Wang R, Mouliswar M, Denman S, Kleban M. Mortality of the Institutionalized Old-Old Hospitalized With Congestive Heart Failure. Arch Intern Med. 1998;158(22):2464-2468. doi:10-1001/pubs.Arch Intern Med.-ISSN-0003-9926-158-22-ioi80065
Congestive heart failure is a major cause of mortality and morbidity in the elderly but the disease impact on the oldest and sickest population has not been defined.
To review the mortality and hospital readmission rate of institutionalized elderly persons with congestive heart failure and to examine the relation of baseline characteristics to subsequent clinical outcomes.
This was a retrospective analysis based on chart review of 231 residents of the Philadelphia (Pa) Geriatric Center (63 congregate housing tenants and 168 nursing home residents) 80 years and older, hospitalized with congestive heart failure from 1989 to 1995. Patients' demographic data and clinical, electrocardiographic, and echocardiographic findings were obtained from their initial (index) hospitalization records. Subsequent outcomes were obtained from their outpatient (nursing home or office) records.
Thirteen percent died during the index hospitalization but the total mortality during the follow-up period was 87%. One hundred forty-six patients (63%) died in the first year with a mean±SD survival of 4±4 months and a readmission rate of 3.9 per patient-year. Eighty-five patients survived the first year with a readmission rate of 1.2 per patient-year and 54 patients subsequently died, with a mean±SD survival of 28±12 months. The first-year decedents and survivors were comparable in sex, age, medical history, and electrocardiographic findings. However, patients who died in the first year, compared with survivors, were more likely to be nursing home residents (81% vs 59%), have New York Heart Association class IV heart failure (54% vs 32%), have impaired left ventricular function by echocardiogram (53% vs 32%), and have renal insufficiency (32% vs 11%).
Very elderly persons with congestive heart failure had a guarded long-term prognosis. Nursing home residency, class IV heart failure, impaired left ventricular function, and renal insufficiency were associated with higher risk for early death and repetitive hospitalizations.
THE HIGH PREVALENCE of congestive heart failure (CHF) among the elderly presents a major public health issue with vast social economic ramifications worldwide. In the United States, it is estimated that about 10% of the population over the age of 80 years have CHF,1 and CHF is the most common cause of hospitalization among Medicare beneficiaries.2 Of all the patients hospitalized with CHF, about half are 75 years and older.3 Congestive heart failure is also the most frequent cause of repetitive hospitalization and death in the elderly.4- 6 However, there are few data on its impact on the fastest growing segment of the population, those 80 years and older residing in long-term care (LTC) facilities.
The Philadelphia (Pa) Geriatric Center houses a large number of very elderly persons with a mean age of 88 years. There are about 350 congregate housing tenants (York Houses) who require some assistance in their daily living activities and 500 nursing home residents. Congestive heart failure is the most frequent cause of hospitalization of this population. We observed that although most patients survived their initial hospitalization, they usually died within a few months after several additional hospitalizations. This retrospective study was performed to examine the clinical outcomes of these patients and to define baseline characteristics and their association with subsequent outcomes.
In 1996, a review of the cardiology consults from 1989 to 1995 identified 241 patients fulfilling the following inclusion criteria: (1) Philadelphia Geriatric Center residents residing in either congregate housing or nursing facility; (2) hospitalized with a primary or secondary diagnosis of CHF as defined with symptoms and signs of CHF and chest radiograph showing pulmonary congestion; and (3) a minimum potential follow-up of 12 months.
The initial hospitalization during which the patient was first seen by the study cardiologist (R.W.) was considered the index hospitalization, which might not have been the patient's first hospitalization for CHF. The admission date of the index hospitalization became the entry date for the study. The follow-up was calculated from the entry date to the date of analysis or death. Patient characteristics were obtained from their index hospitalization records. These included demographic data, clinical, and electrocardiographic and echocardiographic findings. Patients were classified as cognitively intact if they could give a history to the study cardiologist and fully ambulatory if they walked without an assistive device. Renal insufficiency was defined as a serum creatinine level greater than 133 µmol/L. Left ventricular ejection fraction (LVEF) was determined from echocardiograms using the area length method in technically adequate studies and visual estimate in technically limited studies; normal left ventricular function was defined as an LVEF of 0.45 or greater and impaired left ventricular function as an LVEF of less than 0.45. Significant aortic stenosis was defined as an aortic area less than 1 cm2 using the continuity equation. Subsequent outcomes, including readmissions and/or deaths, were obtained from their nursing home or outpatient records.
Deaths due to refractory CHF, acute myocardial infarction, and ventricular tachycardia or fibrillation were classified as cardiac deaths. Deaths were classified as multifactorial when at least an additional disease such as chronic obstructive pulmonary disease, pneumonia, sepsis, cerebrovascular accident, renal failure, dehydration, gastrointestinal tract bleeding, and others contributed to the final demise. A special form of multifactorial death—"do-not-hospitalize" death—was defined as death in a moribund patient with no further hospitalization order as requested by patient and/or family. Sudden death was defined when a patient with stable hemodynamics was found dead with no apparent cause. Although most of these deaths were probably cardiac related, no assumption could be made as none of these had autopsy.
Descriptive statistics were used for reporting the baseline characteristics and clinical outcomes of the study population: qualitative data were expressed as percentages and quantitative data as means±SDs. Comparison of baseline characteristics between first-year decedents and first-year survivors was made using testing for differences between proportions. Statistical significance was defined with a conservative P value of less than .01 to minimize possible type I errors that could result from using a .05 probability level when 25 variables were being compared.
Of the 241 candidates identified, 6 had incomplete data and 4 were lost to follow-up. The remaining 231 patients compri ed the study population. The patient characteristics are summarized in Table 1. All patients were white (Jewish origin), predominantly women, and had a mean age of 89±4 years. Most had a history of coronary artery disease, CHF, hypertension, either singly or in combination. Slightly more than 10% had history of chronic obstructive pulmonary disease, cerebrovascular accident, and/or diabetes mellitus. About one third of patients were cognitively intact and one tenth were fully ambulatory. All patients had either New York Heart Association (NYHA) class III or IV heart failure. About one third had atrial fibrillation. None of the findings from the admission electrocardiograms were completely normal. The echocardiograms were technically limited in 8% of patients but even in these cases a visual estimate of the overall left ventricular function could be made. One hundred six patients (46%) had normal left ventricular function with an LVEF of 0.45 or greater and 125 patients (54%) had poor left ventricular function with an LVEF of less than 0.45. But each group included patients with and without other significant cardiac abnormalities that might account for the heart failure. When grouped separately according to the most likely cause of the heart failure, 29% had diastolic heart failure with an LVEF of 0.45 or greater; 45% had systolic heart failure with an LVEF of less than 0.45, and 26% had miscellaneous cardiac abnormalities, chiefly, aortic stenosis. During the index hospitalization, the patients were managed by the study cardiologist. All patients received intravenous diuretics, digitalis, angiotensin-converting enzyme inhibitors, nitrates, and hydralazine as tolerated according to individual patient need. Three patients were intubated for refractory pulmonary edema; 2 were successfully extubated and 1 died while receiving respirator support. Approximately one fourth developed renal insufficiency.
Patients successfully discharged from the index hospitalization were followed up in the cardiology clinic every 1 to 6 months as warranted by their clinical progress. There was no separate heart failure clinic or distinct LTC heart failure unit for the posthospitalization care of these patients. During a mean follow-up of 1.1 years, 200 of the 231 patients died, for a total mortality of 87%. Only 29 patients (13%) died during the index hospitalization. An additional 117 patients (50%) died within the first 12 months of follow-up, most within 6 months. Therefore, 146 patients died in the first year with a mean survival of 4±4 months; the 1-year mortality was 63%. Eighty-five patients survived the first year after the index hospitalization. Fifty-four patients died during a mean follow-up of 2.4 years, giving an annual mortality of 27%. Thirty-one patients were still alive at the time of analysis, including 3 patients surviving more than 5 years of follow-up. Of a total of 200 deaths, 103 deaths (52%) occurred in the hospitals, including 29 deaths during the index hospitalizations and 74 deaths during the rehospitalizations; 68 were multifactorial deaths, 33 were cardiac deaths (3 with ventricular tachycardia or fibrillation, 11 with acute myocarial infarction, and 19 with refractory CHF); and 2 were sudden deaths. Ninety-seven deaths (48%) occurred in the apartments and nursing homes; 59 were sudden deaths and 38 were do-not-hospitalize deaths.
There were 416 readmissions during a total follow-up period of 248 patient-years, giving a readmission rate of 1.7 per patient-year. Three hundred sixty-six rehospitalizations (88%) were related to cardiac causes and 50 rehospitalizations (12%) were not related to cardiac causes.
In summary, of the 231 patients studied with a mean follow-up of 1.1 years, 13% died during index hospitalization, 84% either died and/or were readmitted at least once, and only 3% survived with no further event.
The 146 first-year deaths can be divided into 4 subgroups: 29 deaths during the index hospitalizations, 52 deaths during subsequent hospitalizations, 40 sudden deaths, and 25 do-not-hospitalize deaths in the apartments and nursing homes. The number of patients in each subgroup was too small for statistical analysis, but the patients were comparable in age, sex, medical history, severity of CHF, and electrocardiographic and echocardiographic findings. Despite the different manners of deaths, the last 3 subgroups had a near identical mean survival after the index hospitalization: 4.7 months for the subsequent hospital deaths, 4.8 months for the sudden deaths, and 4.2 months for the do-not-hospitalize deaths. The first-year decedents had 175 readmissions during a total follow-up of 45 patient-years, yielding a readmission rate of 3.9 per patient-year. The sudden deaths and do-not-hospitalize subgroups had the same readmission rate, 2.6 per patient-year, but the subsequent hospital deaths had a higher readmission rate of 5.4 per patient-year.
The 85 first-year survivors can also be divided into 4 subgroups: 31 patients were still alive at the time of analysis, 22 died in subsequent hospitalizations, 19 were sudden deaths, and 13 were do-not-hospitalize deaths. The number of patients in each group are again too small for statistical analysis, but the patients were also comparable in age, sex, medical history, severity of CHF, and electrocardiographic and echocardiographic findings. The 85 first-year survivors had 241 rehospitalizations during a follow-up of 202 patient-years, yielding a mean readmission rate of 1.2 per patient-year. The 54 late decedents had a mean survival of 28±12 months and a mean readmission rate of 1.3 per patient-year.
Table 2 shows the comparison between the 146 first-year decedents vs the 85 first-year survivors. The 2 groups are comparable in age, sex, medical history, and electrocardiographic findings. Despite some overlap, the first-year decedent group was composed of significantly more nursing home residents (81% vs 59%), more patients with NYHA class IV heart failure (54% vs 32%), impaired left ventricular function (53% vs 32%), and renal insufficiency (32% vs 11%) compared with the first-year survivor group.
The past decade has seen dramatic improvement in the medical management of CHF. Several large clinical trials have demonstrated the efficacy of angiotensin-converting enzyme inhibitors in reducing mortality and hospitalizations for patients with CHF.7- 9 In a single-center study, total 1-year mortality decreased from 33% before 1989 to 16% after 1990, and sudden death decreased from 20% to 8% for patients with advanced CHF.10 However, these data on improved survival are obtained in younger patients participating in clinical trials and may not be applicable to the frail elderly. Our study represents the first attempt to define the clinical outcome of the institutionalized old-old hospitalized with CHF. In our study population, the immediate prognosis appears to be favorable, the short-term prognosis is guarded with repetitive hospitalizations and the long-term prognosis is dismal. A recent study11 on Medicare beneficiaries 65 years and older had an 8% hospital mortality for CHF and 53% of the study sample died or had a readmission during a 6-month follow-up. Our study population, being older and frailer, had a hospital mortality of 13%, and 84% of the patients either died or had a readmission during a mean follow-up of 1.2 years. In a study on LTC residents hospitalized for all causes, the hospital mortality was 12% and 1-year mortality was 40%.12 Our patients had a comparable hospital mortality of 13% but a higher 1-year mortality of 63%.
A more detailed examination of the study results reveals that surviving the first year after index hospitalization predicts a more favorable prognosis. The 85 first-year survivors had an annual mortality of 26.6%, not much different from the 21.5% annual mortality of nursing home residents reported in 1985.13 Our findings are similar to results of an earlier study14 reporting a dramatic early impact of CHF on survival in elderly men. These 94 veterans, with a mean age of 83 years, had a worse survival rate than the control subjects for the first 5 years after diagnosis of CHF, chiefly due to a high first-year mortality (28%) of the CHF group. Thus, the old-old with CHF do not appear to be a homogeneous population with a uniform outcome, and first-year survival demarcates a lower-risk group from a higher-risk group. This distinction has major clinical implications. It has always been a concern that the frail elderly with multiple hospitalizations during end of life consumes a large amount of health care costs.15,16 Apart from financial concerns, hospitalization for an elderly institutionalized patient usually entails transfer to an unfamiliar acute care setting with frequent complications such as increased confusion, falls, and nosocomial infections. In general, an aggressive hospitalization policy for CHF in this population should be warranted if it would prolong life or improve its quality. Our results do not suggest survival benefits. A most remarkable finding is the near identical mean survival of the first-year decedents after their index hospitalization, despite the fact that some died suddenly, some died in subsequent hospitalizations, and some died in LTC facilities after refusing further hospitalizations. It seems that the first-year decedents represent a critically ill, debilitated group that was only temporarily helped by intensive medical regime in the hospital. Further hospitalizations only changed the place of death but did not prolong life. The issue whether hospital death was more desirable than do-not-hospitalize death for this population has never been studied. A recent study on hospital deaths showed that family members perceived that pain or other symptoms were common and troubling to most elderly and seriously ill patients dying in the acute care hospitals.17 They believed that patients preferred comfort measures, but life-sustaining measures were often used. As none of the charts reviewed described the quality of life of the patients, our study did not yield any information on this important issue. It was unclear whether the do-not-hospitalize deaths were truly the peaceful and dignified deaths that the patients and their families had asked for. Also, a recent study18 has shown that a multidisciplinary intervention is effective in preventing the readmission of community-dwelling elderly patients with CHF. Our study population received standard care in their LTC facilities although they did attend regular cardiology follow-up. It remains uncertain whether a similar intensive multidisciplinary approach implemented in the LTC setting with close supervision of dietary sodium intake; emotional and physical stress, particularly mental agitation; daily body weight; blood pressure control; renal function monitoring; and adjustment of diuretics would reduce rehospitalizations and improve long-term prognosis. Further studies are needed.
Another finding of our study is that nearly half of all deaths in patients successfully discharged after the index hospitalization occurred in the LTC facilities. These deaths may be "hidden" from hospital physicians who might not monitor them in the LTC facilities. Thus, the hospital physicians, including house staff, might have a falsely optimistic outlook for the study population. The recent inclusion of geriatrics rotations to the medical resident training programs should enable these younger physicians to have a better understanding of prognosis for frail elderly patients successfully discharged to LTC facilities after hospitalization for CHF.
There are several limitations of our study. First, the study population consisted entirely of Jewish elderly. Most of the family members were highly educated professionals closely involved in patient care decisions. Aggressive treatment approaches were often favored until death. Only 38 of the 231 patients did choose a do-not-hospitalize status. All our patients had an echocardiogram during index hospitalization to evaluate left ventricular function as recommended in the Clinical Practice Guideline for Heart Failure by the American Medical Directors Association.19 A recent retrospective study20 of LTC residents with CHF showed that determination of whether CHF was due to left ventricular systolic or diastolic function was made in fewer than 1% of the patients. It is not clear whether our findings can be generalized to other frail elderly populations with different racial and socioeconomic backgrounds. Second, the index hospitalization was an arbitrary point set for the onset of the study as it represented the availability of the study cardiologist and the systematic echocardiographic examination. In 50% of the cases, the index hospitalization was not the first hospitalization for CHF. Thus, patients entered the study at different stages of CHF. The data, therefore, resembled a cross-sectional sampling of the institutionalized elderly more than a longitudinal study. Furthermore, because of the sedentary lifestyle of the institutionalized elderly, CHF of lesser severity may not have been noted in the study population. All the study patients had either NYHA class III or IV heart failure. The results would not apply to more active community-dwelling elderly hospitalized with lesser degrees of CHF. Finally, being a retrospective study, there was no control of the availability and quality of data collected, eg, none of the index hospitalization charts documented the Mini-Mental State Examination score of the patient and none of the hospital or LTC charts described the manner of death as to whether it was painful or peaceful.
Despite these limitations, our study represents the first attempt to examine the clinical outcome of the institutionalized elderly hospitalized with CHF. In summary, the study population had a low hospital mortality but a guarded long-term prognosis. There were 4 clinical parameters that were predictors of higher first-year mortality: residency in a nursing home, NYHA class IV heart failure, impaired left ventricular function by echocardiogram, and renal insufficiency (serum creatinine level >133 µmol/L). Hospital physicians should use individualized approaches to manage every patient with CHF based on a comprehensive geriatric and cardiac assessment. Prognosis should be discussed with the patient and family and the LTC physician. In patients at high risk for early death, successful discharges may only be temporary reprieves. Patients with better prognosis may benefit from aggressive treatments. Further prospective studies of multiethnic populations with better definition of frailty, comorbidity, and quality of life are needed to provide more information to better manage CHF in the frail institutionalized elderly.
Accepted for publication April 13, 1998.
Reprints: Rebecca Wang, MD, Philadelphia Geriatric Center, 5301 Old York Rd, Philadelphia, PA 19141.