A nomogram for early identification of abo hemolytic disease in neonates with hyperbilirubinemia and abo incompatibility: a retrospective study

Abstract

Purpose

To develop and validate a nomogram for early identification of ABO hemolytic disease of the newborn (ABO-HDN) in neonates with hyperbilirubinemia and maternal–fetal ABO incompatibility, using routinely available clinical and laboratory parameters.

Methods

This retrospective study included 344 neonates with hyperbilirubinemia and maternal–fetal ABO incompatibility who completed the three neonatal hemolysis tests (2023–2024). ABO incompatibility was defined as a mismatch in the ABO blood group system between the mother and the neonate. They were divided into the ABO-HDN group (positive DAT and/or antibody release test) and the isolated hyperbilirubinemia group (all three tests negative). The isolated hyperbilirubinemia group also had ABO incompatibility but did not develop ABO-HDN, and served as the control group. Cases with positive free antibody test alone were excluded to ensure a gold-standard definition for model derivation. Demographic, perinatal, neonatal complications, maternal prenatal complications, and laboratory data (within 24 h of admission) were collected. Missing data were imputed using MICE. LASSO regression selected predictors, followed by multivariate logistic regression to build the nomogram. Model performance was assessed by ROC curves, calibration curves, Decision curve analysis (DCA), clinical impact curve (CIC), and Bootstrap internal validation (1000 resamples).

Results

A total of 344 neonates (154 ABO–HDN, 190 isolated hyperbilirubinemia) were included. LASSO and multivariate logistic regression identified four independent predictors: age at sampling, time of jaundice onset, red blood cell count (RBC), and neonatal blood type (all P < 0.05). The nomogram showed good discrimination (AUC = 0.826, 95% CI: 0.783–0.869), excellent calibration (Hosmer–Lemeshow P = 0.774), and higher net benefit at thresholds 0.01–0.98. Bootstrap–corrected AUC was 0.830.

Conclusion

This nomogram serves as a bedside-friendly decision-support tool for early identification of ABO-HDN in neonates with hyperbilirubinemia and maternal–fetal ABO incompatibility, particularly before hemolysis test results become available. By translating four routinely available parameters into an individualized probability, the model enables clinicians to expedite risk stratification while awaiting confirmatory testing. It provides a quantitative reference that may help clinicians estimate the probability of ABO–HDN and decide whether to intensify monitoring or pursue further evaluation. This practical, low-cost tool addresses an unmet need in routine neonatal care by offering an objective probability without additional testing costs or delays.