End of Life Care in Patients With Heart Failure

Consensus Statement

End-of-Life Care in Patients With Heart Failure

DAVID J. WHELLAN, MD, MHS, 1 SARAH J. GOODLIN, MD, 2 MICHAEL G. DICKINSON, MD, 3 PAUL A. HEIDENREICH, MD, MS, 4 CONNIE JAENICKE, NP, RN-C MSN, 5 WENDY GATTIS STOUGH, PharmD, 6 AND MICHAEL W. RICH, MD, 7 ON BEHALF OF THE QUALITY OF CARE COMMITTEE, HEART FAILURE SOCIETY OF AMERICA

Philadelphia, Pennsylvania; Portland, Oregon; Grand Rapids, Michigan; Palo Alto, California; Minneapolis, Minnesota; Buies Creek, North Carolina; and

St. Louis, Missouri

ABSTRACT Stage D heart failure (HF) is associated with poor prognosis, yet little consensus exists on the care of pa-

tients with HF approaching the end of life. Treatment options for end-stage HF range from continuation of guideline-directed medical therapy to device interventions and cardiac transplantation. However, patients approaching the end of life may elect to forego therapies or procedures perceived as burdensome, or to deactivate devices that were implanted earlier in the disease course. Although discussing end-of-life issues such as advance directives, palliative care, or hospice can be difficult, such conversations are critical to understanding patient and family expectations and to developing mutually agreed-on goals of care. Because patients with HF are at risk for rapid clinical deterioration or sudden cardiac death, end-of-life issues should be discussed early in the course of management. As patients progress to advanced HF, the need for such discussions increases, especially among patients who have declined, failed, or been deemed to be ineligible for advanced HF therapies. Communication to define goals of care for the individual patient and then to design therapy concordant with these goals is fundamental to patient-centered care. The objectives of this white paper are to highlight key end-of-life considerations in patients with HF, to provide direction for clinicians on strategies for addressing end-of-life issues and providing optimal patient care, and to draw attention to the need for more research focusing on end-of-life care for the HF population. (J Cardiac Fail 2014;20:121e134) Key Words: Advanced heart failure, end-of-life care, hospice, palliative care.

Heart failure (HF) with either reduced ejection fraction tend to follow a variable course after the initial insult (eg, (HFrEF) or preserved ejection fraction (HFpEF) is charac-

coronary artery disease, myocardial infarction, genetic con- terized by a broad range of symptoms. Patients with HF

ditions, or environmental factors such as alcohol), but many progress owing to maladaptive remodeling and recurrent damage to the myocardium leading to the development of

From the 1 Department of Medicine, Jefferson Medical College, Phila- delphia, Pennsylvania; 2

worsening symptoms. The American College of Cardiology

Division of General Internal Medicine and Geri- atrics, Portland Veterans Affairs Medical Center and Oregon Health and

Foundation/American Heart Association (AHA) guideline

Science University, Portland, Oregon; 3 Frederik Meijer Heart & Vascular

for the management of HF characterizes HF progression

Institute, Spectrum Health, Grand Rapids, Michigan; 4 Center for Health

into 4 stages, in which stage A includes individuals with

Policy, Center for Primary Care and Outcomes Research, Veterans Affairs

Palo Alto Medical Center, Palo Alto, California; 5 Veterans Affairs Medical

risk factors for HF but without structural heart disease,

Center, Minneapolis, Minnesota; 6 Campbell University College of Phar-

stage B includes persons with structural heart disease

macy and Health Sciences, Buies Creek, North Carolina and 7 Washington

without HF symptoms, stage C represents symptomatic

University School of Medicine, St. Louis, Missouri. Manuscript received July 19, 2013; revised manuscript received

HF, and stage D reflects refractory symptoms despite December 5, 2013; revised manuscript accepted December 6, 2013. 1 guideline-directed medical therapy (GDMT).

Reprint requests: David J. Whellan, MD, MHS, Professor of Medicine,

Although a large body of evidence has accumulated to

Jefferson University, 925 Chestnut Street, Mezzanine, Philadelphia, PA 19107. Tel: 215-955-2007; Fax: 215-503-7420. E-mail: djw150@

guide the management of patients with chronic HF, there is

jefferson.edu

little consensus on the care of these patients near or at the

See page 132 for disclosure information.

end of life. Many factors warrant consideration in this popu-

lation, including prognosis, patient treatment goals, and

http://dx.doi.org/10.1016/j.cardfail.2013.12.003

available treatment options. Discussing end-of-life issues,

122 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 such as advance directives, palliative care or hospice, and de-

in the populations studied, and it highlights the significant vice deactivation, is critical to guiding patient and family ex-

heterogeneity of patients classified as stage D. pectations and helping them to cope with terminal illness and death. Appropriate strategies can be used to improve HF

Heart Failure With Preserved Ejection Fraction symptoms and quality of life throughout HF care, including the end-of-life period. Recognizing that most HF patients die

Up to 50% of HF patients have preserved left ventricular before stage D, it is important to address dying early in the

systolic function, and this form of HF becomes increasingly

course of HF because of the risk of sudden death and poten- more common with advancing age, especially in women. tial need for resuscitative measures. Patients’ preferences

Although symptom severity, hospitalization rates, and prog- regarding end-of-life care should be revisited periodically

nosis are similar in patients with HFpEF to those in patients with HFrEF, as the condition and prognosis evolve. 9 management of patients with HFpEF is

Despite the difficulty and complexities of end-of-life is- compromised by the lack of proven effective therapies. Thus, neither pharmacologic agents nor devices have

sues in HF patients, there is minimal evidence-based guid- been shown to reduce mortality in this population. 1,10 ance to inform the care of this population. The objectives of

this white paper are to highlight key end-of-life considerations in patients with HF, to provide direction for clini-

Predicting Prognosis in Stage D Heart Failure cians on strategies for addressing end-of-life issues and for providing optimal patient care, and to draw attention

The optimal treatment of HF depends in part on the pa- to the need for more research focusing on end-of-life care

tient’s expected survival. Although many medical treat- for the HF population.

ments improve outcomes at all stages of HF, use of ICDs does not clearly improve survival unless life expectancy

Defining the Stage D HF Population is $1 year. Accordingly, clinical guidelines recommend Epidemiology

against, and some payers will not reimburse, the use of ICDs if a patient is expected to live !12 months. In addi-

The prevalence of stage D HF has not been well docu- tion, a patient may choose to discontinue treatments that are mented. Approximately 5.7 million Americans $20 years

only life prolonging (ie, with no impact on symptoms), eg, of age have HF. 2 The proportion of these patients with stage

turning off the ICD function of a device if life expectancy is

D is uncertain, although it has been estimated to be 5% markedly limited or if quality of life is poor. Therefore,

3 e10%.

knowledge of one’s risk of dying in the next year may 300,000e600,000 patients in the USA with stage D HF.

These figures

suggest that there are

help patients and their families select the most appropriate treatment and optimal care setting. A patient with a mark-

Patient Characteristics edly shortened survival and poor quality of life may wish to

be managed outside of the hospital (eg, hospice). istry Longitudinal Module (ADHERE LM) enrolled 1,433

The Acute Decompensated Heart Failure National Reg-

Predicting the outcome for patients with stage D HF is patients with stage D HF. Patients with stage D HF were

challenging. One survey showed that physicians thought younger, more often male, and more likely to have a his-

that they could predict 6-month mortality ‘‘most of the tory of dyslipidemia, coronary artery disease, and chronic

time’’ or ‘‘always’’ in only 16% of their HF patients. 11 Pa- renal insufficiency than other patients hospitalized for

tients also have been poor at predicting their own survival. acute decompensated HF. Stage D patients were also

In a study using the Seattle Heart Failure Model (SHFM; more likely to have a permanent pacemaker or implantable

see below) to estimate life expectancy, patients with

chronic stable HF estimated a survival consistent with actu- tant implications for palliative care or hospice discussions.

cardioverter-defibrillator (ICD), 4 a finding that has impor-

arial data for individuals without HF, 3 years longer than The estimated 1-year survival in this population was

the model predicted. Younger age, increased New York 71.9%, and the estimated 1-year freedom from survival

Heart Association (NYHA) functional class, lower left ven- or hospitalization was 32.9%. This survival rate is higher

tricular ejection fraction (LVEF), and less severe depression than the 6-month survival of 67% in the Evaluation Study

were the most significant predictors of greater overestima- of Congestive Heart Failure and Pulmonary Artery Cathe-

tion by patients. Actual survival was more accurately pre- terization (PAC) Effectiveness (ESCAPE) PAC Registry,

dicted by the SHFM than by the patient’s estimate. which included patients hospitalized for decompensated

Given the difficulty in estimating prognosis, predicting HF who were not randomized into the main ESCAPE trial

survival for patients with HF has become a research priority. but still received PAC. 5 In the medical therapy arm of the

Many large studies have examined patient characteristics Randomized Evaluation of Mechanical Assistance for the

and treatments that are associated with a higher or lower Treatment of Congestive Heart Failure (REMATCH) trial

risk of death. 12,13 Several of these investigations have (a randomized trial of destination ventricular assist device

yielded algorithms to predict survival for patients with HF therapy versus optimal medical management in end-stage

with the use of information commonly available at the

time of a clinical encounter ( Table 1 ). 10e14 They differ in range of outcomes likely reflects underlying differences

patients), estimated 1-year survival was only 25%. 6 This

their outcome (survival to discharge or long-term survival)

Whellan et al 123 Table 1. ESCAPE Discharge Prediction Score

End-of-Life Care in HF Patients

is more difficult to assess prognosis in patients with HFpEF Criteria (Based on Discharge

than with HFrEF. Nonetheless, frequent hospitalizations Measurement)

Score if Yes (No 5 0)

and worsening symptoms, especially NYHA functional class IV, are markers of a downhill trajectory and high mor-

Age O70 y

1 tality within the ensuing 6e12 months, signaling the need

1 to discuss prognosis and end-of-life care options. BUN O90 mg/dL 19 Princi- 6-min walk !300 ft

BUN O40 mg/dL

1 ples of managing symptoms in patients with advanced HF

Sodium !130 mEq/L

1 in the setting of preserved ejection fraction are generally

CPR/mechanical ventilation, yes/no

Diuretic dose O240 mg at discharge, yes/no

1 similar to those in patients with reduced ejection fraction.

No beta-blocker at discharge

Discharge BNP O500 pg/mmol

3 Medical Management at the End of Life

Discharge BNP O1,300 pg/mmol

Total

0e13

Reprinted with permission from O’Connor et al, Triage after hospitalization

In addition to dyspnea, patients with advanced HF often

with advanced heart failure: the ESCAPE (Evaluation Study of Congestive

experience pain, weakness, fatigue, nausea, anorexia, con-

Heart Failure and Pulmonary Artery Catheterization Effectiveness) risk 109

stipation, edema, cough, altered mental status, anxiety,

model and discharge score, J Am Coll Cardiol 2010;55:872e8.

depression, and sleep disorders. 20e22 Management of these symptoms in patients approaching the end of life is briefly reviewed in the following sections.

and the patient population (eg, hospitalized, preserved systolic function). These algorithms are potentially useful for

Dyspnea

end-of-life HF care if they can identify patients at high risk of death during short-term follow-up (ie, 3e6 months), Dyspnea, or the perception of difficulty breathing, is

for whom the approach to management may be altered. The

a hallmark of advanced HF, and may become increasingly SHFM is the most widely used algorithm for predicting

severe in the terminal stage of illness. Management of prognosis, but it should be recognized that it was not derived

dyspnea includes maintenance of euvolemia through judi- from an advanced HF population; therefore, like many

cious use of diuretics in conjunction with dietary sodium models, it may underestimate risk in the advanced HF pop-

and fluid restrictions. 1,10 In patients with systolic HF, ulation. The online version of the SHFM allows one to

renin-angiotensin-aldosterone-system blockade (angio- determine the impact of different treatments on expected

tensin-converting enzyme [ACE] inhibitors, angiotensin re- survival. The model has been validated in numerous HF pa-

ceptor blockers [ARB], mineralocorticoid antagonists) tient populations, and it has been shown to be accurate for

should be titrated as tolerated to optimize cardiac perfor-

predicting risk of death for various patient groups, mance while minimizing adverse effects. Digoxin is

appropriate in patients with dyspnea persisting despite the dividual patients, the discrimination for 1-year mortality has

including those referred for cardiac transplantation. 16 For in-

above measures. 1,10 Similarly, intravenous inotropic ther- varied from 0.68 to 0.81 (ie, good to excellent, where 0.5 in-

apy with dobutamine or milrinone may provide palliation dicates no discrimination and 1.0 indicates perfect predic-

of symptoms and improve quality of life in some patients, tion) depending on the validation cohort. In the original

but this approach requires discussion with the patient and cohort, the model accurately identified patients at the end

family about the potential for increased arrhythmia and mortality. of life who had a 1-year mortality of 84%, although only 1,10

0.4% of the population had such a high risk. A slightly Opioids have demonstrated efficacy and safety for allevi- larger group (3.3%) had a 1-year mortality of 51%, and

ating dyspnea in patients with advanced HF and are consid-

though they are not clearly at the end of life, this informa- ered to be first-choice adjunctive agents. Relatively tion would be important for selecting device therapy.

low dosages, such as 0.05 mg hydromorphone or 1 mg oxy- With the Enhanced Feedback for Effective Cardiac

codone, are often sufficient. Toxic metabolites excreted by

Treatment (EFFECT) model, 17 patients with a score of

the kidneys may accumulate, so patients requiring regular doses of opioids should be converted to fentanyl or metha-

!60 had only 8% 1-year mortality and those with a score of 121e150 had 59% 1-year mortality and therefore would done, which do not have renally excreted metabolites. Ox-

not be candidates for ICDs. Those with a score of O150 ygen is appropriate in patients with hypoxemia, but it has had 79% 1-year mortality and would be considered to be

not been shown to be helpful in patients with preserved

at the end of life. oxygenation. Benzodiazepines may be useful for reducing Patients with HFpEF tend to be older than patients with

anxiety associated with dyspnea. 26 Other therapies, such as HFrEF, and they are more likely to have multiple comorbid

dietary supplements, relaxation techniques, and thermal therapy, are of unproven benefit. illnesses that may affect prognosis. As a result, patients 21

with HFpEF are less likely than patients with systolic HF

Pain

to die from either progressive HF or ventricular tachyar- rhythmias and more likely to die from a noncardiovascular

Although pain is not usually considered to be a symptom

cause, such as pneumonia or cancer. 18 For these reasons, it

of HF, pain is common in advanced HF. 27 When possible,

124 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 the underlying cause of pain should be treated appropriately

with severe constipation, an enema and/or manual disim- (eg, antianginal agents for ischemic chest pain). For other

paction may be necessary to provide relief. pain, opioids are first-line agents, and the dose should be titrated to provide adequate relief, avoiding opioids with

Depression and Anxiety

active metabolites that accumulate in renal impairment. 20,21 Fentanyl may be delivered via oral-buccal or topical routes

Approximately 30%e35% of patients with advanced HF have clinical depression, and the presence of depression

for patients with moderate to severe pain; however, it is not correlates with higher symptom burden and increased risk FDA approved for noncancer pain. Nonsteroidal antiinflam- of adverse outcomes, including hospitalizations and mortal- matory drugs (NSAIDs) should generally be avoided owing ity. 20,21 Depressed patients also tend to be less adherent to to adverse effects on renal function, sodium and fluid reten-

tion, and gastrointestinal side effects. 20,21,28

medications and behavioral interventions. Options for treat- ing depression in patients approaching the end of life include cognitive behavioral therapy, spiritual support (eg,

Fatigue and Weakness clergy), and medications. 20,21 Selective serotonin reuptake Fatigue and weakness in patients with advanced HF are

inhibitors (SSRIs) at low doses are generally considered usually multifactorial, related in part to cardiac insuffi-

to be first-line agents, but these drugs can induce fluid ciency, loss of muscle mass, deconditioning, and comorbid

retention and hyponatremia in patients with renal insuffi- conditions (eg, anemia, thyroid dysfunction, sleep disor-

ciency, 20 and some prolong the QTc interval. The efficacy ders, depression). Cardiac performance should be opti-

of these agents in improving depression scores in patients mized in accordance with current guidelines, and

with stage D HF remains to be established. 32 Tricyclic an- coexisting illnesses, if present, should be treated appropri-

tidepressants, such as nortriptyline or desipramine, are ately. Regular exercise, including aerobic exercise and

alternatives to SSRIs, but side effects, including prolonga- resistance training with light weights, should be encouraged

tion of the QT interval, are relatively common, especially in patients who are able to exercise. 20,21 Stimulants, such as

at higher dosages. 20,21 The onset of antidepressant effect methylphenidate, may be beneficial in some cases. 20,21

is 1e2 weeks or longer with both SSRIs and tricyclics, a major disadvantage in patients approaching the end of

Gastrointestinal Disorders life. In contrast, the onset of action of psychostimulants, Advanced HF is often associated with increased catabo-

such as methylphenidate, is 1e2 days, and these agents may provide significant relief from depression in patients

lism in conjunction with anorexia, leading to cardiac

with advanced HF. 20

cachexia. 29 HF therapy should be optimized, because there Few studies have examined the prevalence and treatment is some evidence that ACE inhibitors and carvedilol have of anxiety in patients with advanced HF, but it is likely that favorable effects on energy metabolism and the manifesta- an increasing proportion of patients experience anxiety as

tions of cardiac cachexia. 29 High-energy nutritional supple-

HF symptoms progress. 20,33 Interventions that enhance pa- ments may be useful for malnutrition, but there is no tients’ and spouses’ sense of control over HF (ie, self-

evidence that they improve clinical outcomes. 29 Similarly,

efficacy) have been shown to reduce emotional distress; appetite-promoting agents, such as megestrol acetate, are such interventions include support groups and HF educa- of uncertain benefit in HF patients, and they can promote

tion. 34

fluid retention. 29

Short-acting benzodiazepines, such as lorazepam, are effective in alleviating anxiety that does not respond

Nausea occurs in some patients with advanced HF, likely

to nonpharmacologic therapies. 20

due to a combination of reduced intestinal perfusion and

medication side effects. 30 A careful review of medications

Sleep Disorders

should be performed, and all nonessential medications should be discontinued. Aspirin, a common cause of

Approximately 50% of patients with HF have sleep- nausea, should be administered at low dosage (75e81

disordered breathing, including central and obstructive mg/d) and with food, and persistent unexplained nausea

sleep apnea. 35 In addition, Cheynes-Stokes respirations should prompt discontinuation of aspirin, at least tempo-

become increasingly common in patients with advanced rarily. Antiemetic agents, such as prochlorperazine or

HF. 35 Continuous positive airway pressure (CPAP) is the ondansetron, should be considered in patients with intrac-

mainstay of therapy for sleep-disordered breathing. 36 How- table nausea, recognizing that these drugs have the potential

ever, many patients with advanced HF are unable to tolerate for inducing significant side effects.

CPAP, and hypoxia may precipitate worsening HF. 37 Constipation is also common in patients with advanced

Nocturnal oxygen therapy has benefits nearly equivalent HF, due in part to decreased food intake, physical inactivity,

to those of CPAP for patients with obstructive sleep and medication side effects (especially from opioids). 30 apnea. 38e40 The utility of other interventions, such as opi-

Treatment of constipation should include consumption of oids and anxiolytic agents, in alleviating the distress associ- high-fiber foods, including vegetables, fruits, and whole

ated with sleep-disordered breathing is unknown. grains. 31 Laxatives, such as polyethylene glycol 3350 and

Insomnia also is a common problem in patients with lactulose, should be administered as needed. In patients

advanced HF. It is often multifactorial, due to orthopnea,

Whellan et al 125 nocturia, sleep-disordered breathing, and psychologic

End-of-Life Care in HF Patients

treatment of nocturnal hypoxia, aggressive diuresis, often

with an intravenous loop diuretic alone or in combination insomnia begins with appropriate management of the un-

distress (ie, anxiety and depression). 41 Treatment of

with a thiazide (eg, metolazone), and optimization of other derlying cause(s) when feasible. Attention to sleep hygiene,

HF medications. 1,10 Dietary salt and fluid restriction is including development of a regular routine before going to

appropriate, and use of NSAIDs and other sodium/fluid re- bed and avoidance of caffeine, alcohol, and excess fluid

taining medications should be avoided. 1,10 Elevating the

legs, when feasible, may afford some relief. Thigh-high odone (25 mg) is effective for inducing sleep with minimal

intake during the evening hours, is also appropriate. 42 Traz-

support stockings may also be beneficial in some cases, side effects but may cause delirium in older patients. Zolpi-

but many patients find them to be uncomfortable and diffi- dem (5 mg) is often effective in patients with more severe

cult to use.

insomnia, but it may induce confusion, falls, or daytime fatigue. Other agents that may be useful in selected cases

Inotrope Infusions

include mirtazapine and nortriptyline. Continuous home inotrope infusions can be a palliative

Confusion and Delirium tool that allows advanced HF patients who are not candidates for, or who do not desire, a mechanical circulatory

Advanced HF has been linked to cognitive impairment, support device (MCSD) or transplantation to be discharged presumably due to impaired cerebral blood flow and/or re-

43 petitive microemboli. home. In patients who develop recurrent symptoms during HF is also one of the most common causes of delirium in hospitalized patients. 44

attempts to wean inotropic therapy, discharge to home

with continuous inotrope infusions are associated with an such as opioids and benzodiazepines, may contribute to

Medications,

initial good level of functioning. In one study, although me- confusion or delirium in HF patients. Treatment includes

dian survival was only 3.4 months, the majority of patients reducing the dosage or eliminating potentially offending

died at home and avoided repeated hospitalization. 47 medications if feasible. In patients with relatively low blood

Continuous home inotrope infusions may be an effective pressure (eg, !100 mm Hg), decreasing the dose of beta-

strategy for palliative management of selected end-stage pa- blockers and/or ACE inhibitors to allow the blood pressure to rise may also be beneficial. 45

tients. In a small cohort of patients, home inotrope infusions

were associated with a 70% reduction of hospital days. 48 to ensure adequate sunlight during the day and avoidance of

Regulating the environment

In a risk-adjusted model, continuous milrinone and do- sleep disturbance at night is also important. In patients with

butamine infusions appeared to be associated with equiva- recurrent or persistent delirium, especially if associated

lent survival rates. 49 The combination of a beta-blocker with agitation or combativeness, small doses of haloperidol

and continuous intravenous milrinone is effective and may be effective, but patients should be monitored closely

may allow optimization of medications and eventual wean- for development of extrapyramidal side effects.

ing of milrinone. 50 The combination of a beta-blocker with dobutamine yields opposing hemodynamic effects and is

Cough

not recommended. 51e53

Cough is another common symptom in patients with In all published reports, continuous inotropic therapy advanced HF. 30 Cough is often worse at night, interfering

was used only after attempts at optimization of nonino- with sleep. Pulmonary congestion and ACE inhibitors are

tropic medications or weaning of inotropes failed. Candi- the most common causes of cough in HF patients. Other

dates for continuous inotropic infusions include those causes include aspiration pneumonitis and bronchitis. 46 who exhibit significant improvement in clinical status on

Treatment of cough includes diuresis in patients with evi- inotropes and who can not be weaned without return of dence of pulmonary congestion and substituting an ARB

disabling symptoms. Patients should be counseled that ino- for an ACE inhibitor if appropriate. Cough suppressants,

tropes are a palliative measure used to facilitate discharge such as dextromethorphan, may be useful for some patients.

home and improved out-of-hospital functioning in the In more severe cases, opioids, especially codeine, are often

period before death, but that inotropes are not associated effective. 46 Excess secretions may also be problematic in

with improved survival.

patients with advanced HF. In most cases, secretions can

be effectively managed with periodic suctioning with the

Discontinuation of Medications

use of a hand-held device. As HF progresses, goals of care may change from a principal focus on extending life to a primary emphasis on con-

Edema trolling symptoms and maximizing quality of life. In this

Patients with advanced HF may have persistent lower ex- context, the role of disease-modifying medications, such as tremity edema and ascites, which, if severe, may be associ-

ACE inhibitors and beta-blockers, may become less impor- ated with considerable discomfort and interfere with

tant than other therapies aimed primarily at palliation of activities, including performance of activities of daily living

symptoms. In such cases, it may be appropriate to reduce (eg, dressing, bathing, toileting). Treatment of marked

the doses or discontinue ACE inhibitors and/or beta- edema and/or ascites in patients with advanced HF includes

blockers, especially if these agents are causing adverse

126 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 effects or are otherwise contributing to impaired quality of

quality of life. For patients who have advanced symptoms life (eg, owing to polypharmacy or increased medication

(NYHA functional class IV) or poor prognostic factors, costs). However, ACE inhibitors may help to control symp-

ICD implantation should be discouraged. Many patients toms through afterload reduction. The impact of discontinu-

will have had an ICD placed some time earlier and it is ing guideline-recommended therapies needs further research.

appropriate that these patients understand the option to deactivate the defibrillator when they approach the end of life.

Comorbid Conditions Patients with HF often have $1 coexisting conditions, Cardiac Resynchronization Therapy

among the most common of which are coronary artery dis- Cardiac resynchronization therapy (CRT) reduces symp- ease, hypertension, diabetes, renal insufficiency, atrial fibrillation, and chronic lung disease. 54,55

toms in selected patients with NYHA functional class IIe

IV systolic HF. 60e64 However, patients with advanced HF detailed discussion of comorbid conditions is beyond the

Although a

accounted for only 4.2% of subjects enrolled in the CRT scope of this paper, in general, comorbidities in HF patients

clinical trials. 65 Subsequent observational reports have at the end of life should be managed in accordance with ex-

had mixed findings. Left bundle branch block (LBBB) pa- isting guidelines, keeping in mind that the principal goal of

tients (n 5 545) who met qualifications for being listed for therapy is to minimize symptoms and maximize quality of

heart transplantation and who instead underwent CRT im- life. Therefore, whereas treatment of ischemic chest

plantation demonstrated 1- and 3-year freedom from HF discomfort and symptomatic atrial fibrillation is appro-

death of 92.3% and 77.3%. 66 These data suggest that priate, aggressive management of hypertension and dia-

CRT implantation is reasonable in advanced HF patients betes is usually not warranted.

with LBBB in lieu of or before listing for heart transplant. Limited nonrandomized series reported benefit from Review of Therapies for Advanced Heart Failure

CRT in inotrope-requiring or inotrope-dependent patients. 67 In one of these studies, 9 out of 10 inotrope-dependent pa-

In addition to GDMT, management of the patient with tients were able to be weaned from inotropes after CRT im- advanced HF often includes device therapy. Unique consid- 68 plantation. However, another series showed that mortality

erations exist for these therapies as patients progress toward was still very high (60% after 9.5 months) in patients who end-of-life care.

were inotrope-dependent at the time of CRT implantation, raising concern about the cost-effectiveness of CRT implan-

Implantable Cardioverter Defibrillators tation in inotrope-dependent patients. 69 Thus, CRT may provide benefit to selected patients with advanced HF, but

Although it is clear that ICDs reduce mortality rates in the potential benefits and risks should be assessed on an in- patients with mild to moderate systolic HF, 56,57 the value

dividual basis.

of ICDs in patients with advanced HF is unproven. In the Sudden Cardiac Death in Heart Failure Trial (SCD-

Cardiac Transplantation

HeFT), the overall group had a significant mortality benefit, but the subgroup of patients with NYHA functional class III

Cardiac transplantation is an established therapy associ- HF did not realize a statistically significant survival

ated with good long-term survival and marked improve- benefit. 56 To date, no study has demonstrated survival

ment in functional capacity in appropriate patients. About benefit in NYHA functional class IV patients. ICD therapy

5,000 patients worldwide receive heart transplants each is therefore indicated for patients with left ventricular sys-

year (2,000e2,200 annually in the USA). The median sur- tolic dysfunction but is not indicated for patients who ‘‘do

vival after heart transplant is 10 years. The overwhelming not have a reasonable expectation of survival with an

majority of heart transplant patients function without phys- acceptable functional status for at least 1 year, even if

ical limitations. 3 These factors suggest that heart transplant

should be considered as an option for appropriately selected is also not indicated for ‘‘NYHA class IV patients with

they meet ICD implantation criteria.’’ 58 ICD implantation

patients with advanced HF.

drug-refractory congestive HF who are not candidates for Transplant patient selection has been summarized by cardiac transplantation or CRT-D.’’ 58 Mancini and Lietz. 3 Patients who are inotrope dependent, Many physicians and patients overestimate the benefit of

have a peak oxygen consumption (VO 2 ) during cardiopul- ICDs. For example, in one study, patients estimated a 50%

monary exercise testing of !10 mL min 1 kg 1 , or an survival benefit rather than the actual 7%e10% mortality 59 SHFM-estimated survival of !80% at 1 year are potential

reduction. Therefore, clinicians should include an honest transplant candidates. Those with a VO 2 of 10e14 mL and evidence-based discussion of the risks and benefits of

min 1 kg 1 or an SHFM-estimated survival of 80%e90% ICDs for patients with advanced HF. The risk of sudden car-

are considered to be of intermediate risk andcan be consid- diac arrest should be balanced against the risks of a nonar-

ered on an individual basis for transplant candidacy. Best rhythmic death (eg, progressive HF with declining

utilization of available donor organs dictates that patients functional capacity) or death from a non-HF cause, the risks

selected for transplant should be both ill enough to require of ICD implantation, and the impact of ICD shocks on

transplantation but well enough to have good long-term

Whellan et al 127 outcome after transplantation. Most transplant centers

End-of-Life Care in HF Patients

advanced HF, discussions about planning for the end of exclude patients O70e75 years of age and those with

life become more important, especially for patients who advanced comorbidities.

have declined, failed, or been deemed to be ineligible for advanced HF therapies.

Mechanical Circulatory Support Devices Basic principles of communication should be applied to MCSDs serve as a bridge to heart transplantation, as

discussions regarding goals of care and end-of-life prefer- destination therapy (DT) in patients who are not candi-

ences, including the use of simple clear language, avoid- dates for heart transplantation, or as a bridge to recovery

ance of euphemisms, and defining technical or medical in selected patients. 6,70e73 Outcomes have improved

terms as they are used. Communication should begin with over time owing to technologic advances and increasing

a query to identify what the patient understands or feels center experience. 74e76 Improved patient outcomes have

in a given situation, followed by providing information to led to significant growth in the use of MCSDs as DT.

educate the patient or correct misunderstandings. The con- Consideration of this therapy in patients with high risk in-

versation should end by asking the patient to explain back dicators is therefore appropriate. Illness severity indicators

what was said and by providing the patient an opportunity for selection of patients for MCSD are similar to those for

to ask questions. This ‘‘Ask-Tell-Ask’’ format ( Table 2 ) is cardiac transplantation and include patients who are 81 the foundation of patient-centered communication. Con-

inotrope dependent, have a maximal oxygen consumption versations may need to be broken into a series of discus- during cardiopulmonary exercise testing of !14 mL

sions, depending on patient and family needs. As many as min 1 kg 1 , or have an estimated 1-year SHFM survival

1 in 5 patients may not want to know their prognosis spe- of !90%. 3

cifically, so asking what information patients want is an Patient outcomes with MCSD are improved by appro-

essential step. 82

priate patient selection. 77 Recognition of high risk features

and calculation of a risk score can aid in patient selection Patient and Family Expectations by identifying individuals who are too sick (ie, HF is too

Many HF patients and their families are unaware of the far advanced or life-limiting comorbidities are present) to

life-limiting potential of HF. 83 Prognostic information realize meaningful benefit. Some patients may be at such

should always be balanced with a statement of the physi- high risk that MCSD implantation would represent ‘‘proce-

cian’s commitment to work with the patient to prolong dural futility.’’ 78 Risk scores have been developed to help

life as well as improve the quality of life, and of the ability assess likelihood of survival after MCSD. Details of patient

of medications and devices to achieve these goals. A con- selection and management are beyond the scope of this pa-

versation about the general course of HF gives clinicians

an opportunity to motivate HF patients and their families guidance has been provided by the International Society for Heart and Lung Transplantation. 80

per but have been summarized by Slaughter et al, 79

and

to actively ‘‘fight’’ HF with medical management and life-

style changes. Partnering with patients and their families to in the care of patients with advanced HF should be familiar

Clinicians involved

help them to identify warning signs of worsening status and with the available risk stratification tools to appropriately

offering them clear instructions about what to do when discuss potential MCSD therapy with their patients. They

symptoms increase can empower patients and help them should be able to use these tools to identify patients who

and their families to cope with the illness. do not meet criteria for MCS as well as those who are

too ill to have sufficient likelihood of good outcome. Pa-

Discussing Dying

tients who might benefit from MCSD implantation should

be educated about the potential risks and benefits to enable Because communication skills for addressing end-of-life informed decision making. Palliative medicine consultation

issues and death are often not incorporated into medical before MCSD placement can facilitate decision making and

training, clinicians caring for HF patients need to acquire provide guidance if adverse events occur or when MCSD

these skills to discuss dying with patients and families. At must be discontinued.

times when death is a greater threat, clinicians should have a conversation about the patient’s status and available

Approaches to End-of-Life Communication in interventions to either allow natural death or attempt to Heart Failure Table 2. ‘‘Ask-Tell-Ask’’ Methodology

Comprehensive HF care integrates education and support for the patient and family, as well as communication and

Ask

‘‘Tell me what you believe is going on in your illness’’ ‘‘As you look back, what has been important in your

decision making with evidence-based therapy throughout

life?’’

the course of illness. 20 Patients with HF are at risk for rapid

‘‘What are your concerns and worries?’’

clinical deterioration (eg, following a myocardial infarction

Tell and Partner ‘‘Heart failure is a disease that can last for years, but that most people die from. My goal is to work with

or other acute illness) or sudden cardiac death, so the pos-

you to do our best to help you ____’’

sibility of dying from HF should be acknowledged early in

Ask

‘‘What are your questions?’’

the course of management. As patients progress to

‘‘Tell me what you understood from our discussion.’’

128 Journal of Cardiac Failure Vol. 20 No. 2 February 2014 forestall death. Because the patient is often new to the clini-

what to do in an emergency.’’ The decision about CPR is cian at such encounters, the conversation should begin by

best initially presented as a big-picture decision of whether asking what the patient understands and expects.

to ‘‘allow natural death’’ or ‘‘try to revive you’’ ‘‘when

A conversation about dying is nearly always a ‘‘bad news your heart stops.’’ Clinicians should avoid presenting a conversation.’’ 84 Clinicians should acknowledge that the

menu of interventions. Rather, a general approach should be topic is sad or distressing, and learn to provide empathetic

identified, with boundaries for limiting therapies. Many pa- responses to patients or families. Examples of empathetic

tients want all potentially effective therapeutic measures as responses include identifying emotions both on the clini-

long as they have a chance to return to independent function. cian’s and patient’s parts, and using ‘‘wish’’ statements

However, many patients prefer being allowed to die rather such as ‘‘I wish things were different.’’

than being kept alive in a condition in which they would not be cognitively intact, eg, after having had significant brain

Clarifying Goals of Care for Heart Failure Patients injury after CPR. When a patient prefers an attempt at CPR, then it is reasonable to ask about conditions under which the

Communication to define the goals of care for an individ- patient would not want prolonged life support. ual patient, and then to identify options and make decisions Some patients will have considered resuscitation prefer- about interventions, are fundamental to patient-centered ences at an earlier stage of their illness. For other patients, care. Goals for care must be set in light of the patient’s clin- advance care planning has been avoided, so the topic can be ical condition, and they need to be reevaluated at turning introduced and the patient and family can be provided writ- points in care, such as decompensation, stabilization, and ten material along with a promise that the clinician will adverse events. revisit the subject at a later date. Patients who choose to Clinicians should begin a conversation about goals of not state a preference should be informed that the default care by specifically identifying the situation and the need intervention in the United States is to attempt CPR. Some to clarify goals. Statements such as ‘‘we have choices, patients will not want to consider the topic, and these indi- and which direction we go will depend on what is important viduals should be asked to designate a proxy who can make to you, as well as what is medically possible’’ or ‘‘we need decisions on their behalf. Even when a surrogate decision to make a decision about what path to take now’’ help to maker is identified, patients should be encouraged to indi- frame the conversation. cate global preferences, if possible. Beginning a discussion of treatment goals by understand- As the patient’s health status evolves throughout their ing what a patient has valued in their life allows the clini- illness, preferences for an attempt at CPR may change. 85 cian to integrate that information into the care plan. Clinicians should revisit preferences for resuscitation with Treatments that are appropriate to the patient’s health state changes in status and with changes in what is medically and HF status can be identified as consistent with their reasonable to offer the patient. Table 3 summarizes values, and presenting these gives the clinician a chance commonly used advance directives for expressing end-of- to check back with the patient.

life care preferences.

In advanced HF, some patients may specifically wish to Designation of a health care proxy, also called ‘‘Durable avoid surgery or to not be hospitalized again, and these Power of Attorney for Health Care,’’ assigns one or more per- preferences clearly direct the path of care, as well as help sons to make decisions on the patient’s behalf (using what is to set goals. A useful concept in conversations to set goals known of the patient’s views and values, or ‘‘substituted judg- of care is ‘‘Hope for the best, and plan for the worst.’’ This ment’’) should they lose capacity, even temporarily. allows clinicians to identify what a patient or their family Living Will and 5-Wishes ( www.agingwithdignity.org/ hope for, as well as to acknowledge that a plan is needed five-wishes.php ) documents provide statements about the for a variety of undesired conditions, such as death, stroke patient’s philosophy for medical treatment if they lose ca- or unconsciousness, prolonged hospitalization in intensive pacity in the future and are not likely to regain it. These di- care, or multisystem organ failure. rectives usually include a statement about life-sustaining

therapy, although some Living Will documents are Advance Care Planning

restricted to ‘‘terminally ill’’ states. Newer directives often stipulate undesired states, such as ‘‘unable to communicate

As part of the ‘‘planning for the worst’’ conversation, the or interact with others,’’ and indicate which therapies might topic of advance care planning can be addressed. A general

be continued or discontinued.

approach to either sudden or progressive HF death can be dis- Many states also have a legal order form for interven- cussed at the time of HF diagnosis or in association with an

tions, the Physician Order for Life-Sustaining Treatments event such as hospitalization or hospital discharge. The

(POLST), which is a transportable order signed by a physi- increased risk of sudden cardiac death in HF patients makes

cian that stays with the patient and is to be followed by it important to have a plan for whether to allow natural death

emergency personnel wherever the patient is living or or attempt cardiopulmonary resuscitation (CPR) when the

receiving health care. This document stipulates whether heart stops. This conversation can be normalized by saying

CPR should be attempted or not, whether the patient should ‘‘all HF patients and their families should have a plan for

be hospitalized and under what circumstances, and

End-of-Life Care in HF Patients

Whellan et al 129

Table 3. Advance Directives

Directive

Content Durable Power of Attorney for

Purpose

Nuances

Assigns decision making when the Health Care or Healthcare Proxy

Designates one or more individuals

Form set by state statute or law;

person is not capable (DPOA/HC)

to make decisions on the patient’s

usually requires a witnessed

behalf should she or he lose

signature

capacity

Living Will

A statement of preferences for care

Language varies by state, but may

Usually specifies life-prolonging

in future states

require 2 physicians to state that

treatment such as ventilation or

nutrition and hydration Five Wishes

the patient is ‘‘terminally ill’’

Legal advance directive in 42 states

May provide guidance even if not a

Includes DPOA/HC and preference

legal advance directive

for types of treatment, level of comfort, approach to care, and information for loved ones

Physician Order for Life-Sustaining

Stipulates whether or not Treatments (POLST)

Transportable order for emergency

Must be signed by a licensed

care; stays with the patient, to be

physician, nurse practitioner, or

cardiopulmonary resuscitation

followed wherever the patient is

physician assistant; designed for

should be initiated, the patient

living or receiving health care;

patients with high likelihood of

should be transported to the

some states have a registry that can

dying but in some states broadly

hospital, and intensity of

be accessed by emergency or

used in long-term care settings

interventions

medical personnel

addresses treatments such as feeding tubes or fluids and an- Medicaid, private insurance or managed care programs, or tibiotics. Some states also permit a written physician order

charity/donations. 90 Although hospice patients are often regarding resuscitation, such as ‘‘Do Not Resuscitate’’ on

cared for in the home, patients can also be cared for in an official prescription. Completing a POLST document

nursing homes or specialty hospice units. Brief ‘‘general or other order set involves identifying the patient’s prefer-

inpatient care’’ in hospitals or nursing homes contracting ences for specific interventions in the future; however, a

with the hospice agency is also permitted. POLST form is not a substitute for a careful discussion of preferences and planning for future care.

Hospice Referral in HF: Statistics and Barriers Historically, hospice has been underutilized for HF pa-

Palliative Care and Hospice tients, and heart disease has accounted for only 11.4% of hospice admissions in recent years. 91 In one study, hospice

Characteristics and Reimbursement utilization in the last 6 months of life for Medicare patients

Hospice and palliative care both improve quality of life with HF increased from 19% to 40%. 92 Yet in the AHA Get through symptom and pain management using a holistic,

With the Guidelines program, the utilization of hospice ser- multi-disciplinary approach. This organized, comprehen-

vices among all patients was !10%, even among those in sive philosophy addresses typical issues that arise

the highest decile of mortality risk. 93 Factors associated throughout care and at end-of-life including psychosocial

with hospice referral included increased age, low systolic and emotional responses to the diagnosis, spiritual issues,

End of Life Care in Patients With Heart Failure

Consensus Statement