Author + information
- Received December 9, 2003
- Revision received February 20, 2004
- Accepted March 16, 2004
- Published online July 7, 2004.
- Peter van der Meer, MD*,* (, )
- Adriaan A Voors, MD*,
- Erik Lipsic, MD*,†,
- Tom D.J Smilde, MD*,
- Wiek H van Gilst, PhD† and
- Dirk J van Veldhuisen, MD, FACC*
- ↵*Reprint requests and correspondence:
Dr. Peter van der Meer, Hanzeplein 1, 9700 RB Groningen, the Netherlands.
Objectives This study aimed to investigate the prognostic importance of plasma erythropoietin (EPO) levels in chronic heart failure (CHF) patients.
Background Anemia is common and is associated with an impaired survival in patients with CHF. Erythropoietin is a hematopoietic growth factor, upregulated in anemic conditions. Little is known about the pathophysiology of anemia in CHF and the prognostic importance of plasma EPO levels in CHF patients.
Methods In 74 patients with CHF (age, 61 ± 2 years; left ventricular ejection fraction, 0.31 ± 0.01; peak oxygen consumption, 19.1 ± 0.6 [mean ± SEM]) and in 15 control patients, hemoglobin levels and plasma concentrations of EPO and brain natriuretic peptide were measured.
Results During a mean follow-up of 3.0 years (range, 2.3 to 5.3 years), 22 patients (30%) died. Anemia was present in 24% of the patients. Multivariate analysis showed that plasma EPO (p = 0.026) and hemoglobin levels (p = 0.005) were independent predictors of survival in this CHF population. We observed only a mild inverse correlation between the logarithm of EPO and hemoglobin levels (r2= 0.08, p = 0.02) in CHF patients, whereas the control group showed a clear significant inverse correlation (r2= 0.44, p = 0.007).
Conclusions Elevated plasma EPO levels are associated with an impaired prognosis independent of hemoglobin levels and other established markers of CHF severity. Furthermore, in the CHF patients, EPO levels poorly correlate with the hemoglobin levels, in contrast with the control group.
Anemia is commonly observed in patients with chronic heart failure (CHF) (1–3). One of the first studies published on anemia and CHF showed in 142 patients with CHF that mean hemoglobin (Hb) concentration decreased from 13.7 g/dl in mild CHF to 10.9 g/dl in severe CHF (4). Several other studies have also examined the prognosis of anemia in patients with CHF (2,3). Hemoglobin levels <12.3 g/dl are associated with worsened symptoms and impaired survival. Although large studies have confirmed these finding, the origin of anemia in CHF remains unclear (5).
In general, in response to anemia, the kidneys produce erythropoietin (EPO), which in turn stimulates red blood cell production (6,7). Previous studies already showed that patients with CHF exhibit elevated levels of EPO (8,9), although the significance of this finding has still to be elucidated. In addition, EPO levels may be correlated with the severity of CHF (8). Therefore, we hypothesized that elevated EPO levels are associated with a poorer prognosis. In the present study, we aimed to establish the prognostic value of both EPO and Hb on mortality in CHF patients.
The prognostic values of EPO and Hb levels were retrospectively assessed in 74 Caucasian CHF patients with stable, mild-to-advanced CHF consecutively admitted to our center (a tertiary referral center), between 1998 and 2000. Referral of patients to our clinic was by general practitioners, cardiologists, local hospitals, or other units of our hospital. Patients were treated for heart failure at the outpatient clinic. Patients were stable on medication for at least three months. In all patients, CHF was diagnosed on the basis of standard criteria, and presence of left ventricular (LV) enlargement or systolic functional impairment by radionuclide ventriculography or echocardiography, according to the European Society of Cardiology guidelines (10). Patients with isolated diastolic dysfunction, valvular disease, myocardial infarction (within 12 weeks), cerebrovascular accident (within 12 weeks), or severe renal failure (creatinine >220 μmol/l) were excluded. To evaluate the EPO response in patients without CHF, we included 15 patients referred to our center with complaints of chest pain or palpitations. Control patients had a mean age of 50.1 ± 4.5 years and had a normal LV function (left ventricular ejection fraction [LVEF] ≥0.60), normal renal function, no clinical signs of inflammation, and no evidence of CHF. All subjects gave informed consent for the protocol, which was approved by the local medical ethics committee. Follow-up data were obtained by review of the medical record or by telephone. The primary end point of the study was all-cause mortality. No cardiac transplantations were observed in this population.
Measurement of brain natriuretic peptide (BNP) and EPO levels
Venous blood samples were taken in the morning to avoid circadian influences. Plasma was stored at −80°C, and BNP concentrations were determined with an immunoradiometric assay (Shionoria, Osaka, Japan). Plasma EPO levels were measured using the IMMULITE EPO assay (DPC, Los Angeles, California), which has been described before (11). The DPC assay consists of a ligand-labeled monoclonal anti-EPO capture antibody, an alkaline phosphatase-labeled polyclonal conjugate antibody, and solid-phase anti-ligand–coated polystyrene beads. The amount of plasma EPO was quantified by chemiluminescent measurement in a luminometer. The assay showed an intra-assay variability of <1%. The impact of age and gender on plasma EPO levels was limited and not significant.
The glomerular filtration rate (GFR) is a standard indicator of renal function. Under steady-state conditions, GFR is estimated from serum creatinine using a formula that accounts for the influence of age and body weight on creatinine production GFR (the Cockroft Gault equation) (12): [140 − age in years) × (body weight in kg)]/(72 × serum creatinine in mg/dl). In women, the value is multiplied by 0.85. This formula is validated and used in several studies of CHF and renal function (13,14).
Data are given as mean ± SEM and as frequencies for categorical variables. We included the following risk factors in our analysis: gender, age, history of hypertension, history of diabetes, etiology of the CHF, LV end-diastolic dimension, calculated GFR, sodium levels, concomitant medication, and severity of CHF assessed by BNP levels, LVEF, and VO2. The Kaplan-Meier method was used to study the influence of several baseline clinical and biochemical variables on survival in the study population. The variables with p < 0.05 in the univariate analysis were used in the multivariate analyses with the use of the Cox backward Wald regression analysis. Hazard ratios (HR) with 95% confidence intervals (CI) demonstrate the risk of death. Pearson and Spearman correlation coefficients were calculated to determine which clinical and biochemical variable had an univariate correlation with log(EPO). All reported probability values were two-tailed, and a p value <0.05 was considered statistically significant. For all statistical analysis, SPSS version 11.0 (SPSS Inc., Chicago, Illinois) was used.
The cohort was 80% male, and age ranged from 26 to 90 years. The New York Heart Association (NYHA) functional class II, III, and IV comprised 37%, 32%, and 31%, respectively. Other baseline characteristics are given in Table 1. Of the 74 patients, 22 (30%) died after 16 to 1,728 days (mean, 621 days), 77.8% of them had a cardiovascular cause of death. Mean follow-up period of the 52 survivors was 1,100 days (range, 844 to 1,934 days; median, 987 days).
Hb and EPO levels
Mean Hb levels were 13.9 g/dl (range, 10.6 to 17.6 g/dl; median, 13.8 g/dl). When we used the generally accepted definition of the World Health Organization (hemoglobin ≤12 g/dl for women, ≤13 g/dl for men), we found that 24.3% of the CHF patients were anemic. Non-survivors had lower Hb levels (p < 0.05) and higher plasma EPO levels than survivors (p < 0.05). When log(EPO) was plotted as a function of Hb, only a modest inverse correlation (r2= 0.08, p = 0.02) was observed in CHF patients (Fig. 1a). To compare the EPO response in non-CHF patients, we obtained blood samples from 15 control patients. Interestingly, we found a clear, significant inverse correlation between log(EPO) and Hb in the control patients (r2= 0.44, p = 0.007) (Fig. 1b).
Furthermore, a modest but statistically significant correlation was observed between plasma EPO levels and the severity of CHF, assessed by BNP levels (r2= 0.14, p < 0.001), NYHA functional class (r2= 0.09, p = 0.01), and VO2(r2= 0.09, p = 0.01). There was no significant correlation between EPO levels and LV end-diastolic dimension and LVEF (r2= 0.003 and r2= 0.04, respectively). Use of medication, including angiotensin-converting enzyme inhibitors, beta-blockers, and diuretics was not significantly correlated to Hb or EPO levels.
Relationship of Hb and EPO with mortality
The Kaplan-Meier survival curve clearly demonstrates increased mortality in patients with anemia, HR 3.28 (95% CI, 1.31 to 8.18; p = 0.01) (Fig. 2). Patients with anemia had a two-year mortality rate of 33%, compared with 14% for non-anemic patients. Furthermore, also elevated EPO levels (≥22.6 mU/ml) are associated with an impaired outcome, HR 2.6 (95% CI, 1.05 to 6.44; p = 0.04) (Fig. 3). Two-year mortality rates were 32% in those with high EPO levels (≥22.6 mU/ml), compared with 16% for those with low EPO levels (<22.6 mU/ml). After adjustment for possible confounders, both Hb and EPO levels remained independently associated with an increased mortality (Table 2).
This is the first study to examine the prognostic value of EPO levels in CHF patients. Our findings demonstrate that increased plasma EPO levels predict an impaired medium- to longer-term survival in patients with CHF. Furthermore, as previously described, lower Hb levels were also associated with increased mortality (2). However, EPO and Hb were both independent prognostic markers. When the correlation between EPO and Hb was further studied in CHF patients and compared with controls, we observed an inadequate increase in EPO levels relative to the hemoglobin levels.
We found that 24% of the CHF patients were anemic, according to World Health Organization criteria, and Hb levels were an independent predictor of all-cause mortality. These findings are well comparable with others studies, which show that anemia is associated with an impaired survival in patients with CHF (3,5,15). A recent study of Horwich et al. (2)also showed that anemia was common in patients with advanced heart failure (NYHA functional class III and IV) and that patients with NYHA functional class IV were more likely to have lower Hb levels. They also showed that Hb levels <12.3 g/dl were associated with an impaired survival in patients with advanced CHF. However, the prognostic value of Hb levels in new cases of heart failure seems to be limited (16). None of these studies assessed the role of endogenous EPO levels in CHF patients.
We observed that elevated plasma EPO levels were associated with severity of CHF. This finding has previously been reported by two other groups (8,9). However, we are the first to show that elevated EPO levels are a prognostic marker for impaired survival in the CHF population. Furthermore, elevated plasma EPO was a risk factor independent of Hb levels.
To further elucidate the effect of EPO on Hb levels in CHF patients, we correlated log(EPO) to Hb levels. Interestingly, we found only a very modest inverse correlation between log(EPO) and Hb levels in CHF patients, while there was a clear correlation in the control group. These findings may indicate a blunted EPO response relative to the Hb levels. This mechanism has already been proposed by others, who suggest that resistance to EPO in the bone marrow may explain the anemia observed in CHF patients (17).
From the modest correlation between EPO and Hb in CHF patients, we hypothesize that sensitivity to EPO may be insufficient in CHF patients. Therefore, one might assume beneficial effects of EPO treatment in CHF patients (18). Several studies already indicated the positive effects of EPO treatment in CHF patients (4,19,20). A study by Silverberg (19)showed that the treatment of anemia in patients with severe CHF, with EPO and intravenous iron, improved cardiac function and quality of life, and markedly reduced hospitalization. Another group observed that EPO treatment in NYHA functional class III to IV patients enhanced exercise capacity and also improved quality of life (20). These studies showed that EPO treatment is both safe and beneficial in anemic CHF patients.
The main limitation of this study was its relatively small size (74 patients, 22 events). Furthermore, we did not measure cytokine levels, which may influence the effect of EPO on Hb levels. We measured EPO and Hb levels on a single point and, thus, can only speculate on its importance over time. Because of these limitations, we regard our study mainly as a hypothesis-generating study. Nevertheless, our findings suggest that lower Hb levels and elevated EPO levels are both independently associated with an impaired survival in CHF patients.
The authors thank Dr. H. L. Hillege for expert statistical advice.
☆ Dr. van der Meer is supported by NWO ZonMW. Dr. Lipsic is supported by GUIDE. Dr. van Veldhuisen is an established investigator of the Netherlands Heart Foundation (grant D97-017).
- brain natriuretic peptide
- chronic heart failure
- confidence interval
- glomerular filtration rate
- hazard ratio
- left ventricular
- left ventricular ejection fraction
- New York Heart Association
- Received December 9, 2003.
- Revision received February 20, 2004.
- Accepted March 16, 2004.
- American College of Cardiology Foundation
- Cromie N,
- Lee C,
- Struthers A.D
- Horwich T.B,
- Fonarow G.C,
- Hamilton M.A,
- MacLellan W.R,
- Borenstein J
- Mozaffarian D,
- Nye R,
- Levy W.C
- Silverberg D.S,
- Wexler D,
- Blum M,
- et al.
- Ezekowitz J.A,
- McAlister F.A,
- Armstrong P.W
- Remme W.J,
- Swedberg K
- Hillege H.L,
- Girbes A.R,
- de Kam P.J,
- et al.
- Al-Ahmad A,
- Rand W.M,
- Manjunath G,
- et al.
- Van der Meer P,
- Voors A.A,
- Lipsic E,
- van Gilst W.H,
- Van Veldhuisen D.J
- Silverberg D.S,
- Wexler D,
- Sheps D,
- et al.
- Mancini D.M,
- Katz S.D,
- Lang C.C,
- LaManca J,
- Hudaihed A,
- Androne A.S