PURPOSE: To measure retinal nerve fiber layer thickness (RNFLT) as a function of distance from the optic nerve head using a confocal scanning laser polarimeter, such as the Nerve Fiber Analyzer (GDx). METHODS: Twenty eyes of 20 healthy patients were recruited for this study. Each patient had a normal visual field and a healthy optic nerve head, which was assessed by slit-lamp biomicroscopy using a 90-diopter lens and by a scanning laser ophthalmoscope. Using the GDx, RNFLT was calculated from 1.1 disc diameters (DD) to 2.5 DD in 0.1-DD increments from the outer edge of the optic disc rim. RNFLT was successively evaluated for the entire annulus surface, for each quadrant, and for every 10 degree sector. RNFLT was calculated in retardation degrees. Differences in RNFLT were calculated by analysis of variance. RESULTS: When the entire RNFLT was considered, the measurements close to the optic nerve head (at 1.1 and 1.2 DD) were found to be significantly (0.05 < or = P < or = 0.01) different from those measured far from the disc (at 2.4 and 2.5 DD). The inferior quadrant had the greatest RNFLT followed by the superior quadrant. When RNFLT was analyzed for every 10 degrees, RNFLT at 1.1 and 1.2 DD was significantly (P < or = 0.001) different from that measured at 2.4 and 2.5 DD in almost all 36 sectors. In the 36 considered sectors, no significant difference was found for all the RNFLT values that were calculated in all the positions of the annulus surface between 1.1 and 1.7 DD. CONCLUSION: From these data, peripapillary RNFLT is shown to be significantly (P < or = 0.001) thinner in the periphery (2.5 DD) than around the optic disc (1.1 DD). However, the lack of any difference in RNFLT from 1.1 to 1.7 DD suggested that the mild variations to locate the optic disc contour line do not change the results of the analysis if the RNFLT is calculated within 1.8 DD.