GENERAL CLINICAL INVESTIGATION/ORIGINAL CONTRIBUTION

Restraint Position and Positional Asphyxia
From the Department of Emergency
Medicine' and the Division of
Pulmonary Medicine, Department of
Medicine, t University of California
San Diego Medical Center, San Diego,
California.
Received for publication
November 11, 1996. Revision received
March 24, 1997. Accepted for
publication May 8,1997.
An abstract of this study was presented at the Society for Academic
Emergency Medicine annual conference, Washington DC, May 1997.
Supported in part by grant No. 941974R from the County of San Diego
and by grant MOl RR0087 from the
General Clinical Research Center,
National Center for Research
Resources, National Institutes of
Health.
Copyright © by the American College
of Emergency Physicians.

Theodore C Chan. MO"

Gary M Vi Ike. MD'
Tom Neuman, MO"'~

Study objective: To determine whether the "hobble" or "hogtie" restraint position results in clinically relevant respiratory
dysfunction.

Jack L Clausen, MOrt

Methods: This was an experimental, crossover. controlled trial
at a university-based pulmonary function laboratory involving 15
healthy men ages 18 through 40 years. Subjects were excluded
for a positive urine toxicology screen, body mass index (8MI)
greater than 30 kg/m2. or abnormal screening pulmonary function
testing (PFT). Forced vital capacity (FVC). forced expiratory volume
in 1 second (FEV 1). and maximal voluntary ventilation (MW) were
obtained with subjects in the sitting, supine, prone. and restraint
positions. After a 4-minute exercise period, subjects rested in the
sitting position while pulse, oxygen saturation, and arterial blood
gases were monitored. The subjects repeated the exercise, then
were placed in the restraint position with similar monitoring.
Results: There was a small. statistically significant decline in
the mean FVC (from 5.31±1.01 L [101 %±1 0.5% of predictedJ to
4.60±.84 L [88%±8.8% of predicted]), mean FEV1 (from 4.31±.53 L
[1 03%±8.4%J to 3.70±.45 L [89%±7.7%]), and mean MW (from
165.5±24.5 L/minute [111 %±17.3%J to 131.1±20.7 L/minute [88%
±16.6%]), comparing sitting with restraint position (all. P<.001).
There was no evidence of hypoxia (mean oxygen tension [Po 2Jless
than 95 mm Hg or co-oximetry less than 96%) in either position.
The mean carbon dioxide tension (Pco2) for both groups was not
different after 15 minutes of rest in the sitting versus the restraint
position. There was no significant difference in heart rate recovery or oxygen saturation as measured by co-oximetry and pulse
oximetry.
Conclusion: In our study population of healthy subjects, the
restraint position resulted in a restrictive pulmonary function pattern but did not result in clinically relevant changes in oxygenation
or ventilation.

[Chan TC, Vilke GM. Neuman T, Clausen JL: Restraint position and
positional asphyxia. Ann Emerg Med November 1997;30:578-586.J

57 8

ANNALS OF EMERGENCY MEDICINE

30:5 NOVEMBER 1997

RESTRAINT AND ASPHYXIA
Chan et al

INTRODUCTION

Prehospital medical and law enforcement personnel often
confront and care for violent, agitated individuals who must
be subdued in order to prevent injury to themselves or
others. A variety of physical restraints are used in this setting to accomplish these goals. Certain physical restraints
used by field personnel have been reported to be detrimental
to the restrained individual, possibly resulting in significant
morbidity and even mortality for the apprehended person.
In the early 1980s, widespread reports of deaths from police
neck "choke holds" resulted in changes in physical restraint
policies nationwide. 1,2
Recent attention has focused on the use of the "hogtie" or
"hobble" restraint position and its possible role in the sudden
deaths of individuals placed in this custody restraint position. 3-6 In this restraint, individuals are placed in the prone
position with their wrists handcuffed or tied together behind
their backs and their ankles bound together and secured to
their wrists. It has been suggested that this position adversely
affects a person's ability to breathe by interfering with chest
wall and abdominal movements necessary for normal
breathing. 7
As a result, sudden, unexpected deaths in persons so
restrained have been attributed to hypoventilatory respiratory failure from body position or "positional asphyxia."3-6
This theory has been based primarily on the work of Reay
et aI,7 who reported that healthy individuals placed in the
restraint position after periods of exercise had prolonged
recovery times for heart rate and oxygen saturation as measured by pulse oximetry
We sought to investigate "positional asphyxia" and respiratory compromise in subjects held in the restraint position.
A two-phase study was conducted in healthy subjects to
assess ventilatory function and gas exchange. In the first
phase, pulmonary function was measured with the subject
in various body positions, including the restraint position,
to determine whether any compromise in ventilatory mechanics occurred, In the second phase, oxygenation and ventilation were monitored by arterial blood gas measurements,
heart rate monitoring, and pulmonary function testing (PFT)
to determine whether any ventilatory compromise or alteration in gas exchange occurred when the subject was placed
in the restraint position after exercise.
MATERIALS AND METHODS

The experimental study design and protocol were reviewed
and approved by the Human Subjects Committee of the
University of California, San Diego. Fifteen healthy male
volunteers between the ages of 18 and 40 years were enrolled

NOVEMBER 1997

30:5

ANNALS OF EMERGENCY MEDICINE

in the study Informed consent was obtained from all individuals who agreed to partiCipate in the study Subjects who
completed the study were financially compensated.
Exclusion criteria included any history of pulmonary
disease (including asthma), cardiac disease, recreational drug
use, or other significant illness or disability that would limit
the ability to perform the exercise regimen required for the
study Individuals with a body mass index (EMI, defined as
the weight in kilograms divided by the square of the height
in meters) greater than 30 kglm 2 were also excluded.
Potential subjects underwent both a urine toxicologic
screening and screening PFT before acceptance into the
study Urine specimens were collected and tested by means
of toxicologic immunoassay for the presence of urinary
metabolites for the follOwing drugs: marijuana, cocaine,
amphetamines, phencyclidine, benzodiazepines, and opiates.
Individuals were excluded if the immunoassay detected any
of these substances in the urine.
Initial screening PFT were performed with spirometry as
outlined in the next paragraph. Peak expiratory flow rates
and lung volumes were measured while potential subjects
were in the sitting position. Forced vital capacity (FVC) and
forced expiratory volume in 1 second (FEV 1) were measured
while the subject was in the sitting pOSition with feet flat on
the floor and back against the chair. Spirometry was performed in accordance with the American Thoracic Society
criteria, including reproducibility within 5% variability on
three repeat measurements. s Abnormal PFT results (FVC and
FEV1) were defined as measurements below the fifth percentile of the normal range (1.65 times the SD for Student's
one-tailed t test) for each subject with a given height, age,
and race. 9 ,10 Abnormal results were verified with repeated
measurements, and subjects with confirmed FVC or FEV 1
values lower than 80% of predicted were excluded.
In phase 1 of the study, subjects underwent PFT in the
follOwing positions: sitting, supine, prone, and restraint. In
the sitting pOSition, the subject sat in a chair with his feet
flat on the floor and his back upright against the back of the
chair. In the supine position, the subject lay flat on his back
on a medical examination table with his arms at his sides
and his legs in full extension with the feet together. In the
prone pOSition, the subject lay flat on his stomach on a medical examination table with his head turned to the side, his
arms at his sides, and his legs in full extension with the feet
together. In the restraint position, the subject lay on his
stomach. To allow secure placement of a radial artery line,
each wrist and hand were taped to an armboard, which was
then taped to the sole of the ipsilateral shoe with the knee
flexed behind the subject (Figure 1). The metacarpal-phalangeal joints of the hands were placed in line with the heels

579

RESTRAINT AND ASPHYXIA
Chan et al

of the feet. This position closely approximates the restraint
position noted in previous studies and case reports in the
prehospital setting. 3 ,4,7 The subject's feet were then taped
together and his head was turned to the side. The order of
positions (supine, prone, or restrained) was randomized
to prevent any potential influence of serial testing on the
measurements obtained. The sitting position was used last
in all cases to allow for comparison with the initial screening
PFT results and to ensure that no Significant change had
occurred (because of repeated measurements) since the initial screening PFT values. Measurements obtained in each of
the four positions included FVC, FEV1 , FEV/FVC% (calculated), and maximal voluntary ventilation (MW). MW was
repeated twice for a duration of at least 6 seconds to ensure
reproducibility
In phase 2 of the study, subjects underwent two exercise
periods and two rest periods. The first rest period occurred
with the subject in the sitting position and the second in the
restraint position; the order of positions was not randomized. During each rest period, serial arterial blood gas (ABG)
measurements, pulse rate, and oxygen saturation by cooximetry and pulse oximetry were recorded according to the
protocol described in the following paragraphs. Additional
PFT measurements were also performed during the rest periods, as described.
To allow for multiple blood gas samples, a 20-gauge radial
arterial line was placed in one of the subject'S wrists. Threelead ECG monitor leads were placed for continuous heart
rate monitoring. Transcutaneous oximetry was measured
by pulse oximeter probes placed on the ear lobe ipsilateral
Figure 1.

Diagram of restraint position. Subjects were placed in the
prone position with hands and wrists taped to armboards
behind the back and secured to the feet with knees flexed.
Diagram by S Manitsas.

580

to the side of the radial arterial line and on the index finger
of the hand contralateral to the radial arterial line.
In the first exercise period, subjects exercised on a cycle
ergometer at 175 W for 4 minutes. ABG samples were drawn
immediately before the start of exercise and immediately
after the end of exercise. Oxygen saturation by ear and finger probes and pulse rate by ECG tracing were recorded at
the start of exercise, at 2 minutes into exercise, and at the
end of exercise.
After the first exercise period the subject rested in the sitting position for 15 minutes. During this first rest period,
blood samples for ABG analysis were obtained at l.5 minutes and 15 minutes into rest. PFT was performed at 3 minutes into the first rest period. Oxygen saturation by both
ear and finger probes and pulse rate by ECG tracing were
recorded every 3 minutes during the first rest phase.
Once the heart rate had returned to less than 100 beats/
minute, the subject underwent a second exercise period,
performing the same level of exercise as in the first period.
Blood samples for ABG analysiS were obtained before the
start of exercise and immediately after the end of exercise.
Oxygen saturation by both ear and finger probes and pulse
rate by ECG tracing were recorded at the start of exercise,
at 2 minutes into exercise, and at the end of exercise.
The subject was placed in the restraint position immediately after the second exercise period and remained in
this position for 15 minutes. Blood samples for ABG analysis were obtained at l.5 minutes and 15 minutes into the
restraint period. PFT was performed at 3 minutes into the
restraint period. Oxygen saturation by both ear and finger
probes and pulse rate by ECG tracing were recorded every
3 minutes during this period. All ABG sampling and analyses were performed in a uniform manner, and results were
verified by repeated testing of each sample on separate
machines that were both internally and manually calibrated
on a daily basis.
Raw PFT measurements were converted to percentages
of predicted values (% predicted) for each subject to allow
for normalization for age, height, and race. 9 ,10 Results are
reported as mean±SD. For phase 1 of the study, one-way
ANOVA for repeated measures, with position as the factor,
and Students t test were used to detect any statistically significant difference in the randomized PFT measurements.
A probability value of less than .05 was considered statistically significant. For phase 2 of the study, two-way ANOVA
for repeated measures, with pOSition and time as the factors,
and Student's t test were used to detect any statistically
significant difference in Po 2 , co-oximetry, Peo 2 , and pulse
rate between the two rest periods (ie, sitting versus restrained
position). A probability value of less than. 05 was considered

ANNALS OF EMERGENCY MEDICINE

30:5

NOVEMBER 1997

RESTRAINT AND ASPHYXIA
Chan et al

statistically significant. We used a computerized statistical
package (PCinfo! [Retriever Data Systems] and Biomedical
Data Programs [Statistical Packages for the Social Sciences])
for these analyses.
RESULTS

Two potential subjects were excluded for abnormal screening PFT measurements, and another individual was excluded
for BMI greater than 30 kglm 2 None of our potential subjects had a urine toxicology screen result that was positive
for recreational drug use.
The results of this study demonstrated significant changes
in both static and dynamic pulmonary function testing with
position and exercise. FVC was significantly decreased
(P<.OOl) in the supine position (mean, 4.95±.93 L [94%
±9.9% of predicted]; range, 3.21 to 7.20 L), the prone
position (4.92±.95 L [94%±10.2%l; range, 3.28 to 7.18 L),
and the restraint position (4.60±.84 L [88%±8.8%]; range,
3.34 to 6.54 L), compared with the sitting position (5.31
±1.01 L [101 %±10.5%]; range, 3.65 to 7.90 L). FVC was

also significantly less (P<.OOl) in the restraint position compared with the supine or the prone position, but the difference between the supine and prone positions was not
significant. Exercise appeared to have no effect on the FVC
in the sitting position (5.32±1.03 L [101 %±1O.5%l; range,
3.88 to 8.00 L; P>.30). However, in the restraint position,
exercise caused a statistically significant increase in the FVC
(4.74±.84 L [91 %± 8.7%]; range, 3.47 to 6.89 L; P<.028).
The Table reveals FVC results in terms of percentage of predicted value and the magnitude of change by position.
FEV 1 decreased similarly to FVC with change in position.
FEV 1 was significantly less (P<.OO 1) in the supine position
(3.99±.54 L [95%±8.7%l; range, 2.81 to 4.79 L), the prone
position (3.94±.52 L [94%±8.8%]; range, 2.83 to 4.63 L),
and the restraint position (3.70±.45 L [89%±7.7%]; range,
2.90 to 4.44 L), compared with the sitting position (4.3l±
.53 L [103%±8.4%]; range, 3.19 to 4.95 L). The FEV 1 was
also less in the restraint position compared with the supine
or the prone position (P<.OOl), but the difference between
the supine and prone positions was not significant. Exercise
appeared to increase the FEV 1 in both the sitting position

Table.
Changes in static and dynamic pulmonary function tests in various positions. Percentages of predicted values were determined after
normalization for height, age, and race in sitting position. 9.10 Probability values reflect significance compared with pre-exercise sitting values.
% of Predicted
Parameter
FVC(L)
Sitting
Postexercise sitting
Supine
Prone
Restraint
Postexercise restraint
FEV, (L)
Sitting
Postexercise sitting
Supine
Prone
Restraint
Postexercise restraint
FEV,/FVC%
Sitting
Postexercise sitting
Supine
Prone
Restraint
Postexercise restraint
MVV
Sitting
Supine
Prone
Restraint

NOVEMBER 1997

30:5

Measurement (Mean±SDI

Mean±SD

Range

5.31±1.01
532±1.03
4.95±.93
4.92±.95
4.60±.84
4.74±.84

101±1O.5
101±1O.5
94±9.9
94±1O.2
88±8.8
91±8.7

91-124
97-125
80-113
82-112
83-102
87-108

0
-7
-7
-13
-10

>30
<.001
<.001
<.001
<.001

431±.53
4.47±.52
3.99±.54
3.94±.52
3.70±.45
3.93±38

103±8.4
107±9.1
95±8.7
94±8.8
89±7.7
94±8.7

94-124
101-103
83-96
83-97
85-93
93-113

+4
-8
-9
-14
-9

<.001
<.001
<001
<.DOl
<.001

82.0±6.41
85.1±8.08
81A±5.61
81.1±6.14
81.2±6.34
83.4±8.68

102±7.1
106±9.2
101±6.0
101±6.8
101±7.0
104±9.8

80-108
82-118
85-103
81-102
81-106
76-114

+4
-1
-1
-1
+2

<.001
>.15
>.15
>.15
>.15

165.5±24.5
151.5±22.2
143.5±20.0
131.1±20.7

111±17.3
1o1±13. 9
96±14.8
88±16.6

90-123
73-113
76-128
65-120

-10
-15
-23

<.001
<.001
<.001

ANNALS OF EMERGENCY MEDICINE

Change From Sitting

P

581

RESTRAINT AND ASPHYXIA
Chan et al

(4.47±.52 L [107%±9.1 %l; range, 3.19 to 4.95 L; P<.OOl)
and the restraint position (3.93±.83 L [94%±8.7%1; range,
3.17 to 4.51 L; P<.OOl). Because there were similar decreases in FEV 1 and FVC, there were no significant changes
in FEV/FVC% (P>.15). The Table reveals FEV 1 and FEV/
FVC% results in terms of percentage of predicted value and
the magnitude of change by position.
MW decreased in a statistically significant fashion
(P<.OOl) from the sitting (165.5±24.5 Uminute [111 %±
17.3%1; range, 128 to 210 Uminute), to the supine (151.5±
22.2 Uminute [101 %±13.9%1; range, 105 to 185 Uminute),
to the prone (143.5±20.0 Uminute [96%±14.8%1; range,
109 to 185 Uminute), to the restraint positions (131.l±20.7
Uminute [88%±16.6%1; range, 93 to 173 Uminute). The
Table reveals MW results in terms of percentage of predicted
value and the magnitude of change by position.
This study also revealed significant changes in gas exchange with exercise in both the sitting and the restraint
position. In the sitting position, gas exchange improved with
exercise (P<.OOl). The P0 2 increased from a baseline of
91.4±7.2 mm Hg (range, 81 to 105 mm Hg) to 108.7±8.0

(P<.OOl).

Immediately after exercise, just before the subject was
placed in the restraint position, the P0 2 had again improved
(P<.012), from a baseline of 102.4±8.7 mm Hg (range, 84
to 115 mm Hg) to 109.8±9.3 mm Hg (range, 93 to 128
mm Hg) after 4 minutes of exercise. As the subjects were
being placed in the restraint position (1.5 minutes after the

HR

Figure 2.

Heart rate recovery during
resting periOds: sitting versus
restraint positions. The Xaxis indicates time into each
rest period (in minutes) after
cessation of exercise. The Yaxis indicates mean heart
rate in beats/minute for the
15 subjects. Standard deviation of the mean is shown by
error bars.

mm Hg (range, 96 to 125 mm Hg) after 4 minutes of exercise. After 1.5 minutes of rest the P0 2 was 122.7±5.8 mm
Hg (range, 110 to 132 mm Hg); after 15 minutes of rest it
had fallen back toward baseline and was 102.4±8.7 mm Hg
(range, 84 to 115 mm Hg). The Peo 2 also changed significantly with exercise (P<.OOl). From a baseline of 38.5±2.8
mm Hg (range, 31 to 41 mm Hg), Peo 2 fell to 34.5±3.9
mm Hg (range, 29 to 41 mm Hg) after 4 minutes of exercise. The Peo 2 was 31.3±2.6 mm Hg (range, 28 to 36
mm Hg) after 1.5 minutes ofrest; after 15 minutes of rest
the Peo 2 remained lower than baseline at 32.9±2.7 mm Hg
(range, 28 to 37 mm Hg). Because the changes in both Pco 2
and P0 2 were statistically Significant, the change in the alveolar-arterial oxygen gradient across time was also significant

170

160

150

140

130

120

110

100

90-L-----,------,------,------,------.------,-----,--1.5

6

9

15

12

Time Into Rest Period (minutes)

582

ANNALS OF EMERGENCY MEDICINE

30:5

NOVEMBER 1997

RESTRAINT AND ASPHYXIA
Chan et al

end of exercise) the P0 2 increased further to 114.0±7.6
mm Hg (range, 103 to 128 mm Hg), and after 15 minutes
in the restraint position the P0 2 was 99 .l±9.3 mm Hg
(range, 86 to 120 mm Hg). Likewise, the Pco 2 fell (P<.OOI)
with exercise, from an initial 32.9±2.7 mm Hg (range, 28
to 37 mm Hg) to 30.7±3.8 mm Hg (range, 24 to 37 mm Hg)
after 4 minutes of exercise, just before the subject was placed
in the restraint position. As the subjects were being placed
in the restraint position the Peo 2 was 31.0±3.4 mm Hg
(range, 26 to 37 mm Hg); after 15 minutes in this position
the Pco 2 remained lower than baseline at 32.7±3.2 mm Hg
(range, 26 to 38 mm Hg).
Comparing the ABG results in the sitting position to the
results as the subjects were being placed in the restraint
position (ie, 1.5 minutes into rest in the sitting position versus 1.5 minutes into the restraint position), the P0 2 increased
to a smaller value with restraint than with sitting 014. O± 7.6
mm Hgversus 122.7±5.8 mm Hg; P<.OI), whereas the Pco 2
was the same in both groups (31.3±2.6 mm Hg versus
31.0±3.4 mm Hg; P>.28). Comparison of ABG results after
15 minutes in the sitting position with those after 15 minutes in the restraint position revealed no significant differ-

ences in either the P0 2 (l02.5±8.7 mm Hg versus 99.1
±9.3 mm Hg; 0.234) or the Pco 2 (32.9±2.7 mm Hg versus
32.7±3.2 mm Hg; P>.30).
Changes in heart rate also occurred with exercise, with
a maximum heart rate of 164±18.9 beats/minute (range,
218 to 180) at the beginning of the sitting rest period and
174±15.3 beats/minute (range, 146 to 198) at the beginning of the restraint rest period. Throughout the two rest
periods, there was no statistically significant difference in
mean heart rate recovery (Figure 2).
Examination of oxygen saturation as measured by cooximetry revealed only minor, nonsignificant increases over
the initial baseline. Comparison between oxygen saturation
values measured by co-oximetry and by pulse oximetry also
revealed no statistically significant differences for any time
period (Figure 3).
DISCUSSION

The term "positional or mechanical asphyxia" has been used
to explain the deaths of certain individuals. Bell et al l l described 30 such cases of individuals whose bodies were
found in positions that caused either external airway obstruc-

% Saturation

Figure 3.

Oxygen saturation over
time. The X-axis indicates
the various time periods
(Exl, first exercise period;
Rl,first rest period in sitting
position; Ex2, second exercise period; R2, second rest
period in restraint position).
The Y-axis indicates mean
oxygen saturation by pulse
oximetry (both finger and
ear monitors) for the 15
subjects. Standard deviation
of the mean is shown by
error bars.

100
98
96
94

92
90
88
__

Ear Probe

86
Finger Probe

84
82
8o~-,-------,-------------------.-------.------------------

Exl

Rl

Ex2

R2

Time Periods

Ext first exercise period; Rt first rest period; Ex2, second exercise period; R2, restraint period.

NOVEMBER 1997

305

ANNALS OF EMERGENCY MEDICINE

583

RESTRAINT AND ASPHYXIA
Chan et al

tion or inadequate ventilatory function. In all cases, there
were no other significant life-ending pathologic findings
and the deaths were attributed to "positional asphyxiation."
Similar asphyxiation deaths have been described involving
the use of vest, jacket, or posey restraints that were accidentally wrapped around the necks of nursing home and geriatric patients and resulted in strangulation. 12-15 There have
been a few reports of ventilatory failure and asphyxiation
caused by restraints that reportedly compressed the chest
and abdomen to the point that mechanical ventilatory function was impaired. 16-18 Recently, attention has focused on
the possible role of "positional asphyxia" in deaths that have
occurred in persons restrained in the "hobble" or "hogtie"
position by law enforcement or prehospital field personnel.
This position places the individual prone with the wrists
and ankles bound behind the back. 2-6 Reay et aP suggested
that this position prevents adequate chest wall expansion
and abdominal and diaphragmatic excursion for normal
ventilatory function and breathing. They postulated that
this inability to expand the thoracic cavity, and disadvantage
in pulmonary mechanics, leads to hypoventilatory respiratory compromise, asphyxiation, and death.
Hirsh 18 argued that the employment of this restraint
position constitutes use of a "potentially lethal force" and
that such deaths should be classified as homicides. However,
most of the reported cases have involved young men in a
state of "excited" or "agitated delirium" as a result of intoxication from recreational drugs or psychiatric illness. In addition, these individuals had often suffered traumatic injuries
before and during placement in the restraint position. Many
have argued that these other factors (ie, intoxication, stress,
and trauma), as opposed to asphyxiation strictly due to body
positioning, played a greater role in causing these deaths. 19,20
Reay et aF studied 10 healthy individuals and found measurable physiologic effects with application of the restraint
position after exercise. They reported that prolonged times
were required for recovery to baseline in the restraint position after exercise, both for heart rate (mean, .40 minutes
longer) and for peripheral oxygen saturation measured by
transcutaneous ear probe (mean, .33 minutes longer). Based
on these findings, they suggested that positional asphyxia
plays an important role in the deaths of persons placed in
the restraint position.
A number of concerns exist with the work of Reay et al.
First, there was no assessment of actual ventilatory function
and respiratory mechanics in subjects placed in the restraint
position. Second, although Reay et al reported a drop in
oxygen saturation (to 85% to 90%) with exercise in their
subjects, previous work has demonstrated improvements in
arterial oxygenation with mild to moderate levels of exer-

584

cise in healthy individuals.21 Third, the preferred method
for assessing arterial blood oxygenation remains ABG measurements. Oxygen saturation measurement by pulse oximetry has been shown to be a potentially inaccurate measure
of arterial oxygenation, particularly during exercise. 22 - 25
In this study, we assessed ventilatory function by PFT
with subjects in various body positions. We found a restrictive pulmonary function pattern in healthy subjects placed
in the restraint position, with small but significant decreases
in the percentage of predicted FVC and FEV l' Associated
with these drops in lung volumes was a corresponding 23%
decrease in percentage of predicted MVV There was no evidence of obstruction and no significant change in FEV /
FVC% from baseline.
Given the fact that PFT measurements as low as 80% of
predicted values are still considered clinically nonnal, these
changes, although statistically significant, are not clinically
relevant. 26 In reviewing the ranges of PFT data we obtained,
none of our subjects had results for FVC and FEV1 lower
than 80% of predicted in any of the positions, including the
restraint position. The range of data for MW did fall below
80% of predicted for certain outlier individuals in the restraint position, although similar decreases below 80% of
predicted were also seen in the supine and prone positions.
In contrast to the studies of Reay et aI, arterial oxygenation measurements were obtained in this study by both ABG
sampling and transcutaneous finger and ear oxygen saturation probes. Based on arterial P0 2 and co-oximetry, we
found that oxygenation increases, rather than decreases,
with exercise. This finding is consistent with previous, wellestablished work on exercise physiology.21,27 In addition,
despite our PFT findings, we found no evidence of hypoxia
while subjects were in the restraint position after exercise.
The improvements in oxygenation occurred in the face of
a more vigorous exercise regimen than that described by
Reay et al. Reay required subjects to exercise on a crosscountry skiing machine only until their heart rates reached
a maximum of 120 beats/minute. The subjects in this study
exercised for 4 minutes continuously on an exercise bicycle,
and the mean heart rate at the end of exercise was 169
beats/minute.
Equally important, despite decreases in MVY, there was
no evidence of hypercapnia either during exercise or during
rest in the restraint position. In fact, mean Peo 2 levels decreased during exercise and remained lower than 40 mm Hg
for as long as 15 minutes during the restraint rest position.
Despite the restrictive pattern demonstrated by PFT, there
was no evidence for ventilatory failure, significant hypoventilation, or asphyxiation as a result of body positioning
while subjects were in the restraint position.

ANNALS OF EMERGENCY MEDICINE

30:5

NOVEMBER 1997

RESTRAINT AND ASPHYXIA
Chan et al

Based on these findings in healthy subjects, we suggest
that factors other than body positioning are more important determinants for the sudden, unexpected deaths that
occur in individuals who are placed in the restraint position.
Recreational drug use (including sympathomimetic, hallucinogenic, and psychomotor stimulant drugs), physiological
stress, hyperactivity, hyperthermia, catechol hyperstimulation, and trauma resulting from struggle may be more important factors in the deaths of these individuals. 19,20,28,29
Although restraints in general increase the psychological
and physiologic stresses on the individual,28,30 there is no
evidence that body position while in the "hogtie" or "hobble"
restraint position as a factor in and of itself causes hypoventilation or asphyxiation.
There are limitations to this study First, we restricted
subjects to healthy men between the ages of 18 and 40 years
with a BMIless than 30 kglm 2; most cases reported in the
literature involve this population. It is not known what effect
positional restraint may have on women, the young, the
elderly, or other individuals with underlying cardiopulmonary
disease or disability It is possible that extremely obese individuals with large abdominal girths and BMIs greater than
30 kglm2 may be at greater risk for development of restrictive pulmonary function pattern as a result of abdominal
compression from body position.
We specifically excluded potential subjects who had a
positive result on urine toxicology screening for recreational
drug use. As noted previously, many of the deaths of restrained individuals involved subjects who were intoxicated
or under the influence of recreational drugs. Stimulants, such
as cocaine and amphetamines, may increase oxygen demand
and muscle fatigue, affecting overall respiratory function.
Although we randomized the order of positions for PFT,
we did not randomize the sitting and restraint position rest
phases after exercise in our study All subjects rested in the
sitting position first, after the initial exercise period, then
exercised again and rested in the restraint position. We believe that 15 minutes of sitting rest should have been adequate time for values to return to baseline after 4 minutes
of exercise. Any residual metabolic or respiratory derangements remaining after the first rest period would bias our
study in favor of detecting Significant abnormalities in the
restraint position. In addition, we did not measure respiratory rate as an indicator of ventilatory status. We believe
that the combination of MW and serial Peo 2 provides a
more appropriate measure of ventilatory status than does
respiratory rate.
This study did not attempt to duplicate exact field conditions under which restraint position deaths have occurred.
Although many such deaths have occurred on gurney mat-

NOVEMBER 1997

30:5

ANNALS OF EMERGENCY MEDICINE

tresses or cushioned car seats in the field, some deaths have
occurred while persons were in the restraint position on
the ground 3 Deaths have also occurred on the floors of
police cars, where the contoured surface may have increased
abdominal compression. 3 In addition, these individuals may
have been subject to forceful apprehension, during which
pressure may have been exerted on their backs while they
were in the restraint position. What effects these differences
may have remain to be determined.
We kept our subjects in the restraint pOSition for 15
minutes after the exercise period. We believe that this was
adequate time to detect any physiologic or respiratory impairment in subjects. It is possible that had our subjects
remained in the restraint position for a longer period we
may have detected more significant alterations in respiratory
phYSiology However, most death of individuals in the restraint position have occurred after only a short period
(often less than 10 minutes) in restraint,3-6 and it has been
suggested by others that this short duration of restraint
positioning should not be fatal. 19
We attempted to reproduce the physiologic effects of
struggle by requiring our subjects to exercise for 4 minutes
before being placed in the restraint position. It is unlikely
that this period of exercise would simulate all the physiologic alterations that may occur vvith struggle and agitation.
In addition, we did not reproduce the effects of trauma and
psychological stress that often occur with apprehended individuals. However, the respiratory mechanics of our subjects
(as evidenced by increased FVC and FEV1) improved with
exercise.
It is possible that a combination of factors, including
underlying medical condition, intoxication, agitation, delirium, and struggle as well as body position, may result in
respiratory compromise that would not be detected by our
study We sought to examine only the role of body position
as a factor affecting pulmonary function and respiratory
phYSiology Although we found that body position by itself
does not result in significant respiratory compromise, further
research is needed on the role of these other factors in the
deaths of individuals placed in the restraint position.
In conclusion, although we found a small restrictive pulmonary function pattern by PFT parameters in subjects who
were placed in the restraint position, we found no evidence
of hypoxia or hypercapnia on serial ABG measurements.
By itself, the restraint position was not associated with any
clinically relevant changes in respiratory or ventilatory function in our study population of healthy individuals with
preserved ventilatory reflexes and normal pulmonary physiology There is no evidence to suggest that hypoventilatory
respiratory failure or asphyxiation occurs as a direct result

585

RESTRAINT AND ASPHYXIA
Chan et al

of body restraint position in healthy, awake, nonintoxicated
individuals with normal cardiopulmonary function at
baseline.

29. Robinson BE, Sucholeiki R, Schocken DD: Sudden death and resisted mechanical restraint: A
case report. JAm Geriatr Soc 1993;41 :424.
30. Pudiak CM. Bozarth MA: Cocaine fatalities increased by restraint stress. Life Sci 1994;55:379.

The authors thank Jeffrey Johnson, Carlos Lopez, and Paul Schragg for their assistance.

REFERENCES
1. Reay OT, Eisele JW: Deaths from law enforcement neck holds. Am J Forensic Med Pathol
1982;3:253.

Reprint no. 47/1/85377
Address for reprints:

2. Luke JL, Reay DT: The perils of investigating and certifying deaths in police custody. Am J
Forensic Med Patho/1992; 13:98.

Theodore Chan, MD

3. Reay OT. Fligner CL, Stilwell AD, et al: Positional asphyxia during law enforcement transport.
Am J Forensic Med Patho/1992;13:90.

University of California San Diego Medical Center

Department of Emergency Medicine
200 West Arbor Drive #8676

4. O'Halloran RL, Lewman LV: Restraint asphyxiation in excited delirium. Am J Forensic Med
Patho/1993; 14:289

San Diego, California 92103

5. Stratton SJ, Rogers C, Green K: Sudden death in individuals in hobble restraints during
paramedic transport. Ann Emerg Med 1995;25:71 O.

Fax 619-543-3115

6. Cowen AR: In a bind. A lawsuit forces LA medics and police to reexamine their policy on applying restraints. Emergency Medical Services 1995;24:54-8.

619-543-6463
E-mail tcchan@ucsd.edu

7. Reay DT. Howard JO, Fligner CL. Ward RJ: Effects of positional restraint on oxygen saturation
and heart rate following exercise. Am J Forensic Med Patho/1988;9:16.
8. American Thoracic Society: Standardization of spirometry: 1994 Update. Am J Respir Crit Care
Med1995;152:1107.
9. Morris JF: Spirometric standards for healthy non-smoking adults. Am Rev Respir Dis 1971;103:57.
10. Bass H: The flow volume loop: Normal standards and abnormalities in chronic obstructive
pulmonary disease. Chest 1973;63:171.
11. Bell MD, Rao VJ, Wetli CV, et al: Positional asphyxiation in adults: A series of 30 cases from
the Dade and Broward county Florida medical examiner offices from 1982 to 1990. Am J Forensic
Med Patho/1992;13:1 01.
12. Dube AH, Mitchell EK: Accidental strangulation from vest restraints. JAMA 1986;256:2725.
13. Katz L: Accidental strangulation from vest restraints. JAMA 1987;257:2032.
14. DiMaio VJM, Dana SE, Bux RC: Deaths caused by restraint vests. JAMA 1986;256:905.
15. Miles S: A case of death by physical restraint: New lessons from a photograph. JAm Geriatr
Soc 1996;44:291.
16. Miles HS, Irvine P: Deaths caused by physical restraints. Gerontologist 1992;32:762.
17. Emson HE: Death in a restraint jacket from mechanical asphyxia. Can Med Assoc J 1994;151:
985.
18. Hirsh CS: Restraint asphyxiation. Am J Forensic Med PathoI1994;15:266.
19. Laposata EA: Positional asphyxia during law enforcement transport. Am J Forensic Med
Patho/1993;14:86.
20. Karch SB: Agitated delirium versus positional asphyxia. Ann Emerg Med1995;26:760
21. Wasserman K, Hansen JE, Sue DY, et al: Normal values, in Wasserman K, Hansen JE, Sue
DY, et al (edsl: Principles of Exercise Testing and Interpretation, ed 2. Philadelphia: Lea &
Febiger, ·1994: 127-128
22. Biebuyck JF: Pulse oximetry. Anesthesiology 1989;70:98.
23. Norton LH, Squires B, Craig NP, et al: Accuracy of pulse oximetry during exercise stress testing. Int J Sports Med 1992;13:523.
24. Hansen JE, Casaburi R: Validity of ear oximetry in clinical exercise testing. Chest 1987;91 :333.
25. Brown DD, Knowlton RG, Sanjab PB, et al: Re-examination of the incidence of exerciseinduced hypoxaemia in highly trained subjects. Sr J Sports Med 1993;27:167.
26. Clausen JL: Pulmonary function testing, in Bordow RA, Moser KM ledsl: Manual of Clinical
Problems in Pulmonary Medicine, ed 4. Boston: Little, Brown, 1996:9-19.

27. Comroe JH (edl: Respiratory adjustments in health. in Physiology of Respiration, ed 2.
Chicago: Yearbook Medical, 1975:234-241.
28. Mirchandani HG, Rorke LB, Sekula-Perlman A, et al: Cocaine-induced agitated delirium, forceful struggle, and minor head injury. Am J Forensic Med Patho/1994;15:95.

586

ANNALS OF EMERGENCY MEDICINE

30:5

NOVEMBER 1997