Adjusting plasma ferritin concentrations in the presence of subclinical inflammation to improve the estimate of response to an iron intervention

Plasma ferritin concentrations are a well‐accepted method of assessing iron status for population surveys and intervention trials. However, ferritin is an acute phase protein (APP) and becomes elevated in the presence of inflammation, creating marked risk for underestimating iron deficiency in populations with high rates of infection. An additional dimension to the problem is in monitoring benefits of an iron intervention when inflammation is present at different time points. Standard practice for handling inflammation in intervention studies has been the removal of data for subjects showing signs of subclinical inflammation, commonly using one of two APP; C‐reactive protein (CRP) and/or á1‐acid glycoprotein (AGP). However, due to significant risk of introducing bias from data removal, there is a need to develop alternative methodologies for handling inflammation inflated ferritin values and to determine which approaches can be used most effectively for intervention trials. Data from an iron biofortified bean intervention in Mexican schoolchildren shows evidence of differential inflammation at baseline and endline. Three different methodologies for adjusting ferritin values will be applied to this data set: data removal, the correction factor published by Thurnham et al. (AJCN, 2010; 92:546–55), and a dataset specific statistical correction factor using CRP and AGP. Support provided by Harvest Plus. Grant Funding Source : Harvest Plus and NIH