Impact of Normal Aging and Progression Definitions on the Specificity of Detecting Retinal Nerve Fiber Layer Thinning.
PURPOSE: To evaluate the specificity of current definitions used to identify progressive change of the average peripapillary retinal nerve fiber layer (RNFL) thickness measurements obtained on optical coherence tomography (OCT) imaging. DESIGN: Prospective observational cohort study. METHODS: Setting: University of California, San Diego. STUDY POPULATION: Seventy-five eyes from 45 normal participants. OBSERVATION PROCEDURE: Patients were seen at an average of 5.7 visits over 3.2 years, to determine the age-related average RNFL thickness changes and longitudinal measurement variability. Slope and variability estimates were used to reconstruct "real-world" OCT imaging measurements with computer simulations. MAIN OUTCOME MEASURE: False-positive rates for progression in normal eyes using different definitions. RESULTS: The estimated normal average RNFL thickness change over time was -0.54 ± 0.23 μm/year (P < .001). Even with a recent definition of progression that appeared to guarantee a high level of specificity by accounting for normal aging (requiring a significant negative slope that was more negative than the 5% lower limit of aging), 18% simulated normal eyes were still falsely identified as having progressed after 5 years of annual testing in a clinical practice scenario. However, this was reduced to 8% and 4% when trend-based analysis of progression was performed after adjustments using the mean and 5% lower limit of normal rates of aging, respectively. CONCLUSIONS: This study highlights how current definitions for detecting RNFL thinning have an unacceptably poor level of specificity, and that more stringent definitions are required to avoid misleading interpretations of progression on OCT imaging in clinical practice.
Wu, Z; Saunders, LJ; Zangwill, LM; Daga, FB; Crowston, JG; Medeiros, FA
Volume / Issue
Start / End Page
Electronic International Standard Serial Number (EISSN)
Digital Object Identifier (DOI)