Adjusting plasma ferritin concentrations to remove the effects of subclinical inflammation in the assessment of iron deficiency : a meta-analysis

Background: The World Health Organization recommends serum ferritin concentrations as the best indicator of iron deficiency (ID). Unfortunately, ferritin increases with infections; hence, the prevalence of ID is underestimated. Objective: The objective was to estimate the increase in ferritin in 32 studies of apparently healthy persons by using 2 acute-phase proteins (APPs). C-reactive protein (CRP) and alpha(1)-acid glycoprotein (AGP), individually and in combination, and to calculate factors to remove the influence of inflammation from ferritin concentrations. Design: We estimated the increase in ferritin associated with inflammation (ie, CRP >5 mg/L and/or AGP >1 g/L). The 32 studies comprised infants (5 studies), children (7 studies), men (4 studies), and women (16 studies) (n = 8796 subjects). In 2-group analyses (either CRP or AGP), we compared the ratios of log ferritin with or without inflammation in 30 studies. In addition, in 22 studies, the data allowed a comparison of ratios of log ferritin between 4 subgroups: reference (no elevated APP), incubation (elevated CRP only), early convalescence (both APP and CRP elevated), and late convalescence (elevated AGP only). Results: In the 2-group analysis, inflammation increased ferritin by 49.6% (CRP) or 38.2% (AGP; both P <0.001). Elevated AGP was more common than CRP in young persons than in adults. In the 4-group analysis, ferritin was 30%, 90%, and 36% (all P < 0.001) higher in the incubation, early convalescence, and late convalescence subgroups, respectively, with corresponding correction factors of 0.77, 0.53, and 0.75. Overall, inflammation increased ferritin by approximate to 30% and was associated with a 14% (CI: 7%, 21%) underestimation of ID. Conclusions: Measures of both APP and CRP are needed to estimate the full effect of inflammation and can be used to correct ferritin concentrations. Few differences were observed between age and sex subgroups. Am J Clin Nutr 2010;92:546-55