Grissom CK, Radwin MI, Harmston CH, Hirshberg EL, Crowley TJ. Respiration During Snow Burial Using an Artificial Air Pocket. JAMA. 2000;283(17):2266-2271. doi:10.1001/jama.283.17.2266
Author Affiliations: Pulmonary Division, LDS Hospital (Dr Grissom), University of Utah (Drs Grissom, Radwin, and Hirshberg), and Black Diamond Equipment Ltd (Mr Harmston), Salt Lake City, Utah; and TJ Crowley Corp, Aurora, Colo (Dr Crowley).
Context Asphyxia is the most common cause of death after avalanche burial. A
device that allows a person to breathe air contained in snow by diverting
expired carbon dioxide (CO2) away from a 500-cm3 artificial
inspiratory air pocket may improve chances of survival in avalanche burial.
Objective To determine the duration of adequate oxygenation and ventilation during
burial in dense snow while breathing with vs without the artificial air pocket
Design Field study of physiologic respiratory measures during snow burial with
and without the device from December 1998 to March 1999. Study burials were
terminated at the subject's request, when oxygen saturation as measured by
pulse oximetry (SpO2) dropped to less than 84%, or after 60 minutes
Setting Mountainous outdoor site at 2385 m elevation, with an average barometric
pressure of 573 mm Hg.
Participants Six male and 2 female volunteers (mean age, 34.6 years; range, 28-39
Main Outcome Measures Burial time, SpO2, partial pressure of end-tidal CO2 (ETCO2), partial pressure of inspiratory CO2(PICO2), respiratory rate, and heart rate at baseline (in open atmosphere)
and during snow burial while breathing with the device and without the device
but with a 500-cm3 air pocket in the snow.
Results Mean burial time was 58 minutes (range, 45-60 minutes) with the device
and 10 minutes (range, 5-14 minutes) without it (P=.001).
A mean baseline SpO2 of 96% (range, 90%-99%) decreased to 90% (range,
77%-96%) in those buried with the device (P=.01)
and to 84% (range, 79%-92%) in the control burials (P=.02).
Only 1 subject buried with the device, but 6 control subjects buried without
the device, decreased SpO2 to less than 88% (P=.005). A mean baseline ETCO2 of 32 mm Hg (range, 27-38
mm Hg) increased to 45 mm Hg (range, 32-53 mm Hg) in the burials with the
device (P=.02) and to 54 mm Hg (range, 44-63 mm Hg)
in the control burials (P=.02). A mean baseline PICO2 of 2 mm Hg (range, 0-3 mm Hg) increased to 32 mm Hg (range, 20-44
mm Hg) in the burials with the device (P=.01) and
to 44 mm Hg (range, 37-50 mm Hg) in the control burials (P=.02). Respiratory and heart rates did not change in burials with
the device but significantly increased in control burials.
Conclusions In our study, although hypercapnia developed, breathing with the device
during snow burial considerably extended duration of adequate oxygenation
compared with breathing with an air pocket in the snow. Further study will
be needed to determine whether the device improves survival during avalanche
Avalanches annually claim about 100 lives in Europe and 40 in North
America.1,2 Approximately 75%
of avalanche deaths are due to asphyxiation; up to 25%, to trauma; and very
few, to hypothermia.3,4 Several
factors may contribute to asphyxia during avalanche burial. First, the airway
may become obstructed with snow. Second, less air is available in snow for
breathing after an avalanche. Snow may be 80% to 90% air in the undisturbed
snowpack, but less than 60% air in avalanche debris.4,5
Third, water in humidified exhaled air condenses and freezes on the snow around
the mouth and face forming a thin "mask" of ice that impairs diffusion of
air.4 Fourth, the weight of snow may cause
chest restriction and impair breathing.6
Time to extrication is the major determinant of survival during avalanche
burial. Falk and colleagues7 report probability
of survival as 92% for persons extricated within 15 minutes, but only 30%
at 35 minutes. They suggest that survival after 35 minutes is dependent on
an air pocket in the snow for breathing. Survival at 90 minutes is 27%, but
decreases to only 3% at 130 minutes, representing avalanche burial victims
with an air pocket succumbing to late asphyxiation.
Even though an air pocket appears to be critical for prolonging survival,
we are aware of no prior studies characterizing the physiology of breathing
with an air pocket after avalanche burial. Preliminary studies from our group8 show that a larger air pocket or an artificial air
pocket device that separates inspired from expired air (AvaLung; Black Diamond
Equipment Ltd, Salt Lake City, Utah)9 may provide
adequate oxygenation for a longer time after burial.
The artificial air-pocket device allows a person to breathe air contained
in snow (Figure 1). A mouthpiece
is connected to a single circuit of respiratory tubing used for inspiration
and expiration. A plastic mesh 500-cm3 air pocket is connected
to the side of the respiratory tubing by two 1-way inspiratory valves. Expired
air passes from the mouthpiece through the respiratory tubing circuit to an
expiratory 1-way valve and then exits around the back away from the air pocket.
The device is built into a vest that is worn over all other clothing.
In this study, we characterized the physiology of breathing after snow
burial with and without an artificial air-pocket device. We hypothesized that
the device would maintain adequate oxygenation and ventilation longer than
an air pocket alone during burial in snow of similar density to avalanche
debris. To test this hypothesis, we studied 8 subjects breathing with the
device during snow burial for up to 60 minutes. As controls, we studied 7
of the same subjects breathing with only an air pocket in the snow during
This study took place at 2385 m elevation (average barometric pressure,
573 mm Hg) in the Wasatch Mountains, Utah, from December 1998 to March 1999.
The experimental set-up simulated avalanche debris and consisted of a large
mound of snow compacted with body weight. Snow density was determined in multiple
sites using a 250-cm3 wedge density cutter (Snowmetrics, Ft Collins,
Colo) that measured the weight of water per cubic meter. Snow density is reported
as a percentage (ie, 300 kg/m3 is 30% density snow or 70% air).
Snow densities for the study burials were 30% to 40% for 4 subjects (moderate-density
group) and 55% to 60% for 4 subjects (high-density group). Reported density
of avalanche debris ranges from 30% for a midwinter dry snow avalanche to
60% or higher for a springtime wet snow avalanche.5
A shoulder-width trench was dug into 1 end of the snow mound, and a sitting
platform was created for the subject so that the subject's head would be 30
cm and the device, 100 cm, under the surface (Figure 2).
Subjects were paid volunteers, 6 men and 2 women, mean age 34.6 years
(range, 28-39 years). All subjects except 1 lived at 1500- to 2500-m elevation.
Subject 7, who declined the control burial, was a healthy, active sea-level
resident who had spent 3 nights at 2500 m prior to burial with the device.
All other subjects were studied first during a burial with the device and
then during a control burial. No subject smoked cigarettes. Two subjects had
asthma: subject 3, atopic asthma and subject 5, exercise-induced asthma. Both
were being treated with β-agonist inhalers, used their inhalers prior
to the study, and experienced no symptoms of asthma. Subject 6 had a history
of hypothyroidism but was not being treated. Subject characteristics are listed
in Table 1. The LDS Hospital Research
and Human Rights Committee approved this study, and written informed consent
was obtained from the volunteers.
Physiologic parameters that were measured at baseline with subjects
breathing ambient air and continuously monitored during the burial studies
included partial pressure of end-tidal carbon dioxide (ETCO2) and
partial pressure of inspiratory CO2 (PICO2) in millimeters
of mercury, respiratory rate (RR), oxygen saturation as measured by pulse
oximetry (SpO2), surface 3-lead electrocardiogram, heart rate,
and skin surface temperature measured by a probe in the axilla. Monitoring
of ETCO2, PICO2, SpO2, and RR was done with
a capnometer (NPB-75; Mallinckrodt, St Louis, Mo) attached in-line with an
airway adapter (Microsteam CO2 Accessory Filterline ICU; Spegas
Industries, Jerusalem, Israel) to the device mouthpiece. A portable monitor
(NPB-4000; Mallinckrodt) was used to track electrocardiogram, heart rate,
SpO2, and body surface temperature. Respiratory rate was also measured
with the portable monitor using transthoracic impedance. A digital pulse oximeter
probe was attached to 1 finger on each hand. Physiologic parameters were observed
continuously and recorded every minute.
During burial with the device, subjects wore the artificial air-pocket
vest over multiple layers of clothing. A warm hood with a face mask and goggles
covered the head. Snow was compacted by hand as subjects sat in the snow mound
trench and were buried completely. Subjects were in communication with the
surface team via intercom. Time 0 of burial was noted when the subject's head
was completely buried. Study burial was terminated after 60 minutes, at the
subject's request, or when SpO2 fell below 84%. An emergency oxygen
backup line was attached to the device mouthpiece and could deliver 15 L/min
of 100% oxygen to increase inspired partial pressure of oxygen and flush CO2.
The study setup for control burials was identical to burials with the
device except that subjects breathed through the device mouthpiece that opened
into a 500-cm3 volume air pocket created in the snow. The capnometer
and emergency oxygen backup line were attached to the device mouthpiece. Control
burial was terminated at the subject's request or when SpO2 fell
Baseline measurements were compared with burial study data using a Friedman
analysis of variance and a Wilcoxon matched pairs test. Measurements at the
end of the device and control burials were compared by Mann-Whitney U test. The number of subjects with adequate oxygenation
(SpO2 ≥88%) at the end of device and control burials were compared
using a χ2 test. Data from the moderate- and high-density snow
groups were compared using a Mann-Whitney U test.
Statistica (StatSoft, 1999 edition, Tulsa, Okla) was used for all statistical
analysis. P<.05 was considered statistically significant.
Data are reported as mean and range.
Subjects remained buried longer with the device vs an air pocket in
the snow without the device (P=.001) (Table 2). Six subjects completed the protocol and remained buried
the full 60 minutes. Subject 1 requested removal at 45 minutes because he
was cold and had started to shiver. This subject was the first studied and
wore less insulating clothing. The investigators terminated the burial with
the device at 56 minutes for subject 6 because of hypoxia. Subject 6 also
had occasional premature ventricular beats during the last minute of burial.
All other subjects had normal sinus rhythm throughout the study. Burial times
for the control study ranged from 5 to 14 minutes. All subjects terminated
the control burial because of dyspnea; 4 subjects reported a sensation of
vertigo, and 3 subjects reported headache.
Data for each subject at baseline and at the end of device and control
burials are shown in Table 2.
Data for ETCO2, PICO2, and SpO2 at baseline
and throughout the device and control burials are shown in Figure 3. Mean baseline SpO2 of 96% (90%-99%) significantly
decreased to 90% (77%-96%) in burials with the device (P=.01), and to 84% (79%-92%) in control burials (P=.02). In device burials, SpO2decreased below 88% in only
1 subject, while in the shorter control burials, SpO2 in 6 subjects
dropped below 88% (P=.005).
Mean baseline ETCO2 of 32 mm Hg (27-38 mm Hg) significantly
increased to 45 mm Hg (32-53 mm Hg) in device burials (P=.02), and to 54 mm Hg (44-63 mm Hg) in control burials (P=.02). Mean baseline PICO2 of 2 mm Hg (0-3 mm Hg) significantly
increased to 32 mm Hg (20-44 mm Hg) in the device burials (P=.01), and to 44 mm Hg (37-50 mm Hg) in the control burials (P=.02). Mean baseline RR of 16/min (7-22/min) did not change
during device burials but significantly increased to 28/min (19-41/min) in
the control burials (P=.03). Mean baseline heart
rate of 71/min (63-90/min) did not change in the device burials but significantly
increased to 97/min (74-126/min) in the control burials (P=.02). P values indicate statistical comparison
of baseline and device or control burial end point data using a Wilcoxon matched-pairs
Body surface temperature significantly decreased only during device
burials (Table 2). Most subjects
felt cold and were starting to shiver at the end of the burial, with subjective
reports ranging from "not cold" in subject 5 to shivering and "very cold"
in subjects 7 and 8.
During burials with the device, PICO2 was significantly higher
in the high-density snow group within 5 minutes (P=.006),
and both PICO2 and ETCO2 were higher in the high-density
snow group throughout the burials (P<.001) (Figure 4). The SpO2 was significantly
lower in the high-density snow group (P<.001),
but the difference was not clinically important because all subjects except
subject 6 maintained adequate oxygenation (SpO2>88%) throughout
burial with the device.
The artificial air-pocket device allowed subjects to maintain adequate
oxygenation for up to 60 minutes during snow burial while a similarly sized
air pocket in the snow resulted in hypoxemia within 5 to 14 minutes. The device
also increased the time required for subjects to reach a clinically significant
degree of hypercapnia. In burials with the device, PICO2 increased
up to 44 mm Hg during 60 minutes, equivalent to a fraction of inspired CO2 (FICO2) of 8%. In the control burials, the same increase
in FICO2 and hypercapnia occurred within 5 minutes.
Worsening hypoxemia occurs with progressive hypercapnia because CO2 displaces oxygen in the alveoli according to the alveolar gas equation.
The artificial air-pocket device maintains adequate oxygenation because it
diverts expired CO2 away from the air pocket and delays the increase
in FICO2 that results in hypercapnia. The limit to breathing with
the device during snow burial, therefore, is related to the rise in FICO2. At the barometric pressure of our study site, when the increase in
FICO2 is sufficient to cause an ETCO2 of greater than
65 mm Hg, cerebral oxygenation will be severely compromised by a partial pressure
of oxygen in the alveoli of 25 to 30 mm Hg.10
Unconsciousness followed by death from asphyxiation will ensue. The device
considerably extends the time required to reach that critical degree of hypoxemia.
Extrapolation of the best-fit linear equation for the mean ETCO2
data suggest that an ETCO2 of 65 mm Hg may occur after about 145
minutes (y=36 + 0.2x) when
breathing with the device during snow burial vs only 15 minutes (y=36 + 2.0x) without the device. Hypercapnia
causes CO2 narcosis when arterial CO2 levels rise to
greater than 75 mm Hg.10 During avalanche burial,
a critical degree of arterial hypoxemia will occur before hypercapnia is severe
enough to cause CO2 narcosis.
Higher-density snow may result in more rapid asphyxia during avalanche
burial because our results suggest that diffusion of CO2 in snow
is inversely related to snow density. Subjects buried in higher-density snow
had higher PICO2 and ETCO2 measurements because CO2 was diffusing more slowly away from the subject vs subjects buried
in moderate-density snow.
Respiratory rates did not significantly change during burial with the
device, but all subjects reported a subjective sensation of breathing with
an increased tidal volume. It is known that in response to hypercapnia, initial
increases in minute ventilation occur with an increase in tidal volume rather
than RR.11 Subjects also reported increased
resistance to expiration while breathing with the device as a result of positive
end-expiratory pressure (PEEP). When a subject is breathing at rest with the
device, PEEP is 2.5 cm H2O; during hyperventilation, PEEP increases
to approximately 5 cm H2O. The increasing PEEP with increasing
flow may have caused subjects to adopt a more efficient pattern of breathing
with a higher tidal volume and lower RR. PEEP also may have increased functional
residual capacity and improved oxygenation.
In both burials, the degree of dyspnea reported by subjects was out
of proportion to the level of hypoxemia, suggesting that hypercapnia contributed
to subjective dyspnea. Hypercapnia is known to contribute to the sensation
of dyspnea.12 All subjects had mountaineering
experience at or above 4000 m and reported that dyspnea during this study
was worse than that experienced acutely at an altitude where SpO2
would be comparable to our end point values.
This study was limited to testing the device with subjects buried in
the sitting position. We have performed 2 additional studies with subjects
using the device while buried in a left-lateral decubitus position and in
an elbow-knee prone position (data not reported). Subjects maintained an SpO2 greater than 90% during a 60-minute burial. Burial in the supine position
may not be as well tolerated. In the preliminary study conducted by our group,9 1 subject in the supine position remained buried for
only 12 minutes because condensation dripping over the nose and mouth made
it difficult to breathe even though SpO2 remained normal.
An artificial air-pocket device provides a significant advantage for
breathing during snow burial because expired CO2 is diverted away
from the air pocket, aspiration of snow is prevented, and an ice mask impermeable
to air does not form. However, during an actual avalanche burial a major issue
will be the ability to place the mouthpiece in the mouth. Stiff respiratory
tubing that allows the mouthpiece to remain in a ready position facilitates
this. Once buried, snow density, clothing, and individual physiologic differences
will influence survival. If the device prolongs survival with longer extrication
times, then hypothermia may become a more important factor because core body
temperature will drop at a rate of about 3°C/h.13
Wearing an artificial air-pocket device does not replace conservative
judgment and appropriate safety precautions when traveling in avalanche terrain.
For persons who are buried in an avalanche, however, our preliminary tests
suggest that such a device may reduce probability of asphyxiation before a