To investigate age-related changes of the retinal nerve fiber layer (RNFL) imaged by a spectral-domain optical coherence tomography (OCT). Prospective, cross-sectional, and longitudinal studies. One hundred normal individuals were recruited for cross-sectional analysis, 35 of whom were randomly selected for longitudinal analysis. The circumpapillary average and quadrant RNFL thicknesses were measured by the Cirrus HD-OCT. In the longitudinal study, participants were followed at 4-month intervals for a mean of 30 months (range, 24-41 months) for RNFL and visual field measurements. Cross-sectional RNFL data were analyzed with multiple linear regression models with adjustment of spherical error, optic disc area, and signal strength. Longitudinal RNFL measurements were analyzed with linear mixed models with fixed coefficients on follow-up duration, baseline RNFL thickness, spherical error, optic disc area, and signal strength. Factors influencing the rate of change of RNFL measurements were analyzed in the interaction terms with "duration" in the linear mixed models. Rates of change of average and quadrant RNFL thicknesses. In the cross-sectional analysis, significant negative correlations were found between age and average (-0.33 μm/year; P = 0.011), inferior (-0.45 μm/year; P = 0.037), and temporal (-0.31 μm/year; P = 0.046) RNFL thicknesses. In the longitudinal analysis, the mean rates of change of average, superior, and inferior RNFL thicknesses were -0.52 (95% confidence interval [CI], -0.86 to -0.17), -1.35 (95% CI, -2.05 to -0.65) and -1.25 μm/year (95% CI, -1.78 to -0.71), respectively, after adjusting for baseline RNFL thickness, spherical error, disc area, and signal strength. There was no detectable RNFL reduction in the nasal and temporal quadrants. The only significant factor influencing the rates of change of RNFL measurements was the baseline RNFL thickness. A greater baseline RNFL thickness was associated with a faster rate of change. Progressive, age-related decline of RNFL thickness can be detected with longitudinal OCT imaging. Rate estimates derived from trend analysis for detection of glaucomatous RNFL progression should be interpreted with reference to the normal ranges of age-related reduction, particularly when the baseline RNFL measurement is large. Copyright © 2012 American Academy of Ophthalmology. Published by Elsevier Inc. All rights reserved.