Physiologically based serum ferritin thresholds for iron deficiency in children and non-pregnant women: a US National Health and Nutrition Examination Surveys (NHANES) serial cross-sectional study

Serum ferritin concentrations are the most widely used indicator for iron deficiency. WHO determined that insufficient data are available to revise the serum ferritin thresholds of less than 12 μg/L for children and less than 15 μg/L for women, which were developed on the basis of expert opinion, to define iron deficiency. We aimed to derive new physiologically based serum ferritin concentration thresholds for iron deficiency in healthy young children and non-pregnant women using data from the US National Health and Nutrition Examination Survey (NHANES).In this serial cross-sectional study, we examined the relationship of serum ferritin with two independent indicators of iron-deficient erythropoiesis, haemoglobin and soluble transferrin receptor (sTfR), in children (12-59 months) and non-pregnant women (15-49 years) using cross-sectional NHANES data from 2003-06, 2007-10, and 2015-18. NHANES is a US national stratified multistage probability sample that includes a household interview followed by a standardised physical examination in a mobile examination centre. We excluded individuals with missing serum ferritin, sTfR, haemoglobin, or white blood cell counts measurements; non-pregnant women with missing C-reactive protein (CRP), alanine aminotransferase (ALT) or aspartate aminotransferase (AST) data were also excluded. In addition, individuals with infection (white blood cell counts >10·0×109/L) and non-pregnant women with possible liver disease (ALT >70 IU/L or AST >70 IU/L) and inflammation (CRP >5·0 mg/L) were excluded. We examined distributions of haemoglobin and sTfR with serum ferritin and used restricted cubic spline regression models to determine serum ferritin thresholds for iron-deficient erythropoiesis.5964 children and 10 462 non-pregnant women had physical examinations and were screened for inclusion in the study, of whom 2569 (43·1%) children and 7498 (71·7%) non-pregnant women were included. At lower serum ferritin concentrations, median haemoglobin concentration decreased as sTfR concentration increased, with each varying in a curvilinear manner. Using restricted cubic spline plateau points to determine the onset of iron-deficient erythropoiesis, the serum ferritin thresholds identified by haemoglobin and sTfR concentrations were not different. For children, the haemoglobin identified serum ferritin threshold was 19·9 μg/L (95% CI 18·8-22·6) and the sTfR identified serum ferritin threshold was 20·0 μg/L (19·4-20·9; p=0·89). For women the haemoglobin identified serum ferritin threshold was 25·2 μg/L (24·2-26·2) and the sTfR identified serum ferritin threshold was 24·0 μg/L (23·3-24·6; p=0·05).The association between two independent indicators of iron-deficient erythropoiesis, haemoglobin and sTfR, identified serum ferritin concentration thresholds of about 20 μg/L for children and 25 μg/L for non-pregnant women, providing physiological evidence of potential new thresholds for consideration when determining the prevalence and distribution of iron deficiency in populations. In healthy children and non-pregnant women, physiologically based thresholds for iron deficiency might be more clinically and epidemiologically relevant than those based on expert opinion. Validation of this physiologically based approach in non-US populations might help the international harmonisation of serum ferritin thresholds for iron deficiency.None.