Iron Deficiency and Red Cell Indices in Patients with Heart Failure: Reply

We would like to thank Zeng for the interest in our study1 investigating iron deficiency (ID) and abnormalities regarding red cell indices (RCI) in patients with heart failure (HF) with either reduced or preserved ejection fraction. In the correspondence, the author raises concerns regarding the appropriateness of the definition of ID [based on ferritin and transferrin saturation (TSAT)] applied in the study. We fully agree with the author that serum ferritin levels can be affected by different confounding factors, such as chronic low-grade inflammation.2 However, we would like to emphasize that associations between iron parameters and inflammation in HF are not that obvious as those observed in e.g. chronic kidney disease. For example, in patients with chronic systolic HF, circulating ferritin and hepcidin (a key regulator of iron metabolism) do not correlate with serum high-sensitivity C-reactive protein, but are inversely related to serum interleukin 6.3 Moreover, the definition of ID based on serum ferritin and TSAT (which was used in our study) is recommended in the 2016 European Society of Cardiology guidelines for the diagnosis and treatment of acute and chronic HF.4 This definition was applied in several observational5, 6 and major interventional7-9 studies regarding ID in HF, and allows to identify patients who could benefit from intravenous iron therapy with ferric carboxymaltose. Furthermore, this definition includes higher cut-off value for serum ferritin (as compared with lower limit of normal for healthy persons without chronic disease) as this protein is an acute phase reactant that increases with inflammation.10 There are attempts to develop more precise and pathophysiology-oriented definitions of ID based on other circulating biomarkers of iron status (such as soluble transferrin receptor mentioned by the author),11 but they were not applied in interventional studies to date, and are not used in decision-making in everyday clinical practice. The author also emphasizes that laboratory measurements in our study were performed in different laboratories of five participating centres (we analysed five cohorts from three countries) and the measurements of RCI are presumably not perfectly comparable. We are aware that the methodology for the measurement of RCI may have varied across these five centres, but we would like to emphasize that in our study we applied the definitions of anaemia and abnormal RCI according to the official documents of the World Health Organization.12, 13 These limits of normal are considered ‘universal’ (as they originate from large epidemiological studies) and are recommended in large, international groups of patients, regardless of the type of blood analyser used in participating centres. Finally, the author emphasizes that one of the possible sources of bias affecting survival analyses is potential therapy with erythropoiesis-stimulating agents or intravenous iron (e.g. ferric carboxymaltose). We would like to stress that no patient received blood transfusions, erythropoietin therapy, or intravenous iron therapy at the time of inclusion in the study.1