Introduction

Heart failure is a frequent, serious and expensive syndrome; 1.5–2 % of Europeans suffer from overt (NYHA II–IV) heart failure, and another 1.5–2 % have asymptomatic (NYHA I) ventricular dysfunction [1]. The prevalence is rising because of the increasing life expectancy of the population and the increasing number of long-term survivors of large myocardial infarctions. On average, the annual mortality rate amounts to 12.5 %, and risk of death or rehospitalization to 40 % [1]. Almost 5 % of hospital admissions are due to heart failure, and about 2 % of health care costs in industrialized countries are caused by heart failure, mainly driven by hospital admissions [1]. Optimal, guideline-recommended management of heart failure may reduce adverse events and possibly the economic burden of this challenging disease.

Current treatment strategies are based on randomized clinical trials (RCTs). It is still unclear, whether these results can be translated into real-world situations with unselected, multimorbid patients. While RCTs evaluate the efficacy of a certain treatment modality under “experimental” conditions, registries evaluate the clinical effectiveness, i.e., the transferability and the transfer of RCT-generated data into clinical practice. Therefore, information from heart failure patients managed in real life is another important element of evidence complementing data derived from RCTs.

This assessment can be achieved by a registry or a heart failure program. Registries, e.g., the ESC-HF-Pilot survey, the Euro Heart Failure Survey, the Swedish Heart Failure Registry, or the Italian Heart Failure Registry, gather and document information from a real life heart failure population [2–7]. Programs, e.g., Adhere, Optimize-HF, Improve-HF, HF-Action, intend to optimize a given situation by specific, pre-defined interventions [8–11]. These registries and programs, however, have certain shortcomings like selective instead of consecutive patient inclusion, heterogeneous populations with de novo or worsening heart failure, inclusion of both, systolic and diastolic heart failure, limited parameters checked, and mostly without or only very short-term follow-up.

The EVIdence based TreAtment in Heart Failure (EVITA-HF) registry was created to focus on a well-defined patient population covered by clear guideline recommendations, well-defined tertiary care heart centers, consecutive patient inclusion, a comprehensive assessment and a 1-year follow-up period. EVITA-HF evaluates demography, comorbidities, diagnostic and therapeutic measures, quality of life, adverse events during an index hospital stay or an index outpatient visit, and at a 1-year follow-up of patients treated for chronic heart failure and an ejection fraction below 40 %.

Here, we report on the characteristics, management modalities and follow-up of the hospitalized patients (n = 1,853) included in the first evaluation cohort (phase I cohort, January 2009 until June 2011) of EVITA-HF.

Methods

EVITA-HF was set up as a web-based registry using an electronic case report form (eCRF). Inclusion criteria were chronic heart failure since at least 3 months, and a documented ejection fraction of 40 % or less. Exclusion criteria were age younger than 18 years or missing consent of the patient. Patients were hospitalized (index hospitalization) in 1 of the 16 participating tertiary care heart centers. All centers offer the whole spectrum of diagnostic and treatment modalities in heart failure. Patients had to be included consecutively. Data management was performed at the Institut für Herzinfarktforschung Ludwigshafen at the University of Heidelberg, Germany. The registry received approval by the institutional ethics committees. Each patient gave an informed written consent. The registry was supported by unrestricted grants from Medtronic, Novartis and Sanofi Aventis.

The eCRF collected baseline information on demographics, presentation, medical history, clinical evaluation and diagnostics, pharmacological treatment and non-pharmacological interventions, quality of life, and adverse events during index hospitalization. In a subset of participating centers, a representative one-year follow-up was performed by phone call and/or contact by the center or general practitioner. Thereby, information was obtained on vital status, adverse events and interventions since index discharge, current health status, pharmacological treatment, and quality of life. The follow-up information of survivors was regarded as 1-year status if obtained between 300 and 450 days after index discharge.

EVITA-HF was started in January 2009. As of June 2011, 1,853 hospitalized patients had been included. The 1-year follow-up as of February 2013 included 1,410 patients.

Statistical analysis

The patient population is described by absolute numbers and percentages with respect to categorical variables and medians with quartiles for continuous variables. The distribution of nominal categorical variables was compared between patient groups by Pearson Chi- square test and Mann–Whitney test, as appropriate. Prescription rates of medication and other binary states at discharge vs. admission and at 1-year FU versus discharge are shown for patients with documented information on both occasions and compared by McNemar’s test.

One-year survival and event-free survival after index discharge were calculated by the product-limit method and demonstrated by Kaplan–Meier curves.

All tests were performed two-sided and p values ≤0.05 were considered significant. The computations were performed using the SAS system (release 9.3, SAS Institute, Inc., Cary, North Carolina).

Results

Baseline demographical data and comorbidities (Table 1)

Median age was 70 years; 24 % of patients were female. The two most frequent etiologies for heart failure were ischemic cardiomyopathy (CMP) in 56 % and dilated CMP in 27 %. A former hospitalization for heart failure was known in 65 % of pts, the first hospitalization being 4 years ago on median. Almost 40 % of pts had a history of myocardial infarction and of revascularization. An implanted device was present in 24 % of pts, in 18 % an ICD or CRT-D. History of hypertension was present in 76 % of pts, diabetes mellitus in 39 %, history of chronic kidney disease in 33 % (creatinine >1.5 mg/dl in 26 % of pts), COPD in 15 %, thyroid dysfunction in 12 %, peripheral artery disease in 10 %, h/o stroke in 9 %, active malignoma in 7 %. Reason for admission was non-elective (non-scheduled) in 68 %, and elective (scheduled) in 32 %. In-hospital mortality amounted to 2.1 %, median length of stay was 9 days (IQR 5–15).

Table 1 Baseline demographics and comorbidities of 1,853 EVITA-HF patients
Full size table

Clinical and technical findings at index presentation (Table 2)

Median BMI amounted to 27.1 kg/m2. Blood pressure amounted to 122/74 mmHg, and heart rate to 80 bpm. NYHA classes were 8 % class I, 29 % in class II, 45 % in class III, and 18 % in class IV. Heart rhythm was sinus in 64 %, atrial fibrillation or flutter in 28 %, and pacemaker stimulation in 11 %. Median QRS width was 110 ms, left bundle branch block was found in 26 % of pts. Median ejection fraction amounted to 30 %. Mitral regurgitation was present in 53 %, pulmonary hypertension in 30 % of pts.

Table 2 Clinical and technical findings at presentation
Full size table

Laboratory data (Table 3)

Median values for hemoglobin (13.4 g/dl), creatinine (1.2 mg/dl), GFR (60 ml/min, MDRD formula), sodium (139 mmol/l), potassium (4.2 mmol/l) and cholesterol (169 mg/dl) were in the normal or near-normal range. Median C-reactive protein was elevated with 1.1 mg/dl. NT-proBNP (measured in 52 % of pts) amounted to 4,330 pg/ml.

Table 3 Laboratory data
Full size table

Medication (Table 4)

At admission, pts were treated with ACE inhibitors (ACEI) or angiotensin receptor blockers (ARB), β-blockers and mineral corticosteroid receptor antagonists (MRA) in 73, 71 and 32 %, respectively. Diuretics were taken by 68 % of pts. Treatment with digitalis was found in 16 % of pts, amiodarone in 9 %. Ivabradin was used very rarely (1 %). Half of the pts were taking aspirin, 15 % ADP-antagonists, and 24 % were anticoagulated. Statins were used in 49 %, oral antidiabetics in 17 %, insulin in 14 %, and antidepressant drugs in 7 %. The discharge medication differed significantly from the medication at admission. The number of pts treated with ACEI/ARB, β-blockers, mineral corticosteroid receptor antagonists and diuretics increased significantly, Table 4. At discharge, 90 % of pts received ACEI/ARB, 90 % β-blockers, 58 % MRA, and 84 % diuretics (p < 0.001 vs. admission for each). The use of digitalis, amiodarone, aspirin, ADP-antagonists, oral anticoagulants, and statins also increased significantly (p < 0.001).

Table 4 Medication and device status at admission and discharge in patients discharged alive
Full size table

Non-pharmacological treatment (Fig. 1)

In 768 of 1,843 pts (42 %), 940 specific, i.e., non-pharmacological procedures were performed: 369 (48 % of the 768 pts) revascularization procedures (PCI 82 %, CABG 18 %), 297 (39 % of pts) device therapies (ICD 62 %, CRT-D 28 %, CRT-P 3 %, pacemaker 7 %), 111 (14 % of pts) cardioversions due to atrial fibrillation or flutter, 36 (5 % of pts) ablation procedures, 68 (9 %) valvular procedures, 14 (1.8 %) IABP or implantation of an assist device, and 45 (6 %) iv inotropes. The device status changed significantly from admission to discharge. At admission, 18 % of hospitalized pts had an ICD or CRT-D implanted, at discharge 33 % (p < 0.001; Table 4).

Fig. 1
figure 1

In 42 % of patients (768/1,843), 940 specific procedures (1.22 per patient) beyond oral heart failure medication were performed during the index hospitalization. N absolute number of procedures,  % refers to the 768 patients

Full size image

One-year follow-up (Tables 5, 6; Fig. 2)

One-year mortality was 16.8 %, death or rehospitalization amounted to 56 %. Death, myocardial infarction or stroke occurred in 18 %. The NYHA status in survivors was significantly improved at follow-up compared to index admission: 30 % in NYHA class III or IV vs. 58 % at baseline (p < 0.0001). During 1-year follow-up, 2.5 % of pts underwent PCI, 1.5 % bypass surgery, 2.4 % valvular surgery or intervention, 1.6 % cardioversion, and 1.7 % ablation procedures. Compared to baseline, general health status, as assessed by the visual analogue scale, was improved in 45 % and unchanged in 36 % of subjects. 42 % had an ICD or CRT device implanted at 1-year follow-up. 6 % of pts had received an ICD or CRT implantation during the 1-year follow-up.

Table 5 Demographics and findings in patients with complete 1-year follow-up (FU)
Full size table
Table 6 Medication at 1-year follow-up
Full size table
Fig. 2
figure 2

Kaplan–Meier curves for all cause death, MACCE (death, myocardial infarction, stroke), and combined endpoint of death or rehospitalization during the 1-year follow-up

Full size image

Adherence to medication after 1 year was significantly lower compared to discharge, but remained high with 85 % (−7 % compared to discharge) of patients taking ACEI/ARB, 86 % (−6 %) β-blockers, 47 % (−13 %) mineral corticosteroid receptor antagonist, and 78 % (−6 %) diuretics (p < 0.001 vs. baseline for all, Table 6).

Discussion

The EVITA-HF registry coherently reflects the contemporary real-world management of chronic heart failure patients with reduced EF treated in tertiary care heart center facilities. The EVITA-HF sample is typical for heart failure, representing an elderly and multimorbid population. During hospital stay these patients received a remarkable optimization of the prognosis-relevant medication. Revascularization and device therapy were the leading specific non-pharmacological procedures performed. At discharge, guideline-recommended drug- and device-therapy was significantly improved. The 1-year follow-up showed a moderate mortality rate, a high drug adherence and a sustained improvement of the NYHA status.

Heart failure registries

There are numerous observational heart failure registries or surveys [2–7, 12–21], heart failure programs [8, 11, 22, 23, 24, 25] and large administrative data sets [26] with published results. However, they differ in the populations included and the clinical situations studied, making comparisons difficult. Most registries included chronic (outpatient) and acute (hospitalized) heart failure patients with both, de novo or worsening heart failure [2, 3, 6, 7, 14]. All except one [11] registry or program included systolic and diastolic heart failure patients [2–8, 12–19, 22, 23, 24]. Thus, a heterogeneous spectrum of heart failure patients is represented in these registries. Follow-up data are sparse and limited to mortality and hospitalization, thus lacking information about NYHA status, drug adherence, or device status [3–5, 9, 23]. The EVITA-HF registry aimed at addressing a very homogeneous, well-defined group of heart failure patients, for which unequivocal guideline recommendations are available, i.e., chronic systolic heart failure with an ejection fraction of 40 % or less. This allows to investigate how institutions may close the gap between guideline recommendations and the real-world setting.

Baseline demographics and clinical findings

Baseline demographics in EVITA-HF reveal a median age of 70 years in hospitalized patients, which is comparable to the average age of 69–73 years reported by other registries [2–9], but differs markedly from the median age of 60–65 years in seven heart failure, landmark randomized controlled trials (RCTs) [27–34]. Despite different inclusion criteria, other heart failure registries revealed similar proportions as those of EVITA-HF regarding ischemic etiology, comorbidities and h/o atrial fibrillation [2–4, 6–9]. Multimorbidity in heart failure patients included in registries is much more prevalent than in RCTs, owing to the higher age and the less restrictive exclusion criteria. Admission may be driven by heart failure, or by another condition (e.g., ischemia, atrial fibrillation or diabetes) with HF being a “comorbidity” (as in Euro Heart Failure Survey, [4, 5]). In the scenario of multimorbidity, heart failure is one condition only among a variety of others determining symptoms, mortality and rehospitalization. In addition, with an increasing number of accompanying diseases, more contraindications to recommended therapies have to be observed (e.g. renal dysfunction and MRA, very old age or frailty and ICD), as well as more drug interactions. Thus, the penetrance of guideline recommendations in real life has other potential limits than suboptimal patient management solely.

SBP in our population was well controlled and considerably lower than in other registries (122 mmHg vs. 129–144 mmHg, [2, 3, 7, 8, 22]), probably explained by the higher drug adherence. Whether this translates into a better outcome is debatable recognizing the U-shaped distribution between systolic blood pressure and mortality [35]. However, in the study of Lee et al., the quintile of systolic heart failure patients (ejection fraction of less than 40 %) with the lowest mortality had systolic blood pressure values between 120 and 139 mmHg [35]. Current guidelines recommend a target heart rate of less than 70 bpm at rest, as higher heart rate is a risk factor in heart failure with reduced EF [36]. Although in EVITA-HF, median heart rate was lower than in other registries (80 vs. 82–89 bpm, [2, 3, 7, 22]), only 27 % of our patients had a heart rate of 70 bpm or less. Thus, heart rate seems to be an under-recognized target for improvement. This should be achieved by either uptitrating the dosage of β-blocker, or adding ivabradin in patients with sinus rhythm [1, 33, 36].

Medication on admission and discharge

At discharge, the percentage of pts treated with ACEI or ARB/BB/MRA was increased highly significant in EVITA-HF to 90 %/90 %/58 %, which is clearly superior to five other registries reporting data from pts with reduced EF. In these populations 68–86 %/41–87 %/21–38 % received ACEI or ARB/BB/MRA, respectively at discharge [5, 6, 8, 15, 22]. The most recent ESC-HF-Pilot registry reports a prevalence for the discharge medication of 78 %/81 %/54 % for the three drug classes, respectively [2, 3]. 20–30 % of patients reveal renal dysfunction as a relative or absolute contraindication for use of MRA. To which extent non-use is due to contraindications or side effects, or due to suboptimal transfer of recommendations remains an unresolved issue and requires further analysis. Use of diuretics increases considerably during hospital stay in all registries and amounts to 82–95 % of pts at discharge [2, 3, 5–8]. Prescription of digitalis is rare in EVITA-HF compared to other registries with 20 vs. 30–43 % [5, 6, 8, 22]. Data for the use of amiodarone are missing in nearly all registries, which is noteworthy, since 14 % of our pts were treated with this drug. Anticoagulation is not well documented in HF registries. For the Adhere International Registry, a considerable underuse of anticoagulation in HF pts with atrial fibrillation was reported [38]. The 30 % rate of anticoagulation in our registry corresponds very well to the 34 % rate of pts with a history of atrial fibrillation or flutter.

A very high prevalence of prognostically relevant drugs at discharge is documented in EVITA-HF. This may explain the very high follow-up drug adherence at 1 year. For ACEI/ARB and for BB the rate of discontinuation over 1 year is 7 and 6 % only, indicating a very stringent compliance of the pts. For MRA only a discontinuation rate of 13 % is noticed. The reasons denote further evaluation. It may reflect the fact that MRA therapy is more complex, most probably due to side effects or unstable renal function leading to discontinuation of the drug. There are reports doubting about a net clinical benefit of MRAs in real life HF pts [14, 37].

Specific treatment procedures during hospital stay

A unique feature of EVITA-HF is the evaluation of specific, non-pharmacological treatment modalities during the hospital stay. In nearly half of the pts more than recompensation or optimization of medical treatment was performed. Leading procedures are revascularization (in most cases PCI) and device implantation in 48 and 39 % of patients receiving more than drug treatment (20 and 16 % of the whole cohort). An EF of 35 % or less was present in 75 % of our pts. Primarily these patients are potential candidates for ICD implantation according to current guidelines [1]. About 60 % of non-ICD patients in our population did not have an indication according to the guidelines, or had a contraindication such as terminal heart failure or frailty.

The high number of specific procedures performed during one index hospital stay reflects the modern treatment modalities in heart failure, focusing more and more on intervention- and device-based therapies in patients receiving near optimal pharmacotherapy already. The indication for an ICD or CRT-D, ablation procedures or a valvular intervention may be found and for the first time discussed during the index admission, but performed later on. This is reflected by the increasing number of pts with an ICD or CRT-D from admission (18 %) to discharge (33 %) to the 1-year follow-up (42 %), and another 10 % of pts receiving revascularization, valvular or electrophysiological procedures between discharge and 1-year follow-up. The penetrance for guideline-recommended ICD therapy in published registries is far from being optimal. In the ESC-HF Pilot survey 33 % of potential candidates did have an ICD [2, 3]. Twenty-one percent of pts hospitalized for worsening heart failure in the IN-HF registry are carrier of an ICD, and 12 % in S-HFR patients with an EF of less than 40 % [6, 7]. Of note, only 4 % of patients included in the randomized, controlled SHIFT study (inclusion criterion EF 35 % or less) had an ICD or CRT device [33]. Given the prevalence of 23-36 % of a LBBB in our registry and others [2, 3, 6] CRT or CRT-D is even more underutilized as a guideline-recommended device therapy and needs improvement.

One-year follow-up

Detailed 1-year follow-up data beyond mortality and morbidity is another unique feature of EVITA-HF. 45 % of pts report an improved general health status compared to baseline, only 19 % deteriorated. The NYHA status is persistently improved with 30 % of pts only remaining in NYHA III or IV, compared to 58 % at index admission. Most probably the high drug compliance is at least one reason for the lasting stabilization of this population. Nevertheless, the rate of rehospitalization is 45 % at 1 year. This is comparable to the 44 %in the ESC-HF-Pilot survey [2, 3]. Other reported rehospitalization rates are higher with 25–36 % in a 4–12 week follow-up [5, 9, 24]. Almost 17 % of our patients received a specific procedure during follow-up (7 % device implantation, 9.8 % non-device procedure, e.g. PCI or CABG, Table 5). The specific reasons for rehospitalization in the other 28 % of our population, e.g. decompensation, comorbidities or independent causes require a further detailed analysis. In-hospital mortality and mortality rates during follow-up are available for several registries. In-hospital mortality of 2.1 % in EVITA-HF compares favourably with 2.5–4.1 % in the OPTIMIZE-HF population with reduced EF [9], 3.2 % in Adhere [8], and 3.3 % in the US-National Inpatient Sample 2009 [26]. The Euro Heart Failure Survey reported in 2003 a 3-month mortality of 13 % [5]. The 1-year mortality in EVITA-HF amounts to 16.8 %. In ESC-HF-Pilot the 1-year mortality for hospitalized patients with reduced EF was 18.6 % [2]. In S-HFR the 1-year mortality in 18.852 patients (both hospitalized and outpatients) with an EF of 40 % or less and a median age of 71 years amounted to 17 % [14]. Independent from the specific causes of mortality, which will be analyzed separately, a 1-year mortality of 16–19 % in an elderly population suffering from moderate to severe systolic heart failure is most probably what we can obtain realistically in a real-world situation even with optimal medical and non-medical therapy.

Limitations

We restricted our inclusion criteria to chronic systolic heart failure, thus excluding de novo heart failure, as well as heart failure with preserved EF. Data are restricted to tertiary care hospitals, and may differ from those of facilities not able to provide the full spectrum of management options. Several issues will be addressed in ongoing analyses of the EVITA-HF registry, e.g. subgroup evaluation for age, sex, renal function, and type of admission. In addition, detailed evaluation of the dosage of prognostically relevant medications and reasons for limited ICD- and CRT-penetrance will be analyzed.

Summary

In conclusion, the EVITA-HF registry confirms clinical epidemiology of existing heart failure registries. Importantly, it adds valuable new information regarding effectiveness of in-hospital management and guideline adherence and offers a detailed 1-year follow-up in patients with chronic systolic heart failure. EVITA-HF is a comprehensive data base evaluating the quality of care in real life for this well-defined population with unequivocal guideline recommendations.