Ferritin Reference Curves and Optimal Curves in Preadolescent Children

Ferritin levels are commonly used to assess iron status. For interpretation, clinicians traditionally rely on age-partitioned reference intervals (RIs), which report 2.5th (lower) and 97.5th (upper) percentiles.

To estimate sex-specific continuous ferritin reference curves (RCs) and optimal curves (OCs) for preadolescents; to create age- and sex-specific RIs and optimal intervals (OIs); to interpret estimates alongside ferritin thresholds for iron deficiency; and to develop an interactive web-based computational tool.

This cross-sectional analysis used data from a longitudinal cohort study. Participants included 4935 children aged 2 weeks to 10 years attending scheduled primary care health supervision visits from June 3, 2008, to February 26, 2020, in Toronto, Ontario, Canada. Data were analyzed from October 16, 2024, to December 17, 2025.

Blood samples were collected and analyzed for levels of ferritin, hemoglobin, and C-reactive protein. Parents completed a questionnaire collecting optimality criteria for iron status.

Sex-specific RCs were estimated using generalized additive models for location, scale, and shape. OCs were derived from a subsample of participants meeting predefined optimality criteria, using multiple imputation to address missing data. RCs and OCs were interpreted in the context of the World Health Organization threshold (<12 ng/mL) and the proposed American Society of Hematology (ASH) threshold (≤20 ng/mL).

RCs were estimated for 4935 children (2613 [52.9%] male) aged 2 weeks to 10 years (median age, 37 [IQR, 18-62] months). OCs were estimated from a subset of 3630 children (1909 [52.6%] male). Curves were highest in early infancy, declined sharply at 1.5 years of age (lower limit, 5-6 ng/mL), then gradually increased at 9 to 10 years of age (lower limit, 15-17 ng/mL). RCs and OCs showed similar trajectories, with slightly higher limits for OCs. The proportion of preadolescent children in the optimal population who would be classified as having iron deficiency in existing (World Health Organization) and proposed thresholds differed with age. Using the ASH threshold in this cohort, 54.7% of males and 37.0% of females would be considered to have iron deficiency at 18 months.

In this cross-sectional study, ferritin RCs and OCs were estimated in a low-risk population of preadolescent children. The lower limits of curves and intervals were substantially lower than physiologically based thresholds to define iron deficiency. These findings further inform the current discussion regarding the distributional vs physiologic approaches to interpreting ferritin measurements and guiding clinical decision-making.