Martina, Jerson R.*; Westerhof, Berend E.†‡; de Jonge, Nicolaas§; van Goudoever, Jeroen‡; Westers, Paul¶; Chamuleau, Steven§; van Dijk, Diederik‖; Rodermans, Ben F. M.#; de Mol, Bas A. J. M.**; Lahpor, Jaap R.*
Continuous-flow left ventricular assist devices (cf-LVADs) are the latest generation mechanical circulatory support systems in the treatment of end-stage heart failure.1–3 These devices are used as a bridge to heart transplantation or recovery or as destination therapy. Presently, they can provide long-term mechanical circulatory support allowing patients to be completely ambulatory.
Particularly, knowledge of arterial blood pressure is important in the management of patients with cf-LVADs as it gives valuable information on organ perfusion pressure and about the interaction between the cf-LVAD and the cardiovascular system. Because cf-LVADs generate flow continuously throughout the cardiac cycle, blood pressure pulsatility is reduced compared to patients without a cf-LVAD. Both the flow through the cf-LVAD and the degree of arterial pulsatility depend on the preload and afterload pressures, pump speed, and left ventricular contractility. As the pressure difference between preload and afterload affects pump output, arterial hypertension can reduce cf-LVAD flow and thereby the degree of ventricular unloading provided by the cf-LVAD. Maintaining an appropriate range of the mean arterial pressure (MAP) could also prevent stroke due to hypertension.4
For the evaluation of patient cardiovascular status during cf-LVAD support, echocardiography has been the primary assessment strategy, in terms of ventricular size, cardiac hemodynamics, and myocardial recovery assessment.5 Often, these evaluations are done through a pump speed change procedure, in which the pump speed is changed in steps. Such a procedure is associated with characteristic echocardiographic responses.6,7 Currently, there is a lack of knowledge regarding the association between echocardiography-derived parameters and arterial blood pressure measurements. Particularly, measurement of blood pressure in patients with a cf-LVAD is difficult because the pulsatility is usually below the level required to measure blood pressure with auscultatory or automated methods.8,9
The current study evaluates the accuracy of noninvasively measured arterial blood pressure waveforms by comparing these with invasive arterial pressure (IAP) after cf-LVAD implantation. Furthermore, we sought to investigate the association between changes in left ventricular size and function derived by echocardiography and arterial blood pressure responses in patients with a cf-LVAD during a pump speed change procedure.
The study was conducted at the University Medical Center Utrecht. The investigation conforms to the principles outlined in the Declaration of Helsinki. The study was approved by the medical ethical committee of the hospital and all participating patients provided written informed consent. Two cohorts of patients were investigated. Cohort A consisted of patients in the intensive care unit (ICU) after receiving cf-LVAD as a bridge to heart transplantation. Of these patients, most (n = 29) received a Heartmate II LVAD (HM II, Thoratec Corp, Pleasanton, CA) while two patients received a Heartware LVAD (HVAD, Heartware, Inc., Miramar, FL). Both devices are continuous-flow pumps and have similar flow properties, resulting in similar physiology during circulatory support. Arterial blood pressure was measured invasively, which served as a reference for the noninvasive measurements. During these measurements, patients were on low dosages of dobutamine (n = 14), milrinone (n = 7), and norepinephrine (n = 4) and vasodilatative drugs (n = 16) including enalapril, captopril, or sildenafil. The fluid balance at the time of measurement was 470 ± 630 ml (range, −980 to 2,500 ml). Additionally, cohort B consisted of outpatients (n = 30) with a HM II cf-LVAD who underwent a pump speed change procedure 6 months post-HM II implantation. This procedure is based on a protocol in which the pump speed of the HM II is reduced in stepwise fashion from 9,000–10,000, to 6,000 rpm by steps of 1,000 rpm for 1 minute per step. Patients were heparinized to limit the risk of pump thrombosis.
Arterial Blood Pressure Waveform Validation
Patients in cohort A received IAP monitoring in the ICU with a radial artery catheter connected to a pressure transducer (PMSET 1DT-XX 1 Safedraw Argon Critical Care Systems, Singapore) and a monitoring system (Spacelabs Healthcare, Issaquah, WA). To avoid inadequate invasive arterial blood pressure wave shape, the invasive arterial catheter was periodically flushed. Simultaneously, noninvasive arterial pressure (NAP) was measured beat-to-beat by the Nexfin monitor (BMEYE BV, Amsterdam, The Netherlands) (Figure 1).10,11
Noninvasive arterial pressure is based on a brachial reconstruction from the finger arterial pressure measured by the finger cuff. To obtain NAP, an appropriate finger cuff size was applied to the mid-phalanx of a finger of the hand on the side of the invasive measurements. To compensate for hydrostatic errors due to level differences between the finger and the heart, the heart reference system of the Nexfin monitor was used. The finger side of the heart reference system was fixated next to the measurement finger and the heart side at the arterial pressure transducer level.12 In the first minute of measurement, the positions of the finger cuff and pressure transducer were checked for possible level errors and signal quality, and if necessary, readjusted. If no NAP signal could be obtained, the cuff was switched to another finger and NAP measurement was reinitiated. Noninvasive arterial pressure is periodically calibrated by the Physiocal algorithm.13 During these short calibrations, NAP is temporarily interrupted for two (or more if required) beats. The analog signals of NAP and IAP were sampled at 200 Hz and stored on a hard disk. For each patient recording, 600 seconds periods of matched NAP and IAP values were selected. Systolic pressure (SYS), diastolic pressure (DIA), mean arterial pressure (MAP), and pulse pressure (PP) values were determined for subsections of 2 seconds (since pulsatility may or may not be present in the blood pressure) of the IAP and NAP recordings, resulting in 300 consecutive datasets per patient.
Pump Speed Change Procedure
Patients in cohort B underwent a pump speed change procedure, 6 months post-LVAD implantation. Throughout the procedure, blood pressure responses were measured at each pump speed, including changes in SYS, DIA, MAP, PP, and the maximum arterial pressure time derivative ([dPart/dt]max) based on the NAP waveform. Meanwhile, patients underwent simultaneous 2D echocardiographic examination using a Philips iE33 machine (Philips Electronics, Eindhoven, The Netherlands). Left ventricular end-diastolic dimension (LVEDD) and left ventricular end-systolic dimension (LVESD) were obtained from the parasternal long axis of the left ventricle at the level of the chordal mitral valve junction. Based on these measurements, left ventricular shortening fraction (LVSF) was computed as percentage change in the LV diastolic dimensions with systole ([LVEDD − LVESD]/LVEDD). Images were stored digitally for offline analysis using Phillips Xcelera analysis software R3.2L1 (Phillips Healthcare, Amsterdam, The Netherlands). Furthermore, aortic valve opening or continuous closure was assessed at all pump speed settings. As long as the aortic valve still opened, patients were denoted as being on partial support. When the aortic valve remained closed, patients were denoted as being on full support. Measurements were done routinely by four experienced laboratory technicians and verified by an independent cardiologist.
Data Analysis and Statistics
To compare SYS, DIA, MAP, and PP from NAP and IAP in cohort A, the Bland and Altman approach was followed. For each patient, the mean and standard deviation (SD) of the 300 averages of NAP and IAP were represented as horizontal coordinate where the SD expressed the range of pressures for each patient (“within-subject variability”). Similarly, the mean and SD of the 300 NAP-IAP differences of each patient represent the vertical coordinate in the Bland–Altman plot where the SD were measures of the individual consistence (“within-subject precision”).10 The Pearson product moment was used to compute coefficients of correlation. To evaluate the accuracy and precision of the NAP, mean ± SD of the differences between NAP and IAP were computed based on the data of the group.
Blood pressure measurements after pump speed change in cohort B were analyzed based on 10 second averages of NAP for each speed setting. Furthermore, the NAP waveform was evaluated for the presence of a dicrotic notch at each pump speed setting. The dicrotic notch was indicated as a slight pressure increase on the descending limb following the maximum arterial pressure.
To examine the association between the echocardiographic and arterial blood pressure parameters, correlation analysis was performed. Receiver operating characteristic (ROC) curves were made to evaluate the specificity and sensitivity of the presence of a dicrotic notch, PP, and (dPart/dt)max to reflect the dynamics of the aortic valve. A Student’s t-test was performed to compare parameters based on patients with partial support at 9,000 rpm (partial support group) and patients with full support at 9,000 rpm (full support group). A one-way repeated measure analysis of variance was performed to analyze the differences between blood pressures and echocardiography parameters at multiple pump speeds. Statistical tests were performed using SPSS 17 (IBM SPSS Statistics 19, IBM Corporation, Somers, NY). A p-value of less than 0.05 was considered significant. Normally distributed continuous data were presented as means ± SD.
NAP Compared to IAP
Cohort A consisted of 23 men and 8 women (total of 31 patients) of 50 ± 11 years of age. Other patients’ characteristics were a height of 177.8 ± 9.5 cm, weight of 76.5 ± 13.7 kg, and body mass index of 24.1 ± 3.5 kg/m2. The mean cf-LVAD flow during measurement was 5.3 ± 1.0 L/min at baseline pump speed (9,500 ± 260 rpm in patients with a HM II device [n = 29] and 3,000 ± 0 rpm in the patients with a HVAD [n = 2]). Noninvasive measurements failed in two patients (1 HM II and 1 HVAD). Thus, data of 29 patients were available for analysis (93% success rate).
The ranges of IAP were 89–116 mm Hg for SYS, 62–84 mm Hg for DIA, 76–93 mm Hg for MAP, and 14–48 mm Hg for PP. The Pearson coefficients of correlation, within-subject variability, and within-subject precision are represented by Table 1. The individual patients’ averages and SDs are shown in the Bland–Altman plots for SYS, DIA, MAP, and PP, respectively (Figure 2). Noninvasive arterial pressure underestimated IAP by less than 10% as shown by Table 2.
NAP Compared to Echocardiography
Patient characteristics and medication in cohort B are shown in Table 3. All characteristic and medication were similar, except for the (dPart/dt)max and the pulsatility index of the HM II LVAD, which were higher in the patients with partial support. None of the patients showed hemodynamic deterioration during the pump speed change procedure.
In general, correlation analysis showed no association between LVEDD, LVESD, or LVSF and noninvasively assessed arterial blood pressure data (Table 4). Left ventricular shortening fraction correlated poorly with PP (r = 0.24; p = 0.005) and (dPart/dt)max (r = 0.25; p = 0.004).
The arterial blood pressure waveform showed characteristic changes in the course of the pump speed change procedure (Figure 3). When the aortic valve did not open, no dicrotic notch was observed on the NAP waveform (specificity >0.99). Subsequently, when the aortic valve truly opened, a dicrotic notch was observed on NAP waveform in 74% of these cases (sensitivity = 0.74), considering all pump speed settings. The ROC curves of Figure 4 show the relationship between the presence of a dicrotic notch, PP, and (dPart/dt)max and true opening of the aortic valve. The area under the curve (AUC) for the presence of the dicrotic notch was 0.87, which can be categorized as a good predictor of aortic valve opening. However, the AUC for PP was 0.64 and (dPart/dt)max was 0.61, respectively. This indicates that both PP and (dPart/dt)max were poor indicators of aortic valve opening.
When comparing the patients in the partial support group (n = 10) with those in the full support group (n = 20), LV dimensions and LVSF responses were not significantly different (Figure 5). Nevertheless, patients with partial support had a significant rise in SYS (p < 0.01), while DIA and MAP were reduced in both groups. In addition, PP and (dPart/dt)max showed a much more prominent rise (both p < 0.001) in the partial support group compared to the full support group while lowering pump speed.
The current study shows the feasibility of simultaneous measurement of noninvasive arterial blood pressure waveforms and echocardiography in patients with cf-LVAD support. The Nexfin monitor provided beat-to-beat NAP data that slightly underestimated IAP, while pulse pressures were accurate. Furthermore, left ventricular dimensions, function, and blood pressure responses were not correlated with one another. The dicrotic notch on the noninvasive blood pressure waveform was a good predictor of aortic valve opening, better than PP and (dPart/dt)max. Meanwhile, patients with partial support at 9,000 rpm had a more prominent rise in SYS, PP, and (dPart/dt)max compared to those at full support while lowering the pump speed.
The Nexfin monitor allowed reliable continuous noninvasive assessment of the arterial blood pressure waveform. Previously, we demonstrated that the Nexfin monitor was capable of measuring arterial blood pressure with reduced pulsatility during cardiopulmonary bypass with good accuracy.14 The Association for the Advancement of Medical Instrumentation has suggested an allowed average pressure difference of ±5 mm Hg and an SD of 8 mm Hg for comparison between the upper arm cuff devices and invasive arterial blood pressure.15 Up till now, no guidelines are available for comparison between invasive and noninvasive continuous blood pressure waveform measurements, but with above standards in mind the noninvasive device gives acceptable results.
In this study, NAP underestimated IAP, while pulse pressures were similar during cf-LVAD support. The differences in continuous noninvasive pressures observed may be related to structural properties within the arterial blood pressure reconstruction algorithm used by the Nexfin monitor. Arterial blood pressure reconstruction involves three basic features.10 First, NAP is generated based on pressure reconstruction. Waveform filtering by an inverse model compensates for arterial blood pressure amplification, caused by pressure wave reflections due to narrowing of arteries of the arm. Second, a population-based level correction formula compensates for frictional losses between the radial arteries down to the finger arteries. Third, pressure calibrations are performed regularly to compensate for changing properties of the vasculature. Calibration procedures were developed with models and criteria derived under (patho) physiological conditions and for a general population, however not for patients with cf-LVADs. A reason for the differences between IAP and NAP could be that the reduced pulsatility in patients with cf-LVAD causes vascular structural changes which need to be accounted for within the pressure reconstruction algorithms applied to generate NAP.
Particularly, Doppler ultrasound of a peripheral artery has been considered an alternative to assess noninvasive blood pressure during cf-LVAD support and showed a relatively high success rate of 94.3% compared to the automatic blood pressure cuff (52.9%), auscultation (14.3%), and palpation (2.9%).16 When compared to invasive blood pressure, MAP by Doppler differed by 0.2 ± 10.5 mm Hg while SYS was overestimated by 8.6 ± 9.5 mm Hg. In a study in patients with the Jarvik 2000 cf-LVAD reported by Myers et al.,8 pressures using the automatic pressure cuff were not obtained in 24 of the 108 occasions (success rate of 78%). The SYS, DIA, and MAP values of the arterial catheter did not correlate with values of the automatic cuff. Nevertheless, PP, when high enough to be detected by the automatic cuff, was not significantly different from invasively measured PP. In comparison to these results, the Nexfin monitor had a success rate of 93% in cohort A and 100% in cohort B. Meanwhile, NAP reflects the entire waveform, with good predictability of aortic valve dynamics through the dicrotic notch.
Both arterial blood pressure assessment and structural characteristics by echocardiography provide vital information to optimize cardiac support in cf-LVAD patients. On one hand, adequate blood pressure levels should be maintained to preserve tissue perfusion, while avoiding high blood pressure, as higher afterload limits pump flow and ventricular unloading and may also induce stroke.9 On the other hand, echocardiography allows structural evaluation of the ventricle, including ventricular unloading, right ventricular function, and valvular pathology, all which play an important role in achieving optimal cardiac support on cf-LVADs. Thus, this suggests that both measurements provide complementary information.
Blood pressure and echocardiography may also be considered to assess suboptimal cf-LVAD support.17 A flat arterial blood pressure waveform with a low PP indicates that left ventricular function is extremely poor or that the pump speed is close to exceeding the available left ventricular volume. On the other hand, a sudden increase of arterial pulsations accompanied by increased ventricular dimensions may reveal the presence of an obstructed cf-LVAD. As NAP waveforms by the Nexfin monitor provided high specificity and moderate sensitivity to reflect the dicrotic notch during aortic valve opening, these measurements could be used for the assessment of aortic valve dynamics independent of echocardiography. Particularly during the regular office visit without planned echo evaluation, the presence of the dicrotic notch on the NAP may be very useful, while the absence of it would require further investigation.
Left ventricular assist device–induced myocardial recovery is a desirable outcome and a multisource assessment strategy to determine the ventricular function could be necessary. The weak association of LVSF and PP and (dPart/dt)max response found in this study perhaps suggests some complementary relationship in reflecting left ventricular systolic functional status. The ability of the ventricle to open the aortic valve at a fixed pump speed, for example, at 9,000 rpm, may be a novel surrogate of left ventricular systolic performance.18 It is plausible that patients with partial support and aortic valve opening may have a better left ventricular function compared to those showing continuous aortic valve closure at the same rotational speeds. Finally, the choice of the pump speed should be such that the ventricle is maximally unloaded; however, not to an extent that the aortic valve remains permanently closed, since aortic valve insufficiency or fusion of the cusps may ensue.19 Altogether, these facts underline the importance of functional evaluation of the aortic valve during cf-LVAD support, which may incorporate both blood pressure waveforms and echocardiography.
A limitation of this study is that we did not perform pump speed change procedures in the ICU to limit risk to the patient during this critical postoperative period. On the other hand, NAP was not compared with IAP during the pump speed change procedures 6 months after cf-LVAD implantation, since invasive measurement during this outpatient evaluation study is not part of standard practice. Hence, 19 patients in cohort A had been included in cohort B. This study also did not compare the Nexfin monitor with other blood pressure modalities (such as blood pressure cuff or Doppler ultrasound). We also could not evaluate the effect of medication on the accuracy of measurement by the Nexfin monitor due to the diversity of drugs administered to each patient during the postoperative period. Meanwhile, the possible effect of the fluid balance was not studied. Finally, echocardiography data were limited to ventricular dimensions rather than ventricular volume.
The Nexfin monitor provided beat-to-beat NAP that slightly underestimated IAP, while PPs were similar. Noninvasive arterial pressure reflected the blood pressure response to pump speed changes and waveform characteristics reflected aortic valve dynamics. However, arterial blood pressure showed no correlation with echocardiographic measures. As a critical component of assessment in patients on cf-LVAD support, Nexfin can provide reliable measurement of the blood pressure waveform and thereby acts as a compliment to the assessment of these patients.
The authors thank Nico Westerhof for his contribution to the preparation of this manuscript. The Nexfin monitor used in this study is owned by the Department of Cardiothoracic Surgery of the University Medical Center Utrecht and was bought with the support of the Van Ruyven foundation.
1. Slaughter MS, Rogers JG, Milano CA, et al.HeartMate II Investigators. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med. 2009;361:2241–2251
2. Lahpor J, Khaghani A, Hetzer R, et al. European results with a continuous-flow ventricular assist device for advanced heart-failure patients. Eur J Cardiothorac Surg. 2010;37:357–361
3. Pagani FD, Miller LW, Russell SD, et al.HeartMate II Investigators. Extended mechanical circulatory support with a continuous-flow rotary left ventricular assist device. J Am Coll Cardiol. 2009;54:312–321
4. Felix SE, Martina JR, Kirkels JH, et al. Continuous-flow left ventricular assist device support in patients with advanced heart failure: Points of interest for the daily management. Eur J Heart Fail. 2012;14:351–356
5. 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
6. Stainback RF, Croitoru M, Hernandez A, Myers TJ, Wadia Y, Frazier OH. Echocardiographic evaluation of the Jarvik 2000 axial-flow LVAD. Tex Heart Inst J. 2005;32:263–270
7. Jakovljevic DG, George RS, Nunan D, et al. The impact of acute reduction of continuous-flow left ventricular assist device support on cardiac and exercise performance. Heart. 2010;96:1390–1395
8. Myers TJ, Bolmers M, Gregoric ID, Kar B, Frazier OH. Assessment of arterial blood pressure during support with an axial flow left ventricular assist device. J Heart Lung Transplant. 2009;28:423–427
9. 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–S39
10. Martina JR, Westerhof BE, van Goudoever J, et al. Noninvasive continuous arterial blood pressure monitoring with Nexfin®. Anesthesiology. 2012;116:1092–1103
11. Eeftinck Schattenkerk DW, van Lieshout JJ, van den Meiracker AH, et al. Nexfin noninvasive continuous blood pressure validated against Riva-Rocci/Korotkoff. Am J Hypertens. 2009;22:378–383
12. Nexfin HD Operator’s Manual: 0086-06. 2008 Amsterdam, The Netherlands BMEYE BV
13. Wesseling KH, De Wit B, Van der Hoeven GMA, Van Goudoever J, Settels JJ. Physiocal, calibrating finger vascular physiology for Finapres. Homeostasis. 1995;36:67–82
14. Martina JR, Westerhof BE, Van Goudoever J, et al. Noninvasive blood pressure measurement by the Nexfin monitor during reduced arterial pulsatility: A feasibility study. ASAIO J. 2010;56:221–227
15. AAMI. American National Standard for Electronic or Automated Sphygmomanometers. 2002 Arlington VA Association for the Advancement of Medical Instrumentation
16. Bennett MK, Roberts CA, Dordunoo D, Shah A, Russell SD. Ideal methodology to assess systemic blood pressure in patients with continuous-flow left ventricular assist devices. J Heart Lung Transplant. 2010;29:593–594
17. Catena E, Milazzo F, Montorsi E, et al. Left ventricular support by axial flow pump: The echocardiographic approach to device malfunction. J Am Soc Echocardiogr. 2005;18:1422
18. Estep JD, Chang SM, Bhimaraj A, Torre-Amione G, Zoghbi WA, Nagueh SF. Imaging for ventricular function and myocardial recovery on nonpulsatile ventricular assist devices. Circulation. 2012;125:2265–2277
19. Cowger J, Pagani FD, Haft JW, Romano MA, Aaronson KD, Kolias TJ. The development of aortic insufficiency in left ventricular assist device-supported patients. Circ Heart Fail. 2010;3:668–674
Nexfin; echocardiography; aortic valve; rotary pump