Utility of umbilical cord blood 25-hydroxyvitamin D levels for predicting bronchopulmonary dysplasia in preterm infants with very low and extremely low birth weight

Abstract

Background and objective: There remains controversy regarding vitamin D deficiency and bronchopulmonary dysplasia (BPD) in very low birth weight (VLBW) and extremely low birth weight (ELBW) preterm infants. This study aimed to determine the prevalence of vitamin D deficiency assessed by umbilical cord blood 25-hydroxyvitamin D [25(OH)D] in preterm infants in northeast China and to evaluate the ability and optimal threshold of 25(OH)D for predicting BPD. Methods: The clinical data of VLBW and ELBW preterm infants with known cord-blood 25(OH)D levels were analyzed retrospectively. Infants were divided into groups based on their cord-blood 25(OH)D levels and BPD diagnosis. Logistic regression was performed to assess the risk factors for BPD and a nomogram was established. Receiver operating characteristic (ROC) curve analysis was used to evaluate the optimal threshold of cord-blood 25(OH)D concentration for predicting BPD. Results: A total of 267 preterm infants were included, of which 225 (84.3%) exhibited vitamin D deficiency and 134 (50.2%) were diagnosed with BPD. The incidence of BPD was lower in the group with a 25(OH)D level of >20 ng/ml than in the other groups (P = 0.024). Infants with BPD had lower cord-blood 25(OH)D levels than those without BPD (11.6 vs. 13.6 ng/ml, P = 0.016). The multivariate logistic regression model revealed that 25(OH)D levels (odds ratio [OR] = 0.933, 95% confidence interval [95% CI]: 0.891–0.977), gestational age (OR = 0.561, 95% CI: 0.425–0.740), respiratory distress syndrome (OR = 2.989, 95% CI: 1.455–6.142), and pneumonia (OR = 2.546, 95% CI: 1.398–4.639) were independent risk factors for BPD. A predictive nomogram containing these four risk factors was established, which had a C-index of 0.814. ROC curve analysis revealed that the optimal cutoff value of 25(OH)D for predicting BPD was 15.7 ng/ml (area under the curve = 0.585, 95% CI: 0.523–0.645, P = 0.016), with a sensitivity of 75.4% and a specificity of 42.9%. Conclusions: A cord-blood 25(OH)D level of <15.7 ng/ml was predictively valuable for the development of BPD. The nomogram established in this study can help pediatricians predict the risk of BPD more effectively and easily.