Using serum hepcidin to identify iron deficiency in previously critically ill patients at time of hospital discharge

Background: Anaemia during critical illness is multifactorial and well studied. There is less data on anaemia during the recovery phase of critical illness and how this relates to long-term outcomes.1 Iron deficiency is the largest and treatable cause of anaemia worldwide, but in critical care, diagnosis of iron depletion is complicated by concomitant inflammation, which raises serum ferritin, the commonest estimate for iron status. Hepcidin has emerged as a key regulator of iron metabolism. Its expression is increased in states of inflammation and iron overload and repressed in iron deficiency, hypoxia and erythroid expansion.2,3 Our aims were:(i)To identify the proportion of patients discharged from critical care who remain anaemic at, or close to, hospital discharge (ii)To identify the proportion of anaemic patients with iron deficiency anaemia (IDA), anaemia of inflammation (AID) and combined anaemia of inflammation and iron deficiency (IDI)(iii)To investigate if serum hepcidin values are significantly lower in the IDI group compared to the AID group. Methods: Anaemia was defined according to WHO guidelines: Males <130 g.L-1 and Females<120 g.L-1. In the presence of inflammation (CRP>8 mg. L-1), a ferritin cut-off of 150 ug.L-1 was used to differentiate between IDI (<150) and AID(>150).3 Serum samples collected as part of RECOVER – a multicenter randomised controlled trial assessing the effects of an intense rehabilitation programme post-ICU discharge on patient outcomes, were analysed.4 We assayed serum hepcidin, soluble transferrin receptor, CRP, ferritin, and other markers of iron status. Statistical significance was defined as p<0.05. Analysis was performed using STATA14.0 (Statacorp, College Station, TX, USA). Results: 117 patient samples were analysed. Proportions of different types of anaemia are shown in Figure 1. Mean (SD) sample collection day was 4.6 (6.08) days prior to hospital discharge and 18.5 (18.5) days post-ICU discharge. 110/117 (94%) patients were anaemic prior to hospital discharge. Overall 89/117 (76%) had an AID and 11/117 (9%) had IDI based on our pre-defined criteria. Geometric mean (95% CI) hepcidin levels were significantly lower in the IDI group compared to the AID group: 9.13 (3.57 – 23.3) ng.mL-1 vs. 28.6 (24.8 – 33.1) ng.mL-1 (p<0.05). Using area under the curve for receiver operator characteristic (AUCROC) curves for our diagnosis of IDI, hepcidin <8 ng.mL-1 had a sensitivity of 36.3% and a specificity of 95.8%, while hepcidin of <19 ng.mL-1 had a sensitivity of 72.7% and a specificity of 71.7% with an AUCROC of 0.722.Conclusion: The majority of patients discharged from critical care, and subsequently hospital, have an AID and a small proportion develop IDI based on current available tests. We have shown that serum hepcidin is a promising diagnostic test of iron deficiency in the context of inflammation. However, the hepcidin levels are elevated to a point where, based on previous studies, patients would not be expected to respond to oral iron.2,5 Prospective studies are therefore needed to further understand hepcidin regulation during the acute and recovery phases of critical illness, to assess if intravenous iron may be beneficial in treating anaemia in patients recovering from critical illness, and determine if serum hepcidin levels could identify those most likely to respond to this therapy.