Abstract 【Abstract】 Objective To construct and validate a predictive model for bronchopulmonary dysplasia (BPD) in premature infants with gestational age ≤32 weeks based on perinatal data, mechanical ventilation status, and ultrasound scores of lung. Method Using a retrospective cohort study approach, a total of 184 preterm infants with gestational age ≤32 weeks admitted to Kunming Children's Hospital from January 2022 o August 2023 were included. According to the incidence of BPD, they were divided into BPD group (n=75) and non-BPD group (n=109). Baseline and perinatal data, lung ultrasound scores on the 3rd, 7th and 14th day after birth were compared between the two groups. Spearman correlation and partial correlation analysis were used to analyze the relationship between lung ultrasound scores and BPD. Multivariate Logistic regression analysis was performed to evaluate the risk factors of BPD in preterm infants. Based on the risk factors of
BPD in preterm infants, the prediction model was established. Bootstrap method, consistency analysis and decision curve analysis were used to evaluate the calibration degree, the consistency of BPD prediction model and the actual situation, and the clinical effectiveness of the model. Statistical methods were performed by independent sample t-test, Mann-Whitney U test, χ2 test or Fisher exact probability method, receiver operating characteristic (ROC)curve analysis. Result The gestational age of the BPD group was lower than that of the non-BPD group, and the proportion of small for gestational age infants and mean platelet volume (MPV) on the 3rd day were higher than those of the non-BPD group, and the duration of invasive and non-invasive mechanical ventilation was longer than those in the non-BPD group (all P<0.05). The ultrasound scores of
lung in the BPD group on the 3rd day , 7th day and 14th day after birth were significantly higher than those in the non-BPD group (all P<0.05). The results of Spearman correlation analysis showed that ultrasound scores of lung on the 3rd day (r=0.782), 7th day (r=0.814) and 14th day (r=0.875) after birth were positively correlated with BPD (all P<0.001). The results of partial correlation analysis showed that the ultrasound scores of lung on the 3rd, 7th and 14th day after birth were still positively correlated with BPD (all P<0.001), and the partial correlation coefficient was still increasing with the extension of birth time (0.740, 0.803, and 0.846, respectively). The results of multivariate Logistic regression analysis showed that gestational age, small for gestational age infants, duration of invasive mechanical ventilation, duration of non-invasive mechanical ventilation, MPV on the 3rd day, and ultrasound scores of lung on the 14th day after birth were independent risk factors for BPD in premature infants with gestational age ≤32 weeks (all P<0.05). Based
on the risk factors screened by multivariate Logistic regression, the predictive model for BPD in premature infants with gestational age ≤32 weeks was developed using R language, with a C-index of 0.820. The ROC analysis showed that the area under the curve (AUC) of the constructed prediction model was 0.820 (95% CI: 0.761-0.879). The internal calibration plot drawn by Bootstrap method showed that the calibration curve was close to the standard curve, with a calibration degree of 0.816, indicating that the constructed prediction model for BPD in premature infants with gestational age ≤32 weeks had good discrimination. External validation showed that the consistency of the prediction model for BPD in premature infants
with gestational age ≤32 weeks with the actual situation was 94.87%, with a Kappa value of 0.870 95% CI: 0.648-1.091), indicating good consistency. The clinical decision analysis curve showed that there was a positive net clinical benefit in predicting the occurrence of BPD in premature infants with a threshold probability of 0.06-0.40 using the constructed prediction model for BPD in premature infants with gestational age ≤32 weeks. Conclusion The prediction model of BPD in
preterm infants with gestational age ≤32 weeks based on perinatal data mechanical ventilation status, and lung ultrasound scores has good discrimination, calibration and clinical effectiveness. The BPD prediction is highly consistent with the actual outcomes. The model may serve as a useful tool for predicting BPD risk in clinical practice.
|
|
Received: 10 October 2024
Published: 01 October 2025
|
|
|
2025, Vol. 13 
