Compared with heart failure (HF) with reduced ejection fraction (HF-REF), the diagnosis of HF with preserved EF (HF-PEF) is more challenging. The aim of the study was to assess the prevalence of HF-PEF among patients hospitalized for HF, to evaluate the pertinence of HF-PEF diagnosis and to compare HF-PEF and HF-REF patients with respect to outcomes. The analysis included 661 Polish patients hospitalized for HF, selected from the European Society of Cardiology (ESC)-HF Long-Term Registry. Patients with an EF of ≥50% were included in the HF-PEF group and patients with an EF of <50% – in the HF-REF group. The primary end point was all-cause death at 1 year. The secondary end point was a composite of all-cause death and rehospitalization for HF at 1 year. HF-PEF was present in 187 patients (28%). Of those 187 patients, mitral inflow pattern was echocardiographically assessed in 116 patients (62%) and classified as restrictive/pseudonormal in 37 patients (20%). Compared with HF-REF subjects, patients with HF-PEF were older, more often female, and had a higher prevalence of hypertension, atrial fibrillation and sleep apnea. Despite lower B-type natriuretic peptide concentrations and lower prevalence of moderate-to-severe mitral regurgitation in patients with HF-PEF, congestive symptoms at admission were as severe as in patients with HF-REF. There were no significant differences in in-hospital mortality between the HF groups. One-year mortality was high in both groups (17% in HF-PEF vs 21% in HF-REF, p = 0.22). There was a trend toward a lower frequency of the secondary end point in the HF-PEF group (32% vs 40%, p = 0.07). In conclusion, in clinical practice, even easily obtainable echocardiographic indexes of diastolic dysfunction are relatively rarely acquired. One-year survival rate of patients with HF-PEF is not significantly better than that of patients with HF-REF.
The prevalence of heart failure (HF) with preserved ejection fraction (HF-PEF) has increased over the last years, with a further increase to be anticipated due to aging of the population and a growing incidence of arterial hypertension, obesity, and type 2 diabetes. In clinical practice, adequate echocardiographic evaluation of diastolic function tends to be neglected, as it necessitates a comprehensive examination, incorporating all relevant 2-dimensional, pulsed-wave Doppler (PWD) and tissue Doppler imaging (TDI) data. Thus, HF-PEF becomes a diagnosis by exclusion, potentially leading to HF misdiagnosis in patients in whom the actual cause of dyspnea or diminished exercise capacity fails to be identified. Another problem in HF-PEF is the choice of optimal pharmacotherapy, as – so far – no treatment has been shown to improve survival in HF-PEF. The aim of the study was to estimate the prevalence of HF-PEF in patients hospitalized for HF decompensation, to validate the pertinence of HF-PEF diagnosis in these patients, and to assess their clinical profile and outcomes in comparison to patients with HF with reduced EF (HF-REF).
Methods
The European Society of Cardiology (ESC) HF Long-Term Registry is an on-going, prospective, international, observational survey, with 211 cardiology centers from 21 European countries participating. The Registry includes both chronic HF patients presenting to ambulatory care clinics and patients admitted to hospital for new-onset or worsening HF. All patients with a diagnosis of HF who are aged ≥18 years are eligible for enrollment. The survey was approved by local ethical review boards according to the regulations of each participating country. A signed, informed consent was obtained from each patient after providing him/her with detailed information on the Registry.
During phase I of the Registry, lasting from May 2011 to April 2013, patients were enrolled on 1 specific day of the week for 12 consecutive months in each of the participating centers. In phase II/III of the Registry (currently on-going), patients are enrolled during 5 consecutive days per trimester. Data on clinical characteristics, diagnostic tests performed, and implemented treatment are collected in the Registry. Electronic case report forms (eCRFs) enable to describe echocardiographically evaluated left ventricular (LV) diastolic function by (1) denominating whether LV filling pattern, assessed by PWD, is restrictive/pseudonormal or not (yes vs no); (2) entering the value of the early (E) to late (A) LV filling velocity ratio (E/A ratio); and (3) entering the value of wave E deceleration time. Information on the presence of LV hypertrophy (LVH) is given dichotomically (yes vs no). It is also possible to enter left atrial (LA) dimension (measured in parasternal long-axis view) and LA volume in the Registry’s eCRF. All patients are followed for 12 months.
The current analysis included Polish patients hospitalized for HF, enrolled during phase I of the Registry. To discriminate between patients with HF-PEF and patients with HF-REF, the analysis included only those patients who had an echocardiographic examination (with EF assessment) performed during index hospitalization. Patients with an EF of ≥50% were included in the HF-PEF group, and patients with an EF of <50% were included in the HF-REF group.
To verify the pertinence of HF diagnosis in patients with EF ≥50%, we assessed whether they met the echocardiographic criteria for HF-PEF according to the 2012 ESC HF guidelines, that is, the presence of LVH and/or LA enlargement (defined as LA dimension of >40 mm and/or LA volume of >34 ml/m 2 ) and/or LV diastolic dysfunction (defined, for the sake of the current analysis, as restrictive/pseudonormal LV filling pattern and/or as E/A ratio of ≥2). We also evaluated serum concentrations of B-type natriuretic peptide (BNP) and of N-terminal pro-BNP (NT-proBNP) in these patients, and, after 2012 ESC guidelines, adopted a threshold of ≥100 pg/ml for BNP levels and of ≥300 pg/ml for NT-proBNP as justifying HF suspicion in patients hospitalized for exacerbation of symptoms suggestive of HF. We applied ESC guidelines, as the Registry was conducted in the European population.
The HF-PEF and HF-REF groups were compared with regard to clinical profile, initial presentation, diagnostic tests results, clinical course and management during index hospitalization, as well as in-hospital and 1-year outcomes.
The primary end point was all-cause death at 1 year. The secondary end point was a composite of all-cause death and hospital readmission for HF worsening at 1 year. We assessed the frequency of the primary and the secondary end points in both HF groups. In addition, we sought to determine predictors of the primary and the secondary end points separately for the HF-PEF and for the HF-REF group.
All statistical analyses were conducted using the SAS software, version 9.2 (SAS Institute Inc., Cary, NC). Normally distributed continuous variables were presented as mean (±SD), whereas ordinal variables and nonnormally distributed continuous variables were presented as median (interquartile range). The HF-PEF and HF-REF groups were compared using the Fisher’s exact test for categorical variables and the Mann–Whitney test for continuous and ordinal variables. Kaplan–Meier curves were plotted for the primary and the secondary end points in both groups. To identify the predictors of the primary and the secondary end points, Cox proportional hazards regression analyses were performed. All variables predictive of the primary or the secondary end points in univariate analyses were consequently included in multivariate models. All tests were 2 tailed. For all tests, a p value of <0.05 was deemed significant.
Results
The final analysis included 661 Polish in-hospital patients with echocardiography performed during index hospitalization: 187 patients (28%) with EF of ≥50% (HF-PEF group) and 474 patients (72%) with EF of <50% (HF-REF group), as shown in Figure 1 . In the studied cohort of 661 patients with HF, 229 subjects (35%) had EF of ≥45%, and 292 patients (44%) had EF of ≥40%.
Baseline characteristics, clinical course of index hospitalization, management and diagnostic tests performed during hospitalization, as well as in-hospital and 1-year outcomes of patients with HF-PEF and HF-REF are presented in Tables 1 to 3 .
| Variable | HF-PEF (n=187) | HF-REF (n=474) | P |
|---|---|---|---|
| Age (years) | 77 (69-84) | 67 (58-76) | <0.0001 |
| Women | 111 (59%) | 110 (23%) | <0.0001 |
| BMI (kg/m 2 ) | 28.0 (25.4-32.4); n=186 | 27.6 (24.9-30.8); n=472 | 0.08 |
| Last known EF before index hospitalization (%) | 56 (50-60); n=102 | 30 (20-38); n=344 | <0.0001 |
| Last known EF (before index hospitalization) ≥50% | 83/102 (81%) | 14/344 (4%) | <0.0001 |
| Primary described as HF-PEF | 20 (11%) | 0 (0%) | <0.0001 |
| Primary ischemic etiology of HF | 62 (33%) | 304 (64%) | <0.0001 |
| Dilated cardiomyopathy | 2 (1%) | 96 (20%) | <0.0001 |
| Tachycardia-related cardiomyopathy | 6 (3%) | 14 (3%) | 0.81 |
| Previous HF diagnosis | 148/186 (80%) | 406/473 (86%) | 0.06 |
| Previous HF hospitalization | 78/186 (42%) | 265/473 (56%) | 0.001 |
| Hypertension | 149 (80%) | 325/472 (69%) | 0.005 |
| Coronary artery disease | 69 (37%) | 293 (62%) | <0.0001 |
| Prior PCI or CABG | 31 (17%) | 201/472 (43%) | <0.0001 |
| Previously implanted ICD | 3 (2%) | 87 (18%) | <0.0001 |
| Previously implanted CRT | 1 (0.5%) | 37 (8%) | <0.0001 |
| Pacemaker | 16 (9%) | 23 (5%) | 0.10 |
| History of atrial fibrillation | 101 (54%) | 204 (43%) | 0.01 |
| Prior stroke or TIA | 24 (13%) | 43 (9%) | 0.15 |
| Peripheral artery disease | 32/186 (17%) | 70/473 (15%) | 0.47 |
| Diabetes | 61 (33%) | 170 (36%) | 0.47 |
| Chronic kidney disease | 52 (28%) | 143 (30%) | 0.57 |
| Sleep apnea | 21/185 (11%) | 15/466 (3%) | <0.0001 |
| Chronic obstructive pulmonary disease | 27 (14%) | 70/473 (15%) | 1.00 |
| Current smoker | 8 (4%) | 77 (16%) | <0.0001 |
| Current or former smoking | 69 (37%) | 307 (65%) | <0.0001 |
| Loop diuretic | 100 (54%) | 319 (67%) | 0.001 |
| Thiazide diuretic | 21 (11%) | 23 (5%) | 0.005 |
| Aldosterone antagonist | 57 (31%) | 250 (53%) | <0.0001 |
| ACE-I | 100 (54%) | 317 (67%) | 0.002 |
| ARB | 24 (13%) | 32 (7%) | 0.02 |
| Nitrates | 22 (12%) | 54 (11%) | 0.89 |
| β-blocker | 133 (71%) | 366 (77%) | 0.11 |
| Calcium channel blocker | 41 (22%) | 55 (12%) | 0.001 |
| Ivabradine | 0 (0%) | 3 (0.6%) | 0.56 |
| Digoxin | 26 (14%) | 97 (21%) | 0.06 |
| Amiodarone | 11 (6%) | 51 (11%) | 0.06 |
| Statin | 97 (52%) | 271 (57%) | 0.23 |
| Antiplatelets | 90 (48%) | 271 (57%) | 0.04 |
| Anticoagulants | 60 (32%) | 156 (33%) | 0.86 |
| Insulin | 28 (15%) | 67 (14%) | 0.81 |
| HF-PEF (n=187) | HF-REF (n=474) | P | |
|---|---|---|---|
| Cardiogenic shock | 2 (1%) | 14 (3%) | 0.26 |
| NYHA class | 3 (3-4) | 3 (3-4); n=471 | |
| I | 0 (0%) | 0 (0%) | |
| II | 41 (22%) | 103/471 (22%) | 0.93 |
| III | 83 (44%) | 212/471 (45%) | |
| IV | 63 (34%) | 156/471 (33%) | |
| Pulmonary rales | 126 (67%) | 298 (63%) | 0.28 |
| Pulmonary congestion/alveolar oedema on chest X-ray | 59/97 (61%) | 136/222 (61%) | 1.00 |
| Right ventricular HF | 17 (9%) | 14 (3%) | 0.002 |
| Peripheral oedema | 101 (54%) | 239 (50%) | 0.44 |
| Systolic blood pressure (mmHg) | 135 (120-160) | 120 (110-140) | <0.0001 |
| Diastolic blood pressure (mmHg) | 80 (70-90) | 75 (70-80); n=473 | 0.0009 |
| Heart rate (b.p.m.) | 80 (70-100) | 80 (70-100) | 0.72 |
| Paced heart rhythm (ECG) | 11/186 (6%) | 56/469 (12%) | 0.02 |
| AF (ECG) | 70/186 (38%) | 111/469 (24%) | <0.0001 |
| AF as a cause of admission | 64 (34%) | 139 (29%) | 0.23 |
| VF or VT as a cause of admission | 10 (5%) | 55 (12%) | 0.01 |
| ACS as a cause of admission | 6 (3%) | 71 (15%) | <0.0001 |
| Uncontrolled hypertension as a cause of admission | 44 (24%) | 38 (8%) | <0.0001 |
| Renal dysfunction as a cause of admission | 20 (11%) | 62 (13%) | 0.44 |
| NT-proBNP (pg/ml) | 1852 (722-5286); n=62 | 4085 (1691-8675); n=203 | 0.0001 |
| BNP (pg/ml) | 217 (116-479); n=33 | 656 (249-1337); n=64 | 0.0004 |
| Serum sodium (mmol/l) | 140 (137-142); n=185 | 139 (137-141); n=472 | 0.005 |
| Serum creatinine (mg/dl) | 1.02 (0.84-1.27); n=185 | 1.14 (0.91- 1.43); n=472 | 0.0003 |
| Hemoglobin (g/dl) | 13.3 (11.8-14.3); n=185 | 13.5 (12.2-14.6); n=469 | 0.03 |
| Echocardiography | |||
| Ejection fraction (%) | 55 (50-60) | 30 (21-38) | <0.0001 |
| LVEDD (mm) | 48 (43-52); n=173 | 60 (54-70); n=444 | <0.0001 |
| Left ventricular hypertrophy | 96/183 (53%) | 165/446 (37%) | 0.0005 |
| E/A ≥2 | 2/52 (4%) | 42/135 (31%) | <0.0001 |
| E/A <1 | 33/52 (64%) | 64/135 (47%) | 0.05 |
| E/A [1-2) | 17/52 (33%) | 29/135 (22%) | 0.13 |
| Restrictive/pseudonormal pattern ∗ | 37/116 (32%) | 150/319 (47%) | 0.006 |
| Deceleration time (ms) † | 203 (140-290); n=27 | 149 (108-218); n=64 | 0.04 |
| LA dimension (mm) | 47 (40-54); n=94 | 48 (43-52); n=229 | 0.71 |
| LA dimension >40 mm | 70/94 (75%) | 193/229 (84%) | 0.71 |
| LA volume (ml) | 73 (45-110); n=43 | 55 (45-90); n=66 | 0.36 |
| LA volume >34 ml/m 2 | 24/43 (56%) | 26/65 (40%) | 0.12 |
| Aortic stenosis ‡ | 25/185 (14%) | 31/466 (7%) | 0.008 |
| Aortic regurgitation ‡ | 13/186 (7%) | 43/465 (9%) | 0.44 |
| Mitral regurgitation ‡ | 79/185 (43%) | 272/466 (58%) | <0.0001 |
| Tricuspid regurgitation ‡ | 72/186 (39%) | 199/466 (43%) | 0.38 |
| Variable | HF-PEF (n=187) | HF-REF (n=474) | P |
|---|---|---|---|
| Inotropic support | 9 (5%) | 82/472 (17%) | <0.0001 |
| Intravenous nitrates | 25 (13%) | 56/471 (12%) | 0.60 |
| Intravenous diuretics | 106 (57%) | 289/472 (61%) | 0.29 |
| Coronary angiography | 24 (13%) | 160/471 (34%) | <0.0001 |
| PCI/CABG | 10 (5%) | 75/472 (16%) | <0.0001 |
| Holter-ECG | 67 (36%) | 164/470 (35%) | 0.86 |
| Exercise test | 18/176 (10%) | 63/463 (14%) | 0.29 |
| Heart rate (b.p.m.) ∗ | 70 (65-80) | 70 (65-80) | 0.89 |
| Systolic blood pressure (mmHg) ∗ | 120 (110-130) | 115 (105-125) | <0.0001 |
| NYHA class ∗ | 2 (2-3) | 2 (2-3) | 0.002 |
| I | 11 (6%) | 16 (4%) | |
| II | 126 (69%) | 268 (59%) | |
| III | 45 (25%) | 165 (36%) | |
| IV | 2 (1%) | 9 (2%) | |
| Serum creatinine (mg/dl) ∗ | 1.05 (0.88-1.31); n=137 | 1.12 (0.94-1.41); n=354 | 0.04 |
| Loop diuretic ∗ | 134 (73%) | 390 (85%) | 0.0004 |
| Thiazide diuretic ∗ | 16 (9%) | 18 (4%) | 0.02 |
| Aldosterone antagonist ∗ | 92 (50%) | 345 (75%) | <0.0001 |
| ACE-I ∗ | 114 (62%) | 377 (82%) | <0.0001 |
| ARB ∗ | 30 (16%) | 33 (7%) | 0.001 |
| Nitrates ∗ | 16 (9%) | 49 (11%) | 0.56 |
| β-blocker ∗ | 151 (82%) | 428 (94%) | <0.0001 |
| Bisoprolol ∗ | 58 (32%) | 143 (31%) | 1.00 |
| Carvedilol ∗ | 29 (16%) | 193 (42%) | <0.0001 |
| Metoprolol ∗ | 41 (22%) | 55 (12%) | 0.001 |
| Nebivolol ∗ | 14 (8%) | 34 (7%) | 1.00 |
| Target β-blocker dose reached † | 10 (5%) | 38 (8%) | 0.25 |
| Calcium channel blocker ∗ | 51 (28%) | 51 (11%) | <0.0001 |
| Ivabradine ∗ | 0 (0%) | 6 (1%) | 0.19 |
| Digoxin ∗ | 33 (18%) | 119 (26%) | 0.03 |
| Amiodarone ∗ | 11 (6%) | 69 (15%) | 0.001 |
| Statin ∗ | 111 (60%) | 332 (73%) | 0.003 |
| Antiplatelets ∗ | 91 (50%) | 304 (66%) | 0.0001 |
| Anticoagulants ∗ | 90 (49%) | 202 (44%) | 0.29 |
| Insulin ∗ | 29 (16%) | 76 (17%) | 0.91 |
| Hospitalization length (days) | 7 (4-10) | 7 (4-12) | 0.26 |
| Death during hospitalization | 3 (1.6%) | 16 (3.4%) | 0.30 |
| Death at 1 year | 30/182 (17%) | 92/438 (21%) | 0.22 |
| Death or rehospitalization at 1 year | 58/182 (32%) | 175/438 (40%) | 0.07 |
∗ At discharge (in patients who survived to hospital discharge, i.e., in 184 patients with HF-PEF, and in 458 patients with HF-REFs).
† That is: bisoprolol ≥10 mg daily, carvedilol ≥50 mg daily, metoprolol ≥200 mg daily, or nebivolol ≥10 mg daily.
Detailed echocardiographic and laboratory characteristics of patients with HF-PEF are presented in Table 2 . Of 187 patients in the HF-PEF group, 144 patients (77%) met the echocardiographic criteria for HF-PEF, as defined in the Methods section (i.e., the presence of LVH, LA dilation, restrictive/pseudonormal LV filling pattern, and/or E/A ratio of ≥2). This was mostly due to the presence of LVH (assessed in 183 patients [98%] and confirmed in 96 patients [51% of the whole HF-PEF group]) and LA dilation (assessed in 137 patients [73%] and confirmed in 94 patients [50% of the whole HF-PEF group]). The PWD mitral inflow velocity pattern was assessed in 116 (62%) of 187 patients in the HF-PEF group, with exact values of E/A ratio given only in 52 (28%) of 187 patients. Merely 37 patients (20% of the whole HF-PEF group) were classified as demonstrating a restrictive/pseudonormal LV filling pattern by the Registry’s investigators. Deceleration time of the E wave was assessed only in 27 (14%) of the 187 patients.
Serum concentrations of NT-proBNP were measured in 62 (33%) of 187 patients with HF-PEF. Of those 62 patients, 57 patients (92%) had an NT-proBNP level of ≥300 pg/ml. Serum concentrations of BNP were evaluated in 33 (18%) of 187 patients with HF-PEF, all of them had a BNP level of ≥100 pg/ml. Of 187 patients from the HF-PEF group, 162 patients (87%) either met the prespecified echocardiographic criteria for HF-PEF or had an NT-proBNP level of ≥300 pg/ml or a BNP level of ≥100 pg/ml.
Of 620 patients with data on 1-year follow-up, 122 patients (20%) reached the primary end point: 30 patients in the HF-PEF group (including 3 patients who died during index hospitalization) and 92 patients in the HF-REF group (including 16 patients who died during index hospitalization), as presented in Table 3 . The secondary end point was reached by 233 patients (38%; Table 3 ). Kaplan–Meier curves for the primary and the secondary end points in both HF groups are plotted in Figures 2 and 3 , respectively. Univariate analyses of predictors of the primary and the secondary end points in both groups are presented in the Supplementary Material (Tables S1 and S2) . Multivariate analyses of predictors of the primary and the secondary end points for the HF-PEF and the HF-REF group are presented in Table 4 . Due to the lack of complete data for some of the patients in the Registry, multivariate models included only those patients for whom all required parameters were available, that is, 177 patients from the HF-PEF group and 420 patients from the HF-REF group for the primary end point analyses, and 176 patients from the HF-PEF group and 421 patients from the HF-REF group for the secondary end point analyses.
