[Application of amplitutude integrated electroencephalography for early diagnosis of brain injury in premature infants].

OBJECTIVE

To study the characteristics of amplitude integrated electroencephalography (aEEG) in preterm infants with brain injury.

METHODS

This study included 62 cerebral damage infants with 28-36 weeks gestational age (GA), and another 51 normal infants in control group, aEEG recording was performed to each infant during the first 48 h of life, the duration of each recording was at least 2 h. The features of aEEG, such as continuity(Co), sleep-wake cycling (Cy) and amplitude of the lower border (LB), were evaluated by semiquantitative analysis and compared between the two groups.

RESULTS

All the aEEG features were found having significantly lower values in brain injuries group (P < 0.05). Logistic regression of aEEG features to the presence of brain injury revealed that only Cy was significantly correlated to the outcome (OR = 0.217, P < 0.05). ROC curve demonstrated Cy of the best sensitivity and specificity with 0.769 AUC. Co, LB yielded 0.677, 0.602 AUC respectively. Correlation analysis of GA to Co, Cy, LB and total score showed significantly correlated, the correlation coefficient for Co, Cy, LB and total scores were 0.546, 0.488, 0.536, 0.588 respectively (P < 0.05).

CONCLUSION

The Cy in the initial aEEG is predictive for brain injury in premature infants with 28-36 weeks GA. The older the GA at birth, the more mature the aEEG pattern in premature neonates.