Comparison of Spectralis-OCT, GDxVCC and GDxECC in assessing retinal nerve fiber layer (RNFL) in glaucomatous patients.

BACKGROUND: Glaucomatous optic neuropathy is characterized by a progressive loss of retinal ganglion cells (RGCs). The defects in the peripapillary retinal nerve fiber layer (RNFL) have been reported to be the earliest sign of glaucoma. We determined the agreement between RNFL thickness assessments from spectral-domain OCT (Spectarlis HRA + OCT; Heidelberg Engeneering, Heidelberg, Germany), scanning laser polarimetry (SLP) with variable cornea compensation (GDxVCC; Carl Zeiss Meditec, Dublin, CA, USA), and SLP with enhanced cornea compensation (GDxECC; Carl Zeiss Meditec, Dublin, CA, USA) in glaucomatous patients. Furthermore, we investigate the influence of typical scan score (TSS) on the results of GDx assessments.

METHODS: The enrolled subjects were devided into different groups by modified HODAPP visual field criteria. The peripapillary RNFL thickness was assessed with the three devices . ANOVA test, Pearson and Spearman correlation coefficient, and Bland-Altman plots were used to analyse the RNFL thickness assessments.

RESULTS: Ninety-two eyes from 92 glaucomatous subjects were analysed. These were divided into four groups: preperimetric glaucoma (n = 26), mild glaucoma (n = 18), moderate glaucoma (n = 21), and severe glaucoma (n = 27). For Spectralis-OCT, the average RNFL thickness (mean ± SD) was 99.25 ± 26.31 μm, 80.52 ± 16.63 μm, 71.59 ± 21.15 μm, and 63.85 ± 20.86 μm for preperimetric, mild, moderate, and severe glaucoma respectively. For GDxVCC, the corresponding assessments were 52.63 ± 8.18 μm, 52.95 ± 10.20 μm, 46.77 ± 10.62 μm, and 49.70 ± 13.34 μm. For GDxECC, the assessments were 49.35 ± 6.52 μm, 45.92 ± 7.21 μm, 42.19 ± 8.00 μm, and 39.53 ± 8.45 μm. All Spectralis-GDxVCC and Spectralis-GDxECC differences were statistically significant by ANOVA test. The differences between GDxVCC and GDxECC were statistically significant only for severe glaucoma. There was a highly significant correlation between Spectralis-OCT and GDxECC, as well as Spectralis-OCT and GDxVCC, in assessing the RNFL thickness. The best instrument agreement was found between GDxECC and Spectralis-OCT. The RNFL thickness assessed with Spectralis-OCT and GDxECC showed a better correlation to visual field defects than GDxVCC. Evaluating GDx assessments with typical retardation pattern GDxVCC and GDxECC showed very similar RNFL thickness results.

CONCLUSIONS: RNFL thickness assessments between GDxVCC, GDxECC, and Spectralis-OCT cannot be directly compared. The assessments are generally higher with Spectralis-OCT than with GDxVCC and GDxECC, because of differences in method of the devices. The atypical retardation pattern has a major impact on the RNFL thickness results of GDx devices. This must be taken into account when evaluating the assessed RNFL thickness results.