Characteristics,Interventions,and Outcomes of Lung Transplant Recipients Co-Managed with Palliative Care

Introduction

Lung transplantation (LT) can improve survival and quality of life for patients with end-stage lung disease, but the 5-year survival rate remains low (<60%) compared with other solid-organ transplant recipients. ¹ The principal threat to survival and quality of life in LT is the onset of chronic allograft dysfunction, a condition with no cure aside from re-transplantation. Transplant recipients may carry a high symptom burden from graft dysfunction, toxicity from immunosuppressive agents, depression, and other causes.^2,3

Palliative care (PC) is recommended for advanced lung disease,^4–6 and some have suggested that PC may help improve symptoms and manage complex illness trajectories among LT candidates and recipients.^2,7,8 However, LT recipients are rarely referred to PC, even after the diagnosis of chronic allograft dysfunction.^2,9 End-of-life care for LT recipients is often aggressive and not comfort oriented. ¹⁰

In our institution, the LT service has increasingly consulted PC to co-manage LT recipients with end-stage graft dysfunction or other terminal conditions. We present the characteristics, PC interventions used, and outcomes of these patients.

Methods

We conducted a single-center, retrospective, cohort study of LT recipients referred by the Toronto Lung Transplant Program at the University Health Network, Toronto, Ontario, Canada, for PC consultation between January 2010 and May 2012. Inclusion criteria were adult LT recipients who were referred for PC consultation for any reason during the study period.

Our center performs approximately 100 adult LTs per year. Approximately 610 post-transplantation patients were being followed in June 2012. Our PC service offers inpatient and outpatient consultation. The service is staffed by three full-time equivalent physicians and one full-time advanced practice nurse. In 2012, the PC service provided consultation to approximately 1000 patients.

We reviewed patient records of LT recipients and abstracted the date of PC referral, LT, and death; patient location; reason for referral; demographic information; Edmonton Symptom Assessment System (ESAS) ¹¹ and Palliative Performance Scale (PPS) ¹² scores; PC interventions; resuscitation orders; and patient outcomes. Information about involvement of other services (e.g., psychiatry, spiritual care, social work) and baseline partial pressure of carbon dioxide (PCO₂) measurements was collected when available. The ESAS is an assessment tool that characterizes the severity of nine common symptoms experienced by patients with terminal illness. It has been validated in cancer patients and is frequently used during assessments of patients with nonmalignant disease. The PPS is a validated assessment tool that characterizes a patient's functional status from 0% (death) to 100% (fully functional). Data were abstracted by a physician with training and experience in respirology, LT, and PC. Data analysis was descriptive. The study was approved by the research ethics board of the University Health Network, Toronto, Ontario (11-0677-BE).

Results

From January 2010 to May 2012, 24 adult LT recipients were referred for PC consultation (Table 1). The majority were referred with chronic allograft dysfunction. The median PPS was 30 and only 8 (33%) patients were able to complete an ESAS. Among the latter, the most significant symptoms (i.e., >5/10) reported were “tiredness,” “shortness of breath,” and “drowsiness.” The remainder did not complete an ESAS, either due to poor functional status, symptoms, or patient refusal.

The details of the PC consultation and interventions provided are included in Table 2. Three-quarters were inpatients at the time of referral, and the most common reason for referral was shortness of breath or end-of-life planning. The median time between transplantation and PC referral was more than 3 years, and the median time between PC referral and death was 2 weeks; 3 of 24 patients were still alive at the time of data collection (a median of 17 [range 0–810] days post-PC consultation). Twenty-two patients (92%) were prescribed opioids over their course of treatment; 14 were already receiving opioids at the time of consultation, whereas 8 had their first opioids prescribed by the PC service. Seven patients were maintained on standing orders for opioids at a median oral morphine equivalent dose of 50 mg per day (range, 6 mg to 240 mg). The remaining 15 patients used short-acting opioids on an “as needed” basis only (actual opioid use was not available). There were no documented episodes of respiratory depression, opioid toxicity, bowel obstruction, or pseudo-obstruction. Some patients were assessed by psychiatry (11/24, 46%), social work (4/24, 17%), and spiritual care (11/24, 46%).

Twelve patients (50%) had documented resuscitation orders prior to PC consultation; 9 had do not resuscitate (DNR) orders, whereas 3 had orders for full cardiopulmonary resuscitation (FC). Following PC consultation, the remaining 12 patients established resuscitation orders; a total of 8 of 24 (33%) had an FC order. Of those who had a documented FC order, 2 were inpatients in the intensive care unit (ICU), and both went on to have a withdrawal of life support within 7 and 14 days, respectively. Three later died on the ward receiving comfort measures. Two died at home with unclear details, and one was still alive at last follow-up.

No patient received cardiopulmonary resuscitation (CPR) at the time of death. Four patients (17%) died in an ICU; 2 of these patients had life support withdrawn (mentioned above), whereas one did not have life support initiated. The fourth patient died at another institution and the details of death were unavailable. Among the 12 whose deaths occurred at our center, 10 (83%) were receiving comfort medications at the time of death.

At the time of review, 21 of 24 of the patients had died. Causes of death were listed as infection (11, 52%), chronic allograft dysfunction (9, 43%), and cirrhosis (1, 5%). Most patients died on the ward (14, 67%), with the remainder dying in the ICU (4, 19%) or at home (3, 14%).

Discussion

The present study is the first published report of a cohort of LT recipients referred for co-management by PC. Most were referred as inpatients, typically when bedbound, for management of dyspnea or end-of-life care. Half were referred to PC more than a week before death, and even though one-third had requested FC at the time of the PC consultation, all patients (or their surrogates) agreed to a DNR order prior to death. None received CPR and most were receiving comfort medications at the time of death.

Although there are limited studies of PC involvement in both solid-organ and stem-cell transplantation, the existing literature suggests that the PC service is rarely consulted despite the high prevalence of significant symptoms or onset of graft dysfunction.^2,3,13,14 Some transplant centers do not have a PC service, ² but there are many barriers to PC consultation even when a service is available.^2,7 Some physicians delay a PC referral until the final days due to a belief that PC is incompatible with the goals of transplant care. Family members are often resistant to PC referral until a poor prognosis is certain. ⁹ In contrast, our findings are consistent with a changing paradigm of PC, in which patients are referred earlier in the course of illness.

By delaying or avoiding PC, many LT recipients receive aggressive care at the end of life, even when the underlying chronic rejection is untreatable.^2,9 One study of a cohort of patients with chronic lung allograft dysfunction found that 16 of 17 (94%) died in an ICU; 3 of these patients received multiple attempts at CPR prior to death. ¹⁵ In contrast, only 4 patients (19%) in our study cohort died in the ICU and none received CPR at the time of death. The high prevalence of aggressive end-of-life care may be due to a low prevalence of advance care planning in this population. ¹⁵ By discussing values and goals, PC consultants may be able to reduce the intensity and improve the quality of care at the end of life.

Similar to previous reports of LT recipients, ¹⁵ a majority of our cohort died in an acute care hospital. PC units and hospices would be appropriate for transplant recipients, but very few patients with end-stage respiratory illness are admitted to a PC unit.^15,16 Transplant teams develop strong relationships with their patients, and both parties may be reluctant to transfer to other facilities even when the goals shift to comfort care.

Our findings are notable also for the common use of opioids for symptom control. Physicians and family members may have concerns about the potential for opioids to compromise survival in patients with respiratory disease. ¹⁷ One study of 20 cystic fibrosis patients who were not transplanted found that only one had received opioids for more than a week prior to death. In contrast, almost all of our patients were receiving opioids, and none were diagnosed with opioid toxicity or had their death attributed to opioid-induced respiratory depression. Opioids are recommended for the treatment of end-stage lung disease and can safely be given to these patients.^18–23

This study has limitations. Its retrospective, single-center design makes it difficult to conclude cause-and-effect relationships between PC interventions and patient outcomes. However, because it is challenging to find an appropriate comparison group for this selected population, previous publications may be the best comparator. Because the end-of-life care described in these LT recipients differs sharply from that in the literature, we suggest that PC co-management had some impact on care. Goals-of-care discussions are inconsistently documented in patient records, and may include more than one physician or more than one encounter. Changes in resuscitation orders can occur some time after the initial PC encounter, and after encounters with other physicians, making it difficult to attribute that change to the PC consultation per se. In addition, we did not have regular reassessments of symptom control after the initial PC assessment, so we cannot conclude that PC referral or opioids improved dyspnea. However, opioids are effective for managing dyspnea, and their usefulness in the study population may be inferred from the fact that most patients continued to take the opioids and were maintained on stable doses.

In conclusion, we found that LT recipients who were referred for co-management by PC typically received comfort medications and avoided the aggressive end-of-life care that is usually reported for this population. PC co-management may represent an important intervention to improve the quality of care for LT recipients nearing the end of life. We plan to study the effect of PC interventions on patient quality of life and the benefits of early PC referral in a future prospective study.