Disordered Iron Homeostasis in Chronic Heart Failure

Journal of the American College of Cardiology © 2011 by the American College of Cardiology Foundation Published by Elsevier Inc.

Vol. 58, No. 12, 2011 ISSN 0735-1097/$36.00 doi:10.1016/j.jacc.2011.04.040

Heart Failure

Disordered Iron Homeostasis in Chronic Heart Failure Prevalence, Predictors, and Relation to Anemia, Exercise Capacity, and Survival Darlington O. Okonko, MBBS,* Amit K. J. Mandal, MBBS,† Constantinos G. Missouris, MD,† Philip A. Poole-Wilson, MD, FMEDSCI* London and Slough, United Kingdom Objectives

The aim of this study was to comprehensively delineate iron metabolism and its implications in patients with chronic heart failure (CHF).

Background

Iron deficiency is an emerging therapeutic target in CHF.

Methods

Iron and clinical indexes were quantified in 157 patients with CHF.

Results

Several observations were made. First, iron homeostasis was deranged in anemic and nonanemic subjects and characterized by diminished circulating (transferrin saturation) and functional (mean cell hemoglobin concentration) iron status in the face of seemingly adequate stores (ferritin). Second, while iron overload and elevated iron stores were rare (1%), iron deficiency (transferrin saturation ⬍20%) was evident in 43% of patients. Third, disordered iron homeostasis related closely to worsening inflammation and disease severity and strongly predicted lower hemoglobin levels independently of age, sex, erythrocyte sedimentation rate, New York Heart Association (NYHA) functional class, and creatinine. Fourth, the etiologies of anemia varied with disease severity, with an iron-deficient substrate (anemia of chronic disease and/or iron-deficiency anemia) evident in 16%, 72%, and 100% of anemic NYHA functional class I or II, III, and IV patients, respectively. Although anemia of chronic disease was more prevalent than iron-deficiency anemia, both conditions coexisted in 17% of subjects. Fifth, iron deficiency was associated with lower peak oxygen consumption and higher ratios of ventilation to carbon dioxide production and identified those at enhanced risk for death (hazard ratio: 3.38; 95% confidence interval: 1.48 to 7.72; p ⫽ 0.004) independently of hemoglobin. Nonanemic iron-deficient patients had a 2-fold greater risk for death than anemic iron-replete subjects.

Conclusions

Disordered iron homeostasis in patients with CHF relates to impaired exercise capacity and survival and appears prognostically more ominous than anemia. (J Am Coll Cardiol 2011;58:1241–51) © 2011 by the American College of Cardiology Foundation

Despite optimal conventional therapy, many patients with chronic heart failure (CHF) remain symptom limited, exercise intolerant, and subject to high rates of mortality. See page 1252

The persistence of such poor outcomes suggests that additional adverse phenomena remain unmitigated by current From *Clinical Cardiology, National Heart & Lung Institute, Imperial College London, London, United Kingdom; and the †Department of Cardiology, Wexham Park Hospital, Slough, United Kingdom. The British Heart Foundation (London, United Kingdom) supported Dr. Okonko (FS/03/104/16341) and Prof. PooleWilson. The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Manuscript received October 18, 2010; revised manuscript received April 14, 2011, accepted April 19, 2011.

therapeutic strategies and endure to accelerate disease progression. Systemic iron homeostasis is frequently deranged in chronic disorders (1–5), but the integrity of iron metabolism in CHF is poorly characterized. Prior analyses are sparse and have largely been conducted in anemia patients (6,7) or based solely on serum iron and ferritin concentrations (8 –10). However, total body iron is distributed across circulating, functional, and storage iron compartments that can be perturbed independently of one another (11–13). Because iron indexes typically reflect individual compartments and are influenced by factors other than iron status, multiple markers are generally needed to adequately gauge iron homeostasis (11,12). To date, no study has comprehensively delineated iron metabolism in anemic and nonanemic patients

1242

Okonko et al. Iron Homeostasis in Heart Failure

with CHF, and none have reported prevalence estimates for the full spectrum of iron disorACD ⴝ anemia of chronic ders. Moreover, the diagnostic disease utility of iron markers in anemic CHF ⴝ chronic heart failure patients with CHF has so far CI ⴝ confidence interval been poorly explored. CRP ⴝ C-reactive protein Iron is obligatory for optimal ESR ⴝ erythrocyte erythrocyte production, and ironsedimentation rate deficient erythropoiesis (IDE) due FERRIC-HF ⴝ Ferric Iron to defectively mobilized (funcSucrose in Heart Failure tional deficiency ⫽ anemia of HR ⴝ hazard ratio chronic disease [ACD]) and/or IDA ⴝ iron deficiency depleted (absolute deficiency ⫽ anemia iron deficiency anemia [IDA]) IDE ⴝ iron-deficient stores is the most common cause erythropoiesis of anemia globally (14 –17). HowIQR ⴝ interquartile range ever, despite the high burden and MCHC ⴝ mean cell adverse implications of anemia in hemoglobin concentration CHF patients (18), the use of conNYHA ⴝ New York Heart ventional iron indexes to distinAssociation guish ACD from IDA and to %HRC ⴝ percent identify IDA in the midst of ACD hypochromic red cells (ACD ⫹ IDA) has not been resTFR ⴝ soluble transferrin ported. Additionally, it is unclear receptor whether the etiologies of anemia TIBC ⴝ total iron-binding vary with disease severity or capacity whether iron deficiency has conseTSAT ⴝ transferrin quences in CHF besides anemia. saturation Iron is quantitatively the most VCO2 ⴝ carbon dioxide production important biocatalyst in human physiology, with critical roles VE ⴝ minute ventilation beyond hemoglobin synthesis (19 –25). Even in the absence of anemia, iron deficiency is a potent substrate for dyspnea and exercise intolerance (26,27) and identifies those in the general population at a heightened risk for mortality (28). In patients with uremia, lower levels of transferrin saturation (TSAT), a more reliable marker of iron deprivation than ferritin in inflammatory cohorts, also correlate to poorer survival (29). Whether lower TSAT relates to mortality independently and more powerfully than anemia in CHF remains unclear. In a broad spectrum of patients with systolic CHF, we sought to characterize: 1) iron metabolism and its predictors; 2) the prevalence of the full spectrum of iron disorders, with an emphasis on discriminating between ACD, IDA, and ACD ⫹ IDA; 3) variations in the etiologies of anemia with New York Heart Association (NYHA) functional class; and 4) the relation of iron deficiency to exercise performance and survival. Abbreviations and Acronyms

Methods Study population. We prospectively appraised clinical and iron indexes in 157 consecutively eligible patients with CHF

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

who attended dedicated heart failure clinics and wards at Wexham Park Hospital (Slough, United Kingdom) and the Royal Brompton Hospital (London, United Kingdom). Control subjects (n ⫽ 22) with no known medical problems or regular medications were also recruited. The diagnosis of CHF was based on a ⬎6-month history of appropriate symptoms and signs and a left ventricular ejection fraction ⱕ45%. Patients with recent myocardial infarctions, those with ongoing non-CHF inflammatory processes, and those using iron, erythropoietic, immunomodulatory, or renal replacement therapies within 3 months of recruitment were excluded. Informed consent was obtained from all subjects, and the local ethics and research governance committees at both institutions approved the study. Laboratory measurements. Peripheral venous blood samples were obtained in the morning from all participants. Full blood counts were measured from K3 ethylenediaminetetraacetic acid samples initially using a Coulter S-Plus Jr. electronic counter (Coulter Electronics, High Wycombe, United Kingdom) at Wexham Park and an ADVIA 120 analyzer (Bayer Diagnostics, Berkshire, United Kingdom) at Royal Brompton. Because of contractual issues, the ADVIA 120 was substituted by a Coulter S-Plus Jr. counter halfway through the study. Percent hypochromic red cells (%HRC) and red cell distribution width were assayed flow cytometrically using the ADVIA 120. Serum iron and total iron-binding capacity (TIBC) were measured colorimetrically (Alpkem RFA analyzer, Alpkem Corporation, Clackamas, Oregon). TSAT was calculated as 100 ⫻ serum iron/TIBC. Serum ferritin was determined using a Cobas Core analyzer (Roche Diagnostics GmbH, Mannheim, Germany). For soluble transferrin receptor (sTFR) concentrations, sera were spun (3,000 rpm for 15 min), separated, and stored immediately at ⫺80°C for later quantification using an enzyme-linked immunosorbent assay (Dade Behring GmbH, Marburg, Germany) with a sensitivity of 0.3 pg/ml. All other biochemical tests were performed using standard techniques. Definitions of anemia and iron deficiency. Anemia was defined as hemoglobin ⬍13 g/dl in men and ⬍12 g/dl in women (30). Iron deficiency was principally defined as TSAT ⬍20% because TSAT is a more reliable marker of iron status than ferritin during inflammation (12,31,32), and diminished circulating iron status is a feature of both absolute (TSAT ⬍20%, low ferritin) and functional (TSAT ⬍20%, normal or high ferritin) iron deficiency. Because ferritin overestimates iron stores in inflammatory cohorts (32,33) prevalence estimates for absolute and functional iron deficiency are given using conventional (30 ␮g/l) (1,12) and pragmatic (100 ␮g/l) (34,35) cutoffs. Prevalence of distinct iron disorders and the origins of anemia. To estimate the prevalence of distinct iron disorders and to discriminate the cause(s) of anemia, we adopted a multiple-measures approach based on TSAT, ferritin, and TIBC or sTFR. Patients with anemia with TSAT ⬍20% had pure IDA if ferritin levels were ⬍30 ␮g/l (1,12).

Okonko et al. Iron Homeostasis in Heart Failure

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

Ferritin levels ⱖ30 ␮g/l identified those with ACD with or without coexistent IDA (1). Because TIBC and sTFR concentrations are augmented by iron store depletion but not by inflammation (11,12,36), their elevated levels were used to identify patients with ACD and concomitant IDA. Two multiple-measures models were constructed. The first used only clinically available indexes (TSAT, ferritin, and TIBC) and was therefore termed the clinical model. The second used TSAT, ferritin, and sTFR and was termed the research model, because sTFR assays are principally research tools. The research model served to validate the results of the clinical model in anemic subjects. Cardiopulmonary exercise testing. Patients underwent a modified Naughton (37) symptom-limited treadmill exercise test. Minute ventilation (VE), oxygen consumption, and carbon dioxide production (Vco2) were monitored on a breath-by-breath basis using a respiratory mass spectrometer (Amis 2000; Innovision, Copenhagen, Denmark). Peak oxygen consumption adjusted for body weight (milliters per minute per kilogram) and the slope of the relation between VE and Vco2 (VE/Vco2) were calculated as previously described (37). Statistical analysis. Data are presented as mean ⫾ SD, frequency (percent), or median (interquartile range [IQR]). Intergroup comparisons were made using Student t test, Pearson chi-square test, the Mann-Whitney U test, 1-way analysis of variance with Fisher protected least significant difference post hoc testing, or nonparametric Kruskal-Wallis analysis of variance as appropriate. To validate TSAT ⬍20% as a marker of iron deficiency, receiver-operating characteristics curve analysis was performed to determine which TSAT or ferritin cutoffs provided optimal sensitivity and specificity for predicting sTFR levels ⱖ1.76 mg/l (best noninvasive marker of cellular iron deprivation) (12,36). Relations between variables were assessed by multiple linear regression using age, sex, ethnicity, CHF etiology, left ventricular ejection fraction, NYHA functional class, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), creatinine, platelet counts, beta-blocker use, and antiplatelet or anticoagulant drug use as covariates. Prognostic associations were determined using Cox proportionalhazards analyses (univariate, bivariate, and trivariate models) only for covariates that were available in all patients. To assess the interaction between iron deficiency and anemia on survival, patients were stratified into 4 groups (nonanemic iron-replete, anemic iron-replete, nonanemic iron-deficient, and anemic iron-deficient). Kaplan-Meier curves for cumulative survival were constructed for each group and differences in survival rates tested using the Cox-Mantel log-rank test. All-cause mortality status was obtained from direct physical or telephone contact with patients or their relatives, from the hospital information systems, and from the U.K. Office of National Statistics. No patient was lost to followup, and the minimum duration of follow-up for all survivors was 1.4 years.

1243

Data were missing for certain indexes, with no imputation performed. Sensitivity analyses revealed no significant difference in baseline characteristics between those who were and were not quantified for each index. Data were analyzed using StatView version 4.5 (Abacus Concepts Inc., Berkeley, California). A p value ⬍0.05 was considered statistically significant, without adjustment for multiple comparisons. Results Baseline characteristics. Patients were similar to controls with respect to age, sex, and ethnicity, but had significantly lower hemoglobin and higher creatinine levels (Table 1). Iron deficiency (TSAT ⬍20%) was evident in 68 patients (43%). Iron-deficient patients did not differ from ironreplete subjects with respect to antiplatelet (p ⫽ 0.92) or anticoagulant (p ⫽ 0.17) drug use or platelet counts, but they had higher NYHA functional class designations, CRP, and white cell counts, and lower hemoglobin and betablocker use. Abnormal iron metabolism in CHF. DIMINISHED CIRCULATING IRON STATUS. Male (22.9 ⫾ 9.5%) and female (20.4 ⫾ 10.4%) patients had similar TSAT (p ⫽ 0.15) (Fig. 1A). Compared with controls, patients with CHF exhibited lower TSAT and serum iron and higher TIBC and sTFR levels (Table 2). Circulating deficits worsened with NYHA functional class and were greater in patients with anemia (Fig. 2A). The mean cell hemoglobin concentration (MCHC) was higher in male (33.1 ⫾ 1.1 g/dl) than in female (32.4 ⫾ 1.5 g/dl) patients (p ⫽ 0.005) (Fig. 1B). Compared with controls, patients with CHF exhibited reductions in MCHC and hemoglobin and increases in %HRC and red cell distribution widths (Table 2). Functional deficits worsened with NYHA functional class and were more marked in patients with anemia (Fig. 2B).

ATTENUATED FUNCTIONAL IRON STATUS.

Median ferritin levels were higher in male (144 ␮g/l; IQR: 48 to 193 ␮g/l) than female (88.9 ␮g/l; IQR: 29 to 118 ␮g/l) patients (p ⫽ 0.01) (Fig. 1C). On comparing all patients with CHF with controls, no significant difference in ferritin or sTFR/ log ferritin ratio was found. However, on stratifying for NYHA functional class, there was a progressive decline in ferritin levels from patients in class I or II (101 ␮g/l; IQR: 54 to 216 ␮g/l) to controls (85 ␮g/l; IQR: 47 to 137 ␮g/l) to patients in class III (70 ␮g/l; IQR: 40 to 123 ␮g/l) and then patients in class IV (59 ␮g/l; IQR: 21 to 121 ␮g/l; analysis-of-variance p ⫽ 0.04). Similarly, the sTFR/log ferritin ratio progressively increased from controls to patients in NYHA functional class I or II, class III, and then class IV (p ⫽ 0.05) (Fig. 2C). Predictors of impaired iron homeostasis. For the circulating iron pool, higher CRP and NYHA functional class SEEMINGLY PRESERVED STORAGE IRON STATUS.

1244

Okonko et al. Iron Homeostasis in Heart Failure

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

Baseline Table 1 Characteristics Baseline Characteristics Patients With CHF

Variable Age (yrs)

Control Subjects (n ⴝ 22) 66 ⫾ 9

All (n ⴝ 157) 71 ⫾ 12

TSAT >20% (n ⴝ 89)

TSAT <20% (n ⴝ 68)

70 ⫾ 10

72 ⫾ 14

Men

68%

72%

79%

63%*

Caucasian

82%

87%

89%

84%

Ischemic cause

—

NYHA functional class

—

64% 2.6 ⫾ 0.8

63% 2.3 ⫾ 0.7

67% 3.0 ⫾ 0.8*

I/II

—

45%

66%

18%

III

—

41%

30%

54%

IV

—

14%

3%

—

32 ⫾ 9

34 ⫾ 9

30 ⫾ 10 125 ⫾ 59

LVEF (%) Creatinine (␮mol/l)

28%

86 ⫾ 21

122 ⫾ 49†

119 ⫾ 40

ESR (mm/h)‡

8⫾6

26 ⫾ 27

20 ⫾ 23

33 ⫾ 31

CRP (mg/l)‡

6 (5–7)

7 (6–10)

12 (7–43)*

8 (6–19)

14.6 ⫾ 1.3

13.1 ⫾ 1.7†

13.7 ⫾ 1.6

12.3 ⫾ 1.6*

Mean cell volume (fl)

90 ⫾ 4

91 ⫾ 6

92 ⫾ 6

89 ⫾ 6*

Mean cell hemoglobin (pg)

30 ⫾ 1

30 ⫾ 2

31 ⫾ 2

29 ⫾ 2*

34.1 ⫾ 1.1

32.9 ⫾ 1.3†

32.6 ⫾ 1.4

33.3 ⫾ 1.2*

Hemoglobin (g/dl)

MCHC (g/dl) White cell count (⫻109/l) Platelet count (⫻109/l)

6.8 ⫾ 1.4

7.9 ⫾ 3.4

7.4 ⫾ 1.8

8.5 ⫾ 4.7*

228 ⫾ 49

231 ⫾ 99

225 ⫾ 62

243 ⫾ 145

Vitamin B12 (pmol/l)

375 ⫾ 234

363 ⫾ 183

340 ⫾ 151

393 ⫾ 215

Red cell folate (␮g/l)

349 ⫾ 122

506 ⫾ 235†

483 ⫾ 191

533 ⫾ 278

ACE inhibitors or ARBs

—

86

88

85

Beta-blockers

—

61

76

42†

Diuretic agents

—

89

89

89

Aspirin

—

44

44

45

Warfarin

—

25

30

20

Treatment

Values are mean ⫾ SD, proportions (%), or median (interquartile range). *p ⬍ 0.05 for comparisons between patients with TSAT ⱖ20% and those with TSAT ⬍20%. †p ⬍ 0.01 for comparisons between patients with CHF and controls. ‡ESR was measured in 9 controls, 20 iron-replete patients, and 14 iron-deficient patients. CRP was measured in 20 controls, 51 iron-replete patients, and 39 iron-deficient patients. ACE ⫽ angiotensin-converting enzyme; ARB ⫽ angiotensin receptor blocker; CHF ⫽ chronic heart failure; CRP ⫽ C-reactive protein; ESR ⫽ erythrocyte sedimentation rate; LVEF ⫽ left ventricular ejection fraction; MCHC ⫽ mean cell hemoglobin concentration; NYHA ⫽ New York Heart Association; TSAT ⫽ transferrin saturation.

were independently correlated with lower TSAT (r2 ⫽ 0.2, p ⫽ 0.0012), and non-Caucasian ethnicity, higher NYHA functional class, and lower beta-blocker use predicted higher sTFR (r2 ⫽ 0.5, p ⫽ 0.0002). For the functional pool, only higher age and NYHA functional class remained correlated with lower MCHC (r2 ⫽ 0.18, p ⬍ 0.0001), and nonCaucasian ethnicity and higher NYHA functional class (r2 ⫽ 0.42, p ⫽ 0.0001) predicted higher %HRC. For iron stores, non-Caucasian ethnicity, female sex, and higher NYHA functional class were related to lower ferritin (r2 ⫽ 0.42, p ⬍ 0.0001), and higher NYHA functional class and non-Caucasian ethnicity predicted higher sTFR/log ferritin ratio (r2 ⫽ 0.55, p ⬍ 0.0001). Prevalence of iron disorders. Receiver-operating characteristic curve analysis confirmed that TSAT (area under the curve 0.72, p ⫽ 0.002), but not ferritin (area under the curve 0.59, p ⫽ 0.28), predicted sTFR ⱖ1.76 mg/l and that a cutoff of 20% was optimal. Consequently, iron deficiency (TSAT ⬍20%) was evident in 43%, 64%, and 30% of all, anemic, and nonanemic patients, respectively. Using a ferritin cutoff of ⬎30 ␮g/l, functional iron deficiency was evident in 31%, 44%, and 22% of all, anemic, and nonane-

mic patients, respectively. Using a cutoff of ⬎100 ␮g/l, it was present in 14%, 23%, and 8% of all, anemic, and nonanemic patients, respectively. Using the FERRIC-HF (Ferric Iron Sucrose in Heart Failure) criteria (ferritin ⬍100 or 100 to 300 ␮g/l with TSAT ⬍20%), iron deficiency was present in 69%, 78%, and 65% of all, anemic and nonanemic patients, respectively. FERRIC-HF– defined functional (ferritin 100 to 300 ␮g/l with TSAT ⬍20%) and absolute (ferritin ⬍100 ␮g/l) deficiencies were evident in 10% and 59% of subjects, respectively. Using the multiple-measures clinical model, distinct diagnoses were possible in 131 subjects (83%) (Fig. 3). Of the total population, 37 (24%) had normal iron status, none had iron overload, 57 (36%) had a nonanemic iron-deficient state, and 36 (23%) had an anemic iron-deficient state. Etiology of anemia in patients with CHF and its variation with disease severity. Anemia was present in 61 patients (39%). Lower TSAT, higher ESR, and female sex were independently associated with lower hemoglobin (r2 ⫽ 0.47, p ⬍ 0.001). Applying the multiple-measures clinical model, ACD, ACD ⫹ IDA, and IDA were evident in 26%, 16%,

Okonko et al. Iron Homeostasis in Heart Failure

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

1245

research model was applicable to 21 patients with anemia. Six (29%) had isolated ACD, 5 (24%) had ACD ⫹ IDA, and 1 (5%) had isolated IDA. Iron status and exercise performance. Twenty-seven patients underwent cardiopulmonary exercise testing. No significant difference in baseline characteristics existed between those who did and did not perform exercise testing. Peak oxygen consumption was lower in iron-deficient than in iron-replete (11.4 ⫾ 2.1 ml/kg/min vs. 14.9 ⫾ 1.7 ml/kg/min, p ⫽ 0.03) patients and positively correlated to TSAT (r ⫽ 0.71, p ⬍ 0.0001), MCHC (r ⫽ 0.63, p ⬍ 0.001), and ferritin (r ⫽ 0.48, p ⫽ 0.01) independently of NYHA functional class and hemoglobin. The VE/Vco2 slope was higher in irondeficient than in iron-replete (54.8 ⫾ 10.6 vs. 44.1 ⫾ 9.1, p ⫽ 0.02) patients and related only to TSAT (r ⫽ ⫺0.43, p ⫽ 0.03). Iron status and survival. Over a median follow-up period of 743 days (IQR: 304 days), 27 patients (17%) died. In univariate analyses, increased mortality was related only to TSAT ⬍20% (hazard ratio [HR]: 3.38; 95% confidence interval [CI]: 1.48 to 7.72; p ⫽ 0.004), NYHA functional class (HR: 2.13; 95% CI: 1.28 to 3.55; p ⫽ 0.004), beta-blocker use (HR: 0.41; 95% CI: 0.19 to 0.87; p ⫽ 0.02), hemoglobin (HR: 0.74; 95% CI: 0.58 to 0.93; p ⫽ 0.01), and creatinine (HR: 1.009; 95% CI: 1.003 to 1.014; p ⫽ 0.001). Survival was unrelated to TSAT as a continuous variable. In bivariate and trivariate prognostic models, iron deficiency predicted mortality independently of all other covariates in each model. On stratifying patients into hematological subsets, there was a significant difference in survival among the groups (Fig. 5). Patients with IDA had a 2-fold greater risk for death than those with nonanemic iron deficiency and a 4-fold greater risk for death than iron-replete patients with or without anemia.

A

B

C

Discussion

Figure 1

Distribution of Iron Markers in Patients With CHF

Sex-specific distribution of markers of circulating (A), functional (B), and storage (C) iron status. CHF ⫽ chronic heart failure; MCHC ⫽ mean cell hemoglobin content; TSAT ⫽ transferrin saturation.

and 16% of patients with anemia, respectively (Fig. 3). Stratification on the basis of NYHA functional class revealed that iron-deficient substrates (ACD and/or IDA) were evident in 17%, 72%, and 100% of patients in classes I or II, III, and IV, respectively, with ACD the dominant etiology in each stratum (Fig. 4). The multiple-measures

Principal findings. We investigated iron metabolism and its clinical significance in anemic and nonanemic patients with systolic CHF and made 3 key observations. First, iron homeostasis was deranged in a predominately functionally iron deficient manner and related closely to NYHA functional class, ethnicity, CRP, and anemia. Second, the etiologies of anemia varied with disease severity, with an iron-deficient substrate increasingly evident with disease progression. Although ACD was more prevalent than IDA, both conditions frequently coexisted. Third, iron deficiency (TSAT ⬍20%) was associated with exercise intolerance and exertional hyperventilation and identified those at an enhanced risk for death independently and more powerfully than anemia. Iron homeostasis in CHF. Although the existence of iron deficiency (6 – 8,10) and potential benefits (34,35) of correcting it in CHF are emerging, little is known about the appropriateness of systemic iron allocation in this

Okonko et al. Iron Homeostasis in Heart Failure

1246

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

Iron Status Table 2 Iron Status Patients With CHF

Variable

Normal Values

Control Subjects (n ⴝ 22)

All (n ⴝ 157)

Nonanemic (n ⴝ 96)

Anemic (n ⴝ 61)

9–30

17 ⫾ 5

14 ⫾ 6*

16 ⫾ 6

11 ⫾ 5†

—

18%

7%

37%

61 ⫾ 8

64 ⫾ 12

64 ⫾ 11

64 ⫾ 13

Circulating iron status Iron (␮mol/l) ⬍9 TIBC (␮mol/l)

45–66

⬍45

—

1%

2%

0%

⬎66

—

35%

23%

55%

28 ⫾ 9

22 ⫾ 10*

25 ⫾ 10

18 ⫾ 8†

⬍20

—

43%

30%

64%

⬎45

—

3%

5%

0%

1.4 ⫾ 0.3

1.7 ⫾ 0.8

1.7 ⫾ 0.9

1.7 ⫾ 0.6

ⱖ1.76

—

29%

26%

33%

n

12

55

34

21

TSAT (%)

sTFR (mg/l)

30–45

⬍1.76

Functional iron status ⱖ13 (male), ⱖ12 (female)

14.7 ⫾ 1.3

13.1 ⫾ 1.7

14.2 ⫾ 1.1

11.3 ⫾ 0.9

Mean cell volume (fl)

80–100

90 ⫾ 4

91 ⫾ 6

92 ⫾ 6

90 ⫾ 7†

Mean cell hemoglobin (pg)

25–35

30 ⫾ 1

30 ⫾ 2

30 ⫾ 2

29 ⫾ 3†

MCHC (g/dl)

32–36

34.1 ⫾ 1.1

32.9 ⫾ 1.3*

33.0 ⫾ 1.1

32.7 ⫾ 1.5

—

20%

17%

25%

11.5–14.5

13.1 ⫾ 0.3

14 ⫾ 2*

14 ⫾ 1

15 ⫾ 2†

13

40

25

15

⬍2.5–5.0

0.4 ⫾ 0.4

3.1 ⫾ 4.7*

2.2 ⫾ 4.0

4.4 ⫾ 5.5

ⱖ5

—

18%

10%

30%

n

15

49

29

20

33 ⫾ 2

32 ⫾ 2

33 ⫾ 2

32 ⫾ 2

15

36

23

13

78 (35–148)

Hemoglobin (g/dl)

⬍32 Red cell distribution width (%) n Hypochromic cells (%)

Reticulocyte hemoglobin (pg)

ⱖ27

n Stored iron status Ferritin (␮g/l)

85 (47–137)

87 (43–149)

88 (48–169)

⬍15

30–300

—

3%

3%

3%

⬍30

—

16%

13%

20% 61%

⬍100

—

59%

58%

⬎300

—

10%

11%

8%

0.7 ⫾ 0.3

1.0 ⫾ 0.7

1.0 ⫾ 0.7

0.9 ⫾ 0.6

12

55

34

21

sTFR/log ferritin

⬎1.00

n

Values are mean ⫾ SD, proportions (%), or median (interquartile range). *p ⬍ 0.05 for comparisons between patients with CHF and controls. †p ⬍ 0.05 for comparisons between anemic and nonanemic patients with CHF. CHF ⫽ chronic heart failure; MCHC ⫽ mean cell hemoglobin concentration; sTFR ⫽ soluble transferrin receptor; TIBC ⫽ total iron-binding capacity; TSAT ⫽ transferrin saturation.

disorder. Normally, iron principally cycles from the circulation to the erythroid marrow to macrophages and back to the circulation, with the flow of iron between compartments matched to achieve equilibrium. Powerful homeostatic mechanisms orchestrate these movements to maintain the size of each iron pool, with the overriding objective of ensuring that circulating levels are sufficient to support normal erythrocyte hemoglobinization (11–13). Here, we found that the circulating pool of iron was contracted by 15% to 35% in patients with CHF, implying that defects in circulatory iron influx and/or efflux existed. In parallel, the function pool was similarly attenuated, and this reflected the decrease in circulating iron, as the means of iron entry into functional sites (i.e., transferrin receptors) were adequate. In contrast, the storage compartment appeared to be expanded by 24% in patients with CHF. Taken together, these data indicate that in the deranged milieu of

CHF, iron is abnormally handled, being preferentially trafficked into storage sites and withdrawn from the circulation and erythroid marrow, triggering IDE. This is evident in anemic and nonanemic subjects and mirrors abnormal iron kinetics in a plethora of inflammatory disorders (1–5). Although iron stores were biochemically enhanced in patients with CHF compared with controls, careful analysis of the data suggests that their progressive depletion occurs with worsening disease. Serum ferritin levels ⬍15 to 30 ␮g/l unequivocally signify exhausted iron reserves (1,11,33). However, as an acute phase reactant, ferritin is 3-fold higher in inflammatory cohorts than in normal subjects for each grade of marrow iron (32,33). While ferritin levels in this study were, on average, higher in patients with CHF than controls, they progressively declined from patients in class I or II to controls to those in class III and then those in class IV. Thus, while ferritin

Okonko et al. Iron Homeostasis in Heart Failure

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

A

Circulating Iron

B

Functional Iron

C

Figure 2

1247

Storage Iron

Iron Metabolism in CHF

Circulating (A), functional (B), and storage (C) iron indexes stratified for New York Heart Association (NYHA) functional class and the presence of anemia. For box-andwhisker plots, whiskers represent 90th and 10th percentiles, upper and lower ends of boxes represent 75th and 25th percentiles, and solid line depicts median. ANOVA ⫽ analysis of variance; MCHC ⫽ mean cell hemoglobin concentration; sTFR ⫽ soluble transferrin receptor.

levels are elevated early in the course of CHF, they do not escalate further, as would be expected if they simply tracked inflammation, but diminish with disease severity. Because elevated ferritin concentrations still correlate with the degree of stainable marrow iron in inflammatory cohorts (33), our findings suggest that a true depletion of iron stores occurs with advancing CHF. This is consistent with data from Nanas et al. (6) and is a feature of other inflammatory conditions (2). Predictors of impaired iron status. Disease severity, as quantified by NYHA functional class, was a powerful and inverse predictor of all iron compartments. Because it did so independently of other measured covariates, iron perturba-

tions might be causally linked to CHF progression. Moreover, the fact that iron markers were related to NYHA functional class, an index of the overall impact of the heart failure syndrome on patients, but not to left ventricular ejection fraction, suggests that iron abnormalities are a consequence of CHF rather than a direct result of myocardial dysfunction. Besides NYHA functional class, non-Caucasian ethnicity also predicted poorer global iron status. This likely reflects the fact that the diet of Southeast Asians, who were the predominant non-Caucasians, is typically high in phosphates and phytates that inhibit iron absorption (38). Predictably, higher CRP was correlated with lower TSAT,

Okonko et al. Iron Homeostasis in Heart Failure

1248

Iron overload

Diagnosis:

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

Increased Iron stores

Normal

n=1

n=37

≥45 >300 ---

20-50 ≥100 ---

Reduced Iron stores

Iron depletion

IDE with depleted stores

IDE with normal stores

ACD

ACD + IDA

IDA

excess

Storage iron Circulating iron Functional iron

30 Frequency (%)

25 20 15 10 5 0

n=0

TSAT(%) Ferritin (µg/L) TIBC (µmol/L ) Anemia Figure 3

>60 >300 ---

n=25

20-30 30-100 -No

n=3

20-30 <30 -No

n=8

n=21

n=16

n=10

n=10

<20 <30 -No

<20 ≥30 -No

<20 ≥30 ≤65 Yes

<20 ≥30 >65 Yes

<20 <30 ≥45 Yes

Prevalence of Distinct Iron States

The full spectrum of iron states is graphically depicted (top row) and quantifed (middle row) using established criteria (bottom row). Adapted from Brittenham (11). ACD ⫽ anemia of chronic disease; IDA ⫽ iron-deficiency anemia; IDE ⫽ iron-deficient erythropoiesis; TIBC ⫽ total iron-binding capacity; TSAT ⫽ transferrin saturation.

consistent with the fact that cytokines divert iron traffic away from the circulation into stores in chronic disorders (1,4). Importantly, we found no relation between iron indexes and antiplatelet or anticoagulant drug use or platelet counts, arguing against a dominant role for occult hemorrhage in iron derangements. Etiology of anemia

ACD ACD + IDA

15

90

IDA

Frequency (%)

80

unrelated to iron

70 60

7

50 8

40 20 10

7

6

30

4

4

4 2 1

0

0

0 NYHA I/II (n=18)

Figure 4

NYHA III (n=25)

NYHA IV (n=15)

Etiology of Anemia Stratified by NYHA Functional Class

Figures above bars correspond to numbers of patients. ACD ⫽ anemia of chronic disease; IDA ⫽ iron-deficiency anemia; NYHA ⫽ New York Heart Association.

Prevalence of iron disorders. Iron disorders are a continuum, and to the best of our knowledge, this is the first study to report prevalence estimates for the full spectrum of anomalies in CHF. In contrast to healthy elders (39), iron-deficient states were more common than those of iron excess, suggesting that CHF per se might play a role in the genesis of iron insufficiency. Interestingly, 29 patients (15.6%) had nonanemic IDE, which can precipitate fatigue, depression, and exercise intolerance (11,26,27). Thus, identifying IDE before the onset of anemia is important, especially in patients with CHF, in whom evidence suggests that anemia is prevalent and adverse (18). Despite such evidence, however, the relation of iron indexes to hemoglobin in CHF remains poorly explored. Iron status and its relation to hemoglobin and the etiologies of anemia. Impaired iron status strongly predicted anemia in this cohort. This is unsurprising given that iron is a critical structural component of hemoglobin (40), a direct trigger of erythroid proliferation (16), an enhancer of globin synthesis (15), and the strongest inducer of aminolevulinic acid synthase, the rate-limiting enzyme of heme production (14). Accordingly, lower TSAT predicted lower hemoglobin independently of higher ESR and female sex. That the ESR was independently predictive of hemoglobin likely reflects the fact that inflammation precipitates anemia

Okonko et al. Iron Homeostasis in Heart Failure

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

A

Figure 5

1249

B

Prognostic Association of Varying Hematological Groups

Kaplan-Meier survival curves for hematological subsets (A) with corresponding hazard ratios (HRs) and 95% confidence intervals (CIs) depicted (B). TSAT ⫽ transferrin saturation.

also by directly suppressing erythroid cells and inducing their resistance to erythropoietin (1). That sex was also independently predictive may reflect the potent influence of androgens on erythropoiesis (41). ACD (with or without IDA) was the dominant substrate for anemia in our cohort, evident in 44% to 53% of subjects. This is similar to the frequencies reported in prior analyses (7,18), concurs with experimental data (42), but supersedes the 33% burden seen in otherwise healthy anemic elders (43). Although less prevalent, IDA was evident in a substantial 29% to 34% of patients with anemia. This is higher than most prior estimates (7,9,10), possibly because of our use of TIBC and sTFR levels to identify IDA in those with ACD. Next, we sought, for the first time, to determine whether the etiologies of anemia varied with CHF severity. We found that as NYHA functional class worsened, the prevalence of ACD and/or IDA rose, with ACD being increasingly complicated by IDA. This is consistent with the known evolution of ACD, whereby persistent hypoferremia ultimately impairs iron storage, giving rise to concomitant IDA that is masked by inflammatory hyperferritemia (1,2,44). Indeed, in an anemic NYHA functional class IV cohort, Nanas et al. (6) found definitive IDA in 73% of patients, whose mean ferritin level was 75 ␮g/l. In this study, only 53% of our anemic NYHA functional class IV patients had IDA, possibly because they were healthier. In contrast, our prevalence estimate for ACD is higher than

that of Nanas et al. (18.9%), probably because we diagnosed ACD using positive criteria, whereas it was a diagnosis of exclusion in that analysis. However, given their mean ferritin levels, many of the Nanas et al. (6) cohort with definitive IDA also likely met biochemical criteria for ACD. Iron status and its relation to exercise performance and survival. Iron-deficient patients were objectively more exercise intolerant and breathless on exertion and harbored a 3-fold escalated risk for death irrespective of whether they were anemic. Indeed, iron-deficient nonanemic patients had a 2-fold greater risk for mortality than iron-replete anemic subjects. This implies that iron deficiency per se might be prognostically more ominous than anemia in CHF. Although lower TSAT may merely reflect other unadjusted adverse factors, a causal prognostic role cannot be excluded. Besides hemoglobin, iron is an obligate component of several metalloproteins that mediate pivotal functions, such as myoglobin (oxygen storage) (19), electron transport enzymes (oxidative phosphorylation) (20,21), ribonucleotide reductase (deoxyribonucleic acid synthesis) (22), monoamine oxidase (neurotransmission) (23), and cyclooxygenases (inflammation) (24). Iron is thus indispensable for life, and cellular iron deprivation triggers G1/S phase mitotic arrest and apoptosis (25). Iron deficiency also drives catecholamine production (45) and ventricular hypertrophy (46), factors that bode unfavorably in CHF. Such factors might therefore underlie the associations between

1250

Okonko et al. Iron Homeostasis in Heart Failure

iron deficiency and adverse prognosis evident here and in other analyses (28,29,47). Study implications. This study has important ramifications. First, it confirms emerging evidence that iron deficiency is another pathophysiological feature of CHF that likely arises from the syndrome itself to drive symptoms, exercise intolerance, and mortality irrespective of anemia. Second, it supports the routine assessment of iron status in patients with CHF, particularly in those with refractory symptoms, advanced disease, or anemia. Third, it suggests that low hemoglobin levels in CHF result predominately from true substrates and are not just spurious consequences of hemodilution. Fourth, it shows that the rational use of 3 simple indexes (TSAT, ferritin, and TIBC) enables the biochemically meaningful discrimination of ACD from IDA and the identification of IDA in the midst of ACD. Decisions regarding the mandatory (IDA and ACD ⫹ IDA) or adjuvant (ACD) need for iron supplementation can therefore be facilitated. Finally, it suggests that even if iron deficiency is not causally related to mortality, its presence might identify those in need of escalated care. Study limitations. First, this study was observational, so causal links between variables cannot be established. Second, it represents data measured at a single time point and does not inform on temporal trends in iron status or anemic etiology. Third, for technical and logistical reasons, indexes such as %HRC and sTFR were quantified only in subsets of patients. However, no difference in clinical characteristics existed between those who were and were not measured. Fourth, brain natriuretic peptide measurements were unavailable, but the prognostic independence of iron deficiency from brain natriuretic peptide was suggested in a prior study (47). Fifth, we did not include patients with acute heart failure or recent infarcts, who may have greater iron derangements (6). Finally, a conventional ferritin criteria (⬍30 g/dl) (1,12) was used to diagnose IDA, despite evidence that higher cutoffs may be better in inflammatory cohorts. Thus, we might have overestimated the prevalence of ACD and underestimated that of IDA in this study. Because the absence of stainable iron in bone marrow aspirates is the optimal marker of iron deficiency, studies predicated on this are needed to define “normal” ferritin levels in CHF. However, by demonstrating that IDA remained substantially prevalent despite the use of a conservative cutoff, we contend that our study further highlights its importance in CHF. Conclusions Iron metabolism is deranged in CHF and is characterized by a diminished circulating and functional iron status in the face of seemingly adequate stores. Disordered iron homeostasis correlates to worsening inflammation, NYHA functional class, and non-Caucasian ethnicity, and increasingly underlies anemia with CHF progression. Iron deficiency, as defined by a TSAT ⬍20%, relates to exercise intolerance

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51

and amplified mortality, and should be targeted in future survival trials. Acknowledgment

This work is dedicated to the memory of Professor PooleWilson, who died suddenly in March 2009. Reprint requests and correspondence: Dr. Darlington O. Okonko, 73D Babington Road, London SW16 6AN, United Kingdom. E-mail: [email protected]. REFERENCES

1. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med 2005;352:1011–23. 2. Goyal R, Das R, Bambery P, Garewal G. Serum transferrin receptorferritin index shows concomitant iron deficiency anemia and anemia of chronic disease is common in patients with rheumatoid arthritis in north India. Indian J Pathol Microbiol 2008;51:102– 4. 3. Darveau M, Denault AY, Blais N, Notebaert E. Bench-to-bedside review: iron metabolism in critically ill patients. Crit Care 2004;8: 356 – 62. 4. Denz H, Huber P, Landmann R, et al. Association between the activation of macrophages, changes of iron metabolism and the degree of anaemia in patients with malignant disorders. Eur J Haematol 1992;48:244 – 8. 5. Andrews NC. Disorders of iron metabolism. N Engl J Med 1999;341: 1986 –95. 6. Nanas JN, Matsouka C, Karageorgopoulos D, et al. Etiology of anemia in patients with advanced heart failure. J Am Coll Cardiol 2006;48:2485–9. 7. Opasich C, Cazzola M, Scelsi L, et al. Blunted erythropoietin production and defective iron supply for erythropoiesis as major causes of anaemia in patients with chronic heart failure. Eur Heart J 2005;26:2232–7. 8. Adlbrecht C, Kommata S, Hülsmann M, et al. Chronic heart failure leads to an expanded plasma volume and pseudoanaemia, but does not lead to a reduction in the body’s red cell volume. Eur Heart J 2008;29:2343–50. 9. Witte KK, Desilva R, Chattopadhyay S, Ghosh J, Cleland JG, Clark AL. Are hematinic deficiencies the cause of anemia in chronic heart failure? Am Heart J 2004;147:924 –30. 10. de Silva R, Rigby AS, Witte KK, et al. Anemia, renal dysfunction, and their interaction in patients with chronic heart failure. Am J Cardiol 2006;98:391– 8. 11. Brittenham GM. Disorders of iron metabolism: iron deficiency and overload. In: Hoffman R, Shattil SJ, Furie B, Cohen HJ, Silberstein LE, McGlave P, eds. Hematology: Basic Principles and Practice. 3rd ed. London: Churchill Livingstone, 2000:397– 428. 12. Cook JD. Diagnosis and management of iron-deficiency anaemia. Best Pract Res Clin Haematol 2005;18:319 –32. 13. Besarab A, Hörl WH, Silverberg D. Iron metabolism, iron deficiency, thrombocytosis, and the cardiorenal anemia syndrome. Oncologist 2009;14:22–33. 14. Melefors O, Goossen B, Johansson HE, et al. Translational control of 5-aminolevulinate synthase mRNA by iron-responsive elements in erythroid cells. J Biol Chem 1993;268:5974 – 8. 15. Hunter AR, Jackson RJ. Control of haemoglobin synthesis. The effects of iron deprivation, cobalt and temperature on the rate and extent of globin synthesis in reticulocytes. Eur J Biochem 1975;58:421–30. 16. Neuwirt J, Hradilek A, Ba´rtek J. The relation of iron uptake to the proliferation of erythroid and hemopoietic cells. Biomed Biochim Acta 1983;42:S177– 81. 17. Stoltzfus RJ. Iron deficiency: global prevalence and consequences. Food Nutr Bull 2003;24:S99 –103. 18. Ezekowitz JA, McAlister FA, Armstrong PW. Anemia is common in heart failure and is associated with poor outcomes: insights from a cohort of 12 065 patients with new-onset heart failure. Circulation 2003;107:223–5. 19. O’Hagan JE. The iron-protein link in haemoglobin and myoglobin. Nature 1959;183:393.

Okonko et al. Iron Homeostasis in Heart Failure

JACC Vol. 58, No. 12, 2011 September 13, 2011:1241–51 20. Oexle H, Gnaiger E, Weiss G. Iron-dependent changes in cellular energy metabolism: influence on citric acid cycle and oxidative phosphorylation. Biochim Biophys Acta 1999;1413:99 –107. 21. Dhur A, Galan P, Hercberg S. Effects of different degrees of iron deficiency on cytochrome P450 complex and pentose phosphate pathway dehydrogenases in the rat. J Nutr 1989;119:40 –7. 22. Jensen KP, Bell CB III, Clay MD, Solomon EI. Peroxo-type intermediates in class I ribonucleotide reductase and related binuclear non-heme iron enzymes. J Am Chem Soc 2009;131:12155–71. 23. Youdim MB, Woods HF, Mitchell B, Grahame-Smith DG, Callender S. Human platelet monoamine oxidase activity in irondeficiency anaemia. Clin Sci Mol Med 1975;48:289 –95. 24. Hemler M, Lands WE. Purification of the cyclooxygenase that forms prostaglandins. Demonstration of two forms of iron in the holoenzyme. J Biol Chem 1976;251:5575–9. 25. Lederman HM, Cohen A, Lee JW, Freedman MH, Gelfand EW. Deferoxamine: a reversible S-phase inhibitor of human lymphocyte proliferation. Blood 1984;64:748 –53. 26. Brownlie T, Utermohlen V, Hinton PS, Haas JD. Tissue iron deficiency without anaemia impairs adaptation in endurance capacity after aerobic training in previously untrained women. Am J Clin Nutr 2004;79:437– 43. 27. Verdon F, Burnand B, Stubi CL, et al. Iron supplementation for unexplained fatigue in non-anaemic women: double blind randomised placebo controlled trial. Br Med J 2003;326:1124 –7. 28. Corti MC, Guralnik JM, Salive ME, et al. Serum iron level, coronary artery disease, and all-cause mortality in older men and women. Am J Cardiol 1997;79:120 –7. 29. Kovesdy CP, Estrada W, Ahmadzadeh S, Kalantar-Zadeh K. Association of markers of iron stores with outcomes in patients with nondialysis-dependent chronic kidney disease. Clin J Am Soc Nephrol 2009;4:435– 41. 30. WHO Scientific Group. Nutritional anaemias. Report of a WHO scientific group. World Health Organ Tech Rep Ser 1968;405:5–37. 31. Cunietti E, Chiari MM, Monti M, et al. Distortion of iron status indices by acute inflammation in older hospitalized patients. Arch Gerontol Geriatr 2004;39:35– 42. 32. Lipschitz DA, Cook JD, Finch CA. A clinical evaluation of serum ferritin as an index of iron stores. N Engl J Med 1974;290:1213– 6. 33. Bentley DP, Williams P. Serum ferritin concentration as an index of storage iron in rheumatoid arthritis. J Clin Pathol 1974;27:786 – 8. 34. Okonko D, Grzeslo A, Witkowski T, et al. Effect of intravenous iron sucrose on exercise tolerance in anemic and non-anemic patients with

35. 36. 37.

38. 39.

40. 41. 42. 43.

44. 45. 46. 47.

1251

symptomatic chronic heart failure and iron deficiency (FERRIC-HF): a randomized, controlled, observer-blinded trial. J Am Coll Cardiol 2008;51:103–12. Anker SD, Comin Colet J, Filippatos G, et al. Ferric carboxymaltose in patients with heart failure and iron deficiency. N Engl J Med 2009;361:1–13. Beguin Y. Soluble transferrin receptor for the evaluation of erythropoiesis and iron status. Clin Chim Acta 2003;329:9 –22. Francis DP, Shamim W, Davies LC, et al. Cardiopulmonary exercise testing for prognosis in chronic heart failure: continuous and independent prognostic value from VE/VCO(2)slope and peak VO(2). Eur Heart J 2000;21:154 –161. Nair KM, Iyengar V. Iron content, bioavailability & factors affecting iron status of Indians. Indian J Med Res 2009;130:634 – 45. Fleming DJ, Jacques PF, Tucker KL, et al. Iron status of the free-living, elderly Framingham Heart Study cohort: an iron-replete population with a high prevalence of elevated iron stores. Am J Clin Nutr 2001;73:638 – 46. Walker BS, Boyd WC. The percentage of iron in haemoglobin. Science 1937;85:360 –1. Shahidi NT. Androgens and erythropoiesis. N Engl J Med 1973;289: 72– 80. Iversen PO, Woldbaek PR, Tonnessen T, Christensen G. Decreased hematopoiesis in bone marrow of mice with congestive heart failure. Am J Physiol Regul Integr Comp Physiol 2002;282:166 –72. Guralnik JM, Eisenstaedt RS, Ferrucci L, Klein HG, Woodman RC. Prevalence of anemia in persons 65 years and older in the United States: evidence for a high rate of unexplained anemia. Blood 2004; 104:2263– 8. Weinstein DA, Roy CN, Fleming MD, et al. Inappropriate expression of hepcidin is associated with iron refractory anemia: implications for the anemia of chronic disease. Blood 2002;100:3776 – 81. Dillmann E, Johnson DG, Martin J, Mackler B, Finch C. Catecholamine elevation in iron deficiency. Am J Physiol 1979;237:R297–300. Rossi MA, Carillo SV. Pathogenesis of cardiac hypertrophy in iron deficiency anaemia: the role of noradrenaline. Br J Exp Pathol 1982;63:269 –77. Jankowska EA, Rozentryt P, Witkowska A, et al. Iron deficiency: an ominous sign in patients with systolic chronic heart failure. Eur Heart J 2010;31:1872– 80.

Key Words: anemia y heart failure y iron.