Redefining Iron Deficiency in Patients With Chronic Heart Failure

Abstract

A serum ferritin level <15 to 20 μg/L historically identified patients who had absent bone marrow iron stores, but serum ferritin levels are distorted by the systemic inflammatory states seen in patients with chronic kidney disease or heart failure. As a result, nearly 25 years ago, the diagnostic ferritin threshold was increased 5- to 20-fold in patients with chronic kidney disease (ie, iron deficiency was identified if the serum ferritin level was <100 μg/L, regardless of transferrin saturation [TSAT], or 100 to 299 μg/L if TSAT was <20%). This guidance was motivated not by the findings of studies of total body or tissue iron depletion, but by a desire to encourage the use of iron supplements to potentiate the response to erythropoiesis-stimulating agents in patients with renal anemia. However, in patients with heart failure, this definition does not reliably identify patients with an absolute or functional iron-deficiency state, and it includes individuals with TSATs (≥20%) and serum ferritin levels in the normal range (20–100 mg/L) who are not iron deficient, have an excellent prognosis, and do not respond favorably to iron therapy. Furthermore, serum ferritin levels may be distorted by the use of both neprilysin and sodium-glucose cotransporter 2 inhibitors, both of which may act to mobilize endogenous iron stores. The most evidence-based and trial-tested definition of iron deficiency is the presence of hypoferremia, as reflected by as a TSAT <20%. These hypoferremic patients are generally iron deficient on bone marrow examination, and after intravenous iron therapy, they exhibit an improvement in exercise tolerance and functional capacity (when meaningfully impaired) and show the most marked reduction (ie, 20%–30%) in the risk of cardiovascular death or total heart failure hospitalizations. Therefore, we propose that the current ferritin-driven definition of iron deficiency in heart failure should be abandoned and that a definition based on hypoferremia (TSAT <20%) should be adopted.

For decades, iron deficiency was considered the cornerstone of the differential diagnosis in patients with anemia, which (in the current era) is often related to a gastrointestinal disorder, poor dietary intake, chronic kidney disease (CKD), or the use of medications that inhibit iron absorption. In recent years, however, it seems likely that a certain proportion of patients with heart failure (HF) are iron deficient, and several double-blind, placebo-controlled trials of oral and intravenous (IV) iron supplementation have been carried out in patients with a reduced ejection fraction.^1–5 The definitions of iron deficiency and the eligibility criteria used in these HF trials were largely informed by earlier experiences in profoundly anemic patients on hemodialysis, but most patients with HF who are at risk of iron deficiency do not have meaningful anemia or advanced CKD.¹ Furthermore, the goals of treating iron deficiency in patients with renal anemia are distinctly different from those that guide the treatment of cardiomyopathy.

IDENTIFICATION OF IRON DEFICIENCY IN PATIENTS WITH HF

Iron-deficiency states historically have been classified as being related to absolute or functional iron deficiency. Absolute iron deficiency is defined by the depletion of total body iron stores, often related to blood loss, poor dietary iron intake, or malabsorption of iron. Functional iron deficiency is defined by the adequacy of total body iron stores, coupled with an impairment in the mechanism by which these stores can be mobilized to sustain the bloodstream delivery of iron to target tissues (eg, to erythroblasts to promote erythropoiesis, or to cardiomyocytes and skeletal muscle to maintain contractile function).

How Can Total Body Iron Stores Become Depleted in Patients With HF?

An absolute iron-deficiency state can occur in patients with HF in several ways. Patients with HF are often receiving antiplatelet or anticoagulant drugs that can cause occult gastrointestinal blood loss.⁶ Patients may also be receiving proton pump inhibitors that reduce the gastric acidity required for iron absorption. Polyphenols in coffee can markedly impair iron absorption in the duodenum.⁷ The use of calcium channel blockers may interfere with membrane-bound channel required for the uptake of iron by enterocytes.⁸ Increases in hepcidin—as might result from hepatic congestion due to increased central venous pressures—may limit duodenal absorption of dietary iron.⁹ Elderly patients with HF are subject to poor dietary intake of iron as a result of anorexia or to iron losses as a result of gastrointestinal malignancy.¹⁰

The diagnosis of absolute iron deficiency conventionally has not been made by the measurement of circulating iron or by bone marrow examination; instead, it has been made by the finding of a serum ferritin level <15 to 20 μg/L.¹¹ Ferritin serves as an intracellular nanocage, which stores excess cytosolic iron in a nonreactive form to be made available when needed (Figure). When levels of biologically reactive cytosolic iron within erythroid precursors are diminished, iron is released from the nanocage by a process of ferritin degradation, the intensity of which is augmented when cytosolic iron is depleted.¹² It is not possible to measure the labile intracellular pool of highly reactive catalytic iron in humans, but it is possible to estimate the expression of ferritin inside cells, because ferritin is actively and proportionately secreted into the bloodstream, primarily from macrophages.^13,14 Very low serum levels of ferritin would imply an accelerated degradation of intracellular levels of ferritin within the reticuloendothelial system, which occurs when cytosolic iron levels are depressed.^14–16 Furthermore, in absolute iron deficiency, the delivery of iron to erythroblasts is insufficient to maintain erythropoiesis, even though erythroblasts represent a privileged destination for circulating iron.¹⁷ For these reasons, absolute iron deficiency is identified by a serum ferritin level <15 to 20 μg/L, generally in combination with a hypochromic microcytic anemia.¹¹

Figure.

Iron regulatory proteins as potential biomarkers for an iron-deficiency state. The top row describes biologic function; the second row, triggers and biologic significance; the third row, diagnostic use; and the bottom row, assay limitations. IV indicates intravenous; and TSAT, transferrin saturation.

How Can Iron Stores Become Sequestered in Patients With HF?

Patients with HF can also develop iron deficiency in the presence of adequate total body iron stores as a result of a functional block on the egress of iron from tissue reservoirs into the bloodstream.¹⁵ This impairment in mobilization is related to upregulation of hepcidin, a protein that inhibits the release of iron into the bloodstream from duodenal enterocytes, hepatocytes, or macrophages that have engulfed senescent erythrocytes.^15,18 Patients with HF often have a low-grade systemic inflammatory state, which stimulates the hepatic synthesis of hepcidin, resulting in trapping of iron, primarily in the reticuloendothelial system.¹⁹ A disproportionate increase in hepcidin is the molecular hallmark of a functional iron-deficiency state.^20,21

In 2 large cohort studies, patients with mild to moderate HF had circulating levels of hepcidin that were twice as high as seen in a healthy control group or for the expected levels based on age and sex.^22,23 This increase was unexpected, because as hypoferremia ensues, hepcidin levels generally decline.^6,15 Despite these reports, measurement of serum hepcidin in the clinical setting is fraught with difficulties (Figure). Assays yield markedly different results, which are dependent on important technical issues. Analytical approaches may differ in the degree to which they detect inactive hepcidin precursors or protein fragments.^24,25

With respect to the effect on circulating levels of iron, the most important site of action of hepcidin appears to be the macrophage and hepatocyte, rather than the enterocyte, because selective hepcidin suppression in nonduodenal cells releases iron and is sufficient to alleviate anemia.^26,27 Macrophage trapping is particularly meaningful, because recycling of iron from senescent erythrocytes is the primary mechanism by which serum iron levels are maintained to sustain red blood cell production.²⁸ A diminished role for duodenal iron uptake in states of heightened hepcidin synthesis may explain why no study has shown a relationship between serum hepcidin and iron malabsorption in HF. In fact, in patients with HF without evidence of systemic inflammation, the absorption of iron has been reported to be enhanced—not impaired.²⁹

Despite the lack of evidence for iron malabsorption, oral iron supplementation in patients with HF produces only modest changes in iron biomarkers, with no little rise in the serum ferritin level.³⁰ The lack of a reactive increase in serum ferritin indicates that oral iron administration produces minimal changes in the intracellular labile pool of reactive iron.²⁹ In fact, only 20% to 25% of patients with HF experience a substantial improvement in iron status after 16 weeks of oral iron therapy.³¹ These results are similar to the disappointing findings with oral iron supplementation in CKD.³² Therefore, it seems likely that, in patients with functional iron deficiency, oral iron is absorbed, but so slowly that it is readily sequestered in hepatocytes or macrophages by the heightened action of hepcidin before it can be delivered to target sites. In contrast, IV iron rapidly achieves high serum iron concentrations, thus bypassing the usual transferrin-based transport mechanisms in the bloodstream and allowing direct delivery to depleted cytosolic pools within target organs.³³

Lack of a Clinical Rationale for Distinguishing Between Absolute and Functional Iron Deficiency in HF

In the absence of a reliable threshold for serum hepcidin, the identification of functional iron deficiency in patients with HF is challenging. Increased levels of hepcidin prevent the egress of iron from maturing erythroblasts, thus preserving erythropoiesis.¹⁷ As a result, in contrast to patients with an absolute iron deficiency, patients typically have hypoferremia (ie, serum iron levels <12–13 μg/L or a transferrin saturation [TSAT] <20%), but with mild anemia or no anemia at all.^11,17 Furthermore, because of the presence of a systemic inflammatory state in functional iron deficiency, the synthesis and secretion of ferritin by macrophages is enhanced.^14,16,34 As a result, serum ferritin levels exceed those seen in absolute iron deficiency (<15–20 μg/L), but in most patients with functional iron deficiency, serum ferritin levels remain in the normal range for healthy people (20–300 μg/L). Therefore, there exists no threshold value for serum ferritin level that identifies patients with a functional iron-deficiency state, and for many patients, the only clinically apparent indicator of the existence of functional iron deficiency is a meaningful therapeutic response to IV iron therapy.

The available evidence, taken together, suggests that patients with HF can experience either absolute or functional iron deficiency, and that the mechanisms for both likely coexist in the majority of iron-deficient patients.^{22,23,35–37} Gastrointestinal iron loss or malabsorption may be compounded by hepcidin-mediated trapping of iron in reticuloendothelial stores. Serum hepcidin levels may decrease with the progression of HF and may be particularly suppressed in patients with acutely decompensated HF^22,37; on the contrary, worsening of kidney function over time may cause increases in serum hepcidin,³⁸ and liver congestion has been hypothesized to promote hepcidin synthesis and prevent hepatocytes from releasing iron.^9,39 Neurohormonal activation in HF may adversely influence the uptake of iron by cardiomyocytes.⁴⁰ These highly dynamic changes suggest that the relative contributions of depleted iron stores and defective iron mobilization to the development of hypoferremia may fluctuate over time. Varying combinations of absolute and functional iron deficiency also appear to be a common occurrence in patients with chronic kidney or inflammatory bowel disease.^41–43

Because the treatment of iron deficiency in HF does not depend on characterization of its pathogenesis, efforts to distinguish between absolute and functional iron deficiency at one particular point in time in a patient’s clinical course have little therapeutic relevance. There is no rationale for identifying the relative contribution of specific pathophysiologic mechanisms for iron deficiency in clinical practice, and there is no reliable noninvasive means of achieving this goal. Nevertheless, the finding of meaningful hypoferremia together with a hypochromic microcytic anemia might prompt a diagnostic workup for a gastrointestinal malignancy.¹⁰

THE CURRENT DEFINITION OF IRON DEFICIENCY IN HF IS WRONG

According to current guidelines,⁴⁴ an iron-deficiency state is defined as a serum ferritin level <100 μg/L (regardless of TSAT) or a serum ferritin level of 100 to 299 μg/L if the TSAT is <20%. Therefore, patients who have a serum ferritin level in the normal range (20–100 μg/L) are considered iron deficient, even if there is no evidence of hypoferremia. Patients with meaningful hypoferremia (ie, TSAT <20%) are not considered iron deficient if the serum ferritin level is ≥300 μg/L.

What Is the Origin of Our Current Definition of Iron Deficiency?

Although this guideline-recommended definition has been frequently used as the basis for eligibility in clinical trials of iron repletion therapy,^1–5 it is difficult to understand why the clinical community believes that it accurately reflects an iron-deficiency state.

Although a serum ferritin level <15 to 20 μg/L was historically used to identify patients with an absolute iron deficiency,^11,45 such low levels of serum ferritin are seen in <10% of patients with HF and absent bone marrow stores,³⁵ because systemic inflammation promotes macrophage ferritin production and secretion.^34,36 The same confounding was observed in CKD⁴⁶; thus, 25 years ago, the diagnostic ferritin threshold was arbitrarily increased 5- to 15-fold (ie, a serum ferritin level <100 μg/L [regardless of TSAT] or a serum ferritin level of <300 μg/L [if the TSAT is <20%] was proposed to be used to identify patients with CKD who had iron deficiency).⁴⁷ However, this recommendation was motivated not by the findings on bone marrow iron examinations or other evidence of iron depletion, but by a desire to encourage the use of iron supplements to potentiate the response to erythropoiesis-stimulating agents (ESAs) in patients with end-stage kidney disease.^11,47 The goal of using IV iron in these patients was not to treat iron deficiency but to prevent brief episodes of intracellular iron dysregulation that might occur during pharmacologically mediated heightened erythropoiesis,⁴⁸ because the principal goal of iron supplementation in patients with CKD receiving ESAs was the treatment of renal anemia.¹¹

Is There Evidence That Our Current Definition of Iron Deficiency Is Mistaken?

In patients with CKD who are not receiving ESAs, the guideline-recommended definition of iron deficiency has only modest positive and negative predictive value in presaging the erythrocytic response to IV iron therapy, even in patients with manifest anemia.^46,49,50 Even more puzzling, the definition assumed that patients with a serum ferritin level of ≥300 μg/L did not have iron deficiency, even though it was known that individuals with renal anemia and high serum ferritin levels exhibit a meaningful erythrocytosis in response to IV iron.^11,49–51

To complicate matters further, many investigators have assumed that a serum ferritin level <100 μg/L represents absolute iron deficiency and that a serum ferritin level of 100 to 299 μg/L with a TSAT <20% represents functional iron deficiency.^23,52–54 It is true that these 2 components of the current definition of an iron-deficiency state may define somewhat different groups of patients: those with a serum ferritin level of 100 to 299 μg/L with a TSAT <20% have greater evidence of systemic inflammation and a lower prevalence of anemia.³⁶ However, no study has shown that a serum ferritin level of 100 to 299 μg/L with a TSAT <20% selectively identifies patients with impaired release of iron from macrophage stores, as compared with others who are deemed to be iron deficient.^11,46,49,50 The 2 components of the current guideline-accepted definition of iron deficiency are arbitrary and are not based on pathogenetic mechanisms.

Specific studies in patients with HF have failed to confirm the validity of the current definition of iron deficiency. One study of bone marrow iron stores in patients with HF⁵⁵ demonstrated that those with a TSAT <20% have iron depletion regardless of the serum ferritin level and that none of those with a serum ferritin level <100 μg/L but with a TSAT >20% had iron depletion. In another report, a serum ferritin level <136 mg/L or TSAT <33% had poor sensitivity for the identification of patients with absent iron stores.⁵⁶ In a third study, a serum ferritin level <128 µg/L was indicative of absent iron stores only when it was seen together with a TSAT <20%.³⁶ On the basis of these bone marrow studies, iron deficiency was most reliably identified when the serum iron level was ≤13 μmol/L or the TSAT was <20%, regardless of serum ferritin levels.^36,55,56

DEVELOPMENT OF A NEW DEFINITION OF IRON-DEFICIENT HF

Given the link between serum iron level ≤13 μmol/L and TSAT <20% with bone marrow stores, it is noteworthy that both thresholds have been shown to have prognostic significance in patients with HF (ie, they each predict the occurrence of cardiovascular death and HF hospitalizations).^53,57–59 In contrast, patients with a serum ferritin level <100 ng/mL but normal serum iron levels and TSAT are deemed iron deficient by the guidelines, but have a favorable prognosis, presumably because serum ferritin levels of 20 to 100 mg/L lie in the normal range for healthy, iron-replete people. Patients with high serum ferritin levels (≥300 μg/L) but low serum iron concentrations (≤13 μmol/L), low TSAT (<20%), or anemia—features commonly seen in functional iron deficiency—are at particularly high risk of adverse HF outcomes,^53,57–60 but they are not considered iron deficient by the guideline definition. Higher serum ferritin levels identify patients at higher risk, because they are indicators of the coexisting inflammatory state that reflects prognostically relevant comorbidities.^53,59–61

Confounding Effects of Foundational Drugs for HF

To complicate matters further, many of the foundational drugs used for the treatment of HF and a reduced ejection fraction may influence hemoglobin, while simultaneously distorting the levels of iron biomarkers that are used to identify an iron-deficiency state. In randomized controlled trials, treatment of patients with HF with an angiotensin-converting enzyme inhibitor has been accompanied by modest decreases in hemoglobin,⁶² and carvedilol (a nonselective β-blocker) is more likely to cause small decreases in hemoglobin than metoprolol tartrate (a selective β-blocker).⁶³ In both instances, the effects on red blood cell counts were seen early in treatment and were sustained for the duration of therapy. These findings are likely the result of the withdrawal of a direct stimulatory effect of angiotensin II and β2 adrenergic receptors on the proliferation of erythroid precursors.^64,65 The action of neurohormonal blockade to lower hemoglobin can be accompanied by direct or reactive changes in iron biomarkers; α₁-adrenergic blockade can directly decrease hepcidin synthesis,⁶⁶ and angiotensin receptor blockade may increase serum ferritin levels.⁶⁷

In marked contrast, when compared with enalapril, long-term use of sacubitril/valsartan is accompanied by higher levels of hemoglobin, and this effect is accompanied by decreases in serum levels of iron, hepcidin, and ferritin.⁶⁸ It might be hypothesized that potentiation of natriuretic peptides could stimulate erythropoietin and erythroid precursors through a cyclic guanosine monophosphate–dependent mechanism,^69,70 but potentiation of cyclic guanosine monophosphate with vericiguat acts to worsen anemia with HF, rather than alleviating it.⁷¹ Instead, the pattern of changes in iron biomarkers with sacubitril/valsartan is consistent with an effect of neprilysin inhibition to mute inflammation,^72–74 thereby interfering with a major stimulus for the synthesis of hepcidin and ferritin. The suppression of serum levels of hepcidin and ferritin by neprilysin inhibition⁶⁸ would serve to ameliorate functional iron trapping, thus promoting iron use to achieve enhanced erythropoiesis.

In randomized controlled trials, sodium-glucose cotransporter 2 (SGLT2) inhibitors also reduced serum hepcidin and ferritin levels and TSAT in patients with HF, and their use is accompanied by an increase in erythropoietin and a rapid and meaningful increase in hemoglobin and hematocrit that is sustained for the duration of treatment.^75,76 As in the case of neprilysin inhibition, it seems likely that an anti-inflammatory effect of SGLT2 inhibition^77,78 is responsible for the suppression of serum levels of hepcidin and ferritin, and, therefore, sufficient iron may be released from macrophage or hepatocyte stores to support and sustain a notable erythrocytic response to the enhanced synthesis of erythropoietin.^79,80 A facilitatory effect on iron mobilization is supported by the finding that patients with HF who fulfill current criteria for iron deficiency nevertheless show a robust erythropoiesis after SGLT2 inhibition, even in the absence of oral or IV iron supplementation.⁷⁵ Given its important limitations, it is not surprising that the current definition of iron deficiency does not distinguish patients who do or do not manifest an erythrocytic response to SGLT2 inhibitors.⁷⁵

The findings with neprilysin and SGLT2 inhibitors suggest that many patients with HF have adequate total body iron stores, but the release of iron from these stores is impaired.^79,80 By acting to lower hepcidin and ferritin synthesis, the anti-inflammatory effects of these drugs alleviate functional iron deficiency, thus promoting iron mobilization and supporting erythropoiesis. However, the action of these drugs on circulating iron biomarkers causes patients to appear to be more iron deficient, if the identification of an iron-deficiency state were dependent on the measurement of serum ferritin or hepcidin levels. These findings raise additional questions about the adequacy of our current definition of iron-deficient HF, especially when applied to patients receiving recommended drug treatments for left ventricular systolic dysfunction.^60,79,80

Defining Iron Deficiency by Response to Treatment Rather Than Iron Stores

Despite the observed relationships between TSAT (but not serum ferritin level) and the prognosis of patients with HF, it is not clear that the ability to predict future HF outcomes should be used to bolster an argument that any given biomarker be used to identify an iron-deficiency state, because hypoferremia or a low TSAT may simply reflect important measured or unmeasured comorbidities and not represent a meaningful therapeutic target. In light of this possibility, it is important to remember that serum iron biomarkers are assessed in the hope that they might provide reliable estimates of cytosolic levels of reactive iron. Therefore, physicians should be able to define the subgroup of patients who have meaningful depletion of this labile reactive pool by identifying the specific individuals who respond particularly well to therapeutic efforts to replete the level of cytosolic iron with the use of IV iron.

If the relevant cytosolic pool of reactive iron resides within cardiomyocytes,^81,82 the proper end point for trials that are designed to assess the effect of iron supplementation would be the magnitude of the reduction in risk of cardiovascular death or HF hospitalizations. On the other hand, if the relevant cytosolic pool resides in skeletal muscle, the proper end point for trials to evaluate the effect of iron supplementation would be improvement in functional capacity and exercise tolerance, particularly in patients with more severe symptoms (ie, class III HF).¹ The symptoms of HF are closely associated with skeletal muscle dysfunction due to inflammation and atrophy, which can be aggravated by iron deficiency.^83–85 A diagnostic approach that is informed by a therapeutic response has also been applied to patients with end-stage kidney disease, in whom an iron-deficiency state has been defined not by the measurement of iron stores but by biomarkers that identify the subset of patients who either show an erythrocytic response to IV iron or a reduction in cardiovascular events.^11,86

It is therefore noteworthy that TSAT, but not serum ferritin, predicts the effect of IV iron to reduce major adverse HF outcomes (Table). In a meta-analysis of 10 randomized controlled trials, the most marked reduction in the combined risk of cardiovascular death and total HF hospitalizations after IV iron was seen in patients with a TSAT <20% (risk ratio, 0.67 [0.49–0.92]), with no observed risk reduction in patients with a TSAT ≥20% (risk ratio, 0.99 [0.74–1.30]).^87,88 The lower the baseline TSAT, the greater the treatment effect (P_interaction=0.019).⁸⁷ In accordance, IV iron reduced the risk of major HF outcomes in IRONMAN (Intravenous Iron Treatment in Patients With Heart Failure and Iron Deficiency) and AFFIRM-AHF (Study to Compare Ferric Carboxymaltose With Placebo in Patients With Acute Heart Failure and Iron Deficiency), with baseline TSATs ≈15%,^3,4 but not in HEART-FID (Randomized Placebo-Controlled Trial of FCM as Treatment for Heart Failure With Iron Deficiency), with baseline TSAT ≈23% to 24%.⁵ Nearly two thirds of patients were enrolled in HEART-FID purely because they had a serum ferritin level within the normal range (20–99 μg/L), often in the absence of anemia⁵; these patients fulfilled current guideline-recommended criteria for an iron-deficiency state, but they did not show a reduction in HF outcomes with IV iron. In IRONMAN, patients who qualified solely because of a serum ferritin level <100 μg/L showed no significant risk reduction with the use of IV iron.⁴ However, a baseline TSAT <20% identified patients with the largest risk reduction with IV iron in a meta-analysis of 10 trials (including IRONMAN and AFFIRM-AHF) and in an individual patient-level analysis of 3 trials (including AFFIRM-AHF and HEART-FID; Table).^87,88

Table.

Influence of Transferrin Saturation and Serum Ferritin Levels at Baseline on the Effect of Intravenous Iron to Reduce the Risk of Cardiovascular Death and Heart Failure Hospitalizations

Although these findings with respect to TSAT might at first appear to be the result of post hoc subgroup analyses, it should be noted that a TSAT <20% represented a distinct prespecified pathway for eligibility for all trials of iron supplementation. Furthermore, the consistency of findings across all available trials based on a common eligibility pathway, coupled with a significant dose–response relationship, provides compelling support, and a strength of evidence that is more persuasive than would be the finding of a positive effect in an isolated dedicated trial with a marginally significant result.

Further analyses are needed to determine whether a TSAT <20% alone might be useful in predicting the benefits of IV iron on symptoms, health status, or exercise tolerance, rather than on cardiovascular death and HF hospitalizations. However, serum ferritin levels do not identify patients with marked exercise intolerance, who are most likely to show an improvement in functional capacity after IV iron.^1,2 In addition, patients with a serum ferritin level >400 μg/dL have not been enrolled in any trial of iron supplementation in HF,⁴ presumably because of concerns that patients with hyperferritinemia have intracellular cytosolic iron overload and might respond adversely to IV iron supplementation.⁸⁹

The finding that TSAT <20% identifies patients with HF who respond favorably to IV iron supplementation should be contrasted with the observations of PIVOTAL (UK Multicentre Open-Label Randomised Controlled Trial of IV Iron Therapy in Incident Haemodialysis Patients),⁸⁶ which evaluated the effects of IV iron supplementation in patients with end-stage kidney disease on hemodialysis receiving ESAs. The primary hypothesis tested was whether IV iron might reduce major cardiovascular events not by alleviating iron deficiency, but by reducing the dose of ESAs, which are known to be produce dose-dependent increases in the risk of cardiovascular death and hospitalizations.^90,91 IV iron acts to circumvent the transient intracellular imbalances in iron homeostasis that can occur during rapid pharmacologically stimulated erythropoiesis, thus potentiating the erythrocytic response to lower doses of ESAs.⁴⁸ PIVOTAL demonstrated that in the absence of demonstrable iron deficiency, IV iron therapy led to a reduction in the dose of ESAs, which was paralleled by a decreased risk of major adverse cardiovascular events. Given the effect of IV iron to spare ESA-related toxicity independent of iron deficiency, the criteria for iron supplementation in PIVOTAL—TSAT <40% and serum ferritin level ≤700 μg/L—cannot be applied to patients with HF and a reduced ejection fraction. Nevertheless, the results of PIVOTAL have encouraged nephrologists to strengthen their commitment to a ferritin-based strategy as a guide to IV iron supplementation, thereby widening the conceptual and pragmatic chasm that exists between the therapeutic goals of patients with HF compared with those on hemodialysis.

Assessment of Soluble Transferrin Receptor as an Alternative Biomarker

An iron-deficiency state is typically accompanied not only by a decrease in TSAT to <20%, but also by an increase in soluble transferrin receptor (sTfR), reflecting the enhanced shedding of the cell-surface TFR1 (transferrin receptor protein 1), which is upregulated when cytosolic levels of iron are low.^92,93 Serum levels of sTfR rise when iron stores are depleted,⁹⁴ and a high sTfR has been used to identify patients with HF who have absent iron stores in the bone marrow.⁵⁶ Some have also proposed calculating a ratio of sTfR to hepcidin or a ratio of sTfR to ferritin to enhance the discrimination of patients with an iron-deficiency state.^95–97 However, circulating sTfR reflects the shedding of TFR1 primarily from erythroblasts, and not cardiomyocytes or skeletal myocytes.^40,98 Furthermore, assays of sTfR are not well-standardized or clinically available, and no study has evaluated the use of measuring sTfR as a means of identifying patients with HF who would show the most marked reduction in HF events or improvements in functional capacity after iron supplementation (Figure). In future trials, the assessment of a ratio of sTfR to TSAT in the bloodstream might be of particular interest, even though this metric would not specifically reflect a myocyte-specific depletion of the labile pool of highly reactive iron.

SUMMARY AND CONCLUSIONS

The current definition of iron deficiency in HF—a serum ferritin level <100 μg/L (regardless of TSAT) or a serum ferritin level of 100 to 299 μg/L if the TSAT is <20%—is based on a definition developed 25 years ago to encourage the use of IV iron supplementation in patients with end-stage kidney disease treated with ESAs. That definition does not reliably identify patients with absent iron stores in the bone marrow, and it includes individuals with TSATs and serum ferritin levels in the normal range (>20% and 20–100 mg/L, respectively) who have an excellent prognosis and do not respond favorably to IV iron therapy.

The most evidence-based definition of iron deficiency is hypoferremia, as evidenced by a TSAT <20%. These patients are demonstrably iron deficient on bone marrow examination, exhibit an improvement in symptoms and functional capacity, and show the largest reduction in the risk of cardiovascular death or HF hospitalization with IV iron therapy. We propose that the current ferritin-driven definition of iron deficiency in HF be abandoned, and that a definition based on hypoferremia (TSAT <20%) be adopted. Further work is needed to confirm, refute, or modify our proposal.

ARTICLE INFORMATION

Sources of Funding

None.

Disclosures

Dr Packer reports personal fees for consulting from 89bio, AbbVie, Actavis, Altimmune, Alnylam, Amarin, Amgen, Ardelyx, AstraZeneca, Attralus, Biopeutics, Boehringer Ingelheim, Caladrius, Casana, CSL Behring, Cytokinetics, Imara, Lilly, Medtronic, Moderna, Novartis, Pharmacocosmos, Reata, Relypsa, and Salamandra, all outside the submitted work. Dr Anker reports grants from Vifor and Abbott Vascular; personal fees for consultancies, trial committee work, or lectures from Vifor, Abbott Vascular, Actimed, Amgen, Astra Zeneca, Bayer, Boehringer Ingelheim, Brahms, Cardiac Dimensions, Cardior, Cordio, CVRx, Cytokinetics, Edwards, Farraday Pharmaceuticals, GlaxoSmithKline, HeartKinetics, Impulse Dynamics, Occlutech, Pfizer, Regeneron, Repairon, Scirent, Sensible Medical, Servier, Vectorious, and V-Wave; and is a named co-inventor of 2 patent applications regarding MR-proANP (DE 102007010834 and DE 102007022367), but does not benefit personally from the patents, all outside the submitted work. Dr Butler reports personal consulting fees from Abbott, American Regent, Amgen, Applied Therapeutic, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Cardiac Dimension, Cardior, CVRx, Cytokinetics, Janssen, Daxor Edwards, Element Science, Eli Lilly, Innolife, Impulse Dynamics, Imbria, Inventiva, Lexicon, LivaNova, Medscape, Medtronics, Merck, Occlutech, Novartis, Novo Nordisk, Pfizer, Pharmacosmos, Pharmain, Roche, Secretome, Sequana, SQ Innovation, Tenex, Tricoq, and Vifor; and honoraria from Novartis, Boehringer Ingelheim-Lilly, Astra Zeneca, Impulse Dynamics, and Vifor, all outside the submitted work. Dr Mentz reports grants from American Regent, AstraZeneca, Amgen, Bayer, Merck, Novartis, Zoll, and Cytokinetics; and personal consulting fees from Pharmacosmos, Vifor, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Merck, Novartis, Abbott, Medtronic, Zoll, Boston Scientific, Cytokietics, Respicardia, Roche, Vifor, Sanofi, and Windtree, all outside the submitted work. Dr Cleland reports grants from Bristol Myers Squibb, British Heart Foundation, Medtronic, Pharma Nord, Vifor, and Pharmacosmos; personal consulting fees from Abbott, Biopeutics, Innolife, NI Medical, Novartis, and Servier; honoraria for committee or advisory boards from Idorsia and Medtronic; honoraria for lectures from AstraZeneca and Boehringer Ingelheim; and stock options or holdings in Heartfelt Limited and Viscardia, all outside the submitted work. Dr Kalra reports grants from Pharmacosmos and the British Heart Foundation; personal consulting fees from Amgen, Boehringer Ingelheim, Pharmacosmos, Servier, and CSL Vifor; and honoraria for lectures from AstraZeneca, Bayer, Novartis, Pfizer, Pharmacosmos, CLS Vifor, and Amgen, all outside the submitted work. Dr Ponikowski reports personal fees for consultancies, trial committee work, or lectures from Astra Zeneca, Bayer, Boehringer Ingelheim, Pfizer, Vifor Pharma, Amgen, Servier, Novartis, Novo Nordisk, Pharmacosmos, Abbott Vascular, Radcliffe Group, and Charite University, all outside the submitted work.