H Mächler, A Lueger, S Huber, P Bergmann, P Rehak, G Stark
aneurysm, aorta, bypass surgery, cardiac, cardio-pulmonary, cardiopulmonary bypass, cardiothoracic, carotid, chest, heart, heart-lung machine, medicine, surgery, thoracic, valve, vascular, vessel
H Mächler, A Lueger, S Huber, P Bergmann, P Rehak, G Stark. The Esophageal-ECG: New applications with a new technique. The Internet Journal of Thoracic and Cardiovascular Surgery. 1998 Volume 2 Number 2.
Introduction: Because of suboptimized techniques, concerning leads and amplification
methods, the esophageal ECG was thus only used to detect atrial arrhythmias. The esophageal ECG
would be suitable for detecting intraoperative myocardial ischemia.
Methods: The esophageal ECG signals are detected with a bipolar esophageal probe. The
signals are passed to a new high-resolution preamplifier (frequency range 0.01-2000 Hz), further on
to an analog-digital-board and finally are visualized on a PC for definite evaluation.
Results: Compared with the surface ECG, the esophageal ECG could detect significantly more
ischemic episodes in both, an animal comparison study (93% versus 47%) (n = 18), as well as in
coronary artery bypass graft (= CABG) surgery. At the beginning of anesthesia 85 % of the high
risk CABG patients showed ischemic episodes with the esophageal ECG (18% detection rate with
the surface ECG), which were correlated with an adverse postoperative outcome (p<0.05). In
addition to that, 8 of 18 CABG patients had a mean of 483 ± 119 high-amplitude, biphasic atrial
components during cardioplegia (on the surface ECG, only one patient showed 26 P-potentials).
Such atrial activities, visually not detectable, were correlated with postoperative supraventricular
arrhythmias in 88 %.
Conclusions: The clinical relevance is that the esophageal ECG represents a convenient
technique with high sensitivity, to monitor intraoperative myocardial ischemia and to detect atrial
activity during cardioplegia.
The concept of esophageal electrocardiography (E-ECG) is not new. Data of using an E-ECG go back to 1906 when Cremer passed an electrode into the esophagus of a professional sword swallower 1. Since then, numerous studies have confirmed the usefulness of this technique. The unipolar E-ECG was first only used for detecting T-waves as well as premature systolic beats 2, 3. In the 70-ies, probes with “pseudo- bipolar” electrodes were used for the first time to analyze atrial arrhythmias 4, 5, 6. For long term registration, an electrode capsule embedded in gelatine 6 was developed that was swallowed by the patients.
Until now, “pseudo-bipolar” esophageal electrodes were connected to conventional surface ECG (S-ECG) amplifiers with their specific filter technique 7.
Because of the lower electrical impedance, the better electrical conductivity, and the close
anatomic range between the heart and the electrode, the signals of the esophageal ECG can be
detected unfiltered and free of artifacts. Therefore, the signals of our new developed E-ECG were
recorded via a custom-made “bipolar” E-ECG probe containing three
circumferential, chloridized silver electrodes (diameter 7 mm each) at a distance of 2,5 cm from
each other. The probe was positioned where the electrodes detected the maximum ventricular
amplitude 30 – 40 cm aborally. The signals were amplified with a self-designed, unfiltered,
battery-supported high-resolution preamplifier (frequency range 0.01-2000 Hz, IEC-601 norm) that
was developed in cooperation with the Institute of Medical Physics Graz and Fa. Paar®, Graz.
Further, the signals were passed on to an analog-digital-board (Axotape®, Axon Instruments, Inc.,
USA) and were finally visualized on a PC for definite evaluation.
The esophageal, bipolar electrocardiogram
The signals obtained from the esophageal ECG showed a low-amplitude, atrial wave form
(LAA) with diastolic atrial activity preceding the P wave in the surface ECG and a steep
(prominent) high-amplitude biphasic wave (P component), which occurred simultaneously with the
P wave in the surface ECG (Fig. 1). Each of the low-amplitude, atrial activities were followed by
the biphasic P component during normal sinus rhythm and were never present in the surface ECG.
The mean amplitude of the early atrial activities was 0,15 ± 0,014 mV (0,12 – 0,25 mV) and
the mean duration was 0,057 ± 0,002 sec. (0,049 – 0,064 sec.). The mean amplitude of the
biphasic P components in the esophageal ECG was 2,1 ± 0,095 mV (1,8 – 2,55 mV) and the
mean duration of the P components was 0,075 ± 0,007 sec. (0,072 – 0,08 sec.). The
amplitude of the P component at the position of maximum deflection compared with the amplitude
of the QRS complex was 1:2. The mean amplitude of the QRS complex in the esophageal ECG was
3,9 ± 3,5 mV (1,9 – 5,1 mV), the mean duration of the QRS complex was 0,051 ± 0,028
sec. (0,028 – 0,089 sec.). The mean amplitude of the T-wave was 1,4 ± 1,7 mV (0,4
– 4,3 mV), the mean duration of the T-wave was 0,120 ± 0,032 sec. (0,076 – 0,160
1. Detection of myocardial ischemia with the esophageal ECG in an animal comparison study
1.1 Introduction The main goals of this study were to evaluate the association between ST segment alternans and myocardial ischemia and /or infarction and to standardize criteria for ischemic changes in the esophageal ECG. Therefore, 18 sheep were anesthetized and monitored. The proximal diagonal branch of the left anterior descending (LAD) coronary artery was ligated to produce myocardial ischemia iatrogenically. 8
1.2 Results The non-ischemic esophageal ECG presented homogenous ST-segments without
any beat-to-beat alternans. Two minutes after the occlusion, inhomogenous ST-segment changes
were detected in 93% (p<0,01) of the cases with the esophageal ECG, whereas the surface ECG
only met the criteria for ischemia in 47%. To calculate the dynamic changes of the ST-segments,
the absolute difference (in millivolts) in the amplitudes of the ST-segments between a mean of five
successive beats and the next beat performed for 200 consecutive beats was determined (Hodges-
Lehmann Point Estimator). The central tendency of the sum of these values was 2000 mV/ms
(p< 0.05) before occlusion and 5000 mV/ms after occlusion (confidence interval 1700/2500
versus 3350/9250). Thus, the connection between the magnitude of ST-segment alternans and
myocardial ischemia could be demonstrated.
2. Detection of intraoperative myocardial ischemia with the esophageal ECG compared with the surface ECG during coronary artery bypass surgery
2.1 Introduction Perioperative myocardial ischemia is associated with adverse postoperative outcome 9,10,11,12. The classic clinical symptoms of acute myocardial ischemia cannot be clearly detected intraoperatively. Hemodynamic and serum enzyme parameters are not specific and not sensitive during open heart surgery. Endo- and epicardial leads 13, 14, as well as signal-averaging methods take great technical efforts and are not applicable in clinical routine.14, 15. The transesophageal echocardiogram would be sensitive intraoperatively, however, too much personnel would be needed for continous monitoring.
Thus, the surface ECG was the best method to detect intraoperative myocardial ischemia in
clinical routine. Sensitivity using a single lead was greatest in V5 (75%) and V4 (61%,) 9.
Combining different leads (V3 /V4 /V5 or II /V4 /V5) increased sensitivity up to 96% 9, 10,
however, the chest leads, especially the anterolateral leads V1, V2, and V3 cannot be used during
open-heart surgery due to sterility reasons.
Therefore, new methods are needed for detecting intraoperative myocardial ischemia with
sufficiently high sensitivity during cardiac surgical procedures. Because of the close anatomic
range between the esophagus and the heart, the esophagus would be suitable for
electrocardiographic lead detection.
2.2 Methods From induction of anesthesia until the end of surgery of 18 “high-
risk” CABG-patients (instable angina pectoris Braunwald class III) ECG signals were
detected with a conventional surface ECG (I, II, III, aVR, aVL, aVF, V4, V5, V6, V2R, V3R, V4R)
and recorded using a 12-channel ECG writer (HP Page Writer Xli®, USA, Frequenz Rate: 0.05-150
Hz). The signals were compared visually with the simultaneously recorded esophageal ECG signals
and evaluated statistically. The criteria from London 9 were used to define ischemic changes in the
2.3 Results To present time-dependent alterations of the ST-segments in the esophageal ECG,
30 consecutive ST-segments were visualized by means of a waterfall diagram.
1. Uniform ST-segments in the surface ECG and homogenous ST-segments in the esophageal
ECG (Fig. 2).
2. Uniform ST-segments in the surface ECG, but ischemic episodes with inhomogenous ST-
segments in the esophageal ECG (Fig. 3).
Morphologically, the esophageal ECG displayed dynamic beat-to-beat alterations in the ST-
segment amplitudes with “notched”, “bifid”,
“dimpled”, or “cloven” ST-segments. Humps, bulges, or
protuberances developed before or just beyond the apex or on the descending limb of the T-wave.
Frequently, the alterations (= inhomogeneities) were concentrated on the ascending limb of the T-
wave. As the ischemia proceeded, the esophageal ECG presented second or third, occasionally
biphasic T-waves in 28% of the cases.
3. Episodes with ischemic ST-segments in the surface ECG and simultaneously ischemic
episodes with inhomogenous ST-segments in the esophageal ECG (Fig. 4).
The signals of the esophageal ECG showed a significant increase in ST-segment alterations
during the ischemic episodes. The surface ECG showed ST-segment elevations or depressions
demonstrating acute myocardial ischemia.
From induction of anesthesia until declamping of the aorta 72% of the “high-
risk” patients had ischemic episodes (duration > 30 sec) with inhomogenous ST-
segments in the esophageal ECG (vs. 11% in the surface ECG, p<0,05). Two patients had
ischemic episodes in both the esophageal ECG and the surface ECG. The clinical relevance of a
pathological esophageal ECG from induction of anesthesia until the beginning of cardiopulmonary
bypass was that multiple ischemic episodes in the esophageal ECG were correlated with adverse
postoperative outcome in 85% (high doses of inotropes, ventricular arrhythmias LOWN class IV b,
implantation of intraaortic balloon pump, death). After cardiopulmonary bypass, ischemic episodes
in the esophageal ECG showed no correlation to the postoperative outcome.
2.4 Discussion In the last few years, monitoring detection of intraoperative myocardial
ischemia received increasingly more attention 9,10,11,12.
In addition, the electromagnetic current dipole theory 17 predicts that reducing the distance between the recording electrodes and the cardiac signals will result in a linear improvement in detection of dynamic changes in the ECG wave form. Consequently, the surface ECG leads, in further distance from the heart are less effective in detecting ST-segment changes than the esophageal ECG leads in close range to the ischemic area. In the literature, the advantages of an esophageal ECG for detection of ischemia are nearly neglected. Thus, only two case reports exist where myocardial ischemia was detected with an E-ECG intraoperatively 18, 19.
The concept of measurement of ST-segment alterations in the esophageal ECG is based on the idea that morphological beat-to-beat changes in the ST-segment reflect increased inhomogeneities in myocardial repolarization. Alternating T-waves are an indicator for the vulnerability of the myocardium in correlation with ischemia or infarction 20, 21, 22. The reason why the pathological E-ECG effects the postoperative outcome only in the phase before the cardiopulmonary bypass can be explained with the idea that high-risk patients are endangered by stress and myocardial ischemia especially at the induction of anesthesia or shortly after that. Also, patients might suffer an ischemia just before surgery that could not be detected.
Limitations of the study exist due to the impossibility to define the time period between the
beginning of ischemia and the appearance of inhomogenous ST-segment intervals. Additionally, it
is impossible to localize the the ischemic area. Further studies should evaluate the correlation
between the magnitude of the inhomogeneities of the ST-segments in the esophageal ECG and the
size of the ischemic area.
The clinical relevance of the present findings is that the esophageal ECG provides a good
indicator for detection of myocardial ischemia. The results prove that the introduced semi-invasive,
high-resolution esophageal ECG represents a more sensitive method to detect intraoperative
myocardial ischemia than the 12 channel surface ECG. The advantage of this method is that
ischemic episodes can be visualized online beat-to-beat. By detecting high-risk patients shortly
after the beginning of anesthesia the perioperative management could be optimized individually
(use of blood cardioplegia, short periods of aortic cross clamping, off-pump surgery).
3. Monitoring of atrial activities in the esophageal-ECG in the potassium-arrested heart
3.1 Introduction Atrial electrical activities during hypothermic potassium-induced cardioplegia are correlated with the occurrence of postoperative supraventricular dysrhythmias 23.
The surface ECG is insufficient for detection, invasive methods are impractical in clinical
3.2 Results During cardioplegic heart arrest eight of 18 patients had a mean of 483 ± 119
biphasic atrial components (mean amplitude: 0.7 ± 0.1 mV) per patient. A mean of 36 ± 6
potentials /min were detected. In addition, six of the 8 patients had a mean of 29 ± 11 low-
amplitude atrial activities per patient, obviously signals of pacing cells in the sinus node. In the
surface ECG only one patient showed 26 P-waves. 88% of the patients with atrial activities
(vs. 30% of the patients without atrial activities) during cardioplegic heart arrest developed
supraventricular tachycardia in the first five postoperative days (p<0.05).
3.3 Discussion Atrial activity during cardioplegic arrest is mainly caused by poor atrial
protection of the cardioplegic solution. Fibers of the sinoatrial node area are more resistant to K+
than atrial myocardium. Additionally, conducting fibers are more resistant to hypothermia than
muscle fibers. With the conventional cardioplegia (potassium concentration of 20 mmol/l) the cells
of the sinus node may stay active until they exhaust their energy-rich substrates. At the end of the
induced arrest the atrial potentials recur as the cardioplegic solution is washed out and hypothermia
is abolished. The protective and paralytic effect of cardioplegia and hypothermia thus is lost; the
electrical activity is an indicator for inadequate atrial protection. Ischemic damage to the atrial cells
and the supraventricular conduction system results in a prolonged recovery of sinus rhythm and
evidently to the development of postoperative supraventricular tachyarrhythmias. The cause may be
an increased automaticity of atrial cells due to, for example, delayed depolarization as a result of
the ischemic damage.