BACKGROUND & AIMS: Iron deficiency (ID) is common in heart failure (HF), may vary in prevalence along the HF trajectory, and is associated with adverse outcomes.  Current guidelines define ID as ferritin <100 µg/L or ferritin 100–299 µg/L when transferrin saturation (TSAT) <20 %. However, this definition was never clinically validated in HF populations, and recent evidence suggests that compared with bone marrow iron staining both a TSAT <20 % or a serum iron ≤13 μmol/L may identify ID and predict prognosis more reliably. Since comparative assessment of the in-hospital prevalence and trajectory and of the prognostic significance of each of these definitions is lacking, we aimed in this study to (1) compare in patients hospitalised with acute HF prevalence rates and in-hospital changes of ID according to the four ID definitions and (2) study their prognostic significance.  

METHODS: Among patients hospitalised with acute HF, who were consecutively recruited into a single-centre prospective cohort study, those with a complete iron status (ferritin, TSAT, serum iron) available within 5 days of admission and within 48 hours of discharge were eligible, if not receiving any intravenous iron supplementation while in hospital. At both time points, ID prevalence rates were determined.  Cox proportional hazard regression analyses and Kaplan-Meier plots were used to assess the prognostic impact of each ID definition regarding 12-month all-cause mortality risk.

Results: Of 632 consecutive patients, 273 were eligible (74±11 years, 41 % female). Overall, median admission ferritin was 152 (83; 276) µg/L, median TSAT 14.6 (10.6; 21.1) % and median serum iron 9.1 (6.3; 12.0) µmol/L. ID by a ferritin <100 µg/L was present in 89 patients (33 %), by ESC guideline criteria in 179 patients (66 %), by a TSAT <20 % in 197 patients (72 %) and by a serum iron ≤13 µmol/L in 218 patients (80 %). Total length of stay was 11 (7; 16) days. ID assessment took place at 3 (2; 4) and 10 (8; 15) days following admission. At the second assessment, ID prevalence had decreased by all definitions: By a ferritin <100 µg/L 73 (27 %) patients had ID (p=0.015 vs. on admission), by ESC guideline criteria 154 (56 %) (p<0.001), by a TSAT <20 % 173 (63 %) (p=0.001) and by a serum iron ≤13 µmol/L 186 (68 %) (p<0.001) (Figure 1). ID definitions by a TSAT <20 % and a serum iron ≤13 μmol/L outperformed ID definitions based on ferritin levels in identifying patients at a higher risk of death compared with their counterparts without ID (Figure 2).  Similar associations with prognosis as shown in Figure 2 were found if iron markers determined on admission were used for ID assessment. 

Conclusions: Prevalence rates of ID vary considerably with different definitions. Since irrespective of definition significant short-term changes of iron variables occur that implicate changes in ID prevalence, the time point of iron assessment matters also in patients hospitalised for acute HF. ID defined by a TSAT <20 % or a serum iron ≤13 µmol/L performed best in identifying patients at increased risk of death. Our findings question the validity of ESC guideline criteria for ID which rely on ferritin levels.