Abstract
Purpose :
To compare the results of indirect ophthalmoscopy versus determination of a red reflex in newborns without any risk factor for retinopathy of prematurity (ROP).
Methods :
A cross sectional study of 71 patients in the intermediate neonatal intensive care unit of the Hospital San Juan de Dios, a third level reference center in San Jose, Costa Rica were initially screened by checking the red reflex between July and September of 2017. Immediately after, they were pharmacologically dilated and indirect ophthalmoscopy was performed. Inclusion criteria included children ≥ 34 weeks of gestational age; birth weight ≥ 1500 g; no history of retinal laser or intravitreal injections and no history of ROP. Patients were excluded if the records were incomplete or if their hemodynamic condition was unstable.
Results :
A total of 142 eyes from 71 patients (40 males, 30 female and 1 undefined) were evaluated. The average birth weight was 2876.5 ± 557 g (range <a href="tel:<a href="tel:4100-1790">4100-1790</a>">4100-1790</a>); the average gestational age was 37.9 ± 2.3 weeks (range 34-41); the average APGAR at 1 min was 7 ± 2 and 8 ± 1 at 5 min; the average age was 4.9 ± 7 days (range 55-0.29).
We calculated a specificity of 89.91%, sensitivity of 71.92%, a positive predictive value of 74.46% and a negative predictive value of 88.42% for the neonatal red reflex. We found a total of 20 eyes with an altered red reflex. Indirect ophthalmoscopy revealed that 4 eyes had leucocoria (2 secondary to optic disc coloboma and the other 2 due to unilateral cataracts), 4 eyes had posterior pole hemorrhages and 4 were false positive results.
Indirect ophthalmoscopy revealed that a total of 31 eyes had posterior pole hemorrhages, 87% of these had a negative red reflex. The most common pattern of hemorrhage seen was Roth spot-like (66% of the cases). Twelve (38.7%) eyes with hemorrhages were associated with sepsis and 2 (6.4%) eyes with congenital syphilis.
Conclusions :
A normal neonatal red reflex does no exclude intraocular pathology. It has a specificity of nearly 90%, but its sensitivity is only 72%. Most retinal hemorrhages are undiagnosed since they can’t be detected with a red reflex.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.