|Home | About | Journals | Submit | Contact Us | Français|
Heart failure (HF) is a common emergency department (ED) presentation and a leading reason for hospitalization. Canadian practice patterns for the management of acute HF have not been well described.
To describe current treatment patterns of patients presenting to the ED with acute HF and investigate whether these treatments influenced outcomes.
A health record review was performed in a 30% random sample of all patients who presented to six EDs in the Capital Health Region (Edmonton, Alberta) with a most responsible diagnosis of acute HF from April 2002, to March 2003.
A total of 448 patients (45% women) with a mean (± SD) age of 75.3±11.2 years were included. Comorbidities included hypertension (55%), coronary artery disease (39%) and previous myocardial infarction (38%). In the first 72 h, patients were most commonly treated with intravenous furosemide (48%), angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (45%), oral furosemide (42%) and salbutamol (38%). Fifty-four per cent of patients were admitted to the hospital, and 20% died or were readmitted within 30 days. Multivariate logistic regression analysis revealed age, history of HF, history of angioplasty and oxygen administration in the ED as independent predictors of death or readmission at 30 days. No medications were associated with decreased readmission or death.
The current treatment patterns for acute HF are mostly symptomatic. Proven efficacious HF therapies remain underused. Future research should focus on the integration of disease management, identifying predictors of admission and readmission, and treatments to reduce rehospitalization.
L’insuffisance cardiaque (IC) est une présentation courante à l’urgence et l’une des principales raisons d’hospitalisation. Les modes de pratique canadiens pour la prise en charge de l’IC aiguë ne sont pas bien décrits.
Décrire les modes de traitement courants des patients qui consultent à l’urgence en raison d’une IC aiguë et explorer si ces traitements influent sur les issues.
Les auteurs ont procédé à une analyse du dossier dans un échantillon aléatoire de 30 % de tous les patients qui s’étaient présentés à six départements d’urgence de la région sanitaire de la capitale (à Edmonton, en Alberta) en raison d’un diagnostic le plus responsable d’IC aiguë entre avril 2002 et mars 2003.
Au total, 448 patients (45 % de femmes) d’un âge moyen (±ÉT) de 75,3±11,2 ans faisaient partie de l’étude. Les comorbidités incluaient l’hypertension (55 %), les coronaropathies (39 %) et un infarctus du myocarde antérieur (38 %). Pendant les 72 premières heures, les patients étaient surtout traités au moyen de furosémide par voie intraveineuse (48 %), d’inhibiteurs ou antagonistes de l’enzyme de conversion de l’angiotensine (45 %), de furosémide par voie orale (42 %) et de salbutamol (38 %). Cinquante-quatre pour cent des patients ont été hospitalisés, et 20 % sont décédés ou ont été réhospitalisés dans les 30 jours. L’analyse de régression logistique multivariée a révélé que l’âge, les antécédents d’IC, d’angioplastie et d’administration d’oxygène à l’urgence étaient des prédicteurs indépendants d’un décès ou d’une réhospitalisation dans les 30 jours. Aucun médicament ne s’associait à une diminution des hospitalisations ou des décès.
Les modes de traitement actuels de l’IC aiguë sont surtout symptomatiques. Les thérapies efficaces démontrées de l’IC demeurent sous-utilisées. Les futures recherches devraient porter sur l’intégration de la prise en charge des maladies, le dépistage des prédicteurs d’hospitalisation et de réhospitalisation et les traitements en vue de réduire les hospitalisations.
Heart failure (HF) is a common and serious condition that affects 200,000 to 300,000 people in Canada (1). HF is a burden on the Canadian health care system; in 2000, it was the second most prominent reason for hospitalization in Canada, with an average length of stay (LOS) of 12.9 days, and accounted for 1.4 million days in the hospital (2). Moreover, HF often coexists with other chronic diseases such as diabetes, chronic obstructive pulmonary disease (COPD) and coronary artery disease, which further contribute to hospitalization and prolonged lengths of stay. Importantly, patients who are admitted to the hospital for HF have readmission rates as high as 30% to 60% within three to six months after the initial discharge (3–5).
There are accepted therapies (eg, diuretics, inotropes, nitrates) (6–9) for acute HF; however, randomized controlled trial evidence for their efficacy is often lacking. Moreover, until recently, there were no clinical guidelines for the management of acute HF. The 2005 publication of the European Society of Cardiology acute HF guidelines focused primarily on the treatment of acute HF symptoms (10). In 2001, the Acute Decompensated Heart Failure National Registry (ADHERE) (11) was initiated in the United States to monitor management of hospitalized HF patients. The ADHERE registry has reviewed 100,000 HF hospitalizations and provided insight into the medical management of this condition. Knowledge of Canadian practice patterns and their relationship to patient outcomes is largely unknown, especially for acute HF. Furthermore, there are no data on patients seen in the emergency department (ED) and discharged directly home (such patients were excluded from ADHERE). Therefore, the objective of the present study was to describe current treatment patterns of patients with acute HF presenting to the ED.
The protocol for the present study was approved by the health research ethics board at the University of Alberta (Edmonton, Alberta). No patient contact occurred after the ED encounter.
A retrospective chart review was performed to examine the treatment of acute HF, defined as treatments received within the first 72 h of presentation to an ED.
The study was conducted at five acute care hospitals and one community health centre in the former Capital Health Region of Alberta, which includes Edmonton and St Albert. All six sites were urban teaching hospitals staffed by full-time emergency physicians who managed 341,951 patients during fiscal year 2002.
Men and women who were older than 18 years of age and presented with acute HF to one of these six EDs were eligible for inclusion in the study. Patients were identified by the medical records code used in the ED (code 150.X of the International Statistical Classification of Diseases and Related Health Problems, 10th revision) that identified HF as the most responsible cause of presentation. Patients were also required to have a valid provincial health number to be included in the study (for analysis of readmission and mortality data). Patient charts were excluded when the diagnosis of HF was not made in the ED (ie, lack of congruency between the ED and hospital discharge diagnosis). In cases of multiple admissions of the same patient during the review period, the earliest visit was defined as the index visit.
All ED visits between April 1, 2002, and March 31, 2003 were eligible for inclusion in the study. Capital Health’s Clinical Performance, Information and Research Unit randomly selected 30% of acute HF patients over the study period to avoid seasonal and physician biases, using a computerized random numbers system. A minimum of 50 patients per site were selected to allow for stable estimates from the smaller EDs.
A standardized form was used to extract data from HF charts and was tested on the first 20 patients. Data elements collected included demographic characteristics, details about the initial presentation to the ED, symptoms prompting the visit, medications self-administered before ED presentation, examinations and investigations performed, treatment provided in the ED and in the hospital within the first 72 h (to separate treatment of acute from chronic HF), and outcomes of care. Each chart was reviewed by centrally trained reviewers. ED or hospital readmission data were available from Capital Health for any of the 17 hospitals in the region.
The primary objective of the present study was to describe current treatment patterns of patients presenting to the ED with acute HF. Secondary objectives included describing treatment administered within the first 72 h of admission to the ED, as well as outcomes of admission to the hospital, LOS in the hospital, LOS in the ED, and readmission and mortality at 30 days. Multivariate logistic regression analysis was conducted to investigate predictors of readmission and mortality.
All data were entered into a database by staff at the Epidemiology Coordinating and Research (EPICORE) Centre at the University of Alberta. Proportions, means with SDs, and medians with interquartile ranges (IQRs) were calculated for descriptive analysis. Logistic regression analysis was used to obtain unadjusted and adjusted ORs (with 95% CIs) of readmission or death within 30 days of the index admission. Multivariate logistic regression analysis was conducted using a backwards stepwise approach. Clinically important variables, including age, sex and smoking status at primary admission, were forced into the multi variate logistic regression model.
The sample size was estimated using Epi Info version 6.0 (Centers for Disease Control and Prevention, USA) for population surveys. The sample size was calculated based on an expected use of diuretics by 84% of patients (from the ADHERE registry) (11). For a confidence level of 95% and a range of ±5%, it was estimated that a sample size of 206 was required. The sample size was increased by adding an additional 264 patients from other EDs in the region to evaluate differences in treatment patterns between EDs in a future analysis.
Of the 483 patients randomly selected for data extraction, 448 patient charts were fully reviewed and included. Of the 35 excluded patients, 17 patient charts were unavailable at the time of data extraction, 12 patients had multiple presentations (therefore, only their initial visit was used in the analysis), three charts had incorrect coding and three charts did not contain a provincial health number. The numbers of patients screened, excluded and included are reported in Figure 1.
The baseline characteristics of patients included in the study are reported in Table 1. The mean age of the patients in the study was 75.3±11.2 years and 45% were women. Smoking status was not documented in 33% of patients.
Medical histories of important comorbidities and risk factors for cardiovascular disease included a documented history of HF (63%), hypertension (55%), coronary artery disease (39%), a previous myocardial infarction (38%), a history of COPD (28%), dyslipidemia (25%) and angina (17%).
Medications administered to patients in the first 72 h after presentation are documented in Table 2. Forty-eight per cent of patients received intravenous furosemide; other medications used were angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (45%), oral furosemide (42%), salbutamol (38%), antiplatelet agents (35%), ipratropium (31%) and anticoagulants (30%). Of note, only one patient received an intravenous inotrope (dopamine). Dobutamine and milrinone were not used.
Fifty-four per cent of patients were admitted to the hospital. The remainder were discharged directly from the ED. Of those admitted to the hospital, the median LOS was 7.5 days (IQR 11.8 days). The median ED LOS for discharged patients was 7.1 h (IQR 6.1 h). Readmission rates were high, with 20% of hospital patients readmitted at 30 days, and 19% of those discharged from the ED readmitted at 30 days (χ2=0.05, degrees of freedom=1; P=0.80). With regard to mortality, 5% of those hospitalized and 1% of those discharged from the ED died 30 days after the index presentation to the ED (Table 3).
With regard to predictors of readmission or mortality at 30 days, age, history of HF and oxygen administration in the ED were significant predictors of these outcomes in the univariate analysis. Age was found to be an independent predictor of readmission or death at 30 days after multivariate analysis (OR 1.03; 95% CI 1.00 to 1.05), as was a history of HF (OR 1.76; 95% CI 1.02 to 3.04), a history of angioplasty (OR 3.48; 95% CI 1.35 to 9.00) and oxygen administration in the ED (OR 1.86; 95% CI 1.01 to 3.43). No medications administered within 72 h of ED admission had any association with reduced readmission or death (Table 4).
Exacerbation of HF is a common reason for ED visits and hospitalizations in Canada (3). Despite this, little is known about how acute HF is managed in Canada. Our review of six EDs has shown that the management of acute HF focuses on symptomatic relief, and includes a variety of treatments for comorbidities such as COPD and coronary artery disease. A particular concern is that less than one-half of these high-risk patients received ACE inhibitors or angiotensin receptor blockers, only one-fifth received a beta-blocker or digoxin, and one-eighth received spironolactone, which are all evidence-based therapies proven to reduce readmissions and/or mortality (12). We observed high rates of readmission, especially in patients discharged directly from the ED. This is notable because individuals discharged directly from the ED represent almost one-half of all ED presentations for HF.
As with most HF populations, the patients in the present study were generally elderly and presented with a substantial range and number of comorbid conditions. More than one-half of the patients selected for the study were reported to have a history of hypertension and 38% had a previous myocardial infarction. Diabetes, coronary artery disease, atrial fibrillation and dyslipidemia were also prevalent in the present patient population. Patients in the current study were approximately the same age as those in ADHERE; however, ADHERE reported a higher proportion of patients who were women (52% versus 45%) and a higher proportion of patients who had a history of hypertension (72% versus 55%). Other common comorbidities reported in the ADHERE registry were similar to our findings, despite the fact that the ADHERE data are restricted to include hospitalized patients only (11).
Not surprisingly, drug therapy in the ED was mostly targeted toward the patient’s symptoms, consisting mostly of furosemide and bronchodilators such as salbutamol and ipratropium bromide. Vasoactive therapies such as nitrates were used relatively infrequently and intravenous inotropes even less so. These results are similar to those reported in the ADHERE registry for inpatients (11). Agents such as dobutamine, dopamine and milrinone were used in 3% to 6% of patients in ADHERE, compared with almost none in our study population. This observation may relate to the adequacy of response to initial therapy or the resource implications (need of higher levels of nursing care) of using intravenous inodilators – a reality of the Canadian health care system.
Although there have been many randomized trials of chronic HF therapy, there is a paucity of such evidence for the management of acute HF. Several groups have issued guidelines, including the European Society of Cardiology (10). Their recommendations are based on elements of patient symptoms such as improving oxygen saturation, controlling pain, eliminating fluid buildup and treating arrhythmias (10). With regard to these recommendations, documentation of oxygen administration was confirmed in 69% of patients in the present study. Analgesia and opioids were not commonly used in our patient population. Vasodilators are recommended in guidelines for patients with a mean arterial pressure of greater than 70 mmHg; however, these agents were only used in 29% of our acute HF patients. The most commonly used vasoactive agent was nitroglycerin administered as a transdermal patch.
After the acute event, optimizing chronic HF medications should be a priority because ACE inhibitors, beta-blockers and spironolactone have all been shown to reduce mortality and hospitalizations, and digoxin has been shown to reduce hospitalizations (13). Interestingly, while 70% of patients were receiving ACE inhibitors or angiotensin receptor blockers on admission, only 45% were documented to be receiving these agents within 72 h after ED admission. These results are consistent with the findings of a comprehensive review of ACE inhibitor use by Bungard et al (14), who reported a weighted average of 53% of patients receiving ACE inhibitors on discharge from the hospital. The low use of ACE inhibitors after 72 h may have contributed to the high readmission rates. Beta-blockers, which have an even greater impact on mortality and hospitalizations than ACE inhibitors, were used in only 22% of patients. While it may not be feasible to initiate beta-blockers in all patients with acute HF, the proportion reported in the sample would be considered suboptimal.
The outcomes of patients presenting to the ED with acute HF are important to examine in further detail, especially within the context of overcrowded Canadian EDs. The acute HF patient often requires extended and comprehensive care, which further congests EDs and reduces patient throughput (15). More than one-half of the patients were admitted to the hospital, with a median LOS of approximately 7.5 days, suggesting significant resource use.
The readmission rates of patients discharged directly from the ED were essentially the same as those who were admitted to the hospital (who were presumably sicker and more aggressively managed). This highlights the need for more comprehensive and evidence-based HF care, rather than simply symptomatic management. Moreover, there was no indication of any patient education on HF and patient self-care. Most urban Canadian EDs are overcrowded and it is likely that education is a low priority in such chaotic environments (16).
Why is education an important issue in this setting? First, previous research has suggested that excessive sodium intake was the most common reason for exacerbation of HF (17). While it may be unrealistic for busy ED physicians to obtain a comprehensive dietary intake history, this may be another role for HF clinics. Second, in addition to dietary history, dedicated clinics are more capable of initiating patient education, as well as initiation/titration of ACE inhibitors and angiotensin receptor blockers, beta-blockers, spironolactone and/or digoxin. In a meta-analysis of studies involving heart dysfunction clinic strategies that incorporated follow-up by a specialized multidisciplinary team (in either a clinic or nonclinic setting), mortality, HF hospitalizations and all-cause hospitalizations were reduced (18). The results of most studies of these specialty clinics suggest that they are also cost effective (18).
Our analysis did not reveal that any medications administered within the first 72 h of ED admission are associated with reduced readmission or mortality at 30 days. This may be because we only collected medication use for the first 72 h of the patient’s stay or that none of the medications used impact 30-day outcomes.
The present study has potential limitations that should be noted. As occurs in many retrospective studies, poor chart documentation was identified, particularly in discharge planning. While this is worth noting, it should not significantly affect the results of the study because ED pharmacological treatments were well documented on each patient chart and outcomes were determined through linked administrative data. Detection bias associated with data abstraction was possible; however, standardized data collection forms were used and all data abstractors were trained by the investigating team. While the study employed several individuals to collect the data, high agreement and a focus on objective data (ie, medication use, LOS and demographics) should ensure the main study results are valid. Relapse data were not readily available due to the use of administrative data for follow-up. This may have underestimated event rates because events (mortality and readmission) that occurred outside the Capital Health Region would have been missed, although we believe this to be a rare occurrence. Fourth, only one region’s experience was examined, and this did not include rural hospitals, other provinces or nonteaching hospitals. The bias associated with this may limit the generalizability of the results; however, if this is the state of affairs in a linked, academic program, it is likely that this approach underestimates the real-world use of proven efficacious therapies. Physician groups in teaching programs may have been exposed to more educational activities, may undertake chart audits more often and may overlap shifts more often, thus encouraging practice standardization. We did not anticipate such a high number of patients seen in the ED and discharged home, and therefore, did not power the study for comparisons of outcomes between the two groups. Finally, no evaluation of left ventricular dysfunction was completed. Therefore, it could not be determined whether HF was due to systolic or diastolic dysfunction (for which the evidence for proven efficacious therapies differs). Of note, this was rarely documented, so it may be that it was not even considered in patient management decisions.
In the present six-hospital review of acute HF, we observed that ED treatment is mostly symptomatic and not evidence-based, even following admission to the hospital. Areas for improvement include increasing the use of proven efficacious medical therapies for chronic HF, more research about treatment efficacy for acute HF and evaluation of different strategies for delivering patient education. The subset of patients who are treated symptomatically and then discharged home appears to be at high risk for readmission, suggesting that physicians may benefit from decision tools to predict poor outcomes and patients could benefit from specialized HF care.
The authors thank the Capital Health Clinical Performance, Information and Research Unit for their assistance, especially Ms Chris Houston. They also thank the staff of the Epidemiology Coordinating and Research (EPICORE) Centre at the University of Alberta for their assistance. The work of Betty Mah Pon (University of Alberta Hospital Health Records), Debbie Boyko (Northeast Community Health Centre, Edmonton), Michelle Pachal (Sturgeon Community Hospital, St Albert) and Brigette Zilkalns (University of Alberta Hospital) is acknowledged. Finally, the authors extend their gratitude to Natalie McMurtry, Margaret Ackman, Nesé Yuksel and Lindsay Torok-Both for their advice and encouragement.
SUPPORT: Dr Rowe is supported by a 21st Century Canada Research Chair in Emergency Airway Diseases from the Government of Canada (Ottawa, Ontario). Dr Tsuyuki is supported by a Merck Frosst Chair in Patient Health Management (University of Alberta). Ms Bresee was supported by a full-time Health Studentship from the Alberta Heritage Foundation for Medical Research.