Work

Clinical and Laboratory Factors Associated with Exchange Transfusion in Neonates with Severe Hyperbilirubinaemia: A Retrospective Cohort Study

Year: 2026
Type: article
Abstract: Background and Objectives: Exchange transfusion (ET) remains a critical intervention for severe neonatal hyperbilirubinaemia, yet laboratory associations with ET receipt require careful interpretation. We examined characteristics associated with ET and post-procedure laboratory changes. Materials and Methods: This retrospective tertiary-centre cohort included 98 neonates evaluated between January 2020 and December 2025 (65 received ET; 33 did not). A four-covariate logistic regression model assessed associations with ET receipt. Discrimination, bootstrap optimism, cross-validated calibration and an exploratory creatinine-containing model were examined. Results: Peak total serum bilirubin (TSB) was higher among ET recipients than non-recipients (33.69 ± 6.31 versus 26.19 ± 3.64 mg/dL; p < 0.001). In the primary model (n = 98), peak TSB was associated with ET receipt (adjusted odds ratio 1.31 per mg/dL, 95% confidence interval 1.16–1.48; p < 0.001). The apparent area under the receiver operating characteristic curve (AUC) was 0.860, decreasing to 0.837 after bootstrap optimism correction. An exploratory six-covariate model (n = 97) yielded an apparent AUC of 0.874 and an optimism-corrected AUC of 0.836. Creatinine showed an adjusted association (odds ratio 1.58 per 0.1 mg/dL, 95% confidence interval 1.07–2.32), but its incremental AUC was uncertain (difference 0.014, 95% confidence interval −0.022 to 0.053). Following ET, bilirubin, platelet count and total calcium decreased. Conclusions: Peak TSB strongly distinguished infants who received ET, but this association partly reflects the bilirubin-based indication for treatment. Creatinine did not demonstrate a reliable incremental discrimination benefit. These findings reinforce guideline-based ET assessment and post-exchange monitoring, not the adoption of new treatment thresholds or prediction tools.
Language: English
Field: Medicine
Open Access status: gold
Estimated APC paid: $3,137
OpenAlex ID: W7225635104
FWCI:
Cites:
Cited by:
Related to: