Diagnostic utility of zinc protoporphyrin to detect iron deficiency in Kenyan preschool children: a community-based survey

Emily M. Teshome, Andrew M. Prentice, Ayşe Y. Demir, Pauline E.A. Andang’o, Hans Verhoef

Zinc protoporphyrin (ZPP) has been used to screen and manage iron deficiency in individual children, but it has also been recommended to assess population iron status. The diagnostic utility of ZPP used in combination with haemoglobin concentration has not been evaluated in pre-school children. We aimed to a) identify factors associated with ZPP in children aged 12-36 months; b) assess the diagnostic performance and utility of ZPP, either alone or in combination with haemoglobin, to detect iron deficiency. We used baseline data from 338 Kenyan children enrolled in a community-based randomised trial. To identify factors related to ZZP measured in whole blood or erythrocytes, we used bivariate and multiple linear regression analysis. To assess diagnostic performance, we excluded children with elevated plasma concentrations of C-reactive protein or α1-acid glycoprotein, and with Plasmodium infection, and we analysed receiver operating characteristics (ROC) curves, with iron deficiency defined as plasma ferritin concentration < 12 μg/L. We also developed models to assess the diagnostic utility of ZPP and haemoglobin concentration when used to screen for iron deficiency. Whole blood ZPP and erythrocyte ZPP were independently associated with haemoglobin concentration, Plasmodium infection and plasma concentrations of soluble transferrin receptor, ferritin, and C-reactive protein. In children without inflammation or Plasmodium infection, the prevalence of true iron deficiency was 32.1%, compared to prevalence of 97.5% and 95.1% when assessed by whole blood ZPP and erythrocyte ZPP with conventional cut-off points (70 μmol/mol and 40 μmol/mol haem, respectively). Addition of whole blood ZPP or erythrocyte ZPP to haemoglobin concentration increased the area-under-the-ROC-curve (84.0%, p = 0.003, and 84.2%, p = 0.001, respectively, versus 62.7%). A diagnostic rule (0.038689 [haemoglobin concentration, g/L] + 0.00694 [whole blood ZPP, μmol/mol haem] >5.93120) correctly ruled out iron deficiency in 37.4%-53.7% of children screened, depending on the true prevalence, with both specificity and negative predictive value ≥90%. In young children, whole blood ZPP and erythrocyte ZPP have added diagnostic value in detecting iron deficiency compared to haemoglobin concentration alone. A single diagnostic score based on haemoglobin concentration and whole blood ZPP can rule out iron deficiency in a substantial proportion of children screened. ClinicalTrials.gov NCT02073149 (25 February 2014).