Power consumption, flow, and PI are routinely used by clinicians to assess native heart function and to optimize pump settings. However, little is known about their reliability, reproducibility, and variability. Basic information on interpreting these parameters is available in the HM II Operating Manual, but we are not aware of any independent or published evaluations of these parameters. Therefore, we assessed HM II controller parameters and concurrent echocardiography during titrations of RPM to low and high values.
Five HM II patients (Table 2) were studied on three to five occasions each, for a total of 20 titrations. The protocol was approved by the local ethics committee, and patients provided written informed consent.
Vital and laboratory parameters were collected and echocardiography was performed before titration. All patients were taking aspirin and warfarin and an INR > 2.0 was ensured before starting titrations. All patients had an implantable cardioverter-defibrillator (ICD).
Titrations were performed from baseline RPM (8,800–9,400) down to 8,000 RPM and back up to baseline and, on separate occasions, up to 10,400 RPM and back down to baseline. Down-titration may confer risk of thrombus formation. The lower limit of 8,000 is deemed safe without additional anticoagulation (Emma Birks, University of Louisville, personal communication). In one patient, down-titration was performed to 6,000 RPM on four occasions, to assess recovery.7,8 This patient received an intravenous bolus of heparin, 200 unit/kg, and an activated clotting time (ACT) of >400 seconds was ensured before the RPM went below 8,000. Up-titration may confer risk of acute RV failure and of suction and ventricular arrhythmias. One study assessed stress echocardiography in 12 patients with RPM settings of 9,200–10,200 at baseline with up-titrations by 1,000 RPM, without complications.9 The highest RPM in this study was 11,300 RPM (Jacob Møller, Rigshospitalet, Copenhagen, personal communication). We chose 10,400 RPM as the upper limit.
Titration was conducted by increasing/decreasing RPM by 200 RPM, the smallest increments permitted by the HM II controller, approximately every 2 minutes. Occasionally, steps were skipped to assess whether this affected readings. During titration down to 6,000 RPM, numerous steps were skipped to minimize time spent at low RPM with high risk for pump thrombus formation or spontaneous bleeding, given heavy anticoagulation. Systolic or mean blood pressure was measured by Doppler 1 minute after every step, patients had continuous electrocardiogram (EKG) and oxygenation monitoring, and echocardiography was performed at baseline, at minimum/maximum RPM, and upon return to baseline RPM. Patients were monitored an additional hour and then discharged. Symptoms, vital, physical, and laboratory parameters were reassessed 1–3 weeks after titration. During weaning down to 6,000 RPM, the patient was held in the hospital until the ACT spontaneously returned to <200 seconds (protamine reversal was not used).
Parameters during titration were assessed and compared in three ways: visually from the monitor display 1 minute after every step, from data logged in the monitor (saved to a flash card), and manually calculated according to the formulas in Table 1. These data were then used to calculate mean and standard deviation during approximately 2 minutes at each RPM for power, flow, and PI. Manual calculations of flow were compared with flow calculated by the controller and saved to the card (and displayed on the monitor). For multiple patient comparisons of repeated measures, one-way analysis of variance (ANOVA) with repeated measures was used, with the Bonferroni method for confidence interval adjustment. The sphericity assumption for the ANOVA was tested with the Mauchly test of sphericity, and if violated, the Greenhouse-Greisser correction was used. p < 0.05 was considered significant. Statistics was performed in Microsoft Excel 2007 (Redmond, WA) and IBM SPSS Statistics 20 (Armonk, NY).
When observing displayed values for power, flow, and PI during titration, they changed frequently (within a few- second intervals). However, changes were generally small, and the observed snapshot values from the screen-matched mean values captured by the controller. During down-titration, “- - -”was frequently displayed for flow. Thoratec states that “− − −”or “+++” get displayed when calculated flow falls outside the expected operational range or acceptable linear region (HeartMate II Operating Manual).
Figure 2 shows data during titration down to 8,000 RPM and back up to baseline for all patients on all occasions. Power declines and returns to baseline smoothly. One patient, on one but not two other occasions, required substantially more power to maintain the set RPM, but the pattern of decline and increase was the same. Blood pressure on this occasion was similar to the two other occasions, and there were no clinical signs of pump thrombus or malfunction. Flow declines and increases similarly but less linearly and with a mirrored pattern. Pulsatility exhibits even less linear and consistent changes. The sharp drop at 9,000 RPM, without corresponding changes in power or flow, is again seen, for all patients on all occasions. The same drop does not occur, for example, in patients starting at 8,800 and going to 8,600 or at any other interval change in RPM.
During titration up and back down for all patients on all occasions (Figure 3), inverse patterns are seen, again with less consistency for PI than for flow and in turn, power, but with lesser changes per increment.
Figure 4 shows flow calculated by the controller compared with flow calculated manually during titration down to 6,000 RPM (approximating zero forward flow)7 and back up for the patient for whom recovery was assessed, on four occasions. Flow calculated by the monitor levels off at 8,000 RPM and remains flat at 2.6 L/min, without any variation between any of the 15 second interval recordings, for all RPM settings ≤ 8,000, except at the very lowest RPM. Manually calculated flow declines and rises relatively steadily. Flow at 6,000 RPM is supposed to approximate zero,7 and the manually calculated flow resembles this theoretical pattern.
Readings from the monitor may be affected by changes in loading conditions. Therefore, we carefully assessed vital parameters during each step of titration and echocardiography before titration, at minimum and maximum RPM, and when returned to baseline (Table 2). There were no findings suggestive of major changes in loading conditions or inflow obstruction, or serious complications during or after the RPM titrations. There were no alarms other than the expected “low flow” alarm at RPM below 8,000. During echocardiography monitoring, we observed expected changes in aortic valve opening and ventricular septal deviation, but no obvious cannula displacement. During down-titration, left ventricular ejection fraction and left ventricular end-diastolic diameter increased slightly, and vice versa during up-titration, as expected. There were no changes in vital (blood pressure, heart rate, saturation, or weight) or laboratory parameters (electrolytes, creatinine, and N-terminal pro-natriuretic peptide) during or 1–3 weeks after titrations, and there were no clinical events. On telemetry during titration, some patients had increased frequency of premature ventricular contractions and two patients had nonsustained ventricular tachycardia (VT) after titrations (both up- and down-titrations). On ICD interrogation weeks after titrations, one patient was noted to have had numerous episodes of nonsustained VT for 1–2 weeks after weaning (down to 6,000 RPM and back to baseline), but no sustained VTs or ICD antitachycardia pacing bursts or shocks.
We show that during down- and up-titration of HM II RPM, the measured and monitor-displayed values for power are consistent, the calculated and displayed values for flow are consistent at RPM settings > 8,000 RPM, the calculated and displayed values for flow are inconsistent and unreliable at RPM settings ≤ 8,000, and the calculated and displayed values for pulsatility are inconsistent and unreliable throughout all RPM settings and also consistently underestimated specifically at the 9,000 RPM setting, a common setting in clinical practice.
The Thoratec HeartMate II Operating Manual states that flow may be unreliable at low RPM settings but does not address PI. Our findings are important for the growing number of clinicians, including those practicing in community settings, who care for patients with an HM II.
Left ventricular assist devices are now accepted therapy for BTT and DT. One-year survival with modern continuous-flow devices approaches 90%.1–3,10–13 The first human HM II implant was in 2000. It received Conformité Européenne mark in 2005 and Food and Drug Administration approval for BTT in 2008 and for DT in 2010. It has been implanted in more than 8,000 patients worldwide (number from Thoratec Inc.) with excellent results.2,3,10,11 For long-term LVADs in general, the cost is reasonable at US$ 36,000–86,000 per life year or quality-adjusted life year in broad populations.1 Indications, contraindications, and selection criteria have evolved, and considerable data are now available to guide clinicians and patients in their decisions.1
The HM II relies on an algorithm rather than direct measurement of flow. Even small changes in pump speed can produce large changes in flow and alterations in ventricular geometry. We routinely use echocardiography to assess pump settings but quantification of HM II flow is difficult and there is limited published data. Echocardiography has been used in one study aimed at optimizing pump settings for assessment of native heart recovery7 and in one case report assessing device thrombosis.14 The HM II flow has, to our knowledge, not been compared to Fick or indicator dilution methods (except for routine perioperative pulmonary artery catheterization), but the HM II flow estimator has been evaluated intraoperatively and shown to be reliable for trends but not absolute values.15 We confirm that the flow calculated by the controller and saved in the log function and displayed online on the monitor is consistent with regard to trends at RPM settings between 8,200 and 9,400. However, at or below 8,000 RPM, the monitor generally displays “− − −” and flow saved in the log function is inconsistent and unreliable. The RPM is rarely set to 8,000 RPM for the long run; our findings are therefore relevant typically for situations where RPM is down-titrated to, for example, assess recovery. Above 9,400 RPM, flow calculations are more erratic and do not display distinct trends, which is relevant for patients maintained in higher RPM ranges.
Pulsatility index is a parameter provided by the HM II monitor that estimates native heart contribution to flow through the device. This is useful for assessing recovery, optimal unloading, and pre- and afterload conditions. We are not aware of any published studies evaluating the HM II PI. We show that the PI is inconsistent and unreliable throughout a range of RPM and that there is a notable and consistent error in the PI specifically at 9,000 RPM. We cannot determine the reasons for these inconsistencies. It is not caused by snapshot variation, as the PI at each RPM was assessed from numerous recordings over time. The PI is calculated from power consumption, which is consistent, suggesting an error in calculation. Thus, the PI should not be relied upon, especially at 9,000 RPM, where many patients are maintained. Other parameters, such as contractility and aortic valve opening on echocardiography are better markers of native heart systolic function and contribution to cardiac output.
Our study has limitations. The purpose was to assess the reliability of values and trends displayed on the HM II monitor. Therefore, we did not assess absolute values for flow or pulsatility. The H-Q curve dictates that changes in flow may be caused by changes in pressure gradient over the pump, but as Table 1 indicates, there were no major changes in blood pressure or echocardiography during titration to suggest changes in pressure gradient or loading conditions. This suggests that our observations are instead indeed attributable to limitations in the HM II algorithms. Mock circuits have been used in other studies6 but do not model the in vivo contractility and pressure differentials. In- or outflow obstruction may alter pressure gradients or loading conditions but clinical and echocardiography data did not indicate any such events.
In summary, parameters displayed on the HM II monitor should be interpreted with caution. Trends are reliable for power throughout all settings and for flow above 8,000 RPM, but unreliable for flow at or below 8,000 RPM and for PI throughout all settings, especially at RPMs above 9,200 and specifically at 9,000 RPM. Despite the rapid and exciting evolution of LVAD technology, there has been little clinician interest in operative and technical aspects of LVADs once they have been brought to market. Similarly, despite the rapidly growing patient population, there is often limited clinician knowledge of pump physiology and patient and device parameters that are important to monitor. The easy availability of these parameters (requires only connecting the LVAD controller to the display) may lead to an overreliance on these data, and clinicians should be encouraged instead to perform rigorous clinical, echocardiographic, and hemodynamic assessments.
1. Lund LH, Matthews J, Aaronson K. Patient selection for left ventricular assist devices. Eur J Heart Fail. 2010;12:434–443
2. Miller LW, Pagani FD, Russell SD, et al. Use of a continuous-flow device in patients awaiting heart transplantation. N Engl J Med. 2007;357:885–896
3. Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med. 2009;361:2241–2251
4. Griffith BP, Kormos RL, Borovetz HS, et al. HeartMate II left ventricular assist system: From concept to first clinical use. Ann Thorac Surg. 2001;71(suppl 3):S116–S120 discussion S114
5. Sheikh FH, Russell SD. HeartMate® II continuous-flow left ventricular assist system. Expert Rev Med Devices. 2011;8:11–21
6. Konishi H, Misawa Y, Fuse K, Sohara Y. Pitfalls in the development of a rotary blood pump controller. ASAIO J. 2001;47:397–400
7. George RS, Sabharwal NK, Webb C, et al. Echocardiographic assessment of flow across continuous-flow ventricular assist devices at low speeds. J Heart Lung Transplant. 2010;29:1245–1252
8. George RS, Yacoub MH, Tasca G, et al. Hemodynamic and echocardiographic responses to acute interruption of left ventricular assist device support: Relevance to assessment of myocardial recovery. J Heart Lung Transplant. 2007;26:967–973
9. Andersen M, Gustafsson F, Madsen PL, et al. Hemodynamic stress echocardiography in patients supported with a continuous-flow left ventricular assist device. JACC Cardiovasc Imaging. 2010;3:854–859
10. Starling RC, Naka Y, Boyle AJ, et al. Results of the post-U.S. Food and Drug Administration-approval study with a continuous flow left ventricular assist device as a bridge to heart transplantation: A prospective study using the INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support). J Am Coll Cardiol. 2011;57:1890–1898
11. Rogers JG, Aaronson KD, Boyle AJ, et al. HeartMate II Investigators: Continuous flow left ventricular assist device improves functional capacity and quality of life of advanced heart failure patients. J Am Coll Cardiol. 2010;55:1826–1834
12. 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
13. Strueber M, O’Driscoll G, Jansz P, Khaghani A, Levy WC, Wieselthaler GM. HeartWare Investigators: Multicenter evaluation of an intrapericardial left ventricular assist system. J Am Coll Cardiol. 2011;57:1375–1382
14. Paluszkiewicz L, Gürsoy D, Spiliopoulos S, et al. HeartMate II ventricular assist device thrombosis-an echocardiographic approach to diagnosis: Can Doppler evaluation of flow be useful? J Am Soc Echocardiogr. 2011;24:350.e1–350.e4
15. Slaughter MS, Bartoli CR, Sobieski MA, et al. Intraoperative evaluation of the HeartMate II flow estimator. J Heart Lung Transplant. 2009;28:39–43
Supplemental Digital Content
Copyright © 2012 by the American Society for Artificial Internal Organs