Abstract:

Objective To investigate the characteristics of early blood glucose fluctuations in premature infants of different gestational ages and to further analyze the value of early blood glucose fluctuation characteristics in assessing periventricular-intraventricular hemorrhage （PIVH）. Methods A retrospective study was conducted， and the medical records of 175 premature infants admitted to Nantong First People's Hospitalfrom January 2023 to December 2025 were collected. They were divided into the early premature infant group （27 - 31 weeks） and the middle-late premature infant group （32 - 36 weeks）. Blood glucose was monitored on the day of admission and within 72 hours. The main parameters were the first blood glucose （ABG）， mean blood glucose （MEBG）， standard deviation of blood glucose （SDBG）， coefficient of variation of blood glucose （CV）， mean amplitude of blood glucose fluctuation （MAGE）. The characteristics of early blood glucose fluctuations in the two groups were compared. Head ultrasound examinations were performed at least twice within 2 weeks of hospitalization. The occurrence of PIVH was classified into the PIVH group and the non-PIVH group. Baseline data of the two groups were collected. Univariate and Logistic regression multivariate analyses were conducted to identify independent predictors of PIVH occurrence. A model was constructed， and the predictive value of single independent predictors and the model for PIVH occurrence was analyzed using the receiver operating characteristic curve （ROC）. Results Compared with the middle-late premature infant group， the ABG、MEBG， SDBG， CV， and MAGE were higher in the early premature infant group， and the differences were statistically significant （P < 0.05）. Compared with the non-PIVH group， the early premature infant group with a gestational age of 27 - 31 weeks， gestational diabetes， gestational hypertension， and the use of vasoactive drugs within 24 hours before delivery were higher， the use of prenatal glucocorticoids was lower， the birth weight， 1-minute Apgar score， 5-minute Apgar score， and hemoglobin level were lower， and ABG， MEBG， SDBG， CV， and MAGE were higher. The differences were statistically significant （P < 0.05）. Logistic regression multivariate analysis showed that the independent predictors of PIVH occurrence in premature infants were gestational age （OR：2.834；95%CI：1.016 - 7.904）， gestational diabetes（OR：2.575；95%CI：1.018 - 6.514）， and CV（OR：1.047；95%CI：1.001 - 1.095）. The ROC curve analysis showed that the AUC values of single independent predictors and the model for predicting PIVH occurrence ranged from 0.630 to 0.760， with the model having the highest predictive value for PIVH occurrence， with an AUC value of 0.760， and the sensitivity and specificity were 60.00% and 80.70%. Conclusions The more intense early blood glucose fluctuations in low-gestational-age premature infants， and a high CV value are independent predictors of PIVH occurrence along with low gestational age， and gestational diabetes. The construction of a model based on early blood glucose fluctuation indicators is helpful for predicting PIVH occurrence and providing important guidance for clinical decision-making and prognosis assessment of premature infants.

Key words: premature infants, gestational age, early blood glucose fluctuations, periventricular-intraventricular hemorrhage