Email this article 
Hide
Show allCASE SERIES
Poor outcomes resulting from ventricular assist devices implanted in hospitals without dedicated ventricular assist device programs
Anelechi C. Anyanwu*, Sean Pinney, Kimmarie Hammond, Kimberly Ashley and David Adams
Department of Cardiothoracic Surgery, Mount Sinai Medical Center, New York, NY, USA
Abstract
Background: There is increasing drive for early implantation of ventricular assist devices (VADs) for patients in cardiogenic shock. Implants in non-VAD centers are often performed by teams without specific expertise in VAD therapy. We review the outcomes of patients transferred to a quaternary VAD center after VAD implants in outside hospitals.
Methods: Retrospective review of 12 patients transferred to our center with VADs implanted in non-VAD centers. Indication for VAD was acute shock (6) and post-cardiotomy shock (6). Median interval from implant to transfer was 3.5 days (range 1–9).
Results: Organ dysfunction was evident in most patients on transfer (medians: creatinine 2 mg/dl, bilirubin 2.6 mg/dl, AST 193 U/l). There were complications directly related to device placement in 9/12 patients resulting in air embolism–one, massive hemorrhage–two, inadequate device function–four, ventricular fibrillation–one, wound necrosis–three, and hypoxia from shunting–one. Ten patients died in intensive care a median of 6 days post-transfer (range 2–35). One died in the hospital on day 143, while one was discharged from the hospital after 139 days.
Conclusions: We observed a high frequency of technical problems related to VAD implantation, almost invariably early mortality, and a very high level of resource use in patients transferred to our center after VAD placement at outside hospitals. The increasingly popular ‘hub and spoke’ approach to VAD therapy may not necessarily be the optimal strategy for managing cardiogenic shock in the community, especially as it leaves some patients with a suboptimal starting point due to surgically related issues.
Keywords:
ventricular assist device;
hub and spoke;
ventricular assist device programs
Received: 9 January 2011; Accepted: 26 April 2011; Published: 15 June 2011
Citation: Mechanical Circulatory Support 2011, 2: 5974 - DOI: 10.3402/mcs.v2i0.5974
Mechanical Circulatory Support 2011. © 2011 Anelechi C. Anyanwu et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Introduction
The ‘hub and spoke’ model to Ventricular Assist Device (VAD) therapy was popularized by Helman and Colleagues (1) who established a regional network to improve survival of patients in post-cardiotomy shock. In this strategy, local centers are encouraged to place VADs for patients in shock and then subsequently transfer them to a regional VAD and transplant center for further management. This approach is now the favored model for providing VAD therapy to patients in cardiogenic shock in the United States and many non-VAD centers have access to short-term VAD technology for this purpose. Several workers have reported varied degree of success with this approach (2–6). Because of publication bias and bias in referral patterns, these few reports could, however, reflect a highly selected snapshot that may not be generalizable. Implants in non-VAD centers are often performed by teams without specific expertise in VAD therapy and patients subsequently transferred to VAD centers for long-term management. Logically, therefore, because of limited experience and expertise, VAD implants in spoke centers should have a low chance of successful outcome. To explore this hypothesis, we retrospectively reviewed patients transferred to our center with VADs implanted in other centers.
Methods
We retrospectively reviewed data on all patients transferred to our center between 2004 and 2008 with VADs implanted at outside centers. The study cohort included 12 patients (median age 55, M:F 8:4). Devices implanted prior to transfer were ABIOMED BVS 5000 (ABIOMED Inc., Danvers, MA)–eight, ABIOMED AB 5000 (ABIOMED Inc., Danvers, MA) – one, Biomedicus (Medtronic Inc., Minneapolis, MN) – one, TandemHeart (Cardiac Assist, Pittsburg, PA) – one, Centrimag (Thoratec Inc., Pleaston, CA)–one. Etiology of cardiomyopathy was ischemic–eight, idiopathic–one, valvular–three, while the indication for VAD was acute shock including post-infarction (6) and post-cardiotomy shock (6). The median interval from implant to transfer was 3.5 days (range 1–9). On arrival, median (range) chemistry values were: creatinine 2 (0.8–5.5) mg/dl, bilirubin 2.6 (0.6–18.8) mg/dl, and AST 193 (57–1734) U/l. Devices were implanted in seven referral centers.
Our center had a no-refusal policy for VAD transfers in the time period. There are two other ‘hub’ quaternary VAD centers in the area that also take patients from the same catchment area. On arrival at our center patients are evaluated for possible recovery. If early recovery is not likely, then the existing VAD is substituted with a long-term device or patients are listed for transplantation. Procedures undertaken at our center were: device explantation (1), conversion to intracorporeal left VAD (5), and conversion to long-term extracorporeal device (4). We did not make changes to the existing device in three patients.
Results
We observed suboptimal quality concerns related to the technical implant in 11 of 12 patients, of which direct complications resulted in 9 patients (Table 1). In surgically implanted devices (n=11), the errors arose largely from improper placement of cannula or poor siting of skin incisions. On arrival of the patient with the single percutaneously placed device (TandemHeart) to our unit, it was shown the patient was severely hypoxic and on inspection the inflow cannula had not been secured adequately at the groin such that the transeptal cannula had dislodged into the right atrium flowing deoxygenated blood through the left VAD – this patient had severe hypoxic injury. In three of the surgical patients, the inflow cannulae had not been secured to the purse string snuggers: in one of these, the left atrial cannula dislodged causing massive air embolism. Poor siting of exit sites for VAD cannulas was noted in four patients (Fig. 1) and poses a particular problem if patients are to be converted to an implantable device if these incisions are too close or near the midline wound, they all coalesce with resultant wound necrosis and infection, which we observed in three of these patients. In one patient, a left ventricular apical inflow cannula had been advanced 9 cm into the left ventricle – this cannula was intermittently obstructed by the mitral valve and irritating the basal septum causing intractable ventricular fibrillation, which was continuous for 4 days prior to transfer (Fig. 2). This patient arrived to us in multiorgan failure 9 days post-implant and within hours the diagnosis was made and the cannula withdrawn and the patient was defibrillated easily.
Fig. 1.
Poor siting of skin exit sites at LVAD placement. Exit sites are too close to each other and too close to central wound. This patient was converted to an implantable VAD but developed necrosis of abdominal wall leading to wound infection and dehiscence and device infection. Head of patient is to the left of the image; the lower aspect of the sternotomy scar is seen as are multiple skin incisions for VAD and chest drainage tubing.
Fig. 2.
Poor positioning of left ventricular inflow cannula. Four chamber view of transesophageal echocardiogram showing inflow cannula (arrows) advanced too far in into the ventricle abutting the ventricular septum and sucking in the anterior leaflet of the mitral valve (AML). There was turbulence and a 60 mmHg gradient going into the inflow cannula. This patient was transferred to our center 9 days post-implant in multiorgan failure–he had been in refractory ventricular fibrillation with suboptimal flows for 4 days pre-transfer.
| |||||||||||||||||||||
Eleven patients died in the hospital with median interval from transfer to death of 6 days. Three patients survived beyond 30 days (35, 139, and 143 days post-transfer). Only the latter two patients were ever discharged from intensive care. Only one patient survived to hospital discharge (at day 139) but he died a few weeks later of an unknown cause.
Discussion
In this paper we have demonstrated suboptimal implantation procedure and poor outcomes with patients transferred to our center with VADs. While other authors have studied this subject (2–6), none has specifically examined the quality of implants done in non-VAD centers. We observed a host of technical problems in patients who had VADs placed in non-VAD institutions. While other workers have reported greater success with over 40% survival to hospital discharge (1, 5, 6), these few reports may not tell the whole story and may reflect biased reporting by those centers with excellent results. Additionally, these centers have been very selective in which patients they accept for transfer, targeting patients with the highest likelihood of survival (4), so their results may not accurately reflect the denominator of patients having VADs in non-transplant centers.
We believe that part of the reason for poor outcomes in this patient subgroup may be the lack of expertise in VAD therapy in the implanting centers. We identified implant related problems in most patients. We believe it is unrealistic to expect such high complex surgery to be reproducibly and effectively undertaken on occasional basis in emergency situations by untrained or inexperienced teams. Indeed, the majority of cardiac surgical teams in the United States goes several years without doing a single VAD implant, so cannot reasonably be expected to provide high quality VAD care. Indeed a recent study of US Medicare patients found the median annual number of VAD implants per hospital was one (7), with several centers going a few years without doing a single VAD implant. Even if surgery can be undertaken, limited understanding of peri-operative VAD management will place these patients at further disadvantage.
Aside from poor outcomes, the ‘hub and spoke’ approach is expensive and carries a large health care resource burden. Most patients spend prolonged periods in intensive care and have repeated operations including changes to durable devices. Outcomes and resource use has not been compared directly between VADs implanted in non-VAD centers versus VAD/transplant centers. However, non-transplant centers do implant the majority of VADs for acute shock as demonstrated by Hernandez et al. (7) who observed that 65% of VADs implanted for post-cardiotomy shock in Medicare patients were implanted in non-transplant hospitals–most of which performed less than five VADs a year. The mean Medicare payment for each post-cardiotomy VAD implant was $111, 769, the total annual payment by Medicare for post-cardiotomy VAD therapy being over 150 million dollars (7). There is also the added expense of stocking rarely used expensive VAD equipment in hospitals across the country.
Our data raise questions as to whether the ‘hub and spoke’ approach to VAD therapy should be revisited. Better education and training of teams in non-VAD centers is required but may have only limited impact as most teams will go through several months or years without placing a VAD and thus may deskill rapidly. We postulate that the cardiogenic shock population may overall be better served by development of systems to facilitate immediate transfer of patients to a VAD center. One strategy worth trialing may be to transport shock patients rapidly to the nearest VAD center, with inotropic and intraortic balloon pump support or, possibly, simple catheter-based assist devices that can be rapidly placed under fluoroscopy, such as the Impella LP2.5 system (Abiomed Europe GmbH, Aachen, Germany). Although some patients may not survive transfer, those patients that survive transfer may have better outcome with VADs being done by experienced teams and there may ultimately be net benefit in terms of lives saved. Precedents for such approaches to specialist emergency care exist such as for the management of neurosurgical emergencies, major trauma, or ruptured aortic aneurysms where the priority is to move the patient to the nearest skilled center rather than have general surgeons in community hospitals attempt heroic life-saving surgery – this approach has led to overall improved outcomes for these emergencies. Many patients who cannot survive transfer are probably unlikely, anyway, to survive emergency surgery by inexperienced teams. This strategy will also lead to better utilization and conservation of scarce health care resources. Obviously, immediate transfer is not practical in post-cardiotomy patients that cannot be weaned off the heart – lung machine or, in the unstable patient, if there are great distances between the spoke and hub hospitals. Other alternative approaches are rapid response teams that can be dispatched to spoke centers to assist with VAD implant and management but this will not likely be a cost-effective or practical approach.
Study limitations include those necessarily imposed by the study design (single center, small sample sized retrospective study). Additionally we do not know the denominator of VADs implanted in the community. There likely was a degree of referral bias and delays in referral seem to have been common in our series. Delayed referral contributes to bad outcomes. In our series, the median time from VAD implant to transfer was 3.5 days – a previous study did not find any survivors in patients who were transferred more than 2 days post-implant (3). These limitations, while valid, do not alter our principal observation that there are often technical errors in implants of VADs in non-transplant centers. To our knowledge, this observation has not been documented previously as prior studies have not specifically sought to examine the quality of the initial VAD implant. The VAD implantation in non-VAD centers is associated with poor outcomes, is expensive, and, we would argue, may not be the optimal strategy for cardiogenic shock in the community, particularly as it leaves many patients with a suboptimal starting point due to a suboptimal implant technique.
Acknowledgements
This study was presented at the International Society for Heart and Lung Transplantation (ISHLT) 29th Annual Meeting and Scientific Sessions. April 22–25, 2009, Paris, France
Conflict of interest and funding
Dr. Adams has received honoraria, consultancy fees, and royalties from Edwards Lifesciences.
References
- Helman DN, Morales DL, Edwards NM, Mancini DM, Chen JM, Rose EA, et al. Left ventricular assist device bridge-to-transplant network improves survival after failed cardiotomy. Ann Thorac Surg 1999; 68: 1187–94. [Crossref]
- Anderson MB, Gratz E, Wong RK, Benali K, Kung RT. Improving outcomes in patients with ventricular assist devices transferred from outlying to tertiary care hospitals. J Extra Corpor Technol 2007; 39: 43–8.
- Gonzalez-Stawinski GV, Chang AS, Navia JL, Banbury MK, Buda T, Hoercher K, et al. Regional referral system for patients with acute mechanical support: experience at the Cleveland Clinic Foundation. ASAIO J 2006; 52: 445–9. [Crossref]
- Kherani AR, Cheema FH, Oz MC, Fal JM, Morgan JA, Topkara VK, et al. Implantation of a left ventricular assist device and the hub-and-spoke system in treating acute cardiogenic shock: who survives? J Thorac Cardiovasc Surg 2003; 126: 1634–5. [Crossref]
- McBride LR, Lowdermilk GA, Fiore AC, Moroney DA, Brannan JA, Swartz MT. Transfer of patients receiving advanced mechanical circulatory support. J Thorac Cardiovasc Surg 2000; 119: 1015–20. [Crossref]
- Haft JW, Pagani FD, Romano MA, Leventhal CL, Dyke DB, Matthews JC. Short- and long-term survival of patients transferred to a tertiary care center on temporary extracorporeal circulatory support. Ann Thorac Surg 2009; 88: 711–7. [Crossref]
- Hernandez AF, Shea AM, Milano CA, Rogers JG, Hammill BG, O'Connor CM, et al. Long-term outcomes and costs of ventricular assist devices among Medicare beneficiaries. JAMA 2008; 300: 2398–406. [Crossref]
*Ani Anyanwu
Associate Professor
Department of Cardiothoracic Surgery
Mount Sinai Medical Center
1190 Fifth Avenue
New York, NY 10029, USA
Tel: 212 659 6820
Fax: 212 659 6818
Email: anelechi.anyanwu@mountsinai.org
Refbacks
- There are currently no refbacks.
Mechanical Circulatory Support eISSN 2000-6993
This journal is published under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License. Responsible editor: Michael Firstenberg.