Longitudinal Changes in Optical Coherence Tomography Angiography (OCTA) Vessel Density and Thickness in Glaucoma Eyes with Focal Lamina Cribrosa Defects

Abstract

Purpose : To determine if the rate of longitudinal changes in vessel density and thickness is faster in glaucoma eyes with focal lamina cribrosa defects (LCD).

Methods : OptoVue Avanti OCTA macula (3 x 3 mm²) and ONH (4.5 x 4.5 mm²) scans and Spectralis optic nerve head (ONH) circle scans were obtained from 74 eyes of 64 glaucoma patients with 3 visits and > 1 year of follow-up from the Diagnostic Innovations in Glaucoma Study (DIGS). LCD were detected using swept-source OCT. Multivariable (including image quality) and weighted multivariable linear mixed models were used to assess differences in rates of change between eyes with and without LCD, with weights derived from propensity scores calculated in a boosted logistic regression, balancing age, baseline MD, and gender between LCD groups.

Results : The median follow-up time (2.7 years, range: 1.0 to 3.5), median number of visits (4) and mean (95% CI) baseline MD (-5.0 (-6.0, -4.0) dB) in the eyes with LCD (n=36) and without LCD (n=38) were similar. In eyes with and without LCD, the rates of change in ONH and macular superficial vessel density were significantly different from zero. A faster decrease in deep layer macula whole image vessel density (wiVD) and parafoveal vessel density (pfVD) was observed in the LCD positive group in the superior hemisphere (wiVD: (LCD rate (95% CI) -0.41%/yr (-0.94%, 0.12%), no LCD rate 0.38%/yr (-0.16%, 0.92%) p = 0.042), pvVD: LCD rate (95% CI) -0.40%/yr (-1.02%, 0.21%), no LCD rate 0.49%/yr (-0.12%, 1.11%) p = 0.044). The circumpapillary retinal nerve fiber layer (RNFL) decreased faster in the LCD positive group (LCD rate (95% CI) -0.69 (-1.17, -0.21) mm/yr, no LCD rate 0.17 (-0.31, 0.66) mm/yr, p=0.017), as did RNFL thickness in the inferior-temporal (LCD rate (95% CI) -1.91 (-3.04, -0.77) mm/yr, no LCD rate 0.01 (-1.16, 1.18) mm/yr, p=0.029). There were no differences found in the rate of ONH vessel density dropout between eyes with and without LCD.

Conclusions : With a relatively short follow-up of 2.7 years, this study suggests that ONH and microvascular dropout is detectable, and that the rate of RNFL thinning and deep layer vessel density loss is faster in eyes with LCD. Longer follow-up is needed to determine the relationship between microvascular dropout and lamina changes.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.