Background
Iron deficiency (ID) is one of the main causes of chronic disease. ID affected more than 10% of the world’s population in 2015 and is responsible for almost half of all anaemia worldwide, which is the first cause of health impairment [
1]. ID anaemia is indeed one of the leading causes of years of living with disability [
1,
2]. However, ID is not only responsible for anaemia, it also causes fatigue and muscle weakness, independently of anaemia [
3]. This has been notably described in patients with heart failure, for whom ID is associated with impaired cardiac function and increased mortality [
4,
5]. ID may also be important in the prognosis of critically ill patients. First ID is frequent on admission in intensive care units (ICU), affecting 20–40% of the patients [
6‐
8]. ID on admission has been associated with longer ICU stay, more renal failure or ICU-acquired infections [
6,
8]. Because critically ill patients are exposed to frequent blood sampling and other causes of blood loss (including surgery, extracorporeal circuit, invasive procedures, etc.) [
9,
10], they are exposed to high iron losses. Daily blood loss in these patients can be as high as 128 ml of blood per day [
10], which represents 2–3 times the daily need for erythropoiesis [
11]. ID is thus expected to be frequent on ICU discharge and may impair the post-ICU rehabilitation. Unfortunately, ID is difficult to diagnose using conventional markers (i.e. ferritin) in the presence of inflammation. Prevalence of ID on discharge is thus probably underestimated, with observed prevalence of less than 10% [
12]. Many markers have been proposed to diagnose ID in the presence of inflammation [
6‐
8,
13]. However, there is no consensus and the best markers (i.e. the percentage of hypochromic red cells) [
14] are not usable in case of blood transfusion, which is very frequent in critically ill patients. Thus, the remaining proposed marker in this context is the ratio between soluble transferrin receptor and the log (ferritin) (sTfR/log ferritin) [
15].
In recent decades, the understanding of iron metabolism has been greatly improved by the discovery of its master regulator, hepcidin [
16]. A low hepcidin level may indicate ID in the critically ill patient [
11,
17]. Because hepcidin may indicate ID and because ID is responsible for muscular weakness and fatigue, we hypothesised that ID diagnosed according to low plasma hepcidin or high sTfR/log(ferritin) on ICU discharge could predict a poorer outcome and quality of life (QOL) in critically ill patients. We used the prospective Frog-ICU cohort [
18,
19] to determine whether ID diagnosis using these markers, measured at ICU discharge, were predictive of one-year mortality and QOL in critically ill patients discharged alive from ICU.
Discussion
In this prospective cohort of critically ill patients we observed that ID diagnosed by low hepcidin or high sTfR/log ferritin ratio is frequent at ICU discharge, affecting almost 40% of the patients and that ID is an independent predictor of one-year mortality. In addition, severe ID defined by hepcidin < 10 ng/l was also an independent predictor of one-year poor physical QOL.
This relatively high prevalence of ID is consistent with the prevalence observed in the general population (around 30–40% in menstruating women or preschool children) [
21,
22] or in populations of patients such as those in the perioperative period (prevalence between 35 and 70%) [
23] or patients with cardiac disease (prevalence around 40–50%) [
24]. However, there are very few data available in critically ill patients, especially at ICU discharge. On admission to ICU, the prevalence of ID varies from 9% to almost 40% [
6,
8,
17,
25‐
27]. These variations in ID prevalence are mainly related to the biological test that is used to diagnose ID. In the general population, ferritin is the gold standard for ID diagnosis [
22]. However, ferritin synthesis is induced by inflammation, decreasing its interest as a diagnostic tool of ID in such conditions. Inflammation is present in virtually all critically ill patients and this probably explains why the prevalence of ID, diagnosed by low ferritin (i.e. ferritin < 100 μg/l) is < 10% on admission to ICU [
27] or at discharge [
12]. This is why other markers are proposed to diagnose ID in the presence of inflammation. Red blood cell indices, such as the percentage of hypochromic red cells or reticulocyte haemoglobin content are considered to be the best variables available to diagnose ID in the presence of inflammation [
14]. Unfortunately, these parameters cannot be used to diagnose ID when patients receive blood transfusions, which is often the case in critically patients. sTfR and its ratio to log(ferritin) has been proposed to diagnose ID in the presence of inflammation [
14,
15]. In this cohort, ID defined by a low sTfR/log(ferritin) ratio makes it possible to diagnose ID in an important number of critically ill patients discharged from ICU and these patients had a poor outcome. However, this assay has some limitations, it is quite expensive and is not standardized [
14]. The reproducibility of measures remains uncertain. Hepcidin quantification appears promising and a number of laboratories are collaborating towards global harmonization of hepcidin assays [
28]. The prevalence of ID identified by hepcidin that we observed in this cohort is consistent with the expected; as the prevalence of ID on admission is around 30–40%, the prevalence of ID on discharge is expected to be the same or higher due to frequent blood loss (even if ID on admission may be also associated with higher mortality). Because hepcidin is higher in patients who have received blood transfusions [
29] and because around 40% of the patients have been transfused, it is possible that more patients had ID. In addition, hepcidin synthesis is linked to the iron stores and to the level of inflammation [
11,
29,
30]. Hepcidin could also be useful to indicate when iron treatment could be necessary, in fact elevated hepcidin prevents the mobilization of iron from stores, by blocking the ferroportin, whereas low hepcidin allows the expression of ferroportin at the cell membrane of macrophages and thus the export of iron from stores [
11,
30].
In this cohort, we observed that severe ID is associated with both increased one-year mortality and poorer physical QOL. This is also consistent with observations that ID is a risk factor for mortality and morbidity in patients with cardiac disease [
4]. Indeed, ID has been shown to be responsible for decreased mitochondrial complex I activity, decreased exercise capacity and left ventricular ejection fraction in an animal model [
31]. ID has been also associated with decreased muscular function (skeletal and myocardial) and this may explain the usual association reported between ID and fatigue [
3,
12,
32].
Our study has some limitations. We used different definitions of ID, but we could also have used other markers such as erythrocyte zinc protoporphyrin [
7,
25]. However, this marker may be influenced by blood transfusion and may not be suitable in these patients (at least at the end of their stay, with a transfusion rate around 40%). We observed an association between low hepcidin and liver disease, but we do not have enough data to further investigate this association. Although the association between ID and poor outcomes is supported by a physiological rationale, our data do not demonstrate that iron treatment is beneficial in these critically ill patients. Furthermore, we do not have blood samples after ICU discharge to determine whether some patients did recover from ID or not and if ID at 3, 6 or 12 months is still associated with poor outcomes. We also do not know if some of these patients have been treated with iron after ICU discharge, but this is not likely since iron is rarely proposed in these patients. We also could have investigated the link between high hepcidin and the outcomes. Indeed, we observed that high hepcidin was also associated with higher mortality (as shown in Additonal file
3: Figure S2 for the highest tertile), and some authors report that high hepcidin (on admission) is associated with blood transfusion for example [
29]. But, although ID is treatable using iron, high hepcidin is not (because anti-hepcidin antibodies are not available yet). It is thus more indicative of more severe inflammation [
29].
Interestingly, treatment of ID using intravenous iron has been proven to improve symptoms of fatigue, physical and mental QOL and even cardiac function [
33,
34]. To date, iron treatment is not recommended for critically ill patients. However, we have already demonstrated in an animal model of critical care anemia with ID, that iron may be used to treat anemia, without toxicity (neither oxidative stress induction nor increased risk of infection) [
35]. Some human data are promising. Intravenous iron does not induce more oxidative stress in critically ill patients compared to healthy volunteers [
36], and it has been shown to improve haemoglobin level at hospital discharge [
37]. The results of a prospective randomized controlled study (i.e. Hepcidan study, NCT02276690) [
38] designed to assess whether diagnosing ID using this mass spectrometry hepcidin measurement on ICU discharge could improve the post-ICU rehabilitation of critically ill patients are expected soon and could help clinicians decide how to diagnose ID in the critically ill and whether iron treatment is useful.
Acknowledgements
This study is based on the FROG-ICU cohort, for which all the FROG-ICU investigators recruited the patients. We thank them very much for this work.
FROG-ICU study group:
Hopital Lariboisiere, Paris, France: N Deye, C Fauvaux, A Mebazaa, C Damoisel, D Payen.
Hopital Saint Louis, Paris, France: E Azoulay, AS Moreau, L Jacob, O Marie.
Hopital Bichat, Paris, France: M Wolf, R Sonneville, R Bronchard.
Hopital Beaujon, Paris, France: I Rennuit, C Paugam.
Hopital, Cochin, Paris, France: JP Mira, A Cariou, A Tesnieres.
Hopital Bicetre, Paris, France: N Dufour, N Anguel, L Guerin, J Duranteau, C Ract.
CHU de Marseille, Marseille, France: M Leone, B Pastene.
Hopital Raymond Poincarré, Garches, France: T Sharshar, A Fayssoyl.
Hopital Saint-Antoine, Paris, France: J-L Baudel, B Guidet.
Hopital Pitié salpétrière, Paris, France: Q Lu, W Jie Gu, N Brechot, A Combes.
CHU Saint Eloi, Montpellier, France: S Jaber, A Pradel, Y Coisel, M Conseil.
Hopital Ambroise Paré, Boulogne Billancourt, France: A Veillard Baron, L Bodson.
CHU Caremeau, Nimes, France: Jy Lefrant, L Elotmani, A Ayral, S Lloret.
Hopital Hopital Jean Minjoz, Besançon, France: S Pily-Flouri, Jb Pretalli.
Clinique Saint Luc, Louvain, Belgique: Pf Laterre, V Montiel, Mf Dujardin, C Berghe.