Reproducibility of Spectral-Domain Optical Coherence Tomography Measurements in Adult and Pediatric Glaucoma

Abstract

To determine the reproducibility of spectral-domain optical coherence tomography (OCT) measurements of peripapillary retinal nerve fiber layer (RNFL) and macular retinal thickness of the central 20-degree area of the posterior pole in adult and pediatric glaucoma and normal subjects. Prospective. Normal, glaucoma suspect and glaucoma eyes of adult and pediatric subjects were recruited. Spectralis OCT scans were obtained 3 times on the same day (intravisit) and 4 weeks later (intervisit). Six individual sectors, average RNFL, and macular retinal thickness using an 8×8 mm grid centered on the foveal pit were measured for all the eyes. The central 16 square areas centered on the foveal pit were called the central macular area and the surrounding 48 square areas were called the perimacular area. Intraclass correlation coefficient, coefficient of variation (COV), and test-retest variability were calculated. Included were 51 adult eyes and 25 pediatric eyes. The intravisit and intervisit intraclass correlation coefficient for the central macular, perimacula, and average RNFL was very good (>0.8) for all adult and pediatric normal and glaucomatous eyes. The mean COVs of normal and glaucomatous eyes were higher for average RNFL (1.5 ± 1.5% adults, 1.0 ± 0.9% pediatric) compared to the central macula (0.56 ± 0.54% adults, 0.68 ± 0.64% pediatric) and perimacular thickness (0.64 ± 0.51% adults, 0.57 ± 0.42% pediatric) (1-way analysis of variance, P<0.05). No difference was noted in the intravisit and intervisit COVs between the adult and pediatric eyes of central macula, perimacula, and average RNFL thickness (unpaired t test, P>0.05). No intervisit reproducibility analysis was available for pediatric glaucoma eyes. The reproducibility of Spectralis spectral-domain-OCT were very good for both RNFL and macular thickness measurements in glaucoma/glaucoma suspect eyes of adult and pediatric subjects.