Iron status and fibroblast growth factor-23 in Gambian children

Abstract

A relationship between iron and fibroblast growth factor-23 (FGF23) metabolic pathways has been proposed. Iron deficiency anaemia is prevalent in The Gambia and concentrations of fibroblast growth factor-23 FGF23 are elevated in a large percentage of Gambian children with rickets-like bone deformity. We speculate that low iron status may be involved in the aetiology of Gambian rickets. The aim of this study was to determine if there was a relationship between haemoglobin, as a marker of iron status, and FGF23 in samples from children with and without a history of rickets-like bone deformities in The Gambia. We conducted a retrospective analysis of studies carried out from 2006 to 2008 in children from a rural community in The Gambia where iron deficiency anaemia is endemic and where elevated circulating concentrations of FGF23 have been found. To investigate the relationship between circulating FGF23 and haemoglobin concentrations we used an age-adjusted linear regression model on data from children <18y of age with a family or personal history of rickets-like bone deformity (BD) (n=108) and from the local community (LC) (n=382). We found that circulating concentration of FGF23 was inversely correlated with haemoglobin concentration. This effect was more pronounced in BD children compared with LC children (interaction: P≤0.0001). Anaemia and elevated FGF23 were more prevalent in BD children compared to LC children (P=0.0003 and P=0.0001 respectively). In conclusion, there is a stronger relationship between FGF23 and haemoglobin in Gambian children with a history of rickets compared to local community children. This study provides support for the contention that iron may be involved in FGF23 metabolic pathways.

Fig. 1

Relationship between age-adjusted haemoglobin (Hb) (g/dl) and age-adjusted FGF23 (RU/ml) (on a logarithmic scale) grouped into BD children (bone deformity) ▴ and LC children (local community) ▿. Age-adjusted Hb was derived by including log_e Hb and age in the regression model separately for each group (BD or LC); evaluating the residual for each subject; adding the residual to log_e (mean group Hb) value; and calculating the antilogarithm. Age-adjusted FGF23 was derived using the same method. The equation of the line derived from multiple regression using unadjusted log_e FGF23 (y) with unadjusted log_eHb and age (x) for: BD children only was log_eFGF23 = [16.05 (SE 3.16)] − [4.28 (SE 1.27) (log_eHb)] − [0.08 (SE 0.03) (age)], R2 = 23.2%, P = 0.001 and LC children only was log_eFGF23 = [7.69 (SE 0.95)] − [1.08 (SE 0.38) (log_eHb)] − [0.08 (SE 0.01) (age)], R² = 10.4%, P = 0.005. All) both BD and LC children together was log_eFGF23 = [9.59 (SE 0.99)] − [1.77 (SE (0.40) (log_eHb)] − [0.09 (SE 0.01) (age)], R² = 15.5%, P ≤ 0.0001 (line not shown). The log_eHb × group interaction was highly significant (P < 0.0007) demonstrating differences between BD and LC in the slopes of the relationship.