Tianjin Medical Journal ›› 2024, Vol. 52 ›› Issue (7): 748-754.doi: 10.11958/20231602

• Clinical Research • Previous Articles     Next Articles

Construction and validation of a risk prediction model for bronchopulmonary dysplasia based on early platelet-related parameters

XUE Yuheng(), MAO Ning, LIU Wenqiang, YANG Qianqian, XU Yan, WANG Jun()   

  1. Department of Neonatology, the Affliated Hospital of Xuzhou Medical University, Xuzhou 221000, China
  • Received:2023-10-25 Revised:2023-12-26 Published:2024-07-15 Online:2024-07-11
  • Contact: E-mail:664586331@qq.com

Abstract:

Objective To develop and validate a risk prediction model based on early platelet-related parameters for bronchopulmonary dysplasia (BPD) in neonates admitted to the neonatal intensive care unit (NICU), and to facilitate early identification and intervention in high-risk populations. Methods Clinical data of 291 preterm infants with a gestational age (GA) ≤32 weeks or a birth weight (BW) <1 500 g, admitted to the NICU, were retrospectively analyzed. Out of these, 214 cases were selected as the modeling group. This group was further categorized into the BPD group (n=76) and the non-BPD group (n=138), based on whether they required oxygen therapy at 28 days post-birth. Perinatal data, platelet-related parameters and other indicators between the two groups. Univariate and multivariate Logistic regression analyses were conducted to identify BPD risk factors, followed by the construction of a nomogram. An additional cohort of 105 preterm infants with GA≤32 weeks or BW<1 500 g, were used to validate the model. This cohort was divided into the BPD group (n=43) and the non-BPD (n=62) group. Receiver operating characteristic (ROC) curve and calibration curve were used to internally verify the efficiency of the prediction model. Results The Logistic regression analysis identified GA, BW, Apgar score at 5 minutes≤7, invasive ventilation, platelet count (PLT) and mean platelet volume (MPV) as significant factors in the model (P<0.05). The constructed nomogram was formulated using R language, and the areas under the ROC curve (AUC) for the three models were 0.908, 0.931 and 0.918, respectively (P<0.05). The verification group was verified by Bootstrap. The calibration curve showed a good fit. The internal validation AUC values of the three models were 0.877, 0.890 and 0.886, respectively. Conclusion GA, BW, invasive ventilation, Apgar score at 5 minutes≤7, MPV and PLT are key risk factors for BPD onset. The risk prediction model based on these indicators can effectively predict BPD, providing clinicians with a valuable tool for early detection and intervention in the development of BPD.

Key words: bronchopulmonary dysplasia, infant, premature, platelet count, risk factors, forecasting

CLC Number: