To assess the reliability of the standard swinging flashlight test (S-SFT) for detection of a relative afferent pupillary defect (RAPD) in comparison to an alternative technique using a direct ophthalmoscope (O-SFT).

RAPD is an essential sign to determine the integrity of optic nerve function and is most commonly evaluated by the S-SFT. However, this test has limitations due to technical factors and interrater variability. Inherent patient characteristics such as miotic pupils and dark irises may also limit detection of a RAPD.

In this prospective validation tool study, 15 raters distributed in three subgroups according to their level of clinical experience (neurologists, neurology residents and medical students) were asked to identify and grade RAPDs using both the S-SFT and O-SFT. Sensitivity and specificity scores were also established for both techniques. Kappa scores were calculated to represent interrater agreement.

Fifteen raters evaluated 23 participants for each RAPD testing method. The examined cohort consisted of patients with RAPD (73.9%) and healthy controls (26.1%). The detection rates did not differ significantly between the S-SFT (70.03%) and O-SFT (66.09%). Among the S-SFT evaluations, detection rates for subgroups of medical students and neurologists were 75.29% and 66.74% respectively without statistically significant difference. Detection rates with the S-SFT increased with higher RAPD grades (1+ 61.57%; 2+ 76.56% and 3+ 90%). The correct grading rate for the S-SFT was only 41.75%. Results showed modest sensitivity and specificity scores for the S-SFT with 73.61% and 60.00% respectively. The S-SFT kappa score showed poor interrater agreement (-0.00275).

This study demonstrated low detection rate and poor interrater agreement for the S-SFT suggesting the need to explore other reliable and easily available methods to evaluate RAPD. The use of the direct ophthalmoscope as an alternative technique did not show improvement in detection rates of RAPD.