OUP user menu

★ editor's choice ★

EuroHeart Failure Survey II (EHFS II): a survey on hospitalized acute heart failure patients: description of population

Markku S. Nieminen, Dirk Brutsaert, Kenneth Dickstein, Helmut Drexler, Ferenc Follath, Veli-Pekka Harjola, Matthias Hochadel, Michel Komajda, Johan Lassus, Jose Luis Lopez-Sendon, Piotr Ponikowski, Luigi Tavazzi
DOI: http://dx.doi.org/10.1093/eurheartj/ehl193 2725-2736 First published online: 25 September 2006


Aims The objective of the EuroHeart Failure Survey II (EHFS II) was to assess patient characteristics, aetiology, treatment, and outcome of acute heart failure (AHF) in Europe in relation to the guidelines on the diagnosis and treatment of AHF published by the European Society of Cardiology.

Methods and results Patients hospitalized for AHF were recruited by 133 centres in 30 European countries. Three thousand five hundred and eighty patients were entered into the database by the end of August 2005. Mean age was 70 years, and 61% of patients were male. New-onset AHF (de novo AHF) was diagnosed in 37%, of which 42% was due to acute coronary syndromes (ACS). Clinical classification according to the guidelines divided AHF patients into (i) decompensated HF (65%), (ii) pulmonary oedema (16%), (iii) HF and hypertension (11%), (iv) cardiogenic shock (4%), and (v) right HF (3%). Coronary heart disease, hypertension, and atrial fibrillation were the most common underlying conditions. Arrhythmias, valvular dysfunction, and ACS were each present as precipitating factor in one-third of cases. Preserved left ventricular ejection fraction (≥45%) was observed in 34%. Valvular disorders were common, especially mitral regurgitation (MR) which was reported on echocardiography in 80% of patients. Median length of stay was 9 days, and in-hospital mortality 6.7%. At discharge, 80% of patients were on angiotensin-converting enzyme-inhibitors or angiotensin receptor blockers, whereas 61% were taking beta-blocker medication.

Conclusion Decompensated HF is the most common clinical presentation of AHF patients. More than one-third of AHF patients do not have a previous history of HF, and new-onset HF is often caused by ACS. Preserved systolic function is found in a substantial proportion of the patients. The prevalence of valvular dysfunction is strikingly high and contributes to the clinical presentation. The EHFS II on AHF verified that the use of evidence-based HF medication was well adopted to clinical practice.

  • Acute heart failure
  • EuroHeart Survey
  • Echocardiography
  • Demographics
  • Treatment
See page 2619 for the editorial comment on this article (doi:10.1093/eurheartj/ehl332)


Heart failure (HF) is one of the most important causes of morbidity and mortality in the industrialized world.1 The prevalence of symptomatic HF is estimated to range from 0.4 to 2.0% in general European population.2 The incidence increases rapidly with age, and in Europe, the mean age of HF population is 74 years.36 Although the number of deaths due to HF has risen generally with the ageing of populations, there seems to be a trend towards improvement in survival, more clearly among men.710

Characteristics, clinical presentation, treatment, and outcomes of HF patients in the acute decompensated phase have not been adequately described, in part because a clear definition of acute HF (AHF) has been lacking. The EuroHeart Failure Survey I (EHFS I) with 11 327 patients described the demographics of acutely hospitalized HF patients as well as those in hospital with possible HF.11,12 The ADHERE registry has data on over 100 000 hospitalizations for AHF from the USA.13 In-hospital mortality was 4 and 7%, in ADHERE and EHFS I, respectively.

The primary objective of EuroHeart Failure Survey II (EHFS II) was to assess characteristics and management of AHF patients across Europe in relation to the guidelines on the diagnosis and treatment of AHF published by the European Society of Cardiology (ESC).14 Specific issues addressed in this second EuroHeart Failure Survey were to evaluate: (i) aetiology and precipitating factors of AHF in Europe in general, (ii) patient characteristics for acute new-onset of HF (de novo AHF) and for acute decompensated chronic HF (ADCHF), and (iii) for clinical groups classified according to ESC AHF guidelines.14 Furthermore, diagnostic processes and therapies including devices were evaluated. Length of stay (LOS) and in-hospital mortality were registered.


The patients were screened at the emergency area, including cardiac care unit (CCU) and intensive care unit (ICU), as well as on ward facilities (internal medicine and cardiology) of the 133 participating hospitals [university hospitals (47%), community or district hospitals (49%), and private clinics (4%)] in 30 European countries. The centres were invited by the ESC EuroHeart network. The participating centres were voluntary and a minimum of 20 patients per centre was expected to be collected. The patient recruitment for this registry and data collection were performed from 21 October 2004 until 31 August 2005. Data collection officers of the participating centres had access to the electronic case report form (CRF) through login on the EuroHeart Survey website for online data entry. The statistical analyses were carried out in agreement with the Helsinki study centre, Finland, and EuroHeart Survey organization in the Biostatistical centre in Ludwigshafen, Germany.

Inclusion criteria

EHFS II recruited patients admitted to hospital (emergency area, internal medicine/cardiology wards, CCU, or ICU) with dyspnoea and verification of HF (new-onset AHF or ADCHF) based on (i) symptoms (dyspnoea) and signs (i.e. rales, hypotension, hypoperfusion, right ventricular HF) of HF and (ii) lung congestion on chest X-ray.

Classification of included patients

The patients were classified by the attending physician according to the guidelines on AHF by the ESC.14 The class of high output HF was not recorded by the EHFS II. ADCHF was defined as worsening of HF in patients with a previous diagnosis or hospitalization for HF and new-onset AHF (de novo AHF) as AHF in patients with no prior history of HF. The severity and pattern of the AHF patients were assessed and classified as follows.

  1. Decompensated HF: dyspnoea or tachycardia and pulmonary congestion or interstitial oedema verified by chest X-ray

  2. Pulmonary oedema: HF accompanied by alveolar oedema in the chest X-ray or with O2 saturation <90% (without supplemental oxygen)

  3. Cardiogenic shock: AHF accompanied by low blood pressure (SBP<90 mmHg) and oliguria (<0.5 mL/kg/h for at least 6 h) or low cardiac index (<2.2 L/min/m2).

  4. HF and hypertension: high blood pressure (>180/100 mmHg) accompanied by symptoms of HF (dyspnoea and tachycardia) and radiological findings of pulmonary congestion or oedema and with preserved left ventricular (LV) function at index hospitalization or before.

  5. Right HF: HF due to right-sided pathophysiology with increased jugular venous pressure and liver size and usually accompanied by peripheral oedema as unique or concomitant to left HF.

Clinical history, symptoms and signs, and medication (admission as well as discharge) were recorded. The pre-hospital medical history including previous HF, coronary heart disease (CHD), hypertension, renal failure (defined as any of the following: patient's serum creatinine recurrently >177 µmol/L (>2.0 mg/dL) at present or in the past or patient on dialysis or with a renal transplant), atrial fibrillation, diabetes, and precipitating factors, such as arrhythmias, acute coronary syndromes (ACS), and valvular disorders, were entered into the CRF. The medical management of HF before and during hospitalization was registered. The most recent echocardiography data were collected. Diastolic dysfunction was classified by the investigator as mild, moderate, or severe according to the generally accepted echocardiographic criteria.15

Statistical analysis

The data are presented as absolute numbers and percentages and medians with interquartile range (IQR). For age and ejection fraction (EF), means and standard deviations are given. Subgroups were compared by χ2 test with respect to dichotomous variables, by Cochran-Armitage test with respect to ordinal categorical variables and by Mann–Whitney U test with respect to continuous variables. Changes in the prescription rates of drugs between admission and discharge were assessed by McNemar's test. A significance level of 0.05 was assumed and all P-values are the results of two-tailed tests. The statistical computations were performed using SAS, version 9.1 (Cary, NC, USA).


The EHFS II collected data from 3580 patients of which 3.5% were collected from Northern Europe, 20.4% from Western Europe, 34.4% from Central Europe, and 42.4% from Mediterranean Europe.

Baseline characteristics of de novo AHF vs. ADCHF patients

The baseline characteristics according to previous history of HF of patients enrolled in EHFS II are shown in Table 1. AHF patients were elderly, mean age 69.9 (SD 12.5) years and 61% male. Males were younger, 67.8 (12.4) years, compared with the females 73.1 (12.0) years (P<0.01). Figure 1 shows the age distribution by decade and gender. In all, there were marked differences in baseline characteristics between the CHF patients (63%) who were hospitalized due to worsening HF (ADCHF) and the new-onset AHF (de novo AHF) patients (37%), (Table 1).

Figure 1

Age distribution by gender in EHFS II.

View this table:
Table 1

Underlying diseases and precipitating factors of EHFS II AHF patients

CharacteristicsTotalADCHFDe novo AHFP-value
Number (%)35802251 (62.9%)1329 (37.1%)
Age, mean (SD)69.9 (12.5)69.5 (12.1)70.5 (13.1)<0.01
Male (%)61.363.757.3<0.001
Underlying diseases (%)
 Diabetes mellitus32.834.430.0<0.01
 Atrial fibrillation/flutter38.746.525.4<0.001
 Previous stroke or TIA13.314.711.0<0.01
 Valvular disease34.443.818.5<0.001
 Renal failure16.820.211.0<0.001
 Chronic obstructive pulmonary disease19.321.515.7<0.001
 Pacemaker implanted9.112.04.3<0.001
 Dilated cardiomyopathy19.325.19.5<0.001
Precipitating factors (on admission)
ACS (%)<0.001
  Unstable angina9.19.97.7<0.05
 Arrhythmia (%)32.432.532.2NS
 Valvular cause (%)26.830.320.8<0.001
 Infection (%)17.619.215.0<0.01
 Non-compliance with therapy (%)<0.001
  • P-value for difference between ADCHF and de novo AHF. TIA, transient ischaemic attack. Renal failure defined as any of the following: patient's serum creatinine recurrently >177 µmol/L (>2.0 mg/dL) at present or in the past or patient on dialysis or with renal transplant; anaemia as reported.

Although there were only minor differences in age and gender distribution between ADCHF and de novo AHF patients, concomitant diseases were significantly more common in patients with ADCHF (Table 1). CHD and hypertension were the most frequent underlying diseases and often co-existent. Dilated cardiomyopathy was present in 25% cases of ADCHF patients. Atrial fibrillation or flutter was extremely prevalent (47%) in ADCHF patients (Table 1). The majority (73%) of ADCHF patients was admitted with decompensated HF. Although decompensated HF was the most common presentation in de novo AHF as well, pulmonary oedema (26%) and cardiogenic shock (7%) were significantly more prevalent than that in ADCHF patients (Figure 2).

Figure 2

Distribution of patients by clinical classification of AHF. Inserted table shows the distribution separately in all patients and patients with de novo AHF, as well as ADCHF. ***P<0.0001 between de novo AHF and ADCHF.

Precipitating factors for AHF hospitalization

The cause of hospitalization for HF differed in de novo AHF and ADCHF groups. ACS was the major precipitating factor in patients with de novo AHF, present in 42% of cases, in which the ACS mostly was due to myocardial infarction (MI). Arrhythmia was equally prevalent as precipitating factor in AHF patients, regardless of previous history of HF. Valvular disorders, infections, and non-compliance to medication were more common in ADCHF (Table 1).

Patient characteristics by clinical class

Decompensated HF was the most common clinical presentation, accounting for about two-thirds of cases (Figure 2). Pulmonary oedema and hypertensive HF were present in 16.2 and 11.4% of patients, respectively. Cardiogenic shock was seen in 3.9% and RV HF in only 3.2% at presentation. Differences between clinical classes are summarized in Table 2.

View this table:
Table 2

Baseline characteristics and precipitating factors by clinical classification of EHFS II patients

CharacteristicsTotalDecomp. HFPulmonary oedemaCardiogenic shockHypert. HFRight HF
No. of patients (% of total)35802340 (65.4)581 (16.2)139 (3.9)407 (11.4)113 (3.2)
Age (years), mean (SD)69.9 (12.5)69.7 (12.8)71.2 (11.5)67.3 (12.7)69.8 (11.2)69.6 (13.4)
Age, IQR62.6–78.762.3–78.764.6–79.759.5–77.261.9–78.463.0–79.5
Male (%)61.362.159.467.660.450.4
Body mass index (kg/m2)26.826.526.926.428.026.6
New-onset AHF (%)37.129.759.664.737.339.8
Hospitalization for HF within last 12 months (%)44.548.033.729.345.146.4
Underlying diseases (%)
 Diabetes mellitus32.830.939.434.334.529.2
 Atrial fibrillation/flutter38.741.328.124.637.758.4
 Previous stroke or TIA13.312.415.711.816.013.3
 Valvular disease34.437.526.218.031.743.8
 Renal failure16.816.615.818.118.717.7
 Chronic obstructive pulmonary disease19.319.219.318.118.027.4
 Pacemaker implanted9.110.65.910.84.98.8
 Dilated cardiomyopathy19.321.811.410.220.215.9
Precipitating factors (on admission)
ACS (%)30.224.749.471.924.414.2
  Unstable angina9.18.610.03.614.32.7
Arrhythmia (%)32.432.929.329.734.533.9
 Valvular cause (%)26.830.224.117.412.632.7
 Infection (%)17.618.517.111.815.617.1
 Non-compliance with therapy (%)22.224.616.97.921.918.1
  • TIA, transient ischaemic attack. Renal failure defined as any of the following: patient's serum creatinine recurrently >177 µmol/L (>2.0 mg/dL) at present or in the past or patient on dialysis or with renal transplant; anaemia as reported.

In decompensated HF, over two-thirds had a history of HF (ADCHF). Valvular disease and atrial fibrillation or flutter were common as underlying diseases. Important precipitating factors were arrhythmias (33%) and valvular causes (30%). Non-compliance to HF medication was also assessed to cause decompensation of HF in almost 25% in this group.

Pulmonary oedema and cardiogenic shock were mostly precipitated by an acute coronary event (49 and 72%, respectively), often with subsequent MI. ST-elevation MI (STEMI) was the cause of ACS in 55% of cases with cardiogenic shock, which is a significantly higher proportion than in any other clinical class. In contrast, non-STEMI was most often present in the pulmonary oedema group, in which hypertension was very prevalent as underlying disease (70%), and diabetes was also more frequent (Table 2).

Almost all (95%) patients presenting with hypertensive HF had a previous history of hypertension, other underlying diseases being similar to other groups in frequency.

In the right HF class, atrial fibrillation and valvular diseases were prevalent as both precipitating factors and underlying conditions, whereas only 38% had a history of CHD.

Clinical and echocardiographic findings

Overall, the median blood pressure was 135/80 mmHg, IQR 110–160/70–90 mmHg. In hypertensive HF, blood pressure was 170/100 mmHg (IQR 145–200/80–110 mmHg), whereas in cardiogenic shock, it was 90/59 mmHg (IQR 75–109/43–68 mmHg). Heart rates were elevated (median 95 bpm, IQR 77–114 bpm), especially in pulmonary oedema and cardiogenic shock in which the median heart rate measured on admission ECG was 100 bpm. Median respiratory rates were ≥20 breaths/min in all classes and in pulmonary oedema as high as 28 breaths/min.

Echocardiography results were available in the majority of patients enrolled in EHFS II, and of these, 89% (2733/3062) had reported LVEF. Overall mean LVEF was 38%, but de novo AHF patients had slightly higher mean LVEF (42 vs. 36%; P<0.0001) compared with ADCHF patients (Table 3). Especially severely depressed LV function (LVEF<30%) was more common in the ADCHF group (34.6 vs. 21.3%; P<0.0001). In contrast, preserved LVEF≥45% was present in 34.3% of the whole study population (42.8% in de novo AHF vs. 29.6% in ADCHF patients, P<0.0001). Left atrial diameter was enlarged to a median value of 47 mm.

View this table:
Table 3

Echocardiographic findings by previous history of HF and by clinical class

VariableAllADCHFDe novo AHFDecomp HFPulm. oedemaCardiog. shockHypert. HFRight HF
ECHO available (n)306219281134199152512132897
LVEDD (mm), median (IQR)*58 (51–65)60 (53–67)55 (48–60)59 (52–66)56 (50–61)55 (48–61)56 (50–61)50 (43–57)
EF% mean (SD)*38 (15)36 (15)42 (16)37 (15)40 (14)33 (15)44 (13)48 (16)
 LVEF<30 (%)29.934.621.334.623.
 LVEF 30–44 (%)35.835.835.935.738.427.937.326.6
 LVEF≥45 (%)34.329.642.829.738.424.050.763.3
LA (mm), median (IQR)*47 (42–52)48 (43–54)45 (40–50)48 (42–54)45 (40–50)43 (40–47)45 (41–50)47 (39–52)
Diastolic dysfunction (%)**
 None or mild49.547.353.348.251.343.853.654.7
Mitral regurgitation (%)*
Tricuspid regurgitation(%)*
 Severe 7.5 9.8 3.7 8.4 2.4 5.5 5.428.3
  • LVEDD, LV end-diastolic diameter; LA, left atrial (transversal) diameter.

  • *P<0.0001 for difference between de novo AHF and ADCHF groups.

  • **P<0.05.

Valvular disorders were common, as shown in Table 3. MR was extremely frequent and reported in four out of five patients. Although usually mild, even moderate to severe regurgitation was found in 43% in the study population (Figure 3). MR was followed in prevalence by tricuspid valve regurgitation(TR) and aortic valve regurgitation (AR). Stenotic valve diseases were less common, aortic stenosis (AS) of moderate to severe degree present in 9%, with no difference between de novo AHF and ADCHF patients.

Figure 3

Graph of prevalence of valvular dysfunction in AHF patients. The graph shows the prevalence of mild vs. moderate to severe valvular dysfunction in the study population overall (A) and separately in de novo AHF (B) and ADCHF patients (C). AS, aortic valve stenosis; MS, mitral valve stenosis; MR, mitral valve regurgitation.

Looking at the differences between clinical classes, moderate to severe MR was present in nearly half of the patients with decompensated HF and was very frequent in patients with cardiogenic shock and pulmonary oedema as well, 41 and 38%, respectively (Table 3). In right HF, MR was less common, whereas TR was found in 84% of cases in this group and 2/3 of TR were assessed as moderate to severe. Pulmonary oedema was associated with moderate to severe AS in 12% of patients. In hypertensive HF, MR was prevalent but mostly mild, whereas stenotic valve disease was uncommon.

Procedures and treatment

Diagnostic investigations and procedures performed during hospitalization are listed in Table 4. BNP and NT-proBNP measurements were recorded in only 16% of study population and there was great variation in sampling rate by centre (0–92% of patients recruited/centre). Transthoracic or transesophageal echocardiography was performed in 90% of patients, of which 85% during the index hospitalization in a median time of 1 day from admission. The remaining echocardiography data were from the preceding year. Pulmonary artery catheter was used for haemodynamic guidance in 5% overall, but in not more than 25% of patients in the cardiogenic shock group.

View this table:
Table 4

Diagnostic investigations and procedures by clinical class

Procedure % performedTotalDecomp. HFPulmonary oedemaCardiogenic shockHypert. HFRight HF
Chest X-ray97.797.698.894.298.896.5
CT scan4.
EP study1.
Holter ECG12.512.413.
Exercise test4.
Arterial Line8.15.914.634.53.56.2
  • ECHO, echocardiography; MRI, magnetic resonance imaging; EP, electrophysiology; PAC, pulmonary artery catheter.

  • aPerformed during index hospitalization.

  • bPerformed during index hospitalization or within 1 year of admission.

In the acute phase, 93% received diuretic treatment, mostly by intravenous (iv) bolus or continuous infusion (Table 5). Nitrate or nitroprusside iv was used in as many as 71% in the pulmonary oedema group. However, their use in hypertensive HF and cardiogenic shock did not differ from the study population overall. Beta-blockers were given intravenously to 10%, mainly in patients with ACS or atrial fibrillation. Amiodarone was used in 18% of cases. Ventilatory support was given in 14% of all patients. The need for endotracheal intubation and mechanical ventilation was high (37%) in cardiogenic shock patients (Table 5). The most frequently used inotropes were dopamine (11%) and dobutamine (10%), and levosimendan was given to 4%. The use of vasoactive medication, as well as treatment with levosimendan and amiodarone, was highest in cardiogenic shock patients, of whom not more than one-third was treated with an intra-aortic balloon pump (IABP). Angiography and percutaneous coronary intervention (PCI) were more often performed in groups with high prevalence of ACS and/or CHD. Thrombolysis was performed in 20%, PCI in 38%, and coronary artery bypass graft (CABG) in 4% of STEMI patients during the index hospitalization. Revascularization procedures were less frequent in NSTEMI patients.

View this table:
Table 5

Acute cardiac care by clinical class

Treatment % performedTotalDecomp. HFPulmonary oedemaCardiogenic shockHypert. HFRight HF
Ventilatory supporta13.98.131.556.17.414.2
Invasive mechanical ventilation5.12.311.
 Iv bolus72.171.781.958.768.658.4
Iv nitrate37.830.470.636.539.78.6
Iv nitroprusside0.
Iv inotrope
Heparin (UFH)18.717.818.845.715.915.2
Blood transfusion5.95.27.910.13.512.4
  • UFH, unfractionated heparin; LMWH, low-molecular-weight heparin; ICD, implantable cardioverter/defibrillator.

  • aVentilatory support=continuous positive airway pressure/non-invasive positive pressure ventilation/invasive mechanical ventilation (intubated).

Cardiovascular medication

Medications on admission and at discharge are listed in Table 6. There was an increase in use of HF medication during hospitalization. Overall, beta-blockers (43% on admission increased to 61% at discharge), angiotensin-converting enzyme (ACE)-inhibitors (55–71%), and aldosterone antagonists (spironolactone/eplerenone, 28–48%) were more often prescribed to patients at discharge.

View this table:
Table 6

Medication on admission and at discharge by history of HF

Prescription ratesAdmissionDischarge
All n=3580ADCHF n=2251De novo AHF n=1329All n=3338ADCHF n=2118De novo AHF n=1220
Cardiovascular medication
 ACE-inhibitors (ACE-I)55.063.340.8*71.1**72.0**69.5**
 ACE-inhibitor or ARB63.171.648.6*80.2**81.1**78.8**
 Beta-blocker (BB)43.246.537.6*61.4**58.9**65.7*,**
 Oral nitrate27.731.820.6*32.9**35.328.7*,**
 Calcium channel blocker17.816.120.814.6**13.4**16.7
 Other vasodilator4.
 Digitalis compound26.634.413.3*31.0**36.521.4*,**
 Antiarrhythmic drug12.916.37.2*17.6**20.6**12.4*,**
Antithrombotic agents
 Vitamin K antagonist24.029.314.8*33.1**36.9**26.6*,**
Other medication
 Lipid regulating drug28.428.328.541.8**38.4**47*,**
 Other CV medication16.618.313.820.9**22.8**17.7**
 Oral antihyperglycaemic therapy17.017.116.917.317.716.7
  • ARB, angiotensin receptor blocker.

  • *P<0.0001 between ADCHF and de novo AHF.

  • **P<0.0001 between admission and discharge.

There was, as expected, a difference in medication on admission between patients with and without a history of HF (Table 6). At discharge, de novo AHF patients less frequently had diuretics, and an aldosterone antagonist was used in only 36% of patients when compared with 54% in ADCHF patients. ACE-inhibitors and angiotensin receptor blockers were prescribed in similar proportions at discharge. Beta-blockers were more frequently used in de novo AHF patients (66 vs. 59% in ADCHF, P<0.001).

In-hospital mortality and LOS

In-hospital mortality in EHFS II was 6.7%. De novo AHF patients had a higher in-hospital mortality compared with ADCHF patients (8.1 vs. 5.8%, P<0.001) (Table 7 and Figure 4). Among clinical groups, in-hospital mortality was extremely high in cardiogenic shock patients (39.6%). In pulmonary oedema and right HF, prognosis was also worse than average. The best survival was seen in hypertensive HF, as almost all patients were discharged alive.

Figure 4

In-hospital mortality in EHFS II by history of HF and clinical class.

View this table:
Table 7

Outcomes and LOS in EHFS II

VariableAllADCHFDe novo AHFDecomp. HFPulm. oedemaCardiog. shockHypert. HFRight HF
In-hospital mortality (n=)239/3580131/2250108/1329116/234053/58155/1396/4079/113
LOS days, median (IQR)9 (6–14)9 (6–14)9 (5–15)9 (6–15)10 (6–15)10 (4–17)8 (6–12)11 (7–17)
% staying in ICU/CCU51.046.359.044.476.092.741.343.6
ICU/CCU stay days, median (IQR)3 (2–5)3 (2–5)3 (2–5)3 (2–5)3 (2–5)4 (2–8)3 (1–5)3 (2–5)
Life-threatening arrhythmias (%)
  • Median LOS reported as for all patients (including in-hospital deaths). Median LOS in ICU/CCU for patients admitted to these units during index hospitalization.

Median LOS in all patients was 9 days (IQR 6–14). Half of the patients were treated in ICU or CCU, and their median stay in ICU/CCU was 3 days (IQR 2–5). Cardiogenic shock and pulmonary oedema patients needed ICU/CCU treatment in a higher proportion compared with other clinical classes. LOS was similar regardless of previous history of HF, and even in different clinical classes, the differences were surprisingly small (Table 7). Taking into account only the patients discharged alive, LOS increased in the cardiogenic shock group to 13 days (IQR 10–19), whereas LOS in other clinical groups did not change substantially; 9 days (IQR 6–14) for decompensated HF, 10 days (IQR 6–15) for pulmonary oedema, 8 days (IQR 6–12) for hypertensive HF, and 12 days (IQR 7–17) for right HF.


EHFS II is specifically targeted on AHF. In EHFS II, only patients hospitalized due to AHF, either de novo AHF or ADCHF, were included, and they were classified according to the current ESC guidelines on AHF. This is the first time this classification was used systematically and describes the frequency and background of various forms of AHF. High output HF was not recorded as it is rare and a not well-recognized patient group. There are very few other studies or surveys which have analysed the aetiology and management of AHF. The largest registry ADHERE13 contains information from individual hospitalizations, and the registry is collected retrospectively based on discharge diagnosis, probably missing some acute new-onset HF patients, especially patients with ACS may not be included in full. A recent paper described AHF patients treated in cardiology wards in Italy, which had more selected and severe patient population (nearly half with pulmonary oedema) compared with the present study.16

In EHFS II, the mean age was similar to previous surveys and registries of AHF patients both in Europe and in the USA.1618 Co-morbidities were abundant, CHD and hypertension being the most common and often co-existing. Nearly 40% of the patients in EHFS II did not have a previous history of HF, in other words, had de novo AHF. In this group, ACS was a major precipitating factor. A significant proportion of AHF patients had preserved LV function, which is in concordance with previous reports11,13,19 stating again that up to 50% of HF patients have preserved systolic function.

Precipitating factors

Arrhythmias, valvular dysfunction and ACS were each present as precipitating factor in almost one-third of AHF hospitalizations. Infections and non-compliance to medication superimposed on these factors, the latter predominantly in patients with a previous history of HF. Arrhythmias were common in all groups and mostly of atrial origin. Atrial fibrillation has previously been reported in high frequency in AHF patients,13,19,20 and in EHFS II, it played a significant role, both as underlying condition and precipitating factor. Nearly half of the ADCHF patients had a history of atrial fibrillation which was also reflected on the use of anticoagulant therapy.

CHD and ACS in AHF

CHD is the most frequent cause of HF and also seen as concomitant disease in patients with ADCHF. EHFS II further demonstrates that ACS was the most important precipitating factor in patients with de novo AHF. In cardiogenic shock and pulmonary oedema, only 40% had a previous history of HF, but more than half of the patients had an ACS as precipitating factor.

The relative distribution of ACS types is also interesting, as non-STEMI was mostly found in the pulmonary oedema group. In contrast, over half of the patients with cardiogenic shock presented with concomitant STEMI. This is in line with the previous reports that STEMI is the most common form of ACS in cardiogenic shock,21 and underlines the importance and need for invasive treatment. In EuroHeart ACS survey, the proportion of ST-elevation ACS decreased with age.22 Furthermore, in their study, the prevalence of pulmonary oedema was low (<5%) in younger age groups but up to 20% of the octogenarians. Indeed, pulmonary oedema patients were older than cardiogenic shock patients in EHFS II as well. Ischaemia leading to systolic or diastolic dysfunction may precipitate pulmonary oedema in the elderly, whereas STEMI more often leads to output failure and cardiogenic shock. Myocardial ischaemia may also provoke ventricular arrhythmias, as observed particularly in the cardiogenic shock group where about one-third of all arrhythmias were of ventricular origin. Despite higher prevalence of chronic CHD, diabetes, and renal failure in ADCHF patients, this higher clinical risk profile for coronary events did not translate into higher frequency of ACS.

Valvular disorders

Valvular dysfunction was reported as the precipitating factor in almost one-third of patients with ADCHF. The prevalence of valvular disorders reported in this survey is far higher than in any previous HF study. History of haemodynamically significant or symptomatic valvular disease was present in one-third of the patients, but on echocardiography more than four out of five had some degree of disturbance in valvular function. In ADCHF patients, regurgitant valvular disorders were significantly more common than in de novo AHF. However, the finding that 72% of de novo AHF and 84% of ADCHF patients had evidence of MR is most important, especially because the grade of MR was assessed as moderate to severe in over 40% of cases. This is not necessarily due to primary valve disease, but may be caused by degeneration of valvular and cardiac structures with both cardiac remodelling and increasing age. Tricuspid regurgitation was common as well and reflects haemodynamic conditions in AHF. The high prevalence of atrioventricular valve dysfunction is possibly overestimated as it was reported by echocardiography done early during the initial hospitalization in a high percentage of patients. The frequency and severity of this valve dysfunction may diminish as a result of therapy and unloading of the ventricle. AS was most common in patients presenting with pulmonary oedema and may contribute to the clinical presentation. No echocardiography follow-up data were recorded in the present study.

Diversity of AHF patients

The differences observed between groups, de novo AHF and ADCHF on one hand and clinical classes on the other hand, have important implications. First, they emphasize the diversity and heterogeneity of AHF patients. Identification of more distinct entities helps us to direct our treatment efforts more correctly, targeting underlying conditions and precipitating factors. Secondly, they highlight the lack of evidence for treatment of major subclasses of AHF, such as patients with preserved LVEF. More precisely defined AHF patient categories, as suggested in the ESC guidelines on AHF, should be used as a base for systematic clinical research to improve prognosis. Further investigations testing the efficacy of different treatment regimens in pre-defined classes are urgently needed.

Diagnostic procedures

Nearly all patients underwent ECG and chest X-ray. Echocardiography was performed during the initial hospitalization in the majority of patients, especially in de novo AHF, or fairly recent echocardiographic data were available for patient evaluation. In-hospital echocardiography was done in most cases within first days of hospitalization. The rate of echocardiography early during hospitalization is amazingly high and clearly higher than that in EHFS I11 and shows excellent adherence to current HF guidelines.1,14 Mean LVEF was reduced in all clinical categories except hypertensive and right HF. However, preserved LVEF was common in all clinical groups, comprising on average one-third of the overall population. As data were recorded in some patients solely before hospitalization, the proportion of patients with preserved LV function may be overestimated. In contrast, early in-hospital echocardiography may overestimate the prevalence of valvular dysfunction as discussed previously. Nevertheless, diastolic dysfunction of at least moderate degree was present in half of the studied subjects. This should be expected in an elderly population with high prevalence of hypertension. Indeed, abnormal relaxation of the left ventricle plays an important role in precipitating AHF.

Management and mortality of hospitalized AHF patients

In the management of AHF, ventilatory support and inotropic therapy were needed in a large proportion of patients with cardiogenic shock, in which IABP was used in not more than one-third of the patients. It seems, though, that the use of IABP in patients with cardiogenic shock has increased during the past few years.21 The use of vasoactive medication in the cardiogenic shock group exceeded that of other clinical classes taken all together. Invasive diagnostic and therapeutic procedures were more common in patients with cardiogenic shock or pulmonary oedema. Still, arterial line was used only in one-third of the patients and pulmonary artery catheter in one-fourth of the patients in cardiogenic shock. Despite intensive treatment efforts, in-hospital mortality in this category remained very high, as has also been previously reported.16,21

With respect to medication, EHFS II provides important information on the state of implementation of guidelines. HF medication was rather well implemented at discharge in EHFS II patients. Although it is difficult to assess contraindications as possible reasons for not prescribing HF medication, it is still possible to point out trends and evaluate adherence to guidelines. An analysis from EHFS I recently found that underutilization of evidence-based medication is only in part explained by the differences between the general HF population and patients enrolled in clinical trials.23 In contrast, HF patients with preserved systolic function are common, and treatment strategies in this group are less well defined. The use of iv beta-blocking agents in EHFS II was mainly in patients with ACS or atrial fibrillation which is in line with the ESC guidelines.14 Diuretics were widely used as basic therapy as recommended, but the use of vasodilators seemed lower than expected in AHF. In contrast, inotropes were rather frequently used. Medication data will be analysed further and more detailed analyses will evaluate, for example, whether the recommendations for the use of iv therapies in the ESC guidelines on AHF were followed.

EHFS II shows that agents acting on the renin–angiotensin system (i.e. ACE-inhibitors, angiotensin receptor blockers, and aldosterone antagonists) are more frequently used than just a few years earlier, when the first HF survey was conducted.12 Furthermore, prescription of ACE-inhibitors and angiotensin receptor blockers to ADCHF patients increased during hospital stay so that four out of five patients were on either medication at the time of discharge. In de novo AHF patients, treatment with these drugs was implemented as well, and at discharge there was no difference in the use of these drugs compared with ADCHF patients. Beta-blocker use has also increased during the past few years. Still, less than half of the patients with previous history of HF were taking a beta-blocker on admission. Beta-blocker medication was initiated in a significant proportion during hospital stay both in ADCHF and in de novo AHF, with a slightly higher prescription rate at discharge in de novo AHF patients.

In-hospital mortality in EHFS II was similar to the European surveys,12,16 and slightly higher than in reports from the USA.13,17 LOS was on average 2 days shorter compared with EHFS I, although differences in study populations preclude drawing any general conclusions about changes in outcomes and management of AHF patients. Interestingly, no clear differences in LOS were seen between groups of clinical severity. When only patients discharged alive were considered, median LOS in cardiogenic shock increased by 3 days and had the longest in-hospital stay. Other groups were not markedly affected. It is still noteworthy that in both cases LOS in de novo AHF and ADCHF was similar. As expected, patients with pulmonary oedema and cardiogenic shock were admitted to ICU/CCU for treatment more often than other clinical classes, but the LOS in these units was on average short.

Limitations of the study

Surveys based on voluntary participation and recruitment of patients have limitations that have to be acknowledged. First, centres attending to the survey may be biased toward bigger and more active hospitals. Secondly, patient recruitment activity differs between countries and centres, and proportions in the study population do not reflect the general population. The limited number of centres and the fact that not all European countries took part in the survey also restricts the possibility to generalize the results, and the data are representative for the participating centres only. Inclusion criteria aimed at identifying patients with verified AHF, but the final diagnosis was not confirmed centrally. The clinical classification was also assessed locally. Some overlap in classifications may occur, as reporting and adherence to classification criteria are dependent on discretion of the participating investigators. Furthermore, no centralized reading of echocardiographic results was done.


EHFS II provides up-to-date information on demographics, characteristics, and underlying conditions of AHF patients as well as aetiology, investigation, and treatment practices of AHF in Europe. It is the first European survey to include patients admitted primarily for AHF. Patients were classified according to whether or not they had a previously known diagnosis of HF. Furthermore, EHFS II allows comparison between groups based on the clinical classification of the ESC guidelines on the diagnosis and treatment of AHF. Demographics and characteristics are well in line with previous reports and with results from the previous EuroHeart Failure Survey.

ACS, valvular dysfunction, and arrhythmias were the most common precipitating factors, with the dominance of each of these factors being somewhat different depending on patient class. The finding of valvular dysfunction in a large majority of patients is remarkable. Adherence to treatment guidelines and the use of HF medication have improved since the first survey on HF in Europe. Outcomes are similar to other reports, the in-hospital mortality being nearly identical to the previous survey. The EHFS II provides comprehensive information for further in-detail analyses on AHF, and follow-up data are currently being collected.


EuroHeart Failure Survey II was endorsed by the Heart Failure Association of the European Society of Cardiology, former Working Group of Heart Failure.

Conflict of interest: none declared.

Appendix: organization of the survey

Heart Failure II Expert Committee: M. Nieminen (Survey Chairman) and V.-P. Harjola (Research Fellow), Finland; D. Brutsaert, Belgium; M. Komajda, France; H. Drexler, Germany; L. Tavazzi, Italy; K. Dickstein, Norway; Piotr Ponikowski, Poland; Jose Luis Lopez-Sendon, Spain; Ferenc Follath, Switzerland.

EuroHeart Survey Team (European Heart House, France): Malika Manini (Operations Manager), Claire Bramley (Data Monitor), Valerie Laforest (Data Monitor), Charles Taylor (Database Administrator).

Statistical analysis centre (Ludwigshafen am Rhein, Germany): M. Hochadel (Statistician).

National Coordinators: Austria, Kurt Huber; Belgium, Guy De Backer; Bulgaria, Vera Sirakova; Czech Republic, Roman Cerbak; Denmark, Per Thayssen; Egypt, Osama Abdel Aziz, Khalid Tammam; Finland, Seppo Lehto; France, François Delahaye; Georgia, Bondo Kobulia; Germany, Uwe Zeymer; Greece, Dennis Cokkinos, Dimitrios Krematisnos; Hungary, Kristof Karlocai; Ireland, Emer Shelley; Israel, Shlomo Behar; Italy, Aldo Maggioni; Lithuania, Virginija Grabauskiene; The Netherlands, Jaap Deckers; Norway, Inger Asmussen; Poland, Janina Stepinska; Portugal, Lino Gonçalves; Russia, Vyacheslav Mareev; Serbia and Montenegro, Zonara Vasilijevic; Slovakia, Igor Riecansky; Slovenia, Miran F. Kenda; Spain, José Luis Lopez-Sendon; Sweden, Annika Rosengren; Switzerland, Peter Buser; Turkey, Tugrul Okay; Ukraine, Oleg Sychov; UK, Peter Schofield.

There was no national coordinator in participating countries which are not mentioned in the above list.

EuroHeart Survey Board Committee: Anselm Gitt (Chairman), Germany; L. Tavazzi, Italy; R. Seabra Gomes, Portugal; J. Marrugat de la Iglesia, Spain; L. Wallentin, Sweden; P. Kearney, Ireland; R. Boyle, UK; M.L. Simoons (Ad hoc Consultant), The Netherlands.

List of Industry Sponsors: Main Sponsors: Abbott Laboratories, GlaxoSmithKline Services Unlimited; Roche Diagnostics GmbH; Sanofi-Synthelabo Groupe.

List of Institutions: Czech Society of Cardiology, Fédération Française de Cardiologie; Hellenic Cardiological Society; Italian Federation of Cardiology; Portuguese Society of Cardiology; Spanish Society of Cardiology; The Netherlands Heart Foundation.

Participating Centres, Investigators, and Data Collection Officers

Armenia: Liana Tumasyan, Armen Nargizyan, Arman Astvatsatryan, Yerevan; Kristine Margaryan, Alina Ohanyan, Yerevan. Austria: Kurt Huber, Gabriele Jakl, Vienna. Belgium: Dirk Brutsaert, Carl Convens, Antwerpen. Bulgaria: Stefan Naydenov Naydenov, Sofia; Assen Rachev Goudev, Kalina Koleva, Sofia; Valentin Petrov, Asen Petrov, Shoumen; Snejana Tisheva, Virginia Jordanova, Nadia Stancheva, Pleven; Vera Sirakova, Yavor Peychev, Varna. Croatia: Duska Glavas, Split; Vjekoslava Raos, Irena Ivanac, Zagreb. Czech Republic: Jiri Vitovec, Hana Pavelcikova, Brno; J. Kautzner, Ivan Malek, Ruzena Jandova, Prague; Petr Zajicek, Valasske Mezirici; Hana Rosolova, Hess Zdenek, Pilsen; Jindrich Spinar, Milan Sepsi, Brno. Egypt: Khalid Ahmed Al Khashab, Fayoum; Aly Saad, Zagazig; Heba Farouk, Cairo. Estonia: Julia Reinmets, Toomas Marandi, Tallinn. Finland: Seppo Lehto, Kirsti Savolainen, Kuopio; Markku Nieminen, Veli-Pekka Harjola, Mervi Pietila, Helsinki. France: Jean Marco, Frederic Petit, Toulouse; Patrick Jourdain, Pontoise; Nicolas Danchin, Sylvie Marinier, Paris; Francois Delahaye, Armelle Delahaye, Lyon; Daniel Galley, Christine Beltra, Albi; Claude Gully, La Roche sur Yon; Michel Komajda, Celine Fratini-Lefol, Paris; Jean-Pierre Bassand, Marie-France Seronde, Denis Pales Espinosa, Besancon. Georgia: Konstantine Nodar Liluashvili, Tbilisi; Gulnara Tabidze, Irina Jashi, Levan Tvildiani, Tbilisi. Germany: Ralph Winkler, Ludwigshafen am Rhein; Thomas Rehm, Duisburg. Greece: Antonios Dimopoulos, G. Athanassopoulos, G. Giamouzis, Natassa Tsiavou, Athens; Athanasios Trikas, C. Stefanadis, Andreas Synetos, Athens; Apostolos Karavidas, Ioannis Fotiadis, Andreas Kaoukis, Dimitrios A. Korres, Athens; Ioannis Vogiatzis, Veria; Alexandros Gotsis, Stella Ntourtsiou, Paraskevi Bozia, Komotini; John Nanas, Stavros Drakos, Athens; Gerasimos Filippatos, Virginia Markou, Athens; Dimitra Kontogianni, D. Babalis, Charalambos Grassos, G. Goranitou, Athens; Alexandros Douras, Lambros Papakosmas, Athanasios Koutroubas, Themistoklis Tsaknakis, Volos; Petros Stravopodis, Antonios Kassimatis, Zakynthos; Efstathia Prappa, John Kandylas, Manolis Paximadakis, Athens; Stavros Kakouros, Margetis Panagiotis, Athens; John Parissis, Katerina Fountoulaki, Dimitrios Farmakis, Athens; P.E. Vardas, Alexander Patrianakos, Heraklion; Ioannis Skoularigis, Nicolaos Prattis, Larisa; M. Zairis, Stefanos Foussas, Nikolaos Patsourakos, Pireus. Hungary: Istvan Preda, Eva Csoti, Budapest; Andrea Badics, Szekesfehervar. Israel: Menashe Epstein, Rehovot. Italy: G. Rosano, Rome; Lorenza Robiglio, Antonio Pesola, Lido di Camaiore; Franco Naccarrella, Giovannina Lepera, Stefano Sdringola Maranga, Bologna; Francesco Gentile, Chiara Benchini, Cinisello Balsamo; Annamaria Nicoletti, Emidio Feraco, Silvio Vena, Cosenza; Pietro Agricola, Crema; Alberto De Bernardi, Maria Teresa Spinnler, Moncalieri; Rita Mariotti, Pisa; Luigi Tavazzi, Carlo Campana, Pavia; Luigi Tarantini, Belluno. Lithuania: Ausra Kavoliuniene, Vaida Mizariene, Kaunas; Virginija Grabauskiene, Ruta Jasaityte, Vilnius; Vitas Vysniauskas, D. Petraskiene, Marijampole; Rasa Raugaliene, Kaunas. Moldova: Eleonora Vataman, Victor Priscu, Kishinau. Poland: Michal Marchel, Grzegorz Opolski, Warszawa; Piotr Ponikowski, Pawel Siwolowski, Wroclaw; Wladyslaw Sinkiewicz, Piotr Sobanski, Bydgoszcz; Zbigniew Gasior, Daniel Jakubowski, Katowice; Beata Malczewska, Nowy Dwor Mazowiecki. Portugal: Ana Oliveira Soares, Rafael Ferreira, Joao Ribeiro, Amadora; Marisa Trabulo, Carnaxide; Pedro Costa Ferreira, Viseu; Aurora Andrade, Vila Nova de Gaia; Rui Teixeira Almeida, Porto; Jose Silva Cardoso, Claudia Santos, Porto. Romania: Cristian Podoleanu, Attila Frigy, Targu-Mures; Maria Dorobantu, Ana-Gabriela Fruntelata, Bucharest. Russian Federation: Mariya Sitnikova, Mariya Trukshina, St Petersburg. Serbia and Montenegro: Masar Gashi, Prishtina UNMIK Kosova; Zorana Vasiljevic, Ana Vojvodic, Belgrade; Milutin Miric, Natasa Markovic, Goran Loncar, Belgrade; Jelica Milosavljevic, Snezana Ivanovic, Jagodina; Zoran Stajic, Zorica Nikolic, Zdravko Mijailovic, Belgrade. Slovenia: Mitja Lainscak, Murska Sobota; Stojan Kariz, Simon Ralca, Katarina Thaler, Izola; Bojan Vrtovec, Gregor Poglajen, Ljubljana. Spain: F. Fernandez Aviles, Manuel Gomez Bueno, Pedro Mota Gomez, Valladolid; Jose Azpitarte Almagro, Raquel del Valle Fernandez, Granada; Francisco Ridocci Soriano, Mercedes Nadal, Victoria Jacas, Valencia; Eduardo de Teresa, Manuel Jimenez-Navarro, Carlos Sanchez-Gonzalez, Antonio Muñoz-Garcia, Malaga; Enrique Galvo Basilio, Anna Diaz Calvo, Barcelona; Jose Luis Lopez-Sendon, Jose Juan Gomez de Diego, Madrid; Manuel Mendez, Nuria Munoz, Madrid; Norbero Herrera Guttierez, Joaquin Ruiz, Juan Ramon Siles Rubio, Cabra; Isabel Antorrena Miranda, Teresa Lozano, Isabel Selles, Pablo Oteo, Villajoyosa; Vicente Montagud Balaguer, Antonio Salvador Sanz, Valencia; Angeles Alonso Garcia, Alfredo Serrano, Madrid; Fransisco Epelde Gonzalo, Terrassa; Fernando Ruiz Rejon, David Marti Sanchez, Madrid; Inmaculada Roldan Rabadan, Madrid; Antoni Bayes Genis, Barcelona; Jose Antonio Gomez Guindal, Puerto del Rosario; Cristobal Lozano Cabezas, Jaen; Miguel A. Rodriguez Garcia, Leon; Josep Comin Colet, Barcelona; Eulalia Roig Minguell, Barcelona; Miguel Angel Ulecia Martinez, Juliana Caballero Gueto, Granada; Maria Luaces Mendez, Pontevedra. Switzerland: Ferenc Follath, Rita Pitsch, Zurich. The Netherlands: J.W. Deckers, Chr. G. Jansen, Rotterdam; Christel Ephraim, Zwolle, Meppel; R. Brons, Meppel; J.H.C. Panis, Rotterdam; C.W. Leenders, Rotterdam; M.J. Veerhoek, M. Kamps, Rotterdam; R. Haan, Harderwijk; L. Baur, Adrie van den Dool, Heerlen; H.J.G.M. Crijns, R. Nieuwlaat, H. Fransen, Maastricht; Theo Kessels, L. Eurlings, Venlo; M.J. De Boer, Zwolle; Herman Broers, Tilburg; C. Werter, Roermond; M.Bijl, S. Versluis, Dordrecht. Tunisia: Salem Kachboura, Zied Belhadj, Ariana. Turkey: Mustafa Kemal Erol, Erzurum; Cihangir Uyan, Bolu; Atila Iyisoy, Ankara. Ukraine: Vasyl Netiazhenko, Oleg Lykov, Kiev.


View Abstract