Frontiers in Pediatrics (Nov 2023)

Accuracy of reticulocyte hemoglobin for diagnosing iron deficiency in former very preterm infants: a population-based cohort study

  • Hudson Barr,
  • Ketan Kulkarni,
  • Ketan Kulkarni,
  • Balpreet Singh,
  • Balpreet Singh,
  • Navjot Sandila,
  • Lisa Morrison,
  • Lori Beach,
  • Satvinder Ghotra,
  • Satvinder Ghotra

DOI
https://doi.org/10.3389/fped.2023.1281513
Journal volume & issue
Vol. 11

Abstract

Read online

BackgroundSerum ferritin (SF) is commonly used to diagnose iron deficiency (ID) but has limitations. Reticulocyte hemoglobin (Ret-He) is being increasingly used for ID diagnosis. This study aimed to assess accuracy of Ret-He for ID diagnosis in former very preterm infants (VPI) at 4–6 months corrected age (CA).MethodsA retrospective population-based cohort study was conducted on all live VPI born between 23 and 30 weeks of gestational age (GA) in Nova Scotia from 2012 to 2018. Infants underwent SF and Ret-He testing at 4–6 months CA. ID was defined using two definitions. The first defined ID as SF < 20 mcg/L at both 4- and 6-months CA, and the second as SF < 30 mcg at at both 4- and 6-months CA. The accuracy of Ret-He for identifying ID was assessed using the area under the receiver operating characteristic curve (AUC).ResultsID was present in 39.7% (62) of 156 infants in the first definition and 59.6% (93) in the second at 4–6 months CA. The AUC of Ret-He for ID diagnosis was 0.64 (p = 0.002) in the first definition and 0.59 (p = 0.04) in the second. The optimal cut-off was 29.4pg in the first and 29.7 in the second definition. The sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) at the 29.4 pg cut-off were 50.0%, 78.7%, 60.8%, and 70.5% for definition 1 and 44.1%, 74.6%, 71.9%, and 47.5% at the 29.7pg cut-off for definition 2.ConclusionRet-He had low diagnostic accuracy for ID diagnosis in former VPI. Caution is advised when using Ret-He alone for ID diagnosis. Further research is needed to establish optimal approaches for identifying ID in VPI.

Keywords