To evaluate hemodynamic and functional changes of the failed left ventricle by Velocity Vector Imaging (VVI) and tissue Doppler, 22 patients with cardiogenic shock supported by extracorporeal life support (ECLS) were imaged during ECLS output variations inducing severe load manipulations.
The following data were acquired: (1) mean arterial pressure, aortic Doppler velocity-time integral, left ventricular end-diastolic volume, and mitral Doppler E wave; (2) tissue Doppler systolic (Sa) and early diastolic (Ea) velocities; and (3) systolic peak velocity (Sv), strain, and strain rate using VVI.
Load variations were documented by a significant decrease in afterload (mean arterial pressure, −21%), an increase in preload (left ventricular end-diastolic volume, +12%; E, +46%; E/Ea ratio, +22%), and an increase in the velocity-time integral (+45%). VVI parameters increased (Sv, +36%; strain, +81%; and strain rate, +67%; P < .05), unlike tissue Doppler systolic velocities (+2%; P = NS). Whatever the ECLS flow, Sa was higher in patients who survived.
VVI parameters are not useful in characterizing the failed left ventricle with rapidly varying load conditions. Tissue Doppler systolic velocities appear to be load independent and thus could help in the management of ECLS patients.
Extracorporeal life support (ECLS) has been proposed for rescuing patients with refractory cardiogenic shock. It has been successfully used as a bridge to myocardial recovery, cardiac transplantation, or the implantation of left ventricular (LV) assist devices in patients with various etiologies of overt cardiac failure.
A venous cannula inserted into the right atrium drains blood from the patient into the pumping mechanism of the ECLS circuit. The blood is oxygenated through a membrane oxygenator and perfused in the aorta by a centrifugal pump via a second cannula. ECLS can be instituted centrally, through the right atrium and the ascending aorta, or peripherally, using the femoral vein and the femoral or axillary artery. During ECLS, right atrial drainage through the venous cannula partially or totally unloads the heart. The external device provides both circulatory and ventilatory functions and infuses oxygenated blood in the aorta; it ensures a contribution to the systemic output that may reach 100%, with output provided by the native heart of zero. By varying the flow of the ECLS centrifugal pump, load conditions are modified : it decreases preload by draining blood directly from the right atrium and increases mean afterload by a steady flow infusion of oxygenated blood in the aorta. ECLS is a model of load variation that simultaneously manipulates preload and afterload and is of physiologic interest. It is therefore suitable for assessing and validating LV parameters independent of load conditions.
The vast majority of LV functional parameters are sensitive to load conditions such as filling pressures and mean aortic pressure. A recently introduced postprocessing of native echocardiographic sequences, Velocity Vector Imaging (VVI; Siemens Healthcare, Erlangen, Germany), might provide less sensitive parameters, but this point is thus far not clearly established. The present study was designed to assess the influence of acute modifications of ECLS on the failed left ventricle by studying VVI echocardiographic parameters by comparison with other two-dimensional and Doppler echocardiographic data. The secondary goal was to evaluate their relationship to patient outcomes.
This study was conducted in accordance with the ethical standards of our hospital’s committee for the protection of human subjects. Informed consent for demographic, physiologic, and hospital outcomes data analyses was not required, because this observational study did not modify existing diagnostic or therapeutic strategies.
This prospective observational study included all patients who received ECLS at the Cardiology Institute of Hôpital La Pitié-Salpêtrière in Paris between February and August 2007. Venoarterial ECLS was initiated under the following circumstances: (1) acute refractory cardiogenic shock complicating acute myocardial infarction, end-stage dilated cardiomyopathy, or fulminant myocarditis; (2) postcardiotomy cardiogenic shock; (3) immediate post-transplantation cardiac graft failure, with elevated pulmonary pressures and right ventricular or LV dysfunction or both; and (4) miscellaneous conditions (e.g., cardiotoxic drug overdose, acute cardiac allograft rejection, persistent cardiac arrest). Patients receiving venovenous ECLS or those with mitral prostheses or severe mitral valvulopathy were not included in the study.
The collected clinical parameters were age, sex, medical history, reason for admission to the intensive care unit, severity at admission assessed by the Simplified Acute Physiology Score II (range, 0–174), type of ECLS (central or peripheral), ECLS duration, and outcome (weaning and survey at 1 month). Hemodynamic status was assessed daily by the invasive measurement of arterial pressure (systolic arterial pressure, diastolic arterial pressure, and mean arterial pressure) and heart rate.
The group of weaned patients was defined as those having the machine removed and not requiring further mechanical support in the following 30 days. Patients who could not be weaned from ECLS either died in the intensive care unit or were bridged to heart transplantation or to a ventricular assist device.
ECLS Management in the Intensive Care Unit
Patients were kept on ECLS for ≥48 to 72 hours. To determine the feasibility of permanent weaning, an ECLS weaning trial was undertaken in patients considered hemodynamically stable (i.e., baseline mean blood pressure > 60 mm Hg while receiving no or low-dose vasoactive agents and a pulsatile arterial waveform maintained for ≥24 hours, and when pulmonary blood oxygenation was not compromised). When a patient tolerated the full weaning trial and had an LV ejection fraction (LVEF) > 25% to 30% under minimal ECLS, ECLS removal was considered. If the patient remained stable after prolonged (15–20 min) complete-circuit clamping in the operating room, the machine was surgically removed and the mediastinum or femoral access surgically repaired. When ECLS weaning was deemed impossible, bridging to a ventricular assist device or transplantation was considered.
ECLS Weaning Trial and Load Manipulations
During the ECLS weaning trial, the patients were tested by means of acute stepwise decrease of support. The patients were exposed simultaneously to increased preload and decreased afterload. ECLS flow was decreased to 66%, 33%, and <10% of the initial output of the device for 10 min at each level. Between each level, ECLS flow was returned to 100% for 15 min. During this transient weaning, data were collected at different levels of ECLS: 100%, 66%, and 33% of the initial value of flow up to, if possible, and <10% of the initial output of the device. At each level, parameters were recorded after 5 min of the ECLS flow decrease.
If mean arterial pressure dropped significantly and constantly to <60 mm Hg during the trial, ECLS flow was returned to 100% of the initial flow, and the trial was stopped.
Population and Load Manipulations
Thirty-three patients were assessed. They were studied daily during 119 attempts at partial weaning. Twenty-two patients tolerated acute decreases to 66% of the initial support and were included in the final analysis. Eleven patients failed to endure <66% of the initial support and thus were excluded from the final analysis. All 22 patients who tolerated variations of ECLS output could be imaged. A total of 70 load conditions were stored and processed. The baseline characteristics of the 22 included patients are summarized in Table 1 .
|Age (y)||48 (20–71)|
|Reason for ECLS implantation|
|Simplified Acute Physiology Score II||66 ± 23|
|Central (vs femoral) ECLS||11|
|ECLS duration||7 (2–22)|
|Weaning from support||11|
Echographic Examination and Conventional Doppler Echocardiographic Data
All echocardiographic examinations were performed daily at each level of ECLS flow by transthoracic echocardiography (Acuson Sequoia; Siemens Healthcare). The apical view was used to obtain standard four-chamber, three-chamber, and two-chamber long-axis views of the left ventricle. Two-dimensional black-and-white sequences corresponding to two beats were stored with a frame rate of 70 frames/sec in Digital Imaging and Communications in Medicine format for subsequent offline analysis. These sequences were used to measure LV end-diastolic volume (LVEDV) and to calculate LVEF using Simpson’s rule. Transmitral early peak (E) and late diastolic velocities were recorded with pulsed Doppler in a four-chamber view. Spectral Doppler tissue imaging placed at the lateral and the septal mitral annulus measured early and late annular diastolic peak myocardial velocities. Then, pulsed Doppler was performed on the aortic valve to measure the aortic velocity-time integral (VTI). Systolic velocity (Sa) and early diastolic (Ea) peak were obtained with the use of spectral Doppler tissue imaging at the mitral annular level (septal and lateral) ( Figure 1 ). LV filling pressures and their variations were estimated with LVEDV and the E/Ea ratio.
VVI Principles and Measurements
VVI is a recent technique for obtaining velocity, strain, and strain rate measurements. It analyzes motion by tracking speckles (natural acoustic markers) in the two dimensions of the ultrasonic image sequences. The displacement of each speckle is followed from one image to the next. This yields the systolic peak velocity (Sv), strain, and strain rate imaging. It has emerged as a quantitative technique to estimate myocardial function and contractility. Strain is a dimensionless parameter representing deformation of an object, relative to its original shape, and is expressed as the percentage (or fractional) change from the original dimension. Strain rate is the local rate of deformation or strain per unit time, which equals velocity difference per length unit.
VVI uses all of the information available in the images. The algorithm incorporates tracking of the mitral plane, the inward and outward motion of the border, and the motion of tissue along the border.
For the purpose of this study, echographic data were archived in Digital Imaging and Communications in Medicine format and then transferred for offline postprocessing using VVI software. On an end-diastolic four-chamber long-axis view of the left ventricle, the endocardial contour was manually traced and used as an initial reference by the postprocessing software. The quality of the tracking was verified for each loop, and new endocardial traces were recorded when necessary. For each single cardiac cycle, longitudinal myocardial velocities (systolic and diastolic), longitudinal systolic strain, and longitudinal strain rate (systolic) were calculated by the software and collected at the mitral annulus of the apical view for the study ( Figure 2 ).
Because the majority of the measured parameters showed non-Gaussian distributions, data are expressed as medians with interquartile ranges (IQRs). Variations are expressed in absolute and relative (Δ%) values. Echocardiographic data during ECLS flow variations were compared using a nonparametric equivalent of analysis of variance (Wilcoxon’s test). Categorical variables were compared using χ 2 tests. Changes of clinical and echocardiographic parameters from 100% to minimal ECLS flow were compared between weaned and nonweaned patients with repeated-measures analysis of variance. The Bonferroni-type correction was performed to compare echocardiographic data among different ECLS flows. In other cases, statistical significance was defined as P < .05.
Of the 22 patients who received ECLS, 11 were ultimately weaned from the device. Among the 11 nonweaned patients, three had LV assist devices, one underwent transplantation, and seven died. Twelve patients (11 weaned and one who underwent transplantation) survived at 1 month.
Hemodynamic Changes during ECLS Variations
The initial support was 4.0 L/min (IQR, 3.0–4.5 L/min) and could be decreased by a mean of 83%. This decrease led to a drop in mean arterial pressure from 95 mm Hg (IQR, 82–103 mm Hg) to 75 mm Hg (IQR, 68–89 mm Hg), an increase in LVEDV from 95 mL (IQR, 72–132 mL) to 108 mL (IQR, 86–149 mL), and an increase in the lateral E/Ea ratio from 5.9 (IQR, 3.5–8.0) to 7.2 (IQR, 5.3–9.4) ( P < .05; Table 2 ). Heart rate remained unchanged.
|ECLS flow (L/min)||4.0 (3.0 to 4.5)||0.7 (0.6 to 1.1)||−83 †|
|MAP (mm Hg)||95 (82 to 103)||75 (68 to 89)||−21 †|
|SAP (mm Hg)||113 (95 to 174)||101 (87 to 182)||−11 ∗|
|DAP (mm Hg)||75 (63 to 143)||57 (48 to 128)||−24 †|
|HR (beats/min)||98 (80 to 109)||100 (80 to 109)||+2|
|Lateral E/Ea ratio||5.9 (3.5 to 8.0)||7.2 (5.3 to 9.4)||+22 †|
|LVEF (%)||15.0 (10.0 to 25.0)||17.5 (10.0 to 30.0)||+17 †|
|VTI (cm)||8.0 (5.9 to 11.1)||11.6 (9.2 to 13.1)||+45 †|
|LVEDV (mL)||95 (72 to 132)||108 (86 to 149)||+12 †|
|E (cm/sec)||45.0 (22.3 to 62.8)||65.8 (47.1 to 82.2)||+46 †|
|Tissue Doppler lateral Ea (cm/sec)||7.5 (5.9 to 9.3)||8.6 (7.2 to 11.0)||+15 †|
|Tissue Doppler lateral Sa (cm/sec)||5.2 (4.1 to 6.3)||5.3 (4.6 to 6.5)||+2|
|Lateral systolic Sv (cm/sec)||1.1 (0.6 to 1.6)||1.5 (1.2 to 2.1)||+36 †|
|Lateral systolic strain (%)||−2.7 (1.5 to 4.9)||−4.9 (3.2 to 8.9)||+81 ∗|
|Lateral systolic strain rate (sec −1 )||−0.3 (0.2 to 0.5)||−0.5 (0.3 to 0.7)||+67 ∗|
Conventional Doppler Echocardiographic Data: Initial Values and Variations
The conventional echocardiographic values recorded at 100% and at minimal ECLS flow and the percentages of the variations are presented in Table 2 . When ECLS flow decreased, LVEF and the VTI increased ( P < .001), whereas lateral Sa remained unchanged, from 5.2 cm/sec (IQR, 4.1–6.3 cm/sec) to 5.3 cm/sec (IQR, 4.6–6.5 cm/sec).
The conventional Doppler echocardiographic data recorded at different steps of ECLS flow are reported in Table 3 .
|Variable||ECLS flow (%)|
|Number of patients||22||22||19||13|
|MAP (mm Hg)||95 (82 to 103)||85 (74 to 94) †||77 (68 to 88) †||70 (67 to 84) †|
|Lateral E/Ea ratio||5.9 (3.5 to 8.0)||6.2 (4.4 to 9.0) ‡||8.3 (4.9 to 9.3) †||9.1 (4.8 to 9.5) †|
|LVEF (%)||15.0 (10.0 to 25.0)||15.0 (10.0 to 30.0)||17.5 (10.0 to 30.0) †||30.0 (15.0 to 35.0) †|
|VTI (cm)||8.0 (5.9 to 11.1)||10.2 (6.9 to 11.9) †||12.2 (7.9 to 12.8) †||12.4 (9.9 to 13.5) †|
|E (cm/sec)||45.0 (22.3 to 62.8)||57.1 (36.8 to 74.4) †||64.9 (47.8 to 82.7) †||65.8 (47.1 to 85.3) †|
|Tissue Doppler lateral Ea (cm/sec)||7.5 (5.9 to 9.3)||8.4 (7.0 to 9.7) ‡||8.0 (7.5 to 11.1) †||8.1 (6.2 to 12.7) §|
|Tissue Doppler lateral Sa (cm/sec)||5.2 (4.1 to 6.3)||5.3 (4.3 to 6.4)||5.5 (4.6 to 6.6)||5.3 (5.0 to 6.2)|
|Lateral systolic Sv (cm/sec)||1.1 (0.6 to 1.6)||1.5 (0.8 to 2.5) †||1.6 (1.2 to 2.0) †||1.7 (1.2 to 2.1) †|
|Lateral systolic strain (%)||−2.7 (1.5 to 4.9)||−5.5 (4.0 to 9.1) §||−6.2 (4.5 to 7.9) §||−5.9 (4.3 to 10.0) §|
|Lateral systolic strain rate (sec −1 )||−0.3 (0.2 to 0.5)||−0.5 (0.4 to 0.6) †||−0.5 (0.4 to 0.7) †||−0.5 (0.3 to 0.7) ∗|
VVI Parameters: Initial Values, Variations, and Comparisons with Doppler Velocities
The variations in VVI parameters are summarized in Table 2 . The longitudinal velocities were modified by ECLS flow variations ( P < .01), and their increases exceeded 20%. Strain and strain rate were sensitive to load conditions ( P < .05). When ECLS flow decreased, lateral systolic strain increased from −2.7% (IQR, 1.5% to 4.9%) to −4.9% (IQR, 3.2% to 8.9%) and the lateral systolic strain rate from −0.3 sec −1 (IQR, 0.2 to 0.5 sec −1 ) to −0.5 sec −1 (IQR, 0.3 to 0.7 sec −1 ) ( P < .01).
The VVI data recorded at different steps of ECLS flow are reported in Table 3 .
Comparing tissue Doppler systolic velocities at the level of the mitral annulus, we found significantly lower values with the use of VVI ( P < .01).
Comparison of Conventional Echocardiographic and VVI Parameters at Maximal and Minimal ECLS Flow according to Weaned Status
The positioning of ECLS (central or peripheral) had no impact on weaning (seven weaned patients with central ECLS vs four in the peripheral group, P = .41).
At maximal ECLS flow, successfully weaned patients had higher aortic VTIs, LVEFs, and Sa ( Table 4 ). However, indices of LV filling pressure (pulsed Doppler mitral E velocity, tissue Doppler Ea velocity, and E/Ea ratio) and VVI parameters did not differ significantly between the two groups. At minimal ECLS flow, the same differences were observed ( Table 4 ).
|Variable||Maximal ECLS flow||Minimal ECLS flow|
|Weaned||Not weanable||Weaned||Not weanable|
|ECLS flow (L/min)||3.3 (2.8 to 4.1)||4.3 (3.8 to 5.1) ∗||0.7 (0.6 to 0.9)||0.8 (0.6 to 1.5)|
|MAP (mm Hg)||96 (82 to 107)||94 (83 to 101)||78 (70 to 92)||75 (67 to 84)|
|Lateral E/Ea ratio||6.3 (3.1 to 8.7)||5.0 (3.5 to 6.4)||8.4 (6.2 to 9.9)||6.4 (5.2 to 7.4)|
|LVEF (%)||25.0 (15.0 to 38.8)||10.5 (5.0 to 11.3) †||27.5 (20.0 to 43.8)||10 (5.0 to 15.0) †|
|VTI (cm)||10.5 (7.4 to 12.2)||6.2 (4.6 to 8.8) †||12.8 (11.0 to 14.2)||9.5 (7.0 to 11.6) ‡|
|E (cm/sec)||54.4 (22.0 to 69.2)||37.1 (24.6 to 53.4)||66.4 (50.7 to 87.5)||60.8 (46.1 to 74.3)|
|Tissue Doppler lateral Ea (cm/sec)||7.8 (6.3 to 9.3)||6.8 (5.1 to 8.9)||8.5 (7.2 to 10.7)||8.8 (7.4 to 11.3)|
|Tissue Doppler lateral Sa (cm/sec)||6.0 (5.1 to 6.8)||4.7 (3.6 to 5.2) ∗||6.2 (5.2 to 7.3)||4.6 (3.9 to 5.3) ∗|
|Lateral systolic Sv (cm/sec)||1.5 (1.1 to 1.7)||0.9 (0.6 to 1.4)||1.9 (1.5 to 2.2)||1.4 (1.1 to 1.8)|
|Lateral systolic strain (%)||−3.0 (2.2 to 4.1)||−2.0 (1.5 to 4.9)||−5.1 (3.2 to 14.3)||−4.6 (3.4 to 6.7)|
|Lateral systolic strain rate (s −1 )||−0.4 (0.2 to 0.7)||−0.3 (0.2 to 0.4)||−0.5 (0.4 to 0.8)||−0.5 (0.3 to 0.7)|