Purpose or Case Report: The purpose of this study is to determine whether a decrease in hematocrit is predictive of an intracranial hemorrhage (ICH) on neonatal head ultrasound (HUS).
Methods & Materials: This is a retrospective study of 136 neonatal HUS between 2005 and 2017 at a single institution. The indication for the HUS was categorized as being related to hematocrit (e.g. “drop in hematocrit”) or unrelated to hematocrit. The medical record was also reviewed for the change in hematocrit in the 48 hours prior to HUS, a decrease in hematocrit, gestational age, number of days since birth, prematurity, and abnormal neurologic exam. The association between these variables and the presence or absence of hemorrhage on the HUS was analyzed using STATA 12.1.
Results: Twenty-one percent (N=29/136) of neonates had an ICH on their HUS. Studies that were ordered with an indication related to hematocrit were less likely to have an ICH on HUS compared to those with alternative indications for the study (odds ratio 0.35, p=0.018). The mean change in hematocrit (p=0.95), a decrease in hematocrit (p=0.30), and an abnormal neurological examination (p=0.25) were not associated with ICH on HUS. However, a lower gestational age (mean difference 4.47, p<0.001) and number of days since birth (mean difference 25.1, p=0.01) were both significantly predictive of ICH. Those with an ICH were also three times more likely to be premature, though this association was only borderline significant (odds ratio 3.03. p=0.09).
Conclusions: A decrease in hematocrit in the preceding 48 hours is not predictive of an ICH on neonatal HUS. Those with ICH on HUS had a significantly lower gestational age and were more likely to be born premature.