| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Hydraulics Laboratory, Institute Biomedical Technology; and
Department of Intensive Care Unit, Ghent University Hospital and International Research Center, Ghent University, Belgium
Address correspondence and reprint requests to J Poelaert, MD, PhD, Department of Intensive Care Unit, 5K12 IE, Ghent University Hospital, De Pintelaan 185, B9000 Gent, Belgium. Address e-mail to jan.poelaert{at}rug.ac.be
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
IMPLICATIONS: We evaluated the effects of load changes on systolic mitral annular velocity (Sm) by Tissue Doppler Imaging velocity. Our results show that Sm is dependent on increases in preload and cannot be recommended as a variable of ventricular contractile performance in critically ill patients where significant changes in ventricular filling occur.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In the isolated fiber, contractility is an intrinsic property of cardiac muscle independent of afterload and length and influenced only by changes in the perfusate and temperature of the muscle bath (1). An ideal index that could measure and translate these concepts and make them applicable to the intact heart should therefore be independent of loading conditions (preload and afterload), heart size and mass, and sensitive to changes in inotropy (2). In this way, it should be possible to more accurately tailor supportive therapy in hemodynamically unstable patients. Moreover, because goal-directed therapy is associated with improvement of outcome (3), a load-independent index of ventricular contractility may permit a more accurate assessment and specifically tailored therapy.
The assessment of the systolic velocity of the mitral annulus (Sm) with pulsed-wave tissue Doppler imaging (TDI) allows for an estimation of global left ventricular (LV) systolic function similar to conventional two-dimensional or M-mode imaging, providing information on the velocity of the descent of the base as an estimate of global systolic ventricular function (4). Sm has been correlated with peak positive dP/dt and ejection fraction in patients with dilated cardiomyopathy and hypertensive heart disease (5,6) and is sensitive to changes in inotropy (7) or acute ischemia (8).
It is evident that a prerequisite for an index that reflects the contractile performance of the ventricle is its independence from loading conditions. The knowledge of load dependency of ventricular performance indices is an important issue in clinical practice and must be assessed before general use. We tested the hypothesis that Sm was load independent by imposing different preload and afterload conditions in patients after coronary artery bypass grafting.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
On arrival in the operating room, leads II and V5 electrocardiogram (ECG) monitoring was initiated. All patients received a brachial artery catheter, as previously described (Laboratoire Plastimed, St. Leu-la-Forêt, France). Its characteristics were previously tested and correlate to those of a typical fluid-filled catheter system in clinical use (9). After the induction of anesthesia, a transesophageal echocardiography (TEE)-probe (Omniplane II, Hewlett Packard 5-MHz probe, Andover, MA) was inserted and connected to an echocardiograph (Hewlett Packard Sonos 2500).
Exclusion criteria were reduced preoperative or postoperative LV ejection fraction (<60%) or regional wall motion abnormalities, mitral annular calcification, any esophageal or gastric pathology diagnosed as a contraindication to TEE, hemodynamic instability at admission to intensive care unit, intraaortic balloon pump, concomitant valvular diseases, supraventricular or ventricular rhythm disturbances, and postoperative administration of large doses of inotropes (dobutamine or dopamine >5 µg · kg–1 · min–1).
Upon arrival in the intensive care unit, patients were mechanically ventilated (fraction of inspired oxygen [FIO2] = 0.6), remained sedated after a standard protocol (propofol, 0.5–2.0 mg · kg–1 · h–1), and were randomly assigned to a single study group. In Group 1, 12 patients (Group 1a) underwent an increase of afterload and preload with a bolus of phenylephrine (Ph; 50–100 µg IV), and 12 patients (Group 1b) underwent a decrease of afterload and preload with a bolus of nitroglycerine (NTG; 300–500 µg IV). Using either drug, a change in systolic blood pressure of more than 20% from baseline was the goal. In all patients of Group 2, preload was changed with a rapid infusion of 500 mL of a gelatin solution (Gelofusin®) administered for more than 15 min. No changes in medications or IV fluid infusion were allowed during the study. In all patients, heart rate was kept constant by AAI pacing (80 ± 2 bpm).
The echocardiographic experimental protocol was initiated as soon as hemodynamic stability of the patient was present for at least 15 min, ending after all variables were measured. Care was taken that the three images required for echocardiographic and hemodynamic evaluation (2-chamber, 4-chamber, and deep transgastric) were recorded at the same value of arterial blood pressure (including mean, systolic, and diastolic). Digital recordings of a minimum of three consecutive cardiac cycles were stored for off-line analysis.
The TDI was recorded by using the pulsed Doppler software present in the echocardiograph. The spectral Doppler signal filter settings were adjusted at the lowest wall filter and at the minimum optional gain. The lateral mitral annulus, which was the myocardial area of interest, was aligned to be parallel to the sampling cursor, and the sweep speed was set at 50 mm/s. After stabilization of the mean arterial blood pressure (MAP), measurements were performed in apnea during three consecutive cardiac cycles. All variables were obtained at baseline and after the loading change. One investigator (RA), who performed all echocardiographic procedures, was blinded to the different loading intervention. The sample volume of TDI was placed at the lateral side of the mitral annulus in the mid-esophageal 4-chamber view (Fig 1a). The Sm is the peak negative systolic wall motion wave at the level of the T-wave of the ECG, as shown in Figure 1b.
|
Hemodynamic profile data in the different loading conditions were collected as previously described (11). In brief, MAP data were recorded together with a limb lead of the ECG on a workstation in conjunction with the Doppler aortic flow. All data were collected during apnea at baseline and after loading alteration. Customized software (Vision Bloodflow®, Belgium) enabled simultaneous acquisition of echocardiographic images and MAP. For the evaluation of the load profile of each patient, we analyzed off-line continuous-wave Doppler flow across the aortic valve and the MAP tracing using customized software written in Matlab 6.5 (The Mathworks, Natick, MA). Resistance (R) is calculated as MAP/cardiac output. Arterial elastance (Ea) was calculated as Pes/SV. The ST segment of the ECG and regional wall motion abnormalities were used to monitor a possible ischemia induced by pressure changes or drug intervention.
Because ventricular contractile performance can be altered by afterload itself, we estimated preload-adjusted maximal power (PAMP; a load-independent index of contractility (12)) to exclude that load-induced alteration in ventricular contractility could be responsible for the changes in Sm. Data of ventricular performance of all the patients during the different loading conditions were collected as previously described (11). Power (PWR) can be calculated from simultaneous recordings of pressure and flow through the ejection period. In the absence of mitral regurgitation or aortic valve dysfunction, PWR can be calculated as PWR = Plv(t) x Q(t), where Plv(t) is instantaneous ventricular pressure, and Q(t) is instantaneous aortic flow.
If we assume that ventricular pressure during the ejection can be approximated to the aortic blood pressure (Pao), the maximal value of PWR (PWRmax) can be described as PWRmax = Pao(t) x Q(t) max.
PWRmax was calculated off-line with Matlab 6.5 (The Mathworks) where the continuous-wave Doppler flow across the aortic valve was analyzed (semiautomatic contour tracing) and time aligned with the MAP waveform. PAMP was then calculated as PAMP = PWRmax/EDV2.
Statistical evaluation was performed using a personal computer-based package (SPSS 11.0, SPSS Inc, Chicago, IL). All data are shown as mean ± SD. Analysis of variance was used for initial statistical analysis. A Wilcoxon signed-rank sum test was applied to compare the results with the baseline values. Statistical significance was accepted at P < 0.05. Statistical power analysis (with a statistical power of 0.8 and a statistical significance of 0.05) showed that at least 4947 patients had to be evaluated to demonstrate statistically significant differences in the setting of Ph and 551–700 patients in the setting of NTG.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Tables 1–3 summarize the results of hemodynamics and Sm obtained at different loading conditions. All the variables, used as a monitoring of the loading intervention, changed significantly from baseline according to physiological behavior (increases in Ea and vice versa; for increases in EDV, no changes in arterial load variables were detected). The impact of a single change in one of the variables that describe arterial load (R and Ea) was not possible to evaluate (as all are changing with changes in afterload) and was not the goal of the investigation. No changes in ST segments of ECG or in regional wall motion were detected during the study.
|
|
|
In the first group (n = 24), all arterial load variables changed significantly from baseline. In Group 1a (Table 1), Ea and R increased significantly after the bolus of Ph. Also, there was an increase of EDV. SV in this group did not change significantly.
In Group 1b (Table 2), Ea, R, EDV, and SV decreased significantly. Despite the important changes in afterload in Groups 1a and 1b, Sm did not change (Fig. 2, A and B). In Group 2 (Table 3), both EDV and SV increased significantly from baseline. Sm significantly increased with the increase of preload (Fig. 2c). Load manipulation did not induce changes of PAMP (Table 4).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Sm by TDI reflects the peak velocity of myocardial fiber shortening oriented in the longitudinal direction and is believed to give an estimation of the global systolic function of the LV with a high feasibility and reproducibility (13,14). In our study, we found that for an increase of preload, there was a concomitant significant increase of Sm. This phenomenon can be simply justified by the length-dependent activation (Frank-Starling mechanism), which endows the ventricle with performance characteristics such that the heart ejects whatever volume it receives during diastole. Thus, an increase of the length of the myocardial fibers caused by an increase of EDV will lead to an increase of SV, the velocity of shortening, and also an increase of Sm.
Alternatively, our observations are also explained using the force-velocity-length relationship as the physiological background of Sm. Several studies (15–17) linked the fiber velocity or the contractile element to preload and ventricular tension during the complete contraction cycle, and even the so-called Vmax (the ideal velocity at zero load), as elegantly demonstrated by Pollack (17), is preload-dependent.
The independence of Sm on afterload seems contradictory. Assuming the force-velocity-length relationship is the background of Sm, it is to be expected that a decrease of arterial load (by infusion of NTG; Group 1b) and of ventricular tension or wall stress would have resulted in an increase of Sm. However, if the Frank-Starling relationship is the main determinant of the velocity of shortening of the mitral annulus, the decrease of EDV that we found and that is physiologically induced by NTG, would have resulted in a decrease of Sm. What we observe is likely the integrated effect of the two mechanisms; the effects of a decrease of preload on the fiber velocity (resulting in a decrease of the velocity of shortening and EDV) will be balanced by the concomitant effects of a decrease of afterload (resulting in an increase of the velocity because of a decrease of ventricular tension and ESV), and the net result is the unchanged value of the velocity of Sm.
The other result of our study is that Sm is also independent of an increase of afterload induced by an infusion of Ph, although we anticipated (as also previously detected by others (18)) a decrease of the velocity of shortening of the ventricle and Sm. The sudden increase of afterload has a complex impact on the hemodynamic behavior of the ventricle that is presently still not fully comprehended. First, EDV shows a significant increase with Ph, but the Frank-Starling mechanism this time is not able to produce an increase of Sm (as in Group 2) because of the sudden increase of afterload. The same increase of arterial load, following the force-velocity-length mechanism, would have led to a decrease of Sm, but this was also not observed. Therefore, a balance between afterload match and preload reserve may be responsible for our observations and can be a manifestation of good LV function (19).
Other explanations are all possible. Several studies have focused on the important changes in contractility during increases of afterload. The myofilament Ca2+ sensitivity, the force-generating capacity, or the homeometric autoregulation (Anrep effect) (20) are all physiologic effects that could be involved during an acute increase of afterload. The net results of these phenomena are an increase of contractility for an increase of afterload (21–24). In our study, we determined whether these factors contributed to the independence of Sm to the infusion of a bolus of Ph using a load-independent index of contractility (PAMP) (11,12). The results suggest that the load manipulation in our patients did not change the contractile performance of the ventricle because PAMP was not changed (Table 4).
Although we were able to demonstrate preload dependence, it is not feasible to entirely eliminate the chance of a type II error with regard to afterload independence of Sm. Exclusion of a type II error would require very large sample sizes (see Methods) that are impractical in the context of a clinical study with our design and methodology.
Some limitations could arise from the estimation of ventricular volumes with TEE, which does not allow a good visualization of the ventricular apex, even with the best quality images. Also, we used the modified Simpson rule to estimate ventricular volumes, which is a simplification of the Simpson rule obtained in 2 different 2D planes (4-chamber and 2-chamber). However, these possible limitations are present both in baseline conditions and after alteration of the loading conditions, so it is unlikely that they have a major impact on the overall evaluation of changes in EDV.
We used PAMP to estimate if manipulation of load can change the ventricular contractile performance, but it has to be emphasized that no studies are available on the sensitivity of PAMP to load. The lack of any change in PAMP could be the result of a low sensitivity of PAMP to detect changes in contractility induced by changes in load.
Because of wave travel and reflection, peripheral pressure (measured with a fluid-filled catheter) is only a poor estimate of true central Pao and ventricular pressure. This bias was partially bypassed using a long catheter inserted into the brachial artery that allowed us to estimate a pressure closer to the central aorta (25), although we acknowledge that the use of a long fluid-filled catheter may also induce distortion of the measured signal.
We preferred Ph and NTG for load manipulation. This method has the bias of a lack of any information of isolated changes in afterload without any change in preload. Nevertheless, in clinical practice (in animal or human research), these drugs are widely used for afterload manipulation and for the evaluation of the effects of load on ventricular function (24,26).
Although we enrolled only patients with good preoperative LV function (ejection fraction >60%), the values of Sm in baseline conditions are less than those of normal subjects (27). This could be the result of the important changes in preload that are present in all postoperative cardiac surgical patients. The baseline MAP differed between the patients in the group of NTG and in the group of PH, despite random group assignment. This possible bias cannot explain or change the response to the drug intervention because we analyzed the differences within the same group from baseline and after the change in loading conditions. Autonomic reflexes were not blocked in the present study, as generally is the case for human evaluation. Thus, reflex activation may have contributed to preload and afterload responses in our patients, although this effect could be reduced by keeping the heart rate constant by the external pacemaker.
By not having all subjects undergo the same protocol with a randomized sequence of interventions, it is difficult to compare the results between groups. This simplified study design was preferred to avoid the impact of repeated changes in loading on ventricular function.
From the results of this study, the TDI-derived systolic wave (Sm) at the level of the mitral annulus is dependent on increases in preload and insensitive to loading changes, as induced by the action of Ph and NTG (either a simultaneous increase or decrease of afterload and preload).
|
Acknowledgments |
|---|
The authors wish to thank G. Van Maele for his statistical advice.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Hori, T. Kano, F. Hoshi, and S.-i. Higuchi Relationship between tissue Doppler-derived RV systolic function and invasive hemodynamic measurements Am J Physiol Heart Circ Physiol, July 1, 2007; 293(1): H120 - H125. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. H. Y. Ie and R. Zietse Evaluation of cardiac function in the dialysis patient--a primer for the non-expert Nephrol. Dial. Transplant., June 1, 2006; 21(6): 1474 - 1481. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Poelaert The Failing Heart Under Stress: Echocardiography is an Essential Monitoring Tool in the Intensive Care Unit Seminars in Cardiothoracic and Vascular Anesthesia, March 1, 2006; 10(1): 111 - 115. [Abstract] [PDF]
|
|
|
|
|
|
E. H. Y. Ie, R. Krams, W. B. Vletter, R. W. Nette, W. Weimar, and R. Zietse Myocardial contractility does not determine the haemodynamic response during dialysis Nephrol. Dial. Transplant., November 1, 2005; 20(11): 2465 - 2471. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Bendjelid Impact of Volume Loading on Systolic Mitral Annular Velocity (Sm) by Tissue Doppler Imaging Anesth. Analg., March 1, 2005; 100(3): 901 - 901. [Full Text] [PDF]
|
|
|
|
|
|
R. Ama', P. Segers, C. Roosens, T. Claessens, P. Verdonck, and J. Poelaert Impact of Volume Loading on Systolic Mitral Annular Velocity (Sm) by Tissue Doppler Imaging Anesth. Analg., March 1, 2005; 100(3): 901 - 901. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
