Agreement and repeatability of infrared pupillometry and the comparison method.

OBJECTIVE: To evaluate the accuracy and repeatability of the widely used comparison method of measuring pupil size.
DESIGN: Cross-sectional study. PARTICIPANTS: Fourteen eyes of seven healthy myopic subjects were examined. INTERVENTION: Two examiners made two repeated measures of pupil diameters of 14 eyes using Rosenbaum card comparison pupillometry and infrared pupillometry. Subjects fixated on a distant visual acuity chart, and pupils were measured under three luminance conditions. The agreement and inter-rater repeatability of both methods were determined. MAIN OUTCOME MEASURES: Outcomes were pupil diameters, limits of agreement, and coefficient of repeatability of two examiners.
RESULTS: The mean difference between the two techniques ranged from 0.3 to 0.5 mm. The limits of agreement within two standard deviations ranged from 2.4 to 2.8 mm. Coefficient of repeatability ranged from 0.6 to 1.4 mm for infrared pupillometry and 1.0 to 1.2 mm for Rosenbaum pupillometry. Inter-rater repeatability of Rosenbaum pupillometry was consistently pupil diameter biased. Pupil diameters measured with the Rosenbaum method were consistently larger than diameters measured with the infrared technique for both examiners under all luminance conditions.
CONCLUSIONS: Results indicate that although the mean difference in techniques was small, the range of the agreement between the Rosenbaum and the infrared techniques was large. The Rosenbaum method consistently overestimated pupil diameters and was subject to inter-rater repeatability bias. Rosenbaum pupillometry may not be appropriate when accurate pupil measurements are required. The results have implications for many clinical trials in ophthalmology, including those evaluating refractive surgery that use Rosenbaum comparison pupillometry.