| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Departments of *Anesthesiology, Cardiac Surgery, Biostatistics, and Medicine, Montreal Heart Institute, Montreal, Quebec, Canada
Address correspondence and reprint requests to Pierre Couture, MD, Department of Anesthesiology, Montreal Heart Institute, 5000 Belanger St. E., Montreal, Quebec, Canada, H1T 1C8. Address e-mail to p.couture{at}sympatico.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
IMPLICATIONS: In the presence of low cardiac output, the clinician’s ability to identify which patients are more likely to benefit from volume administration to improve hemodynamic status while avoiding fluid overload is important. The analysis of Doppler measurement of the mitral flow as an indirect indicator of the individual diastolic pressure/volume relationship may be useful to predict the intravascular volume responsiveness in patients undergoing coronary artery bypass graft surgery.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Preload can also be estimated with transesophageal echocardiography (TEE) through direct measurement of cavity dimensions as estimates of LV end-diastolic volume. LV end-diastolic area (EDA) more accurately reflects LV preload than PCWP (3). However, even if the measurement of EDA with TEE is potentially useful to monitor cardiac filling (4–6), a study by Tavernier et al. (7) suggests that EDA estimated with TEE is not clearly superior to PCWP to guide fluid therapy in patients with sepsis-induced hypotension. However, pulsed Doppler examinations of the mitral flow (MF) and pulmonary venous flow (PVF) can provide estimation of LV filling pressure and are sensitive to changes in preload (8–10). We have previously documented that respiratory variation of the PVF could be of some benefit in evaluating the preload (11). However, the value of MF Doppler variables as a predictive index of responsiveness of cardiac output (CO) to changes in preload has not been evaluated in the cardiac surgical patient.
Because the goal of preload optimization is to increase CO (an index of cardiac performance), the ability of a preload index to predict changes in CO in response to fluid challenge is important. In this study, we tested the ability of hemodynamic variables, TEE-derived EDA, and MF Doppler measurements to predict responsiveness of the CO to changes in preload. Accordingly, we simultaneously measured standard hemodynamic variables, the EDA and Doppler measurements of the transmitral and PVF, and changes in CO during acute preload variations in patients undergoing coronary artery bypass grafting surgery (CABG).
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Sufentanil (5–10 µg/kg), midazolam (0.05–0.1 mg/kg), and pancuronium (0.1 mg/kg) were used to induce anesthesia. Anesthesia was maintained by incremental doses of opioids and pancuronium and was supplemented with isoflurane as needed. Lungs were ventilated by intermittent positive-pressure ventilation with a fraction of inspired oxygen of 1.0 by use of an Ohmeda volume-cycled ventilator (Ohmeda, Helsinki, Finland), with a respiratory rate of 10 breaths/min, a tidal volume of 10 mL/kg, and an inspiratory/expiratory ratio of 1:2, which was kept constant for the entire experiment. Fluid and medications were administered as needed according to the patient’s clinical status.
Monitoring included a five-lead electrocardiogram, a pulse oximeter, an infrared CO2 analyzer, and peak inspiratory pressure. A radial artery catheter and a 4-lumen 7.5F pulmonary artery catheter (Baxter, Irvine, CA) were inserted. All pressures were monitored with heparinized saline-filled polyethylene catheters attached to pressure transducers (Transpac 42587-05; Sorenson, Chicago, IL) referenced to atmospheric pressure and positioned at the midthoracic level. A 5.0-MHz TEE omniplane probe interfaced with a phased array imaging system (Sonos 1500; Hewlett-Packard, Andover, MA) was inserted after the induction of general anesthesia.
After the induction of anesthesia, we performed the TEE examination, which included a short-axis transgastric view at the midpapillary level, a midesophageal four-chamber view, and color flow Doppler imaging of the mitral valve to detect any unsuspected significant mitral disease. Hemodynamic measurements consisted of heart rate; systolic, diastolic, and mean arterial pressure; and systolic, diastolic, and mean pulmonary artery pressure. The PCWP was determined at end-expiration. CO was measured in triplicate by use of the thermodilution technique (Sirecust 1281; Siemens, Danvers, MA).
Nearly simultaneously, we first obtained a baseline transgastric short-axis view of the LV at the midpapillary level, followed by a pulsed Doppler examination of PVF and MF. To measure PVF, we positioned the Doppler sample volume (2-mm width) in the left upper pulmonary vein approximately 1 cm proximal to its entrance into the left atrium and used color Doppler flow to sample maximal flow. When necessary, to minimize the angle between the Doppler beam and the pulmonary vein’s long axis, we rotated the transducer array of the omniplane probe as far as needed from the horizontal plane (11). This axis was maintained throughout the PVF examination. Mitral inflow velocities were measured at the tip of the mitral leaflets.
Two independent observers were involved: one recorded hemodynamic variables and the other, blinded to the hemodynamic data, simultaneously recorded the pulsed Doppler and two-dimensional echocardiographic images. TEE evaluations were performed by an anesthesiologist board certified in perioperative TEE who was not in charge of the patient.
The measurements were taken after the positioning of the arterial and venous CPB cannulas and systemic heparinization, before the beginning of CPB, with the pericardium and chest opened. The protocol consisted of the measurement of hemodynamic, two-dimensional echocardiographic, and Doppler signals (dependent variables) during preload variation (independent variable). These variables were measured during three periods, as follows: 1) control measurements, without any preload manipulation; 2) during a decrease in preload, obtained by draining the patient’s blood into the cardiotomy reservoir after systemic heparinization to decrease the mean arterial pressure by 20%–30% of the control value or to a minimum of a systolic arterial pressure of 95 mm Hg during 1 or 2 min; and 3) during an increase in preload, by transfusion of crystalloids from the CPB circuit with the arterial cannula to obtain a PCWP of 15–18 mm Hg. The quantity of crystalloids transfused was noted.
To validate our findings obtained in the pilot study, we performed the same hemodynamic and echocardiographic measurements before and after a rapid-volume infusion of 500 mL of 10% pentastarch (Pentaspan®; Dupont Pharma Inc., Mississauga, Ontario, Canada) given over 5 min, with the chest opened, during the dissection of the internal mammary artery, in 36 additional patients with varying systolic and diastolic function.
From the LV midpapillary short-axis view, we measured the EDA, end-systolic area, and fractional area change by manual tracing of the endocardium by the leading edge to leading edge technique (13). Three consecutive beats were used and averaged independently of the respiratory cycle. From the PVF velocity tracings, we measured peak systolic (S wave) and peak diastolic (D wave) velocities, the peak velocity of flow reversal during atrial contraction (A wave), and their respective velocity/time integrals (VTI S, D, and A). The VTI S was measured from the onset of forward flow after the peak R wave on the electrocardiogram to the cross-over of that wave with the zero line. The VTI D was measured from the onset of the second wave to its cross-over with the zero line. The VTI of flow reversal caused by atrial contraction (VTI A) was measured from onset to termination of negative flow velocity in late diastole. From the MF velocities, the following variables were recorded: peak velocity of early diastolic filling wave (E) and late diastolic filling wave (A), their respective VTI E and A, the ratios of these velocities or of these integrals (E/A; VTI E/VTI A), and the deceleration time of the E wave (dt). The dt was measured as the time between the peak E velocity and the return to zero baseline of early diastolic flow velocity (if the velocity did not return to baseline, the deceleration slope was extrapolated to baseline). After anesthetic induction, diastolic function was characterized according to the Canadian Consensus Recommendations for the Measurement and Reporting of Diastolic Dysfunction by Echocardiography (14). In the Validation group, we arbitrarily defined as "fluid responsive" those patients who demonstrated an increase in stroke volume (SV) of >20% (15). Our experience with the use of TEE and our inter- and intraobserver variability has been previously published for the evaluation of systolic and diastolic function (11,16,17).
The average of three consecutive measurements was used. All the TEE images were recorded on 0.5-in. VHS videotape for further off-line analysis.
For the pilot study, analysis of variance for repeated measures was used to analyze the evolution of the hemodynamic, two-dimensional, and Doppler TEE measurements during preload variation, followed by Tukey’s tests for multiple comparisons when differences were found. Linear regression analysis was used to evaluate the relationship between Doppler mea-surements before volume infusion and the increase in CO. Stepwise multiple linear regression was used to identify the influence of hemodynamic and MF Doppler variables on the increase of CO after volume infusion.
For the validation study, Student’s t-test was used to compare the hemodynamics before and after the infusion of 500 mL of 10% pentastarch. To assess the ability of hemodynamic variables and echocardiographic measurements to discriminate between positive and negative responses to fluid challenge, receiver operating characteristic (ROC) curves were generated. In this Validation group, the area under the ROC curve represented the probability that a random pair of responders and nonresponders after fluid infusion would be correctly ranked by the Doppler variable. A value of 0.5 indicates that the screening mea-sure is no better than chance, whereas a value of 1 implies perfect performance. Quantitatively, the closer the plot is to the upper left corner, the higher the overall predictive power of the criterion. A P value <0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Thirty-six patients were included in the validation study. We evaluated 28 men and 8 women, with a mean age of 61.1 ± 10.2 yr and a mean weight of 77.7 ± 21.8 kg. Thirty-three patients had CABG, and three had aortic valve replacement. Twenty-seven patients were taking ß-adrenergic blockers, 13 were taking calcium channel blockers, and 14 took angiotensin-converting enzyme inhibitors up to the time of surgery. Seven patients were diabetic, and 23 had hypertension. After anesthetic induction, 22 patients were classified as having normal diastolic function, 5 had a relaxation abnormality pattern, 6 had a pseudonormal pattern, and 3 had restrictive diastolic patterns. The mean ejection fraction was 44.9% ± 10.2%. Fourteen patients had an ejection fraction <40%. We excluded 3 of 72 measurement periods for the PVF that were considered unsuitable for analysis. The remaining data were included for statistical analysis.
In the pilot study, the average amount of blood drained or infused to modify preload was 211 ± 87 mL and 176 ± 149 mL, respectively. The effect of preload variation in the pilot study is shown in Table 1. There was a significant change in CO and cardiac index (CI) during preload variation (2.48 ± 0.34, 2.14 ± 0.36, and 2.88 ± 0.37 L · min-1 · m-2 for the control, decreased, and increased preload, respectively; P < 0.0001), with no other significant change in hemodynamic variables. There was a significant change in the mitral VTI E/A ratio during preload variation (1.16 ± 0.43 vs 1.73 ± 0.67 for the decreased and increased preload, respectively; P < 0.05). There was no other significant change in other hemodynamic, two-dimensional echocardiographic, or other MF and PVF Doppler variables.
|
|
|
|
The VTI E/A was the most important variable to explain the change in CO. The removal of this variable led to a lack of correlation of the model.
Hemodynamic and echocardiographic variables measured before and after the fluid infusion during the validation study are shown in Table 2. The performance of hemodynamic and MF Doppler variables to predict an increase in SV of >20% after the infusion of 500 mL of 10% pentastarch was evaluated by constructing ROC curves. The areas under the ROC curves for PCWP, central venous pressure, mean pulmonary artery pressure, EDA, and CI were not statistically different from 0.5. However, the area under the ROC curve for the MF E/A ratio was 0.71 (95% confidence interval, 0.54 to 0.88; P < 0.05), and the optimal threshold value given by ROC analysis for the E/A ratio was 1.26 (Table 3). A ratio of <1.26 before the fluid infusion had a sensitivity of 75%, a specificity of 60%, a positive predictive value of 60%, and a negative predictive value of 75% for a >20% increase in SV after fluid infusion.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
To confirm our findings, we validated our preliminary results in 36 additional patients receiving 500 mL of 10% pentastarch with a range of systolic and diastolic function. By use of ROC curves, we assessed the ability of hemodynamic and echocardiographic variables to predict the responders and nonresponders. In this heterogeneous group of patients, only the E/A ratio of the mitral valve flow could predict the response to fluid challenge, with an area under the ROC curve of 0.71. The ROC analysis identified an optimal threshold value for the E/A ratio of 1.26, although the sensitivity and specificity were only moderate (75% and 60%, respectively) to predict responsiveness to fluid infusion. The clinical relevance of these findings is that the MF Doppler E/A ratio may help the clinician to identify patients who are more likely to benefit from volume administration by improvement in CO.
In a previous study of surgical patients undergoing hypervolemic hemodilution (18), volume administration resulted in significant increases in SV and EDA, which increased up to a threshold beyond which no further increases were observed. The PCWP, however, continued to increase. In a study in dogs (5), both the LV EDA and PCWP increased in response to volume administration, and these changes correlated well with changes in CO. Further volume administration did not result in any further increase in EDA or CO, although PCWP progressively increased. These studies suggest that a threshold or end point for fluid administration could be determined for EDA. This threshold in EDA in response to volume administration is influenced by diastolic function of the LV. Our observations are supported by Tousignant et al. (15), who found that nonresponders to fluid infusion had a lower LV compliance, because EDA changed little, whereas the PCWP increased.
The value of the mitral E/A ratio to predict preload responsiveness could be related to the diastolic function of the LV. This ratio was decreased in the reduced preload state, which is similar to the MF observed in relaxation abnormalities (14). After an increase in preload, this diastolic pattern reverted to a normal E/A ratio. As shown by our results, the lower the E/A ratio, the more likely that the patient will have an increase in CO during preload enhancement. These observations may suggest that the diastolic mitral pattern characterized by impaired relaxation could be associated with fluid responsiveness. Conversely, an increase in CO is less likely to occur after volume infusion in patients with a high E/A ratio (>1.5), which is associated with a restrictive pattern (14). Consequently, measurements of the E/A ratio may be useful to determine the position of the diastolic pressure/volume curve of the LV. Lower E/A ratios suggest a leftward position on this curve, and a higher ratio, a rightward position, where LV compliance is reduced (Fig. 2).
|
In clinical practice, measurement of EDA with TEE is considered to be the best method to determine the LV preload (3). Reich et al. (6) reported the ability to identify a decrease in blood volume ranging from 5% to 8% with a sensitivity varying from 80% to 95% and a specificity of 80% in a pediatric population undergoing cardiac surgery. In adult cardiac surgery, Cheung et al. (4) concluded that TEE-derived EDA was able to detect a decrease of 2.5% in effective blood volume, or approximately 1.75 mL/kg. We observed positive, but not significant, changes in EDA with the acute preload variations performed in the pilot study and in the validation study. The change in EDA would probably have been significant if more patients had been included. Other factors can explain our results, including the magnitude and the rapid rate of change in the effective blood volume obtained and differences in patient population. More recently, consistent with these observations, Tavernier et al. (7) reported that LV diastolic dimension measured as EDA did not predict the response to volume loading as accurately as systolic pressure variation during mechanical ventilation in patients with hypotension and sepsis. Tousignant et al. (15) also observed that it was not possible to establish a threshold of EDA below which a large proportion of patients responds to volume administration in a cardiac surgical group, with the responders being distributed over a wide range of EDA.
This study has several limitations. First, the results apply to patients undergoing CABG surgery with the pericardium and the chest opened and cannot be generalized to other situations. Second, many factors other than preload variations can affect MF and PVF, including diastolic dysfunction (26). The diastolic function of the patients of the pilot study was normal and was unchanged in a given patient during the short study period. The patients in the Validation group had heterogenous systolic and diastolic function; this could partly explain the moderate sensibility of the E/A ratio of the MF to predict preload responsiveness after fluid challenge. Third, we did not try to define an optimal preload in which a peak in CI occurs and represents a meaningful end point for IV fluid administration. Finally, our definition of responders as those who increase SV to >20%, although previously used by other investigators (15), is arbitrary. Although this increase in SV is clinically important, our definition of responders may have been too restrictive, because some patients with an E/A ratio <1.26 may have had a less important increase (from 10% to 20%, for example), which remains clinically significant, particularly in a patient with a borderline hemodynamic status.
We conclude that MF Doppler measurement is superior to hemodynamic and two-dimensional echocardiographic variables to predict the increase in CO after fluid challenge in patients undergoing cardiac surgery. The analysis of the MF can be viewed as an indirect indicator of the individual diastolic pressure/volume relationship for each patient. This may be useful to predict the volume responsiveness for optimization of the hemodynamic status in patients undergoing CABG.
|
Acknowledgments |
|---|
The authors thank Hewlett-Packard, Pointe-Claire, Quebec, Canada, for an equipment grant. We also thank the cardiac surgeons of the Montreal Heart Institute for their collaboration; Annie Laprade for secretarial assistance; and Micheline Roy, Raymonde Garant, and Linda Dufresne for their assistance as research assistants.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Solus-Biguenet, M. Fleyfel, B. Tavernier, E. Kipnis, J. Onimus, E. Robin, G. Lebuffe, C. Decoene, F. R. Pruvot, and B. Vallet Non-invasive prediction of fluid responsiveness during major hepatic surgery Br. J. Anaesth., December 1, 2006; 97(6): 808 - 816. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Y. Denault, P. Couture, J. Buithieu, F. Haddad, M. Carrier, D. Babin, S. Levesque, and J.-C. Tardif Left and right ventricular diastolic dysfunction as predictors of difficult separation from cardiopulmonary bypass: [La dysfonction ventriculaire diastolique gauche et droite comme predicteur des difficultes de sevrage de la circulation extracorporelle]. Can J Anesth, October 1, 2006; 53(10): 1020 - 1029. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
