Abstract
Background
Despite the ability of left ventricular assist device as destination therapy (DT-LVAD) to prolong survival for many patients with advanced heart failure (HF), little is known about the eventual end-of-life care that patients with DT-LVAD receive.
Methods and Results
All patients undergoing DT-LVAD at the Mayo Clinic in Rochester, Minnesota from January 1, 2007 through September 30, 2014 who subsequently died prior to July 1, 2015 were included. Information about end-of-life care was obtained from documentation in the electronic medical record. Of 89 patients that died with a DT-LVAD, the median (25th–75th percentile) time from LVAD implantation to death was 14 (4–31) months. The most common causes of death were multiorgan failure (26%), hemorrhagic stroke (24%), and progressive HF (21%). Nearly half (46%) of patients saw palliative care within one month prior to death; however, only 13 (15%) patients enrolled in hospice a median 11 (range 1–315) days prior to death. Most patients (78%) died in the hospital, of which 88% died in the intensive care unit. In total, 49 patients had their LVAD deactivated prior to death, with all but 3 undergoing deactivation in the hospital. Most patients died within an hour of LVAD deactivation, and all within 26 hours.
Conclusions
In contrast to the general HF population, most patients with DT-LVAD die in the hospital and very few utilize hospice. Further work is needed to understand these differences and to determine whether patients with DT-LVAD are receiving optimal end-of-life care.
Keywords: heart failure, left ventricular assist device, end-of-life care, morbidity/mortality
Left ventricular assist devices (LVAD) are increasingly being utilized as destination therapy (DT-LVAD) for patients with advanced heart failure (HF) who are not candidates for orthotopic heart transplantation due to age, comorbidities, or personal choice. DT-LVAD has been shown to decrease mortality compared with medical therapy1, and can result in improved quality of life for many patients1, 2. While DT-LVAD can prolong survival for many patients with advanced HF, all patients will eventually die. Many patients live for several years after DT-LVAD, sometimes developing another chronic illness or acute event that ultimately takes their life. Others may continue to struggle with right HF or multiorgan dysfunction after DT-LVAD, resulting in death within weeks to months of device placement.
While it is known that the one-year mortality after DT-LVAD is around 20%3, little is known about the details and events preceding death. For example, in community patients with HF, we know that 22% die in the hospital and 42% enroll in hospice prior to death4, but knowledge of how death in patients with DT-LVAD compares has not been reported. Furthermore, while LVAD deactivation at the end-of-life is a well-accepted practice5, we know very little about how often and in what context it occurs. While a recent study reviewed the causes of death in patients after LVAD and noted that they vary by temporal proximity with implant6, the details surrounding death remain undescribed. This information is critical to providing optimal care for patients with DT-LVAD throughout their clinical course, but particularly as they approach end-of-life.
To address these gaps in knowledge, the goal of this analysis was to systematically examine the deaths in patients treated with DT-LVAD at a single academic center.
METHODS
Identification of Patients
All patients undergoing DT-LVAD at the Mayo Clinic in Rochester, Minnesota from 2007 through September 30, 2014 who subsequently died prior to July 1, 2015 were included. Deaths were as noted in the electronic medical record; all notifications of death of our patients implanted with LVADs are entered into the medical record in our practice. As our primary interest was in reviewing patients who died with an LVAD in situ, we excluded patients who were implanted as DT, but subsequently received a heart transplant or had their LVAD explanted prior to death. The Health Insurance Portability and Accountability Act (HIPAA) permits research using decedent records without authorization, waiver or de-identification of data when the protected health information is used solely for research. As this study met these criteria, the Mayo Clinic IRB did not require review of this protocol, in accordance with the Code of Federal Regulations, 45 CFR 46.
Patient Characteristics at the Time of LVAD Implantation
All data was abstracted from documentation in the electronic medical record. Clinicians’ diagnoses were used to define hypertension and peripheral vascular disease. Chronic obstructive pulmonary disease (COPD) was defined based on clinician’s diagnosis or greater than mild obstructive lung disease on pulmonary function testing. Cerebrovascular disease was defined as a prior history of stroke or transient ischemic attack, prior carotid revascularization, or moderate or greater carotid stenosis on ultrasound pre-LVAD. Diabetes mellitus was defined by physician diagnosis or use of diabetes medications. Body mass index (BMI) (kg/m2) was calculated using height and weight pre-LVAD and obesity included a BMI ≥30 kg/m2. All laboratory data were taken from the day prior to LVAD surgery. The Leitz Miller score7 (developed to predict in-hospital mortality, although never validated for use in patients with continuous flow LVADs) was calculated using pre-operative variables as: 7 points for platelets ≤148,000/μL, 5 points for albumin ≤3.3 g/dL, 4 points for international normalized ratio (INR) >1.1, 4 points for use of vasodilators, 3 points for mean pulmonary artery pressure <25mm Hg, 2 points for aspartate transaminase (AST) >45 U/mL, 2 points for hematocrit ≤34%, 2 points for blood urea nitrogen >51 U/dL, and 2 points for no intravenous inotropes.
Cause of Death
The cause of death was determined based on review of documentation in the electronic medical record, including autopsy reports, death certificates and electronic clinical notes.
Circumstances Preceding Death
Palliative care consultations in the month prior to death and whether patients enrolled in hospice were reviewed. At our institution, palliative care consultation is a routine part of the DT-LVAD pre-implantation workup, and ongoing palliative care availability continues longitudinally in accordance with published guidelines8. For patients who were hospitalized at other institutions prior to death, information about the timing and details of hospitalization were based on notes from physicians and LVAD coordinators in the medical record. It is standard practice at our institution for LVAD coordinators to document, in detail, each phone call that they receive about a patient with an LVAD. LVAD deactivation was defined as turning off the LVAD prior to the occurrence of clinical death. Implantable cardioverter defibrillator (ICD) deactivation was defined as turning off therapies for ventricular arrhythmias (tachy therapies).
Clinical Course Prior to Death
Rizzieri et al previously proposed three end-of-life trajectories for patients dying with DT-LVAD9, 10. Their classification included one trajectory where patients died very early after LVAD (we will refer to as Early course) due to postoperative complications or other issues. They described this to include patients who remained hospitalized after LVAD and died within 90 days of implantation. Their second trajectory included patients that survive for a period of time after LVAD but do not improve clinically and follow a similar quality of life and survival trajectory as patients with end-stage HF managed medically. We interpreted this to include patients who continue to struggle with organ failure such as right HF or renal failure requiring prolonged or recurrent readmissions from the time of LVAD until death (Persistent course). The final trajectory they describe included patients who derive quality of life and survival benefit from LVAD but later develop a gradual decline in quality of life due to a serious complication such as infection, progression of comorbidity or heart failure, or the development of a new terminal condition such as cancer. This “Terminal” end-of-life course is similar to that described among patients in the general population dying of cancer11. In extensively reviewing our population of patients dying with a DT-LVAD, we noted a unique end-of-life trajectory that was not represented in Rizzieri’s classification. Some patients experienced an improvement in quality of life after LVAD therapy, but then suffered an unexpected, acute event (such as intracranial hemorrhage) that led to an abrupt decline in health and resulted in death within 14 days; we term this the Acute course. After a detailed review of documentation of in the electronic medical record, each patient was categorized into the course that best reflected their documented experience from the time of LVAD implantation until death (Early, Persistent, Terminal, Acute, Figure 1A).
Figure 1. Clinical Course from DT-LVAD Implantation to Death.
Figure 1A depicts the change in quality of life over time after DT-LVAD in patients experiencing four clinical courses prior to death. Those experiencing an early postoperative death (A, Early course) do not derive an improvement in quality of life after LVAD, die within 90 days of implantation and remain hospitalized until the time of death. Patients experiencing a Persistent course (B) continue to struggle with heart and/or other organ (renal, hepatic, pulmonary) failure despite LVAD therapy resulting in prolonged or recurrent readmissions and, eventually, death. Other patients experience an initial functional and symptomatic improvement after LVAD resulting in an improved quality of life, but later face either a new life-limiting illness (such as cancer or progressive infection) or new-onset organ failure that results in a slow decline in quality of life preceding death (C, Terminal Course). Finally, some patients experience an initial improvement in quality of life after LVAD, but then suffer an acute event (such as hemorrhagic stroke or LVAD device malfunction) that results in death within hours to days (D, Acute course). The magnitude and duration of improvement in quality of life after DT-LVAD for patients experiencing Terminal (C) and Acute (D) end-of-life courses varies from patient to patient.
Figure 1B depicts the estimated proportion of all patients undergoing DT-LVAD who will experience each of the four clinical courses (Early, Persistent, Terminal, Acute) from the time of DT-LVAD implantation to death.
Statistical analysis
Patient characteristics were summarized using mean with standard deviation, median with interquartile range or number (%) where appropriate. Analyses were performed using STATA version 10.0.
RESULTS
Study Population
A total of 166 patients were implanted with DT-LVAD during the study period, of which 13 (7.8%) subsequently received a heart transplant (n=11) or had their LVAD explanted (n=2), 64 (38.6%) were alive at last follow-up, and 89 (53.6%) died and were included in the analysis. The characteristics of the deceased patients at the time of DT-LVAD implantation are shown in Table 1. Most patients were men (80.7%), and had ischemic cardiomyopathy as the primary etiology of their HF (58.4%). In total, 84 (94.4%) had the HeartMate II device, while 3 patients received a HeartWare (3.4%) and 2 patients had a HeartMate XVE (2.2%).
Table 1.
Pre-LVAD Characteristics of Decedents
| Characteristic | Overall (n=89) |
|---|---|
| Age, years | 64.5 (10.7) |
| Male, N(%) | 71 (80.7) |
| Ischemic etiology of HF, N(%) | 52 (58.4) |
| Prior sternotomy, N(%) | 54 (60.7) |
| Comorbidities, N(%) | |
| Hypertension | 53 (59.6) |
| Diabetes | 41 (46.1) |
| Peripheral vascular disease | 25 (28.1) |
| COPD | 15 (16.9) |
| Cerebrovascular disease | 32 (36.0) |
| Obese (BMI≥30 kg/m2) | 39 (44.3) |
| Laboratory Data, median (IQR) | |
| Total bilirubin, mg/dL | 1.0 (0.8, 1.7) |
| Aspartate aminotransferase, U/mL | 37 (26, 62) |
| Alanine aminotransferase, U/mL | 27 (18, 37) |
| Albumin, g/dL* | 3.8 (3.3, 4) |
| Platelet count, *1000/μL | 148 (108, 190) |
| Blood urea nitrogen, U/dL | 30 (21, 41) |
| Creatinine, mg/dL | 1.4 (1.1, 1.8) |
| INR | 1.3 (1.1, 1.4) |
| Leitz-Miller score* | 10.8 (5.6) |
| >Moderate RV dysfunction, N(%)* | 23 (26.4) |
| Primary Caregiver Relationship to Patient | |
| Spouse | 78 (87.6) |
| Adult child | 7 (7.9) |
| Other | 4 (4.5) |
COPD= chronic obstructive pulmonary disease, INR-international normalized ratio, IQR= interquartile range, RV= right ventricular
Two patients were missing albumin, two were missing quantification of RV dysfunction and eight were missing a Leitz-Miller score. All remaining variables were complete.
Characteristics Preceding Death
The patient characteristics and circumstances at the time of death are shown in Table 2. The date of death was available on all patients. In two patients, no clinical details around the time of death were available, and only very limited information was available on two additional patients. Follow-up time from DT-LVAD implantation to death ranged from 1–2366 days, with a median (25th–75th percentile) of 14 (4–31) months (Figure 2). The mean age at death was 66.1 years. Most patients died in the hospital (77.6%), of whom 68.2% were hospitalized at a Mayo Clinic hospital. In total 18 (20.2%) died during the initial LVAD implantation hospitalization (i.e. were never discharged after surgery); this reflects an index in-hospital mortality rate of 10.8% overall (18 of 166 patients implanted with DT-LVAD), which is similar to the national average12. Among those 18 patients, the median time from LVAD implantation to death was 16 days (range 1–270 days), with 5 dying within a week of implantation. For patients who died in the hospital, but not during the index hospitalization where the LVAD was implanted, death occurred a median of 4 days after admission (range 0–60 days), with 14 of 48 (29.2%) dying within 24 hours of admission.
Table 2.
Circumstances at Death of 89 DT-LVAD Patients
| Characteristic | Missing | Summary |
|---|---|---|
| Age at death, years, mean (SD) | 0 | 66.1 (10.7) |
| Time from LVAD implant to death, years, mean (SD) | 0 | 1.7 (1.7) |
| In hospital death, N(%) | 4 | 66 (77.6) |
| Time hospital admission to death, days, median (IQR) | 0* | 6 (2, 24) |
| ICU care during terminal hospitalization, N(%) | 5* | 55 (91.7) |
| Died in an ICU, N(%) | 9* | 50 (87.7) |
| Resuscitation during terminal hospitalization, N(%) | 3* | 13 (20.6) |
| Palliative Care consult last month of life, N(%) | 9 | 37 (46.3) |
| Enrolled in hospice, N(%) | 5 | 13 (15.3) |
| Time hospice to death, days, median (IQR) | 1 | 10.5 (7.5, 25.5) |
| LVAD deactivated prior to death, N(%) | 8 | 49 (60.5) |
| Time LVAD deactivation to death, hours | All within 26 hours | |
| ICD deactivated prior to death, N(%)† | 20† | 43 (66.2) |
| Autopsy, N(%) | 12 | 24 (31.2) |
ICD= implantable cardioverter defibrillator, ICU= intensive care unit, IQR= interquartile range, LVAD= left ventricular assist device, SD= standard deviation
Of the 66 patients who were known to be hospitalized at the time of death
Of the 85 patients with an ICD. In 20 patients it was unknown if the ICD was deactivated prior to death, as they were hospitalized or cared for at another institution and this information was not documented in the medical records we received.
Figure 2. Time from DT-LVAD Implantation to Death Among Decedents.
The proportion of decedents that died by months after LVAD implantation is shown. As many patients implanted with DT-LVAD in the study period were still alive at the end of the study period, these proportions do not represent expected survival after DT-LVAD.
Among the 66 patients who died in-hospital, 57 patients had a known location of death in the hospital (intensive care unit [ICU] vs. non-ICU bed), and 50 (87.7%) of these patients died in the ICU. Of the 45 patients that died at Mayo Clinic Rochester hospitals, 40 (88.9%) died in the ICU, 33 (73.3%) were mechanically ventilated and 13 (28.9%) received hemodialysis in the 48 hours preceding death. In addition, 40 (88.9%) transitioned to comfort-focused care prior to death; but the remaining 5 were “Full Code” status and died despite resuscitative efforts. Of the 40 patients that transitioned to comfort-focused care, 33 (82.5%) transitioned to this approach within 24 hours of death. Despite a transition to comfort-focused care, this did not necessarily mean the LVAD was deactivated; rather, the patients’ care was focused on comfort and not escalating care further. One patient remained hospitalized for more than 3 days after transitioning to a comfort-directed approach to care (7 days from transition to death) and his care needs were believed to be too complex for safe discharge to another setting.
Cause of Death
The cause of death was available in 86 (96.6%) patients. The most common causes of death were multiorgan failure (25.6%), hemorrhagic stroke (24.4%), and HF (20.9%, Figure 3). Four patients died of bleeding; two of which were in the immediate postoperative setting. Two patients died of a known LVAD mechanical issue that occurred in a non-hospitalized setting (one ran out of batteries, one during controller exchange).
Figure 3. Cause of Death in Patients with DT-LVAD.
The number of patients (X axis) experiencing various causes of death and their associated clinical course prior to death are shown. The 85 patients with a known clinical course and cause of death are represented.
Palliative Care and Hospice
While we previously reported that 89% of patients at our institution had palliative care consultation prior to DT-LVAD implantation13, only about half (46.3%) of patients saw palliative care in consultation within one month prior to death. All patients were hospitalized at the time of palliative care consultation, though 4 (10.8%) patients were subsequently discharged from the hospital prior to death. The median time from palliative care consultation to death was 9 days and the median number of visits per patient in the last month of life was 4 (range 1–15). In total, only 13 (15.3%) patients enrolled in hospice a median of 11 (range 1–315) days prior to death, including 4 patients that had cancer, 2 with stroke, 2 with HF, and 5 with multiorgan failure. Of those that enrolled in hospice, 8 received hospice care at home, 4 received hospice care in an inpatient hospice facility and 1 in a skilled nursing facility. Only 1 patient received hospice care (home) through a hospice agency that was located in close proximity to our institution. Only one patient that enrolled in hospice died in the hospital, while the remaining 12 patients died in the outpatient setting. The patient who survived the longest in hospice care chose to enroll after receiving a diagnosis of metastatic solid organ malignancy, and lived an additional 10 months after enrollment.
LVAD Deactivation
Information was not available on whether the LVAD was deactivated prior to death in 8 (9.0%) patients. Of the remaining 81 patients, 49 (60.5%) had their LVAD deactivated prior to death and the remaining 32 patients did not have their LVAD deactivated. Of those who underwent LVAD deactivation, the median time from LVAD implantation to death was 419 (range 1–2366) days. The most common causes of death were multiorgan failure (n=18, 36.7%), hemorrhagic stroke (n=15, 30.6%), heart failure (n=7, 14.3%) and infection (n=4, 8.2%) In total, 3 (6.1%) patients had the LVAD deactivated at home, while the remainder had LVAD deactivation in the hospital. The final decision to deactivate the LVAD was usually made by the patient’s family after discussion with the clinical team (85.7% of deactivations), rather than the patient (8.2%) as the patient was frequently unconscious or lacked decision-making capacity. Of the 49 patients that underwent LVAD deactivation prior to death, the time from deactivation to death (in hours) was known in 47 patients. The vast majority (42 of 47, 89.4%) died within 1 hour of deactivation. The remaining patients survived 4, 9, 11, 17, and 26 hours after LVAD deactivation. All 12 patients who had their LVAD deactivated during their initial (implant) hospitalization died within one hour of LVAD deactivation.
Clinical Course Prior to Death
A total of 86 patients were categorized into one of four groups by their clinical course from the time of LVAD implantation until death. The remaining 3 patients could not be completely categorized due to a lack of detailed information around the time of death though all 3 died several years after DT-LVAD and thus did not experience Early deaths. Clinical courses observed were Early, Persistent, Acute and Terminal in 18 (20.9%), 17 (19.8%), 31 (36.0%), and 20 (23.3%) patients, respectively. The median time from LVAD implantation to death was shortest in Early course patients (16 days), followed by patients with Persistent (201 days), Acute (602 days) and Terminal (892 days) clinical courses. In hospital death occurred in 100%, 76.5%, 74.2% and 63.2% of patients experiencing Early, Persistent, Acute, and Terminal courses. In the patients with an Acute course prior to death, the time from the acute event until death was <24 hours in 19 (61.3%) patients, 1–7 days in 9 (29.0%) patients, and 8–14 days in 3 (9.7%) patients. Of the 49 patients who had their LVAD deactivated prior to death, clinical courses observed were Early, Persistent, Acute and Terminal in 12 (24.5%), 10 (20.4%), 16 (32.7%), and 11 (22.4%). All patients who enrolled in hospice prior to death experienced a Persistent or Terminal course. Most patients experiencing Early and Persistent courses had deaths due to HF and multiorgan failure (Figure 3). Hemorrhagic stroke accounted for 60.0% of deaths in those with an Acute course. The most common causes of death in patients with a Terminal course were cancer (25.0%) and multiorgan failure (40.0%).
The estimated proportion of all patients undergoing DT-LVAD who fit each of the four clinical courses is shown in Figure 1B. These estimates are based upon the distributions observed among decedents, but were adjusted to account for the fact that many patients who had DT-LVADs implanted in the study period are still alive. By including only decedents in this analysis, we have overestimated the proportion of patients implanted with DT-LVADs who have clinical trajectories with shorter times from LVAD implantation to death (i.e. the early postoperative deaths) and underestimated the proportion who would follow the longer survival trajectories (i.e. terminal condition and acute event deaths). Accounting for this, we estimate that, in all patients undergoing DT-LVAD, ~10% would experience Early deaths (consistent with national outcomes12), 20% would experience a Persistent course, 30% Terminal, and 40% Acute.
DISCUSSION
In this single-center study, the vast majority of patients with a DT-LVAD died in the hospital and most in an ICU setting. While half of patients saw palliative care in consultation in the month before death, very few enrolled in hospice. LVAD deactivation prior to death was common, and these patients died shortly after deactivation.
Our findings reveal that patients with DT-LVAD die differently than other patients with HF. Patients with DT-LVAD are far less likely to enroll in hospice (15.3%) and more frequently die in the hospital (77.6%) than either local Olmsted County decedents with HF (42.2% hospice enrollment, 22.4% in-hospital death)4 or 2007 Medicare beneficiaries dying with HF (38.1% hospice enrollment, 35.2% in-hospital death)14. The comparatively low rates of hospice enrollment and high in-hospital mortality rates seen in patients with DT-LVAD are also divergent compared to other patients dying with chronic diseases such as cancer and COPD15. These outcomes are important as prior studies have demonstrated that patients who die in an ICU or hospital experience worse quality of life and more physical and emotional distress at the end-of-life compared with patients who die at home with hospice services16. Furthermore, their bereaved caregivers suffer more psychiatric illness16 and are less satisfied with the end-of-life care received by their loved one17. Whether these findings are true for patients with DT-LVAD and their caregivers remains to be determined.
There are several potential reasons why end-of-life care for DT-LVAD patients may be so different than patients with other chronic diseases. First, patient preferences may differ. Patients with DT-LVAD have, by definition, chosen to have a mechanical heart pump implanted and to remain in situ for the rest of their life. Thus, DT-LVAD patients are a selected population who have shown that they were willing to undergo invasive procedures in order to stay alive, and may be more likely to choose life-prolonging therapies in other clinical situations, though this is worthy of further examination. Second, in our experience, it can be challenging to find hospice agencies and skilled nursing facilities that are willing to accept patients with LVADs due to a lack of comfort and training in managing the device. While further work is needed to determine the most appropriate end-of-life options to support patients with DT-LVAD and their caregivers, hospice has been shown to be beneficial to patients dying with a variety of health conditions, and patients with DT-LVAD may also benefit from their services. As such, it will be important to determine ways to provide LVAD education and support to hospice agencies so that DT-LVAD patients may benefit from their services. Third, it may be more challenging to prognosticate in patients with a DT-LVAD compared to other patients with HF. For example, acute deaths from etiologies such as hemorrhagic stroke are common and can be difficult to predict. Finally, clinicians may not be routinely engaging patients with DT-LVAD and their families in goals of care discussions. Clinicians often hesitate to have these discussions for many reasons, including a fear of destroying hope, sensing that the patient or family are not ready, or because they lack confidence in their ability to approach these important conversations18. However, these conversations are critically important to understanding the patient’s goals, values and preferences so that we can provide goal-concordant care at the end-of-life.
The involvement of palliative care clinicians in the care of patients with HF approaching the end-of-life has been a key part of our practice. Palliative care consultations were obtained in a similar proportion of patients dying with DT-LVAD (47%) and those dying with HF (44%)4 in our community. While palliative care clinicians are routinely involved in preparedness planning for patients prior to DT-LVAD13 in accordance with guidelines8, they were often re-consulted only in very close proximity to death. While in our local community, increased consultations from palliative care specialists at the end-of-life coincided with increased use of hospice services and decreased rates of in-hospital death in patients with HF4, we observed very low rates of hospice enrollment and high rates of in-hospital death for patients with DT-LVAD despite frequent input from palliative care specialists. This may reflect the fact that patients with DT-LVAD are choosing to continue with life-prolonging therapies even when prognosis is grim more often than other patients with HF. As previously noted, it may also reflect the challenge of transitioning patients with DT-LVAD to the outpatient setting at the end-of-life due to the lack of adequate training of skilled nursing facilities and hospice agencies in LVAD management. Further work is needed to understand the optimal role of palliative care specialists in the longitudinal care of patients with DT-LVAD.
DT-LVADs were deactivated in most patients prior to death and death occurred within 26 hours in all cases. By comparison, one previous study reported on time from LVAD deactivation to death in 17 patients and death occurred in <20 minutes in all cases.19 The time from deactivation to death will inevitably vary based upon native cardiac function and by whether the aortic valve was sutured close at the time of LVAD; however, this information has important implications on how patients and families are counselled regarding what to expect after LVAD deactivation20.
Limitations
There are limitations that are important to consider when interpreting these data. First, this is a single center study and results from other centers may differ. Some details about deaths that occurred in other hospitals were unavailable. Quality of life was not routinely assessed during follow-up after LVAD. However, there is no LVAD-specific quality of life measure available and heart failure-specific quality of life measures such as the Kansas City Cardiomyopathy Questionnaire may not adequately capture domains that impact quality of life in patients with LVADs. As such, patients were categorized into end-of-life trajectories based upon available clinical data, which could result in misclassification.
Clinical Implications and Future Directions
These data provide a first step toward quantitatively understanding the end-of-life care of patients dying with DT-LVAD. The next step would be to examine whether end-of-life care for DT-LVAD follows a similar pattern at other institutions, and to examine reasons for differences if they exist. Second, we need a better understanding of the reasons that hospice use is so low and in-hospital death is so high in this population. A recent qualitative study of 8 bereaved caregivers of patients with LVAD found that caregivers felt confused about the involvement of other care providers, such as palliative care and hospice clinicians, at the end of their loved one’s life21. Further qualitative inquiry to include various stakeholders such as physicians, LVAD coordinators, social workers, patients, and caregivers are needed to fully delineate the factors that may be contributing to our observations. Finally, better models to predict death and poor quality of life in patients after DT-LVAD implantation are needed, which may enable clinicians to provide more accurate prognostic information to patients so that they may make informed decisions about their care.
Conclusions
In contrast to patients dying with HF and other chronic conditions, most patients with DT-LVAD die in the hospital and very few enroll in hospice. Potential reasons that patients with DT-LVAD may experience different end-of-life care than other patients with HF exist, including that they often experience sudden changes in health status prior to death from acute events and that arranging for safe out-of-hospital care at the end-of-life can be particularly challenging. However, further work is needed to understand these differences and to determine how to optimize end-of-life care for patients with DT-LVAD.
Clinical Perspective.
In this analysis of 89 decedents with DT-LVAD at a single center, we found that most patients died as a result of multiorgan failure, hemorrhagic stroke, or heart failure. About one-third of decedents did well for a period of time after LVAD, but then experienced an unexpected, acute event that led to an abrupt decline in health and resulted in death. In contrast with the general heart failure population, the vast majority of LVAD recipients died in the hospital, most often in the intensive care unit, and very few patients enrolled in hospice. LVAD deactivation prior to death was common, and most patients died within an hour of deactivation. These data provide a first step toward understanding the current state of end-of-life care of patients dying with DT-LVAD. Future work is needed to understand the reasons that hospice use is so low and in-hospital death is so high in this population. Ultimately, efforts toward determining optimal end-of-life care in patients with DT-LVAD are needed.
Acknowledgments
SOURCES OF FUNDING: Dr. Dunlay is supported by an NIH Career Development Award (K23 HL116643).
Footnotes
DISCLOSURES: None.
References
- 1.Rose EA, Gelijns AC, Moskowitz AJ, Heitjan DF, Stevenson LW, Dembitsky W, Long JW, Ascheim DD, Tierney AR, Levitan RG, Watson JT, Meier P, Ronan NS, Shapiro PA, Lazar RM, Miller LW, Gupta L, Frazier OH, Desvigne-Nickens P, Oz MC, Poirier VL. Long-term use of a left ventricular assist device for end-stage heart failure. N Engl J Med. 2001;345:1435–43. doi: 10.1056/NEJMoa012175. [DOI] [PubMed] [Google Scholar]
- 2.Kato NP, Okada I, Imamura T, Kagami Y, Endo M, Nitta D, Fujino T, Muraoka H, Minatsuki S, Maki H, Inaba T, Kinoshita O, Nawata K, Hatano M, Yao A, Kyo S, Ono M, Jaarsma T, Kinugawa K. Quality of Life and Influential Factors in Patients Implanted With a Left Ventricular Assist Device. Circ J. 2015;79:2186–92. doi: 10.1253/circj.CJ-15-0502. [DOI] [PubMed] [Google Scholar]
- 3.McIlvennan CK, Magid KH, Ambardekar AV, Thompson JS, Matlock DD, Allen LA. Clinical outcomes after continuous-flow left ventricular assist device: a systematic review. Circ Heart Fail. 2014;7:1003–13. doi: 10.1161/CIRCHEARTFAILURE.114.001391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dunlay SM, Redfield MM, Jiang R, Weston SA, Roger VL. Care in the last year of life for community patients with heart failure. Circ Heart Fail. 2015;8:489–96. doi: 10.1161/CIRCHEARTFAILURE.114.001826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mueller PS, Swetz KM, Freeman MR, Carter KA, Crowley ME, Severson CJ, Park SJ, Sulmasy DP. Ethical analysis of withdrawing ventricular assist device support. Mayo Clin Proc. 2010;85:791–7. doi: 10.4065/mcp.2010.0113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Stulak JM, Mehta V, Schirger JA, Aaronson KD, Joyce LD, Daly RC, Pagani FD, Maltais S. Temporal Differences in Causes of Mortality After Left Ventricular Assist Device Implantation. Ann Thorac Surg. 2015;99:1969–72. doi: 10.1016/j.athoracsur.2015.01.036. discussion 1972–4. [DOI] [PubMed] [Google Scholar]
- 7.Lietz K, Long JW, Kfoury AG, Slaughter MS, Silver MA, Milano CA, Rogers JG, Naka Y, Mancini D, Miller LW. Outcomes of left ventricular assist device implantation as destination therapy in the post-REMATCH era: implications for patient selection. Circulation. 2007;116:497–505. doi: 10.1161/CIRCULATIONAHA.107.691972. [DOI] [PubMed] [Google Scholar]
- 8.Feldman D, Pamboukian SV, Teuteberg JJ, Birks E, Lietz K, Moore SA, Morgan JA, Arabia F, Bauman ME, Buchholz HW, Deng M, Dickstein ML, El-Banayosy A, Elliot T, Goldstein DJ, Grady KL, Jones K, Hryniewicz K, John R, Kaan A, Kusne S, Loebe M, Massicotte MP, Moazami N, Mohacsi P, Mooney M, Nelson T, Pagani F, Perry W, Potapov EV, Eduardo Rame J, Russell SD, Sorensen EN, Sun B, Strueber M, Mangi AA, Petty MG, Rogers J. The 2013 International Society for Heart and Lung Transplantation Guidelines for mechanical circulatory support: executive summary. J Heart Lung Transplant. 2013;32:157–87. doi: 10.1016/j.healun.2012.09.013. [DOI] [PubMed] [Google Scholar]
- 9.Rady MY, Verheijde JL. Ethical challenges with deactivation of durable mechanical circulatory support at the end of life: left ventricular assist devices and total artificial hearts. J Intensive Care Med. 2014;29:3–12. doi: 10.1177/0885066611432415. [DOI] [PubMed] [Google Scholar]
- 10.Rizzieri AG, Verheijde JL, Rady MY, McGregor JL. Ethical challenges with the left ventricular assist device as a destination therapy. Philos Ethics Humanit Med. 2008;3:20. doi: 10.1186/1747-5341-3-20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Murray SA, Kendall M, Boyd K, Sheikh A. Illness trajectories and palliative care. BMJ. 2005;330:1007–11. doi: 10.1136/bmj.330.7498.1007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Khazanie P, Hammill BG, Patel CB, Eapen ZJ, Peterson ED, Rogers JG, Milano CA, Curtis LH, Hernandez AF. Trends in the use and outcomes of ventricular assist devices among Medicare beneficiaries, 2006 through 2011. J Am Coll Cardiol. 2014;63:1395–404. doi: 10.1016/j.jacc.2013.12.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Verdoorn B, Swetz KM, Luckhardt AJ, Dunlay SM. Preparedness planning for patients receiving left ventricular assist device as destination therapy: how best to measure change in culture? J Pain Symptom Manage. 2015;40:404–05. doi: 10.1016/j.jpainsymman.2016.10.372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Unroe KT, Greiner MA, Hernandez AF, Whellan DJ, Kaul P, Schulman KA, Peterson ED, Curtis LH. Resource use in the last 6 months of life among medicare beneficiaries with heart failure, 2000–2007. Arch Intern Med. 2011;171:196–203. doi: 10.1001/archinternmed.2010.371. [DOI] [PubMed] [Google Scholar]
- 15.Teno JM, Gozalo PL, Bynum JP, Leland NE, Miller SC, Morden NE, Scupp T, Goodman DC, Mor V. Change in end-of-life care for Medicare beneficiaries: site of death, place of care, and health care transitions in 2000, 2005, and 2009. JAMA. 2013;309:470–7. doi: 10.1001/jama.2012.207624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wright AA, Keating NL, Balboni TA, Matulonis UA, Block SD, Prigerson HG. Place of death: correlations with quality of life of patients with cancer and predictors of bereaved caregivers’ mental health. J Clin Oncol. 2010;28:4457–64. doi: 10.1200/JCO.2009.26.3863. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Teno JM, Clarridge BR, Casey V, Welch LC, Wetle T, Shield R, Mor V. Family perspectives on end-of-life care at the last place of care. JAMA. 2004;291:88–93. doi: 10.1001/jama.291.1.88. [DOI] [PubMed] [Google Scholar]
- 18.Dunlay SM, Foxen JL, Cole T, Feely MA, Loth AR, Strand JJ, Wagner JA, Swetz KM, Redfield MM. A survey of clinician attitudes and self-reported practices regarding end-of-life care in heart failure. Palliat Med. 2015;29:260–7. doi: 10.1177/0269216314556565. [DOI] [PubMed] [Google Scholar]
- 19.Brush S, Budge D, Alharethi R, McCormick AJ, MacPherson JE, Reid BB, Ledford ID, Smith HK, Stoker S, Clayson SE, Doty JR, Caine WT, Drakos S, Kfoury AG. End-of-life decision making and implementation in recipients of a destination left ventricular assist device. J Heart Lung Transplant. 2010;29:1337–41. doi: 10.1016/j.healun.2010.07.001. [DOI] [PubMed] [Google Scholar]
- 20.Gafford EF, Luckhardt AJ, Swetz KM. Deactivation of a left ventricular assist device at the end of life #269. J Palliat Med. 2013;16:980–2. doi: 10.1089/jpm.2013.9490. [DOI] [PubMed] [Google Scholar]
- 21.McIlvennan CK, Jones J, Allen LA, Swetz KM, Nowels C, Matlock DD. Bereaved caregivers perspectives on end-of-life experience with a left ventricular assist device: “You know. he probably can’t die”. JAMA Int Med. 2016;176:534–9. doi: 10.1001/jamainternmed.2015.8528. [DOI] [PubMed] [Google Scholar]