Measurement of iron deficiency in infants, implications for screening in a high risk population
Hawkins, Clare A.
MetadataShow full item record
Objective. To evaluate the measurement properties of hematological and biochemical indicators of iron deficiency (ID). Design. Secondary data analysis of a longitudinal data set collected for a randomized controlled trial of iron supplemented formula conducted in 1991. Analysis of a data-set of four venous samples drawn from 225 infants at ages 6, 9, 12 and 15 months. Setting. Aboriginal children living in the urban core of Winnipeg. Patient selection. Non-breastfeeding mothers who agreed with informed consent to participate in an infant formula iron-fortification trial $<$2 months of age. Measurement and results. Serum Ferritin (SF), Transferrin Saturation (TS) and Free Erythrocyte Protoporphyrin (FEP) were compared with Hemoglobin (HB), Mean Corpuscular Volume (NCV) and Red Cell Distribution Width (RDW). Non parametric receiver operator characteristic curve (ROC) analysis was conducted as a cut-off independent method to compare the iron tests. The areas under the curve were measured to contrast the diagnostic strength of each test. The analysis was repeated with a series of surrogate "gold standards". At the same assessment age, SF and hematological tests were used. The longitudinal nature of the dataset was used to determine if a low value of a test at one sample interval predicts low values of other tests at the next sample interval. Predicting which infants would have a HB rise in response to iron administration was also evaluated. Conclusions. HB performed well at predicting the response to iron supplementation. TS was the most accurate early indicators of ID in infants. Age appropriate cut-off values for FEP and MCV are recommended. Recommendations are made for screening and primary prevention in high risk, remote populations.