Full text of DOI:10.1080/10903120290938805

EVALUATION OF AN ELECTRONIC ESOPHAGEAL DETECTOR DEVICE IN
PATIENTS WITH MORBID OBESITY AND PULMONARY FAILURE
Timothy R. Wolfe, MD, Edward J. Kimball, MD, L. Lazarre Ogden, MD, Pat Schafer, MD,
Stephen C. Hartsell, MD, Scott Richardson, MD, Matthew R. Moog, MD, Richard Barton, MD
ABSTRACT
Undetected esophageal intubation can result in per-
manent injury or death.1 Clinical confirmation is oper-
Objective. Undetected esophageal intubation can result in
permanent injury or death. Clinical confirmation of tube
ator-dependent and may be misleading.2,3 End-tidal
location may be misleading. Adjunctive methods should be
used to supplement clinical judgment. Unfortunately, end-
tidal carbon dioxide may misidentify properly placed tra-
cheal tubes in low perfusion situations, while esophageal
detector devices (EDDs) may misidentify properly placed
tracheal tubes in situations where little airway dead space
exists (morbid obesity, pulmonary failure). This study eval-
uated a modified EDD (the electronic esophageal detector
device, or EEDD) designed to eliminate the problem of
misidentified tracheal intubations. Methods. Intubated mor-
bidly obese or pulmonary failure patients were eligible for
study entry. All endotracheal tubes (ETTs) were confirmed
to be tracheal by waveform capnography and clinical judg-
ment prior to study entry. Following consent, all patients
were attached to the EEDD and a “measurement” was made
to determine the “location” of their ETTs. Probability of
misidentifying a tracheal intubation in these high-risk pop-
ulations was calculated using a log-normal distribution
method. Results. Twenty-seven morbidly obese patients
and 37 pulmonary failure patients were entered. The EEDD
correctly identified all tracheal intubations in these patients,
giving a false-negative rate of zero. The probability of
misidentifying a tracheal intubation in the combined group
was 0.06%. Conclusion. This study demonstrates that the
EEDD reliably identifies tracheal intubations in situations
where standard EDDs may fail. However, future studies
must determine the reliability of this device for identifica-
tion of esophageal intubations and the reliability of this
device in the less controlled emergency department and pre-
hospital settings. Key words: intubation confirmation;
esophageal detector device; esophageal intubation; obesity;
pulmonary failure.
carbon dioxide (ETCO2) and esophageal detector
devices (EDDs) offer methods other than clinical judg-
ment for intubation confirmation. Measurements of
ETCO2 and EDDs are very accurate at identifying
esophageal intubations. However, false-negative
results occur with both ETCO2 and EDDs in the pre-
hospital setting. (False-negative result is defined as an
endotracheal tube that is located in the trachea, but the
confirmation device identifies the tube as esophageal
in location. For example, a false-negative ETCO2 result
would be failure to detect CO2, whereas a false-nega-
tive EDD result would be failure to aspirate air.) The
most common reason for false-negative ETCO2 is car-
diac arrest,4,5 while morbid obesity accounts for the
majority of false-negative EDD findings.6,7 In fact,
Lang et al. found a 30% misidentification rate of prop-
erly placed endotracheal tubes (ETTs) in the morbidly
obese using a bulb model EDD.6 Bronchoscopy in
these patients demonstrated airway collapse pre-
sumed to be due to the weight of the patient’s chest,
with resultant loss of airway dead space and available
air for aspiration. In addition to morbid obesity, other
factors such as bronchospasm, fluid-filled airway,
right mainstem intubation, third trimester pregnancy,
and infants also can cause reduced residual volumes
and subsequent false-negative EDD results.7–12
This study evaluated a modification of the EDD. The
modification was specifically designed to eliminate
the false-negative results found with the standard
EDD. The modification, called an electronic esopha-
geal detector device (EEDD), consists of an aspirating
device, a pressure decay measuring transducer (meas-
ures drop in pressure per unit of time), and a micro-
processor that interprets and identifies the pressure
decay within the ETT as either esophageal or tracheal.
Because EDD technology is based on the differences in
resistance to aspiration between the esophagus (col-
lapsed fibromuscular tube with significant resistance
to aspiration) and the trachea (rigid structure full of
air that is easily aspirated), these differences are pro-
found in most patients. The purpose of this study was
to determine the accuracy of this modified device in
two populations with a high incidence of false-nega-
tive results: morbidly obese patients and critically ill
patients suffering pulmonary failure.
PREHOSPITAL EMERGENCY CARE 2002;6:59–64
Received April 2, 2001, from the Division of Emergency Medicine
(TRW, SCH), Department of Surgery (EJK, RB), and Department of
Anesthesiology (LLO, PS, SR, MRM), University of Utah School of
Medicine, Salt Lake City, Utah. Revision received July 9, 2001;
accepted for publication July 13, 2001.
Presented at the American College of Emergency Physicians
Research Forum, Las Vegas Nevada, October 1999.
Wolfe Tory Medical Inc. supplied the device and necessary supplies
for this study. Dr. Wolfe is the president and a shareholder of the
company that manufactured the device used in this study.
Address correspondence and reprint requests to: Tim Wolfe, MD,
Division of Emergency Medicine, 1150 Moran Building, 75 North
Medical Drive, Salt Lake City, UT 84132. e-mail: <wolfeman@
qwest.net>.
59

60
PREHOSPITAL EMERGENCY CARE JANUARY / MARCH 2002 VOLUME 6 / NUMBER 1
degradation than the raw area. Based on these prelim-
inary data, the accept/reject point for this study was
prospectively defined as 80%. In other words, any
tracing that had an area of ≥80% was “rejected” and
identified as esophageal in location. Time zero was
defined as 0.1 of a second after the aspiration started
to reduce human variability at the release of the aspi-
ration chamber and to eliminate the initial small drop
found in preliminary esophageal tracings.
Morbidly obese patients were identified during the
preoperative anesthesia evaluations conducted in the
preoperative clinic. Height in centimeters and weight
in kilograms were calculated from the actual patient
weights and self-reported heights. Morbid obesity
was defined as a body mass index >35 kg/m2 deter-
mined by the body mass index equation BMI = (kg)/
(ht in meters).2 Inclusion criteria were morbid obesity,
adult patient aged 18 years or older undergoing rou-
tine elective surgery, and American Society of
Anesthesiologists physical status 1–3. Exclusion crite-
ria were pregnancy and prisoner status. Appropriate
candidates were consented preoperatively.
FIGURE 1. The electronic esophageal detector device.
Patients included in the study were monitored with
the standard monitors, including: pulse oximetry,
heart rate, electrocardiography (ECG), noninvasive
MATERIALS AND METHODS
This was a prospective, observational study conduct-
ed at the University of Utah in Salt Lake City, Utah.
The University of Utah Institutional Review Board
approved the study. Written consent was obtained
from all participants or their families. Due to the com-
plexities of obtaining informed consent from family
members, patients were entered at the convenience of
the investigators rather than consecutively.
The EEDD consists of several components (Fig. 1): a
15-mm internal diameter ETT adapter with an internal
O-ring seal that prevents air leak around the connec-
tion; a 50-mL spring-loaded aspiration chamber that
generates a maximum negative pressure of 175 mm
Hg when attached to an occluded 8.0 ETT; a pressure
transducer integrated into the aspiration chamber that
measures pressure changes over a predetermined time
and converts them to a digital electronic signal; a
coiled extension tubing (1/8 inch internal diameter, 62
inches long, 12.5 mL dead space) that attaches the
device to the ETT; a microprocessor that compiles and
analyzes the pressure signature; and a set of three
lights that indicate esophagus (red), trachea (green), or
error due to connection failure or compression failure
(yellow).
A preliminary derivation study of 40 healthy elec-
tive surgical patients suggested that an area under the
curve (AUC; reduction in mm Hg per second) would
most accurately distinguish tracheal from esophageal
intubations (Fig. 2). This total area was converted into
a percentage area of the “box” created by drawing a
horizontal line from the highest initial pressure point.
This allowed a more consistent measure of pressure
FIGURE 2. Derivation data comparing esophageal and tracheal intu-
bation tracings. Upper, horizontal tracings are from the esophagus;
lower, rapidly dropping tracings are from the trachea. AUC = area
under the curve.

Wolfe et al. ELECTRONIC ESOPHAGEAL DETECTOR DEVICE
61
TABLE 1. Patient Demographics
acquired, the probability of misidentification of a
future properly placed tracheal tube in this patient
population (i.e., the probability of the AUC’s exceed-
ing 80%) was calculated using the log-normal density
function. This is then reported as a p-value. The mean
AUCs for morbidly obese and ICU patients were com-
pared using Student’s t-test.
Intensive
Care Unit
Morbid Obesity
Weight in kilograms (SD)
137.7 (28.0)
46.0 (8.17)
21.3% (17.7)
94.1 (32.2)
32.2 (9.2)
12.5% (9.6)
Body mass index in kg/m2 (SD)
Percent area under the curve (SD)
Probability of misidentifying a
tracheal intubation
0.6%
0.001%
RESULTS
Morbid Obesity Subgroup
blood pressure, and ETCO2 monitors. Patients were
preoxygenated and underwent anesthetic inductions
according to the primary anesthesia care team. No
alterations were made with regard to anesthetic
planned technique or planned surgery. No emergent
intubations were studied.
Once the anesthesia provider had secured the ETT
with tape, the EEDD was occluded and tested to
ensure no error existed. After confirming that the
device functioned properly, the ETT was tested by one
of the anesthesiologists participating in the study. All
patients were receiving 100% oxygen after induction
and intubation. A disconnection from the circuit was
necessary to test the device. The device testing took
less than 5 seconds for each measurement. Patients
were reconnected to the anesthesia circuit immediate-
ly after each test. Pressure degradation points and
AUC data were checked to see that they were record-
ed and saved.
Twenty-seven patients were entered into the morbid
obesity study arm. Patient demographics are listed in
Table 1. Pressure degradation curves are presented in
Figure 3. The mean AUC for this group was 21.3% (SD
= 17.7). The EEDD correctly identified 100% of these
patients’ ETT locations. However, there were two out-
lying pressure tracings in this group (marked with tri-
angles and crosses in Fig. 3). The percent AUCs for
these two patients were 76.1% and 45.1%, respective-
ly. Both these patients had substantially higher BMIs
than the mean for the group (61.70 and 55.21 versus
group mean of 46.04). Nevertheless, their percent
AUCs fell below the predefined reject setting and their
ETTs were properly identified as tracheal in location.
There were no yellow lights to indicate data error and
no red lights to indicate esophageal tube location.
There were no desaturations during the testing period
and no adverse events occurred.
Critically ill, intubated patients admitted to a terti-
ary care, 12-bed surgical intensive care unit (ICU)
were also eligible for entry into the study. Exclusion
criteria were age under 18 years, pregnancy, prisoner
status, inability of the patient or family member to
consent, and belief of the intensivist that the patient
was too unstable to disconnect from the ventilator for
5 to 10 seconds. Informed consent was obtained from
the patient or family member.
Prior to patient testing the EEDD was occluded and
tested to ensure no “error” existed. After confirming
that the device functioned properly, consented
patients were disconnected from the ventilator circuit
and attached to the EEDD via the adapter and exten-
sion tubing. A single EEDD aspiration test was con-
ducted and patients were then reconnected to the ven-
tilator circuit. The entire process took less than 5 sec-
onds. All patients were monitored with pulse oxime-
try, invasive or noninvasive blood pressure monitor-
ing, and continuous ECG rhythm monitoring.
Statistical analysis was performed using the
Statistica software package (version 5.10A, Statsoft,
Inc., Tulsa, OK). The main outcome measured for both
groups was the frequency that the EEDD misidenti-
fied a true tracheal tube as being esophageal in loca-
tion (false-negative rate). The data distribution fit a
log-normal distribution function (Kolmogorov-
Smirnov test not significant, p < 0.20). Using the data
ICU Subgroup
Thirty-seven patients were entered from into the ICU
study arm. The EEDD correctly identified 100% of
these patients’ ETT locations. Patient demographics
are listed in Table 1. Pressure degradation curves are
presented in Figure 4. The mean percent AUC for this
group was 12.5% (SD = 9.6). There was one outlying
pressure tracing in this group. This patient was noted
to be biting the ETT closed during the test and, until
he relaxed and a bite block was placed, could not be
bag-ventilated and he suffered oxygen desaturation.
The percent area for this patient was 66.2%. There
were no yellow lights to indicate data error and no red
lights to indicate esophageal tube location. No other
adverse events or clinically important vital sign fluc-
tuations occurred.
The combined group mean AUC was 16.2% (SD =
14.2). The mean AUC data were substantially different
between these two subgroups. Morbidly obese
patients had a significantly higher mean AUC than
did the ICU subgroup: 21.3% (SD = 17.7) versus 12.5%
(SD = 9.6); p = 0.0132. These findings suggest there is
either less air available for aspiration or greater resist-
ance to aspiration in the morbidly obese patient, a
finding consistent with other authors who have noted
this group to account for the majority of false-negative

62
PREHOSPITAL EMERGENCY CARE JANUARY / MARCH 2002 VOLUME 6 / NUMBER 1
critically ill patients suffering from respiratory failure
using 4 seconds as the bulb re-inflation time cut-off.8
Kasper and Deem found the false-negative rate
dropped to 1% in the critically ill if 10 seconds was
allowed for refilling of the bulb.13
We found that the EEDD used in the current study
had a false-negative rate of zero in these two groups.
Interestingly, although the EEDD properly identified
all ETTs, the pressure degradation data support the
notion that morbidly obese patients have greater
resistance to aspiration of air than do respiratory fail-
ure patients or elective surgical patients. In fact, even
with our improved sensitivity, two of 27 morbidly
obese patients approached the false-negative cut-off
programmed into our decision algorithm. Neverthe-
less, we found that the EEDD reliably identified tra-
cheally located ETTs in 100% of patients, suggesting it
may have a place in confirming ETT location in situa-
tions where clinical judgment and ETCO2 results may
be inadequate or unavailable.
Although this study specifically addresses the accu-
racy of the EEDD in identifying tracheal intubations,
the identification of esophageal intubations is actually
the critical issue to prevent inadvertent injury or
death.1–3,14 However, a device that fails to properly
identify tracheal intubations will cause confusion
when it indicates that an ETT is improperly placed.
The currently available inexpensive detection devices,
EDD and colorimetric ETCO2, while very accurate at
identifying esophageally placed ETTs, may give mis-
leading results when a tube is properly placed in the
trachea.4–6,8,10 This false-negative result may cause the
clinician to doubt the answer the device provides in
the future when the tube may in fact be in the esopha-
gus.
FIGURE 3. Negative pressure degradation curves for the morbidly
obese patients.
EDD results.6,8 Using the log-normal distribution
method, the probability of misidentifying a tracheal
intubation in the combined groups was found to be
0.0006.
The EEDD begins to address many issues surround-
ing EDD reliability. It uses a decision model that elim-
inates subjective interpretation of the results, it uses a
device with an O-ring seal that ensures proper attach-
ment to the ETT and prevents inadvertent air leaks,
and it uses a volume of aspirate that has been found to
reduce the incidence of false negatives while not
increasing the incidence of false positives.10,15 These
preliminary results suggest that a highly sensitive
pressure transducer combined with a properly pro-
grammed decision model may be able to accurately
identify tracheal intubations in situations where pre-
vious clinical studies suggest that EDD technology
often fails—morbid obesity and severe pulmonary
failure. Although these patients are the minority
encountered in the prehospital setting, they are not
rare and can cause some consternation when the EDD
results are in conflict with clinical judgment.
DISCUSSION
Lang et al. studied 54 morbidly obese patients with a
mean BMI of 43.9 ± 12.3 kg/m2 using a homemade
bulb-type EDD. They considered a failure to reinflate
in 4 seconds when the ETT was within the trachea to
be a false-negative result. The false-negative rate in
their study was 30% if the bulb was compressed and
then attached, and 11% if the bulb was attached and
then compressed. The cause of noninflation was felt to
be due to reduced volumes of air available for aspira-
tion in the upper airway due to collapse of the major
airways from the weight of these patients’ chests.6
Other authors have also noted morbid obesity to be
one of the more common causes of false-negative EDD
results.7,10
The second most common cause of false-negative
EDD results is in patients who have reduced volumes
of air for aspiration due to obstruction by fluid-filled
airways, inadvertent mainstem bronchus intubation,
and severe bronchospasm.7–10 Czinn et al. found the
false-negative rate to be as high as 13% in a group of
While these results are promising, this is only the
first step. Many unanswered questions regarding this
technology remain. The most obvious question is
whether the ability to accurately identify tracheal intu-

Wolfe et al. ELECTRONIC ESOPHAGEAL DETECTOR DEVICE
63
bations in situations that often “mimic” the esophagus
will affect the device’s ability to accurately identify
esophageal intubations. If the improved sensitivity for
tracheal intubations leads to reduced specificity and
misidentification of esophageal intubations, the
results are meaningless and the decision algorithm
will require further modification. Our preliminary
derivation data suggest this is unlikely (Fig. 1).
However, until we compare this device with estab-
lished methods of confirming ETT (EDD and ETCO2)
and further data are generated comparing esophageal
and tracheal intubations, we cannot be certain of this
assumption.
A second similar question is whether the clinical set-
ting in which this study occurred mimics real life. In
this study all patients were already known to be prop-
erly intubated, and already were undergoing ventila-
tion. In the prehospital setting, this device would like-
ly be used immediately after intubations to ensure
proper tube positioning prior to ventilation. Previous
authors note that ventilation with an ETT or simply
squeezing 60 mL of air into the ETT prior to utilization
of the EDD results in a reduced incidence of false-neg-
ative results.6 However, other reports suggest that sig-
nificant overventilation of an ETT that is actually
esophageal can result in false-positive results with free
air return in an EDD.8 While this is exceedingly
uncommon, the obvious implication is that it is prefer-
able to use an EDD prior to ventilation, not after ven-
tilation.
FIGURE 4. Negative pressure degradation curves for the patients in
the intensive care unit.
Other problems exist regarding the design of this
product and the ideal device. The ideal instrument
would be inexpensive, be durable, have unlimited
shelf life, require no power source, be very simple yet
extremely accurate in all situations, give a rapid
answer, and be a continuous monitor. This device is
not a continuous monitor in its current design. In
addition, it depends on a power source and on sophis-
ticated electronic equipment. Though technology has
come a long way, this can be construed as a problem,
especially in the prehospital setting where significant
temperature extremes and physical abuse of the
device can be expected and may result in inadvertent
failure to perform. Finally, with increasing sophistica-
tion come increased costs. Nevertheless, the potential
exists for this to be an important and viable technolo-
gy in the setting of emergent intubation, especially if a
high incidence of cardiopulmonary arrest is occurring.
On the other hand, although colorimetric ETCO2 alone
and standard EDD technology alone lead to occasion-
al misidentifications of ETT location, the simplest and
most cost-effective solution may be combined use of
both technologies in concert with clinical judgment.16
These technologies work by different and independ-
ent principles and are complementary.2
failure of some emergency department physicians,
emergency medical services agencies, and their med-
ical directors to admit that esophageal intubation is a
serious problem and aggressively address this situa-
tion.2,17 This resistance appears to be changing as evi-
denced by new Advanced Cardiac Life Support
guidelines mandating use of alternate technologies to
assist in verifying ETT location.14,18 If confirmed in
further studies, use of this new EEDD or the use of
combined ETCO2 and EDD devices in conjunction
with clinical judgment will likely result in near-elimi-
nation of undetected esophageal intubation with rare
instances of removal of properly placed ETTs.
CONCLUSION
This study demonstrates that the EEDD can reliably
identify tracheal intubations in situations where stan-
dard EDDs may fail. However, prior to implementation
of this technology in a clinical decision model, future
studies must determine the reliability of this device for
identification of esophageal intubations (double-blind
trial with both tracheal and esophageal tubes) and the
reliability of this device in less controlled emergency
department and prehospital settings.
Perhaps the largest problem with missed
esophageal intubation is not the technology but the

64
PREHOSPITAL EMERGENCY CARE JANUARY / MARCH 2002 VOLUME 6 / NUMBER 1
The authors acknowledge the assistance and support of Anne
Weiland for help with data analysis and statistical calculations.
J. The esophageal detector device: accuracy and reliability in
difficult airway settings. Prehosp Disaster Med. 1996;11:60-2.
10. Marley CD Jr, Eitel DR, Anderson TE, Murn AJ, Patterson GA.
Evaluation of a prototype esophageal detection device. Acad
Emerg Med. 1995;2:503-7.
References
11. Baraka A, Khoury PJ, Siddik SS, Salem MR, Joseph NJ. Efficacy
of the self-inflating bulb in differentiating esophageal from tra-
cheal intubation in the parturient undergoing cesarean section.
Anesth Analg. 1997;84:533-7.
12. Haynes SR, Morton NS. Use of the oesophageal detector device
in children under one year of age. Anaesthesia. 1990;45:1067-9.
13. Kasper CL, Deem S. The self-inflating bulb to detect esophageal
intubation during emergency airway management.
Anesthesiology. 1998;88:898-902.
1. Caplan RA, Posner KL, Ward RJ, Cheney FW. Adverse respira-
tory events in anesthesia:
Anesthesiology. 1990;72:828-33.
a
closed claims analysis.
2. White S, Slovis C. Inadvertent esophageal intubation in the
field: reliance on a fool’s “gold standard.” Acad Emerg Med.
1997;4:89-91.
3. Katz SH, Falk JL. Misplaced endotracheal tubes by paramedics
in an urban emergency medical services system. Ann Emerg
Med. 2001;37:32-7.
14. Cummins RO, Hazinski MF. Guidelines based on the principle
‘First, do no harm’: new guidelines on tracheal tube confirma-
tion and prevention of dislodgment. Circulation. 2000;102:380-4.
15. Wafai Y, Salem MR, Czinn EA, Barbella J, Baraka A. The self-
inflating bulb in detecting esophageal intubation: effect of bulb
size and technique used [abstract]. Anesthesiology. 1993;79:
A496.
16. Ghabash M, Choueiry P, Matta M. Aspirated air capnography
with esophageal detector device to confirm tracheal intubation
in rapid sequence induction. Middle East J Anesthesiol. 1998;14:
381-6.
17. Ginsburg WH Jr. When does a guideline become a standard?
The new American Society of Anesthesiologists guidelines give
us a clue. Ann Emerg Med. 1993;22:1891-6.
18. American Heart Association. Section 3: Adjuncts for oxygena-
tion, ventilation, and airway control. Circulation. 2000;102
(suppl I):95-104.
4. Bozeman WP, Hexter D, Liang HK, Kelen GD. Esophageal
detector device versus detection of end-tidal carbon dioxide
level in emergency intubation. Ann Emerg Med. 1996;27:595-9.
5. Varon AJ, Morrina J, Civetta JM. Clinical utility of a colorimet-
ric end-tidal CO2 detector in cardiopulmonary resuscitation and
emergency intubation. J Clin Monit. 1991;7:289-93.
6. Lang DJ, Wafai Y, Salem MR, Czinn EA, Halim AA, Baraka A.
Efficacy of the self-inflating bulb in confirming tracheal intuba-
tion in the morbidly obese. Anesthesiology. 1996;85:246-53.
7. Wafai Y, Salem MR, Joseph NJ, Baraka A. The self-inflating bulb
for confirmation of tracheal intubation: incidence and demogra-
phy of false negatives [abstract]. Anesthesiology. 1994;81:
A1304.
8. Czinn EA, Wafai Y, Salem MR. Efficacy of the self-inflating bulb
for confirmation of emergency tracheal intubation in critically ill
patients [abstract]. Anesthesiology. 1995;83:A266.
9. Kapsner CE, Seaberg DC, Stengel C, Ilkhanipour K, Menegazzi