Mary Clare Derfler, RN, MSN; Mark Jacob, RN, MSN; Robert E. Wolf, MS; Frank Bleyer, MD; Paul J. Hauptman, MD

Am J Geriatr Cardiol 13(6):299-304, 2004. © 2004 Le Jacq Communications, Inc.

Posted 11/29/2004

Abstract

The care of the end-stage patient has not been extensively studied, and little is known about best care practices. Therefore, using new definitions for mode of death due to heart failure, we performed a retrospective chart review of records from a university-based heart failure disease management program to characterize the population of patients dying from heart failure and to define clinical predictors that identify patients who will likely die of metabolic and/or progressive causes. Of 74 deaths recorded over a 60-month period, 17.6% and 21.3% were deemed to be metabolic or progressive, respectively. Utilization of resources was considerable, and only a small number of patients died while in hospice. Patients who required continuous inotropic support and those with preexisting renal failure were at highest risk for non-sudden cardiac death. We conclude that prospective identification of patients at risk for metabolic and progressive heart failure death is possible. The numbers of these patients is likely to increase in an era of implantable cardioverter-defibrillators. Intervention studies designed to evaluate and improve strategies that emphasize symptom control should target this group.

Introduction

Congestive heart failure (HF) is primarily a disease of the elderly.^[1-4] Despite improvements in medical therapy for both chronic and acute decompensated HF,^[5-7] the increasing prevalence of risk factors for the disease and the aging of the general population have given rise to estimates that individuals free of HF at age 40 have a lifetime risk of 21.0% for men and 20.3% for women.^[8] Since upwards of 15% of cases are considered severe (or New York Heart Association [NYHA] class IV) and mortality has not changed demonstrably for patients once the signs and symptoms of advanced disease develop,^[9-11] the need for management approaches at the end-of-life will increase. Nevertheless, the care of the end-stage patient has not, to date, been extensively studied, and little is known about best care practices. This knowledge gap is especially noteworthy because only a very small fraction of patients qualify for, or have access to, heart transplantation or the implantation of ventricular-assist devices as destination therapy, and many elderly patients are not enrolled in clinical trials.^[12]

For those patients who have exhausted standard medical therapy, palliative and/or hospice care represent important options,^[13] and hence insight into the terminal phases of the disease may have implications for the selection of appropriate therapies. However, little is known about the way in which end-stage patients progress to death.

Traditionally, the mode of death in HF has been defined by reductive terminology used in clinical trials, namely sudden cardiac (if the patient was last seen alive within 24 hours of death) or progressive HF death.^[14] However, criticisms about this scheme exist and reflect uncertainty about terminal events.^[15-17] For example, the Assessment of Treatment with Lisinopril And Survival (ATLAS)^[17] trial investigators state "sudden death is notoriously difficult to define...but is usually thought of as a death attributed to an opportunistic arrhythmia." Moreover, in adjudicating mode of death in clinical trials, interobserver discordance has been described.^[18] As a consequence, prior definitions may not accurately describe terminal events in patients with HF. Further, with the advent of implantable defibrillators, changes in the way in which patients die may be occurring with a shift toward progressive heart and multisystem failure and away from sudden arrhythmic events.^[19] Many of these deaths may result from metabolic causes, such as renal and hepatic failure brought about by organ hypoperfusion or congestion; they also may not be painful or difficult, especially if dyspnea is not prominent or is easily controlled.

Therefore, we sought to construct new functional definitions for mode of death and characterize the population of patients dying from HF in a university-based disease management (DM) program. Specifically, we hypothesized that the terminal phase for a considerable proportion of patients is characterized by metabolic complications. Further, we sought to define clinical predictors that identify patients who will likely die of metabolic and/or progressive causes, thereby providing a basis for critical evaluations of the timing and type of palliative and hospice care that might be appropriate to administer in this population.

We performed a retrospective chart review covering a 60-month period (October 1998-September 2003). The setting was a university-based HF disease management (HFDM) program staffed by two full-time Master's-prepared nurses and physicians with specialty training in HF, with a referral base including urban St. Louis city/county and rural eastern Missouri and southern Illinois. The Saint Louis University Institutional Review Board approved the study.

Patient Population

Patients were included if they had a primary diagnosis of congestive HF in the absence of hypertrophic cardiomyopathy; history of solid organ transplantation or evaluation for transplantation; metastatic or terminal forms of cancer; or a primary neurologic, neuromuscular, or pulmonary diagnosis. We excluded HF patients who were actively enrolled in a drug or device study. Patients had to attend at least one HFDM office visit within the 12 months preceding the date of death to be included in this analysis. Date of death was derived from medical records or, when not available, through the Social Security file (accessed at http://www.rootsweb.com).

Definitions of Mode of Death

We divided the mode of death into distinct categories as follows: metabolic (creatinine >4.0 mg/dL within 1 month of death or creatinine >3.0 mg/dL before referral to hospice or progressive hepatic failure); progressive (NYHA class IV symptoms without renal or hepatic failure); sudden arrhythmic (patient last seen alive within 1 hour of death ands unexpected); unwitnessed presumed sudden and arrhythmic (patient last seen alive within 72 hours of death with no new symptoms);^[20] other (includes trauma and sepsis); and unknown.

For the definition of comorbidities, chronic renal failure was defined by the presence of a creatinine >2.0 mg/dL measured more than once over a 2-month period; other variables were atrial fibrillation, hypertension, insulin-dependent diabetes mellitus, nicotine use, and prior coronary artery bypass graft surgery or implantation of an implantable cardioverter-defibrillator. Durations of time from first diagnosis of HF to date of first referral to the HFDM program and from first HFDM program visit to death and/or hospice were calculated; the numbers of outpatient cardiology visits, hospitalizations, hospital days, and cardiac diagnostic procedures or therapeutic interventions (including echocardiograms, stress tests, and right and left heart catheterizations) were tabulated.

Adjudication of Mode of Death

Three investigators (MCD, MJ, PJH) who had some clinical knowledge of the patients separately reviewed records and decided on the mode of death according to the established definitions. In the event of any disagreement, the group as a whole reviewed the records and consensus was achieved.

An investigator (FB) who had no prior knowledge of the patients or their clinical course performed a separate independent review of randomly selected cases. The definitions of mode of death were described in detail with, and clarifications were provided to, the reviewer during the adjudication process, as needed. Of 30 death records reviewed, there was agreement on 28 cases. In one case, a death classified as progressive was deemed to be metabolic by the independent reviewer; in the second case, a death classified as metabolic was deemed to be progressive. If details on the events surrounding the patient's demise could not be detected from any clinical document or chart, the mode of death was considered to be unknown.

Data Analysis

Data shown are mean ± standard deviation for patient age, durations of time from diagnosis to referral to the HF program, referral to hospice and death, and number of DM visits and hospitalizations per patient. Predictors of mode of death from metabolic or metabolic and progressive causes were determined using logistic regression, which controlled for the effects of key clinical covariates, including etiology and duration of HF; chronic atrial fibrillation; use of inotropic therapy (dopamine, dobutamine, milrinone); chronic renal failure; patient age; frequency of antecedent hospitalization; and presence of an implantable cardioverter-defibrillator.

A total of 74 deaths were recorded, distributed among three physician providers. The mean age was 57.7±15.4 years; 55 were male, 10 were African American, and 64 were Caucasian. The etiology of HF was ischemic in 45% and nonischemic including familial dilated, valvular, and idiopathic in 55%.

Comorbidities and concurrent conditions included chronic atrial fibrillation (n=12), chronic renal failure (n=20), hypertension (n=20), prior coronary artery bypass graft surgery (n=20), insulin-dependent diabetes mellitus (n=19), and chronic nicotine abuse (n=30). Fourteen patients had an implantable cardioverter-defibrillator at the time of death; none of these patients had the device inactivated before death. The left ventricular end-diastolic diameter measured 5.6±2.8 cm, measured at a mean of 365.3±411.4 days before death or hospice referral.

The time from first diagnosis to first referral to the DM program was 939±1386 days, and the time from first DM program visit to death was 1083±1044 days. A total of seven patients were referred to hospice; the time from first DM visit to initiation of hospice was 784±1325 days with survival of 11.0±11.3 days after referral. The mean age of the patients referred to hospice was significantly greater than the mean age of patients not in hospice (73.3±8.5 years vs. 56.1±15.1 years, p<0.01). Twenty-two patients were treated with chronic IV dobutamine (n=15) or milrinone (n=7) therapy for a mean of 203±224 days. None of the patients were treated with dopamine or a combination of two or more inotropic drugs. Scheduled intermittent IV inotropic therapy was not employed in any patient.

In the 12 months preceding death, utilization of resources included 6.26±5.91 outpatient visits/patient to the HF clinic and 2.39±2.18 hospitalizations (total hospital days: 1298; mean and median number of hospital days per patient: 17.5±21.1 and 10, respectively). The mean number of antemortem cardiac procedures/patient during this period was lower for the sudden death (unwitnessed presumed sudden: 1.00; sudden: 1.29) than for progressive (2.75) or metabolic (2.92) modes of death and was not related to patient age.

Mode of death was metabolic in 13 (17.6%), progressive in 16 (21.6%), sudden in 14 (18.9%), unwitnessed presumed sudden in 7 (9.5%), other in 14 (18.9%), and unknown in 10 (13.5%) patients (Table). Place of death was distributed nearly equally between home (n=27; 37%) and hospital (n=31; 42%), with a minority of patients dying in a skilled nursing facility or other location (n=9; 12%). We were unable to determine the place of death in an additional 9% (Figure).

Figure. Distribution of place of death. SNF=skilled nursing facility

In a multivariable analysis that excluded patients who did not have a known mode of death, the significant predictors of progressive or metabolic (nonarrhythmic) death were the use of inotrope (odds ratio [OR] 15.6, 95% confidence interval [CI] 2.9-83.3) and the presence of preexisting chronic renal failure (OR 8.0, 95% CI 1.4-45.1). There was a trend suggesting that the presence of greater than two antecedent hospitalizations in the year before death also predicted a metabolic or progressive death (OR 5.0, 95% CI 1.1-23.2). Gender was not a contributing factor. When the analysis compared metabolic death alone to all other causes, the same factors were identified as significant predictors.

Despite improvements in the outcome of patients with congestive HF related to the introduction of angiotensin-converting enzyme inhibitors,^[5,21] β-adrenergic blockers,^[6,22] and aldosterone antagonists,^[20,23] mortality remains high in the elderly cohort. Although much is known about factors associated with an increased risk of death from HF,^[24,25] there is uncertainty about how to describe the mode of death in this population,^[15,18] as reflected by heterogeneity of the definitions in the two broadest categories employed in clinical trials (sudden death and death from progressive HF). Within each category, there is also interobserver discordance. For example, Ziesche et al.^[18] demonstrated that a subgroup of Studies of Left Ventricular Dysfunction (SOLVD) investigators disagreed with the adjudication of death in the Vasodilator Heart Failure Trial-II (V-HeFT): the kappa statistic for interobserver agreement was only 0.22. The implications for clinical trial design, as well as clinical care, are significant. For the former, the ability of investigators to delineate the mechanism of action and clinical relevance of an intervention may depend on appropriate classification of terminal events. For the latter, uncertainties about mode of death add to the literature that points to significant gaps in physician knowledge about HF, including patient preferences^[26] and the power of prognostic markers to predict outcomes within the last 6 months of life.^[13] Therefore, the development of greater insight into the nature of disease progression has the potential to impact the process of selection of care options when the condition of the HF patient deteriorates.

Previously, the ATLAS^[17] investigators performed an adjudication of mode of death based on a classification scheme that included sudden unexpected (including witnessed and unwitnessed events, death observed but not attributed to acute "intractable" myocardial infarction, and resuscitated sudden death leading to demise shortly thereafter), HF death, myocardial infarction-related, and other cardiovascular etiologies (including pulmonary embolus, cerebrovascular accident, or death associated with a vascular procedure); the bulk of the deaths in the trial were associated with the first two categories. Another set of specific definitions for fatal events were proposed by Pitt and colleagues,^[20] who defined progressive HF death as one due to an increase in HF symptoms demonstrated by a change in medication for the treatment of HF which, in turn, is reflected by a 50% increase in dose of oral medication, addition of a new class of medication, or addition of IV medication. However, the time frame for these changes are not made clear, and there is room for different interpretation about the implications of a medication change, such as an increase in angiotensin-converting enzyme inhibitor.

We devised a classification that emphasizes progression of the sequelae of pump dysfunction, with a focus on end-organ hypoperfusion, in light of data suggesting that renal function is a potent predictor of outcomes.^[25,27,28] We avoid the less defined terms "HF death" and "medication change" to define worsening, since the addition or dose escalation of β-adrenergic blockers or angiotensin-converting enzyme inhibitors is often encouraged in stable patients to meet evidence-based therapeutic goals.^[29]

We found that more than 40% of HF deaths in a tertiary HFDM program are due to progressive disease or renal/hepatic failure. Many patients were treated for extended periods of time in the DM setting and were high utilizers of both inpatient and outpatient resources. Only a small minority was referred to hospice, and the length of stay was short. However, the number of patients dying with metabolic complications is likely to increase given the greater use of implantable cardioverter-defibrillators.^[19] Despite small sample size, these patients were more likely to be on inotropic therapy and have preexisting baseline chronic renal failure. Both variables point to the presence of low cardiac output syndrome. Further, there was a trend suggesting that a history of frequent antecedent hospitalizations was associated with a greater likelihood of a metabolic death.

These findings may influence the way in which caregivers approach patients who are failing conventional therapies. Since many patients have metabolic deaths (20% of those with known modes of death) and these can be predicted based on the presence of chronic renal dysfunction and the need for inotropic therapy, the initiation of palliative care^[13] should be strongly considered in patients who fit this profile. Further, interventions can be designed to target this high-risk group, including the withdrawal of aggressive diagnostic testing and emphasis on symptom control. However, it is noteworthy that fewer than 15% of HF patients in our series were referred to hospice, similar to data from the Acute Decompensated Heart Failure National Registry (ADHERE).^[30] The reasons for this are unclear, but may reflect the fact that while length of stay in hospice is similar for other diseases,^[31] uncertainties exist about the ability of current models to accurately predict HF mortality within 6 months.^[32] Further, the use of inotropes in the hospice setting is controversial.^[33]

Limitations

This was a retrospective, single-center review using definitions that have not previously been validated. The patient population is predominantly, but not entirely, elderly (mean age 57.7 years), reflecting the university setting in which the DM program exists. Therefore, this analysis may not translate to the larger HF population in the community setting where the prevalence of HF in patients over the age of 75 is far greater. In that context, it is noteworthy that the patients referred to hospice in this series were considerably older than the mean for the entire cohort.

Blinded adjudication was performed on only 40% of our sample, but the rate of agreement was significant. Small sample size as a consequence of the single-center design and missing data may affect the likelihood of detection of other important parameters, such as gender in the multivariable analysis. However, with regard to missing data, it is unlikely that mode of death can be established in every patient in any clinical or research setting.

We believe that prospective identification and/or stratification of HF patients at risk for metabolic death, and a greater understanding of the natural history of metabolic impairment, are key elements in choosing appropriate levels of care for those at or near the end-of-life. In addition, elucidation of physicians' and patients' preconceptions about the mode of death with HF may form the basis for studies designed to evaluate interventions and improve strategies in symptom control and palliative care. Patients most likely to benefit from those interventions are those whose trajectory suggests an inexorable course toward a metabolic demise.

CME Information

The print version of this article was originally certified for CME credit. For accreditation details, contact the publisher. Le Jacq Communications, Inc., Three Parklands Drive, Darien, CT 06820-3652.

1. Croft JB, Giles WH, Pollard RA, et al. Heart failure survival among older adults in the United States. Arch Intern Med. 1999;159:505-510.
2. Gillum RF. Epidemiology of heart failure in the United States. Am Heart J. 1993;126:1042-1047.
3. Popovic JR, Kozak LJ. National Hospital Discharge Survey: annual summary 1998. Vital and health statistics. Series 13, Number 148. Washington DC: Government Printing Office, September 2000 (DHHS publication number [PHS] 2000-1719).
4. Kupari M, Lindroos M, Iivanainen AM, et al. Congestive heart failure in old age: prevalence, mechanisms and 4-year prognosis in the Helsinki Ageing Study. J Int Med. 1997;241:387-394.
5. Garg R, Yusuf S. Overview of randomized trials of angiotensin converting enzyme inhibitors on mortality and morbidity in patients with heart failure. JAMA. 1995;273:1450-1456.
6. Packer M. Bristow MR, Cohn JN, et al. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. US Carvedilol Heart Failure Study Group. N Engl J Med. 1996;334:1349-1355.
7. Publication committee for the VMAC Investigators. Intravenous nesiritide vs. nitroglycerin for treatment of decompensated congestive heart failure. A randomized controlled trial. JAMA. 2002;287:1531-1540.
8. Lloyd-Jones DM, Larson MG, Leip EP, et al. Lifetime risk for developing heart failure: The Framington Heart Study. Circulation. 2002;106:3068-3072.
9. Senni M, Tribouilloy CM, Rodeheffer RJ, et al. Congestive heart failure in the community: trends in incidence and survival in a 10-year period. Arch Intern Med. 1999;159:29-34.
10. Senni M, Tribouilly CM, Rodeheffer RJ, et al. Congestive heart failure in the community: a study of all incident cases in Olmsted County, Minnesota in 1991. Circulation. 1998;98:2282-2289.
11. Ho KKL, Anderson KM, Kannel WB, et al. Survival after the onset of congestive heart failure in Framingham heart study subjects. Circulation. 1993;88:107-115.
12. Heiat A, Gross CP, Krumholz HM. Representation of the elderly, women and minorities in heart failure clinical trials. Arch Intern Med. 2002;162:1682-1688.
13. Goodlin SJ, Hauptman PJ, Arnold R, et al. Consensus statement: palliative and supportive care in advanced heart failure. J Card Fail. 2004;10:200-209.
14. Hinkle LE, Thaler HT. Clinical classification of cardiac deaths. Circulation. 1982;65:457-464.
15. Narang R, Cleland JG, Erhardt L, et al. Mode of death in chronic heart failure: a request and proposition for more accurate classification. Eur Heart J. 1996;17:1390-1403.
16. Packer M. Lack of relation between ventricular arrhythmias and sudden death in patients with chronic heart failure. Circulation. 1992;85:150-156.
17. Poole-Wilson PA, Uretsky BF, Thygesen K, et al. Mode of death in heart failure: findings from the ATLAS trial. Heart. 2003;89:42-48.
18. Ziesche S, Rector TS, Cohn JN. Interobserver discordance in the classification of mechanisms of death in studies in heart failure. J Card Fail. 1995;1:127-132.
19. Stevenson WG, Epstein LM. Predicting sudden death risk for heart failure patients in the implantable cardioverter-defibrillator age. Circulation. 2003;107:514-516.
20. Pitt B, Remme W, Zannad F, et al. Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N Engl J Med. 2003;348:1309-1321.
21. The SOLVD investigators. Effect of enalapril on mortality and the development of heart failure in asymptomatic patients with reduced left ventricular ejection fractions. N Engl J Med. 1992;327:685-691.
22. MERIT-HF Study Group. Effect of metoprolol CR/XL in chronic heart failure: metoprolol CR/XL randomized intervention trial in congestive heart failure. Lancet. 1999;353:2001-2007.
23. Pitt B, Zannad F, Remme WJ, et al. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study. N Engl J Med. 1999;341:709-717.
24. Aaronson K, Schwartz JS, Chen TM, et al. Development and prospective validation of a clinical index to predict survival in ambulatory patients referred for cardiac transplant evaluation. Circulation. 1997;95:2660-2667.
25. Fonarow GC, Abraham WT, Adams K, Adhere Scientific Advisory Board and Investigators. Risk stratification for inhospital mortality in heart failure using classification and regression tree methodology: analysis of 33046 patients in ADHERE. Circulation. 2003;108:IV-693:3151.
26. Califf RM, Vidaillet H, Goldman L. Advanced congestive heart failure: what do patients want? Am Heart J. 1998;135: S320-S326.
27. Gottlieb SS, Abraham W, Butler J, et al. The prognostic importance of different definitions of worsening renal function in congestive heart failure. J Card Fail. 2002;8:136-141.
28. Gattis WA, O'Connor CM, Gallop DS, et al. Predictors of 60-day death or rehospitalization in patients admitted with decompensate heart failure: insights from the IMPACT-HF Trial and Registry. Circulation. 2003;108:IV-695:3154.
29. Anonymous. Heart Failure Society of America guidelines for management of patients with heart failure caused by left ventricular systolic dysfunction—pharmacological approaches. J Card Fail. 1999;5:357-382.
30. ADHERE Registry Web site. Available at: http://www.adhereregistry.com. Accessed December 18, 2003.
31. Christakis NA, Escarce JJ. Survival of Medicare patients after enrollment in hospice programs. N Engl J Med. 1996;335:172-178.
32. Fox E, Landrum-McNiff K, Zhong Z, et al. Evaluation of prognostic criteria for determining hospice eligibility in patients with advanced lung, heart, or liver disease. SUPPORT Investigators. JAMA. 1999;282:1638-1645.
33. Rich MW, Shore BL. Dobutamine for patients with end-stage heart failure in a hospice program? J Palliat Med. 2003;6:93-97.

Reprint Address

Address for Correspondence: Paul J. Hauptman, MD, Division of Cardiology 15-FDT, Saint Louis University Hospital, 3635 Vista Avenue, St. Louis, MO 63110. E-mail: hauptmpj@slu.edu