Hematological parameters of bronchopulmonary dysplasia in preterm infants: a meta-analysis.

Background: Bronchopulmonary dysplasia (BPD) represents the most widespread and severe form of chronic pulmonary disease in preterm infants. Studies have revealed an association between BPD and hematological parameters (HPs); however, the findings are inconsistent. Objectives: This study utilized a systematic review and meta-analysis to summarize the association. Methods: The Web of Science, Cochrane Library, Embase, and PubMed were retrieved up to May 4, 2024, with an update on April 10, 2025. Studies investigating the correlation between HPs and BPD in preterm infants were included. This study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The quality of the studies was evaluated via the Newcastle-Ottawa Scale. Stata 18 and MetaDisc were utilized for statistical analysis. Results: This meta-analysis encompassed 20 studies with a total of 4,752 participants and investigated the association between HPs and BPD in neonates. Statistically significant differences were found between BPD and non-BPD infants for eight parameters: neutrophil (NEU) count, hemoglobin (HGB), monocyte count, hematocrit (HCT), neutrophil-to-lymphocyte ratio (NLR), red blood cell (RBC) count, platelet (PLT) count, and systemic inflammatory response index (SIRI). Further correlation analysis and effect size evaluation revealed that, while some parameters differed between groups, the association between key parameters and BPD risk was not significant. Specifically, the odds ratios (ORs) for HCT (OR = 1.33, 95% confidence interval [CI]: -0.11 to 2.77), PLT (OR = 1.0, 95% CI: 0.98-1.03), and HGB (OR = 1.38, 95% CI: 0.42-4.49) did not reach statistical significance, suggesting these three parameters may not be independent influencing factors for BPD risk. Concurrently, diagnostic performance analysis demonstrated limited discriminatory ability for NLR, with a receiver operating characteristic-area under the curve (ROC-AUC) of 0.670 (standard error [SE] = 0.054), and for PLT, with an ROC-AUC of 0.675 (SE = 0.067). Conclusions: The current study indicated significant differences in NEU, HGB, HCT, RBC, PLT, NLR, and SIRI between BPD and non-BPD patients, with elevated NEU, NLR, and SIRI and reduced HGB, HCT, RBC, and PLT. However, this study had its limitations. Further analysis requires more multicenter, large-sample prospective studies. Systematic Review Registration: https://www.crd.york.ac.uk/PROSPERO/, identifier CRD42024552716.