Fetal hemoglobin decline predicts retinopathy of prematurity and mediates the effect of red blood cell transfusion
In brief
In preterm infants, falling fetal hemoglobin accounts for 72% of the transfusion-ROP link
Among 403 infants born before 32 weeks, 11.7% developed retinopathy of prematurity; each 10-point drop in fetal hemoglobin by day 14 was linked to 34% higher odds. The decline over 28 days statistically mediated 71.8% of the transfusion association, and an internally validated risk model performed well. Whether preserving fetal hemoglobin prevents ROP remains untested.
- Journal
- Neonatology (Q1)
- Published
- 10 October 2026
- Study design
- Prospective / inception cohort
- Evidence level
- Level 2, Moderate (CEBM 2b)
- Authors
- Laura Torrejón-Rodríguez, Alba Cayuela-Hernández, Alejandro Pinilla-González, Nerea Valles-Murcia, Abel Albiach-Delgado, Julia Kuligowski, et al.
- PMID
- 42858294
- DOI
- 10.1159/neo/adsag022
Why clinicians should know about it
- Picked for Neonatology (paper of the day, 12 October 2026): HbF decline predicts ROP risk
Abstract
BACKGROUND: Retinopathy of prematurity (ROP), the leading cause of preventable childhood blindness in preterm infants, is associated with red blood cell transfusions from adult donors. Whether this risk is driven by the resulting loss of fetal hemoglobin (HbF) remains unclear. We assessed whether postnatal HbF decline mediates the transfusion-ROP association and developed a predictive tool based on HbF kinetics to estimate individual ROP risk. METHODS: We conducted a prospective cohort of infants below 32 weeks' gestation (2021-2026). HbF was measured at birth and on days 7, 14, and 28. A LASSO-regularized logistic model identified independent ROP predictors and generated a probability-based prediction tool. Causal mediation analysis, adjusted for gestational age, assessed whether HbF decline mediates the transfusion-ROP association. RESULTS: Among 403 infants, ROP developed in 11.7%. Day-14 HbF decline and gestational age were retained as main predictors of ROP. Each 10-percentage-point drop in HbF by day 14 increased ROP odds by 34% (OR, 1.34; 95% CI, 1.11-1.63; P<0.001). The model showed strong discrimination on internal validation (2000-repetition bootstrap with cross-validation: AUC 0.90; NPV 98.1%). In mediation analysis, HbF decline over the first 28 days accounted for 71.8% of the total effect of transfusion on ROP (P=0.003), suggesting HbF loss, not transfusion per se, drives this association. CONCLUSIONS: Postnatal HbF decline is an independent predictor of ROP and the dominant pathway linking transfusion to retinal injury in preterm infants, supporting further research on HbF preservation as a preventive strategy. Our HbF kinetics-based tool showed high discriminative accuracy (AUC 0.90), enabling early, individualized risk stratification.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.