Full text of DOI:10.1007/BF03017329

CARDIOTHORACIC ANESTHESIA, RESPIRATION AND AIRWAY
393
Utility of esophageal Doppler as a minimally inva-
sive hemodynamic monitor: a review
[Revue de l’utilité du Doppler œsophagien comme moniteur hémodynamique
à effraction minimale]
Kevin B. Laupland MD FRCPC,* Colin J. Bands MB CHB FRCPC†
Purpose: The current bedside “gold standard” for cardiac output
(CO) monitoring is thermodilution using a pulmonary artery
catheter (PAC) but there is a number of risks associated with its use.
The primary objective of this review was to evaluate the utility of
esophageal Doppler (ED) as a minimally invasive monitor of CO.
Source : Nous avons procédé à une recherche dans Medline, pour la
période de 1966 à 2001, et à une revue des citations pour les études
comparant l’usage périopératoire du DO et de la thermodilution avec
CAP pour le DC chez des patients gravement malades.
Constatations principales : Vingt-cinq articles traitaient de la
comparaison des mesures du DC avec DO et CAP chez une grande
diversité de patients. Une bonne corrélation globale a été notée entre
le DC déterminé par le DO et la thermodilution (n = 18 études, R
médian = 0,89, étendue de 0,52 à 0,98) et biais minimal (n = 13,
médiane -0,01, étendue de 1,38 à 2 L·min–1). La précision du DO n’a
été que bonne dans l’ensemble comme l’ont montré les limites de la
concordance. La technique du DO a été trouvée efficace, car elle
détecte les modifications du DC noté par thermodilution, et fiable,
puisqu’elle peut démontrer une faible variation intra-observateur et
interobservateur. Les articles révèlent que le DO est facile à insérer, ne
requérant qu’une formation minimale, et sans risque, ne présentant
pas de complications significatives connues.
Source: Medline literature search from 1966 to 2001 with citation
review for studies comparing ED to PAC thermodilution for CO in
perioperative and critically ill patients.
Principal findings: Twenty-five publications were identifed com-
paring ED and PAC measurement of CO in a broad range of
patients. There was a good overall correlation between CO deter-
mined by ED and thermodilution (n = 18 studies, median R =
0.89, range 0.52 to 0.98) and minimal bias (n = 13, median -0.01,
range 1.38 to 2 L·min–1). The precision of ED was only fair overall
as assessed by limits of agreement. The ED technique was found to
be responsive in detecting changes in thermodilution CO and was
reliable demonstrating both low intra- and inter-observer variation.
ED was reportedly easy to insert after minimal training and was
safe, with no significant complications identified.
Conclusion : Le DO est un appareil de mesure périopératoire du DC
pratique, fiable et valide pour des patients gravement malades.
D’autres études auprès d’un plus grand nombre de patients
demeurent nécessaires pour déterminer si la précision limitée qui a été
observée est inhérente à la technique, aux diagnostics des patients
étudiés ou aux échantillonnages de faible effectif.
Conclusion: ED is a practical, reliable, and valid device for mea-
suring CO in perioperative and critically ill patients. Further studies
with larger numbers of patients are needed to determine if the lim-
ited precision observed is inherent to the technique, the diagnoses
of patients studied, or the small sample sizes.
HE use of pulmonary artery catheters
(PAC) for hemodynamic monitoring of
perioperative and critically ill patients has
Objectif : La thermodilution avec un cathéter artériel pulmonaire
(CAP) représente la référence actuelle en monitorage de chevet du
débit cardiaque (DC), mais elle comporte certains risques. L’objectif
principal de la présente revue était d’évaluer l’utilité du Doppler
œsophagien (DO) comme moniteur du DC à effraction minimale.
1
T
become increasingly criticized. Since the
introduction of the flow-directed balloon-tipped PAC
three decades ago by Swan et al. there has been an
explosion in its use.2 Despite concerns regarding its
From the Departments of Medicine,* Critical Care,*† and Anesthesia,† Peter Lougheed Centre and the University of Calgary, Calgary,
Alberta, Canada.
Address correspondence to: Dr. Colin J. Bands, Department of Anesthesia and Critical Care, Peter Lougheed Centre, 3500, 26th Avenue
N.E., Calgary, Alberta T1Y 6J4, Canada. Phone: 403-291-8315; Fax: 403-291-1491; E-mail: colin.bands@calgaryhealthregion.ca
Dr. Laupland was supported by the 2000 Bayer Healthcare/Canadian Institutes of Health Research/Canadian Infectious Diseases Society
Research Fellowship and a clinical fellowship award from the Alberta Heritage Foundation for Medical Research.
Accepted for publication September 27, 2001.
Revision accepted January 16, 2002.
CAN J ANESTH 2002 / 49: 4 / pp 393–401

394
CANADIAN JOURNAL OF ANESTHESIA
evaluation since.9,10 Initially, suprasternal transthoracic
ultrasound/Doppler probes were used for determin-
ing CO but they were not widely adopted because
probe position instability limited their use for repeat-
ed measures over extended periods of time.11
Esophageal probes were recognized to have two sig-
nificant advantages over suprasternal probes. The first
was that the smooth muscle tone of the esophagus is
a natural means of maintaining the probe in position
for repeated measures. The second was that the esoph-
agus is in close anatomical proximity to the aorta so
that signal interference from bone, soft tissue, and
lung is minimized. Over the past three decades ED has
evolved from an experimental technique to a relative-
ly simple bedside procedure with the latest models
incorporating both Doppler and echo-ultrasound in a
single probe.
Determination of CO using ED is based on the fol-
lowing principles.12 A Doppler flow probe is used to
directly measure flow velocity (Vf). The cross section-
al area of the aorta (CSAa) is then determined either
by ultrasonic measurement or estimated using pub-
lished nomograms. Aortic blood flow (ABF) is then
calculated based on the product of Vf and CSAa.
Because the ED probe measures ABF in the descend-
ing aorta, the measured ABF is always less than CO as
a result of blood flow to aortic arch branches.
Although it varies among patients and disease states,
ABF is typically 70% of CO.12 Therefore, ED probes
directly measure ABF but only estimate CO.
There are a number of considerations regarding the
accuracy of ED for estimating CO. First, turbulence
arising from thoracic aortic aneurysms, intra-aortic
balloon pumps, and aortic valve disease may affect the
Vf profile and lead to invalid results. However, this is
a theoretical concern that has not been adequately
studied. Second, the assumption of a constant propor-
tion of blood flow to the descending aorta may not be
valid such as in the setting of aortic coarctation, aortic
cross-clamp, acute bleeding, or pregnancy.13,14 Finally,
the CSAa may vary considerably between patients and
within the same patient in different disease states.
Some ED models estimate CSAa using nomograms
based on patient gender, age, and body surface area or
use a single transthoracic ultrasonic measure at the
start of the study. In the former case, significant error
in the actual aortic diameter may occur as a result of
biological variation and in both cases error may arise
from changing states of hydration, stress, vasoactive
medication use, or from application of an aortic cross-
clamp.11,15,16 In an attempt to minimize this error, an
echo-ED model that contains both a Doppler and
ultrasound (M-mode) probe has been developed.12
safety and lack of empiric evidence for benefit, it has
become a standard of care for managing high-risk
perioperative and critically ill intensive care unit (ICU)
patients in many centres. In the 1990’s a number of
non-randomized cohort and case-control studies asso-
ciating the use of PAC with increased mortality were
published with increasing concern raised about its
widespread use.1,3,4
Sandham and colleagues on behalf of the Canadian
Critical Care Clinical Trials Group recently reported
on a randomized controlled trial of PAC use in 1994
high-risk geriatric surgical patients.5 They found no
difference in the mortality rate (7.8% PAC vs 7.7%
control) or length of hospitalization among patients
treated with the aid of a PAC as compared to those
managed without this device. However, patients ran-
domized to PAC use showed a trend toward less renal
impairment but suffered a significantly higher rate of
procedural complications and pulmonary emboli.
Although the results of this trial must be generalized
cautiously to all high-risk perioperative and critically ill
patients, this trial provides good evidence that PACs
do not inherently increase mortality. However, it is
not clear whether the potential advantages of less
organ dysfunction and improved diagnostic certainty
with PAC outweigh the significantly increased compli-
cations associated with its use.
Non-invasive or minimally invasive hemodynamic
monitoring techniques may provide an alternative to
the use of PAC. However, to be adopted, these tech-
niques, at a minimum, need to demonstrate that they
are safe and provide accurate hemodynamic informa-
tion. A number of non-invasive or minimally invasive
methods have been proposed and include esophageal
Doppler (ED), transesophageal echocardiography,
arterial waveform analysis, thoracic impedance, and
modified Fick techniques.6–8 ED is well suited to the
perioperative or ICU environment because it is a rela-
tively simple technique that does not require special-
ized training or complex equipment. The primary
objective of this report was to perform a semi-struc-
tured, comprehensive review of the utility of ED as a
minimally invasive measure of cardiac output (CO) in
perioperative and critically ill adult patients. Since
many critical care physicians and anesthesiologists may
not be familiar with this technique, a brief review of
theoretical and practical aspects of ED and its use as a
cardiac preload measure is also presented.
ED: theoretical considerations
ED was first proposed as a minimally invasive hemo-
dynamic measure in the 1970’s and it has undergone
significant technological advancement and clinical

Laupland et al.: ESOPHAGEAL DOPPLER MONITORING
395
With the echo-ED, real-time measurements of Vf and
CSAa are made and continuously measured ABF is dis-
played.
patients. Search terms included PAC, Swan Ganz
catheterization, ED, and Doppler echocardiography.
Studies were limited to English language, human, and
age > 12 yr. Abstracts were screened and full-length
articles were retrieved for all relevant studies. Review
articles and bibliographies of identified studies were
searched in an attempt to identify other studies missed
by the Medline search strategy.
ED: practical considerations
Probe placement
The ED probe is placed via the oral or nasal route and
may be left in situ for days to weeks.17 The ideal probe
tip location is at the level between the fifth and sixth
thoracic vertebrae because at that level the aorta is
adjacent and parallel to the esophagus.18 This location
is achieved by superficially landmarking the distance to
the third sternocostal junction anteriorly and is
approximately 30 to 40 cm in the average adult. After
insertion, the ED probe is then rotated on its axis to
achieve an optimal signal prior to taking measure-
ments. Contraindications to ED use include known
esophageal anatomical abnormalities such as malfor-
mations, tumours, strictures or varices, recent
esophageal or upper airway surgery, or acute
esophagitis. Because of the mild discomfort associated
with placing the probe (comparable to insertion of a
nasogastric or orogastric drainage tube) and maintain-
ing it in fixed position, patients monitored with an ED
require adequate sedation.
Measures of agreement
In this review the parameters of Bland and Altman
were chosen in preference to the product-moment
correlation coefficient (R) to assess agreement
between COTD and COED or ABF.24 The product-
moment correlation coefficient is merely a measure of
the degree of linear association between the two
methods and is a poor measure of agreement or the
degree to which the two techniques measure the same
absolute value. In Bland–Altman analysis, the differ-
ence between the two measures (COTD–COED) is
plotted against their mean [½(COTD + COED)]. The
mean difference or “bias” is a measure of how well the
two techniques agree on average.25 A measure of pre-
cision or range of agreement for a given individual is
expressed as the 95% “limits of agreement”. This is
calculated as the bias two standard deviations of the
differences. Practically, if 95% limits of agreement
(LOA) are within clinically acceptable limits the two
techniques may be used interchangeably.24
Safety and ease of use
ED appears to be associated with few complications and
minimal training is required to insert and achieve an
adequate signal in most patients. Although not formal-
ly timed in studies, ED has been reported to typically
take less than five minutes to place and obtain a clear
signal.17,19,20 Two studies have formally evaluated the
role of experience on the quality of results and found
that approximately ten to 12 patients were required for
an operator to develop consistent efficiency and accura-
cy.21,22 No major adverse events related to the use of ED
were reported in any of the literature included in this
review. Furthermore, one author has anecdotally
reported no significant complications with his use of
ED in more than 500 patients.17
Identified studies
A total of 25 full-length articles and abstracts were iden-
tified that fulfilled the search criteria and are shown in
the Table. Nineteen reports evaluated ED13,14,21,22,26–39
and six16,20,40–43 assessed echo-ED COED/ABF vs PAC
COTD respectively. However, one data set was published
twice,42,43 and one abstract that was more than five years
old was excluded from review based on journal policy.44
No studies were identified that directly compared ED
CO/ABF with echo-ED CO/ABF. The identified stud-
ies varied considerably in patient population and in qual-
ity (Table). Twelve studies were from Europe, nine from
the United States, one from Australia, and one from
South Africa. There was a broad range of patients
enrolled including perioperative (cardiac, aortic recon-
structive, and general surgery) and medical and surgical
ICU patients. The studies were generally small with a
median of 20 (range 9 to 60) patients. It is not practical
to report a detailed critical appraisal of each of the study
methodologies in this review. However, despite its
importance to the validity of their results, it is notable
that only a minority of the studies described an adequate
blinding process.13,26,28,29,39 Furthermore, a number of
ED as a valid measure of CO
Search strategy
CO determination by thermodilution using a PAC is
the current bedside “gold standard” method and has
been shown to be comparable to other valid measures
including the Fick and dye dilution techniques.23
Since PAC thermodilution cardiac output (COTD) is
the most widely used standard, a Medline search from
1966 to February 2001 was conducted to identify
published studies comparing COTD and ED CO
(COED) or ABF in perioperative and intensive care

396
CANADIAN JOURNAL OF ANESTHESIA
TABLE Summary of studies for ED CO
Study population
(year)
Modalities compared
Number of patients
(paired measurements)
Results
Limitations
Reference
ICU, Lyon, France
(1985)
Echo esophageal
Doppler (ED) vs
thermodilution
cardiac output
21 (300)
R = 0.98
Not blinded, no
Bland-Altman (BA)
analysis
(42, 43)
(COTD
)
Cardiac surgery,
Boston, USA
1986
ED (Ultracom,
16 (372)
23 (420)
R = 0.92
Not blinded, no
BA analysis, small
study
(32)
Lawrence Medical
Systems, Camarillo
USA) vs COTD
Elective surgery,
Seattle, USA (1987)
ED (Ultracom) vs
COTD
R = 0.67 overall, after
the first ten patients
R = 0.85. Bias (limits
of agreement, LOA)
-0.16 (-0.97, 0.65)
L·min–1
Not blinded, results (22)
statistical analysis
not well described
Cardiovascular
surgery, Stanford,
USA (1988)
ED (Lawrence 3000,
Lawrence Medical
Systems) vs COTD
9 (25)
R = 0.68
Not blinded, small
study, LOA
reported inconsistently
in article
(35)
(36)
Bias -0.1 L·min–1
Surgical or ICU
patients, London
UK (1989)
ED (prototype) vs
COTD
38 (238)
Percent bias (LOA)
for changes in COTD
0.6% (-13.5%, 14.7%)
No R between
COTD and ED or
BA analysis reported,
not blinded
Elective surgery,
Los Angeles, USA
(1989)
ED (Ultracom) vs
COTD
14 (246)
23 (96)
R = 0.75
Not blinded,
small study
(30)
(37)
Bias -1.38 L·min–1
Postcardiac
ED (Accucom I,
Datascope Corp.
Paramus, USA) vs
COTD
R = 0.55
Not blinded,
no BA analysis
surgery ICU,
Zurich, Switzerland
(1990)
Noncardiac surgery,
New Haven, USA
ED (Accucom I and
II) vs COTD
43 (923)
Accucom I: R = 0.72,
bias (LOA) -0.4
Exact values for
BA analysis not
(33)
(-2.9, 2); Accucom II
R = 0.91, bias (LOA)
reported (estimated
from graphs), not
-0.2 (-1.6, 1.2) L·min–1 blinded
Cardiac surgery,
Sydney, Australia
(1991)
ED (Lawrence 3000)
vs COTD
11 (106)
39 (489)
Bias -1.0 LOA (-4.6,
2.6) L·min–1.
Small study, not
blinded
(38)
(13)
Abdominal aortic
surgery, New
Haven, USA
ED (Accucom II)
vs COTD
Baseline R = 0.94, X-
clamp R = 0.72,
Exact values for
BA analysis not
reported (estimated
from graphs)
postclamp R = 0.88.
Approx. Bias (LOA)
baseline -0.3 (-1.7,
1.1), X-clamp 0.7 (-1.9,
3.3), postclamp -0.1
(-1.6, 1.4) L·min–1.
Postop cardiac
surgery patients
Zurich, Switzerland
(1993)
ED (Accucom II) vs
COTD
16 (140)
R = 0.52, bias (LOA)
-0.37 (-3.8, 3.0)
L·min–1.
Not blinded, small
study
(34)

Laupland et al.: ESOPHAGEAL DOPPLER MONITORING
TABLE continued
397
Study population
(year)
Modalities compared
Number of patients
(paired measurements)
Results
Limitations
Reference
Cardiac surgery,
Belfast, Northern
Ireland (1997)
ED (ODM II,
Doptek) vs
continuous COTD
11 Group I
(probe not
readjusted) 5
Group II (probe
readjusted)
Bias (LOA) for
Group I = 0.8 (-2.2,
3.8) and Group II =
0.14 (-0.58, 0.85)
L·min–1
Small study
(29)
ICU patients,
Nimes, France
ED (ODM I) vs
COTD
11 (107) training,
49 (320) evaluation
Training and evaluation Unclear if adequate
(21)
R = 0.53 and
blinding in evaluation
phase
R = 0.89, bias (LOA)
1.2 (-2.0, 4.4) and
0.1 (-2.1, 2.3)
L·min–1 respectively
Medical ICU,
Paris, France
Paris, France) vs COTD
Echo-ED (Dynemo-
3000, Sometec Inc.,
20 (80)
22 (60)
46 (136)
48 (171)
R = 0.80
R = 0.92
Not blinded,
no BA analysis
(41)
(16)
(39)
(20)
Medical ICU,
Nice, France
(1998)
Echo-ED (Dynemo-
3000) vs COTD
BA analysis
not reported
3 ICUs in Paris,
France (1998)
ED (ODMI) vs COTD
R = 0.95, bias (LOA)
0.24 (-1.56, 2.04)
L·min–1
Cardiac or
Echo-ED (Dynemo
3000) vs COTD
R = 0.90
No BA analysis,
not blinded, abstract
publication
abdominal surgery,
Stuttgart, Germany
(1998)
Surgical ICU
Detroit, USA
(1998)
ED (EDM, Deltex
Medical Inc.) vs COTD
10
R = 0.92
No BA, analysis,
not blinded, abstract
publication, small
study
(27)
ICU, Bobigny,
France (1999)
3 Surgical ICUs
in New Orlean’s
USA (1999)
ED (ODM II) vs
continuous COTD
ED (Deltex) vs
COTD
10 (145)
14 (118)
Bias (LOA) -0.01
(-0.97, 0.96) L·min–1
R = 0.77
Small study
(26)
(31)
Not blinded,
no BA analysis,
small study
Off pump cardiac
surgery, Stuttgart,
Germany (1999)
Echo-ED (Dynemo
3000) vs COTD
10 (50)
17
R = 0.89
Not blinded,
abstract publication,
small study
(40)
(14)
Bias (LOA) 0.77
(-0.49, 2.03) L·min–1
Preeclampsia, Cape
Town, South Africa
ED vs COTD
Bias (LOA) 2.0
Not blinded, type
of ED and ABF
correction factor
not specified,
(-1.0, 5.0) L·min–1
small study
Elective cardiac
surgery, Dallas,
USA (2000)
ED (Deltex) vs COTD
vs direct aortic flow
probe
34 (160)
R = 0.77 for ED-CO
vs flow probe
comparable to COTD
Did not report
correlation or
agreement directly
with COTD
(28)
ICU = intensive care unit; ABF = aortic blood flow.

398
CANADIAN JOURNAL OF ANESTHESIA
the studies were limited because they only reported cor-
relation coefficients and did not perform Bland–Altman
analysis of their results.24
Applicability
The proportion of patients for whom a successful
insertion and signal was obtained was 97% among the
558 patients included in this review. This rate, howev-
er, refers to those patients that were enrolled in the
studies and does not necessarily reflect the applicabili-
ty rate of consecutive patients presenting to the oper-
ating room or ICU. Reasons cited for ED failure
included interference from a nasogastric tube, lung
disease and particularly adult respiratory distress syn-
drome requiring high levels of positive end-expiratory
airway pressure, and poor signal quality or inability to
COED vs COTD
In general, the studies showed good correlation
between the ED derived ABF or CO and COTD. In the
18 studies that reported the correlation coefficient, the
median was 0.89 (range 0.52 to 0.98). There was a ten-
dency for a higher correlation to be observed with the
newer generations of ED and echo-ED (Table). Among
the 13 studies that reported a measure of bias, the
median was -0.01 (range -1.38 to 2) L·min–1. It is not
possible to estimate composite LOA without the raw
data from each of the studies. Boulnois and Pechoux12
pooled three studies16,20,41 of echo-ED and found
among 90 (311 matched measurements) patients a cor-
relation coefficient of 0.89, bias of 0.06 L·min–1, and
LOA -2.21 to 2.33 L·min–1.
maintain
a stable signal for unspecified rea-
sons.13,16,20,21,30,33,37
ED for preload assessment
Although CO is the most valuable hemodynamic para-
meter, assessment of ventricular filling is also believed
to be important in management of perioperative and
critically ill patients. Unlike the use of PAC thermod-
ilution for CO determination, there is unfortunately
no bedside gold standard for determining optimal
ventricular filling. In the absence of a good measure,
the pulmonary capillary wedge pressure (PCWP)
using a PAC is commonly used. However, the PCWP
is subject to a number of technical, disease related, and
interpretation problems and its validity has not been
proven in critically ill human beings.45,46 ED wave-
form analysis has been increasingly evaluated as a
method for determining optimum cardiac preload. A
detailed description of ED Vf waveform shape analysis
is beyond the scope of this article, but it has been
reviewed extensively by Singer.17 The key preload
parameter of interest is the flow time, that is the time
required from start of waveform upstroke to return to
baseline. Since the flow time is heart rate dependent it
is typically corrected (FTc) by dividing the flow time
by the square root of the cycle time (normal range
approximately 330 to 360 msec).17 Few studies have
attempted to evaluate the FTc as a preload measure
and are detailed in the following two paragraphs.
Singer and Bennett investigated FTc as a measure of
ventricular filling by placing an ED and a PAC in 43
mechanically ventilated ICU and cardiothoracic surgery
patients in London, UK.47 After baseline measures, the
patient’s ventricular filling was manipulated either by
increasing it with iv fluid loading if hypovolemic
(PCWP < 8 mmHg) or by decreasing it with iv nitrates
if either normovolemic (PCWP 8–20 mmHg) and
hypertensive or hypervolemic (PCWP > 20 mmHg).
Patients who had decreased preload from intraoperative
hemorrhage were also followed. They observed a
matched increase in PCWP and FTc with fluid loading
Responsiveness
Five studies specifically quantified ED’s ability to detect
changes in COTD. Studies typically defined a significant
change in COTD as at least a 12–15% difference from
baseline as per Stetz et al.23 Valtier et al. found good
agreement between changes in COTD and COED R =
0.9, bias 0 L·min–1, LOA -1.7 to 1.7 L·min–1).39 Cariou
and colleagues calculated a sensitivity of 75% and speci-
ficity of 95% for echo-ED to detect a 13% change in
COTD in medical ICU patients.41 In patients undergo-
ing aortic reconstructive surgery, Perrino found that
ED detected changes of 15% COTD with a sensitivity of
89% without and 79% during cross-clamping of the
aorta.13 Two studies calculated percent bias (LOA) for
changes in COTD of 0.6% (-13.5%, 14.7%) and -0.66%
(-20.1%, 18.8%).34,36 Penney et al. found that COED
correctly tracked COTD changes in 13 of 16 women
with pre-eclampsia but they did not quantify the mag-
nitude.14 Kumar and colleagues reported similar results
in 11 of 14 anesthetized surgical patients.30
Reliability
Six studies assessed the reproducibility of ED both
between (inter-observer) and within (intra-observer)
operators.16,30,32,36,39,43 Inter-observer variability was
assessed with echo-ED in two studies with 11 and 22
patients that found differences with repeated measures in
stable patients to be 0.75% and 3.3% respectively
between two operators blinded to each other’s find-
ings.16,43 Intra-observer variability as assessed by calculat-
ing the coefficient of variation in four studies was 8%,
5.7%, 3.8%, and 3.3% for ED as compared to 12%, 8.3%,
6.2%, and 4.6% for thermodilution respectively.16,32,36,39

Laupland et al.: ESOPHAGEAL DOPPLER MONITORING
399
among all 18 patients with hypovolemia. Similarly, all
normovolemic patients had a concordant decrease in
PCWP and FTc. However, while the PCWP decreased
in all 12 hypervolemic patients, the FTc increased ini-
tially in 11/12 patients (reflecting optimal CO) prior to
decreasing. Thus, FTc appears to be a useful measure to
direct optimal ventricular loading.
In the other study by Mythen et al., a prospective,
randomized, open study was conducted with 60 elec-
tive cardiac surgery patients in London, UK.48 The
control patients received usual care and the treatment
patients were given repeated boluses of 6% hydrox-
yethyl starch to maximize stroke volume based on ED
and central venous pressure. Compared to control, the
treatment group had a lower incidence of gut hypop-
erfusion as measured by gastric intramucosal pH <
7.32 (7% vs 56%, P < 0.001), less major complications
(zero vs six patients, P = 0.01), and shorter mean
lengths of ICU (one vs 1.7 days, P = 0.023) and over-
all hospital stay (6.4 vs 10.1 days, P = 0.011). The
main limitation of this study is that it was not blinded
and knowledge of the patients’ treatment allocation
may have influenced either management or ascertain-
ment of outcome. Furthermore, as with the Sinclair
study, the authors did not adequately describe a mech-
anism by which protocol patients would be able to be
discharged earlier. Despite the limitations of these two
studies, they are important contributions to the ED
literature because they showed improvements in clini-
cal outcomes with ED usage.
Two other studies have been undertaken to try to
assess FTc as a measure of ventricular filling. DiCorte et
al. in a blinded fashion compared pulmonary artery
diastolic (PAd) pressure and FTc by ED to end-diastolic
short-axis area (EDA) as measured by transesophageal
echocardiography in 20 patients undergoing coronary
bypass surgery.28 They found a modest correlation
between FTc and EDA (R = 0.49) and a poor correla-
tion with PAd pressure and EDA (R = 0.10). Although
this provides some evidence that FTc may be a better
measure of preload than PAd it is not a clinically relevant
comparison because PCWP is virtually always used in
preference to PAd. Madan and colleagues conducted a
study among 14 surgical ICU patients comparing ED
with PAC.31 They found a better correlation between
FTc (R = 0.52) and COTD than with PCWP (R = 0.2)
and COTD. Further study is needed to define the role of
FTc based optimization of ventricular filling.
Summary and conclusions
ED is a minimally invasive technique that has a good
safety profile, is easy to use, and requires only minimal
training. When compared to the current bedside gold
standard of bolus COTD using a PAC in a broad range
of perioperative and critically ill patients, ED was found
to be reliable, responsive to changes, and have good
agreement with low bias in the majority of studies.
However, wide limits of agreement were observed with
ED in several studies raising concerns about the preci-
sion of this technique. Further clinical investigation
with larger numbers of varied patients is required to
determine if the limited precision observed is inherent
to the technique, related to patients’ diagnoses, or is
simply a function of the small study sample sizes used.
It is unlikely that ED will replace PAC at present
because of the concerns regarding limited precision and
reduced applicability. However, use of ED may lead to
a reduction of the number of patients undergoing inva-
sive monitoring with its associated risks. Future clinical
investigations will hopefully further define its role in
optimization of ventricular filling and in improving the
outcome of perioperative and critically ill patients.
ED driven protocols
Improved patient outcome?
In assessing any new technology or treatment an
important criteria for its adoption is whether it can be
shown to decrease morbidity and/or mortality. Two
studies have shown improved outcomes with the use
of ED perioperatively.48,49 Sinclair et al. conducted a
randomized controlled trial with 40 geriatric patients
undergoing fractured hip repair under general anes-
thesia in London, UK.49 All patients had an ED insert-
ed but were randomized to either an intraoperative
volume optimization protocol (goal FTc > 350 msec
and optimized stroke volume) vs usual care.
Anesthesiologists were blinded to ED values but not
to the iv volume administered. Protocol patients
received significantly more fluid, had higher mean FTc
and CO values, but had similar blood pressure and
heart rates compared to controls intraoperatively.
Protocol patients had significant reductions in the
main outcome measures of median time to be
declared fit for discharge of ten vs 15 days and length
of hospitalization of 12 vs 20 days as compared to con-
trol respectively (P < 0.05 for both). An important
potential limitation of this study is that protocol
patients may have been recognized (unblinded) based
on the increased amount of fluids given and possibly
discharged earlier.
References
1 Pulmonary Artery Catheter Consensus Conference
Participants. Pulmonary artery catheter consensus con-
ference: consensus statement. Crit Care Med 1997; 25:
910–24.

400
2 Swan HJC, Ganz W, Forrester J, Marcus H, Diamond
CANADIAN JOURNAL OF ANESTHESIA
noninvasive measurement of aortic blood flow in criti-
cally ill patients with a new esophageal echo-Doppler
system. J Crit Care 1998; 13: 177–83.
G, Chonette D. Catheterization of the heart in man
with use of a flow-directed balloon-tipped catheter. N
Engl J Med 1970; 283: 447–51.
17 Singer M. Esophageal Doppler monitoring of aortic
blood flow: beat-by-beat cardiac output monitoring.
Int Anesth Clin 1993: 99–125.
18 Histand MB, Wells MK, Reeves JT, Sodal IE, Adamson
HP, Wilson JT. Ultrasonic pulsed Doppler trans-
esophageal measurement of aortic haemodynamics in
humans. Ultrasonics 1979; 17: 215–8.
19 Rodriguez RM, Berumen KA. Cardiac output measure-
ment with an esophageal Doppler in critically ill emer-
gency department patients. J Emerg Med 2000; 18:
159–64.
20 Klein G, Emmerich M, Maisch O, Dummler R. Clinical
evaluation of non-invasive monitoring aortic blood
flow (ABF) by a transesophageal echo-Doppler device.
Anesthesiology 1998; 89: A446 (abstract).
3 Connors AF Jr, Speroff T, Dawson NV, et al. The effec-
tiveness of right heart catheterization in the initial care
of critically ill patients. JAMA 1996; 276: 889–97.
4 Zion MM, Balkin J, Rosenmann D, et al. Use of pul-
monary artery catheters in patients with acute myocar-
dial infarction. Analysis of experience in 5,841 patients
in the SPRINT registry. Chest 1990; 98: 1331–5.
5 Sandham JD, Hull RD, Brant R, and the Canadian
Critical Care Clinical Trials Group Rationale. A ran-
domized controlled trial of pulmonary artery catheter
use in 1994 high-risk geriatric surgical patients. Am J
Respir Crit Care Med 2001; 163: A16 (abstract).
6 Singer M. Cardiac output in 1998. Heart 1998; 79:
425–8.
21 Lefrant J-Y, Bruelle P, Aya AGM, et al. Training is
required to improve the reliability of esophageal
Doppler to measure cardiac output in critically ill
patients. Intensive Care Med 1998; 24: 347–52.
22 Freund PR. Transesophageal Doppler scanning versus
thermodilution during general anesthesia. An initial
comparison of cardiac output techniques. Am J Surg
1987; 153: 490–4.
23 Stetz CW, Miller RG, Kelly GE, Raffin TA. Reliability
of the thermodilution method in the determination of
cardiac output in clinical practice. Am Rev Respir Dis
1982; 126: 1001–4.
24 Bland JM, Altman DG. Statistical methods for assess-
ing agreement between two methods of clinical mea-
surement. Lancet 1986: 307–10.
25 Altman D. Practical Statistics for Medical Research.
London, UK: Chapman & Hall, 1991.
7 Newman DG, Callister R. The non-invasive assessment
of stroke volume and cardiac output by impedance car-
diography: a review. Aviat Space Environ Med 1999;
70: 780–9.
8 Troianos CA, Porembka DT. Assessment of left ventric-
ular function and hemodynamics with transesophageal
echocardiography. Crit Care Clin 1996; 12: 253–72.
9 Side C, Gosling R. Non-surgical assessment of cardiac
function. Nature 1971; 232: 335–6.
10 Daigle RE, Miller CW, Histand MB, McLeod FD,
Hokanson DE. Nontraumatic aortic blood flow sensing
by use of using an ultrasonic esophageal probe. J Appl
Physiol 1975; 38: 1153–60.
11 Huntsman LL, Stewart DK, Barnes SR, Franklin SB,
Colocousis JS, Hessel EA. Noninvasive Doppler determi-
nation of cardiac output in man. Clinical validation.
Circulation 1983; 67: 593–602.
26 Baillard C, Cohen Y, Fosse JP, Karoubi P, Hoang P,
Cupa M. Haemodynamic measurements (continuous
cardiac output and systemic vascular resistance) in criti-
cally ill patients: transesophageal Doppler versus con-
tinuous thermodilution. Anaesth Intensive Care 1999;
27: 33–7.
27 Cuschieri J, Rivers E, Caruso J, et al. A comparison of
transesophageal Doppler, thermodilution and Fick car-
diac output measurements in critically ill patients. Crit
Care Med 1998; 26(suppl): A62 (abstract).
28 DiCorte CJ, Latham P, Greilich PE, Cooley MV,
Grayburn PA, Jessen ME. Esophageal Doppler monitor
determinations of cardiac output and preload during
cardiac operations. Ann Thorac Surg 2000; 69:
1782–6.
29 Krishnamurthy B, McMurray TJ, McClean E. The peri-
operative use of the oesophageal Doppler monitor in
patients undergoing coronary artery revascularization.
12 Boulnois JG, Pechoux T. Non-invasive cardiac output
monitoring by aortic blood flow measurement with the
Dynemo 3000. J Clin Monit 2000; 16: 127–40.
13 Perrino AC Jr, Fleming J, LaMantia KR.
Transesophageal Doppler cardiac output monitoring:
performance during aortic reconstructive surgery.
Anesth Analg 1991; 73: 705–10.
14 Penny JA, Anthony J, Shennan AH, de Swiet M, Singer
M. A comparison of hemodynamic data derived by pul-
monary artery flotation catheter and the esophageal
Doppler monitor in preeclampsia. Am J Obstet
Gynecol 2000; 183: 658–61.
15 Lavandier B, Muchada R, Cathignol D. The impor-
tance of measuring the diameter of the aorta in perop-
erative blood flow monitoring. V International
Symposium on Biomedical Engineering 1994,
Santiago: Spain, 1994: 258–9 (abstract).
16 Bernardin G, Tiger F, Fouché R, Mattéi M. Continuous

Laupland et al.: ESOPHAGEAL DOPPLER MONITORING
401
43 Lavandier B, Cathignol D, Muchada R, Xuan BB,
Anaesthesia 1997; 52: 624–9.
Motin J. Noninvasive aortic blood flow measurement
using an intraesophageal probe. Ultrasound Med Biol
1985; 11: 451–60.
30 Kumar A, Minagoe S, Thangathurai D, et al.
Noninvasive measurement of cardiac output during
surgery using a new continuous-wave Doppler
esophageal probe. Am J Cardiol 1989; 64: 793–8.
31 Madan AK, UyBarreta VV, Aliabadi-Wahle S, et al.
Esophageal Doppler ultrasound monitor versus pul-
monary artery catheter in the hemodynamic manage-
ment of critically ill surgical patients. J Trauma 1999;
46: 607–11.
32 Mark JB, Steinbrook RA, Gugino LD, et al. Continuous
noninvasive monitoring of cardiac output with
esophageal Doppler ultrasound during cardiac surgery.
Anesth Analg 1986; 65: 1013–20.
44 Lefrant JY, Aya G, de la Coussaye JE, et al. Comparison
of cardiac output measured by esophageal Doppler vs
thermodilution. Intensive Care Med 1992; 18(suppl
2): S177 (abstract).
45 Raper R, Sibbald WJ. Misled by the wedge? The Swan-
Ganz catheter and left ventricular preload. Chest 1986;
89: 427–34.
46 Calvin JE, Driedger AA, Sibbald WJ. Does the pul-
monary capillary wedge pressure predict left ventricular
preload in critically ill patients? Crit Care Med 1981; 9:
437–43.
47 Singer M, Bennett ED. Noninvasive optimization of left
ventricular filling using esophageal Doppler. Crit Care
Med 1991; 19: 1132–7.
33 Perrino AC Jr, Fleming J, LaMantia KR.
Transesophageal Doppler ultrasonography: evidence for
improved cardiac output monitoring. Anesth Analg
1990; 71: 651–7.
48 Mythen MG, Webb AR. Perioperative plasma volume
expansion reduces the incidence of gut mucosal hypop-
erfusion during cardiac surgery. Arch Surg 1995; 130:
423–9.
49 Sinclair S, James S, Singer M. Intraoperative intravascu-
lar volume optimisation and length of hospital stay
after repair of proximal femoral fracture: randomised
controlled trial. BMJ 1997; 315: 909–12.
34 Schmid ER, Spahn DR, Tornic M. Reliability of a new
generation transesophageal Doppler device for cardiac
output monitoring. Anesth Analg 1993; 77: 971–9.
35 Siegel LC, Shafer SL, Martinez GM, Ream AK, Scott
JC. Simultaneous measurements of cardiac output by
thermodilution, esophageal Doppler, and electrical
impedance in anesthetized patients. J Cardiothorac
Anesth 1988; 2: 590–5.
36 Singer M, Clarke J, Bennett ED. Continuous hemody-
namic monitoring by esophageal Doppler. Crit Care
Med 1989; 17: 447–52.
37 Spahn DR, Schmid ER, Tornic M, et al. Noninvasive
versus invasive assessment of cardiac output after car-
diac surgery: clinical validation. J Cardiothorac Vasc
Anesth 1990; 4: 46–59.
38 Stein MS, Barratt SMcG, Purcell GJ. Intraoperative
assessment of the Lawrence 3000 Doppler cardiac out-
put monitor. Anaesth Intensive Care 1991; 19: 251–5.
39 Valtier B, Cholley BP, Belot J-P, de la Coussaye J-E,
Mateo J, Payen DM. Noninvasive monitoring of cardiac
output in critically ill patients using transesophageal
Doppler. Am J Respir Crit Care Med 1998; 158:
77–83.
40 Wagner G, Kalckhoff A, Emmerich M, Westhof FB,
Klein G. Noninvasive monitoring cardiac output (CO)
during off-pump coronary artery bypass grafting (OP-
CABG). New York Post Graduate Assembly. New
York: 1999: 115 (abstract).
41 Cariou A, Monchi M, Joly L-M, et al. Noninvasive car-
diac output monitoring by aortic blood flow determi-
nation: evaluation of the Sometec Dynemo-3000
system. Crit Care Med 1998; 26: 2066–72.
42 Muchada R, Cathignol D, Lavandier B, Lamazou J,
Haro D. Aortic blood flow measurement. Am J
Noninvas Cardiol 1988; 2: 24–31.