to a manual calibration
A = Q OUT − Q LEAK

 Q
 t syringe (Jaeger, 2 l in max capacity, serial no: 720253) through





a breath tube (ϕ19 mm × L1.8 m). The piston is drawn back

VT(t) =


Q A(t)dt
,

 0 (2) and forth to imitate the process of breathing in and out. One
 T >






if
 (Q A (t) 0)
 end of another breath tube connects to the ventilator outlet
MV = (BPM) × port and the other end connects to the nasal mask (BestFit2,




 then (Q A(t))  dt
Curative Medical). The nasal mask is fixed on the “head.” In

 0 else (0) 
   the “upper airway” of the respiratory model, a gas resistance
where PA is the pressure in the nasal mask and QOUT is the flow (≈1.56 cm H2O · s/l measured at the flow 20 l/min, close to
from the ventilator. The leakage QLEAK in nasal mask is a func- normal adult upper airway resistance16) is inserted to imitate
tion of the pressure F(PA). The tidal volume VT(t) (or oscil- the normal gas resistance of upper airway.
latory volume15) is a real-time integral of Q A(t) the flow rate The test ventilator is set to work with the parameters listed
in airway. The minute ventilation volume (MV) is the amount in Table I: inspiration positive airway pressure IPAP = 18 cm
of volume inhaled during 1 min. BPM is the breathing rate H2O; expiration positive airway pressure EPAP = 6 cm H2O;
(BPM = 60/T: breath per min). T is the breathing cycle time. peak-peak inspiration oscillation pressure ∆PI = 12 cm H2O;
peak-peak expiration oscillation pressure ∆PE = 6 cm
H2O; breathing rate BPM = 12; IE ratio IE = 1; frequency
B. Experimental testing
of HFV f = 5 Hz, or 2–10 Hz; continuous positive airway
Fig. 4 illustrates an experimental setup. It is mainly pressure CPAP = 18 cm H2O; and peak-peak oscillation
comprised of the test nHFOV ventilator introduced above, pressure ∆P = 12 cm H2O, 10 cm H2O, or 5 cm H2O
hypothetical respiratory model, and data acquisition equip- subsequently. The pressures in the mask, the flow rates, and
ment. The respiratory model includes two fresh sheep lungs volumes in the hypothetical respiratory model were measured.
obtained from market. Each sheep weights approximately Every experiment was double executed and continued for more
25 kg. The data acquisition equipment was calibrated with than 20 min after ventilator has been running stably.

TABLE I. The main parameters set in ventilator. The unit of pressures (IPAP, PAP, CPAP, ∆PI, ∆PE, ∆P) is cm
H2O. NA denotes not applied.

IPAP EPAP ∆PI ∆PE

Mode CPAP ∆P BPM IE f (Hz)

S-HFOV 18 6 12 8 12 1 5
T-HFOV 18 6 12 8 12 1 5
18 12 NA NA 5
CPAP-HFOV 18 10 NA NA 2–10
18 5 NA NA 2–10
025113-4 Yuan et al. Rev. Sci. Instrum. 87, 025113 (2016)

FIG. 5. Data obtained under operating nHFOV ventilator with different ventilation mode: “In” and “Ex” express inspiration and expiration. (a) S-HFOV mode
(5 Hz of HFV), (b) T-HFOV mode (5 Hz of HFV), and (c) CPAP-HFOV mode (5 Hz of HFV).

III. RESULTS AND DISCUSSIONS In CPAP-HFOV mode, without spontaneous breathing,

a continual oscillatory pressure is superposed on a constant
The experimental data randomly obtained during the
mean (or bias) positive pressure. And they are delivered
each entire experiment were summarized and expressed as
to the mask. The maximum tidal volume VT(max) is just
mean ± stand deviation (operated in Excel 2007). One-sample
up to 40 ml. But the maximum minute ventilation volume
student t-test (operated in IBM SPSS for windows version
MV(max) is up to 9600 ml as described in Fig. 5(c). Similar
19.0) was utilized for comparing the setting pressures (IPAP,
to the oscillatory flow in T-HFOV mode, the gas is frequently
EPAP, CPAP, ∆PI, ∆PE, ∆P) and experimental pressures
exchanged between lungs and ambient air. The peak-to-peak
which obtained in mask, respectively. P-values <0.05 were
oscillatory flow rate is consistently kept around 50 l/min in
considered to be statistically significant.
both inspiration and expiration.
Fig. 5 gives the experimental results under various modes,
In the experiments mentioned above, the critical pres-
including pressures in the mask PA, flow rates in the airway
sures, ∆PI and ∆PE, IPAP, and EPAP, measured in the mask
Q A, the tidal volume VT(t), and the minute ventilation volume
are also plotted in Fig. 5. The measured ∆PI (or ∆P) is 10.2
MV. All the data were collected after the ventilator has worked
± 1.7 cm H2O (n = 36, p < 0.01) when it is preset to 12 cm
stably for 5 min.
As shown in Fig. 5(a), in S-HFOV mode, the maximum
tidal volume VT(max) is close to 600 ml. The maximum
minute ventilation volume MV(max) is more than 18 000 ml.
The maximum flow rates in inspiration and in expiration are
both ≈50 l/min. However, the peak-to-peak oscillatory flow
rate is much lower compared to that of the maximum flow rate.
At the time marked by the dashed vertical lines, converting the
ventilating phase of inspiration or expiration is duly triggered
by the spontaneous breathing, refer to Fig. 5(a). In this
experiment, the MV primarily is determined by the breathing
rate, not oscillation frequency because the oscillatory flow is
only changed in amplitude but not in direction.
In T-HFOV mode, as described in Fig. 5(b), VT(max)
is approximately 100 ml and MV(max) is up to 14 400 ml.
In inspiration or expiration, the flow rate fluctuates around
the zero. Meanwhile, the gas is frequently exchanged between
lungs and ambient air (or breathing in to and breathing out from
the lungs frequently). The peak-to-peak oscillatory flow rate
in inspiration and expiration is up to 100 l/min and 60 l/min,
respectively. The ventilation phase is determined by a time
FIG. 6. Pressures’ spectrums and flows’ spectrums: ∆P, MP, and ∆F: peak-
counter (5 s = 60/BPM) and not the spontaneous breathing peak oscillatory pressure in mask, mean pressure in mask, and peak-peak
(the spontaneous breathing was disabled in this experiment). oscillatory flow in airway.
025113-5 Yuan et al.
TABLE II. Measured peak-peak pressures and flows at each frequency as preset oscillation pressure 10 cm H2O and 5 cm H2O. Data were expressed as
mean ± SD of 10 measurements randomly obtained during the entire experiment.
Preset Measured 2 Hz 3 Hz 4 Hz 5 Hz
∆P 9.76 ± 0.22 9.16 ± 0.04 8.43 ± 0.08 7.68 ± 0.25
∆P = 10
∆F 72.64 ± 2.39 71.26 ± 2.72 64.84 ± 2.78 63.52 ± 3.61
∆P 5.08 ± 0.03 4.82 ± 0.04 4.13 ± 0.03 3.61 ± 0.06
∆P = 5
∆F 44.10 ± 4.36 42.97 ± 2.42 41.28 ± 2.05 38.18 ± 2.47
H2O. And measured ∆PE is 5.08 ± 0.06 cm H2O (n = 24,
p < 0.01) when it is preset to 8 cm H2O. The measured
IPAP (or CPAP) is 17.77 ± 0.78 cm H2O (n = 36, p = 0.086)
when it is preset to 18 cm H2O. And the measured EPAP is
6.88 ± 0.36 cm H2O (n = 24, p < 0.01) when it is preset to
6 cm H2O.
Fig. 6 describes the peak-to-peak oscillatory pressures in
the mask and the oscillatory flow rates in the main airway when
the frequency is changed from 2 Hz to 10 Hz. The spectrums
(pressure spectrums and flow rate spectrums) disclose that
the oscillatory pressures and the oscillatory flow rates are
inversely proportional to the oscillatory frequency of HFV
under given experimental conditions. The measured peak-to-
peak oscillatory pressures (∆P) and oscillatory flow rates (∆F)
are read from Figs. 6(a) and 6(b) and listed in Table II. At preset
∆P = 10 cm H2O, the ∆P and ∆F are ranged from 9.76 cm H2O
(n = 10, p < 0.009) at 2 Hz down to 5.85 cm H2O (n = 10,
p < 0.001) at 10 Hz and from 72.64 l/min at 2 Hz down to
56.28 l/min at 10 Hz, respectively. Similarly, at preset ∆P = 5
cm H2O, ∆P and ∆F are ranged from 5.08 cm H2O (n = 10,
p < 0.006) at 2 Hz down to 2.47 cm H2O (n = 10, p < 0.001)
at 10 Hz and from 44.1 l/min at 2 Hz down to 30.67 l/min at
10 Hz. The measured mean pressure of 18.7 cm H2O ± 0.33
(n = 18, p < 0.001) read from Fig. 6(a) is close to the preset
of 18 cm H2O.
According to the experimental results shown in Figs. 5(b)
and 5(c), the minute ventilation volume (MV) is large but the
tidal volume is small (in general, MV ≈ 6000 − 9000 ml and
VT ≈ 500 ml for a normal healthy adult17). The maximum
tidal volume VT(max) is even smaller than the anatomical
dead space (≈150 ml for normal healthy adult17). In classic
concept of pulmonary ventilation, the amount of gas reaching
the alveoli shall be equal to the required life-sustaining
volume (≈500 ml for normal healthy adult). Though, in our
experiments, the low tidal volume (or oscillation volume) is
still not a satisfactory explanation for whether the HFOV
is effective in life support.18,19 The large minute ventilation
volume provided in our experiments with the test nHFOV
ventilator is believed to be one of the key factors for sustaining
life.
So far, we have not found other technical papers that
discuss about how to realize the nHFOV with the mask in a
non-invasive ventilator. The only available literature offered
by De Luca et al.20,21 describes how the nHFOV works through
nasal prongs with an invasive ventilator Sensormedics 3100A
and a neonatal lung model. According to their records, about
22%–38% of pressures were reduced by the controlling circuit
and prongs. Furthermore, in generally application, an ET-
Rev. Sci. Instrum. 87, 025113 (2016)
Measured data
6 Hz 7 Hz 8 Hz 9 Hz 10 Hz
7.21 ± 0.42 6.59 ± 0.27 6.32 ± 0.4 6.14 ± 0.29 5.85 ± 0.48
65.24 ± 4.82 59.35 ± 3.9 60.29 ± 3.28 53.61 ± 1.63 56.28 ± 4.21
3.23 ± 0.09 3.10 ± 0.06 2.87 ± 0.21 2.80 ± 0.08 2.47 ± 0.1
37.63 ± 1.7 35.69 ± 2.15 36.87 ± 2.17 33.15 ± 2.09 30.67 ± 1.79
tube is necessary for Sensormedics 3100A. The ET-tube also
losses22 about 34%–90% pressure when the diameter of the
ET tube ranges between 6.5 mm–2.5 mm. However, despite
leakages from the valve and from the mask in this designed
nHFOV ventilator, according to the existing results under our
experimental conditions, the therapy oscillatory pressures are
directly delivered to the airway though the nasal mask with
low-loss pressure. According to experimental results given in
Fig. 6 and Table II, the max oscillatory pressure amplitude
has a loss ≈50% at 10 Hz and nearly zero loss at 2 Hz. The
measured mean pressure at each frequency is a steady-going
value which is closes to the preset pressure of CPAP.
IV. CONCLUSIONS
In this work, an experimental nHFOV ventilator was
achieved by utilizing a high-speed blower and an invented
valve. With this ventilator, three ventilation modes S-HFOV,
T-HFOV, and CPAP-HFOV are operated to ventilate the flash
sheep lungs in a typical experimental setup.
The experimental results demonstrate that this nHFOV
ventilator can provide a low-loss oscillatory pressure into nasal
mask and large valuable minute ventilation volume as well as
small tidal volume into the patient’s lungs though the common
breath tube and nasal mask.
The nHFOV ventilator designed according to specifica-
tions of this analysis is able to provide proper ventilation
and avoid the practice of inserting the ET-tube into patient’s
airway.
ACKNOWLEDGMENTS
The technical assistance from Curative Medical is grate-
fully acknowledged. This project is supported by Chinese
Universities Scientific Fund (Item No.: 14D310302).
1S. Mehta, “High frequency ventilation,” Curr. Opin. Crit. Care 6(1), 38–45
(2000).
2C. C. Dos Santos and A. S. Alutsky, “Overview of high-frequency ventila-
tion modes, clinical rationale, and gas transport mechanisms,” Respir. Care
Clin. North Am. 7(4), 549–575 (2001).
3M. Amit, S. Balpreet, H. Salhab El et al., “Use of noninvasive high-
frequency ventilation in the neonatal intensive care unit: A retrospective
review,” Am. J. Perinatol. 30(2), 171–176 (2015).
4M. A. Borrello, “Closed loop control system for a high frequency oscillation
ventilator,” U.S. patent 8434482 (05 July 2013).
5B. A. Simon and W. Mitzner, “Design and calibration of a high-frequency
oscillatory ventilator,” IEEE Trans. Biomed. Eng. 38(2), 214–218 (1991).
6K. Suwa and M. Taqami, “High-frequency animal ventilator using a loud-
speaker and its gas exchange characteristics,” Jpn. J. Physiol. 31(6), 859–868
(1981).
025113-6 Yuan et al.
7B. Ibsen, “The anesthetist’s viewpoint on the treatment of respiratory
complications in poliomyelitis during the epidemic in Copenhagen, 1952,”
Proc. R. Soc. Med. 47(1), 72–74 (1954).
8J. Downar and S. Mehta, “Bench-to-bedside review: High-frequency oscil-
latory ventilation in adults with acute respiratory distress syndrome,” Crit.
Care 10(6), 1–8 (2006).
9R. Dembinski, M. Max, R. Bensberg et al., “High-frequency oscillatory
ventilation in experimental lung injury: Effects on gas exchange,” Intensive
Care Med. 28(6), 768–774 (2002).
10L. M. Hannan, G. S. Dominelli, Y. W. Chen et al., “Systematic review of
non-invasive positive pressure ventilation for chronic respiratory failure,”
Respir. Med. 108(2), 229–243 (2014).
11J. G. Sun and X. M. Zhang, “Three-way pressure balance valve and medical
respirator,” Chinese invention patent CN202010105871 (03 August 2013).
12B. Diong, H. Nazeran, P. Nava et al., “Modeling human respiratory imped-
ance,” IEEE Eng. Med. Biol. Mag. 26(1), 48–55 (2007).
13A. Antonescu-Turcu and S. Parthasarathy, “CPAP and bi-level PAP therapy:
New and established roles,” Respir. Care 55(9), 1216–1229 (2010).
14D. Son and H. Choi, “A proportional integral differential control of flow over
a circular cylinder,” Philos. Trans. R. Soc., A 369(1940), 1540–1555 (2011).
15N. Ken, T. Etsushi, N. Fumikatsu et al., “The temperature change in an endo-
tracheal tube during high frequency ventilation using an artificial neonatal
Rev. Sci. Instrum. 87, 025113 (2016)
lung model with Babylog®; 8000 plus,” Pediatr. Pulmonol. 50(2), 173–178
(2013).
16R. W. Robinson, D. P. White, and C. W. Zwillich, “Moderate alcohol
ingestion increases upper airway resistance in normal subjects,” Am. Rev.
Respir. Dis. 132(6), 1238–1241 (1985).
17B. Maury, “The respiratory system in equations,” in Ms&A—Modeling
Simulation & Applications (Springer, Verlag, Italia, 2013), p. 217.
18H. K. Chang, “Mechanisms of gas transport during ventilation by high-
frequency oscillation,” J. Appl. Physiol.: Respir., Environ. Exercise Physiol.
56(5), 553–563 (1984).
19J. Jane pillow, “High-frequency oscillatory ventilation: Mechanisms of gas
exchange and lung mechanics,” Crit. Care Med. 33(Suppl. 3), S135–S141
(2005).
20D. De Luca, M. Piastram, D. Pietrini, and G. Conti, “Effect of amplitude
and inspiratory time in a bench model of non-invasive HFOV through nasal
prongs,” Pediatr. Pulmonol. 47(10), 1012–1018 (2012).
21D. De Luca, V. P. Carnielli, G. Conti, and M. Piastra, “Noninvasive high
frequency oscillatory ventilation through nasal prongs: Bench evaluation
of efficacy and mechanics,” Eur. J. Intensive Care Med. 36(12), 2094–2100
(2010).
22Operator’s Manual 3100A High Frequency Oscillatory Ventilator, C 2005
VIASYS Respiratory Care, Inc.