Early Deaths in Patients With Heart Failure Discharged From the Emergency DepartmentCLINICAL PERSPECTIVE
A Population-Based Analysis
Background— Although approximately one third of patients with heart failure (HF) visiting the emergency department (ED) are discharged home, little is known about their care and outcomes.
Methods and Results— We examined the acute care and early outcomes of patients with HF who visited an ED and were discharged without hospital admission in Ontario, Canada, from April 2004 to March 2007. Among 50 816 patients (age, 76.4±11.6 years; 49.4% men) visiting an ED for HF, 16 094 (31.7%) were discharged without hospital admission. A total of 4.0% died within 30 days from admission, and 1.3% died within 7 days of discharge from the ED. Although multiple (≥2) previous HF admissions (odds ratio [OR], 1.64; 95% CI, 1.14 to 2.31), valvular heart disease (OR, 1.37; 95% CI, 1.00 to 1.84), peripheral vascular disease (OR, 1.41; 95% CI, 1.00 to 1.93), and respiratory disease (OR, 1.33; 95% CI, 1.08 to 1.63) increased the risk of 30-day death among those discharged from the ED, presence of these conditions did not increase the likelihood of admission. Patients were more likely to be admitted if they were older (OR, 1.08; 95% CI, 1.06 to 1.10 per decade), arrived by ambulance (OR, 2.02; 95% CI, 1.93 to 2.12), had a higher triage acuity score (OR, 4.12; 95% CI, 3.84 to 4.42), or received resuscitation in the ED (OR, 2.85; 95% CI, 2.68 to 3.04). In those with comparable predicted risks of death, subsequent 90-day mortality rates were higher among discharged than admitted patients (11.9% versus 9.5%; log-rank P=0.016).
Conclusions— Patients with HF who are discharged from the ED have substantial risks of early death, which, in some cases, may exceed that of hospitalized patients.
Received June 8, 2009; accepted January 13, 2010.
Heart failure (HF) is a leading reason for hospitalization, physician visits, and healthcare costs,1 with a large proportion of resource utilization attributable to emergent care and hospitalizations.2,3 The global implications of HF are underscored by the high lifetime risk of the condition in North America4 and the rising HF incidence around the world attributable, in part, to increasing economic globalization.5,6 Approximately 2.5% of the total healthcare budget in Europe is attributable to HF, and hospital-based care continues to be a major component, comprising ≈70% of this amount.5,7 The emergency department (ED) is the point of first contact for patients with acute HF, and in North America, there are >1 million visits to the ED for HF annually, representing a 20% increase during the past decade.8
Clinical Perspective on p 228
Acute decompensated HF has been increasingly recognized as a distinct entity from that of chronic stable HF, yet relatively little is known about the presentation and outcomes of this condition.9 Patients with acute decompensated HF funnel toward clinical presentation in the ED with the possible need for hospital admission.9 Although many previous studies have examined disease epidemiology and outcomes after hospitalization,10–12 there has been little study of the totality of patients presenting to the ED with an acute decompensation of HF. There is escalating interest in the emergent care of patients with HF because of the increasingly important role of care provided in this setting, stemming from high rates of utilization, overcrowding, and growing pressure to discharge patients from the ED to home. Thus, much of the current literature portrays only a partial view of HF care because the large group of patients who are assessed in the ED and discharged home have been largely excluded from previous studies.
In this study, we describe the characteristics of a population sample of patients with HF in the ED setting and compare the lethal outcomes and predictors of mortality among those who were admitted to the hospital or discharged from the ED. We hypothesized that even among patients discharged from the ED, and thus considered “safe” for discharge, a significant number of adverse outcomes would occur within days of discharge from the ED.
ED information was obtained using the National Ambulatory Care Reporting System database, which contains information on all ED visits. Hospitalizations were evaluated using the Canadian Institute for Health Information Discharge Abstract Database (CIHI-DAD). Both the National Ambulatory Care Reporting System and CIHI-DAD used the International Classification of Diseases, 10th Revision, Canadian Enhancement (ICD-10-CA) or 9th Revision, Clinical Modification (ICD-9-CM) coding systems. Procedures were determined using both the CIHI-DAD and Ontario Health Insurance Plan physician billing database. We used the 2006 Canadian Census to determine the socioeconomic status quintile based on forward sortation address of the postal code. All data were linked using the patients’ unique, encrypted health card numbers.
We examined patients aged ≥20 years who visited an ED in Ontario from April 1, 2004, to March 30, 2007, with a primary diagnosed condition of HF. Both those patients who were discharged from the ED and those who were admitted to the hospital from the ED were evaluated. In those with multiple ED visits during the study period, the first episode was selected as the index HF visit.
Determination of Cardiac and Noncardiac Conditions
We determined the presence of cardiac conditions (eg, HF and myocardial infarction) and noncardiac comorbidities by examining all diagnoses in the CIHI-DAD and the same-day surgery database in the 3 years before the index ED visit (see the online-only Data Supplement Appendix 1 for ICD-9-CM and ICD-10-CA codes). Previous procedures (eg, coronary artery bypass graft surgery, percutaneous coronary intervention, and implantable cardioverter defibrillator) were determined using the Canadian Classification of Interventions, Canadian Classification of Procedures, and Ontario Health Insurance Plan codes (supplemental Appendix 1). Noncardiac comorbid conditions were classified using the Charlson comorbidity classification system.13
Presentation Features and Care
We examined the mode (eg, ambulance versus ambulatory) and time of presentation (eg, daytime, 0801 to 1600 hours; evening, 1601 to 2400 hours; and night, 0001 to 0800 hours) and the Canadian Triage Acuity Scale, which signified the perceived acuity of the ED patient. Patients with Canadian Triage Acuity Scale scores 1 (resuscitation) or 2 (emergent) were considered high acuity; score 3 (urgent), medium acuity; and scores 4 (less urgent) and 5 (nonurgent) were low acuity.14 The impact of resuscitation provided in the ED was assessed using Ontario Health Insurance Plan fee codes for critical care delivery, cardiorespiratory resuscitation, or endotracheal intubation for cardiopulmonary arrest or prearrest situations. The length of stay in the ED also was examined and calculated as the time from presentation to the time of disposition from the ED (eg, time of discharge or admission).
The main outcome of this study was mortality, which was determined using the Registered Persons Database for all deaths and the CIHI-DAD for in-hospital deaths. Time to death was determined starting from the time of discharge from the ED to either the patient’s residence (nonadmitted) or to the hospital (admitted).
Clinical and demographic characteristics were compared between discharged and admitted patients. Continuous variables are reported as mean±SD, unless otherwise stated. Student t test and the χ2 test were used to compare continuous and categorical variables, respectively. We examined factors associated with the decision to admit or discharge patients from the ED, using multiple logistic regression analysis. Because our intent was to examine the joint effects of these predictors, we adjusted for all covariates in the multivariable model.
We determined the predicted probabilities of death in discharged and admitted patients and determined the degree of overlap of predicted death probabilities using multiple logistic regression analysis. We constructed predictive models for death within 7 and 30 days of discharge from the ED using multiple logistic regression. These models included demographic variables (age, sex, and socioeconomic status quintile) and cardiovascular history of previous HF, myocardial infarction, peripheral vascular disease, cerebrovascular disease, coronary artery bypass graft surgery, percutaneous coronary intervention, implantable defibrillator, or permanent pacemaker. We also included presenting features (eg, paramedic versus ambulatory, Canadian Triage Acuity Scale, and registration time), ED features (eg, length of stay and resuscitation status), and all noncardiac comorbidities in the mortality prediction models. The cardiac and noncardiac conditions and procedures included in the predictive models for death are shown in supplemental Appendix 1. One-way interactions of all covariates with age and sex also were included. Discrimination of the logistic regression models were assessed using the C statistic, and we examined model fit using the Hosmer-Lemeshow goodness-of-fit statistic.
We compared mortality rates in patients with predicted probabilities of death within the same range, defined empirically by the 75th percentile of risk for discharged patients and the 25th percentile for admitted patients. Thus, we compared the highest risk quartile of discharged patients and the lowest risk quartile of admitted patients, an overlapping cohort in whom there may have been equipoise in the decision to admit to the hospital or discharge from the ED based on predicted risks of death. These ranges were considered to represent plausible thresholds that defined overlapping predicted probabilities of death. We used Kaplan-Meier analysis to determine cumulative mortality among discharged and admitted patients with overlapping predicted probabilities of death and compared the survival curves using the log-rank statistic. Sensitivity analyses were conducted by varying the thresholds of predicted risk probabilities and thus allowing for inclusion of larger numbers in the discharged and admitted cohorts. We used Cox regression analysis with time-varying covariates to estimate the mortality effect of ED visits or hospitalizations for HF occurring over time, and we censored our time-to-event analyses at 90 days or on the last follow-up date of March 31, 2008 (whichever occurred earlier). A 2-sided P<0.05 was considered statistically significant. All analyses were performed using SAS version 9.1.3.
A total of 78 642 ED visits for HF were examined. Of the 50 816 unique first patient visits to the ED (mean age, 76.4±11.6 years; 49.4% men), 16 094 (31.7%) were discharged home without hospital admission and 34 722 (68.3%) were admitted to the hospital directly from the ED. The status of all study patients at 30 days was determined, with a total of 48 570 person-months of follow-up examined. Baseline characteristics of discharged and admitted study patients are shown in Table 1, and features at ED presentation are shown in Table 2. The majority of patients discharged from the ED (99.2%) were discharged by the physician, and only a few were discharged against medical advice (0.7%) or left without being seen (0.1%). A small proportion of those discharged from the ED were referred for palliative care (0.3% within 7 days and 0.8% of 7-day survivors within 30 days).
Multivariable Predictors of Admission Versus Discharge
The significant multivariable predictors of admission to the hospital are shown in Table 3 (all model covariates, including both significant and nonsignificant, are shown in supplemental Appendix 2). Although arrival by ambulance, evening or nighttime presentation, acuity, and receipt of resuscitation were all associated with hospital admission, a history of myocardial infarction or HF and most noncardiac comorbidities had little influence on disposition from the ED.
The 7-day mortality rate of patients discharged from the ED was 1.3% (n=203), with 108 (53.2%) deaths occurring out of hospital and 95 (46.8%) occurring in hospital. The 30-day mortality rate among patients discharged from the ED was 4.0% (n=649), with 275 (42.4%) deaths occurring out of hospital and 374 (57.6%) occurring in hospital. Among patients initially admitted to the hospital directly from the index ED visit, the mortality rate was 5.7% at 7 days and 12.3% at 30 days.
Predicted Death in Discharged and Admitted Patients With HF
The predicted probability of death was 1.2±1.9% at 7 days and 4.0±4.3% at 30 days among patients discharged from the ED. Predicted probabilities of death in admitted patients were 5.7±4.5% at 7 days and 12.3±7.8% at 30 days. All covariates were included in the multivariable models for mortality, and there was no lack of model fit (Hosmer-Lemeshow, P>0.30).
Observed Mortality Rates
Histograms of predicted probabilities of 7-day and 30-day death are shown in Figures 1 and 2⇓, respectively. There was a significant overlap in the predicted probabilities of death, with a large proportion of discharged patients having predicted probabilities that were similar to that of admitted patients. Mortality rates in admitted and discharged patients with overlapping predicted probabilities of 7-day and 30-day death are shown in Figures 3 and 4⇓. Mortality comparisons were performed for patients discharged from the ED or admitted to the hospital with overlapping predicted probabilities of 7-day death between 1.3% and 2.5% and between 4.8% and 6.5% for 30-day death. A total of 7549 (14.9%) patients of whom 1696 were discharged and 5853 were admitted had predicted risks of 7-day death that overlapped in the range of 1.3% to 2.5%. A total of 4893 (9.6%) patients of whom 1189 were discharged and 3704 were admitted had predicted risks of 30-day death that overlapped in the range of 4.8% to 6.5%.
For the 7-day cohort, mortality was significantly higher in patients who were discharged from the ED, with 90-day mortality rates of 12.4% for discharged and 9.4% for admitted patients (log-rank P<0.001) (Figure 3). Similar findings were observed for the 30-day cohort (Figure 4), with mortality rates of 11.9% for discharged and 9.5% for admitted patients at 90 days (log-rank P=0.016). The rates of repeated ED visits or hospitalizations for HF were significantly higher in those who were discharged directly from the ED. Event rates at 30 days were 14.0% (discharged from the ED) and 7.5% (discharged from the hospital) among patients who survived the first 7 days after HF onset (P<0.001). Event rates at 90 days were 20.3% (ED) and 16.6% (hospitalized) among patients who survived the first 30 days after HF onset (P<0.001). In a time-varying Cox model adjusted for age, sex, and Charlson comorbidity score, the hazard ratio for death was 3.65 (95% CI, 3.46 to 3.84) for each additional ED visit or hospitalization that occurred within the first 90 days after discharge from the ED (P<0.001).
In sensitivity analyses, we examined greater numbers of patients by including discharged patients with ≥65th percentile (n=3944) and admitted patients with ≤35th percentile (n=11 320) of risk, with predicted 7-day mortality rates between 0.79% and 3.12%. Mortality rates remained higher in patients discharged from the ED than in those admitted to the hospital at 90 days (12.2% versus 9.1%, P<0.001). Similarly, expanding to discharged patients with ≥65th percentile (n=3605) and admitted patients with ≤35th percentile (n=9746) of risk, with predicted 30-day mortality rates between 3.43% to 7.88%, mortality rates at 90 days remained higher in patients who were discharged from the ED than in those admitted to the hospital (12.2% versus 9.7%, P<0.001).
Predictors of Death
Mortality rates according to patient characteristics are shown in Table 4. Mortality increased with the number of previous HF admissions, mode and time of presentation, and receipt of resuscitation in the ED. Significant multivariable predictors of 30-day mortality are presented in Table 5, which are age, male sex, arrival by ambulance, ≥2 previous HF admissions, valvular heart disease, peripheral vascular disease, respiratory disease, and longer ED length of stay. Higher initial triage acuity was associated with lower 30-day death. A subset of the 30-day covariates (Table 5) also predicted increased 7-day mortality, including older age (odds ratio [OR], 1.55 per 10 years; 95% CI, 1.31 to 1.84; P<0.001), male sex (OR, 1.63; 95% CI, 1.20 to 2.20; P=0.002), arrival by ambulance (OR, 5.07; 95% CI, 3.58 to 7.24; P<0.001), ≥2 previous HF admissions (OR, 2.26; 95% CI, 1.25 to 3.90; P=0.005), dementia (OR, 2.27; 95% CI, 1.44 to 3.46; P<0.001), metastatic cancer (OR, 3.27; 95% CI, 1.22 to 7.25; P=0.008), and longer ED length of stay (OR, 1.07 per 5 hours; 95% CI, 1.00 to 1.13; P=0.022). The C statistics for the 7-day and 30-day models were 0.806 and 0.755, respectively.
Previous studies pertaining to the hospital-based care of patients with HF have largely included those who were hospitalized. Relatively little is known about the outcomes of those who are discharged home from the ED. In this study, we examined population-based data from a large cohort of patients presenting to the ED with symptoms of acute decompensated HF, and we found that 4.0% of nonadmitted patients with HF died within 30 days of discharge from the ED. There was a wide variation in the predicted risks of death among discharged patients with HF, suggesting that clinicians currently lack tools to determine the anticipated risk of adverse outcomes in the ED. Indeed, a substantial proportion of discharged patients had risks of death that were comparable with their counterparts who were admitted to the hospital. They also had high rates of repeated ED visits and hospitalizations for HF occurring within 90 days, and the postdischarge events conferred increased mortality risk as we had observed in a previous evaluation.15
Previous studies have examined almost exclusively patients with HF who have been admitted to the hospital.10,12,16,17 Therefore, most of the previous studies portray a partial view of HF care because patients who are assessed in the ED and discharged home often have been excluded. There have been few previous studies of patients with HF discharged directly from the ED. The Acute Decompensated Heart Failure National Registry Emergency Medicine study reported characteristics of patients presenting to the ED with acute HF.18 Others have reported that approximately one third of patients with HF are discharged from the ED, with 30-day mortality rates that were similar to our observations.19 Our study adds to the current HF literature that many patients who are discharged from the ED exhibited high-risk features, with a substantial early mortality rate. Indeed, higher rates of death were observed in patients discharged from the ED than in patients admitted with comparable short-term predicted probabilities of death.
Several factors influenced the decision to admit patients with HF to the hospital. Some of these factors were related to disease severity, but variations in care were observed. Although the majority of patients who received resuscitation, arrived by ambulance, and had higher triage acuity scores were admitted, a significant proportion of these patients were discharged from the ED. Longer durations of ED stay conferred higher risk in the discharged group, suggesting that the empirical strategy of observation of patients with HF in the ED to ensure that it is safe to discharge the patient does not necessarily discriminate well those who are at low risk of subsequent death. In patients discharged from the ED, the mode of arrival and longer ED stay were multivariate predictors of mortality; however, higher initial triage acuity was associated with improved survival. The seemingly paradoxical effects of higher triage acuity could be related to the intensity of initial investigations and postdischarge follow-up care, which may have been provided preferentially to those who were perceived to be sicker at the time of presentation. Previous HF admissions directly reflect HF severity,15 whereas other conditions, such as valvular heart disease, peripheral vascular disease, and respiratory disease, may contribute to mortality through multiple mechanisms. However, these conditions did not increase the likelihood of admission to the hospital, illustrating the need for further study of patients with HF in the emergency setting. Finally, our findings could not be explained by an intention for palliation because only a small percentage of patients who were discharged from the ED received palliative care.
Because the crude mortality rate was lower among patients discharged from the ED than those who were admitted, some may argue that current empirical decision making is effective. However, the mortality rate of patients with HF is high compared with the parallel context of patients with chest pain who visit the ED and are discharged, in whom the 30-day mortality rate is <1%.20,21 Despite the substantially lower rate of death in ED patients with chest pain, great efforts are being applied to further reduce this risk with the use of clinical prediction rules,22 biomarkers,23 and imaging.21,24 In contrast, for patients with HF who have higher rates of death, there is little evidence to guide decision making in the ED.
From the public health perspective, the implications are large. For every 25 discharged patients, 1 died within 30 days, and for every 80 discharged patients, >1 died within 7 days; thus, we estimate that thousands of early deaths occur after discharge from the ED in North America annually. From a clinical standpoint, our study suggests that although management for the ED is influenced somewhat by the acuity of presentation, the decision to admit or discharge is based largely on clinical judgment. There is a need for validated tools to assist physicians to better determine the mortality risk of patients with HF who present to the ED and to guide decision making regarding the safety of discharge from the ED. Potentially, predictive tools might be used to identify high-risk patients who could benefit from early interventions, including expeditious referral to an HF clinic or specialist care, home visits by nurses, and cardiac investigations.
Our study had some notable limitations. We used large administrative databases that did not include clinical variables such as blood pressure or echocardiographic information. Despite this limitation, our broad, population-based analysis identified a gap in HF care for which future clinical research is needed to improve decision making. We did not examine the uptake or impact of brain natriuretic peptide in the ED setting; however, emerging data suggest that brain natriuretic peptide is of untested value for outcomes assessment in the ED at the broad population level25,26 or for guiding therapeutic decisions.27 The contribution of acute HF drug therapy was beyond the scope of this study, but conversely, the relative benefits of emergent use of pharmacological agents for acute decompensated HF have not been fully delineated. Finally, although we examined the contribution of multiple patient characteristics, presentation features, and facets of the treatments received, as yet unmeasured pre- and post-ED factors also may have influenced the results.
In conclusion, patients with HF who are discharged from the ED have an appreciable early mortality risk beginning as soon as 7 days postdischarge. Indeed, some patients who were discharged from the ED had predicted and observed mortality risks that were comparable with or exceeded those of hospitalized patients. These early findings suggest that there is a need for further clinical evidence to guide risk stratification in the ED and to assist in decision making with regard to the safety of direct discharge of patients with HF from the ED.
Sources of Funding
The Institute for Clinical Evaluative Sciences is supported in part by a grant from the Ontario Ministry of Health and Long-Term Care. The opinions, results, and conclusions are those of the authors, and no endorsement by the Ministry of Health and Long-Term Care or by the Institute for Clinical Evaluative Sciences is intended or should be inferred. This research was supported by an operating grant from the Canadian Institutes of Health Research (CIHR grant MOP 86718). Dr Lee is a clinician-scientist of the Canadian Institutes of Health Research. Dr Austin is career investigator of the Heart and Stroke Foundation of Ontario. Dr Tu is a career investigator of the Heart and Stroke Foundation of Ontario and a Canada Research Chair in health services research.
The online-only Data Supplement is available at http://circheartfailure.ahajournals.org/cgi/content/full/CIRCHEARTFAILURE.109.885285.
- ↵Rosamond W, Flegal K, Friday G, Furie K, Go A, Greenlund K, Haase N, Ho M, Howard V, Kissela B, Kittner S, Lloyd-Jones D, McDermott M, Meigs J, Moy C, Nichol G, O'Donnell CJ, Roger V, Rumsfeld J, Sorlie P, Steinberger J, Thom T, Wasserthiel-Smoller S, Hong Y. Heart disease and stroke statistics—2007 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2007; 115: e69–e171.
- ↵Liu P, Arnold M, Belenkie I, Howlett J, Huckell V, Ignazewski A, LeBlanc MH, McKelvie R, Niznick J, Parker JD, Rao V, Ross H, Roy D, Smith S, Sussex B, Teo K, Tsuyuki R, White M, Beanlands D, Bernstein V, Davies R, Issac D, Johnstone D, Lee H, Moe G, Newton G, Pflugfelder P, Roth S, Rouleau J, Yusuf S. The 2001 Canadian Cardiovascular Society consensus guideline update for the management and prevention of heart failure. Can J Cardiol. 2001; 17 (suppl E): 5E–25E.
- ↵Lloyd-Jones DM, Larson MG, Leip EP, Beiser A, D'Agostino RB, Kannel WB, Murabito JM, Vasan RS, Benjamin EJ, Levy D. Lifetime risk for developing congestive heart failure: the Framingham Heart Study. Circulation. 2002; 106: 3068–3072.
- ↵McMurray JJ, Stewart S. Epidemiology, aetiology, and prognosis of heart failure. Heart. 2000; 83: 596–602.
- ↵Sanderson JE, Tse TF. Heart failure: a global disease requiring a global response. Heart. 2003; 89: 585–586.
- ↵Remme WJ, McMurray JJ, Rauch B, Zannad F, Keukelaar K, Cohen-Solal A, Lopez-Sendon J, Hobbs FD, Grobbee DE, Boccanelli A, Cline C, Macarie C, Dietz R, Ruzyllo W. Public awareness of heart failure in Europe: first results from SHAPE. Eur Heart J. 2005; 26: 2413–2421.
- ↵Adams KF Jr, Fonarow GC, Emerman CL, LeJemtel TH, Costanzo MR, Abraham WT, Berkowitz RL, Galvao M, Horton DP. Characteristics and outcomes of patients hospitalized for heart failure in the United States: rationale, design, and preliminary observations from the first 100,000 cases in the Acute Decompensated Heart Failure National Registry (ADHERE). Am Heart J. 2005; 149: 209–216.
- ↵Diercks DB, Fonarow GC, Kirk JD, Jois-Bilowich P, Hollander JE, Weber JE, Wynne J, Mills RM, Yancy C, Peacock WF. Illicit stimulant use in a United States heart failure population presenting to the emergency department (from the Acute Decompensated Heart Failure National Registry Emergency Module). Am J Cardiol. 2008; 102: 1216–1219.
- ↵Gibler WB, Cannon CP, Blomkalns AL, Char DM, Drew BJ, Hollander JE, Jaffe AS, Jesse RL, Newby LK, Ohman EM, Peterson ED, Pollack CV. Practical implementation of the guidelines for unstable angina/non-ST-segment elevation myocardial infarction in the emergency department: a scientific statement from the American Heart Association Council on Clinical Cardiology (Subcommittee on Acute Cardiac Care), Council on Cardiovascular Nursing, and Quality of Care and Outcomes Research Interdisciplinary Working Group, in collaboration with the Society of Chest Pain Centers. Circulation. 2005; 111: 2699–2710.
- ↵Udelson JE, Beshansky JR, Ballin DS, Feldman JA, Griffith JL, Handler J, Heller GV, Hendel RC, Pope JH, Ruthazer R, Spiegler EJ, Woolard RH, Selker HP. Myocardial perfusion imaging for evaluation and triage of patients with suspected acute cardiac ischemia: a randomized controlled trial. J Am Med Assoc. 2002; 288: 2693–2700.
- ↵Schneider HG, Lam L, Lokuge A, Krum H, Naughton MT, De Villiers SP, Bystrzycki A, Eccleston D, Federman J, Flannery G, Cameron P. B-type natriuretic peptide testing, clinical outcomes, and health services use in emergency department patients with dyspnea: a randomized trial. Ann Intern Med. 2009; 150: 365–371.
- ↵Pfisterer M, Buser P, Rickli H, Gutmann M, Erne P, Rickenbacher P, Vuillomenet A, Jeker U, Dubach P, Beer H, Yoon SI, Suter T, Osterhues HH, Schieber MM, Hilti P, Schindler R, Brunner-La Rocca HP. BNP-guided vs symptom-guided heart failure therapy: the Trial of Intensified vs Standard Medical Therapy in Elderly Patients With Congestive Heart Failure (TIME-CHF) randomized trial. J Am Med Assoc. 2009; 301: 383–392.
Patients with acute heart failure (HF) often present to the emergency department (ED) for care, and approximately one third are discharged home without hospital admission. In this study of 50 816 patients with HF, we found that death occurred in 1.3% within 7 days and 4.0% at 30 days after discharge from the ED. There was overlap in predicted probabilities of death among those who were discharged from the ED and those admitted to the hospital, suggesting that equipoise exists in the decision to admit or discharge the patients with HF from the ED. Among those with similar predicted probabilities of death, observed 90-day mortality was significantly higher among patients who were discharged from the ED compared with those who were admitted to the hospital (P<0.001 and P=0.016 for 7-day and 30-day predicted probability cohorts, respectively). Repeat ED visits or hospitalization for HF within 90 days occurred in 20.3% of those initially discharged and 16.6% of those initially hospitalized (P<0.001). Recurrent ED visits or hospitalizations after initial discharge conferred a 3.6-fold increase in risk of death. Older age, male sex, arrival by ambulance, ≥2 previous HF hospitalizations, valvular heart disease, peripheral vascular disease, respiratory disease, and longer length of ED stay were predictors of death after discharge from the ED. These early findings suggest that there is a need for further clinical evidence to guide risk stratification in the ED and to assist in decision making regarding the safety of direct discharge of patients with HF from the ED.