Heart failure (HF) is the most common reason for hospitalization in patients older than age 65 and is the most common reason for rehospitalization in the same group. It accounts for more costs to Medicare than any other single disease entity. The money spent annually on HF in the United States exceeds the gross domestic product of Kuwait.

The prevalence of HF has risen sharply, resulting in part from the rapid growth of the elderly population, improved survival following acute myocardial infarction and treatment of chronic hypertension (1). As expected, the number of HF hospitalizations has also risen. In 2004, there were over 1 million HF hospitalizations in the United States and another 1 million in Europe. It is estimated that these hospitalizations account for more than 75% of the 46 billion dollars spent each year on the care of HF patients (2).

Our understanding of the demographics of HF has evolved significantly over the last decade. If clinical trials were to define the HF population, the typical patient might be described as outlined in Table 7-1. However, large databases show that clinical trials do not reflect the true HF population. The Acute Decompensated Heart Failure National Registry (ADHERE) has contributed importantly to our current understanding of HF. This registry includes hospitalization information on patients discharged with a diagnosis of, or primary treatment for, acute decompensated heart failure (ADHF) (Diagnostic Related Group 127). The registry contains information on greater than 200,000 ADHF hospitalizations and can be examined to determine current patient demographics and treatment regimens for ADHF. Evaluation of this data set indicates that the average ADHF patient is much older, more often female, and burdened by more coexistent disease than previously appreciated (see Table 7-1).

The ADHERE registry has also provided important information regarding the process of care. Most patients ultimately diagnosed with ADHF present to the emergency department (ED). Of the overall inpatient ADHF population, 78% were admitted through the ED. In the hospitalized population, the chief complaint is remarkably consistent, with 89% reporting shortness of breath. Once admitted to the hospital, 79% of ADHF patients are placed in either a telemetry bed (66%) or an intensive care unit (ICU) (13%). These admission destinations represent some of the greatest costs for the U.S. hospital system and suggest opportunities for improved management strategies. As a consequence, in an effort to decrease the cost associated with inpatient admission, reimbursement strategies to encourage short stay observation unit treatment have been structured. Usually managed as part of an emergency visit, observation unit admissions increased approximately 100% between 2003 and 2004. Finally, even in the patient destined for hospitalization, because most admissions arise following an ED visit, appropriate ED evaluation and intervention are required to initiate quality care.

Recently a second iteration of the ADHERE registry was begun. Specifically designated as an emergency module (ADHERE-EM), this registry records data on the emergency medicine encounter and will ultimately track outcomes for 60,000 ED ADHF patients. Additional features have been added to the ADHERE-EM registry. One of the most important improvements is the ability to track visits longitudinally, so that the long-term impact of diagnostic and treatment interventions can be evaluated. Also unique to the ADHERE-EM registry is the requirement that participating institutions implement a disease management strategy at the beginning of data collection. Participating centers are required to implement at least three of the first five disease management strategies listed in Table 7-2.

Historically, hospitals have approached the financial aspects of their medical operations by utilizing a departmental framework. In this manner the hospital is divided into units, with each being responsible for its own costs and profits. Although this is an attractive model for management of a chain of independent gas stations, the application to a medical facility, where reimbursements do not necessarily return equally to all participants involved in a patient’s outcome, can result in impediments to quality care. For example, the department of cardiology may wish to use an expensive therapy because it is able to shorten cardiology care unit (CCU) length of stay by 1 day and improve their DRG remuneration margin. However, the pharmacy may be less than enthusiastic due to the fact that they must provide this medication at increased costs to their budget, but not receive any direct pharmacy department value from a shorter CCU admission.

Disease management is therefore the principle of removing artificial barriers to improved patient outcomes so that all the participants involved in ADHF care are similarly invested and rewarded based on the final clinical outcome, rather than their individual departmental financial result. This approach is particularly attractive in chronic disease management, such as ADHF, where a number of studies have demonstrated shorter length of hospitalization, a lower revisit frequency (3,4), and a decreased mortality when disease management has been applied to HF (3). In a recent meta-analysis of 29 trials that included over 5,000 HF patients, they demonstrated that when multidisciplinary team follow-up was used, there was a reduction of both mortality and HF hospitalizations by approximately 25%; mortality RR 0.75 (95% CI 0.59 to 0.96), HF hospitalizations RR 0.74 (95% CI 0.63 to 0.87) (5). One of the important parameters of disease management is that no specific therapy or device is mandated. It is merely the provision of care across departmental lines, with the shared goals of increasing quality patient outcomes.

Initial data from the ADHERE-EM registry has already identified trends in ADHF patients that were not readily apparent from historical data. First, as would be suspected intuitively, ED patients appear to have greater severity of illness than the hospitalized cohort. Although patients admitted to the hospital by the ED constitute nearly 80% of the inpatient cohort, this population has a severity of illness that is diluted by the 20% of patients representing direct admissions from a non–acute care environment (e.g., a physician’s office). Increased severity of illness in the ED ADHF population is manifest by greater rates of renal insufficiency, higher mean blood urea nitrogen (BUN) and creatinine, higher B-type natriuretic peptide (BNP) levels, and importantly, increased troponin levels. The median troponin in the ED patient with ADHF has been reported to be 0.06 ng/mL for troponin I, and 0.03 ng/mL for troponin T. Although controversial, these findings have clear outcome implications in the ED ADHF population.

In an analysis of 70,000 ADHF patients from the ADHERE data set, only 6.3% had an elevated troponin. However those with increased troponin had markedly worsened acute adverse outcomes that included longer hospitalization, longer ICU time, an increased rate of the composite of endotracheal intubation, intra-aortic balloon counterpulsation, and coronary artery bypass grafting, as well as a fourfold increase in acute mortality (Table 7-3) (6). In a separate analysis of 14,000 ADHF patients (7) with elevated troponin, those receiving vasodilator therapy had in-hospital mortality rates of only 5%, compared to 22% when inotropic therapy was used, despite risk adjustment for systolic and diastolic blood pressure, BUN, creatinine, sodium, heart rate, dyspnea at rest, and age. An elevated troponin is clearly a marker of near-term serious adverse outcomes in ED patients with ADHF.

The ADHERE-EM registry has also provided insight into the challenge of diagnosis in the ED environment. If the initial ED admitting diagnosis is compared to the diagnosis obtained following the complete hospitalization, when all testing and patient responses to therapy can be considered, the diagnosis
is concordant in only 83% of patients. This suggests that the ED misdiagnosis rate in patients ultimately determined to have ADHF is 17%. This has clear implications for the selection of therapies and interventions.

This misdiagnosis rate is consistent with other literature. In the Breathing Not Properly trial (8), the correct diagnosis of HF, based on history and physical in over 1,500 patients presenting to the ED with acute dyspnea, was 75%. When BNP results were considered, the correct diagnosis rate increased to 81.5%. Their suggested misdiagnosis rate of 18.5% is consistent with the error rate as determined by the ADHERE-EM registry.

Thus, the potential for misdiagnosis must be considered in the setting of adverse outcomes. The VMAC trial (9) was a double-blind, randomized, standard therapy controlled study of 489 ADHF patients receiving either nitroglycerin or nesiritide. Of these, 1% of the nitroglycerin group and 0.5% of the nesiritide cohort sustained symptomatic hypotension within 3 hours. This low rate of symptomatic hypotension has clinical outcomes in the ED. If, after an intravenous vasodilator is administered to a patient with suspected ADHF, symptomatic hypotension occurs, the possibility of misdiagnosis must be considered. Vasodilator-induced hypotension occurs in 1 of 100 ADHF patients, whereas misdiagnosis is seen in nearly one in five patients ultimately found to have ADHF. Additionally, although the rate of symptomatic hypotension from a vasodilator being given to patients with an ADHF mimic is unknown (e.g., pulmonary embolus, pneumonia), this cohort could reasonably be expected to sustain symptomatic hypotension. Therefore, when faced with symptomatic hypotension, the clinician must first consider that misdiagnosis is numerically much more likely than a primary drug effect. In this situation the drug should be terminated immediately and the differential diagnosis reconsidered.

A number of challenges exist that result in ED misdiagnosis. This is the consequence of the lack of sensitivity and specificity of history and physical exam findings, an elderly patient population with numerous coexistent pathologies (e.g., chronic obstructive pulmonary disease [COPD] and HF), environmental factors (a loud and sometimes chaotic ED), limits to testing in the ED, and the compressed time course of an ED visit.

As has been demonstrated by the ADHERE registry, approximately three-quarters of patients with ADHF represent repeat visits of previously established disease. However, because the ADHF population has a median age of approximately 75 years, concurrent disease can represent a significant challenge in determining which pathology has prompted the current presentation. In one study, the sensitivities and specificities of common physical findings were determined (Table 7-4) (10), and none have sufficient accuracy to be considered definitive.

Furthermore, testing in the ED is limited to those interventions that can be performed rapidly and that have limited invasiveness. Consequently, workups are limited to electrocardiogram (ECG), blood testing, and radiologic studies. An ECG should be performed in all ED-suspected ADHF patients. Although it is of value to diagnose acute ischemia in the breathless patient, it is so insensitive for the detection of ADHF as to be functionally useless.

A chest radiograph should be obtained in all suspected ADHF patients. It has better accuracy than the ECG, but must be interpreted with caution. X-rays cannot exclude abnormal left ventricular (LV) function, but can eliminate other diagnoses (e.g., pneumonia). The x-ray findings of HF are, in descending order of frequency: dilated upper lobe vessels, cardiomegaly, interstitial edema, enlarged pulmonary artery, pleural effusion, alveolar edema, prominent superior vena cava, and Kerley lines (11). Because abnormalities lag the clinical appearance by hours, therapy is not withheld pending a film.

In chronic HF patients, which is the common ED presentation, the chest x-ray (CXR) signs of congestion have unreliable sensitivity, specificity, and predictive value in identifying patients with high pulmonary capillary wedge pressure (PCWP) (12). In one analysis, radiographic pulmonary congestion was absent in 53% of patients with mild to moderately elevated PCWP (16 to 29 mm Hg) and in 39% of those with markedly elevated PCWP (>30 mm Hg).

Cardiomegaly can suggest an HF diagnosis, and a cardiothoracic ratio >60% correlates with increased 5-year mortality (13), but this cannot help in the compressed time frame of ED decision making. Ultimately, the CXR has poor sensitivity for cardiomegaly (14). In fact, in patients with echocardiographically proven cardiomegaly, 22% had cardiothoracic ratios <50% (14). The poor detection of cardiomegaly by the CXR has been explained by intrathoracic cardiac rotation.