OUP user menu

Reassessing treatment of acute heart failure syndromes: the ADHERE Registry

Mihai Gheorghiade, Gerasimos Filippatos
DOI: http://dx.doi.org/10.1093/eurheartj/sui008 B13-B19 First published online: 15 March 2005

Abstract

Acute heart failure syndromes (AHFS) are the leading cause of hospitalization in persons aged over 65 years, costing an estimated $12.7 billion per year in the United States alone. Acute Decompensated Heart Failure National Registry (ADHERE) was established in the United States in 2001, and now represents the largest database on acute heart failure in the world, with over 100 000 patients enrolled to date. The first analyses of data from the registry are now in the public domain, and offer useful insights into the characteristics, management, and outcomes of AHFS patients. The ‘real-life’ patient population represented in ADHERE is older, contains more women, and has a higher incidence and complexity of co-morbidities than individuals studied in clinical trials. Patient management varies substantially, and often does not utilize the full range of technical resources and clinical knowledge available. It is also notable that among patients admitted with worsening chronic heart failure, at least one-third are not receiving well-established life-saving medications. The information gathered so far by ADHERE has highlighted key areas for improvement in the management of both acute and chronic heart failure.

  • Acute disease
  • congestive heart failure
  • classification/epidemiology
  • Hospitalization/statistics and numerical data
  • Patient registry
  • United States

Introduction

Acute heart failure syndromes (AHFS) cause almost 1 million hospitalizations annually in the United States,1 placing an enormous social and financial burden on society. Indeed, heart failure is the leading cause of hospitalization in persons aged over 65 years, and in 2003, inpatient management of AHFS cost an estimated $12.7 billion in the United States.2 Although European data on acute heart failure are limited, the EuroHeart Failure Survey36 and the majority of studies suggest that heart failure is also a large and growing public health burden throughout this continent. Despite the substantial survival and financial impact of the disease, there are few randomized trials studying the efficacy and safety of different therapies for AHFS. Over the past few years, the extent of the worldwide public health burden of AHFS has been brought into sharp focus by a number of studies, fuelling an interest in defining optimal treatment and patient management strategies.

A key step towards designing relevant research trials to evaluate treatment of AHFS is the generation of a database of patients hospitalized with heart failure. Acute Decompensated Heart Failure National Registry (ADHERE) is one such database, which was established in October 2001 and is now the most extensive acute heart failure registry in the world, with more than 100 000 patients enrolled to date.2,7 The registry is an observational study involving more than 275 sites throughout the United States, under the guidance of the ADHERE Registry Scientific Advisory Committee (Table 1).7 Patients over the age of 18 years who are admitted to an acute care hospital with a primary or secondary discharge diagnosis of heart failure are eligible for inclusion. Details of medical history, clinical presentation, laboratory tests, medical management, and health outcomes are collected from hospital discharge medical records.

View this table:
Table 1

The ADHERE Registry Scientific Advisory Committee2

InvestigatorInstitutionLocation
William T. AbrahamThe Ohio State University Heart CenterColumbus, OH
Kirkwood F. Adams JrUniversity of North CarolinaChapel Hill, NC
Robert L. BerkowitzHackensack University HospitalHackensack, NJ
Maria Rosa CostanzoMidwest Heart SpecialistsNaperville, IL
Teresa De MarcoUniversity of CaliforniaSan Francisco, CA
Charles L. EmermanThe Cleveland ClinicCleveland, OH
Gregg C. FonarowAhmanson—UCLA Cardiomyopathy CenterLos Angeles, CA
Marie GalvaoMontefiore Medical CenterBronx, NY
J. Thomas HeywoodLoma Linda University Medical CenterLoma Linda, CA
Thierry H. LeJemtelAlbert Einstein HospitalBronx, NY
Lynne Warner StevensonBrigham and Women's HospitalBoston, MA
Clyde W. YancyUniversity of Texas, Southwestern Medical CenterDallas, TX

The database has several goals: first, to describe demographics and clinical characteristics of AHFS patients; secondly, to characterize current management of hospitalized patients with AHFS; thirdly, to generate hypotheses for randomized controlled trial design; and most importantly, to evaluate and improve the quality of care for these patients. The first analyses of data from the ADHERE Registry entered the public domain in 2003, and have already provided important insights into characteristics and management of patients with AHFS.

The real value of ADHERE lies in identifying and acting on those findings that can make a significant impact on clinical management of AHFS. With this aim in mind, this paper highlights some of the most important insights to have emerged to date. Unless stated otherwise, figures quoted here refer to patients enrolled in the registry who were discharged from an acute care hospital between October 2001 and January 2004 (n=105 388).8

Patient characteristics

The characteristics of patients included in the ADHERE database highlight an important caveat: that the populations studied in heart failure clinical trials often differ substantially from ‘real-world’ patients with heart failure. For example, the median age of patients enrolled in the registry was 75 years and 52% of patients were female.2 This is consistent with other recent hospital surveys but contrasts with the majority of clinical trial participants, both in the United States and Europe, who tend to be younger and predominantly male.9

Similarly, although the majority of patients with significant co-existing morbidity may be excluded from clinical trials, a high degree of co-morbidity was shown among patients in the ADHERE database. Many patients had a history of hypertension (72%) or coronary artery disease (57%); 44% had diabetes, 35% had a history of hyperlipidaemia, 31% had atrial fibrillation, and 31% had chronic obstructive pulmonary disease or asthma. The high frequency of these diagnoses within the ADHERE database underscores the complex challenges faced by physicians when managing patients who may suffer from multiple conditions that contribute to a poor prognosis.

Systolic function

Interestingly, among those patients in whom left ventricular ejection fraction (LVEF) was measured, only 47% had LVEF <40%.2 An analysis of patients enrolled up to January 2003 found that patients with LVEF >40% were older, more likely to be female or have diabetes, and less likely to have coronary heart disease or myocardial infarction than those with compromised LVEF (Table 2).10 This subgroup of patients presented with an increased incidence of rales, peripheral oedema and hypertension, a lower incidence of hypotension, and diminished elevation in B-type natriuretic peptide (BNP) levels. Although the ADHERE data indicate that patients with relatively preserved systolic function represent almost half of all patients with AHFS, these patients would have been excluded from most recent clinical trials.9

View this table:
Table 2

Clinical characteristics of patients with relatively preserved vs. impaired systolic function10

CharacteristicsLVEFP-value
>40% (n=7829)<40% (n=8245)
Age (years)73.969.9
Female sex (%)6340
History of CAD (%)5261<0.0001
History of MI (%)4761<0.0001
History of diabetes (%)4541<0.0001
Rales on presentation (%)71690.002
Peripheral oedema on presentation (%)7064<0.0001
Serum BNP level on presentation (pg/mL)647918<0.0001

CAD, coronary artery disease; MI, myocardial infarction.

Renal dysfunction

Signs of renal dysfunction were apparent in a substantial proportion of patients: 30% had a history of chronic renal insufficiency, 13% had an admitting diagnosis of renal insufficiency or failure, and one-fifth of those whose serum creatinine was determined had levels >2.0 mg/dL.2 Importantly, compromised renal function was associated with poor prognosis: classification and regression tree analysis of risk factors for in-hospital mortality found that the best single predictor for mortality was an admission blood urea nitrogen level of >43 mg/dL.11 Along with low admission systolic blood pressure, high serum creatinine—another well-established indicator of poor renal function—was also a predictor of in-hospital mortality.

Congestion

Only 2% of patients had a systolic blood pressure (SBP) <90 mmHg, with 48% of patients classified as normotensive (SBP: 90–140 mmHg) and 50% hypertensive (SBP >140 mmHg). Most patients had signs of congestion on presentation (Table 3), and heart failure, recorded in 93% of patients, was the most common admitting diagnosis. Dyspnoea was the most common symptom, and more than one-third of patients experienced dyspnoea at rest while being examined. In addition, many patients exhibited rales or peripheral oedema, and 75% had pulmonary congestion on chest X-ray. The majority of patients were severely symptomatic [New York heart association (NYHA) Class III or IV], and 20% were in atrial fibrillation on admission. The median level of BNP was 667 pg/mL in patients who underwent this assessment.

View this table:
Table 3

Signs and symptoms of congestion on presentation in patients enrolled in the ADHERE database, October 2001–January 2004 (n=105 358)8

Signs/symptomsProportion of patients (%)
Any dyspnoea89
NYHA Class III or IV87
Pulmonary congestion by CXR75
Rales68
Peripheral oedema66
LVEF <40%47
Dyspnoea at rest34
Fatigue32

CXR, chest X-ray.

On discharge, 52% of patients reported that they were asymptomatic, 37% were improved (but still symptomatic), <1% were unchanged, and <1% were worse. However, congestion was not substantially improved in many patients during their inpatient stay—one-third had lost only 0–5 lb in weight on discharge and 16% had actually gained weight (Table 4). The tendency to discharge patients without significant improvement of congestion may reflect a dissociation between patients' self-assessment of their condition and physician-determined signs of AHFS. Although this suboptimal treatment of congestion appears to provide short-term improvement in symptoms, it could well be a major contributor to the high readmission rates.12 This is a matter of great concern, given the mounting evidence that congestion should be a major treatment target in patients with AHFS.13

View this table:
Table 4

Body weight change at discharge in patients enrolled in the ADHERE database, October 2001–January 2004 (n=51 013)8

Change in body weight at dischargeProportion of patients (%)
Decrease of >20 lb (>9 kg)7
Decrease of 15–20 lb (6.8–9 kg)6
Decrease of 10–15 lb (4.5–6.8 kg)13
Decrease of 5–10 lb (2.3–4.5 kg)24
Decrease of 0–5 lb (0–2.3 kg)33
Increase of 0–5 lb (0–2.3 kg)11
Increase of 5–10 lb (2.3–4.5 kg)3
Increase of >10 lb (>4.5 kg)2

Work-up and treatment

Diagnosis and patient management

Considering the complex physiology underlying the presentation of these patients, one noteworthy observation from the ADHERE data is the relative lack of comprehensive work-up. In particular, recently developed technologies for deriving diagnostic and prognostic information seem to be underused. Electrocardiogram, chest radiography, assessments of systolic blood pressure, serum sodium, and serum creatinine were performed in >90% of patients. However, LVEF, BNP levels, and NYHA class were determined in only 58, 35, and 11% of patients, respectively. Increases in LVEF and BNP have been shown to be predictors of increased rehospitalization and mortality in patients with AHFS,13 and their use may identify patients with increased mortality risk. Data from the ADHERE Registry suggest that improvements in patient assessment could make an important contribution for optimizing the treatment of the condition.

A surprising finding is that patient management varies substantially, depending on the hospital department providing treatment—worryingly, this disparity appears to translate into significant differences in patient outcome.14 Although 78% of patients in the ADHERE Registry were initially treated in an emergency department (ED), a substantial proportion (21%) were admitted directly for inpatient care. There was an unexpected difference in time-to-treatment between patients first given intravenous vasoactive therapy in the ED (median time of 1.1 h after admission) and those whose vasoactive therapy was not initiated until admission to an inpatient unit (median time of 22.2 h after admission). This had substantial effects on outcomes: early administration of intravenous vasoactive therapy in the ED was associated with significantly lower mortality, fewer transfers to the intensive care or critical care unit, and decreased requirement for invasive procedures. Furthermore, patients who received intravenous vasoactive treatment in the ED had shorter in-hospital and intensive/critical care unit median length of stay, and fewer of these patients required prolonged hospitalization. These findings require further investigation, given that patients included in the analysis received a wide range of vasoactive drugs with differing toxicity and efficacy profiles.

Use of inotropes

Previous studies have indicated that the use of inotropic agents may be associated with increased risk of adverse events and increased mortality, particularly in patients with preserved left ventricular function.15 Therefore, it has been recommended that positive inotropes should only be used in patients who absolutely require inotropic support for low cardiac output.13 With this in mind, it is concerning that 9% of patients in the ADHERE Registry received a positive inotrope (dobutamine or milrinone) as part of their inpatient care, although only 2% of patients were hypotensive. This suggests that >75% of patients who receive these potentially deleterious drugs may have no clinical indication for their use, and underlines the importance of developing consensus guidelines for management of patients with AHFS.

One analysis of ADHERE data provides evidence that patients with preserved systolic function may respond differently to treatment with positive inotropes compared with those with impaired LVEF.16 In this subgroup analysis, patients with preserved systolic function who received inotropic agents had a significantly longer length of hospital stay (mean: 12.9 vs. 9.6 days; P<0.0001) and higher mortality rate (19 vs. 14%; P<0.002) than all other inotrope-treated patients. Furthermore, among patients with preserved systolic function, those who received inotropic agents had more than double the length of stay (12.9 vs. 5.8 days; P<0.0001) and a nine-fold increase in mortality rate (19 vs. 2%; P<0.0001) compared with those who were not treated with inotropes. As patients with preserved systolic function are often excluded from clinical trials, this suggests that the adverse outcomes associated with these drugs could potentially affect a greater proportion of patients in ‘real-life’ populations than in clinical studies. These data highlight the need to examine the effects of therapies on this important, but perhaps overlooked, patient subgroup. Designing clinical studies with this aim should, therefore, be a key focus of future research.

Use of intravenous diuretics and vasoactive therapy

During their inpatient stay, 88% of patients in ADHERE were treated with intravenous diuretics. However, despite the ability of these drugs to provide rapid symptomatic improvement, several studies have found an association between high-dose diuretic use and adverse outcomes.1719 Consistent with these findings, a recent analysis of ADHERE data found that patients who received intravenous diuretics had a longer overall length of stay (P<0.001) as well as a longer ICU length of stay (P<0.001) and slightly higher mortality (non-significant) than patients who did not receive these drugs.20 A recent study also found that chronic use of diuretics was associated with increased mortality among patients in ADHERE who had renal insufficiency or earlier stages of renal disease: high creatinine levels and chronic treatment with diuretics were strong independent predictors of mortality.21 Moreover, it has been shown that high-dose isosorbide dinitrate is more effective than high-dose furosemide, in terms of need for mechanical ventilation and frequency of myocardial infarction, in patients with pulmonary oedema.22 Although high-dose intravenous diuretic use may indicate patients with more severe heart failure, this finding emphasizes the need for large-scale randomized, controlled trials of diuretics in AHFS. Despite the widespread use of diuretics, there are very few clinical data to support their safety and efficacy in AHFS.

The nitrovasodilator nesiritide, a recombinant exact copy of BNP, was used in 10% of patients in the ADHERE Registry, and a further 11% received either nitroglycerin or nitroprusside. The mortality associated with nesiritide has been reported to be similar to that of nitroglycerin and lower than that of dobutamine.23,24 Similarly, an analysis of ADHERE data found that treatment with nesiritide within the first 24 h of hospitalization for AHFS was associated with significantly lower in-hospital mortality than treatment with milrinone or dobutamine, and similar mortality to treatment with nitroglycerin (Figure 1).25

Figure 1 Comparison of in-hospital mortality in patients treated with nesiritide vs. nitroglycerin, milrinone, or dobutamine.25

Outcomes and quality of care

The median hospital stay for patients in the ADHERE Registry was 4.3 days and in-hospital mortality was 4%. This mortality rate is above that reported in several large heart failure clinical trials [e.g. 2.3% for the placebo group in the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME CHF) trial, who received standard care as deemed appropriate by their physician26]. Again, this may reflect the selective nature of clinical trial inclusion criteria—which often exclude patients with co-existing disease—and emphasizes the value of studies on non-restricted populations, such as the ADHERE Registry.

Performance measured by Joint Commission on Accreditation of Healthcare Organizations (JCAHO) quality of care indicators was often substandard and varied substantially among sites. These quality of care indicators were developed to identify standards of acceptable care in patients hospitalized with heart failure. Left ventricular function was measured in 82% of patients (JCAHO indicator HF-2), but only 63% of eligible patients received an angiotensin-converting enzyme inhibitor (ACE-I) on discharge (HF-3), and just 28 and 31% of patients received a complete set of discharge instructions (HF-1) and smoking cessation counselling (where appropriate, HF-4), respectively.27 These data highlight a clear need to develop treatment guidelines and communicate these widely in order to standardize and improve quality of care.

Long-term patient management

As indicated by the poor performance on JCAHO HF-3, the use of chronic heart failure medications is disappointingly low. US practice guidelines advocate the prescription of ACE-Is and beta-blockers in all patients with chronic heart failure who do not have contraindications for their use.28,29 Despite these recommendations, the use of pre-hospitalization chronic heart failure medications among ADHERE patients was relatively low. Most of the patients (70%) were on diuretics as outpatients. However, only 53% were on either ACE-Is or angiotensin-receptor blockers (ARBs) and only 48% were on beta-blockers. Considering that 75% of patients had a history of heart failure, this suggests that about one-third of eligible individuals are not receiving these medications in the outpatient setting, despite their substantial benefits being well-documented. Even after hospitalization for AHFS, prescription of ACE-Is and beta-blockers remains comparatively low. While 86% of patients were prescribed diuretics, only 69% were given either ACE-Is or ARBs, and only 59% received beta-blockers. A recent study found that in-hospital initiation of beta-blocker therapy increased the rate of use at 60 days after randomization without increasing side-effects or length of stay.30 In light of this finding, prescription of chronic heart failure therapies is clearly an area where there is room for achieving significant improvements in current practice that could provide substantial long-term benefits for patients.

The lessons learned so far

Patient registries are recognized as vital tools in the drive to understand and optimally manage AHFS, providing comprehensive data on patient characteristics, treatments, and outcomes. In the short time since its inception, the ADHERE database has already revolutionized the way we think about heart failure and its management. Along with European initiatives, such as the EuroHeart Failure Survey programme36 and Etude d'observation Francaise de l'Insuffisance Cardiaque Aiguë (EFICA),31 ADHERE has brought several important issues surrounding AHFS into sharp focus (Table 5).

View this table:
Table 5

Key issues highlighted by ADHERE

Patient characteristics
  • Unlike clinical trial populations, ‘real-life’ patients with AHFS are a heterogeneous population: they tend to be relatively old, normo- or hypertensive, and have a high incidence of co-morbidities including diabetes, atrial fibrillation, coronary artery disease, hypertension, and renal dysfunction

  • Over 50% of patients have a LVEF >40%—but patients with preserved systolic function are rarely included in clinical trials

  • Renal dysfunction is a common finding and a major predictor of poor prognosis

Diagnosis and treatment
  • Despite the complex physiology underlying presentation, work-up is not always complete and the available diagnostic technology is under-utilized

  • Congestion is not well treated and warrants greater clinical focus

  • Treatment requires a multidisciplinary approach

  • Many patients with no apparent clinical indication received positive inotropes, even though these drugs are associated with increased mortality

Outcomes and quality of care
  • Performance on JCAHO quality of care indicators requires substantial improvement

  • Many patients receive treatment which affords temporary relief but may ultimately contribute to the high readmission rates

Long-term patient management
  • The implementation of basic life-saving chronic heart failure therapies is not optimal (e.g. beta-blockers)

First, ADHERE has taught us that patients admitted with acute heart failure belong to a heterogeneous population. Most clinical trials are based on patients with compensated systolic dysfunction and few co-morbidities. The ADHERE data show that 53% of patients with AHFS have relatively preserved systolic function. Moreover, these patients tend to be older, predominantly female, and are more likely to have a history of hypertension or diabetes. A multidisciplinary approach—involving control of hypertension, diabetes, and renal failure—is therefore essential to the ultimate treatment of heart failure.

Secondly, it is apparent that treatment of AHFS could be improved substantially. Advanced prognostic and diagnostic techniques that are now available are under-utilized, and patient management varies considerably from site to site across the United States, and between hospital departments. Up to 7% of patients are receiving positive inotropes, drugs that are associated with an increased incidence of poor outcomes, without any clinical indication to support their use. It is clear that more trials and definitive guidelines are needed to establish and promote best practice in the management and treatment of patients with AHFS.

Thirdly, the medical profession is failing to provide life-saving therapies for patients with chronic heart failure. While ADHERE shows that the majority of patients are symptomatically improved on discharge, body weight was not decreased in about half of the patients, and a similar proportion face readmission within 6 months.32,33 It is disappointing to find that many patients are not on essential medications such as ACE-I and beta-blockers. In addition, only 28% of patients receive basic discharge instructions on diet, weight monitoring, activity level, medications, and follow-up.27 An ADHERE-based toolkit has been developed to provide a structured approach to the management of patients with AHFS.2 The toolkit, in the form of standardized algorithms, orders, and discharge checklists, should help bridge the gap between evidence-based and ‘real-life’ medicine.

It is also worth noting that a comparison of heart failure registry data from the United States and Europe reveals some interesting differences between these two regions. For example, average length of hospital stay for patients in the EuroHeart Failure Survey was 11 days, with an in-hospital mortality rate of 6.9%, compared with 8 days and 4%, respectively, for the ADHERE Registry.2,3 The striking difference in length of stay could reflect either varying degrees of morbidity between the two patient populations or differences in patient management (e.g. whether patients with worsening congestive heart failure are admitted to hospital or treated in the emergency room). Further studies are needed to distinguish between these and other possible explanations. These findings underline the importance of conducting not only observational studies to quantify variation in patient populations and their management around the world but also international clinical trials to examine the effect of these differences on the efficacy of therapeutic interventions.

Although European guidelines for acute heart failure have been presented, many of the recommendations are, of necessity, based on empirical observations rather than on randomized trial data. Development of comprehensive evidence-based treatment guidelines for AHFS therefore remains an important future goal.34 In view of the large economic and morbidity burden caused by this disease, and the substantial variation in patient management highlighted by ADHERE, development of guidelines, design of robust, large-scale, randomized trials of therapies, and formation of a consensus on optimal patient management are essential next steps in the modernization of treatment. The ADHERE Registry provides an invaluable opportunity to reassess and improve the management of AHFS.

Acknowledgements

Meeting session and publication supported by an unrestricted educational grant from GlaxoSmithKline Ltd. M.G. has acted as a consultant to GlaxoSmithKline Ltd and Otsuka America Pharmaceutical, Inc., and currently holds research grants from Scios Inc.

References

View Abstract