Author + information
- Received December 28, 2001
- Revision received March 4, 2002
- Accepted March 11, 2002
- Published online June 5, 2002.
- Tamara B Horwich, MD*,
- Gregg C Fonarow, MD, FACC†,* (, )
- Michele A Hamilton, MD, FACC†,
- W.Robb MacLellan, MD, FACC† and
- Jeff Borenstein, MD‡
- ↵*Reprint requests and correspondence:
Dr. Gregg C. Fonarow, Ahmanson–UCLA Cardiomyopathy Center, UCLA Division of Cardiology, 47-123 CHS, 10833 Le Conte Avenue, Los Angeles, CA 90095-1679, USA.
Objectives This study aimed to evaluate the relationship between anemia and heart failure (HF) prognosis.
Background Although it is known that chronic diseases, including HF, may be associated with anemia, the impact of hemoglobin (Hb) level on symptoms and survival in HF has not been fully defined.
Methods We analyzed a cohort of 1,061 patients with advanced HF (New York Heart Association [NYHA] functional class III or IV and left ventricular ejection fraction [LVEF] <40%) referred to a single center for evaluation and management. The Hb level was drawn at time of initial evaluation. Patients were divided into quartiles of Hb: Hb <12.3; Hb 12.3 to 13.6; Hb 13.7 to 14.8; Hb >14.8 g/dl.
Results Mean Hb was 13.6, and values ranged from 7.1 to 19.0 g/dl. The Hb groups were similar in age, medication profile, LVEF, hypertension, diabetes, smoking status and serum sodium. Lower Hb was associated with an impaired hemodynamic profile, higher blood urea nitrogen and creatinine, and lower albumin, total cholesterol and body mass index. Patients in the lower Hb quartiles were more likely to be NYHA functional class IV (p < 0.0001) and have lower peak oxygen consumption (PKVo2) (p < 0.0001). Survival at one year was higher with increased Hb quartile (55.6%, 63.9%, 71.4% and 74.4% for quartiles 1, 2, 3 and 4, respectively). On multivariate analysis adjusting for known HF prognostic factors, low Hb proved to be an independent predictor of mortality (relative risk 1.131, confidence interval 1.045 to 1.224 for each decrease of 1 g/dl).
Conclusions In chronic HF, relatively mild degrees of anemia are associated with worsened symptoms, functional status and survival.
Heart failure (HF) is an increasingly important cause of morbidity and mortality, with prevalence in the U.S. recently estimated near 5 million. Although deaths from myocardial infarction (MI) and stroke have decreased by 50% over the past decade, mortality from HF has been steadily rising despite advances in medical and surgical therapy (1). Identification of modifiable risk factors for HF morbidity and mortality may lead to improvement in the clinical management of HF.
Recent reports have suggested that mild to moderate anemia is a prevalent condition in the HF patient population (2,3), and severe, chronic anemia has been associated with the de novo development of HF (4). However, the impact of the anemic state on symptoms and prognosis in patients with established HF has not been well-described in the medical literature. Current guidelines and reviews do not include anemia as a prognostic factor or treatment goal in HF (5,6).
The effect of hemoglobin (Hb) level on HF outcomes has been studied in certain patient populations. Anemia was a risk factor for development of HF as well as rehospitalization for HF in patients with end-stage renal disease (ESRD) (7). In patients with asymptomatic left ventricular dysfunction or mild to moderate HF, a recent analysis of patients enrolled in the Studies Of Left Ventricular Dysfunction (SOLVD) identified low Hb as a predictor of mortality independent of renal insufficiency (2). However, in HF patients who present the most significant management challenge to physicians, those with New York Heart Association (NYHA) functional class III and class IV, the interaction among Hb level, symptoms and prognosis is unknown.
Recently, several small clinical studies have investigated the correction of anemia with erythropoietin and iron as a treatment for HF. These initial studies have shown favorable results including improvement in left ventricular ejection fraction (LVEF), NYHA functional class and exercise capacity (8–10). Before larger trials are undertaken, it is necessary to have a more complete understanding of anemia’s role in the progression of advanced HF. To explore this issue further, we examined the relationship among Hb level, HF patient characteristics and symptoms, and HF survival in a cohort of NYHA functional class III and class IV HF patients of multiple etiologies.
The study subjects were 1,733 patients referred to a single university medical center for heart transplant evaluation between 1983 and 1999. All subjects were followed in a comprehensive HF management program, as previously described (11). Medical record review was approved by the University of California-Los Angeles, Medical Institutional Review Board. Patients with LVEF >40% (n = 106), NYHA functional class <3 (n = 58) and those without an initial Hb level, “dry” weight or adequate follow-up data (n = 508) were excluded from analysis. The final study population consisted of 1,061 subjects.
Both Hb and hematocrit (Hct) levels were determined at time of initial presentation. All specimens were analyzed in Centers for Disease Control-approved laboratories. The NYHA functional class was assessed at the time of initial presentation. Medical treatments recorded were those implemented after baseline hemodynamic evaluation. Hemodynamic variables utilized in the analyses were the optimal values recorded after pulmonary artery catheter-tailored medical therapy, as these hemodynamic measurements have been shown to best correlate with survival (12). Laboratory testing, electrocardiography, echocardiography and cardiopulmonary exercise tests all occurred within three months of initial referral; later values were excluded from our analysis. Creatinine clearance was calculated by the Cockcroft-Gault formula using the patient’s “dry” weight after hemodynamically guided therapy. Hypertension, diabetes and smoking histories were based on medical record review.
Death was the primary end point in this study. Deaths were classified as sudden death, HF death or death secondary to MI. Death was considered sudden if it was unexpected based on the patient’s clinical status and if it occurred out of the hospital within 15 min of the onset of unexpected symptoms or during sleep. Death during hospitalization for worsening congestive symptoms was considered an HF death. Urgent heart transplants (Status I) were analyzed as deaths, under the assumption that these patients would have died without a transplant. Nonurgent transplants (Status II) were considered a nonfatal end of follow-up.
Data are presented as mean ± SD for continuous variables and as frequencies for categorical variables. Patients were divided into quartiles of Hb. Differences in baseline characteristics among quartiles were analyzed using analysis of variance for continuous variables and Pearson’s chi-square test for categorical variables. To evaluate differences between survivors and nonsurvivors, we used an independent sample ttest for continuous variables and Pearson’s chi-square test for categorical variables. Both one-year and five-year survival curves were calculated with the Kaplan-Meier method, and differences between the curves were evaluated with the log-rank statistic. We assessed the relationship between baseline variables and mortality using a Cox proportional hazards survival model (SPSS for Windows, version 10.0.5). Hazard ratios (relative risk [RR]) with 95% confidence intervals (CI) demonstrate the risk of death when a variable is present. The multivariate Cox model included the following variables seen to be predictors of mortality on univariate analysis as well as other known factors for HF mortality: age, gender, body mass index (BMI), left ventricular end diastolic dimension, hypertension history, diabetes history, smoking history, serum sodium, albumin, creatinine and HF etiology.
Baseline characteristics of the cohort
The cohort was 77% male, and ages ranged from 16 to 82 years. The NYHA functional class III and IV patients comprised 35% and 65% of the population, respectively, and mean LVEF was 22%. Etiologies of HF were ischemic (50%), idiopathic (38%) and valvular (4.5%); the remaining 7.5% included alcohol-induced, hypertrophic and postpartum cardiomyopathy.
Mean Hb was 13.6 ± 1.9 g/dl and Hb ranged from 7.1 to 19.0 g/dl. The mean Hb level for women was 12.7 ± 1.8 g/dl and for men was 13.8 ± 1.9 g/dl. Of both men and women in the cohort, 30% were considered anemic when anemia was defined as Hb <13 g/dl in men and Hb <12 g/dl in women (13). Hemoglobin level >17 g/dl was seen in 3% of men and 1% of women.
Relationship between hb level and baseline characteristics
Differences among the Hb quartiles are detailed in Table 1. Age, past medical history, LVEF, etiology of HF, serum sodium and medication usage were similar among the Hb groups. Patients in the lower quartiles were more likely to be women. Lower Hb levels were associated with greater symptoms, as evident by more patients being NYHA functional class IV (75.3%, 68.0%, 56.7% and 59.1% for quartiles 1, 2, 3 and 4, respectively, p < 0.0001). Lower BMI and albumin and impaired renal function were seen in the lower Hb quartiles. The hemodynamic profile of the patients in the lower quartiles were characterized by lower blood pressure, but higher heart rate and pulmonary capillary wedge pressure (Table 1). Greater impairment in exercise capacity was demonstrated in HF patients with lower Hb concentrations. Peak oxygen consumption (PKVo2) on cardiopulmonary exercise testing was 12.8 ± 4.6, 12.4 ± 4.5, 13.8 ± 4.7 and 14.6 ± 5.4 ml/kg per min for quartiles 1, 2, 3 and 4, respectively, p < 0.0001.
Relationship between hb and mortality
There were 212 deaths by year one, and 360 deaths at five-year follow-up. At one year, progressive HF death accounted for 98 deaths (46%), while 89 deaths were sudden, 5 occurred secondary to MI and 20 occurred from unknown or other causes. At the end of one year, 247 of the 1,061 patients had received heart transplants (126 urgent, Status 1; 148 nonurgent, Status 2). At five-year follow-up, 405 patients received heart transplants.
A low baseline Hb level in this cohort of advanced HF patients proved to be a significant predictor of subsequent mortality. The Hb level was significantly higher in patients alive at one year compared to those who had died or had undergone urgent transplant at one year. Table 2details the differences in baseline characteristics between survivors and nonsurvivors at one year. On univariate Cox regression analysis, each 1 g/dl decrease in Hb was associated with a 16% increased risk of death. Hemoglobin was similarly predictive of mortality at five-year follow-up (data not shown). Survival rates steadily declined as Hb quartile decreased, as shown in the Kaplan-Meier survival curves for the four Hb quartiles (Fig. 1). This stepwise trend was preserved after the cohort was further subdivided by Hb deciles, with no evidence of a U-shaped relationship (Fig. 2).
After adjusting for variables found to be significant predictors of mortality on univariate analysis, Hb level maintained its significant prognostic value. On multivariate analysis, the risk of mortality increased by 13% for each decrease by 1 g/dl in Hb (RR 1.131, CI 1.045 to 1.224). Independent predictors of mortality, as determined by multivariate analysis, are detailed in Table 3.
Kaplan-Meier survival curves were recalculated in clinically significant subgroups to determine whether Hb maintained its prognostic value in these subsets of HF patients. The association between low Hb and increased mortality was preserved in subsets of patients with and without coronary artery disease (CAD) as well as patients both older and younger than the median age of 53 years (Fig. 3). Because healthy women have lower Hb levels than healthy men, we separated each gender group into quartiles before recalculating survival curves. Hemoglobin was predictive of mortality in both men and women, yet the association was stronger in men (Fig. 3). The relationship between Hb and mode of death was also investigated, revealing that Hb was a significant predictor of progressive HF death (p = 0.00001), but not sudden death (p = 0.73). After exclusion of patients who received heart transplants, higher mortality rates were still observed in patients with lower Hb (p < 0.005) (Fig. 3). On multivariate analysis, lower Hb remained a significant predictor of increased mortality (RR 1.129, CI 1.018 to 1.251).
Our study, which examined Hb level in a large, well-characterized cohort of advanced HF patients, demonstrates that Hb is significantly associated with symptoms, exercise capacity, and prognosis in patients with advanced HF. Several findings deserve emphasis. First, although the relationship between anemia and adverse outcomes has been described in patients with ESRD (7)and patients with asymptomatic left ventricular dysfunction or mild to moderate HF (2), our study extends this finding to patients with advanced HF. The relationship between anemia and mortality persists in subgroups based on gender, age and HF etiology, and it is independent of other established HF prognostic factors. Furthermore, based on our analysis of Hb deciles, we have observed that mortality decreases in linear fashion as Hb level increases, without excess mortality at the highest Hb levels.
A second point of emphasis is the relatively mild degrees of anemia associated with increased morbidity and mortality; significantly impaired survival was seen in women with Hb <11.6 g/dl and men with Hb <12.6 g/dl, values that may not be considered significant or associated with increased risk in clinical practice. Third, patients with decreased Hb were characterized by an unfavorable hemodynamic profile, worse symptoms of HF, as described by NYHA functional class, and greater impairment in exercise capacity, as quantified by PKVo2. Based on this analysis it would be predicted that correction of anemia would be associated with an improvement in exercise capacity of approximately 2 ml/kg/min, which has previously been shown to be clinically relevant (14). Interestingly, that was the magnitude of improvement in exercise capacity recently observed in a small trial of 12 weeks of erythropoietin therapy in patients with anemia and HF (10).
Prior epidemiologic studies that have included Hb or Hct in their analyses provide supportive evidence for the significant relationship between anemia, HF and prognosis. Anemia was associated with an increased risk of death or HF rehospitalization in California patients hospitalized for HF between 1991 and 1992 (15). Furthermore, the Framingham Heart Study found Hct to be a significant risk factor for increased HF incidence (16).
Although the present study describes a remarkable association between anemia and HF mortality, the pathophysiologic relationship between Hb and HF progression requires further study. Hemoglobin may be a marker of poor prognosis in HF or, conversely, it may play a causative role in HF progression. Several potential explanations for the association between low Hb and poor prognosis deserve exploration.
Anemia of chronic disease
Increased levels of plasma cytokines such as tumor necrosis factor in HF patients are associated with higher NYHA functional class (17)and increased HF mortality (18). Anemia in advanced HF may be a marker of increased levels of circulating cytokines and cytokine receptors. It is known that anemia commonly seen in chronic inflammatory states, such as some infections, neoplasm and rheumatologic conditions, is mediated by inflammatory cytokines such as tumor necrosis factor, interleukin-1 and the interferons (19). It may be that the anemia in HF is similar to that seen in other chronic diseases and is likewise mediated by elevated circulating cytokines. As with anemia of other chronic disease states, increased erythropoietin levels have been found in patients with chronic heart failure (20).
In our study, patients in the lower Hb quartiles were characterized by markers of malnutrition, such as lower levels of albumin and lower BMI (Table 1). Malnutrition is a common causative factor in the development of anemia in non-HF populations, such as the elderly (21). It is possible that the poor prognosis of anemia in HF stems from malnutrition due to cardiac cachexia, a catabolic state seen in advanced HF and associated with increased HF mortality (22). However, Hb was predictive of mortality independent of BMI and albumin levels.
Low Hb may be a marker of volume overload in poorly compensated, high-risk HF patients. Previous studies have demonstrated that decreased Hb or Hct in HF is reflective of plasma volume expansion and not necessarily indicative of decreased red blood cell volume (23). Because Hb was predictive of mortality independent of pulmonary capillary wedge pressure after hemodynamically guided diuresis, hemodilution alone cannot fully account for these findings.
Anemia and ischemia
Although it has been shown that the resting, healthy human heart can withstand acute, severe anemia without sustaining myocardial ischemia (24), the presence of CAD may significantly impair the heart’s ability to tolerate low levels of Hb (25). A recent observational study showed reduction in 30-day mortality in patients with Hct <33% who received blood transfusion during hospitalization for MI compared to those who were not transfused (26). Anemia in HF patients may predispose to myocardial ischemia, repetitive stunning, apoptosis, and necrosis, thus contributing to the progression of ventricular dilation and clinical HF.
The kidney and its related hormonal mechanisms play a fundamental role in the pathophysiology of HF. Impaired renal function is associated with both anemia and worse prognosis in patients with HF. Although lower Hb was associated with higher blood urea nitrogen and creatinine and lower creatinine clearance, the increased mortality risk seen remained after adjustment for renal function and other HF prognostic factors. Similarly, in the SOLVD analysis the increased mortality risk with anemia in HF was independent of renal function impairment (2). It should be noted, however, that in a randomized clinical trial involving patients with ESRD and HF, normalization of Hct (42%) with erythropoietin and intravenous (IV) iron dextran compared to maintaining lower Hct levels (30%) was associated with a trend for increased nonfatal MIs and mortality (RR 1.3; 95% CI 0.9 to 1.9, p = NS) (27).
Our study is retrospective and examines a selected population of HF patients with advanced disease referred for transplant evaluation. Hemoglobin levels were assessed at a single point in time, and thus we cannot comment on the importance in change in Hb over time. Treatment of anemia with transfusion or other medications during the time course of the study is also not documented or included in the analysis, though it was not routine to treat these patients with transfusions, iron or erythropoietin. Data on the use of beta-blockers or addition of medicines after the time of initial presentation are not included in the analysis. We also do not have data on cytokine or erythropoietin levels, or direct measures of plasma volume or red blood cell mass, information that would help in understanding the pathophysiologic role of anemia in HF progression.
Lower Hb is associated with greater functional impairment, worse exercise capacity, and increased mortality in this cohort of advanced HF patients of multiple etiologies. Further studies aimed at understanding the interaction between Hb and HF progression are needed. Initial studies suggest that treatment of anemia is beneficial in HF (8). Additional randomized, placebo-controlled studies are warranted to test the hypothesis that correction of anemia improves symptoms and exercise capacity in advanced HF and also to determine whether treatment of anemia has an impact on HF survival.
☆ This study was supported by the Ahmanson Foundation, Los Angeles, California, and Amgen, Thousand Oaks, California.
- body mass index
- coronary artery disease
- confidence interval
- end-stage renal disease
- heart failure
- left ventricular ejection fraction
- myocardial infarction
- New York Heart Association
- peak oxygen consumption
- relative risk
- Studies Of Left Ventricular Dysfunction
- Received December 28, 2001.
- Revision received March 4, 2002.
- Accepted March 11, 2002.
- American College of Cardiology Foundation
- Al-Ahmad A.,
- Rand S.M.,
- Manjunath G.,
- et al.
- Stone R.M.,
- Ridges K.R.,
- Libby P.
- Cowburn P.J.,
- Cleland J.G.,
- Coats A.S.,
- et al.
- Hunt S.A.,
- Baker D.W.,
- Chin M.H.,
- et al.
- Silverberg D.S.,
- Wexler D.,
- Sheps D.,
- et al.
- Silverberg D.S.,
- Wexler D.,
- Blum M.,
- et al.
- Mancini D.,
- Katz S.,
- Lamanca J.,
- et al.
- Fonarow G.C.,
- Stevenson L.W.,
- Walden J.A.,
- et al.
- ↵(1968) World Health Organization. Nutritional anaemias: Report of a WHO scientific group (World Health Organization, Geneva, Switzerland).
- Stevenson L.W.,
- Steimle A.E.,
- Fonarow G.C.,
- et al.
- Kannel W.B.
- Rauchhaus M.,
- Doehner W.,
- Francis D.P.,
- et al.
- Deswal A.,
- Petersen N.J.,
- Feldman A.M.,
- Young J.B.,
- White B.G.,
- Mann D.L.
- Androne A.S.,
- Hryniewiz K.,
- Hudaihed A.,
- Mancini D.,
- Katz S.D.
- Levy P.S.,
- Kim S.J.,
- Eckel P.K.,
- et al.