Assessment of cardiovascular volume status by transoesophageal echocardiography and dye dilution during cardiac surgery

Abstract

Conventional evaluation of cardiovascular volume status by filling pressures is unreliable in critically ill patients. Measurements of left ventricular end diastolic area index by transoesophageal echocardiography and of intrathoracic blood volume index by dye indicator dilution are new approaches to this problem. In this study, different indices of cardiovascular volume status were analysed to define their relation during the pronounced haemodynamic changes associated with systemic inflammation after cardiopulmonary bypass. Correlations were performed with left ventricular end diastolic area index, intrathoracic blood volume index, central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP). Data from 15 patients receiving coronary artery bypass grafts were compared after induction of anaesthesia and in the intensive care unit. Spearman's correlation coefficient for perioperative absolute changes in left ventricular end diastolic area index and intrathoracic blood volume index was 0.87 (P<0.05). However, an increase in intrathoracic blood volume index by 125 mL m⁻² was necessary to maintain a baseline left ventricular end diastolic area index. Absolute values of all variables varied widely, with the only significant correlation found between CVP and PCWP. Changes in CVP and PCWP did not correlate with changes in left ventricular end diastolic area index or intrathoracic blood volume index. Provided simultaneous baseline measurements are available and a supranormal intrathoracic blood volume index compensates for the haemodynamic changes in systemic inflammation, left ventricular end diastolic area index and intrathoracic blood volume index may substitute for each other during the evaluation of cardiovascular volume status in patients with stable cardiac function.

Introduction

Maintaining an adequate cardiovascular volume status is important during the haemodynamic management of both anaesthetized and critically ill patients. However, the parameters used to assess cardiovascular filling in the clinical scenario do not include volumes, but pressures, that can be influenced by many factors. It is, therefore, not surprising that central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP) have been shown to be unreliable for the evaluation of cardiovascular volume status ^[1,2].

The intrathoracic blood volume (ITBV) determined by the indicator dilution technique has been proposed as a more reliable alternative to CVP and PCWP as a guide to volume therapy ^[1,3]. ITBV is a true measure of volume and comprises the end diastolic blood volumes of the right atrium, right ventricle, left atrium, left ventricle, the pulmonary blood volume and part of the systemic vascular blood volume, depending on the site of measurement.

An estimate of cardiovascular filling is also provided by transoesophageal echocardiography (TEE)^[4]. TEE allows for determination of the left ventricular end diastolic area (EDA). Changes in EDA seem to correlate with graded acute hypovolaemia in cardiac surgical patients ^[5]. A linear relation between changes in EDA and volume deficits was found in patients with both normal left ventricular function and left ventricular wall motion abnormalities. Reich et al.^[6] identified changes in EDA by manipulation of blood volume in paediatric patients. However, EDA can only provide information about left ventricular filling, and it is not clear how this would relate to the cardiovascular volume status if the patient were to develop a massive systemic inflammatory response with possible changes in cardiac function and systemic vasodilation. (Table 1)

Table 1: Biometrical data

The purpose of this study was to evaluate the relation between different measures of cardiovascular volume status during large changes in preload and cardiovascular function that occur during anaesthesia for cardiac surgery. It was hypothesized that EDA index (EDAI) and ITBV index (ITBVI) would correlate with each other perioperatively under different preload conditions.

Methods

This observation study was approved by the ethics committee of the University of Münster. Informed consent was obtained from each patient. The study was performed in 15 patients who underwent anaesthesia for coronary artery bypass grafting. Patients without sinus rhythm or with an intracardiac shunt or significant cardiac valve insufficiency, diagnosed preoperatively, were excluded.

All patients received flunitrazepam 1-2 mg for pre-medication. Anaesthesia was induced by the administration of propofol (1.0 mg kg⁻¹) and sufentanil (1 μg kg⁻¹) and maintained with propofol (2 mg kg⁻¹ h⁻¹) and sufentanil (1-2 μg kg⁻¹ h⁻¹). Tracheal intubation was performed after muscle relaxation with pancuronium bromide 0.1 mg kg⁻¹. The patients were mechanically ventilated with a mixture of O₂ in air (F_iO₂ = 0.5).

After induction of anaesthesia, a 7.5-French thermodilution pulmonary artery catheter was introduced through the right internal jugular vein and advanced further into the pulmonary artery wedge position. A combined 4-French fibreoptic thermistor catheter with an additional lumen for the measurement of arterial blood pressure (Fa. Pulsion, Munich, Germany) was positioned through a 5-French introducer sheath into the right femoral artery and advanced proximally by 20 cm. Correct positioning of the fibreoptic catheter was verified continuously by optical control of pulsatility of the reflected light signal.

Cardiopulmonary monitoring included TEE and the measurement of intravascular fluid volumes by the indicator dilution technique, as described below. Measurements were performed under routine treatment of the patient without interventions for study purposes. TEE and indicator dilution technique were performed by two independent investigators who were blinded to the co-investigators' results. The results of TEE and indicator dilution were presented to the anaesthesiologist in charge of the patient to guide treatment with vasoactive agents and fluid replacement therapy. Simultaneous measurements by both techniques were performed after induction of anaesthesia (baseline) and again 60-190 min later in the ICU after cardiopulmonary bypass. At this time, the patients were still sedated with propofol (1.3 ± 0.4 mg kg⁻¹ h⁻¹) and mechanically ventilated.

Measurement and calculation of intravascular volumes by indicator dilution

Bolus injections of 10-15 mL of the dye indicator, indocyanine green (ICG; 1 mg mL⁻¹ in ice-cold 5% glucose solution), into the right atrium were performed in triplicate spread randomly over the respiratory cycle. The dose of the dye was chosen to yield an appropriate dye dilution curve. The blood temperature curve was recorded in the pulmonary artery and in the distal aorta. The dye dilution curve was recorded in the distal aorta only. All indicator dilution curves were then processed further by a commercially available indicator dilution computer (COLD system; Pulsion Medizintechnik, Munich, Germany). Cardiac output was derived from pulmonary and aortic thermodilution curves according to the Stewart-Hamilton procedure in a standard fashion. The average value of pulmonary and aortic cardiac output was used for further calculations of cardiac index (CI), stroke volume index (SVI) and systemic and pulmonary vascular resistance indices (SVRI and PVRI respectively).

Intrathoracic blood volume was derived from the mean transit time of ICG (mtt ICG) and cardiac output ^[7,8].

ICG binds rapidly to plasma proteins and is therefore confined to the intravascular space. Calculating the volume of distribution of ICG therefore provides the intravascular fluid compartment between the right atrium and the 'systemic vascular tree' that is reached by ICG during mtt ICG. This fluid volume was corrected for body surface area (BSA) and presented as ITBVI. (Equation 1) The right partial volume (RPV) was derived from the mean transit time of the temperature signal travelling from the site of injection to the tip of the pulmonary artery catheter. RPV was then normalized for BSA and presented as RPVI. (Equation 2) The left partial volume index (LPVI) resulted from the difference between ITBVI and RPVI.

Measurement of EDA by transoesophageal echocardiography

After induction of anaesthesia and endotracheal intubation, a 5-mHz oesophageal transducer connected to an echocardiograph (Hewlett-Packard, Palo Alto, CA, USA) was advanced to obtain a cross-sectional short axis view of the left ventricle at the level of the mid-papillary muscles ^[9].

End diastolic and end systolic contours were traced off-line using the same system. The R-wave of the electrocardiogram was taken to determine the end diastolic frame. End systole was defined as the minimal cross-sectional area.

Three cardiac cycles were analysed for each measurement of ITBVI. As a mean ITBVI was calculated from three single ITBVI measurements, each EDAI provided in the graphs represents a mean of nine cardiac cycles.

The endocardial border was traced manually using the leading edge method. The anterolateral and post-eromedial papillary muscles were included in the area determinations (American Society of Echocardiography). EDA was normalized for body surface area (EDAI).

Statistics

Data are presented both as means±standard deviations (tables) and from individual patients (figures). Differences between pre- and post-operative haemodynamic measurements (Table 3) were evaluated statistically using the Wilcoxon signed-rank test. Statistical significance was defined as P<0.05.

Table 3: Haemodynamic indices before and after coronary artery bypass grafting (CABG)

The relation between two variables was tested with Spearman's correlation coefficient for non-parametric variables.

Regression analysis was performed to describe these relations if the correlation coefficient r_s was greater than 0.6.

Results

Information about the period of extracorporeal circulation and the post-operative use of vasoactive agents is provided in Table 2.

Table 2: Data concerning extracorporeal circulation and vasoactive drug therapy during haemodynamic measurements

Preoperative evaluation

Indices of ventricular filling varied over a wide range after induction of anaesthesia (Table 3). Correlations between the absolute values of different indices were poor, with the exception of CVP and PCWP (Fig. 1).

Fig. 1: Preoperative relations between intrathoracic blood volume index (ITBVI)/left ventricular end diastolic area (EDAI, first graph), pulmonary capillary wedge pressure (PCWP)/ITBVI (second graph), PCWP/EDAI (third graph) and PCWP/central venous pressure (CVP, fourth graph). Each of the four variables showed a wide range of values. Only CVP and PCWP showed a significant positive correlation.

There was also no significant correlation between SVI and EDAI, ITBVI, PCWP or CVP.

Post-operative evaluation

Marked haemodynamic alterations had occurred by the time the patients were re-evaluated in the intensive care unit (Table 3): SVRI had decreased, and CI had increased in conjunction with a rise in heart rate.

Again, CVP and PCWP were the only indices that showed a strong correlation (Fig. 2).

Fig. 2: Post-operative relations between intrathoracic blood volume index (ITBVI)/left ventricular end diastolic area index (EDAI, first graph), pulmonary capillary wedge pressure (PCWP)/ITBVI (second graph), PCWP/EDAI (third graph) and PCWP/central venous pressure (CVP, fourth graph). Each of the four variables showed a wide range of values. A significant positive correlation was found between CVP and PCWP.

Perioperative evaluation

When comparing pre- and post-operative mean values, EDAI remained unchanged, whereas ITBVI and its partial volumes, RPVI and LPVI, showed a significant increase (Table 3). However, a close correlation (r_s = 0.87) was observed for absolute changes between EDAI and ITBVI (Fig. 3). An increase in ITBVI by 125 mL m⁻² was necessary to maintain a baseline EDAI.

Fig. 3: Perioperative changes in the relations between intrathoracic blood volume index (ITBVI)/left ventricular end diastolic area index (EDAI, first graph), pulmonary capillary wedge pressure (PCWP)/ITBVI (second graph), PCWP/EDAI (third graph) and PCWP/central venous pressure (CVP, fourth graph). A significant correlation was found for perioperative changes in EDAI and ITBVI. In spite of a higher ITBVI in 13 out of 15 patients, 10 out of 15 patients had a lower EDAI than at baseline.

Perioperative changes in PCWP and CVP did not correlate with concomitant changes in EDAI or ITBVI. There was a poor, but statistically significant, correlation between PCWP and CVP (r_s = 0.56).

Changes in SVI showed positive correlations with changes in both EDAI (r_s = 0.72) and ITBVI (r_s = 0.61), perioperatively, but not with changes in PCWP or CVP.

Discussion

The goal of this study was to assess the relation between different parameters of cardiovascular volume status. A Spearman correlation coefficient of 0.87 was found for changes in EDAI and ITBVI measured immediately after the induction of anaesthesia (baseline) and during systemic inflammation after cardiac surgery in the ICU.

ITBVI reflects the blood volume between the injection site at the right atrium and the site of detection in the aorta. ITBVI is therefore a measure of both cardiac and vascular volumes. ITBVI was 882 ± 156 mL m⁻² or 21.7 ± 3.9 mL kg⁻¹ after induction of anaesthesia in the present study. London and colleagues ^[10] reported a cardiopulmonary blood volume of 741 mL m⁻² in normotensive humans. Hachenberg et al.^[11] found an intrathoracic blood volume of 18.4 ± 2.7 mL kg⁻¹ in cardiac surgical patients after induction of anaesthesia. The slightly higher value in the present study may be explained by the positioning of the fibreoptic catheter 10-20 cm more distally in the aorta than may have been used by the other investigators ^[10,11]. Hence, a higher vascular blood volume should have contributed to ITBVI in the present study.

Mean post-operative EDAI showed a trend towards lower values than at baseline, while ITBVI was significantly elevated after surgery. EDAI was decreased in two-thirds of patients, whereas ITBVI was increased in 13 out of 15 patients. The increase in ITBVI was caused by a higher RPVI and LPVI.

RPVI represents the right heart volume plus that part of the pulmonary intravascular volume that has been reached by ICG during mtt_{(thermopulmo)}. The increase in RPVI was most probably caused by an increase in right heart volume. LPVI consists of that part of the pulmonary intravascular volume that has not been reached by ICG during mtt_{(thermopulmo)}, plus the left heart volume and that part of the systemic arterial intravascular volume that has been reached by ICG during mtt_(ICG). Hypothetically, several mechanisms may have contributed to the marked elevations in ITBVI in the presence of reductions or minor increases in EDAI. In the low pressure system of the pulmonary circulation, static pulmonary blood volume may have increased perioperatively without marked changes in pulmonary artery pressure. Moreover, the volume of the arterial 'vascular tree' reached by ICG during mtt_(ICG) may have increased during a state of systemic vasodilation. Left ventricular compliance may have decreased perioperatively, thereby shifting changes in left ventricular volume to a higher left ventricular pressure. Perioperative changes in the shape of the left ventricle caused by myocardial oedema formation may have resulted in a different relation between EDAI and left ventricular volume.

There was a poor correlation between absolute EDAI and ITBVI. EDAI is determined by the dimensions of the left ventricle. These depend on ventricular size and myocardial function such as left ventricular diastolic relaxation, but also on ventricular compliance, vascular volume and tone and on intrathoracic pressure. ITBVI represents the whole intravascular volume between the right atrium and the distal aorta and is, therefore, much less dependent on left ventricular size than EDAI. It is therefore conceivable that there was no relation between the absolute values of these totally different physiological parameters.

In order to explore a potential relation between EDAI and ITBVI, changes in both parameters were compared and showed a correlation coefficient of 0.87.

If the effects of myocardial stunning on left ventricular compliance and diastolic relaxation were similar in all patients after extracorporeal circulation, the perioperative relation between EDAI and ITBVI, demonstrated here, could indeed represent intrathoracic blood volume dependency of EDAI. Similar schemes for vasoactive drugs were given to all patients (Table 2) and were therefore unlikely to change the intrathoracic blood volume dependency of EDAI. Changes in EDA correlated well with changes in vascular volumes in previous studies ^[5,6].

It was not unexpected to find no significant correlation between vascular filling pressures and EDAI or ITBVI. PCWP and CVP measure vascular pressures. As in the case of EDAI, these depend on factors other than vascular volume, such as vascular tone, myocardial function and intrathoracic pressure. A comprehensive review of the factors influencing PCWP has been provided by Tuman et al.^[12]. PCWP and CVP did not seem to be a useful measure of cardiovascular volume status in this study and have also been shown to be unreliable in other investigations ^[1,2,13]. The good correlation between CVP and PCWP is consistent with the absence of high stage right or left ventricular insufficiency or atrioventricular valve insufficiency in the patients studied. The latter were also excluded by TEE.

Perioperative alterations in SVI correlated with those of EDAI and ITBVI. An increase in SVI in combination with an elevated ITBVI has been found in patients with the adult respiratory distress syndrome ^[3]. The authors suggested that ITBVI may, therefore, be a useful parameter for the evaluation of preload. SVI will also increase if afterload is decreased. In the present study, the perioperative haemodynamic management after cardiopulmonary bypass was associated with an elevated ITBVI. These results are consistent with Hachenberg's data obtained in the same setting ^[11]. In his investigation, intrathoracic blood volume increased by more than 25% after cardiopulmonary bypass when systemic vascular resistance was markedly decreased.

TEE is being used increasingly for the cardiovascular monitoring of critically ill patients in the intensive care unit and during anaesthesia ^[14,15]. Besides information on vascular filling and preload, a wide variety of other useful information is provided by this technique, including that on global and regional cardiac function, cardiac valves, atrial thrombi or pericardial effusion ^[14,15]. TEE has been shown to produce a high diagnostic yield and important information for prompt therapeutic decision-making during cardiac anaesthesia as well as in the ICU and has, therefore, been advocated as a valuable technique for the evaluation of the diseased heart ^[9,15]. However, it requires the presence of an experienced investigator and the availability of expensive equipment. This may limit the repeated use of TEE. Other parameters may, therefore, be needed to assess the cardiovascular volume status in anaesthetized and critically ill patients, including those recovering from cardiac surgery.

Although the results of the present study were only of an observational character, suggestions about the management of cardiovascular volume status by EDAI and ITBVI during cardiac anaesthesia can be drawn. If preoperative ITBVI is available, post-operative optimization of cardiovascular volume status may be guided further by ITBVI. Supranormal ITBVI may be required once a marked reduction in systemic vascular resistance has occurred. The current data imply that ITBVI can be increased by more than 100 mL, if a baseline EDAI is desired. TEE should be conducted for specific evaluation of the heart, if haemodynamic stabilization cannot be achieved despite marked increases in ITBVI with cardiac output being inadequately low.

In conclusion, changes in ITBVI and EDAI showed a high correlation coefficient (0.87) during the specific conditions studied. However, an increase in ITBVI of 125 mL m⁻² had to compensate for the haemodynamic changes of systemic inflammation if a baseline EDAI was desired post-operatively. Our data suggest that, with this supranormal ITBVI provided and simultaneous baseline measurements available from both techniques, the two parameters could have been used in substitution for each other in these patients with a stable cardiac function. If myocardial function is compromised echocardiographic assessment seems appropriate, as it provides specific information about the heart. Cardiac filling pressures do not seem to be useful parameters of cardiovascular volume status in this setting. Further studies are needed to investigate the potential role of ITBVI in patients after complicated cardiac surgery.