Criteria for Iron De ﬁ ciency in Patients With Heart Failure

BACKGROUND Guidelines on heart failure (HF) de ﬁ ne iron de ﬁ ciency (ID) as a serum ferritin < 100 ng/mL or, when 100-299 ng/mL, a transferrin saturation (TSAT) < 20%. In ﬂ ammation (common in HF) may hinder interpretation of serum ferritin. OBJECTIVES This study sought to investigate how different de ﬁ nitions of ID affect its prevalence and relationship to prognosis in ambulatory patients with chronic HF. METHODS Prevalence, relationship with patients ’ characteristics, and outcomes of various ID de ﬁ nitions were evaluated among patients with HF referred to a regional clinic (Hull LifeLab) from 2001 to 2019. RESULTS Of 4,422 patients with HF (median age 75 years [range: 68-82 years], 60% men, 32% with reduced left ventricular ejection fraction), 46% had TSAT < 20%, 48% had serum iron # 13 m mol/L, 57% had serum ferritin < 100 ng/mL, and 68% ful ﬁ lled current guideline criteria for ID, of whom 35% had a TSAT > 20%. Irrespective of de ﬁ nition, ID was more common in women and those with more severe symptoms, anemia, or preserved ejection fraction. TSAT < 20% and serum iron # 13 m mol/L, but not guideline criteria, were associated with higher 5-year mortality (HR: 1.27; 95% CI: 1.14-1.43; P < 0.001; and HR: 1.37; 95% CI: 1.22-1.54; P < 0.001, respectively). Serum ferritin < 100 ng/mL tended to be associated with lower mortality (HR: 0.91; 95% CI: 0.81-1.01; P ¼ CONCLUSIONS Different de ﬁ nitions of ID provide discordant results for prevalence and prognosis. De ﬁ nitions lacking speci ﬁ city may attenuate the bene ﬁ ts of intravenous iron observed in trials while de ﬁ nitions lacking sensitivity may exclude patients who should receive intravenous iron. Prespeci ﬁ ed subgroup analyses of ongoing randomized trials address this issue.

I ron deficiency (ID), inferred from the results of blood tests, is common in chronic heart failure (HF) and, in the presence or absence of anemia, is associated with poorer quality of life, exercise capacity, and prognosis. [1][2][3][4] Many definitions of ID have been proposed, but consensus is lacking on which should be used in clinical practice for patients with HF.
The World Health Organization defines ID as a serum ferritin <15 ng/mL, and most clinical laboratories define ID as <30 ng/mL. However, international guidelines on HF define ID as a serum ferritin <100 ng/mL or, when ferritin is 100-299 ng/mL, a transferrin saturation (TSAT) <20%. 5 an increase in serum ferritin may occur even in the presence of ID. 9 Bone marrow biopsy, the gold-standard for diagnosing ID, may demonstrate ID even when ferritin is high. 10 Observational studies suggest that serum iron concentration and TSAT might both be more strongly associated with prognosis than serum ferritin and might be a better guide to which patients benefit from IV iron. [10][11][12][13][14] Accordingly, we investigated the prevalence, associations, and prognostic significance of ID using diverse criteria in a large cohort of patients attending an HF clinic.   with HF. 10 We also explored prevalence and outcomes of additional thresholds (ferritin <30 ng/mL, ferritin <100 ng/mL, ferritin <300 ng/mL, serum iron #10 mmol/L, and TSAT <30%).  Values are n (%) or median (IQR). Guideline ID criteria were ferritin <100 ng/mL or TSAT <20% if ferritin was 100-299 ng/mL. ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin receptor blocker; BB ¼ beta-blocker; BMI ¼ body mass index; COPD ¼ chronic obstructive pulmonary disease; eGFR ¼ estimated glomerular filtration rate; HFmrEF ¼ heart failure with mid-range ejection fraction; HFpEF ¼ heart failure with preserved ejection fraction; HFrEF ¼ heart failure with reduced ejection fraction; hs-CRP ¼ high-sensitivity C-reactive protein; ID ¼ iron deficiency; IHD ¼ ischemic heart disease; MRA ¼ mineralocorticoid receptor antagonist; NT-proBNP ¼ N-terminal pro-B-type natriuretic peptide; NYHA ¼ New York Heart Association; TSAT ¼ transferrin saturation.

METHODS
older, were more likely to be women, and were more likely to have hypertension and atrial fibrillation, but despite these differences, they had a lower median plasma NT-proBNP and similar renal function.

PREVALENCE OF ID AND ANEMIA AND ASSOCIATION
WITH PATIENT CHARACTERISTICS. The prevalence of ID ranged from 44% to 68% depending on the definition ( Table 1)  Compared with those without anemia, those with anemia had a higher prevalence of ID, irrespective of subgroups. *Significant difference (P < 0.05) between patients subgroups; †P ¼ NS. HFmrEF ¼ heart failure with mid-range ejection fraction; HFpEF ¼ heart failure with preserved ejection fraction; HFrEF ¼ heart failure with reduced ejection fraction; hs-CRP ¼ high-sensitivity C-reactive protein; LVSD ¼ left ventricular systolic dysfunction; NT-proBNP ¼ N-terminal pro-B-type natriuretic peptide; NYHA ¼ New York Heart Association functional class.
Masini et al Compared with those with a higher TSAT, those with a TSAT <20%, had a lower hemoglobin concentration, which was similar across the range of serum ferritin. Patients with a TSAT <20% and serum ferritin $300 ng/mL had the highest prevalence of anemia. These patients were more likely to be men, had lower body mass index, had higher median NT-proBNP and high-sensitivity C-reactive protein, and had worse renal function compared with those with ferritin <300 ng/mL (Supplemental Table 1). The guideline criteria for ID were met by 3,011 (68%) patients, but 1,079 (36%) of them had a TSAT >20% ( Figure 2). In univariable analysis, lower TSAT and serum iron, but higher serum ferritin, were associated with a higher all-cause and CV mortality (  Table 4).
As categorical variables, TSAT <20% (vs $20%) and serum iron #13 mmol/L (vs >13 mmol/L) were associated with a higher all-cause and CV mortality in the univariable model (  Iron Deficiency Criteria in Heart Failure mortality and tended to be associated with lower allcause mortality. Ferritin <300 ng/mL was associated with both lower all-cause and CV mortality ( Table 2).  Figure 6 shows the unadjusted HRs for all-cause death for each decile of TSAT, serum iron, and ferritin.

No significant interaction was found in the adjusted models between definitions of ID and HF
phenotypes for all-cause or cardiovascular mortality.
Patients with a serum ferritin $300 ng/mL and a TSAT <20%, who do not fulfill guideline criteria for ID, had the highest risk of death, while those with a ferritin <100 ng/mL with a TSAT $20%, who do fulfill HF guideline criteria for ID, had a similar prognosis to those with ferritin 100-299 ng/mL and TSAT $20% In the human body, iron is central to hematopoietic (erythropoiesis, oxygen transport, and storage) and nonhematopoietic processes (substrate utilization and mitochondrial energy production). 18   system as a result of inflammation, and subclinical gastrointestinal blood loss, which might be common in those taking antithrombotic agents. 21 To our knowledge, this is the first study to assess systematically the prevalence and prognostic implications of different definitions of ID across all HF phenotypes in a large population of patients with chronic HF. Our findings confirm that ID is common in patients with HF, in line with previous reports. [2][3][4]11 However, many previous studies have included only patients with HFrEF and assessed ID using only the guideline definition. We extend these findings and suggest that many patients with HFpEF also have ID, which might reflect their advanced age, the high proportion of women, and the high comorbidity burden.
The relationship between serum iron biomarkers and adverse outcomes in those with ID remains uncertain. ID defined by guideline criteria was an independent predictor of mortality in a study of 546 patients with HFrEF and in another study of 1,506 patients with HF, of whom 87% had HFrEF (LVEF #45%). 4,22 However, others have questioned both the diagnostic and prognostic utility of the guideline criteria in patients with HF. [10][11][12][13][14] In a European multicenter study of 1,821 patients with chronic HF, a TSAT <20% but not a ferritin <100 ng/mL independently predicted mortality. 12 In another study, in which ID was defined by bone marrow iron staining, 10 only a serum iron #13 mmol/L or TSAT #19.8%, but not guideline criteria, predicted bone marrow ID in patients with HFrEF (defined by an EF #45%; n ¼ 42) undergoing coronary artery bypass. In a previous analysis of the Hull LifeLab cohort, the highest quintiles of ferritin had the worst all-cause and CV mortality. 13 The current analysis includes many more patients, has longer These findings support the use of TSAT rather than ferritin to select those more likely to benefit from IV .20 SHR: 0.98 (95% CI: 0.85-1.14) Kaplan-Meier cumulative curves for all-cause mortality and cumulative incidence curves using Fine and Gray method for cardiovascular (CV) mortality according to guideline iron deficiency (ID) criteria. SHR ¼ subdistribution HR.   Serum iron is almost entirely transferrin bound, and therefore a close association between serum iron and TSAT is not surprising. 13  Kingdom. 38 We expect that few patients had received IV iron prior to June 2019. We did not investigate for the presence of genetic mutations associated with ID; however, they are rare and international guidelines do not recommend specific genetic testing for patients with ID. We enrolled patients during a period of 20 years, and evidence-based therapies for HF have evolved. However, applying sensitivity analysis, the period of enrolment did not affect our main results.
This is a single-center study including a predominantly white British population, and therefore data should be extrapolated with caution to a population with characteristics different from ours.

CONCLUSIONS
Irrespective of how it is defined, ID is common in patients with HF. When defined by current guideline criteria, ID was not associated with poor outcome; indeed, lower serum ferritin concentrations were associated with a better survival. TSAT <20% and serum iron #13 mmol/L were associated with a higher mortality and this was independent of HF phenotype.  Iron Deficiency Criteria in Heart Failure F E B R U A R Y 1 , 2 0 2 2 : 3 4 1 -3 5 1