Contemporary heart failure (HF) therapies have not resulted in improved outcomes among patients with HF and preserved left ventricular ejection fraction (LVEF). We sought to evaluate the differential effect of LVEF on long-term mortality after hospitalization for acute decompensated HF in a real-world setting. All-cause mortality at 4 years after hospitalization for HF was assessed by LVEF (categorized as preserved [≥50%], mildly [40% to 49%], moderately [30% to 39%], and severely [<30%] reduced) among 1,620 patients enrolled in the Heart Failure Survey in ISrael. Among the study patients, 30% had preserved LVEF and 20%, 25%, and 25%, had mild, moderate, and severe reductions in LVEF, respectively. Multivariate analysis showed that patients with preserved LVEF had a similar risk of long-term mortality as patients with mild or moderate reduction in LVEF (hazard ratio [HR] 0.92 [p = 0.40] and 1.01 [p = 0.90], respectively) while severely reduced LVEF conferred increased increase rate compared with preserved LVEF (HR 1.20, p = 0.04). Interaction term analysis showed that the risk associated with severely reduced LVEF was evident only among patients ≤75 years (HR 1.49, p = 0.003), whereas among older patients, there was no difference in the risk of long-term mortality between those with preserved versus severely reduced LVEF (HR 1.02 [p = 0.86]; p value for age-by-LVEF interaction = 0.03). In conclusion, patients hospitalized for HF who have preserved LVEF experience similar long-term mortality as patients with mild or moderate reductions in LVEF, whereas severely reduced LVEF remains an independent predictor of long-term mortality in this population. The differential effect of LVEF on long-term mortality is significantly attenuated in the older age group.
Data from surveys of patients with heart failure (HF) have demonstrated an inconsistent association between left ventricular ejection fraction (LVEF) and long-term outcomes in patients with HF. In contrast, limited data from randomized controlled trials suggest that only patients with severe reductions in LVEF experience worse prognosis compared with those with HF who have preserved LVEF. However, these trials have enrolled a generally healthier and younger patient population because of stringent exclusion criteria and, thus, are subject to selection bias, especially in the older age group. Therefore, there are limited contemporary real-world data on the association between LVEF and long-term survival after hospitalization for acute HF. The present study was carried out among 1,620 patients enrolled in the Heart Failure Survey in ISrael (HFSIS) and was designed to (1) compare the risk of long-term mortality between HF patients with preserved LVEF and those with varying degrees of cardiac dysfunction in a real-world setting and (2) evaluate the association between LVEF and survival in the older age group.
Methods
The design and methods of the HFSIS have been described previously. Briefly, the survey, conducted in March and April 2003, included 4,102 patients diagnosed with HF, admitted to 93 of the 98 internal medicine departments and 24 of the 25 cardiology departments in all 25 public hospitals in Israel. The criteria used for diagnosing HF were symptoms of HF (at rest or during exercise) and objective evidence of cardiac dysfunction at rest. Acute HF was defined as a rapid onset or change in the signs or symptoms of HF, resulting in the need for urgent therapy. Acute HF may be either acute de novo or acute exacerbation of chronic HF. Of the total 4,102 patients, 2,212 were hospitalized with a diagnosis of either acute or acute exacerbation of HF. Echocardiography during the index hospitalizations was performed in 1,620 (63%) of the patients who comprised the present study population.
For the present study, LVEF was categorized as follows: preserved ≥50%, mildly impaired 40% to 49%, moderately impaired 30% to 39%, and severely impaired <30%. Age was dichotomized at the approximate median value of 75 years. Thus, elderly patients were defined as those >75 years of age. Diabetes mellitus was defined by 1 of the following criteria: a history of diabetes mellitus obtained from medical records, admission blood glucose ≥200 mg/dl, or the use of antidiabetic agents (on admission or discharge). New York Heart Association (NYHA) functional class was determined according to the functional status and symptoms before the index hospitalization.
Renal function was categorized using Modification of Diet in Renal Disease formula for estimated glomerular filtration rate as follows: good renal function ≥50 ml/min/1.73 m 2 and impaired renal function <50 ml/min/1.73 m 2 .
The end point of this study was 4-year all-cause mortality, obtained through the Israeli National Population Register. Every effort was made to ensure accurate and reliable profiling data, which included standardizing HF and data validation definitions. The protocol was approved by the ethics committee at each of the participating hospitals.
Baseline characteristics of the patients were compared according to LVEF at baseline. Comparison of categorical variables was performed with chi-square analysis and that of continuous variables was performed with the 1-way analysis of variance test. The cumulative probabilities of all-cause mortality during 4 years of follow-up were assessed according to the Kaplan-Meier method, with comparison of cumulative events by the log-rank test.
Multivariate Cox proportional hazards regression analysis was applied to measure predetermined demographic and clinical factors presumed to be associated with mortality outcome. The prespecified covariates in addition to the aforementioned LVEF categories included age, gender, NYHA score, hemoglobin, blood sodium, systolic blood pressure, estimated glomerular filtration rate, and history of diabetes mellitus. Interaction term analysis was used to assess the effect of age (dichotomized at the median value of 75 years) on the association between LVEF and long-term mortality. In this analysis, age-by-LVEF category interaction terms were added to the multivariate model. The statistical software used was SPSS, version 19.0 (SPSS Inc., Chicago, Illinois). p Value <0.05 was considered significant.
Results
Among 1,620 patients, 30% had preserved LVEF and 20%, 25%, and 25% had mild, moderate, and severe reductions in LVEF, respectively ( Table 1 ).
| Variable | LVEF | p Value ∗ | |||
|---|---|---|---|---|---|
| ≥50% (n = 484), % | 40%–49% (n = 323), % | 30%–39% (n = 412), % | <30% (n = 401), % | ||
| Demographic characteristics | |||||
| Age (yrs), mean ± SD | 76 ± 5.8 | 73 ± 10.7 | 73 ± 10.5 | 73 ± 10 | <0.001 |
| Age >75 yrs | 56 | 43 | 46 | 43 | <0.001 |
| Women | 64 | 4 | 33 | 27 | <0.001 |
| BMI, mean ± SD | 29 ± 6.0 | 29 ± 7.4 | 27 ± 4.4 | 26 ± 4.6 | <0.001 |
| Vascular risk factors | |||||
| Smoking | 23 | 33 | 34 | 39 | <0.001 |
| Hypertension | 82 | 75 | 68 | 59 | <0.001 |
| Dyslipidemia | 32 | 42 | 40 | 38 | 0.01 |
| Diabetes mellitus | 40 | 46 | 49 | 47 | 0.06 |
| NYHA class III–IV | 41 | 38 | 24 | 58 | <0.001 |
| In-hospital features | |||||
| PCI | 8 | 19 | 22 | 15 | <0.001 |
| TC >170 mg/dl | 43 | 47 | 44 | 33 | 0.002 |
| SBP <140 mm Hg | 36 | 42 | 48 | 64 | <0.001 |
| Na <130 mEq/dl | 7 | 4 | 7 | 8 | 0.11 |
| Hb <11 g/L | 33 | 29 | 27 | 23 | 0.01 |
| eGFR (ml/min), mean ± SD | 52 ± 24 | 57 ± 25 | 53 ± 24 | 51 ± 23 | 0.01 |
| Co-morbidities | |||||
| Ischemic heart disease | 61 | 84 | 92 | 88 | <0.001 |
| Stroke | 11 | 14 | 11 | 13 | 0.57 |
| PVD | 9 | 8 | 14 | 15 | 0.002 |
| COPD | 20 | 19 | 18 | 19 | 0.86 |
| Heart failure origin | |||||
| Ischemic | 51 | 79 | 89 | 82 | <0.001 |
| Diabetic | 22 | 23 | 23 | 25 | 0.78 |
| Hypertensive | 53 | 42 | 34 | 32 | <0.001 |
| Cor pulmonale | 8 | 3 | 1 | 3 | <0.001 |
| Cardiomyopathy | 12 | 9 | 9 | 25 | <0.001 |
| Valvular | 36 | 32 | 23 | 30 | <0.001 |
| Home medications | |||||
| Aspirin | 47 | 60 | 61 | 61 | <0.001 |
| Clopidogrel | 3 | 7 | 4 | 6 | 0.05 |
| Anticoagulant | 20 | 16 | 14 | 20 | 0.05 |
| ACEI | 44 | 53 | 50 | 54 | 0.03 |
| β Blockers | 47 | 52 | 49 | 52 | 0.35 |
| Furosemide | 58 | 54 | 58 | 70 | <0.001 |
| Spironolactone | 11 | 7 | 14 | 25 | <0.001 |
| Digoxin | 10 | 9 | 11 | 22 | <0.001 |
| Calcium antagonists | 35 | 26 | 21 | 14 | <0.001 |
| Amiodarone | 10 | 11 | 12 | 14 | 0.35 |
∗ Denotes the overall statistical difference among the 4 LVEF groups.
Patient characteristics, within each LVEF group, significantly differed in most aspects of demographics, clinical history, medication, and in-hospital features. Notably, patients with preserved LVEF were older and more obese than those with any reduction in LVEF and comprised a higher rate of women and a lesser frequency of patients with ischemic heart disease. The frequency of vascular risk factors such as smoking and dyslipidemia was inversely correlated with LVEF. However, the frequency of hypertension was directly correlated with LVEF, and the rate of diabetes mellitus did not differ significantly among the 4 LVEF groups. Notably, baseline NYHA functional class was similar between patients with mild to moderate reductions in LVEF and those with preserved LVEF but was significantly more advanced among those with severely reduced LVEF.
Patient groups showed a variability of in-hospital parameters as well ( Table 1 ). Patients with reduced LVEF exhibited lower levels of systolic blood pressure and total cholesterol. In contrast, the rate of patients with renal failure and low sodium levels during the index hospitalization was similar between the 4 LVEF groups ( Table 1 ).
Medical therapy differed significantly among patient groups. Cardiac and HF medications, including aspirin, nitrates, statins, digoxin, angiotensin-converting enzyme inhibitors (ACEI) or angiotensin receptor blocker (ARB), and spironolactone therapy, were administered at increased frequency with decreasing LVEF, whereas the rate of therapy with calcium channel and α blockers was directly correlated with LVEF ( Table 1 ).
Kaplan-Meier analysis ( Figure 1 ) showed that the cumulative probabilities of survival at 4 years of follow-up were 38% in the preserved LVEF group, 45% in the mildly reduced LVEF group, and 39% in the moderately reduced LVEF group. In contrast, patients with severely reduced LVEF exhibited significantly lower survival rates throughout follow-up, with a mortality rate of 73% at 4-years of follow-up (log-rank p value <0.001 for the overall difference among the groups). Consistent with the univariate findings, multivariate Cox proportional hazards regression modeling, adjusted for multiple established predictors of mortality, showed that the risk of long-term mortality was not significantly different between patients with HF with preserved LVEF and those with mildly (hazard ratio [HR] 0.92, p = 0.41) or moderately reduced LVEF (HR 1.01, p = 0.90). In contrast, HF patients with severely reduced LVEF experienced a significant 20% (p = 0.04) increased risk for 4-year mortality compared with patients with preserved LVEF ( Table 2 ). Consistent results were obtained after further adjustment for baseline medications including β blockers, ACEI or ARB, calcium channel blockers, statins, diuretics, anticoagulants, insulin, and hypoglycemic, demonstrating a similar risk of long-term mortality between patients with preserved LVEF and those with mild or moderate reductions in LVEF and increased risk among those with severely reduced LVEF (HR 1.22, 95% confidence interval 1.02 to 1.45, p = 0.03).
| Variable | HR (95% CI) | p Value |
|---|---|---|
| LVEF class | 0.04 | |
| Preserved | 1 | |
| Mildly reduced vs preserved | 0.92 (0.76–1.11) | 0.41 |
| Moderately reduced vs preserved | 1.01 (0.85–1.21) | 0.90 |
| Severely reduced vs preserved | 1.20 (1.01–1.44) | 0.04 |
| Age >75 yrs | 1.94 (1.7–2.2) | <0.001 |
| Women | 0.88 (0.77–1.01) | 0.06 |
| Diabetes mellitus | 1.14 (1–1.3) | 0.05 |
| NYHA class III–IV | 1.57 (1.38–1.79) | <0.001 |
| Na <130 mEq/L | 1.53 (1.22–1.92) | <0.001 |
| SBP <140 mmHg | 1.27 (1.11–1.45) | <0.001 |
| eGFR <50 ml/min | 1.68 (1.48–1.92) | <0.001 |
| Hb <11 g/L | 1.24 (1.08–1.42) | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
