Echocardiography is used to guide pump speed settings in patients supported with a continuous-flow left ventricular (LV) assist device (CF-LVAD). The aim is to secure LV unloading, a sufficient cardiac output (CO), intermittent aortic valve (AV) opening, and at the same time avoid septal shift/arrhythmic events.1 It has been suggested that changes in LV end-diastolic diameter (LVEDD) as a function of pump speed (revolutions per minute [RPM])—performed as a ramp study—are indicative of pump function and malfunction and that LVEDD and RPM are inversely linear related in pumps without obstruction to flow.2–4 Pulmonary capillary wedge pressure (PCWP) is the key to describing LV unloading; however, this parameter is not routinely used because of the invasive nature of the required procedure. Contrary echocardiography and LVEDD measures are easy accessible and used in current clinical practice. The relation between pressure (PCWP) and the currently used echocardiography-derived surrogate of LV volume, LVEDD, as a function of RPM is unknown in CF-LVAD patients. The aim of this study was to examine the correlation between LVEDD and LV filling pressure as a function of LVAD RPM speed change. Furthermore, we measured hemodynamics during submaximal exercise (with baseline and increased pump speed) and assessed the potential effect on the pressure–volume relation.
In this prospective study, patients underwent ramp testing with simultaneous measurements of hemodynamics by Swan-Ganz catheter and cardiac dimensions and AV opening by transthoracic echocardiography.
All CF-LVAD patients (HeartMate II; Thoratec Corp., Pleasanton, CA) treated at the University Hospital Rigshospitalet at the time of the study were considered for inclusion. Participants had to be 18 years or older, anticoagulation had to be well regulated (international normalized ratio >1.8 and <3), and informed consent to participate had to be provided. Exclusion criteria were a support duration less than 1 month and unstable status with need of inotropes. The trial was approved by the Ethics Committee of Copenhagen in accordance with the Helsinki declaration (Project no. H-3-2013-010) and registered at ClinicalTrials.gov (Identifier: NCT01851889).
Measurements and Interventions
In the study period (May 29, 2013 until December 20, 2013), all included CF-LVAD patients underwent ramp testing. Ten patients were included in the study. The ramp protocol started at usual pump setting (ramp-base) and then went from 8,000 RPM (ramp-low) increasing by 400 RPM/5 minutes until reaching 12,000 RPM or suction/arrhythmic event (ramp-high) as used in previous studies.2
The 12 ramp steps in the protocol (baseline and 8,000–12,000 RPM in +400 RPM/step) each lasted 5 minutes—2 minutes to reach equilibrium at the new pump setting followed by 3 minutes of measurements. At each step, hemodynamics and echocardiographic dimensions were recorded simultaneously by the same two experienced physicians, who were both blinded from their colleagues’ data by monitors facing away from each other.
The study was finalized by a light exercise test (25 Watt) where hemodynamic measurements and echocardiographic dimensions were collected. Exercise was performed on a tilt-table ergometer bike (ebike; GE Healthcare, Fairfield, CT), permitting measurements in supine position. Data were measured pre-exercise (pre-exer) in the supine position as well as during exercise at two ramp steps: “baseline pump speed” (exer-base) and “baseline with +800 RPM” (exer-inc). Workload was constant (25 Watt), and each of the two exercise steps was 4 minutes long—1 minute to reach equilibrium and 3 minutes of measurements. Current studies are investigating the possibility of enhancing functional capacity in future generation LVADs with automatic speed change functions in response to loading conditions. Thus pressure–volume relations during exercise with increased pump speed are of general interest, and the chosen increase of +800 RPM in this study has previously been shown clinically safe.5,6
Central pressures, CO, and mixed venous blood oxygen saturation (SvO2) were measured by right heart catheterization (RHC), via the right internal jugular vein, using a Swan-Ganz catheter (Edwards Lifesciences, Irvine, California). Cardiac output was measured three times at each ramp step by standard thermodilution with cold water injection, and the average result was recorded.
Echocardiography was performed, as part of the trial protocol, according to the current guidelines3 on an Philips iE33 cardiac ultrasound system (Philips Healthcare, Best, The Netherlands). During each level of the ramp test, pulsed and continuous wave Doppler images were acquired in the parasternal long-axis view. Left ventricular dimensions were measured from parasternal M-mode images. Aortic valve opening was assessed using M-mode over the AV in the parasternal long-axis view. At least 10 consecutive cardiac cycles were reviewed, and the frequency of AV opening recorded as the percentage. Aortic regurgitation was classified as mild, moderate, or severe.7
In addition to heart rate, mean arterial blood pressure (MAP) was measured with a noninvasive blood pressure cuff at each ramp step. Systemic vascular resistance (SVR) and pulmonar vascular resistance (PVR) were subsequently calculated as SVR = ((MAP − central venous pressure)/CO)) and PVR = ((mean pulmonary arterial pressure − PCWP)/CO)). The pressure difference across the pump (ΔPP) was estimated as ΔPP = MAP − PCWP, which we found appropriate as no severe mitral regurgitation or shunts were present in the cohort.
Near-infrared spectroscopy (NIRS) (INVOS Oximeter; Somanetics Corp., Troy, MI) was used to measure tissue oxygenation cerebrally (ScO2) and peripherally (SmO2).
Data were examined using SAS 9.3 Statistical Software (Cary, NC). To estimate the association between pressure and pump speed for all patients, we plotted PCWP/RPM, which showed pressure to be linearly related to RPMs. To examine the association between PCWP and LVEDD in the individual subject, PCWP was plotted against LVEDD for each patient with a fitted linear regression line, and the correlation was calculated. To be able to compare our data with former ramp studies, we also computed both LVEDD and PCWP slopes for each patient with pump speed steps given numeric values (8,000 RPM = 1, 8,400 RPM = 2, 8,800 = 3 etc.) as used previously.2 To estimate the overall relation between pressure and volume during ramp testing, we subsequently calculated the correlation of the LVEDD and PCWP slopes.
Differences between groups were examined using two-sided Student’s t-test, paired or unpaired as appropriate. In all statistical tests used (paired and unpaired t-test and test for nonzero correlation), no major deviations from normality were observed. Data were examined with histograms and qq-plots. Statistically significant results were defined by a two-tailed p value less than 0.05. Data are presented as mean ± standard deviation unless otherwise stated.
Patient characteristics are shown in Table 1. Indication for pump implantation was destination therapy in three patients and bridge-to-transplantation (BTT) in seven patients, overall with a mean support duration of 668 ± 594 days. New York Heart Association functional classification was I to IIIa, assessed by the principal investigator on the day of the study. Mean LV ejection fraction was 11 ± 7% and pro-brain natriuretic peptide was 204 ± 240 pmol/L (ref <24.8). Baseline pump speed (ramp-base) was 9,300 ± 241 RPM, at which 3 of 10 patients had AV opening (2 complete and 1 partial). Ramp-low was 8,000 ± 0 RPM and ramp-high was 12,000 RPM in all except in three patients due to suction at the penultimate step at 11,600 RPM in two patients and nonsustained ventricular tachycardia (VT) at 11,200 RPM in one patient. Lowering pump speed returned LV unloading to normal. Development of suction in patients 3 and 5 was associated with LVEDD decreasing to 39% and 35% of baseline LVEDD (3.9 and 4.1 cm, respectively, at ramp-high) while LVEDD remained relatively unchanged in patient 9 (5% decrease and 6.1 cm at ramp-high) before VT. Patients without event reaching 12,000 RPM in ramp-high had an average LVEDD decrease of 17 ± 19% (final LVEDD mean 5.8 ± 1.8 cm). Pulmonary capillary wedge pressure at ramp-high did not differ between patients with and without event (7 ± 0.8 vs. 6.4 ± 3.4 mm Hg).
Figure 1 shows the relation between PCWP and LVEDD in each patient and the correlations, which varied from R2 = 0.24 to 0.92, are stated. From ramp-low to ramp-high, PCWP decreased 13 ± 4 mm Hg and LVEDD decreased 1.1 ± 1.2 cm. At ramp-low, ramp-base, and ramp-high, PCWP was 20 ± 4, 14 ± 4, and 7 ± 3 mm Hg (p < 0.001 for all comparisons) and LVEDD 6.6 ± 1.0, 6.7 ± 0.9, and 5.5 ± 1.7 cm (p < 0.05 for all comparisons but ramp-low versus ramp-base).
Values of the slopes depicting the change in PCWP and LVEDD for each individual patient are shown in Table 2. The overall correlation between slopes was positive and statistically significant, R2 = 0.53 (p = 0.02) (Figure 2). If changes in pressures were correlated to squared changes in LVEDD—recognizing that such a relation may not be linear—a better correlation was not demonstrated, R2 = 0.34 (p = 0.08). Subjects with ischemic cardiomyopathy (ICM) (n = 4) had a correlation between PCWP and LVEDD slopes of R2 = 0.70 (p = 0.16) vs. R2 = 0.59 (p = 0.07) in non-ICM subjects (n = 6).
Hemodynamic responses as a function of pump speed are shown in Table 3. Cardiac output increased by 1.79 ± 1.77 L/minute from ramp-low to ramp-high (4.51 ± 1.11 vs. 6.29 ± 1.37 L/minute, p = 0.014). None of the trial participants had significant tricuspid regurgitation, at most mild. Mean arterial blood pressure, SVR, and PVR did not change significantly from ramp-low to ramp-high, whereas CVP decreased by 1.3 ± 0.7 mm Hg (10 ± 3 vs. 9 ± 3 mm Hg; p = 0.004) and ΔPP increased 20 ± 24 mm Hg (65 ± 6 vs. 85 ± 25 mm Hg, p = 0.026). Baseline ΔPP (72 ± 15 mm Hg) was not related to LVEDD slope (p = 0.55), whereas baseline CVP (9 ± 3 mm Hg) was (p = 0.04). Complications after RHC were not encountered except for a minor hematoma in one patient.
The AV was completely or partly open in 7 of 10 patients at ramp-low of which the AV subsequently closed in five patients at 9,440 ± 408 RPM with a mean PCWP of 18 ± 3 mm Hg. The AV remained open throughout the ramp test in two patients. At steps with AVs open, ΔPP was 71 ± 8 mm Hg vs. 83 ± 19 when AVs were closed (p = 0.0001).
SvO2 increased from 60.0 ± 7.2% to 70.0 ± 4% from lowest to highest pump speed (p = 0.003), cerebral tissue oxygenation measured by NIRS increased from 55 ± 11% to 58 ± 10% (p = 0.07), whereas tissue oxygenation peripherally was unchanged; 55 ± 9% to 55 ± 14% (p = 0.96).
Hemoglobin was in the lower range of normal (8.2 ± 1.3 mmol/L). Haptoglobin was below detection threshold in all, whereas lactate dehydrogenase (LDH) was elevated in all, highest in patient 2. Plasma free hemoglobin was normal for all except for a slight elevation in two patients.
Table 4 shows exercise-induced changes. During light exercise, CO increased significantly when exercising with +800 RPM (exer-inc) compared with exercise at baseline pump speed (exer-base), 7.7 ± 1.8 vs. 6.6 ± 1.6 L/minute; delta CO (inc versus base) 1.1 ± 1.0 L/minute (p = 0.017), whereas the exercise-induced increase in PCWP was not affected by increasing pump speed. Pulmonary capillary wedge pressure increased 6 ± 4 mm Hg from pre-exercise to exer-base (p = 0.001), whereas LVEDD decreased by 0.1 ± 0.3 cm (p = 0.38), correlation between changes, R2 = 0.26 (p = 0.13).
Patient 2 in the study was clinically suspected for pump thrombosis and presented with hemoglobinuria. Ramp testing revealed an LVEDD slope of −0.03 with MAP 69 mm Hg, constantly opening AV, high pulsatility index, LDH elevation (>4 × upper reference), and plasma free hemoglobin above upper limit. Substantial ventricular recovery subsequently allowed the patient to be explanted where thrombosis in the outflow cannula was confirmed visually. Pulmonary capillary wedge pressure in this patient, at 8,000 and 12,000 RPM, was 13 and 6 mm Hg, respectively. Excluding this patient with pump thrombosis did not change the result of the correlation analysis between PCWP and LVEDD slopes significantly (R2 = 0.50; p = 0.03). Last, omitting patient 2 decreased average LDH (Table 1); 409 ± 126 U/L (460 ± 193 U/L) but did not impact on the remaining biochemistry.
The 10 ramp tests resulted in pump speed regulation in three (increase by 467 ± 249 RPM) and adjustment in medical therapy in four patients (increased loop-diuretics in one, increased beta-blocker in one, and adjusted phosphodiesterase 5 inhibitor dose in two). Patients (4, 5, and 10) had pump speed increased after clinically assessing that an increased CO would optimize the circulatory support without risking suction or arrhythmic events. The three patients were BTT and the AVs were already closed at baseline pump speed before upregulating RPMs. Patients (5, 7, 8, and 10) had medical therapy adjusted in response to pressures indication overhydration and to regulate systemic arterial hypertension and afterload.
In this study, the correlation between pressure and volume changes as a function of pump speed in CF-LVAD patients was investigated. To our knowledge, it is the first study to report directly on this relation.
We found that changes in PCWP as a function of RPM, which is the key parameter to describe LV unloading, was only weakly positively correlated with changes in LVEDD and that the difference in the relation was large between patients. This indicates that both intrinsic and external factors have to be taken into account when optimizing settings of LVAD devices with echocardiography. Thus, factors such as ventricular elastance, pericardium, pleural pressure, afterload, and volume status must be considered when analyzing LVEDD changes as a function of pump speed.
In some cases, a constant unchanged LVEDD as a function of RPM is suspected to be attributed to obstruction of pump flow, either due to thrombosis or cannula displacement/kinking; however, in this case, both LV pressure and volume unloading will be affected resulting in flat/less steep PCWP and LVEDD slopes. Recently, it has been recommended to use echo-guided ramp protocols when optimizing CF-LVAD pump speed and as a guidance to confirm pump thrombosis, the latter being an adverse event requiring immediate intervention.8–10 Using a novel ramp protocol, Uriel et al.2 found that pumps with obstruction to flow were associated to an LVEDD slope of −0.16 or greater (corresponding to a less steep response to pump speed than expected). In the study, however, only pumps suspected of thrombosis were inspected for clot validation postexplant, thus not confirming the internal appearance of the “normal pumps” with LVEDD slopes less than −0.16. Our findings differ from those of Uriel et al. as 6 out of 10 had an LVEDD slope of −0.16 or greater of which only one was clinically suspected of pump thrombosis. The remaining five patients did not present with heart failure symptoms, biochemical markers indicating hemolysis, or power spikes/increasing power requirements. Furthermore, in these patients, progressive LV unloading with increasing pump speed was confirmed with a decrease in PCWP. The implication of this finding is that LVEDD slope cannot be viewed as a single diagnostic test to detect pump dysfunction. Thus, in some cases, RHC might help guide the physician in distinguishing, for example, an LV with low compliance from a malfunctioning pump and, in turn, which clinical action to launch.
That both intrinsic and external factors should be taken into account, when analyzing LVEDD/RPM, is in keeping with the findings of Adataya et al.,11 who recently presented two case reports showing that changes in loading conditions can affect the individual LVEDD/RPM slope. We also found that loading conditions in the form of baseline central venous pressure were related to the LVEDD slope.
SvO2, measured in the pulmonary artery, increased 10% from lowest to highest pump speed, whereas tissue oxygenation did not change significantly in the same RPM range. SvO2 describes the balance between oxygen delivery and oxygen consumption and will increase in response to enhanced CO, hemoglobin, and saturation, whereas increased oxygen consumption will diminish SvO2. Thus, the significant rise in SvO2 was expected, as CO and, in turn, pulmonary perfusion increased in response to higher pump speeds throughout the ramp tests, whereas oxygen requirements remained stationary (resting patient). SvO2 does not reflect any particular tissue oxygenation, but the overall global oxygen extraction and cannot exclude tissue hypoxia or hyperoxia in individual organs.12 Formation of intestinal arteriovenous malformations in the course of continuous-flow support could be hypothesized to contribute to the lack of increase in tissue oxygenation, both in cerebral tissue and skeletal musculature, even though CO increased, as seen in the current study. Indeed, in an animal study, Tuzun et al.13 found evidence of intestinal arteriovenous shunting opening under continuous-flow without intestinal tissue perfusion defect when comparing baseline pump settings with low and high continuous-flow modes. Another possibility is a Type II error, where NIRS failed to detect a difference in tissue oxygenation in the different flow situations. Limitations associated with NIRS measurements are well described including insufficient light shielding, optode displacement,14 and inaccurate measurements associated with changes in blood flow velocity and erythrocyte shape and position when increasing pump speed. Finally, peripheral cardiovascular factors associated with advanced HF, including deconditioned skeletal muscle, endothelial dysfunction, and changes in the neurohormonal system might challenge accurate NIRS measurements.
Our study is limited by its small sample size and thus the results must be read with appropriate caution. The results may be viewed as the first pilot study examining the pressure–volume relations in CF-LVAD patients during ramp testing. Clearly, our findings must be confirmed in larger prospective cohorts where the relative influence of the different covariates can be estimated statistically. However, due to the rigorous setup of the study, we feel that our results are sufficiently valid to call for caution in the interpretation of echo-guided ramp protocols. The study was prospective, and the two physicians measuring hemodynamics and echocardiographic dimensions were both blinded. Due to our small center size, all patients fitting the inclusion criteria were enrolled, thus omitting selection bias. Last, we waited 2 minutes at each ramp step before taking any measures and, thus allowing time for equilibrium at the new speed step to be reached.
In conclusion, LVEDD is not an accurate measure of unloading in patients supported with CF-LVADs. Changes in LVEDD during an echo-guided ramp study should be interpreted with caution as factors other than LVAD function may determine the relation between pump speed and LV volume.
The authors thank ventricular assist device coordinator Lene Larsen for valuable help in facilitating practical concerns and for controlling the power base unit during the studies.
1. Slaughter MS, Pagani FD, Rogers JG, et al. Clinical management of continuous-flow left ventricular assist devices in advanced heart failure
. J Heart Lung Transplant. 2010;29:S1–39
2. Uriel N, Morrison KA, Garan AR, et al. Development of a novel echocardiography ramp test for speed optimization and diagnosis of device thrombosis in continuous-flow left ventricular assist devices: The Columbia ramp study. J Am Coll Cardiol. 2012;60:1764–1775
3. Estep JD, Stainback RF, Little SH, Torre G, Zoghbi WA. The role of echocardiography and other imaging modalities in patients with left ventricular assist devices. JACC Cardiovasc Imaging. 2010;3:1049–1064
4. Kato TS, Colombo PC, Nahumi N, et al. Value of serial echo-guided ramp studies in a patient with suspicion of device thrombosis after left ventricular assist device
implantation. Echocardiography. 2014;31:E5–E9
5. Brassard P, Jensen AS, Nordsborg N, et al. Central and peripheral blood flow during exercise with a continuous-flow left ventricular assist device
: Constant versus
increasing pump speed: A pilot study. Circ Heart Fail. 2011;4:554–560
6. Jung MH, Hansen PB, Sander K, et al. Effect of increasing pump speed during exercise on peak oxygen uptake in heart failure
patients supported with a continuous-flow left ventricular assist device
. A double-blind randomized study. Eur J Heart Fail. 2014 doi: 10.1002/ejhf.52
7. Lancellotti P, Tribouilloy C, Hagendorff A, et al. Recommendations for the echocardiographic assessment of native valvular regurgitation: An executive summary from the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging. 2013;14:611–644
8. Kirklin JK, Naftel DC, Kormos RL, et al. Fifth INTERMACS annual report: Risk factor analysis from more than 6,000 mechanical circulatory support patients. J Heart Lung Transplant. 2013;32:141–156
9. Shah P, Mehta VM, Cowger JA, Aaronson KD, Pagani FD. Diagnosis of hemolysis and device thrombosis with lactate dehydrogenase during left ventricular assist device
support. J Heart Lung Transplant. 2014;33:102–104
10. Fine NM, Topilsky Y, Oh JK, et al. Role of echocardiography in patients with intravascular hemolysis due to suspected continuous-flow LVAD thrombosis. JACC Cardiovasc Imaging. 2013;6:1129–1140
11. Adataya S, Masri C, John R, Eckman P. Loading conditions influence reliability of the echocardiographic RAMP test in continuous-flow left ventricular assist devices. J Heart Lung Transplant. 2013;32:1142–1144
12. Cariou A, Monchi M, Dhainaut JF. Continuous cardiac output and mixed venous oxygen saturation monitoring. J Crit Care. 1998;13:198–213
13. Tuzun E, Chorpenning K, Liu MQ, et al. The effects of continuous and intermittent reduced speed modes on renal and intestinal perfusion in an ovine model. ASAIO J. 2014;60:19–24
14. Madsen PL, Secher NH. Near-infrared oximetry of the brain. Prog Neurobiol. 1999;58:541–560
Keywords:Copyright © 2015 by the American Society for Artificial Internal Organs
left ventricular assist device; hemodynamics; assisted circulation physiology; heart failure