Nonpulsatile continuous-flow left ventricular assist devices (LVADs) have significantly alleviated the morbidity and mortality risks related to advanced heart failure. They are used either as a bridge to transplantation until a donor heart is available, as a bridge to recovery, or as destination therapy. The use of LVADs is currently on the rise because of the large discrepancy between the number of patients listed for heart transplantation and the scarcity of available donor hearts.1
Fluid administration plays a key role in the hemodynamic management of post-LVAD implantation patients during their postoperative care, when clinical determination of the intravascular volume can be extremely difficult. On the one hand, uncorrected hypovolemia leads to inappropriate infusion of vasopressors and could cause organ hypoperfusion, whereas excessive fluid administration and over-resuscitation could lead to increased complications and mortality.2–4 Therefore, a reliable assessment of a patient’s intravascular fluid status together with an accurate prediction of fluid responsiveness are crucial. To predict fluid responsiveness successfully, a change in preload needs to be created, while measuring the subsequent changes in cardiac output.
Passive leg raising (PLR) is a noninvasive method that provides a fast and accurate way to guide fluid management in a wide spectrum of critically ill patients.5,6 Passive leg raising is done by placing the patient in a supine position with legs elevated to 45 degrees, thereby creating a transient endogenous fluid challenge that mimics fluid administration.7,8
In this prospective study, we tested the hypothesis that PLR early after LVAD implantation, using the LVAD monitor flow variation to assess cardiac output, helps predict fluid responsiveness in these critically ill patients.
Materials and Methods
This prospective study was approved by the Ethics Committee of the Sheba Medical Center in August 2016 and was conducted in the cardiac surgery intensive care unit (ICU) of the Sheba Medical Center (Tel Hashomer, Israel) between October 2016 and November 2017.
Twenty consecutive adult patients who underwent nonpulsatile continuous-flow HeartMate III LVAD (St. Jude Medical Inc, St. Paul, MN) implantation were included in the study. Written informed consent was obtained from every patient before surgery.
The first set of measurements (baseline) was obtained immediately (within 30 minutes) after surgery when the patient (in a supine position) arrived at the ICU. These measurements included LVAD flow (Figure 1), arterial blood pressure, heart rate, central venous pressure (CVP; through the right jugular vein), end-tidal carbon dioxide (ETCO2), intraoperative fluid balance, cardiac output obtained from the measured thermodilution curves using a pulmonary artery Swan-Ganz catheter (Edwards Lifesciences, Irvine, CA), and systemic vascular resistance (SVR) obtained from the formula: cardiac output = mean arterial pressure − CVP/SVR. Immediately after baseline measurements, PLR was performed for 2 minutes as previously described.9 The PLR effect was assessed with a continuous measurement of the LVAD flow, arterial blood pressure, heart rate, CVP, and ETCO2. All medications, including vasopressors, were kept constant during this period. Patients with an increase of 15% or more in LVAD flow during the PLR maneuver were classified as volume responders. Patients with <15% change in LVAD flow were classified as nonresponders. Volume responders were treated with fluid administration according to our ICU protocol, followed by repeat measurements 1 hour later. Those with a 15% increase in LVAD flow after the fluid challenge were classified as fluid responders.
Data are presented as mean ± standard deviation if normally distributed or as median (interquartile range). Categorical variables are given as frequencies and percentages. Continuous variables were tested with the t-test for normal distribution, and Mann–Whitney–Wilcoxon test for abnormally distributed variables. A χ2 test was used for comparison of categorical variables between PLR responders and nonresponders, and fluid responders and nonresponders.
To identify factors associated with PLR responsiveness among the entire cohort, a multivariable logistic regression model was constructed. Candidate covariates are presented in Tables 1–3. Variables associated with PLR responsiveness (p < 0.2) were included in a stepwise analysis. Results are presented as odds ratio and 95% confidence interval.
Statistical significance was assumed when the null hypothesis could be rejected at p < 0.05. All p values are the results of two-sided tests. Statistical analyses were conducted using R Core Team (2015), a language and environment for statistical computing, R Foundation for Statistical Computing, Vienna, Austria, (https://www.R-project.org/; version 3.4.1).
A power analysis was conducted to investigate whether the sample size was adequate to show differences between fluid responders and nonresponders using a two-sided Z-test. We defined α to be significant if p was <0.05 and β <20%. With this definition, we found that with a sample size of 20 patients, we had power of 87.98% to reject the null hypothesis for the primary outcome (efficacy of PLR maneuver to predict fluid responsiveness).
Mean patient age was 54 years, and 80% were male. All patients had severe left ventricular dysfunction preoperatively and underwent an elective LVAD implantation because of decompensated heart failure (New York Heart Association III-IV functional class score). Preoperatively, four patients had severe right ventricular dysfunction, eight had moderate dysfunction, four had mild dysfunction, and four had normal right ventricular function. Patients with moderate or more than moderate reduction in right ventricular function were treated with levosimendan (Simdax; Orion Corporation, Espoo, Finland) 48 hours before the LVAD implantation as part of the preparation protocol. Before surgery, a lower extremity venous ultrasound-Doppler study showed normal results in all patients. All baseline characteristics are presented in Table 1. Mean operation time was 248 minutes, and mean cardiopulmonary bypass time was 89 minutes. The operation was performed on a beating heart in 18 patients (Table 2).
After performing the PLR maneuver, 11 patients (55%) were classified as responders and 9 (45%) as nonresponders. Baseline measurements including mean LVAD flow, cardiac output, ETCO2, mean arterial pressure, and CVP were similar between the responders and nonresponders, whereas SVR was significantly higher in the responders (p = 0.003; Table 3). Furthermore, patients who were volume responders had a lower intraoperative fluid balance compared with the nonresponders (224 ml vs. 1,095 ml, respectively; p = 0.126).
The mean elevation in ETCO2 after the PLR maneuver was 4.6 mm Hg in the volume responders and 1.1 mm Hg in the nonresponders (p < 0.001). All of the responders had an elevation of at least 2 mm Hg in their ETCO2, whereas only two of the nonresponders had an elevation of 2 mm Hg or more (p = 0.002). The CVP was raised by 35% in the responders and by 8% in the nonresponders (p = 0.008). Although the mean arterial pressure was raised by 5% in the responders, it decreased by 2% in the nonresponders after the PLR maneuver (p = 0.012; Table 4). Logistic regression with stepwise selection analysis demonstrated that only high SVR (above 1,250 dynes·sec/cm5) was significantly associated with leg raising responsiveness (p = 0.025; Figure 2).
Responders to the PLR maneuver were treated with an average of 1,586 ± 986 ml of Ringer’s lactate solution during the first hour after PLR. All of them demonstrated an increase in LVAD flow (an average rise of 24%) and in ETCO2 (an average rise of 4.3 mm Hg). However, two patients from the responders group had less than a 15% increase in LVAD flow (11% and 8%) after fluid administration and were actually non-fluid responders by the LVAD flow variation. Furthermore, these two patients had a lower increase in their ETCO2 (a rise of only 2.5 vs. 4.7 mm Hg; p = 0.348) compared with the other nine patients who underwent PLR and were volume and fluid responders.
There was no difference in the early outcomes between PLR responders and nonresponders. There were no cases of in-hospital mortality or stroke in the current cohort. One patient who was a nonresponder to PLR underwent reopening after surgery because of bleeding. The mean ventilation time was 44.8 ± 82.5 hours, mean ICU stay was 90.6 ± 114 hours, and mean hospital stay was 14.2 ± 6.2 days.
The main finding of this study is that the PLR maneuver using the LVAD monitor flow chart is an accurate and reliable tool to predict fluid responsiveness in patients early after LVAD implantation. Left ventricular assist device flow variation is correlated with changes in cardiac output induced by the PLR maneuver and could provide a noninvasive method for predicting fluid responsiveness.
Central venous pressure, central venous oxygen saturation, Swan-Ganz catheter, arterial pressure waveform, ultrasound, echocardiography, and other noninvasive cardiac output monitoring may assist in evaluating the intravascular fluid state and predict fluid responsiveness.10,11 In cases of low sensitivity and lack of availability, it has been reported in some previous studies that only half of hemodynamically unstable patients are volume responders12 and that PLR is a noninvasive validated dynamic method to guide fluid management in critically ill patients.13,14 Continued assessment of cardiac output can be performed by ETCO2 variations,15 using the Vigileo Flo-Trac system (Edwards LifeSciences, Irvine, CA),16,17 and in patients with a LVAD by the device flow shown on the display unit (Figure 1).
In the current study, the population was highly selective and included only patients with LVAD support, and, therefore, it was possible to measure cardiac output continuously by the LVAD flow chart. Although ETCO2 variation has been shown to be a validated surrogate to cardiac output during PLR, with a specificity of 100% and sensitivity of 71%,18 it has some limitations: 1) because carbon dioxide production (VCO2) is not constant, being dependent on cell metabolism and other variables (such as fever and shivering), ETCO2 variation is able to monitor only during short time periods, and, therefore, would fail to reflect the change in cardiac output; 2) the effect of pulmonary diseases, such as chronic obstructive pulmonary disease, on the validity of ETCO2 could bring into question its reliability; 3) patients need to be deeply sedated or paralyzed to eliminate fluctuation of minute ventilation and prevent any spontaneous breathing activity, whereas LVAD flow variation can be carried out on awake and nonmechanical ventilated patients.
The HeartMate II LVAD flow estimator has demonstrated potential variation in flow estimation. The correlation between flow and current in axial pumps, such as the HeartMate II, is not quite linear over the full flow range and therefore less accurate for flows below 3 L/minute. In a study of estimated flow accuracy in patients on 20 HeartMate II, Slaughter et al.19 concluded that the estimated flow values of the device can provide accurate directional information for trend purposes but less accurate information for absolute values of pump flows. Unlike the axial HeartMate II pump, the HeartMate III is a centrifugal pump that has a linear current-to-flow relationship across the full range of operating pump flows, providing more accurate flow estimation. Furthermore, the HeartMate III LVADs uses the patient’s hematocrit level as a surrogate for the patient’s blood viscosity, a factor that influences flow estimator accuracy. In our current report, we entered updated hematocrit values to the LVAD settings before the measurements. Furthermore, the two measurements were performed within a few minutes of each other, and, therefore, any changes in the hematocrit values were unlikely.
This prospective study provides a simple, non–time-consuming, noninvasive, and safe tool to predict fluid responsiveness by PLR maneuver and LVAD flow variation. The PLR maneuver induces a transient shift of venous blood from the lower limbs and the abdominal compartment towards the heart, increasing cardiac preload.20 Although fluid responders are on the steep part of their Frank-Starling curve and fluid administration induces a significant increase in cardiac output, patients who are classified as nonresponders are assumed to be further along the Frank-Starling curve, toward the flat part, where they will not gain from fluid administration, and where, as a result, there will be no improvement in their cardiac output. Should fluid be given to the nonresponders, the result would likely be worsening edema.
One of the findings of this study was the higher SVR among responders, reflecting their hypovolemic state. Although there was a significant difference in the SVR, there were only modest changes in arterial pressure and LVAD flow between the groups, meaning that these patients were hemodynamically stable. These hypovolemic patients had sufficient compensation at the time the measurements were taken, thus causing no significant reduction in arterial pressure or LVAD flow. However, it is assumed that a continuing hypovolemic state would have predisposed these patients to a decompensation state if not treated or if treated incorrectly (i.e., by vasopressors). Among PLR responders, an increase in preload by the PLR maneuver led to a significant rise in the LVAD flow, arterial pressure, CVP and ETCO2, a phenomenon not seen in those patients classified as nonresponders. Using the PLR maneuver, hypovolemic patients can be treated early on, thereby avoiding any effect on hemodynamic parameters.
We used the volume challenge test as a standard reference method to decide whether a PLR-responder patient was fluid responsive or not. Using an increase of 15% of the LVAD flow as a cutoff value after the fluid challenge, we identified that 82% of the PLR responders were also fluid responders. In our cohort, a cutoff rise of 17.5% in the LVAD flow after PLR would have given 100% sensitivity for the PLR maneuver for fluid responsiveness. Those who were volume responders by the PLR maneuver, but fluid nonresponders by the fluid challenge and had a <15% increase in the LVAD flow, had an increase of 11% and 8%, respectively, after fluid administration, with no volume overload adverse effect such as pulmonary edema. In our cohort as in previous reports, CVP variation is not considered to be a variable that predicts fluid responsiveness.21
Right ventricular failure is common after LVAD implantation and is associated with higher mortality. Our cohort included four patients with severe, eight with moderate, and four with preoperative mild right ventricular dysfunction. Fluid management is extremely important to maximize right ventricular function. Although the right ventricle is dependent on precise preload to produce optimal function, it is sensitive to volume overload that can result in functional deterioration. Therefore, a precise volume evaluation is critical in the management of patients with varying degrees of right ventricular dysfunction, and, thus, it is essential to identify those patients who will benefit from resuscitation and those who will not. To the best of our knowledge, the current study is the first to show that PLR is a reliable method for regulating volume evaluation in LVAD-implanted patients early after implantation of the device.
This study has some limitations. First, because of the small cohort size, the study is underpowered to evaluate the positive predictive value of PLR for fluid responsiveness in LVAD-implanted patients. Second, to avoid pulmonary edema, the nonresponders were not treated with fluid administration. As a result, we could not calculate the false-negative amount and, therefore, could not check the negative predictive value of the PLR maneuver by LVAD flow variation. Third, the HeartMate III LVAD flow estimator was used as a surrogate to cardiac output, and as a result, the potential flow change by the PLR maneuver could have been within measured accuracy error limitations.
Passive leg raising maneuver is a noninvasive and easy to perform method to assess fluid responsiveness in patients after LVAD implantation. The current study is the first to show the accuracy of PLR in LVAD patients using the LVAD flow for cardiac output monitoring displayed on an LVAD flow display screen and validated by ETCO2 variation.
1. Patel CB, Cowger JA, Zuckermann A. A contemporary review of mechanical circulatory support. J Heart Lung Transplant 2014.33: 667–674.
2. Marik PE, Monnet X, Teboul JL. Hemodynamic parameters to guide fluid therapy
. Ann Intensive Care 2011.1: 1.
3. Monnet X, Marik PE, Teboul JL. Prediction of fluid responsiveness: an update. Ann Intensive Care 2016.6: 111.
4. Boyd JH, Forbes J, Nakada TA, Walley KR, Russell JA. Fluid resuscitation in septic shock: a positive fluid balance and elevated central venous pressure are associated with increased mortality. Crit Care Med 2011.39: 259–65.
5. Cherpanath TG, Geerts BF, Lagrand WK, Schultz MJ, Groeneveld AB. Basic concepts of fluid responsiveness. Neth Heart J 2013.21: 530–536.
6. Xu J, Peng X, Pan C, et al. Fluid responsiveness predicted by transcutaneous partial pressure of oxygen in patients with circulatory failure: a prospective study. Ann Intensive Care 2017.7: 56.
7. Monnet X, Teboul JL. Passive leg raising. Intensive Care Med 2008.34: 659–663.
8. Saugel B, Kirsche SV, Hapfelmeier A, et al. Prediction of fluid responsiveness in patients admitted to the medical intensive care unit. J Crit Care 2013.28: 537.e1537.e–9.
9. Monnet X, Rienzo M, Osman D, et al. Passive leg raising predicts fluid responsiveness in the critically ill. Crit Care Med 2006.34: 1402–1407.
10. Garg M, Sen J, Goyal S, Chaudhry D. Comparative evaluation of central venous pressure and sonographic inferior vena cava variability in assessing fluid responsiveness in septic shock. Indian J Crit Care Med 2016.20: 708–713.
11. Oord M, Olgers TJ, Doff-Holman M, Harms MP, Ligtenberg JJ, Ter Maaten JC. Ultrasound and NICOM in the assessment of fluid responsiveness in patients with mild sepsis in the emergency department: a pilot study. BMJ Open 2017.7: e013465.
12. Marik PE, Cavallazzi R, Vasu T, Hirani A. Dynamic changes in arterial waveform derived variables and fluid responsiveness in mechanically ventilated patients: a systematic review of the literature. Crit Care Med 2009.37: 2642264–7.
13. Cavallaro F, Sandroni C, Marano C, et al. Diagnostic accuracy of passive leg raising for prediction of fluid responsiveness in adults: systematic review and meta-analysis of clinical studies. Intensive Care Med 2010.36: 1475–1483.
14. Cherpanath TG, Hirsch A, Geerts BF, et al. Predicting fluid responsiveness by passive leg raising: a systematic review and meta-analysis of 23 clinical trials. Crit Care Med 2016.44: 981–991.
15. Young A, Marik PE, Sibole S, Grooms D, Levitov A. Changes in end-tidal carbon dioxide and volumetric carbon dioxide as predictors of volume responsiveness in hemodynamically unstable patients. J Cardiothorac Vasc Anesth 2013.27: 681–684.
16. Biais M, Vidil L, Sarrabay P, Cottenceau V, Revel P, Sztark F. Changes in stroke volume induced by passive leg raising in spontaneously breathing patients: comparison between echocardiography and Vigileo/FloTrac device. Crit Care 2009.13: R195.
17. Zimmermann A, Kufner C, Hofbauer S, et al. The accuracy of the Vigileo/FloTrac continuous cardiac output monitor. J Cardiothorac Vasc Anesth 2008.22: 388–393.
18. Monnet X, Bataille A, Magalhaes E, et al. End-tidal carbon dioxide is better than arterial pressure for predicting volume responsiveness by the passive leg raising test. Intensive Care Med 2013.39: 93–100.
19. Slaughter MS, Bartoli CR, Sobieski MA, et al. Intraoperative evaluation of the HeartMate II flow estimator. J Heart Lung Transplant 2009.28: 39–43.
20. Rutlen DL, Wackers FJ, Zaret BL. Radionuclide assessment of peripheral intravascular capacity: a technique to measure intravascular volume changes in the capacitance circulation in man. Circulation 1981.64: 146–152.
21. Marik PE, Baram M, Vahid B. Does central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest 2008.134: 172–178.