Defining ferritin clinical decision limits to improve diagnosis and treatment of iron deficiency: A modified Delphi study

Abstract

Background Iron deficiency is highly prevalent worldwide and is an issue of health inequity. Despite its high prevalence, uncertainty on the clinical applicability and evidence‐base of iron‐related lab test cut‐offs remains. In particular, current ferritin decision limits for the diagnosis of iron deficiency may not be clinically appropriate nor scientifically grounded. Methods A modified Delphi study was conducted with various clinical experts who manage iron deficiency across Canada. Statements about ferritin decision limits were generated by a steering committee, then distributed to the expert panel to vote on agreement with the aim of achieving consensus and acquiring feedback on the presented statements. Consensus was reached after two rounds, which was defined as 70% of experts rating their agreement for a statement as 5 or higher on a Likert scale from 1 to 7. Results Twenty‐six clinical experts across 10 different specialties took part in the study. Consensus was achieved on 28 ferritin decision limit statements in various populations (including patients with multiple comorbid conditions, pediatric patients, and pregnant patients). For example, there was consensus that a ferritin <30 μg/L rules in iron deficiency in all adult patients (age ≥ 18 years) and warrants iron replacement therapy. Conclusion Consensus statements generated through this study corresponded with current evidence‐based literature and guidelines. These statements provide clarity to facilitate clinical decisions around the appropriate detection and management of iron deficiency.

Ferritin Cutoffs and Diagnosis of Iron Deficiency in Primary Care

“…We conducted all analyses from October 2, 2023, to May 29, 2024, using the statistical software R, version 4.3.2 (R Foundation for Statistical Computing). 22 We assessed the association of ferritin cutoff choice with the incidence of iron deficiency diagnoses using the following 3 cutoffs recommended by guidelines and expert panels for the adult general population 4,11,[23][24][25][26][27] : 15, 30, and 45 ng/mL. For each cutoff and patient, we defined an iron deficiency event as the first measurement of ferritin concentrations below that cutoff during the study period and calculated patient-time as the number of days from (and including) inclusion to (and including) the earliest of either an iron deficiency event, if any, or the end of the study period.…”

Accuracy of reticulocyte hemoglobin for diagnosing iron deficiency in former very preterm infants: a population-based cohort study

“…For the main analysis (definition 1) in this study, ID was defined as SF level <20 mcg/L at both 4 and 6 months CA, as completed as standard practice by the present study hospital’s laboratory. As a secondary analysis (definition 2), ID was defined using SF level <30 mcg/L at both 4 and 6 months CA, as recommended by a recent modified Delphi study for preterm infants ( 16 ). The study was approved by the Research Ethics Board of the IWK Health (Approval 1026862).…”

Recommendations for diagnosis, treatment, and prevention of iron deficiency and iron deficiency anemia

“…Many authors suggested that the diagnostic accuracy of ferritin could be improved by increasing the cutoff to 30 μg/L. 6 , 7 , 8 Such a value has a higher sensitivity (from 85% to 92%) and unchanged specificity (98%), and it is the most accepted threshold used to establish a diagnosis of ID, even in the mildest cases. 9 However, little evidence is available from high‐quality studies to justify specific thresholds.…”

“…The WHO defines ID as serum or plasma ferritin levels <15 ug/L in children older than 5 years, adolescents, and adults and less than 12 ug/L in children younger than 5 years (WHO, 2022). Many authors suggested that the diagnostic accuracy of ferritin could be improved by increasing the cutoff to 30 μg/L 6–8 . Such a value has a higher sensitivity (from 85% to 92%) and unchanged specificity (98%), and it is the most accepted threshold used to establish a diagnosis of ID, even in the mildest cases 9 .…”