Higher prevalence of fundus haemorrhages in early-screened (NEST Study) as compared to late-screened (SUNDROP Study) newborn populations

Abstract

Background/aims: To determine whether timing of ophthalmic screening influences prevalence of neonatal fundus haemorrhages. We compared the prevalence of fundus haemorrhages in two populations: term newborns screened early (less than 72 hours) and preterm newborns screened late (4-11 weeks). Additionally, we reviewed the literature on timing and prevalence of newborn haemorrhages.

Methods: Retrospective observational cohort study. Infants who underwent wide-angle ophthalmic digital imaging over one overlapping year in the Newborn Eye Screen Testing (NEST) or Stanford University Network for Diagnosis of Retinopathy of Prematurity (SUNDROP) programme were included. The PubMed database was filtered to include English-language articles dating back to 1950. Nine articles were selected for review based on inclusion of the prevalence of newborn fundus haemorrhages at multiple time points.

Results: A total of 202 patients received early imaging in the NEST cohort and 73 patients received late imaging in the SUNDROP cohort. In the NEST cohort, 20.2% of newborns had haemorrhages. In contrast, we found haemorrhages in only one case or 1.4% of the SUNDROP cohort. Using prevalence data from nine additional studies, we developed a predicted probabilities model of newborn haemorrhages. Per this model, the probability of seeing a haemorrhage if you screen an infant at 1 hour is 18.8%, at 2 weeks is 2.9% and at 1 month is 0.28%.

Conclusion: We found a significant difference in the prevalence of fundus haemorrhages between the early-screened NEST cohort and the late-screened, preterm SUNDROP cohort. Likely, this difference is due to the transient nature of most newborn haemorrhages.

Keywords: child health (paediatrics); epidemiology; imaging; public health; retina.