Mark S. Hynes, MD
Department of Anaesthesia
Ottawa Heart Institute / Civic Hospital and University of Ottawa
Objectives:
Transesophageal echocardiography is a comparative newcomer to the list of monitors used in anaesthesia and intensive care. Although cardiac ultrasound was pioneered in the mid 1950's it was only with the development of newer less bulky probes which permitted two dimensional imaging of the heart in the late 1980's that the use of echocardiography became really feasible in the OR and ICU. The explosion of literature in anaesthesia journals since then atests to its widespread use. It has been advocated as an indispensible monitor in managing critically ill surgical patients, but what can it realistically offer?
TEE can be used to monitor:
It has long been recognized that myocardial ischemia produces regional wall motion abnormalities (RWMA) and that these changes in regional myocardial function occur shortly after the onset of ischemia and long before any surface ECG changes. The first systematic study using TEE to detect these RWMA and compare TEE detected ischemic events to surface ECG evidence of ischemia in patients who either had documented coronary artery disease (CAD) or were felt to be at high risk for CAD and who were undergoing anaesthesia and surgery was by Smith et al1. They showed that RWMAs were four times more frequent than ECG changes suggestive of ischemia and that one third of these RWMAs persisted until the end of surgery (when TEE and ECG monitoring stopped). Of the four perioperative myocardial infractions (MIs), three occurred in patients with persistent RWMAs, and one in a patient with a transient RWMA. Intraoperative ECG changes were not predictive of perioperative MI. All patients who did have ischemic ECG changes also had new RWMAs. Several other studies in patients undergoing coronary artery bypass graft (CABG) surgery followed and are tabulated below.
Table 1 TEE Ischemia in Cardiac Surgery
Reference |
Study Population |
Monitoring and Sampling Protocol |
Post CPB RWMA |
ECG Changes |
Outcome |
|
|
Selected and random sampling |
|
|
6 adverse outcomes all in patients with post bypass RWMA ECG changes not predictive |
|
|
|
|
|
19 had post bypass RWMA |
|
|
|
|
|
3 had post bypass RWMA |
The results of these studies substantiated the findings of Smith et al.:
What about noncardiac surgery?
The studies involving noncardiac surgery are listed in Table 2. Although the early work suggested the same kinds of results in noncardiac surgery there were considerable methodological flaws. Investigators were often not blinded to the TEE data and frequently the results of ECG monitoring were not reported. London et al8 adressed these shortcomings and described the natural history of intraoperative RWMA in patients who had or were at a high risk for having CAD undergoing noncardiac surgery. They found a very poor concordance between new RWMAs and ECG evidence of ischemia. Neither was associated with the other, and neither predicted adverse ischemic cardiac outcomes. A later study from the same institution11 confirmed this lack of agreement between TEE and ECG evidence of ischemia and showed that TEE added little incremental value to preop clinical data and intraoperative two channel ECG monitoring in identifying patients at high risk for perioperative ischemic outcomes.
Table 2 TEE as a Monitor of Myocardial Ischemia in
Noncardiac Surgery
Reference |
Study Population |
Protocol & Sampling |
New RWMA |
Persistent |
Hypokinetic |
ECG Changes |
Outcome |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 MIs (8%) all in pts with persistent RWMAs |
|
|
|
|
|
|
|
2 MIs (4%) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
So what conclusions are we left with? There is no doubt that TEE permits us to detect RWMA in a significant number of high risk patients undergoing noncardiac surgery. How-ever, the more rigorous studies suggest that these RWMA are of little prognostic value. Why?
Several questions and possible answers come to mind
1) Are all RWMA caused by ischemia? No
There are nonischemic causes of RWMA. Conduction abnormalities, acute changes in loading conditions, changes in adrenergic tone and anaesthetic agents can all cause real or apparent RWMAs. Myocardium which has been chronically ischemic ("hibernating") or is recovering from an ischemic insult ("stunned") can show impaired systolic function even though the ischemic event is over and myocardial blood flow has returned to normal. A huge range (0-150%) in segmental wall thickening and inhomogeneity in ventricular contraction has been demonstrated by transthoracic echo in young healthy males without any clinical evidence of ischemic heart disease. Obviously the likelihood of an ischemic cause of a RWMA depends on the incidence of ischemic heart disease in the population being studied.
2) If ischemia is the likely cause of the RWMA Are all RWMA of equal significance? Probably not.
Animal and human studies suggest that hypokinesis is not a specific enough marker for transmural myocardial infarction and may be too sensitive to reliably indicate significant myocardial ischemia. Akinesis or dyskinesis are much more likely to signify a substantial reduction in regional blood flow, and be associated with an MI.
3) Even if TEE detected intraoperative RWMAs are indicative of myocardial ischemia is that of prognostic significance? The answer to this is not as obvious as one might think.
The literature suggests that intraoperative ischemia is associated with an increased risk of perioperative MI and other ischemic complications in cardiac surgery. However, the vast majority of episodes of intraoperative ischemia do not result in MI. In their much quoted paper 12 Sloggoff and Keats showed that intraoperative ischemia increased the postoperative reinfarction rate from 2.5% to 6.9%. However, that means that in almost 93% of patients the ischemic events were benign. In non cardiac surgery the situation is even less clear. Most investigators have shown that preoperative and postoperative ischemia are associated with adverse cardiac outcomes, but intraoperative ischemia is less predictive. The patient's perioperative ischemic pattern is predicted by his preop ischemic risk. Associations between pre or postoperative ischemia and adverse cardiac outcomes do not necessarily entail cause and effect. It is much more likely that they are associated because of a common etiology - say the severity of the patient's ischemic heart disease which is what ultimately dictates perioperative risk. All we are seeing is that sick patients have bad disease and a worse prognosis.
As our understanding of the pathophysiology of ischemic heart disease improves we realize that the causes of myocardial ischemia and infarct are quite different. Myocardial ischemia is caused by an imbalance between myocardial oxygen supply and demand, while myocardial infarction (and the other unstable coronary syndromes) are caused by thrombosis on a fissured or ruptured atherosclerotic plaque. They both occur in patients with CAD, but their pathophysiologic mechanisms are completely different. It is not necessary to implicate one as causing the other.
Even if TEE is ultimately shown to be better at detecting myocardial ischemia and predicting outcome in noncardiac surgery than more con-ventional monitors there are drawbacks to its use:
1) It is invasive enough that it can only be used after induction and has to be removed before emergence. Thus the induction, intubation, emergence and postoperative periods are missed. As these are often periods of haemodynamic instability this is not inconsequential.
2) It does not distinguish between ischemic and nonischemic causes of RWMA. Work is ongoing on the use of myocardial perfusion contrast agents which may change this.
3) There are areas which cannot be imaged with monoplane probes. Newer generation probes which permit multiplane imaging have changed this. However, most of the studies were done with monoplane probes which may explain why some patients had ECG evidence of myocardial ischemia without RWMAs. Even with multiplane imaging only one two dimensional slice can be monitored at a time, and events in other planes can be missed. Work is going on on three dimensional echo which may change this.
4) There is not a great correlation between real time analysis even by trained anaesthetists working in a busy OR and experts doing off line analysis. Improved image storage to permit on line comparisons with current and stored images, and automated endocardial tracking systems may change this shortly.
5) A major time and training committment are required to master the technique and understand its limitations.
1) Systolic Function - Preload
Ideally we would like to be able to measure left ventricular end diastolic volume (LVEDV). In the operating room we rely on the measurement of intracardiac and intravascular pressures (CVP, PAD, PCWP) and hope that we can infer volume changes from changes in these pressures. However, the assumptions about ventricular compliance and cardiac filling - which are probably acceptable in healthy patients without cardiac or pulmonary disease - we make while doing this become more and more tenuous as patients become sicker. It is possible to measure directly LVEDV with TEE. To do this, however, assumptions are made to equate the LV with a mathematical model. The more complex this model is, the more accurate the result. Measurements in multiple planes and axes are required. The exercise is far too time consuming to be of any value in on line analysis. Fortunately for anaesthetist/echocardiographers most of the volume change in the LV occurs in the short axis. Short axis area changes have been well correlated with radionuclide estimates of LVEDV during surgery,13 and these short axis area changes can be qualitatively detected by experienced anaesthetist/echocardiographers with both a high sensitivity and specificity.14 Newer software which permits the automated detection of the interface between the endocardium and the LV intracavitary blood (ABD) have made the on-line measurement of LV short axis area possible and practical, with an excellent correalation between the ABD results and expert off line analysis. Overall, TEE provides a much better assessment of LV volume than either CVP or pulmonary artery catheter (PAC) data.
2) Systolic Function - Afterload
In combination with invasive arterial pressures TEE can provide the measurements necessary to calculate meridional wall stress which is a much better approximation of afterload than the more commonly used parameters such as SVR off PAC data. However, the measurements are tedious and time consuming. Although a useful reasearch tool, the technique is of little value in on line analysis in the operating room.
3) Systolic Function - Contractilty
TEE can be used to asses myocardial contractility both qualitatively by assessing the vigour of contraction and visually estimating ejection fraction (EF), and quantitatively by various measurements. Most of these measurments are ejection phase indices of contractility (things like ejection fraction, fractional area change, fractional shortening, cardiac output, velocity of circumferential shortening etc.) and are not load independent. That is, they can change in response to preload or afterload changes and not necessarily because of changes in contractilty. For the most part the measurements involved are too time consuming to make them of any clinical value in the operating room. An important exception is the End-Systolic Pressure Volume Relation (ESPVR). This is a plot of LV end -systolic pressure against LV end-systolic volume at various preload levels. It was developed by Suga and Sagawa in the 1970's and - at least over the normal range of afterloads- is load independent. Interfacing ABD measured short axis area changes (as a measure of LV volume) with invasive arterial pressures15 it is possible to create on-line ESPVRs. Although this technique is currently only experimental it offers the promise of easily obtainable on line measures of LV contractility.
4) Diastolic Function
TEE permits the easy measurement of flow velocities across the mitral valve and in the pulmonary veins which fill the left atrium. The patterns and absolute values of these measurements can be used to infer delayed LV relaxation and decreased LV compliance. However, the significance and specificity of these patterns is still being debated by echocardiographers. It is also unlikely that diastolic function changes acutely (unless caused by myocardial ischemia, and since these techniques detect global and not regional diastolic functional changes their sensitivity in detecting myocardial ischemia is exceedingly limited) in the OR, and that anaesthetists will ever make therapeutic decisions based on any of this information.
In summary, if TEE is compared to the competition (which for haemodynamic monitoring means a PAC) the results are:
TEE PAC
Preload +++ +
Afterload Lab +++ 0
OR 0 +
Contractilty +++ +
Diastolic function + +
I feel that TEE wins hands down as a haemodynamic monitor that, especially in guiding volume and inotropic therapy. The sicker the patient is the more it excels over conventional monitors. Recent reports in the literature confirm this. In a group of high risk patients undergoing cardiac surgery16 TEE lead to at least one change in anaesthetic management in 69% of patients, while in an ICU17 a major change in medical management was instituted in 29% of patients (and in 27% of patients with a PAC in place) because of TEE findings.
TEE as a Monitor for the Presence of Air, Fat and Other Assorted Junk
TEE is extremely sensitive at detecting air, fat and other emboli as they pass through or get lodged in the heart. The literature is filled with case reports of intraoperative pulmonary or venous air emboli detected by TEE. However, it is so sensitive that it picks up microscopic bubbles which are too small to see arising from a fast running IV and of no significance. This makes the predictive value of detecting intravascular air quite low. Although studies have shown that all patients undergoing total knee arthroplasty have showers of tiny emboli and some have larger macroscopic embolic masses (presumed to be thrombi) upon release of the tourniquet, these could not be related to any clinical picture suggestive of pulmonary emboli. The role of TEE as a monitor for emboli has yet to be addressed in any outcome studies.
I have attempted to review briefly the monitoring potential of TEE. In the hands of trained echocardiographers it is also a very powerful diagnostic tool. The close proximity of the heart to the esophagus permits ultrasound transducers positioned there to obtain images of exceptional clarity. The diagnostic information obtained can be of inestimable value in the management of critically ill patients with and without cardiac disease. Because of this immense diagnostic potential the use of TEE has spilled over from the echo lab to the operating room and intensive care unit. In these settings it is often used by noncardiologists who claim that they are using it as a "monitor" and, therefore, do not feel that they need the same kind of training and experience as echocardiographers who use TEE as a diagnostic aid. I feel that this separation is totally artificial. To "monitor" a patient's volume status is to diagnose hyper-, hypo- or euvolemia and to "monitor" wall motion is to diagnose hypo-, dys- or akinesis. What is one physician's monitor is clearly another's diagnosis. I do not think that TEE's use should be restricted to cardiologists. Anyone who understands cardiac physiology and pathophysiology (and that includes anaesthetists and intensivists) can understand the physiology that is behind the image and be trained as an echocardiographer. However, a dedicated period of study with trained echocardiographers is required to master the imaging techniques, recognize normal variants and abnormal patholgy and understand its limtations
Questions