- Email: [email protected]
Multicentre Canadian Experience With the HeartWare Ventricular Assist Device: Concerns About Adverse Neurological Outcomes
Canadian Journal of Cardiology 30 (2014) 1662e1667
Clinical Research
Multicentre Canadian Experience With the HeartWare Ventricular Assist Device: Concerns About Adverse Neurological Outcomes Jamil Bashir, MD,a Jean-Francois Legare, MD,b Darren H. Freed, MD, PhD,c Anson Cheung, MD,a Vivek Rao, MD,d and Mustafa Toma, MDe a
University of British Columbia, Division of Cardiovascular Surgery, Vancouver, British Columbia, Canada b
c
Dalhousie University, Division of Cardiovascular Surgery, Halifax, Nova Scotia, Canada
University of Manitoba, Section of Cardiac Surgery, Department of Surgery, Winnipeg, Manitoba, Canada d e
University of Toronto, Division of Cardiac Surgery, Toronto, Ontario, Canada
University of British Columbia, Division of Cardiology, Vancouver, British Columbia, Canada
ABSTRACT
RESUM E
Background: The HeartWare left ventricular assist device (HVAD; HeartWare Inc, Framingham, MA) was first implanted in Canada in 2010. We performed a multicentre analysis of the real world outcomes associated with its use. Methods: Between May 2010 and January 2013, 4 Canadian centres inserted a total of 72 HVADs in 71 patients. Data were collected prospectively and analyzed retrospectively for the 1-year estimate of the principal outcome of transplant, explant for recovery, or death in patients who had a bridge to transplantation indication. Adverse event rates were estimated as events per patient-year (PPY). Results: In the 67 patients who received the HVAD with the indication of bridge to transplant, 26 (38.8%) received a successful transplant, 2 (3%) received an explant for recovery, and 10 (14.9%) patients died during support. Median follow-up time with the HVAD was 6.9 months (range, 2 days to 30.4 months). Despite having 74% of the patients
Introduction : Le dispositif d’assistance ventriculaire gauche Heart te implante pour la Ware (HVAD; HeartWare Inc., Framingham, MA) a e alise une analyse première fois au Canada en 2010. Nous avons re sultats en temps re el associe s à son utilisation. multicentrique des re thodes : Entre mai 2010 et janvier 2013, 4 centres canadiens ont Me un total de 72 HVAD chez 71 patients. Les donne es ont e te implante es re trospectivement quant à recueillies prospectivement et analyse sultats de transplantation, d’exl’estimation à 1 an des principaux re cupe ration ou de de cès des patients chez plantation à la suite d’une re qui il s’agissait d’un « pont vers la transplantation ». Les taux ve nements inde sirables ont e te e value s comme des e ve nements d’e e (PPA). par patient-anne sultats : Chez les 67 patients qui avaient reçu le HVAD comme Re ne ficie d’une « pont vers la transplantation », 26 (38,8 %) ont be ussie, 2 (3 %) ont eu une explantation à la suite transplantation re
Continuous flow left ventricular assist devices (LVADs) have largely replaced older generation devices and outcomes have steadily improved in the bridge to transplantation (BTT) and destination therapy (DT) populations.1-4 The HeartWare LVAD (HVAD; HeartWare, Framingham, MA) is a nextgeneration centrifugal and continuous flow pump designed for intrapericardial implantation. This pump was the subject of investigation in the ADVANCE trial, which was a noninferiority trial comparing the HVAD with the Heartmate II
(Thoratec, Pleasanton, CA) in a group of patients who were contemporaneously implanted and registered in from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) registry. The HVAD was developed with potential advantages over previous devices. Its smaller size might improve the ease of implantation by reducing tissue dissection and the smaller diameter of the outflow graft might facilitate aortic anastomosis. It has a single moving impeller that is suspended by passive magnets and a hydrodynamic thrust-bearing system to reduce the risk of mechanical failure and wear.5 Despite more than 3500 devices implanted worldwide, however, an understanding of the performance of the device in a real-world setting is still lacking. In Canada, the HVAD has only been used with the indication of BTT or bridge to candidacy until 2012 when the
Received for publication June 4, 2014. Accepted July 27, 2014. Corresponding author: Dr Jamil Bashir, St Paul’s Hospital, Suite 459, 1081 Burrard St, Vancouver, British Columbia 6Z 1Y6, Canada. Tel.: þ1604-806-9668; fax: þ1-604-806-9917. E-mail: [email protected] See page 1666 for disclosure information.
http://dx.doi.org/10.1016/j.cjca.2014.07.746 0828-282X/Ó 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.
Bashir et al. Canadian HVAD Experience
1663
with Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) scores of level 1 and 2 at the time of implantation, the rate estimate for survival at 1 year was 86.3% (95% confidence interval, 76.7-93.3). With 48.2 total patient years of support, the rates of ischemic and hemorrhagic strokes were 0.21 and 0.19 events PPY, respectively. Women made up 40% of the cohort and an adverse neurologic event occurred with an event rate of 0.38 PPY in women. Conclusions: The HVAD adequately supports acutely ill heart failure patients until the time of transplant or recovery. A high incidence of adverse neurologic outcomes might be related to the large percentage of female patients, the high INTERMACS levels, or unknown factors; further surveillance is required.
cupe ration et 10 (14,9 %) patients sont morts durant le soud’une re dian des patients portant un HVAD e tait de tien. Le temps de suivi me tendue, 2 jours à 30,4 mois). En de pit des 74 % de patients 6,9 mois (e ayant des scores de stades 1 et 2 selon l’INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) au moment de tait de 86,3% l’implantation, l’estimation du taux de survie après 1 an e (intervalle de confiance à 95 %, 76,7-93,3). Pour un nombre total de es de 48,2 avec soutien, les taux d’accident vasculaire patient-anne re bral ische mique et he morragique ont respectivement e te de 0,21 ce ve nement PPA. Les femmes constituaient 40 % de la cohorte et 0,19 e ve nement neurologique inde sirable apparaissait selon un taux et l’e ve nement de 0,38 PPA chez les femmes. d’e quat aux patients soufConclusions : Le HVAD offrait un soutien ade frant d’une insuffisance cardiaque en phase aiguë jusqu’au moment cupe ration. Une incidence e leve e de la transplantation ou de la re sirables serait lie e au pourcentage e leve d’issues neurologiques inde leve s selon l’INTERMACS ou à des facteurs de patientes, aux stades e inconnus. D’autres populations devraient être soumises à une surveillance.
devices were also approved for use with the indication of DT in Ontario. All patients who received the device had advanced heart failure during maximal medical therapy and were thought to be unable to survive without LVAD support. All centres implanting the HVAD were established LVAD centres that had previously implanted the Heartmate II (Thoratec) device. Because of the early and widespread adoption of this new technology, we engaged in a study of the HVADs implanted in Canada. The Canadian experience and outcomes of patients receiving the HVAD is relevant and unique compared with previous reports because of the publicly funded health care system, variability in funding for ventricular assist devices (VADs) across provinces, lower proportion of patients who received an implant for DT, and tendency to reserve mechanical circulatory support (MCS) for the sickest patients (ie, INTERMACS 1/2). To our knowledge, this represents the largest multicentre examination of the performance of these pumps outside of a clinical trial setting.6
form of neurologic complications (ischemic stroke, hemorrhagic stroke, and other), bleeding (gastrointestinal or other), and right ventricular failure. The secondary outcomes were calculated on the entire cohort of patients receiving the pump including the DT patients. An ischemic stroke was defined as a new neurologic deficit that was accompanied by a new corresponding defect on computed tomography (CT) scan of the brain. A hemorrhagic stroke was defined as a new neurologic deficit or symptoms with corresponding CT of the brain showing new intracranial bleeding. Gastrointestinal bleeding was defined as that requiring transfusion or change in the anticoagulation protocol. Right ventricular failure was defined as failure that required right VAD or inotropes for > 5 days after the initial implant. All centres followed the prescribed anticoagulation protocol of acetylsalicylic acid with warfarin and transesophageal echo use was routine in every case to rule out ventricular thrombus and other pathology. International normalized ratio (INR) management was performed by community physicians in conjunction with the implanting institution. Institutional preference for the target range of INR varied from 1.5-2.5 in some to 2.5-3.5 in other centres. Furthermore, intravenous unfractionated heparin or low molecular-weight heparin were used in selected patients who remained nontherapeutic for an extended period of time (typically 4-7 days) unless contraindicated because of recent active bleeding. Doppler-measured blood pressure targets were consistent among institutions and followed guideline pressures of 70-80 mm Hg for the nonpulsatile patient.7 It is routine practice to clearly determine the patients’ neurological status before implantation of the pump. This typically involves a CT examination of the head and/or a clinical evaluation by a neurologist. Informed consent was received from each patient.
Methods All patients who had an HVAD implanted in 4 Canadian centres (Vancouver, Toronto, Halifax, and Winnipeg) from May 13, 2010 until January 30, 2013 were included in the study. An additional 30 days was allowed for initial postoperative events and therefore data collection was complete as of February 28, 2013. Most Canadian centres did not use the HVAD device exclusively and continued to use the Heartmate II (Thoratec) device. Device selection was left to the discretion of the implanting surgeon. Data from Canadian centres that implanted the HVAD were collected locally and then centrally analyzed. Although the data were prospectively collected in each centre, it was analyzed retrospectively. The primary outcome was the proportion of patients who, at 12 months, survived to transplantation, had explantation of their pump for recovery, or who had ongoing mechanical support to February 28, 2013. This analysis was done on the BTT population only. The preoperative acuity of the patients was determined according to their INTERMACS score. Secondary outcomes included frequency of adverse events in the
Statistical analysis Baseline patient characteristics are presented as percentages for discrete variables and medians with 25th and 75th percentiles for continuous variables. For the primary outcome of
1664
Canadian Journal of Cardiology Volume 30 2014
Table 1. Baseline characteristics of the study population (n ¼ 71) Characteristic Median age (Q1, Q3), years Male sex, n (%) Heart failure aetiology, n (%) Dilated CMO Ischemic CMO Other INTERMACS level, n (%) 1 2 3 4 Median LVEF (Q1, Q3) ECMO, n (%) Destination therapy, n (%)
Value 53 (47, 62) 43 (60) 31 (44) 25 (35) 15 (21) 22 28 14 7 19% 12 4
(31) (39) (20) (10) (13, 21) (17) (6)
CMO, cardiomyopathy; ECMO, extra-corporeal membrane oxygenation; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; LVEF, left ventricular ejection fraction; Q, quartile.
transplantation, explant, or death, a competing risk analysis was used to estimates rates and to plot the cumulative incidence curves of each event for patients waiting for transplantation.8 All 71 patients were included in the secondary outcomes analysis to characterize adverse events during use of the HVAD pump. Adverse events are presented as frequencies and events per patient-year (event PPY).
Results A total of 72 HVAD devices were implanted in 71 patients in the 4 Canadian study centres. Four of the patients received the device for DT indication and 67 patients received it for a BTT indication. Baseline characteristics of the entire cohort are presented in Table 1. Female patients made up 40% of the study population and the median age was 53 years. The median ejection fraction before implant was 19%. The aetiology of heart failure was dilated cardiomyopathy in 44% (31 of 71), ischemic cardiomyopathy in 35% (25 of 71), and a mixture of other causes in the remaining. A large proportion (70%) of the patients in the total cohort were INTERMACS level 1 and 2 before implant and no patient had an INTERMACS level > 4. Extra-corporeal membrane oxygenation (ECMO) was used before HVAD implantation in 12 of the 71 (16.9%) patients. In the 67 patients examined using the competing risks analysis for primary outcome, the rate estimate for survival at 30 days was 91.0% (95% confidence interval, 82.7-96.4) and at 1 year was 86.3% (95% confidence interval, 76.7-93.3) (Fig. 1). Throughout the follow-up period, heart transplantation occurred in 26 of the 67 patients (38.8%) and 2 patients (3%) had their devices explanted because of myocardial recovery. Support was ongoing at last follow-up for 29 (43.3%) patients and 10 (14.9%) patients died during the follow-up period. Of the 10 patients who died, 4 died from new intracranial bleeding (2 within the first 30 days and 2 after 30 days), 4 died of ischemic stroke (2 perioperative strokes and 2 after 30 days), and 2 died of multisystem organ failure. One of the patients who died from intracranial bleeding had 2 open heart surgeries in the 10 days before implant and might have had heparin-induced
Figure 1. Competing risks analysis. Survival of patients accounting for transplants and recovery is shown in red. Rate of ongoing support is shown in orange. Cumulative incidence of death, transplant, and explant are shown in black, green, and grey, respectively.
thrombocytopenia. One additional patient died from multiple small embolic strokes that started 3 weeks after implant and a final patient died from a global ischemic neurologic insult after an arrest that occurred during a tracheostomy. In the 71 patients analyzed for secondary outcomes in the form of adverse events, the median number of months with LVAD support was 6.9 with a range of 2 days to 30.4 months. The total number of patient-years with LVAD support was 48.2 years. Two patients experienced a pump thrombosis although only 1 pump required exchange because the other patient received a transplant within 48 hours of presentation. The median time to extubation was 2 days (interquartile range [IQR], 1-5) and the median length of stay in the intensive care unit was 9 days (IQR, 6-19). The median hospital length of stay was 28 days (IQR, 20-46). The frequencies of serious adverse events are described in Table 2. Ischemic stroke occurred in 10 of the 71 patients (14.1%) and hemorrhagic stroke occurred in 9 of the 71 patients (12.7%). One patient had an ischemic stroke and a hemorrhagic stroke that were 1 month apart. The event PPY rates for ischemic stroke and hemorrhagic stroke were 0.21 and 0.19, respectively. Other neurologic events (including peripheral neuropathy and transient ischemic attack) occurred
Table 2. Rates of adverse events in counts and events PPY Adverse Event Pump thrombosis Ischemic stroke Hemorrhagic stroke Other neurologic RV failure GI bleeding RVAD
Patients With Event
Event Rate PPY
3 10 9 5 18 7 1
0.06 0.21 0.19 0.11 0.37 0.15 0.02
Median time with HVAD support 6.9 months; total support 48.2 patientyears. GI, gastrointestinal; HVAD, HeartWare left ventricular assist device (HeartWare Inc, Framingham, MA); PPY, per patient-year; RV, right ventricular; RVAD, right ventricular assist device.
Bashir et al. Canadian HVAD Experience
1665
Table 3. Rates of adverse neurologic events in counts and events PPY stratified according to timing of event (0-30 days and > 30 days) Adverse Neurologic Event Ischemic stroke Hemorrhagic stroke
Patients With Event at 0-30 Days
Patients With Event at > 30 Days
Event Rate PPY at 0-30 Days
Event Rate PPY at > 30 Days
4 2
6 7
0.80 0.40
0.14 0.16
PPY, per patient year.
in 5 patients (7.0%) for an event PPY of 0.11. Gastrointestinal bleeding occurred in 7 patients (9.9%) for an event PPY of 0.15. Right ventricular failure was the most common adverse event, affecting 18 patients (25.4%) for an event PPY of 0.37, although only 1 patient received a right VAD. Neurologic events were noted to occur at a higher rate in women compared with men. Ischemic stroke and intracranial bleeding occurred in 7 of 43 men (16.3%) (0.24 events PPY) and 11 of 28 women (39.3%) (0.38 events PPY). Most patients who died (8 of 10) had had an associated neurologic event in the form of either an ischemic stroke, intracranial bleeding, or both. Six of the neurologic events occurred early after implant (< 30 days) and 13 of them were late ( 30 days). Four of the 6 (67%) early neurologic events were fatal whereas only 3 of the 13 (23%) later events were fatal (P ¼ 0.03). Timing of the neurologic events with event rates are presented in Table 3. Of the 12 patients who had been supported with ECMO before HVAD implant, 3 had fatal neurologic events in the first 3 weeks of LVAD support. Discussion In this multicentre experience with the HVAD, the 1-year survival was 86.3%, despite a significant proportion of patients (74%) having early or true cardiogenic shock (INTERMACS 1 and 2) at the time of implantation. Unfortunately, however, the neurological event rate was surprisingly high and women fared worse outcomes when compared with men. The 1-year survival rate in our cohort (86%) is similar to the 86% survival at 360 days in the ADVANCE trial cohort.3 Transplantation occurred in 38.8% of our population and a significant number remained on support at 1 year, with the longest duration of support being 30 months. Postoperative outcomes were in keeping with how ill the patients were at the time of device implantation. The median time to extubation was 2 days but the median length of stay in the intensive care unit was 9 days. The median length of stay in-hospital was longer (28 days) than that in the ADVANCE trial (20 days). The HVAD was approved for use by the United States Food and Drug Administration in November 2012 principally on the basis of the ADVANCE trial results. This was not a randomized trial but compared outcomes to the INTERMACS registry of continuous-flow LVAD devices.3 The survival outcomes in this trial were excellent, with 86% of patients alive at 1 year and there were relatively few adverse events. However, the acuity of patients in the trial was also lower than our patient population because most (70%) patients were INTERMACS level 3. Strueber et al. examined 50 HVAD recipients from 5 international centres
and showed a 1-year survival of 84% with most of the patients being INTERMACS level 3 at the time of implantation.9 A recently published single-centre experience with 50 patients showed a 1-year survival of 78% with few adverse events.6 Because of the paucity of real-world experience with the HVAD, we conducted a study using prospectively collected data from 4 centres that had implanted the device in Canada. Patients who had the device implanted in Canada frequently presented with acute decompensation and most patients receive the device with the indication of BTT or bridge to candidacy while being acutely ill. Because 74% of the BTT patients in this study were INTERMACS 1 and 2, we describe a cohort that is more acutely ill than the patients enrolled in the described clinical trials or smaller multicentre studies. As the use of the device was adopted by acute heart failure units across the country, a natural predilection developed for implanting this device in women because of its smaller size. Therefore, 40% of patients who received the LVAD in our study were female. This is the largest proportion of women found in any LVAD series to our knowledge.1-3,6,8-10 This highlights the truly real-world nature of our experience with the device. An unexpectedly large number of patients suffered from an adverse neurologic event. The rates of hemorrhagic and ischemic stroke (0.19 and 0.21 events PPY, respectively) are significantly greater in our cohort of patients than other studies in this area including trials of the Heartmate II (Thoratec) in DT.3,8,10,11 These neurologic events were significant because 7 of the 10 patients in our cohort who died had suffered a stroke or intracranial bleeding or both. Four of these fatal events were perioperative (within first 30 days), however, and might have been related to a diverse set of risks including those incurred at the time of pump implant. In particular, these risks included a high rate of ECMO use before implant because this is undoubtedly associated with an increased incidence of early stroke.12 Also, a number of the patients with early events were extremely ill and had numerous surgeries, a profound inflammatory response, and hypotension. An additional 2 nonfatal neurologic events occurred within the first 30 days of implant. Excluding the events that occurred in the first 30 days, the rate of neurologic events remained high with a particularly high rate of hemorrhagic stroke after 30 days (0.16 events PPY). Furthermore, 7 of the 9 hemorrhagic events occurred in women and many of these events occurred well after the acute phase of illness was over and the patients were ostensibly recovered. The occurrence of late events and especially hemorrhagic stroke in women was highly consistent across institutions. This consistency within multicentre data argues against the issue being due to chance
1666
or small patient numbers and requires investigation and evaluation. A neurologic event occurred in 39.3% of women in this study. Increased rates of adverse events in the female VAD population have recently been described in 2 large series. Investigators from Baylor found over double the rate of hemorrhagic stroke in women with the Heartmate II (Thoratec) device in the BTT population (0.10 vs 0.04 events PPY).13 Also, an examination of outcome differences in the INTERMACS registry found female sex to be associated with an increased hazard of a first neurological event with a hazard ratio of 1.44.14 The cause of a greater incidence of hemorrhagic stroke in the female VAD population is unknown. Large epidemiologic studies in North America of patients without VADs do not seem to show an increased risk of stroke (ischemic or hemorrhagic) in women.15 Although an argument can be made that the high event rates in our series are at least partially because of the preponderance of women, these event rates are still significantly greater than the 0.10 total event PPY rate described for women with the Heartmate II (Thoratec) device and might be because of additional factors. Hypertension and anticoagulation are closely associated with an increased risk of hemorrhagic stroke in the general population,15,16 and undoubtedly contribute to the risk of intracranial bleeding in the LVAD population. A mean arterial pressure target of 80 mm Hg might be higher than necessary and differential targets might be beneficial for men and women. Also, all manufacturers of continuous-flow LVADs currently recommend a management strategy that includes anticoagulation. Although target INRs from the institutions in our study varied from 1.5-2.5 to 2.5-3.5, we were unable to find an association between the institutional target INR and risk of bleeding because the event rates were quite similar across institutions. The most appropriate strategy for anticoagulation with the HVAD is unknown although some institutions have begun to target the lower range. The nonpulsatile nature of continuous-flow VADs has been anecdotally linked to bleeding from acquired von Willebrand disease (VWD).17 Although the potential combination of a high nonpulsatile mean arterial pressure and an increased INR likely contribute significantly to the risk of hemorrhagic stroke in the female LVAD population, this effect might be exacerbated by other factors such as the rheology of different pump designs or the effect of rheology on the creation of acquired VWD. It is possible that the HVAD pump design potentiates acquired VWD or has other rheological effects. Thromboembolic events were also significant in this study and a smooth inflow cannula has been associated with embolic phenomenon in previous-generation VADs.17 The manufacturer of the HVAD switched from a nonsintered (smooth) to sintered inflow cannula in March of 2011 and the first of these pumps were implanted in Canada in late 2012. No relationship was found between the neurologic events and the type of inflow cannula in our study. Although our observational data show a significant pattern of increased neurologic events in the acutely ill and particularly female HVAD patient, the cause of this is unknown. These results will need to be verified in larger studies. The patients in our series were gravely ill and this undoubtedly contributed to adverse outcomes in the perioperative period.
Canadian Journal of Cardiology Volume 30 2014
It is uncertain, however, whether the acuity of presentation had an effect on late adverse neurologic outcomes. Finally, it is unknown whether factors uniquely related to the HVAD pump itself have contributed to the incidence of adverse neurologic events. Careful and frequent titration of mean arterial pressure and INR are, regardless, essential in the LVAD population. Limitations Although the data were analyzed retrospectively, they were collected prospectively, which should reduce recall bias. The difference in neurological events between men and women are hypothesis-generating only because this outcome was not prespecified a priori. The relationship between anticoagulation regimens and the incidence of adverse neurologic outcomes is not clear because individual INR values closely preceding an adverse event were not uniformly available. Also, the relationship between adverse outcomes and the medical management of these patients, particularly blood pressure management, is not well understood because of the lack of accurate recording of mean arterial pressure before an event, especially in the out-patient setting. Conclusions The HVAD continues to show the ability to support patients in the BTT population with a high success rate. High rates of adverse neurologic events need to be further investigated and might be related to the greater proportion of women in our cohort, the high INTERMACS score at the time of implantation, or other factors including pump design. Disclosures A. Cheung: consultant and speakers bureau for HeartWare Inc; V. Rao, consultant for HeartWare Inc; the remaining authors have no conflicts of interest to disclose. References 1. Pennington DG, Mcbride LR, Peigh PS, Miller LW, Swartz MT. 8 years experience with bridging to transplantation. J Thorac Cardiovasc Surg 1994;107:472-81. 2. Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009;361:2241-51. 3. Aaronson KD, Slaughter MS, Miller LW, et al. Use of an intrapericardial, continuous-flow, centrifugal pump in patients awaiting heart transplantation. Circulation 2012;125:3191-200. 4. Porepa LF, Starling RC. Destination therapy with left ventricular assist devices: for whom and when? Can J Cardiol 2014;30:296-303. 5. HeartWare Ventricular Assist System. Available at: http://www. heartware.com/products-technology/pump-design. Accessed March 16, 2013. 6. Dell’Aquila AM, Schneider SR, Schlarb D, et al. Initial clinical experience with the HeartWare left ventricular assist system: a single-center report. Ann Thorac Surg 2013;95:170-7.
Bashir et al. Canadian HVAD Experience 7. 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(suppl 4):S1-39. 8. McGiffin DC, Naftel DC, Kirklin JK, et al. Pediatric heart transplant study group. Predicting outcome after listing for heart transplantation in children: comparison of Kaplan-Meier and parametric competing risk analysis. J Heart Lung Transplant 1997;16:713-22. 9. Strueber M, O’Driscoll G, Jansz P, et al. Multicenter evaluation of an intrapericardial left ventricular assist system. J Am Coll Cardiol 2011;57: 1375-82. 10. Struber M, Sander K, Lahpor J, et al. Heartmate II ventricular assist device; early European experience. Eur J Cardiothorac Surg 2008;34: 289-94. 11. Starling SC, Naka Y, Boyle AJ, et al. Results of the post-US FDA approval study with a continuous flow left ventricular assist device as a bridge to heart transplantation. J Am Coll Cardiol 2011;57:1890-8. 12. Cheng R, Hachamovitch R, Kittleson M, et al. Complications of extracorporeal membrane oxygenation for treatment of cardiogenic shock and
1667
cardiac arrest: a meta-analysis of 1,866 adult patients. Ann Thorac Surg 2014;97:610-6. 13. Bogaev RC, Pamboukian SV, Moore SA, et al. Comparison of outcomes in women versus men using a continuous-flow left ventricular assist device as a bridge to transplantation. J Heart Lung Transplant 2011;30: 515-22. 14. Hsich EM, Naftel DC, Myers SL, et al. Should women receive left ventricular assist device support? Findings from INTERMACS. Circ Heart Fail 2012;5:234-40. 15. O’Donnell MJ, Xavier D, Liu L, et al. Risk factors for ischaemic and intracerebral haemorrhagic stroke in 22 countries (the INTERSTROKE study): a case-control study. Lancet 2010;376:112-23. 16. Hart RG, Tonarelli SB, Pearce LA. Avoiding central nervous system bleeding during antithrombotic therapy: recent data and ideas. Stroke 2005;36:1588-93. 17. Dassanayaka S, Slaughter MS, Bartoli CR. Mechanistic pathway(s) of acquired von Willebrand syndrome with a continuous-flow ventricular assist device: in vitro findings. ASAIO J 2013;59:123-9.
Clinical Research
Multicentre Canadian Experience With the HeartWare Ventricular Assist Device: Concerns About Adverse Neurological Outcomes Jamil Bashir, MD,a Jean-Francois Legare, MD,b Darren H. Freed, MD, PhD,c Anson Cheung, MD,a Vivek Rao, MD,d and Mustafa Toma, MDe a
University of British Columbia, Division of Cardiovascular Surgery, Vancouver, British Columbia, Canada b
c
Dalhousie University, Division of Cardiovascular Surgery, Halifax, Nova Scotia, Canada
University of Manitoba, Section of Cardiac Surgery, Department of Surgery, Winnipeg, Manitoba, Canada d e
University of Toronto, Division of Cardiac Surgery, Toronto, Ontario, Canada
University of British Columbia, Division of Cardiology, Vancouver, British Columbia, Canada
ABSTRACT
RESUM E
Background: The HeartWare left ventricular assist device (HVAD; HeartWare Inc, Framingham, MA) was first implanted in Canada in 2010. We performed a multicentre analysis of the real world outcomes associated with its use. Methods: Between May 2010 and January 2013, 4 Canadian centres inserted a total of 72 HVADs in 71 patients. Data were collected prospectively and analyzed retrospectively for the 1-year estimate of the principal outcome of transplant, explant for recovery, or death in patients who had a bridge to transplantation indication. Adverse event rates were estimated as events per patient-year (PPY). Results: In the 67 patients who received the HVAD with the indication of bridge to transplant, 26 (38.8%) received a successful transplant, 2 (3%) received an explant for recovery, and 10 (14.9%) patients died during support. Median follow-up time with the HVAD was 6.9 months (range, 2 days to 30.4 months). Despite having 74% of the patients
Introduction : Le dispositif d’assistance ventriculaire gauche Heart te implante pour la Ware (HVAD; HeartWare Inc., Framingham, MA) a e alise une analyse première fois au Canada en 2010. Nous avons re sultats en temps re el associe s à son utilisation. multicentrique des re thodes : Entre mai 2010 et janvier 2013, 4 centres canadiens ont Me un total de 72 HVAD chez 71 patients. Les donne es ont e te implante es re trospectivement quant à recueillies prospectivement et analyse sultats de transplantation, d’exl’estimation à 1 an des principaux re cupe ration ou de de cès des patients chez plantation à la suite d’une re qui il s’agissait d’un « pont vers la transplantation ». Les taux ve nements inde sirables ont e te e value s comme des e ve nements d’e e (PPA). par patient-anne sultats : Chez les 67 patients qui avaient reçu le HVAD comme Re ne ficie d’une « pont vers la transplantation », 26 (38,8 %) ont be ussie, 2 (3 %) ont eu une explantation à la suite transplantation re
Continuous flow left ventricular assist devices (LVADs) have largely replaced older generation devices and outcomes have steadily improved in the bridge to transplantation (BTT) and destination therapy (DT) populations.1-4 The HeartWare LVAD (HVAD; HeartWare, Framingham, MA) is a nextgeneration centrifugal and continuous flow pump designed for intrapericardial implantation. This pump was the subject of investigation in the ADVANCE trial, which was a noninferiority trial comparing the HVAD with the Heartmate II
(Thoratec, Pleasanton, CA) in a group of patients who were contemporaneously implanted and registered in from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) registry. The HVAD was developed with potential advantages over previous devices. Its smaller size might improve the ease of implantation by reducing tissue dissection and the smaller diameter of the outflow graft might facilitate aortic anastomosis. It has a single moving impeller that is suspended by passive magnets and a hydrodynamic thrust-bearing system to reduce the risk of mechanical failure and wear.5 Despite more than 3500 devices implanted worldwide, however, an understanding of the performance of the device in a real-world setting is still lacking. In Canada, the HVAD has only been used with the indication of BTT or bridge to candidacy until 2012 when the
Received for publication June 4, 2014. Accepted July 27, 2014. Corresponding author: Dr Jamil Bashir, St Paul’s Hospital, Suite 459, 1081 Burrard St, Vancouver, British Columbia 6Z 1Y6, Canada. Tel.: þ1604-806-9668; fax: þ1-604-806-9917. E-mail: [email protected] See page 1666 for disclosure information.
http://dx.doi.org/10.1016/j.cjca.2014.07.746 0828-282X/Ó 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.
Bashir et al. Canadian HVAD Experience
1663
with Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) scores of level 1 and 2 at the time of implantation, the rate estimate for survival at 1 year was 86.3% (95% confidence interval, 76.7-93.3). With 48.2 total patient years of support, the rates of ischemic and hemorrhagic strokes were 0.21 and 0.19 events PPY, respectively. Women made up 40% of the cohort and an adverse neurologic event occurred with an event rate of 0.38 PPY in women. Conclusions: The HVAD adequately supports acutely ill heart failure patients until the time of transplant or recovery. A high incidence of adverse neurologic outcomes might be related to the large percentage of female patients, the high INTERMACS levels, or unknown factors; further surveillance is required.
cupe ration et 10 (14,9 %) patients sont morts durant le soud’une re dian des patients portant un HVAD e tait de tien. Le temps de suivi me tendue, 2 jours à 30,4 mois). En de pit des 74 % de patients 6,9 mois (e ayant des scores de stades 1 et 2 selon l’INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) au moment de tait de 86,3% l’implantation, l’estimation du taux de survie après 1 an e (intervalle de confiance à 95 %, 76,7-93,3). Pour un nombre total de es de 48,2 avec soutien, les taux d’accident vasculaire patient-anne re bral ische mique et he morragique ont respectivement e te de 0,21 ce ve nement PPA. Les femmes constituaient 40 % de la cohorte et 0,19 e ve nement neurologique inde sirable apparaissait selon un taux et l’e ve nement de 0,38 PPA chez les femmes. d’e quat aux patients soufConclusions : Le HVAD offrait un soutien ade frant d’une insuffisance cardiaque en phase aiguë jusqu’au moment cupe ration. Une incidence e leve e de la transplantation ou de la re sirables serait lie e au pourcentage e leve d’issues neurologiques inde leve s selon l’INTERMACS ou à des facteurs de patientes, aux stades e inconnus. D’autres populations devraient être soumises à une surveillance.
devices were also approved for use with the indication of DT in Ontario. All patients who received the device had advanced heart failure during maximal medical therapy and were thought to be unable to survive without LVAD support. All centres implanting the HVAD were established LVAD centres that had previously implanted the Heartmate II (Thoratec) device. Because of the early and widespread adoption of this new technology, we engaged in a study of the HVADs implanted in Canada. The Canadian experience and outcomes of patients receiving the HVAD is relevant and unique compared with previous reports because of the publicly funded health care system, variability in funding for ventricular assist devices (VADs) across provinces, lower proportion of patients who received an implant for DT, and tendency to reserve mechanical circulatory support (MCS) for the sickest patients (ie, INTERMACS 1/2). To our knowledge, this represents the largest multicentre examination of the performance of these pumps outside of a clinical trial setting.6
form of neurologic complications (ischemic stroke, hemorrhagic stroke, and other), bleeding (gastrointestinal or other), and right ventricular failure. The secondary outcomes were calculated on the entire cohort of patients receiving the pump including the DT patients. An ischemic stroke was defined as a new neurologic deficit that was accompanied by a new corresponding defect on computed tomography (CT) scan of the brain. A hemorrhagic stroke was defined as a new neurologic deficit or symptoms with corresponding CT of the brain showing new intracranial bleeding. Gastrointestinal bleeding was defined as that requiring transfusion or change in the anticoagulation protocol. Right ventricular failure was defined as failure that required right VAD or inotropes for > 5 days after the initial implant. All centres followed the prescribed anticoagulation protocol of acetylsalicylic acid with warfarin and transesophageal echo use was routine in every case to rule out ventricular thrombus and other pathology. International normalized ratio (INR) management was performed by community physicians in conjunction with the implanting institution. Institutional preference for the target range of INR varied from 1.5-2.5 in some to 2.5-3.5 in other centres. Furthermore, intravenous unfractionated heparin or low molecular-weight heparin were used in selected patients who remained nontherapeutic for an extended period of time (typically 4-7 days) unless contraindicated because of recent active bleeding. Doppler-measured blood pressure targets were consistent among institutions and followed guideline pressures of 70-80 mm Hg for the nonpulsatile patient.7 It is routine practice to clearly determine the patients’ neurological status before implantation of the pump. This typically involves a CT examination of the head and/or a clinical evaluation by a neurologist. Informed consent was received from each patient.
Methods All patients who had an HVAD implanted in 4 Canadian centres (Vancouver, Toronto, Halifax, and Winnipeg) from May 13, 2010 until January 30, 2013 were included in the study. An additional 30 days was allowed for initial postoperative events and therefore data collection was complete as of February 28, 2013. Most Canadian centres did not use the HVAD device exclusively and continued to use the Heartmate II (Thoratec) device. Device selection was left to the discretion of the implanting surgeon. Data from Canadian centres that implanted the HVAD were collected locally and then centrally analyzed. Although the data were prospectively collected in each centre, it was analyzed retrospectively. The primary outcome was the proportion of patients who, at 12 months, survived to transplantation, had explantation of their pump for recovery, or who had ongoing mechanical support to February 28, 2013. This analysis was done on the BTT population only. The preoperative acuity of the patients was determined according to their INTERMACS score. Secondary outcomes included frequency of adverse events in the
Statistical analysis Baseline patient characteristics are presented as percentages for discrete variables and medians with 25th and 75th percentiles for continuous variables. For the primary outcome of
1664
Canadian Journal of Cardiology Volume 30 2014
Table 1. Baseline characteristics of the study population (n ¼ 71) Characteristic Median age (Q1, Q3), years Male sex, n (%) Heart failure aetiology, n (%) Dilated CMO Ischemic CMO Other INTERMACS level, n (%) 1 2 3 4 Median LVEF (Q1, Q3) ECMO, n (%) Destination therapy, n (%)
Value 53 (47, 62) 43 (60) 31 (44) 25 (35) 15 (21) 22 28 14 7 19% 12 4
(31) (39) (20) (10) (13, 21) (17) (6)
CMO, cardiomyopathy; ECMO, extra-corporeal membrane oxygenation; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; LVEF, left ventricular ejection fraction; Q, quartile.
transplantation, explant, or death, a competing risk analysis was used to estimates rates and to plot the cumulative incidence curves of each event for patients waiting for transplantation.8 All 71 patients were included in the secondary outcomes analysis to characterize adverse events during use of the HVAD pump. Adverse events are presented as frequencies and events per patient-year (event PPY).
Results A total of 72 HVAD devices were implanted in 71 patients in the 4 Canadian study centres. Four of the patients received the device for DT indication and 67 patients received it for a BTT indication. Baseline characteristics of the entire cohort are presented in Table 1. Female patients made up 40% of the study population and the median age was 53 years. The median ejection fraction before implant was 19%. The aetiology of heart failure was dilated cardiomyopathy in 44% (31 of 71), ischemic cardiomyopathy in 35% (25 of 71), and a mixture of other causes in the remaining. A large proportion (70%) of the patients in the total cohort were INTERMACS level 1 and 2 before implant and no patient had an INTERMACS level > 4. Extra-corporeal membrane oxygenation (ECMO) was used before HVAD implantation in 12 of the 71 (16.9%) patients. In the 67 patients examined using the competing risks analysis for primary outcome, the rate estimate for survival at 30 days was 91.0% (95% confidence interval, 82.7-96.4) and at 1 year was 86.3% (95% confidence interval, 76.7-93.3) (Fig. 1). Throughout the follow-up period, heart transplantation occurred in 26 of the 67 patients (38.8%) and 2 patients (3%) had their devices explanted because of myocardial recovery. Support was ongoing at last follow-up for 29 (43.3%) patients and 10 (14.9%) patients died during the follow-up period. Of the 10 patients who died, 4 died from new intracranial bleeding (2 within the first 30 days and 2 after 30 days), 4 died of ischemic stroke (2 perioperative strokes and 2 after 30 days), and 2 died of multisystem organ failure. One of the patients who died from intracranial bleeding had 2 open heart surgeries in the 10 days before implant and might have had heparin-induced
Figure 1. Competing risks analysis. Survival of patients accounting for transplants and recovery is shown in red. Rate of ongoing support is shown in orange. Cumulative incidence of death, transplant, and explant are shown in black, green, and grey, respectively.
thrombocytopenia. One additional patient died from multiple small embolic strokes that started 3 weeks after implant and a final patient died from a global ischemic neurologic insult after an arrest that occurred during a tracheostomy. In the 71 patients analyzed for secondary outcomes in the form of adverse events, the median number of months with LVAD support was 6.9 with a range of 2 days to 30.4 months. The total number of patient-years with LVAD support was 48.2 years. Two patients experienced a pump thrombosis although only 1 pump required exchange because the other patient received a transplant within 48 hours of presentation. The median time to extubation was 2 days (interquartile range [IQR], 1-5) and the median length of stay in the intensive care unit was 9 days (IQR, 6-19). The median hospital length of stay was 28 days (IQR, 20-46). The frequencies of serious adverse events are described in Table 2. Ischemic stroke occurred in 10 of the 71 patients (14.1%) and hemorrhagic stroke occurred in 9 of the 71 patients (12.7%). One patient had an ischemic stroke and a hemorrhagic stroke that were 1 month apart. The event PPY rates for ischemic stroke and hemorrhagic stroke were 0.21 and 0.19, respectively. Other neurologic events (including peripheral neuropathy and transient ischemic attack) occurred
Table 2. Rates of adverse events in counts and events PPY Adverse Event Pump thrombosis Ischemic stroke Hemorrhagic stroke Other neurologic RV failure GI bleeding RVAD
Patients With Event
Event Rate PPY
3 10 9 5 18 7 1
0.06 0.21 0.19 0.11 0.37 0.15 0.02
Median time with HVAD support 6.9 months; total support 48.2 patientyears. GI, gastrointestinal; HVAD, HeartWare left ventricular assist device (HeartWare Inc, Framingham, MA); PPY, per patient-year; RV, right ventricular; RVAD, right ventricular assist device.
Bashir et al. Canadian HVAD Experience
1665
Table 3. Rates of adverse neurologic events in counts and events PPY stratified according to timing of event (0-30 days and > 30 days) Adverse Neurologic Event Ischemic stroke Hemorrhagic stroke
Patients With Event at 0-30 Days
Patients With Event at > 30 Days
Event Rate PPY at 0-30 Days
Event Rate PPY at > 30 Days
4 2
6 7
0.80 0.40
0.14 0.16
PPY, per patient year.
in 5 patients (7.0%) for an event PPY of 0.11. Gastrointestinal bleeding occurred in 7 patients (9.9%) for an event PPY of 0.15. Right ventricular failure was the most common adverse event, affecting 18 patients (25.4%) for an event PPY of 0.37, although only 1 patient received a right VAD. Neurologic events were noted to occur at a higher rate in women compared with men. Ischemic stroke and intracranial bleeding occurred in 7 of 43 men (16.3%) (0.24 events PPY) and 11 of 28 women (39.3%) (0.38 events PPY). Most patients who died (8 of 10) had had an associated neurologic event in the form of either an ischemic stroke, intracranial bleeding, or both. Six of the neurologic events occurred early after implant (< 30 days) and 13 of them were late ( 30 days). Four of the 6 (67%) early neurologic events were fatal whereas only 3 of the 13 (23%) later events were fatal (P ¼ 0.03). Timing of the neurologic events with event rates are presented in Table 3. Of the 12 patients who had been supported with ECMO before HVAD implant, 3 had fatal neurologic events in the first 3 weeks of LVAD support. Discussion In this multicentre experience with the HVAD, the 1-year survival was 86.3%, despite a significant proportion of patients (74%) having early or true cardiogenic shock (INTERMACS 1 and 2) at the time of implantation. Unfortunately, however, the neurological event rate was surprisingly high and women fared worse outcomes when compared with men. The 1-year survival rate in our cohort (86%) is similar to the 86% survival at 360 days in the ADVANCE trial cohort.3 Transplantation occurred in 38.8% of our population and a significant number remained on support at 1 year, with the longest duration of support being 30 months. Postoperative outcomes were in keeping with how ill the patients were at the time of device implantation. The median time to extubation was 2 days but the median length of stay in the intensive care unit was 9 days. The median length of stay in-hospital was longer (28 days) than that in the ADVANCE trial (20 days). The HVAD was approved for use by the United States Food and Drug Administration in November 2012 principally on the basis of the ADVANCE trial results. This was not a randomized trial but compared outcomes to the INTERMACS registry of continuous-flow LVAD devices.3 The survival outcomes in this trial were excellent, with 86% of patients alive at 1 year and there were relatively few adverse events. However, the acuity of patients in the trial was also lower than our patient population because most (70%) patients were INTERMACS level 3. Strueber et al. examined 50 HVAD recipients from 5 international centres
and showed a 1-year survival of 84% with most of the patients being INTERMACS level 3 at the time of implantation.9 A recently published single-centre experience with 50 patients showed a 1-year survival of 78% with few adverse events.6 Because of the paucity of real-world experience with the HVAD, we conducted a study using prospectively collected data from 4 centres that had implanted the device in Canada. Patients who had the device implanted in Canada frequently presented with acute decompensation and most patients receive the device with the indication of BTT or bridge to candidacy while being acutely ill. Because 74% of the BTT patients in this study were INTERMACS 1 and 2, we describe a cohort that is more acutely ill than the patients enrolled in the described clinical trials or smaller multicentre studies. As the use of the device was adopted by acute heart failure units across the country, a natural predilection developed for implanting this device in women because of its smaller size. Therefore, 40% of patients who received the LVAD in our study were female. This is the largest proportion of women found in any LVAD series to our knowledge.1-3,6,8-10 This highlights the truly real-world nature of our experience with the device. An unexpectedly large number of patients suffered from an adverse neurologic event. The rates of hemorrhagic and ischemic stroke (0.19 and 0.21 events PPY, respectively) are significantly greater in our cohort of patients than other studies in this area including trials of the Heartmate II (Thoratec) in DT.3,8,10,11 These neurologic events were significant because 7 of the 10 patients in our cohort who died had suffered a stroke or intracranial bleeding or both. Four of these fatal events were perioperative (within first 30 days), however, and might have been related to a diverse set of risks including those incurred at the time of pump implant. In particular, these risks included a high rate of ECMO use before implant because this is undoubtedly associated with an increased incidence of early stroke.12 Also, a number of the patients with early events were extremely ill and had numerous surgeries, a profound inflammatory response, and hypotension. An additional 2 nonfatal neurologic events occurred within the first 30 days of implant. Excluding the events that occurred in the first 30 days, the rate of neurologic events remained high with a particularly high rate of hemorrhagic stroke after 30 days (0.16 events PPY). Furthermore, 7 of the 9 hemorrhagic events occurred in women and many of these events occurred well after the acute phase of illness was over and the patients were ostensibly recovered. The occurrence of late events and especially hemorrhagic stroke in women was highly consistent across institutions. This consistency within multicentre data argues against the issue being due to chance
1666
or small patient numbers and requires investigation and evaluation. A neurologic event occurred in 39.3% of women in this study. Increased rates of adverse events in the female VAD population have recently been described in 2 large series. Investigators from Baylor found over double the rate of hemorrhagic stroke in women with the Heartmate II (Thoratec) device in the BTT population (0.10 vs 0.04 events PPY).13 Also, an examination of outcome differences in the INTERMACS registry found female sex to be associated with an increased hazard of a first neurological event with a hazard ratio of 1.44.14 The cause of a greater incidence of hemorrhagic stroke in the female VAD population is unknown. Large epidemiologic studies in North America of patients without VADs do not seem to show an increased risk of stroke (ischemic or hemorrhagic) in women.15 Although an argument can be made that the high event rates in our series are at least partially because of the preponderance of women, these event rates are still significantly greater than the 0.10 total event PPY rate described for women with the Heartmate II (Thoratec) device and might be because of additional factors. Hypertension and anticoagulation are closely associated with an increased risk of hemorrhagic stroke in the general population,15,16 and undoubtedly contribute to the risk of intracranial bleeding in the LVAD population. A mean arterial pressure target of 80 mm Hg might be higher than necessary and differential targets might be beneficial for men and women. Also, all manufacturers of continuous-flow LVADs currently recommend a management strategy that includes anticoagulation. Although target INRs from the institutions in our study varied from 1.5-2.5 to 2.5-3.5, we were unable to find an association between the institutional target INR and risk of bleeding because the event rates were quite similar across institutions. The most appropriate strategy for anticoagulation with the HVAD is unknown although some institutions have begun to target the lower range. The nonpulsatile nature of continuous-flow VADs has been anecdotally linked to bleeding from acquired von Willebrand disease (VWD).17 Although the potential combination of a high nonpulsatile mean arterial pressure and an increased INR likely contribute significantly to the risk of hemorrhagic stroke in the female LVAD population, this effect might be exacerbated by other factors such as the rheology of different pump designs or the effect of rheology on the creation of acquired VWD. It is possible that the HVAD pump design potentiates acquired VWD or has other rheological effects. Thromboembolic events were also significant in this study and a smooth inflow cannula has been associated with embolic phenomenon in previous-generation VADs.17 The manufacturer of the HVAD switched from a nonsintered (smooth) to sintered inflow cannula in March of 2011 and the first of these pumps were implanted in Canada in late 2012. No relationship was found between the neurologic events and the type of inflow cannula in our study. Although our observational data show a significant pattern of increased neurologic events in the acutely ill and particularly female HVAD patient, the cause of this is unknown. These results will need to be verified in larger studies. The patients in our series were gravely ill and this undoubtedly contributed to adverse outcomes in the perioperative period.
Canadian Journal of Cardiology Volume 30 2014
It is uncertain, however, whether the acuity of presentation had an effect on late adverse neurologic outcomes. Finally, it is unknown whether factors uniquely related to the HVAD pump itself have contributed to the incidence of adverse neurologic events. Careful and frequent titration of mean arterial pressure and INR are, regardless, essential in the LVAD population. Limitations Although the data were analyzed retrospectively, they were collected prospectively, which should reduce recall bias. The difference in neurological events between men and women are hypothesis-generating only because this outcome was not prespecified a priori. The relationship between anticoagulation regimens and the incidence of adverse neurologic outcomes is not clear because individual INR values closely preceding an adverse event were not uniformly available. Also, the relationship between adverse outcomes and the medical management of these patients, particularly blood pressure management, is not well understood because of the lack of accurate recording of mean arterial pressure before an event, especially in the out-patient setting. Conclusions The HVAD continues to show the ability to support patients in the BTT population with a high success rate. High rates of adverse neurologic events need to be further investigated and might be related to the greater proportion of women in our cohort, the high INTERMACS score at the time of implantation, or other factors including pump design. Disclosures A. Cheung: consultant and speakers bureau for HeartWare Inc; V. Rao, consultant for HeartWare Inc; the remaining authors have no conflicts of interest to disclose. References 1. Pennington DG, Mcbride LR, Peigh PS, Miller LW, Swartz MT. 8 years experience with bridging to transplantation. J Thorac Cardiovasc Surg 1994;107:472-81. 2. Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009;361:2241-51. 3. Aaronson KD, Slaughter MS, Miller LW, et al. Use of an intrapericardial, continuous-flow, centrifugal pump in patients awaiting heart transplantation. Circulation 2012;125:3191-200. 4. Porepa LF, Starling RC. Destination therapy with left ventricular assist devices: for whom and when? Can J Cardiol 2014;30:296-303. 5. HeartWare Ventricular Assist System. Available at: http://www. heartware.com/products-technology/pump-design. Accessed March 16, 2013. 6. Dell’Aquila AM, Schneider SR, Schlarb D, et al. Initial clinical experience with the HeartWare left ventricular assist system: a single-center report. Ann Thorac Surg 2013;95:170-7.
Bashir et al. Canadian HVAD Experience 7. 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(suppl 4):S1-39. 8. McGiffin DC, Naftel DC, Kirklin JK, et al. Pediatric heart transplant study group. Predicting outcome after listing for heart transplantation in children: comparison of Kaplan-Meier and parametric competing risk analysis. J Heart Lung Transplant 1997;16:713-22. 9. Strueber M, O’Driscoll G, Jansz P, et al. Multicenter evaluation of an intrapericardial left ventricular assist system. J Am Coll Cardiol 2011;57: 1375-82. 10. Struber M, Sander K, Lahpor J, et al. Heartmate II ventricular assist device; early European experience. Eur J Cardiothorac Surg 2008;34: 289-94. 11. Starling SC, Naka Y, Boyle AJ, et al. Results of the post-US FDA approval study with a continuous flow left ventricular assist device as a bridge to heart transplantation. J Am Coll Cardiol 2011;57:1890-8. 12. Cheng R, Hachamovitch R, Kittleson M, et al. Complications of extracorporeal membrane oxygenation for treatment of cardiogenic shock and
1667
cardiac arrest: a meta-analysis of 1,866 adult patients. Ann Thorac Surg 2014;97:610-6. 13. Bogaev RC, Pamboukian SV, Moore SA, et al. Comparison of outcomes in women versus men using a continuous-flow left ventricular assist device as a bridge to transplantation. J Heart Lung Transplant 2011;30: 515-22. 14. Hsich EM, Naftel DC, Myers SL, et al. Should women receive left ventricular assist device support? Findings from INTERMACS. Circ Heart Fail 2012;5:234-40. 15. O’Donnell MJ, Xavier D, Liu L, et al. Risk factors for ischaemic and intracerebral haemorrhagic stroke in 22 countries (the INTERSTROKE study): a case-control study. Lancet 2010;376:112-23. 16. Hart RG, Tonarelli SB, Pearce LA. Avoiding central nervous system bleeding during antithrombotic therapy: recent data and ideas. Stroke 2005;36:1588-93. 17. Dassanayaka S, Slaughter MS, Bartoli CR. Mechanistic pathway(s) of acquired von Willebrand syndrome with a continuous-flow ventricular assist device: in vitro findings. ASAIO J 2013;59:123-9.