Screening for iron deficiency anemia by dietary history in a high-risk population

Background. Iron deficiency anemia (IDA) in young children is important to identify because of its adverse effects on behavior and development. Because of costs and inconvenience associated with blood test screening and the decline in prevalence of IDA, the Institute of Medicine and the Centers for Disease Control and Prevention recommend that blood test screening for IDA be targeted to children first identified by dietary and health history. Objective. To evaluate a parent-completed dietary and health history as the first stage of 2-stage screening for IDA. Design and Methods. A cross-sectional study was conducted in inner-city clinics in children 9 to 30 months old having routine anemia screening as part of a scheduled visit. Parents completed a questionnaire and children had venous blood sampling for complete blood count and ferritin. Anemia was defined as Hb <11.0 g/dL. Iron deficiency (ID) was defined as ferritin <10 mu g/L or mean corpuscular volume <70 fL and red cell distribution width >14.5%. Children were categorized into 1 of 4 groups: iron-sufficient, not anemic (ISNA); iron-sufficient, anemic (ISA); iron-deficient, not anemic (IDNA); and iron-deficient anemic (IDA). The questionnaire consisted of 15 dietary items in domains of infant diet, intake of solid food, intake of beverages, and participation in the Special Supplemental Nutrition Program for Women, Infants, and Children together with 14 historical items in domains of birth history, recent illness, chronic medical conditions, history of anemia, and maternal history. Analysis was performed on individual items, domains, and combinations of selected items. Results. In the 282 study subjects, the prevalence of anemia (35%), IDNA (7%), and IDA (8%) did not vary significantly by age. Among individual historical and dietary questions, maternal history of anemia and drinking >2 glasses of juice per day identified the highest proportion of children with IDA: 50% sensitivity (95% confidence interval [CI]: 16,81) and 77% sensitivity (95% CI: 54,89), respectively. However, specificities for these questions were 60% (95% CI: 55,65) and 22% (95% CI: 17,27), respectively. Domains of questions with the highest sensitivity for IDA were beverage intake (91%; 95% CI: 68,99) and intake of solid food (91%; 95% CI: 68,99). However, specificities of the domains were only 14% (95% CI: 10,18) and 29% (95% CI: 24,35), respectively. The dietary items used by Boutry and Needlman were 95% (95% CI: 77,99) sensitive but only 15% (95% CI: 11,19) specific for IDA. The recommendations of the Centers for Disease Control and Prevention for health and dietary screening were 73% (95% CI: 56,92) sensitive and 29% (95% CI: 24,35) specific for IDA. The individual questions, domains of questions, and interdomain groups of questions had similar sensitivity and specificity for anemia and ID (IDA + IDNA). Conclusion. In this high-risk population, neither individual nor combinations of parental answers to dietary and health questions were able to predict IDA, anemia, or ID well enough to serve as a first-stage screening test.