Purpose: A test of validation study was conducted to investigate the use of a novel computerized portable pupillometer to detect and quantify relative afferent pupillary defects (RAPDs).
Materials and methods: Binocular pupillary response curves were recorded in patients with RAPDs clinically graded by an examiner (n = 32) and in normal subjects (n = 31) with RAPDs simulated using quantifiable dimmed light intensities. In 14 normal subjects, testing was repeated within two months. Pupillary constriction amplitude (CA), velocity (CV) and onset latency (COL) were used to calculate RAPDs.
Results: RAPDs in normal subjects were 0.16 ± 0.12 log units (LU) (range = 0-0.38). In retested normals, inter-visit variability was 0.21 ± 0.12 LU (range = 0-0.42). Significant correlation was found between RAPD values and dimmed light intensity in normal subjects (Pearson's r = 0.87, p < 0.0001) and between clinician and pupillometer grading of RAPDs in patients (r = 0.81, p < 0.0001). Using the upper limit of the one-sided 95% confidence interval (CI) of the ratio of percentage change in CA (LU) as determined from normals, 21/23 (91%) patients with RAPDs ≥ 0.5 LU were distinguished from normals. The area under the receiver operating characteristic curve for distinguishing RAPDs ≥ 0.5 LU was 0.98 (95% CI = 0.95-1.00). RAPDs calculated using CA and CV correlated more strongly with the clinician's grading compared to COL (Steiger's test p < 0.0001).
Conclusions: This novel pupillometer correlated strongly with an expert examiner's clinical grading of RAPDs and detected clinically significant RAPDs with high sensitivity and specificity, suggesting it may have a prominent role as an objective clinical tool in the screening of patients with vision loss.
Keywords: Diagnostic tests; pupil; pupillometry; relative afferent pupillary defect; swinging flashlight test.