Acute heart failure (AHF) with preserved left ventricular ejection fraction (PLVEF) represents a significant part of AHF syndromes featuring particular characteristics. We sought to determine the clinical profile and predictors of in-hospital mortality in patients with AHF and PLVEF in the Acute Heart Failure Global Survey of Standard Treatment (ALARM-HF). This survey is an international observational study of 4,953 patients admitted for AHF in 9 countries (6 European countries, Mexico, and Australia) from October 2006 to March 2007. Patients with PLVEF were defined by an LVEF ≥45%. Of the total cohort, 25% of patients had PLVEF. In-hospital mortality was significantly lower in this subgroup (7% vs 11% in patients with decreased LVEF, p = 0.013). Candidate variables included demographics, baseline clinical findings, and treatment. Multivariate logistic regression analysis showed that the variables independently associated with in-hospital mortality included systolic blood pressure at admission (p <0.001), serum sodium (p = 0.041), positive troponin result (p = 0.023), serum creatinine >2 mg/dl (p = 0.042), history of peripheral vascular disease and anemia (p = 0.004 and p = 0.015, respectively), secondary (hospitalization for other reason) versus primary AHF diagnosis (p = 0.043), and previous treatment with diuretics (p = 0.023) and angiotensin-converting enzyme inhibitors (p = 0.021). In conclusion, patients with AHF and PLVEF have lower in-hospital mortality than those with decreased LVEF. Low systolic blood pressure, low serum sodium, renal dysfunction, positive markers of myocardial injury, presence of co-morbidities such as peripheral vascular disease and anemia, secondary versus primary AHF diagnosis, and absence of treatment with diuretics and angiotensin-converting enzyme inhibitors at admission may identify high-risk patients with AHF and PLVEF.
Limited data are available regarding demographics, clinical profile, management, and predictors of in-hospital mortality in patients with acute heart failure (AHF) and preserved left ventricular ejection fraction (PLVEF). Previous studies have indicated that these patients have different clinical features and outcomes from patients with HF and decreased LVEF. However, recent AHF registries have suggested that, although demographics and treatment at time of hospital discharge differ between the 2 groups, clinical presentation at admission is similar and rates of mortality and morbidity are equally high according to follow-up data. The identification of patients with AHF and PLVEF at high risk for in-hospital mortality is crucial because of the need for aggressive monitoring and intervention. We sought to determine the predictors of in-hospital mortality among baseline characteristics of these patients in the Acute Heart Failure Global Survey of Standard Treatment (ALARM-HF).
Methods
The ALARM-HF is a hospital-based observational study including patients hospitalized for AHF syndromes in Europe, Latin America and Australia. More specifically, this survey enrolled 4,953 patients admitted to cardiology departments (67%) or intensive care units (33%) for AHF in 9 countries (6 European countries, Turkey, Mexico, and Australia). Participating hospitals were a representative sample according to geographic region, hospital magnitude (by number of beds), sector (public vs private), and type (academic vs nonteaching status). Data collection was conducted from October 2006 to March 2007.
This database contains thorough patient characteristics including demographic data, medical history, initial clinical evaluation, diagnostic procedures, treatment patterns, and in-hospital outcomes. Patient case-report forms were completed at or close to discharge every 5 to 8 consecutive patients, with diagnoses to be made at discharge according to the definition and classification of European Society of Cardiology guidelines. These also included cause of AHF (primary vs secondary), co-morbidities, precipitating factors, echocardiographic data, and details of intravenous drugs (timing and location of initiation, dosage, duration). The protocol was approved by the institutional review board of each participating center; however, written informed consent of patients was not required for registry entry. All patient-related variables, clinical diagnoses, and outcomes used standardized definitions.
SPSS 13.0 (SPSS, Inc., Chicago, Illinois) was used. Data are reported as counts and percent nonmissing values for categorical variables or mean ± SD for quantitative variables. Categorical variables were compared between patients with preserved and depressed LVEF using chi-square tests, and quantitative variables were compared using t test or analysis of variance Wilcoxon rank-sum tests. Receiver operating characteristics analysis was performed to evaluate the predictive value of variables and identify cut-off values. Various models of in-hospital mortality and of postdischarge rehospitalization or mortality have been developed to adjust for significant covariates. Multivariate logistic regression analysis was used to identify independent prognostic predictors of in-hospital mortality. Demographics, clinical and laboratory findings at presentation, medical history, cardiovascular risk factors, and background medication were forced into the regression model. Odds ratios and corresponding 95% confidence intervals are reported for each covariate. For all tests, which were 2-sided, a p value ≤0.05 was considered as indicating statistically significant differences.
Results
Clinical characteristics and outcomes in overall patients with AHF enrolled in ALARM-HF registry have been presented previously. This database included 4,953 patients hospitalized for AHF syndromes in Europe, Latin America, and Australia (588 in France, 617 in Germany, 679 in Italy, 700 in Spain, 623 in the United Kingdom, 255 in Greece, 628 in Turkey, 601 in Mexico, and 262 in Australia) at 666 hospitals from October 2006 to March 2007. In 3,283 patients (66%) of this cohort, LVEF was quantitatively determined. AHF with preserved LVEF represented a significant proportion concerning 25% of these patients. Data regarding the patient subpopulation without an in-hospital assessment of LVEF are clinically distinct, reflect an apparent higher-risk profile, and have many similarities to those of respective patient populations of other registries. These data are not presented in the tables because these are not the focus of this report.
Differences in baseline characteristics and clinical presentation between patient with AHF and PLVEF and those with decreased LVEF are presented in Table 1 . Patients with PLVEF more frequently developed symptoms of HF at hospital (14% vs 11% for decreased LVEF) or were admitted at the emergency room independently (18% vs 14% for decreased LVEF) and less frequently admitted by ambulance (44% vs 47% for decreased LVEF, p = 0.010 for all equations). Compared to patients with LV systolic dysfunction, those with PLVEF were more frequently women (p <0.001), had higher percent de novo AHF rather than acute exacerbation of chronic HF (p <0.001), but lower prevalence of acute coronary syndromes as the cause of AHF (p <0.001). Moreover, patients with PLVEF had less often cardiogenic shock, but more frequently hypertensive HF and right HF (p <0.001 for all equations). Prehospitalization New York Heart Association class was better in these patients (class IV, 29% vs 41%, p <0.001).
| PLVEF | Decreased LVEF | p Value | |
|---|---|---|---|
| (n = 837) | (n = 2,446) | ||
| Gender | <0.001 | ||
| Women | 52% | 30% | |
| Men | 48% | 70% | |
| Acute heart failure diagnosis | NS | ||
| Primary acute heart failure | 77% | 78% | |
| Secondary acute heart failure | 23% | 22% | |
| Heart failure presentation | <0.001 | ||
| Acutely decompensated heart failure | 55% | 64% | |
| De novo acute heart failure | 45% | 36% | |
| Acute heart failure classification | <0.001 | ||
| Acutely decompensated heart failure | 34% | 41% | |
| Cardiogenic shock | 6% | 13% | |
| High cardiac output | 1.0% | 0.8% | |
| Hypertensive acute heart failure | 13% | 5% | |
| Pulmonary edema | 38% | 37% | |
| Right heart failure | 8% | 3% | |
| New York Heart Association class (before hospitalization) | <0.001 | ||
| I | 1.4% | 1.1% | |
| II | 12% | 6% | |
| III | 40% | 35% | |
| IV | 29% | 41% | |
| Cardiovascular co-morbidities | |||
| Atrial fibrillation/flutter | 26% | 24% | NS |
| Cardiomyopathy | 8% | 17% | <0.001 |
| Peripheral vascular disease | 8% | 9% | NS |
| Coronary artery disease | 20% | 35% | <0.001 |
| Congestive heart failure | 27% | 36% | <0.001 |
| Heart valvular disease | 19% | 14% | <0.001 |
| Obesity | 30% | 26% | 0.02 |
| Diabetes mellitus | 42% | 44% | NS |
| Dyslipidemia | 41% | 45% | 0.019 |
| Arterial hypertension | 72% | 69% | 0.046 |
| Noncardiovascular co-morbidities | |||
| Anemia | 15% | 13% | NS |
| Depression | 8% | 9% | NS |
| Hyponatremia | 4% | 6% | 0.04 |
| Renal dysfunction | 26% | 30% | 0.023 |
| Symptoms and signs | |||
| Dyspnea | 72% | 74% | NS |
| Fatigue | 47% | 43% | NS |
| Orthopnea | 55% | 59% | 0.02 |
| Peripheral edema | 40% | 44% | 0.042 |
| Jugular venous distension | 37% | 44% | <0.001 |
| Rales | 57% | 65% | <0.001 |
| Weight gain | 24% | 29% | 0.003 |
| Renal function | <0.001 | ||
| Anuria | 3% | 6% | |
| Oliguria | 33% | 38% | |
| Normal diuresis | 50% | 43% | |
| Positive cardiac markers | |||
| Troponin T | 22% | 35% | <0.001 |
| Cardiac isoenzyme of creatinine phosphokinase | 16% | 25% | <0.001 |
| Precipitating factors | |||
| Acute coronary syndromes | 24% | 38% | <0.001 |
| Arrhythmia | 27% | 26% | NS |
| Noncompliance | 12% | 14% | NS |
| Infection | 15% | 14% | NS |
| Risk factors | |||
| Smoker | 20% | 25% | 0.003 |
| Former smoker | 26% | 38% | <0.001 |
| New York Heart Association class (at discharge) | <0.001 | ||
| I | 31% | 19% | |
| II | 44% | 46% | |
| III | 8% | 12% | |
| IV | 0.4% | 1.1% |
Concerning clinical examination, patients with PLVEF less frequently had orthopnea (p = 0.02) and peripheral edemas (p = 0.042). Rales, jugular venous distension, and weight gain were mostly observed in patients with decreased LVEF ( Table 1 ). Patients with PLVEF also exhibited higher systolic blood pressure (SBP) at admission (p <0.001) as presented in Table 2 . Levels of troponin T and creatine kinase-MB were also less frequently positive in patients with PLVEF compared to those with decreased LVEF ( Table 1 ).
| PLVEF | Decreased LVEF | p Value | |
|---|---|---|---|
| Vital signs | |||
| Systolic blood pressure at admission (mm Hg) | 146 ± 42 | 129 ± 38 | <0.001 |
| Systolic blood pressure at discharge (mm Hg) | 125 ± 20 | 116 ± 22 | 0.002 |
| Systolic blood pressure change (mm Hg) | −23 ± 36 | −14 ± 31 | <0.001 |
| Heart rate at admission (beats/min) | 108 ± 28 | 107 ± 26 | NS |
| Oxygen saturation at admission (%) | 88 ± 7 | 88 ± 8 | NS |
| Body weight | |||
| Weight at admission (kg) | 76 ± 18 | 80 ± 16 | NS |
| Weight change (kg) | −3.4 ± 3.6 | −4.3 ± 4.3 | 0.03 |
| Body mass index (kg/m 2 ) | 27 ± 5 | 28 ± 5 | NS |
| Laboratory findings | |||
| Sodium at admission (mmol/L) | 137 ± 8 | 136 ± 7 | NS |
| Sodium at discharge (mmol/L) | 138 ± 7 | 137 ± 6 | NS |
| Sodium change (mmol/L) | 0.8 ± 6.8 | 0.6 ± 6.6 | NS |
| Serum creatinine at admission (mg/dl) | 1.5 ± 1.6 | 1.6 ± 1.6 | NS |
| Serum creatinine change (mg/dl) | −0.1 ± 1.5 | 0.0 ± 1.6 | NS |
| Urine volume (ml) | 1,609 ± 979 | 1,506 ± 1,064 | 0.027 |
| Troponin T (ng/mL) | 0.3 ± 0.4 | 0.4 ± 0.5 | <0.001 |
| Cardiac isoenzyme of creatinine phosphokinase (U/L) | 0.2 ± 0.4 | 0.4 ± 0.5 | <0.001 |
Co-morbidities, cardiovascular and noncardiovascular, are also presented in Table 1 . Patients with PLVEF were more likely to be obese (p = 0.02) and hypertensive (p = 0.046) and had higher prevalence of valvular disease (p <0.001). Renal dysfunction and hyponatremia were also significantly less frequent in patients with PLVEF (p = 0.023 and p = 0.04, respectively). No significant difference in prevalence of atrial fibrillation, diabetes mellitus, anemia, or depression was observed between the 2 groups.
Differences in treatment methods during hospitalization in patients with or without PLVEF are listed in Table 3 . Most patients from the 2 groups received diuretics; however, patients with decreased LVEF received more frequently angiotensin-converting enzyme (ACE) inhibitors or angiotensin II receptor blockers and β blockers or digitalis. Moreover, the more frequent ischemic cause of HF in patients with decreased LVEF may interpret the increased use of aspirin and clopidogrel in these patients (p <0.001). Calcium channel blockers were the only agents administered more often in PLVEF patients rather than patients with decreased LVEF (p <0.001).
| PLVEF | Decreased LVEF | p Value | |
|---|---|---|---|
| Diuretics | 96% | 97% | NS |
| Diuretics per os | 71% | 75% | 0.012 |
| Diuretics intravenously | 90% | 92% | 0.031 |
| Angiotensin-converting enzyme inhibitors | 57% | 68% | <0.001 |
| Angiotensin II receptor blockers | 47% | 54% | <0.001 |
| Beta blockers | 42% | 53% | <0.001 |
| Nitrates per os/transdermal therapeutic system | 22% | 26% | 0.012 |
| Nitrates intravenously | 40% | 43% | NS |
| Calcium channel blockers | 3.1% | 0.7% | <0.001 |
| Other vasodilators | 4% | 2% | 0.001 |
| Digitalis | 25% | 31% | 0.001 |
| Aspirin | 45% | 59% | <0.001 |
| Clopidogrel | 14% | 21% | <0.001 |
| Continuous positive airway pressure | 8% | 10% | 0.035 |
| Ventouri mask | 15% | 16% | NS |
| Percutaneous coronary intervention | 9% | 17% | <0.001 |
| Coronary artery bypass graft surgery | 3% | 4% | NS |
| Intra-aortic balloon pump | 3% | 7% | <0.001 |
| Pacemaker | 4% | 2% | 0.025 |
| Adrenaline | 2% | 4% | 0.003 |
| Dobutamine | 17% | 26% | <0.001 |
| Dopamine | 9% | 15% | <0.001 |
| Noradrenaline | 5% | 4% | NS |
| Amiodarone | 3% | 3% | NS |
Clinical outcomes during hospitalization are demonstrated in Table 4 . AHF patients with PLVEF experienced better in-hospital outcome rather than patients with decreased LVEF, concerning lower in-hospital mortality (7% vs 11%, p = 0.013) and decreased transfer to rehabilitation centers. Furthermore, higher percentage of patients with PLVEF were discharged home (65% vs 62% for decreased LVEF, p = 0.013). No difference was observed in the proportion of patients discharged to intensive care unit or transferred to other hospitals between the 2 groups.
| PLVEF | Decreased LVEF | p Value | |
|---|---|---|---|
| Outcome | 0.013 | ||
| Death | 6.9% | 10.8% | |
| Transfer from intensive care unit to cardiology ward | 10.3% | 10.3% | |
| Discharged to rehabilitation center | 5.6% | 6.5% | |
| Discharged required to health care facility | 6.7% | 5.1% | |
| Discharged to other hospital | 5.3% | 5.0% | |
| Discharged home | 65.2% | 62.2% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
