Anaemia predicts iron homoeostasis dysregulation and modulates the response to empagliflozin in heart failure with reduced ejection fraction: the EMPATROPISM-FE trial

Abstract

Background and aims: Sodium-glucose cotransporter 2 inhibitors (SGLT2i) impact iron metabolism in patients with heart failure but mechanisms are incompletely understood. This post hoc analysis explored interrelations between iron homeostasis, cardiac structure/function, exercise capacity, haematopoiesis, and sympathetic activity at baseline, and the effects of 6-month treatment with empagliflozin vs. placebo by anaemia status in EMPATROPISM-FE study participants.

Methods: Myocardial iron content (MIC, estimated by cardiac magnetic resonance T2* imaging), left ventricular (LV) volumes and LV ejection fraction (LVEF), exercise capacity, laboratory iron markers (LIM), haemoglobin/haematocrit, erythropoietin, and plasma norepinephrine were determined at baseline and 6 months.

Results: At baseline, 24/80 participants (30%) had anaemia (haemoglobin < 13/<12 mg/dL in men/women). Patients with vs. without anaemia had higher T2* (indicating lower MIC, P < .001), lower peak oxygen consumption (VO2max, P = .024) and hepcidin (P = .017), and higher erythropoietin (P = .040) and norepinephrine (P = .016). Across subgroups, lower MIC correlated with higher LV volumes (P < .01) and norepinephrine (P < .001), and lower LVEF (P < .01), VO2max (P < .001) and haemoglobin/haematocrit (P < .001). Associations with LIM were poor (all P > .10). Empagliflozin increased MIC (P < .012), improved exercise capacity, and activated haematopoiesis. Changes in LIM and norepinephrine suggested progressive systemic iron depletion and sympatholysis. LV reverse remodelling was greater in individuals with anaemia.

Conclusions: Dysregulated cellular iron uptake/availability may be a shared mechanism in myocardial structural/functional impairment, reduced exercise capacity, and restricted haematopoiesis in heart failure, which are worse in patients with anaemia, and improve with empagliflozin. Empagliflozin increases MIC and decreases norepinephrine. Given this inverse association, sympatholysis may help explain the diverse cardiac and systemic benefits from SGLT2i therapy.

Clinical trial registration: NCT03485222 (www.clinicaltrials.gov).

Keywords: Anaemia; Cardiac remodelling; Empagliflozin; Exercise capacity; Heart failure; Iron homoeostasis; Neurohormonal activation.

Structured Graphical Abstract

Among 80 participants of the EMPATROPISM-FE study, who had symptomatic heart failure and a left ventricular ejection fraction (LVEF) < 50%, 30% had anaemia at baseline. At baseline and 6 months after randomization to empagliflozin or placebo, all patients underwent cardiac magnetic resonance imaging (MRI) for determination of T2* as an estimate of myocardial iron content and for quantification of cardiac morphology and function, and had assessment of exercise capacity using maximum oxygen consumption (peak VO₂) and 6 min walking distance (6MWD) and laboratory evaluation of systemic iron status, norepinephrine, and haematopoiesis (left panel). Comparison of subgroups at baseline revealed that patients with vs. without anaemia had higher T2* [indicating lower myocardial iron content (MIC)], peak VO₂, and hepcidin, and higher erythropoietin and norepinephrine (mid panel, top). Across subgroups, lower MIC correlated with higher left ventricular (LV) volumes and norepinephrine, and lower LVEF, peak VO₂ and haemoglobin/haematocrit, while associations with systemic iron status (e.g., transferrin saturation) were poor (mid panel, bottom). Empagliflozin increased MIC and haematopoiesis, and decreased norepinephrine, regardless of anaemia status. LV reverse remodelling, an increase in erythropoietin, and progressive systemic iron depletion were greater in individuals with anaemia, while exercise capacity improved more in those without anaemia (right panel). Close interrelations between MIC and norepinephrine levels suggest that sympatholytic effects of sodium-glucose cotransporter 2 (SGLT2) inhibitors might help explain their diverse cardiac and systemic benefits. LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume.

Figure 1

Correlations and regression lines for baseline myocardial T2* and markers of systemic iron status, hepcidin, red blood cell indices, norepinephrine, and erythropoietin. Pearson correlation coefficients are displayed for subgroups with and without anaemia. Myocardial T2* as a surrogate of tissue iron content was unrelated to ferritin and soluble transferrin receptor concentrations (A, B), hepcidin (C), and erythropoietin (E), but was associated across subgroups with haemoglobin and norepinephrine (D, F)

Figure 2

Correlations and regression lines for baseline myocardial T2* and indices of left ventricular structure and function and exercise capacity. Pearson correlation coefficients are displayed for subgroups with and without anaemia. Myocardial T2* as a surrogate of tissue iron content was associated across subgroups with LV volumes and LVEF (A–C), and with peak VO₂ (E). Associations were stronger in the anaemic subgroup. T2* and LV mass showed no significant interrelation (D), and T2* and 6 min walking distance appeared related only in the non-anaemic subgroup (F). LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LV mass, left ventricular mass; peak VO₂, maximum oxygen consumption

Figure 3

Correlations and regression lines for baseline hepcidin and erythropoietin levels with indices of systemic iron status, with haemoglobin, and between each other. Pearson correlation coefficients are displayed for subgroups with and without anaemia. Graphs illustrate associations between hepcidin and ferritin, transferrin, and haemoglobin, (A–C), and between erythropoietin and soluble transferrin receptor, haemoglobin, and hepcidin levels (D–F)

Figure 4

Changes in myocardial T2*, laboratory iron markers and hepcidin after 6-month treatment with empagliflozin or placebo by anaemia status. Variables are reported as change from, or multiple of, baseline values, as appropriate. P-values for the interaction term (P_interaction) and the treatment effects in subgroups by anaemia status (P_treatment, italic font) are given for effects with a P_interaction < .100. For effects with a P_interaction ≥ .100, P_treatment-values refer to the effects of empagliflozin vs. placebo in the entire cohort (regular font). All results are adjusted for baseline values and anaemia status. CI, confidence interval

Figure 5

Changes in indices of left ventricular structure/function (A) and of exercise capacity (B) in patients with vs. without anaemia after 6-month treatment with empagliflozin or placebo. Variables are reported as change from baseline. P-values for the interaction term (P_interaction) and the treatment effects in subgroups by anaemia status (P_treatment, italic font) are given for effects with a P_interaction < .100. For effects with a P_interaction ≥ .100, P_treatment-values (regular font) refer to the effects of empagliflozin vs. placebo in the entire cohort. All results are adjusted for baseline values and anaemia status. CI, confidence interval; LV, left ventricular; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; VO₂, oxygen consumption

Figure 6

Changes in red blood cell indices and erythropoietin in patients with vs. without anaemia after 6-month treatment with empagliflozin or placebo. Variables are reported either as change from, or multiple of, baseline values, as appropriate. P-values for the interaction term (P_interaction) and the treatment effects in subgroups by anaemia status (P_treatment, italic font) are given for effects with a P_interaction < .100. For effects with a P_interaction ≥ .100, P_treatment-values refer to the effects of empagliflozin vs. placebo in the entire cohort (regular font). All results are adjusted for baseline values and anaemia status. CI, confidence interval