Low ferritin correlated well with low bone marrow iron stores, but no feasible ferritin cutoff ruled out iron deficiency anaemia with high certainty.

Bone marrow aspiration with iron staining is the gold standard for diagnosing iron deficiency anaemia. However, because it is invasive, clinicians instead usually rely on iron markers, such as serum iron, iron binding capacity, transferrin saturation, ferritin, mean corpuscular volume (MCV) and, occasionally, soluble transferrin receptor. To identify the best-performing parameter, investigators retrospectively compared results of these tests and bone marrow aspiration in 6610 patients with haematological disorders.

Ferritin correlated best with bone marrow iron stores and was superior to transferrin saturation, levels of soluble transferrin receptor, transferrin and haemoglobin, and MCV. A ferritin cutoff of 30 mcg/L was highly specific for iron deficiency anaemia in men (99%) and women (97%) but poorly sensitive (35% and 54%, respectively). No combination of tests performed better than ferritin alone. The lowest ferritin cutoffs allowing for 95% negative predictive value were 1750 mcg/L for women and 4967 mcg/L for men.

Comment: When ferritin is low, a diagnosis of iron deficiency is straight­forward, although reference ranges vary, and ‘low’ is defined differently across laboratories and institutions. This study confirms high specificity and positive predictive value when ferritin is below 30 mcg/L. Unfortunately, however, there is no feasible ferritin cutoff for ruling out iron deficiency with high negative predictive value.

A common clinical conundrum arises when we try to identify which patients with inflammatory disease – with low transferrin saturation and normal to increased ferritin – have concurrent iron deficiency anaemia. Although many haematologists use a ferritin level above 100 mcg/L as an indicator of sufficient iron stores, the accuracy of this cutoff is unclear. In these cases, we must use clinical judgement based on risk factors for iron deficiency (e.g. bleeding, malabsorption) and less-utilised tests (soluble transferrin receptor), which unfortunately can underperform.

Brady L. Stein, MD, MHS, Associate Professor of Medicine, Division of Hematology/Oncology, Northwestern University Feinberg School of Medicine, Chicago, USA.

Lahtiharju T, et al. Ferritin outperforms other biomarkers in predicting bone marrow iron stores in patients with hematologic disorders. Blood Adv 2025; 9: 1608-1617.

This summary is taken from the following Journal Watch titles: Oncology and Hematology, General Medicine, Ambulatory Medicine.