Intraobserver reproducibility of retinal nerve fiber layer measurements using scanning laser polarimetry and optical coherence tomography in normal and ocular hypertensive subjects.

PURPOSE: To evaluate quantitatively the intraobserver reproducibility of measurements of the retinal nerve fiber layer (RNFL) in healthy subjects and an ocular hypertensive population using two nerve fiber analyzers. METHODS: Sixty eyes of normal (n=30) and ocular hypertensive subjects (n=30) were consecutively recruited for this study and underwent a complete ophthalmologic examination and achromatic automated perimetry. RNFL were measured using scanning laser polarimeter (GDx-VCC) and optical coherence tomography (OCT Model 3000). Reproducibility of the RNFL measurements obtained with both nerve fiber analyzers were compared using the coefficient of variation. RESULTS: In both groups the authors found fair correlations between the two methods in all ratio and thickness parameters. The mean coefficient of variation for measurement of the variables ranged from 2.24% to 13.12% for GDx-VCC, and from 5.01% to 9.24% for OCT Model 3000. The authors could not detect any significant differences between healthy and ocular hypertensive eyes, although in normal eyes the correlations improved slightly. Nev-ertheless, the testretest correlation was slightly better for GDx-VCC than for OCT Model 3000 (5.55% and 7.11%, respectively). CONCLUSIONS: Retinal mapping software of both nerve fiber analyzers allows reproducible measurement of RNFL in both healthy subjects and ocular hypertensive eyes, and shows fair correlations and good intraobserver reproducibility. However, in our study, GDx showed a better testretest correlation. (Eur J Ophthalmol 2004; 14: 523-30).