Reevaluating 30-day head ultrasound screening for preterm infants in the era of decreasing periventricular leukomalacia

Abstract

Objective

Neonatal brain injury is a potentially devastating cause of neurodevelopmental impairment. There is no consensus, however, on the appropriate timing and frequency of routine head ultrasound (HUS) screening for such injuries. We evaluated the diagnostic utility of routine HUS screening at 30 days of life (“late HUS”) for detecting severe intraventricular hemorrhage (IVH) or cystic periventricular leukomalacia (c-PVL) in preterm infants with a negative HUS before 14 days of life (“early HUS”).

Methods

Single-center retrospective cohort analysis of infants born at ≤ 32 weeks gestational age (GA) admitted to the University of Nebraska Medical Center NICU from 2011–2018. Demographics, HUS and MRI diagnoses were abstracted from clinical records. Fisher’s exact test and t-test assessed associations between categorical and continuous variable, respectively.

Results

205 infants were included—120 very preterm (28–32 weeks GA) and 85 extremely preterm (<28 weeks GA). Negative predictive value of early HUS for predicting any clinically significant anomalies (severe IVH or c-PVL) on late HUS was 100% for extremely and 99.2% for very preterm infants. Term-equivalent MRI detected previously undiagnosed c-PVL in 16.7% of the 24 patients that received MRI; all infants with new c-PVL on MRI had severe IVH on early HUS.

Conclusion

Following negative early HUS, late HUS detected significant new abnormalities in one infant. These data suggest that in a unit with low prevalence of c-PVL, 30-day HUS may have limited clinical utility following negative screening. In infants with abnormal early HUS, clinicians should consider obtaining term-equivalent MRI screening to detect c-PVL.

Keywords:

• Brain injury
• intraventricular hemorrhage (IVH
• )
• neonate
• neuroimaging
• ultrasound

Acknowledgements

The authors would like to thank Dr. Purnima Guda from the UNMC Center for Clinical & Translational Research for support during data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

R.K. was supported by the Alpha Omega Alpha Carol L. Kuckein Student Research Fellowship.