343 - Serum Ferritin Values in Neonates <29 Weeks Gestation Are Highly Variable and Do Not Correlate with Reticulocyte Hemoglobin Content

Poster Number: 343
Publication Number: 343.122

Timothy M. Bahr, Intermountain Healthcare, Salt Lake City, UT, United States; Sylvia Tan, RTI International, WASHINGTON, DC, United States; Mariana Baserga, University of Utah, Park City, UT, United States; Robert D. Christensen, University of Utah School of Medicine, Salt Lake City, UT, United States; Robin K. Ohls, University of Utah, Salt Lake City, UT, United States

Background: Serum ferritin and reticulocyte hemoglobin content (RET-He) are both validated methods for iron deficiency screening among children and adults, but lack rigorous confirmation among extremely low gestational age neonates (ELGANs). We measured simultaneous serum ferritin and RET-He values among ELGANs and compared each with measurements of iron-deficient erythropoiesis.

Objective: To quantify correlation between serum ferritin and RET-He in an ELGAN cohort. Additionally, to determine which correlates with erythrocyte mean corpuscular volume (MCV), a measurement of erythrocyte microcytosis that occurs during iron-limited erythropoiesis.

Design/Methods:
This was a secondary analysis of the NICHD Neonatal Research Network Newborn Darbepoetin trial, a randomized, masked trial of weekly darbepoetin vs. placebo dosing of neonates 23 0/7-28 6/7 weeks gestation to increase red cell mass and improve developmental outcome. Study data were included in this analysis only for placebo recipients who had a serum ferritin, a RET-He, and an MCV measurement measured at baseline (0-2 days; RET-He and MCV only), early (11-17 days of age) or late (38-45 days of age) time periods. All subjects received parenteral and enteral iron per study protocol, and were transfused according to study guidelines. Spearman correlation analyses were performed, and were adjusted for gestational age, SGA, and the clustering within study site.

Results: A total of 102 subjects (26.7±1.5 weeks, 875±218 grams at birth) were evaluated. RET-He and ferritin did not positively correlate at early or late timepoints (Figure 1A and 1B). Neither relationship was significant after adjusting for GA, SGA, and the clustering within site. A significant positive correlation was observed between RET-He and MCV at baseline, early, and late time points (r=0.25, 0.32, and 0.39, respectively; late time period data shown in Figure 2A); these relationships were confirmed during adjusted analyses. Ferritin inversely correlated with MCV (r=-0.25 and r=-0.55 during early and late time periods, respectively); results were not significant after adjusted analyses (late time period data shown in Figure 2B).Conclusion(s): Ferritin and RET-He did not correlate in this ELGANs population. The reliability of ferritin as a marker of iron sufficiency among ELGANs was poor. RET-He is correlated with MCV, a marker of microcytic iron-limited erythropoiesis. We speculate that RET-He is a more accurate measure of iron insufficiency in preterm infants.
Relationship between simultaneous obtained RET-He and ferritin measurementsScatter plots showing the relationship between simultaneous obtained RET-He and serum ferritin concentration measurements in neonates born at less than 29 weeks gestation who did not receive darbepoetin alpha. A regression line is drawn on each plot. (A) Paired values obtained at or before 27 days of life; (B) paired values obtained on or after 28 days of life.