Position of exhalation port and mask design affect CO2 rebreathing

during noninvasive positive pressure ventilation*

Guilherme P. P. Schettino, MD, PhD; Sunisa Chatmongkolchart, MD; Dean R. Hess, PhD, RRT;
Robert M. Kacmarek, PhD, RRT

Objective: Noninvasive positive pressure ventilation may be
considered a first line intervention to treat patients with hyper-
capnic respiratory failure. However, CO2 rebreathing from the
ventilator circuit or mask may impair CO2 elimination and load
the ventilatory muscles. This study was conducted to evaluate the
effect of exhalation port location and mask design on CO2 re-
breathing during noninvasive positive pressure ventilation.
Design: Lung model evaluation.
Setting: Experimental laboratory of a large university-affiliated
hospital.
Subjects: A dual-chamber test lung was used to simulate the
ventilatory mechanics of a patient with obstructive lung disease.
Intervention: Hypercapnic respiratory failure (end-tidal CO2 of 75
mm Hg) and obstructive lung disease were simulated in a double-
chamber lung model. A facial mask (inner volume of 165 mL) with
exhalation port within the mask (Facial-MEP) or the same mask with
exhalation port in the ventilator circuit (Facial-WS) and a total face
mask with exhalation port within the mask (inner volume 875 mL,
Total Face) were tested during continuous positive airway pressure
and pressure support ventilation provided by a single-limb circuit
ventilator at the same frequency and tidal volume.
Measurements and Main Results: A capnometer and a flow
transducer were placed in the lung model upper airway to mea-
sure the volume of CO2 rebreathed and tidal volume (VT). The
inspiratory load was estimated from the pressure variation in the
lung model driving chamber (PDR). Volume of CO2 rebreathed was
smaller during Facial-MEP compared with the other masks in all
tested conditions (p < .001). The VT and PDR necessary to de-
crease end-tidal CO2 20% (from 75 to 60 mm Hg) was different
among the tested masks (Facial-MEP, VT 701 ⴞ 9 mL, PDR 8.1 ⴞ
0.1 cm H2O/sec; Facial-WS, VT 745 ⴞ 9 mL, PDR 10.2 ⴞ 0.1 cm
H2O/sec; Total Face, VT 790 ⴞ 12 mL, PDR 11.4 ⴞ 0.2 cm H2O/sec,
p < .001).
Conclusion: Facial-MEP with its exhalation port within the
mask and the smallest mask volume demonstrated less re-
breathed CO2 and a lower PDR than either the Facial-WS or
Total Face masks. Additional studies are necessary to confirm
if mask design can clinically affect patient’s inspiratory effort
during noninvasive positive pressure ventilation. (Crit Care
Med 2003; 31:2178 –2182)
KEY WORDS: noninvasive mechanical ventilation; mask; exhala-
tion port; CO2 rebreathing; work of breathing

N oninvasive positive pressure
ventilation (NPPV) has been
successfully applied to pa-
tients with acute or acute on
chronic respiratory failure of various eti-
ologies (1). However, the best outcome
associated with this technique has been
reported in patients with hypercapnic re-
spiratory failure (2– 4). When applied to
these patients, NPPV increases alveolar
ventilation, decreases PaCO2, and unloads
*See also p. 2247.
From the Department of Anesthesia and Critical
Care and Respiratory Care (GPPS, SC), Massachusetts
General Hospital, Harvard Medical School; and Harvard
Medical School and Respiratory Care Services (DRH,
RMK), Massachusetts General Hospital, Boston, MA.
Supported, in part, by FAPESP, Brazil (GPPS).
Address requests for reprints to: Robert M. Kac-
marek, PhD, RRT, Respiratory Care, Ellison 401, Mas-
sachusetts General Hospital, 55 Fruit Street, Boston,
MA 02114. E-mail: rkacmarek@partners.org
Copyright © 2003 by Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000081309.71887.E9
the ventilatory muscles. However, CO2
rebreathing from the mask or the single-
limb ventilator circuit used in most of the
ventilators designed for NPPV can impair
CO2 elimination and ventilatory muscle
unloading (5).
The use of positive end-expiratory
pressure (PEEP) or continuous positive
airway pressure (CPAP) or the placement
of a nonrebreathing valve in the circuit
can reduce rebreathing from the ventila-
tor circuit (5–7). However, PEEP or CPAP
should not affect mask-related rebreath-
ing if the mask is placed distal to the
circuit valve. Nevertheless, no data are
available regarding the influence of mask
design, particularly for masks with large
inner volume and alternative exhalation
port (EP) positions (within the mask vs.
in the circuit), on the dynamics of CO2
rebreathing.
This bench study was conducted to
evaluate the dynamics of CO2 rebreathing
during NPPV. A lung model with me-
chanics similar to patients with obstruc-
tive lung disease was used to simulate
hypercapnic respiratory failure. Masks
with two different inner volumes along
with two EP positions were evaluated. We
hypothesized that placement of the EP
within the mask reduces the amount of
CO2 rebreathing during NPPV.
METHODS
Respiratory System Model. A dual-cham-
ber test lung (Adult TTL; MI Instruments,
Grand Rapids, MI) was used to simulate the
ventilatory mechanics of a patient with ob-
structive lung disease (Fig. 1). One chamber
simulated the patient’s lung (lung chamber,
LC) and the other chamber the chest wall and
diaphragm (driving chamber, DC). The com-
pliance of LC and DC were set at 150 mL/cm
H2O. An adult-sized fiberglass mannequin
head was connected to the LC by using a
low-compliance circuit with deadspace vol-
ume of 120 mL (8). A linear resistor (20 cm
H2O·L⫺1·sec⫺1; Hans-Rudolph, Kansas City,
MO) was placed at the entrance of the LC.

2178 Crit Care Med 2003 Vol. 31, No. 8

 The driving and lung chambers were con-
nected to each other by a metal connector as
detailed by Yamada and Du (9). A ventilator
(7200; Puritan Bennett, Carlsbad, CA) was
used to power the DC creating spontaneous
inspiratory efforts in the LC. One hundred
percent CO2 entered the LC via a flowmeter to
simulate CO2 production as described by Lam-
potang et al. (10).
A single-limb circuit ventilator designed
for NPPV and commercially available (BiPAP
Synchrony; Respironics, Murrysville, PA) was
used throughout the experiment to ventilate
the LC.
Masks and Exhalation Ports. The following
combinations of masks and exhalation ports
were tested (Fig. 2): a) facial mask (Adult Me-
dium; ResMed, San Diego, CA) with a mask
inner volume of 165 mL when attached to the
mannequin’s face and the exhalation port
within the mask (Facial-MEP); b) the same
facial mask with the mask exhalation port oc-
cluded and a Whisper Swivel II exhalation port
(Respironics) placed between the ventilator
circuit and the mask (Facial-SW); and c) total-
face mask (Respironics) with a mask inner
volume of 875 mL when attached to the man- Figure 1. Illustration of the experimental setup. A two-chamber lung model was used to simulate
nequin’s face and the exhalation port located spontaneous breathing. CO2 was delivered to the lung chamber. A capnometer and pressure and flow
in the front of the mask (Total Face). The transducers were placed at the output of the mannequin upper airway, allowing measurement of the
masks were tightly secured to the manne- volume of CO2 rebreathed each breath and tidal volume.
quin’s face to guarantee the minimum possi-
ble airleak through the mask-face interface.
Data Acquisition and Measurements. A
nonaspirating-type capnometer with flow and
pressure transducers (CO2SMO; Novametrix
Medical Systems, Wallingford, CT) was placed
in the circuit between the mannequin and LC,
allowing continuous recording of end-tidal
CO2 (PetCO2) and LC tidal volume (VT; Fig. 1).
The circuit deadspace volume proximal (upper
airway) and distal to the capnometer was 50
and 70 mL, respectively. CO2 rebreathed
(CO2REB, mL/breath) was electronically calcu-
lated based on the area defined by the product Figure 2. Illustration of the tested masks and exhalation ports (EP, arrow). A, facial mask (Adult
of the inspiratory VT and the CO2 concentra- Medium, ResMed) with the EP within the mask. B, facial mask (Adult Medium, ResMed) with the mask
tion (CO2%). EP occluded and a Whisper Swivel II (Respironics) placed in the ventilator circuit. C, total face mask
A pressure transducer (Ventrak, Novame- (Respironics) with the EP within the mask.
trix) recorded the pressure generated inside
the DC by the driving ventilator. The pressure-
time product of the DC (PDR) for each breath were applied using Facial-MEP, Facial-SW, because by protocol VT in the LC needed to
was calculated by integrating the area of the and Total Face masks. vary as we evaluated the PDR required to vary
pressure time curve above end-expiratory The influence of mask design on PDR dur- PetCO2 by 20% with each mask. This could not
pressure during the inspiratory phase. ing partial correction of hypercapnia was eval- be accomplished with volume control ventila-
The capnometer and the pressure and flow uated during pressure support ventilation tion since the DC and the LC were connected.
transducers were calibrated before the exper- (PSV). First, PDR was decreased to 30% of its Statistical Analysis. After a period of stabi-
iment as recommended by the manufacturer. baseline value, and then the tested ventilator lization, five consecutive breaths for each ex-
Experimental Protocol. The effects of EP (BiPAP Synchrony) was attached to each of the perimental setting were recorded on a per-
location and mask inner volume on CO2REB evaluated masks. This ventilator was set as sonal computer for off-line analysis. Two-way
were assessed during CPAP of 4 and 8 cm H2O. follows: PEEP of 4 cm H2O and the minimum analysis of variance for repeated measures
The driving ventilator was set in pressure con- PSV level necessary to ensure a VT of 600 mL with Tukey’s post hoc analysis (SPSS 10.0;
trol, PEEP 4 cm H2O, frequency 18/min, and in the LC for each of the tested masks. PDR was SPSS, Chicago, IL) was used to compare the
inspiratory/expiratory ratio 1:3, and the pres- then progressively increased, while PSV and f CO2REB when the studied masks (Facial-MEP,
sure level was adjusted to generate a VT of 600 were maintained constant, increasing the VT Facial-WS, and Total Face) were tested with
mL in the LC (baseline PDR). Next, the CO2 of the LC and, as a result, decreasing PetCO2 by the two different CPAP levels (4 and 8 cm
flow rate into the lung chamber was adjusted 20% (from 75 to 60 mm Hg). PetCO2, CO2REB, H2O). One-way analysis of variance for re-
to establish a PetCO2 of 75 mm Hg and kept PDR, and lung chamber VT were compared. peated measures was used to compare the
constant throughout all the experimental set- We used PC, as opposed to volume control CO2REB recorded when the model was tested
tings. CPAP levels of 4 and 8 cm H2O then with a sign wave flow pattern, to power the DC without mask (baseline) and during CPAP

Crit Care Med 2003 Vol. 31, No. 8 2179

measurements. One-way analysis of variance
for repeated measures also was used to com-
pare the CO2REB, VT, and PDR during PSV. The
results are presented as mean ⫾ SD, and p ⬍
.05 was considered significant.
RESULTS
All protocol steps were successfully
performed. No auto-triggering or missed
inspiratory efforts were observed during
any experimental setting.
There was no significant difference in
CO2REB between baseline without a mask
(8.0 ⫾ 0.4 mL) and Facial-MEP (8.1 ⫾ 0.3
mL and 7.7 ⫾ 0.4 mL, respectively, for
CPAP of 4 and 8 cm H2O), showing that
virtually no CO2 was rebreathed from the
circuit or mask when Facial-MEP was
used (Fig. 3). The use of Facial-MEP re-
sulted in less CO2REB compared with Fa-
cial-WS (11.3 ⫾ 0.3 mL and 10.1 ⫾ 0.5
mL, respectively, for CPAP of 4 and 8 cm
H2O) or Total Face (11.4 ⫾ 0.5 mL and
10.5 ⫾ 0.4 mL, respectively, for CPAP of
4 and 8 cm H2O) at both tested CPAP
levels (p ⬍ .001). For the same CPAP
level, CO2REB was similar with Facial-WS
and Total Face (Fig. 4), and increasing
CPAP level from 4 to 8 cm H2O decreased
CO2REB for Facial-WS (p ⫽ .001) and To-
tal Face (p ⫽ .04) but not for Facial-MEP.
Most of the CO2 rebreathed during CPAP
came from the mannequin upper airway
and not from the ventilator circuit or
mask as illustrated in Figure 3.
Figure 3. CO2 rebreathing (CO2REB) during spon-
taneous ventilation without mask (dashed line)
and with continuous positive airway pressure
(CPAP) of 4 and 8 cm H2O. End-tidal CO2 was set
a 75 mm Hg, and VT of 600 mL was generated in
the lung chamber in all tested conditions.
Squares, facial mask with exhalation port within
the mask (Facial-MEP); circles, facial mask with
Whisper Swivel II exhalation port in the circuit
(Facial-WS); triangles, total face mask (Total
Face). *p ⬍ .001 for difference among Facial-
MEP vs. other masks with the same CPAP level;
†p ⫽ .001 for Facial-WS between CPAP of 4 and
8 cm H2O; ‡p ⫽ .04 for Total Face between CPAP
of 4 and 8 cm H2O.
2180
The PSV level necessary to ensure a VT
of 600 mL when PDR was set at 30% of its
baseline value was 19 cm H2O above
PEEP (peak inspiratory pressure of 23 cm
H2O) for all tested masks. At this setting,
PetCO2 and CO2REB were lower with Fa-
cial-MEP compared with the other two
masks (p ⬍ .001, Table 1). Facial-WS and
Total Face had similar PetCO2 and CO2REB
during PSV ventilation.
The target PetCO2 (60 mm Hg) was
reached during PSV (PEEP ⫽ 4 cm H2O
and PSV ⫽ 19 cm H2O) when the VT was
increased by increasing PDR. However,
CO2REB (6.4 ⫾ 0.2, 9.0 ⫾ 0.2, and 10.7 ⫾
0.2 mL/breath, respectively, for Facial-
MEP, Facial-WS, and Total Face), PDR
(8.1 ⫾ 0.1, 10.2 ⫾ 0.1, and 11.4 ⫾ 0.2 cm
H2O/sec), and VT (701 ⫾ 9, 745 ⫾ 9, and
790 ⫾ 12 mL) were different for each
tested mask (p ⬍ .001; Fig. 5).
DISCUSSION
The most important findings of this
study can be summarized as follows: a)
Figure 4. Volume-based capnograms for each
tested mask during continuous positive airway
pressure (CPAP). Solid black line, facial mask
with exhalation port within the mask (Facial-
MEP); dashed line, facial mask with circuit Whis-
per Swivel (Facial-WS); gray line, total face mask
(Total Face); CO2%, CO2 concentration; VT, in-
spiratory tidal volume. The area under the curve
(i.e., the CO2%-VT product), illustrates the vol-
ume of CO2 rebreathed for each tested mask
during CPAP 4 cm H2O, end-tidal CO2 75 mm
Hg, and VT 600 mL. Note that the CO2% tracing
drops to zero during inspiration when Facial-
MEP was tested but not with Facial-WS and Total
Face.
Table 1. CO2 dynamics during pressure support ventilation
Facial-MEP
PETCO2 mm Hg 72.4 ⫾ 0.1a
CO2REB mL/breath 7.0 ⫾ 0.2a
VT mL 610 ⫾ 5
Facial-MEP, facial mask with exhalation port within the mask; Facial-WS, facial mask with Whisper
Swivel II exhalation port in the circuit; Total Face, total face mask; PETCO2, end-tidal CO2; CO2REB,
total volume of CO2 rebreathed per breath; VT, tidal volume.
a
p ⬍ .001 for differences between Facial-MEP vs. others.
The exhalation port located within the
mask was more effective in clearing CO2
from the mask and circuit than the Whis-
per Swivel II exhalation port located
within the circuit; and b) the presence of
the exhalation port within the mask and a
small mask inner volume reduced the
inspiratory pressure time product during
attempts to correct hypercapnia.
Mask Design and CO2 Rebreathing.
Although standard intensive care unit
ventilators can be used for noninvasive
ventilation, ventilators specifically de-
signed for mask ventilation have been
used more frequently because of their
ability to compensate for system leaks
and their low cost (7, 11). Most of these
ventilators are flow-generating units that
are capable of providing CPAP or PSV/
PEEP and operate with a single circuit as
both inspiratory and expiratory limb. To
avoid excessive CO2 exhalation back into
the circuit, an exhalation port or a non-
rebreathing valve must be placed between
the circuit and the mask or within the
mask. End-expiratory pressure level, ex-
piratory time, VT, and exhalation port or
valve types have been reported to affect
CO2 rebreathing during NPPV (5–7).
However, no data addressing the effects
of exhalation port position and mask in-
ner volume on CO2 rebreathing are avail-
able.
Ferguson and Gilmartin (6) studied
CO2 rebreathing during NPPV using a
single-limb circuit ventilator and a nasal
mask in volunteers and stable patients
with chronic ventilatory failure. The vol-
ume of CO2 exhaled into the circuit, and
consequently the volume of CO2 inhaled
back from the circuit, was greater when
using the Whisper-Swivel exhalation port
compared with two other nonrebreathing
valves. This difference was more evident
when PEEP was ⬍8 cm H2O. The design
of the current experiment prevented us
from differentiating the volume of CO2
rebreathed from the mask or the patient’s
upper airway from that rebreathed from
Facial-WS Total Face
74.6 ⫾ 0.9 75.7 ⫾ 1.4
9.9 ⫾ 0.5 10.4 ⫾ 0.3
611 ⫾ 4 606 ⫾ 9
Crit Care Med 2003 Vol. 31, No. 8

Figure 5. CO2 rebreathed during each breath
(CO2REB, mL/breath) and simulated inspiratory
effect (PDR, cm H2O per second) during pressure
support ventilation required to decrease end-tidal
CO2 by 20% (from 75 to 60 mm Hg). Facial-MEP,
facial mask with exhalation port within the mask;
Facial-WS, facial mask with Whisper Swivel II
exhalation port in the circuit; Total Face, total
face mask. *p ⬍ .001 for differences among
masks for CO2REB. †p ⬍ .001 for differences
among masks for PDR.
the ventilator circuit. Lofaso et al. (5)
demonstrated in a lung model that ⬎50%
of the expired VT can accumulate in the
single-limb circuit at end-expiration
when a Whisper Swivel exhalation port
was tested with PEEP ⬍2 cm H2O. Our
data showed that increasing CPAP level
from 4 to 8 cm H2O had only a small
effect on decreasing CO2REB when masks
with exhalation ports within the mask
(Facial-MEP and Total Face) were tested,
but during the use of a circuit exhalation
port (Whisper Swivel II) the CPAP level
positively affected CO2 clearing. It is im-
portant to note that all CO2 could not be
cleared from the circuit and mask even
when a high CPAP level (8 cm H2O) was
set if a circuit exhalation port or a mask
with a large inner volume was used. It is
important also to consider the flow
through the exhalation port at end-
exhalation. There are different flows from
the three exhalation ports used at the two
end-expiratory pressure levels evaluated
(12–14) (Table 2). As a result, flow
through the exhalation port as well as
position of the port must be considered.
In general, one would expect CO2 clear-
ance to be directly related to flow from
the exhalation ports on the mask. In ad-
dition, leak at the interface of the mask
with the patient’s face, provided it is not
excessive, should also facilitate CO2 clear-
ance regardless of exhalation port loca-
tion or mask volume. Nonrebreathing
and plateau exhalation valves have been
recommended to decrease CO2 rebreath-
ing during NPPV when a low CPAP or
PEEP level is applied. However, the use of
Crit Care Med 2003 Vol. 31, No. 8
Table 2. Flows through expiratory ports at end-exhalation (L/min)
Masks
ResMed Mirage (Facial-MEP)
Whisper Swivel II (Facial-WS)
Total Face
nonrebreathing valves may increase the
expiratory resistance imposed by the ven-
tilator (7).
In our study, the calculated CO2REB
represented the total amount of CO2 re-
breathed from the ventilator circuit,
mask, and mannequin’s upper airway and
not simply the CO2 rebreathed from the
circuit as described in previous experi-
ments. Our results demonstrate that the
majority of the CO2 rebreathed came
from the upper airway (8 mL/breath) and
not from the ventilator circuit or mask
(ⱕ3.4 mL/breath for all tested masks dur-
ing CPAP of 4 cm H2O). We measured
CO2 as close to the exit of the airway from
the mannequin head as possible. The vol-
ume of the airway from the lips to the
capnometer equaled 50 mL. Actually,
CO2REB during Facial-MEP was similar to
that recorded during spontaneous
breathing without a mask, showing that
the vast majority of CO2 exhaled back
into the mask and circuit was cleared
before the next breath. We believe that
the amount of CO2REB observed during
Facial-MEP and CPAP 8 cm H2O, which
was slightly lower than CO2REB without a
mask, was a result of CO2 cleared from
the upper airway. Jet mixing created by
the flow through mask exhalation port
may have cleared some CO2 from the
model upper airway during this setting.
The amount of CO2REB during Facial-
MEP and Facial-SW (both masks with the
same mask inner volume) clearly indi-
cates that the exhalation port within the
mask more efficiently avoids CO2REB than
using a circuit exhalation port.
Mask Design and Inspiratory Effort.
Facial masks (oro-nasal), as opposed to
nasal masks, have been recommended for
patients with acute or acute on chronic
respiratory failure (15). However, dis-
comfort related to wearing a facial mask
and airleak around the mask, particularly
leak flow directed toward the eyes, are
important limitations of applying NPPV.
In an effort to minimize these side ef-
fects, new devices (i.e., masks or helmets)
have been developed, but with the cost of
increasing mask inner volume (16, 17).
End-Expiratory Pressure
4 cm H2O 8 cm H2O
24 36
15 24
20 32
Using a mask with a large inner volume
may theoretically impair ventilatory mus-
cle unloading during NPPV by two differ-
ent mechanisms: by increasing the vol-
ume of compressed gas, which can
decrease the ventilator’s performance
and, more importantly, by increasing the
volume of CO2 rebreathed from the mask
each breath.
A similar VT was generated in the LC
with the three tested masks during pres-
sure support ventilation while the PSV
level and PDR were kept constant. This
data demonstrated that mask design (ex-
halation port position and mask inner
volume) has a minimal effect on ventila-
tor performance. However, the PDR and
VT necessary to decrease the PetCO2 from
75 to 60 mm Hg were highest during the
use of a mask with greater inner volume
and CO2REB (Fig. 5). When masks with
the same inner volume were compared
(Facial-MEP vs. Facial-WS), the presence
of the exhalation port within the mask
decreased the CO2REB and allowed partial
correction of hypercapnia with a smaller
increase in VT and PDR. These data dem-
onstrated that small differences in
CO2REB may result in increases in venti-
latory load. Our data suggest that the
inspiratory effort of patients ventilated
noninvasively with the same level of pres-
sure support would be smaller when a
Facial-MEP is applied compared with a
Facial-WS or a Total Face mask. This
observation is consistent with the results
presented by Lofaso et al. (5), who
showed that CO2 rebreathing during the
use of a single-limb circuit ventilator in-
creased VT, minute ventilation, and work
of breathing in invasively ventilated pa-
tients.
Limitations. The major limitation of
this study is that it was performed on a
lung model and not in patients. However,
it may be impossible to obtain similar
data from patients. The sizes of the exha-
lation ports of the Facial-MEP, Facial-
WS, and Total Face masks were different.
This makes it more difficult to defini-
tively draw specific conclusions from the
study. All currently available facial masks
2181

B
ased on the re-
sults of this study,
masks with exha-
lation ports located within
the mask and smaller inner
volume masks can be ex-
pected to minimize the re-
breathing of CO 2 during
noninvasive ventilation.
were not evaluated. Since we only tested
a select PetCO2 and VT, we cannot state
that all PCO2 and VT measures would re-
spond the same way. However, we would
expect that CO2 rebreathing would be di-
rectly related to PetCO2 and indirectly re-
lated to VT. Finally, we do not know if the
increase in CO2 rebreathed or inspiratory
pressure time product measured in this
experiment is clinically relevant.
CONCLUSION
Based on the results of this study,
masks with exhalation ports located
within the mask and smaller inner vol-
ume masks can be expected to minimize
the rebreathing of CO2 during noninva-
sive ventilation. However, this study only
evaluated three mask configurations,
2182
making extrapolation of our data to all
masks impossible. Additional laboratory
and clinical studies are necessary to con-
firm these observations.
REFERENCES
1. International consensus conference in inten-
sive care medicine: Noninvasive positive
pressure ventilation in acute respiratory fail-
ure. Am J Respir Crit Care Med 2001; 163:
283–291
2. Brochard L, Mancebo J, Wysocki M, et al:
Noninvasive ventilation for acute exacerba-
tion of chronic obstructive pulmonary dis-
ease. N Engl J Med 1995; 333:817– 822
3. Kramer N, Meyer TJ, Meharg J, et al: Ran-
domized, prospective trial of noninvasive
positive pressure ventilation in acute respi-
ratory failure. Am J Respir Crit Care Med
1995; 151:1799 –1806
4. Keenan SP, Kernerman PD, Cook DJ, et al:
Effect of noninvasive positive pressure venti-
lation on mortality in patients admitted with
acute respiratory failure: A meta-analysis.
Crit Care Med 1997; 25:1685–1692
5. Lofaso F, Brochard L, Touchard D, et al:
Evaluation of carbon dioxide rebreathing
during pressure support ventilation with air-
way management system (BiPAP) devices.
Chest 1995; 108:772–778
6. Ferguson GT, Gilmartin M: CO2 rebreathing
during BiPAP ventilatory assistance. Am J
Respir Crit Care Med 1995; 151:1126 –1135
7. Lofaso F, Brochard L, Hang T, et al: Home
versus intensive care pressure support de-
vices. Experimental and clinical comparison.
Am J Respir Crit Care Med 1996; 153:
1591–1599
8. Schettino GPP, Tucci MR, Sousa R, et al:
Mask mechanics and leak dynamics during
noninvasive pressure support ventilation: A
bench study. Intensive Care Med 2001; 27:
1887–1891
9. Yamada Y, Du H-L: Analysis of the mecha-
nisms of expiratory asynchrony in pressure
support ventilation: A mathematical ap-
proach. J Appl Physiol 2000; 88:2143–2150
10. Lampotang S, Gravenstein N, Banner MJ, et
al: A lung model of carbon dioxide concen-
trations with mechanical spontaneous venti-
lation. Crit Care Med 1986; 14:1055–1057
11. Bunburaphong T, Imanaka H, Nishimura M,
et al: Performance characteristic of bilevel
pressure ventilators. A lung model study.
Chest 1997; 111:1050 –1060
12. Product Literature ResMed Disposable Mi-
rage Full Face Mask. No. 2000 16849/20010.
San Diego, CA, ResMed
13. Product Literature Whisper Swivel II. No.
1001515 JBM 8/27/99. Murrysville, PA, Re-
spironics
14. Product Literature Respironics Total Face
Mask. No. 1002964 JBM 10/15/99. Murrys-
ville, PA, Respironics
15. Navalesi P, Fanfulla F, Frigerio P, et al: Phys-
iologic evaluation of noninvasive mechanical
ventilation delivered with three types of
mask in patients with chronic hypercapnic
respiratory failure. Crit Care Med 2000; 28:
1785–1790
16. Criner GJ, Travaline JM, Brennan KJ, et al:
Efficacy of a new full face mask for noninva-
sive positive pressure ventilation. Chest
1994; 106:1109 –1115
17. Antonelli M, Conti G, Pelosi P, et al: New
treatment of acute hypoxemic respiratory
failure: Noninvasive pressure support ven-
tilation delivered by helmet—A pilot con-
trolled trial. Crit Care Med 2002; 30:
602– 608
Crit Care Med 2003 Vol. 31, No. 8