Disease Severity Threshold (Tipping Point) at which RNFL and GCIPL Glaucoma Progression Rates Fall

Abstract

Purpose : Ganglion cell-inner plexiform layer (GCIPL) and retinal nerve fiber layer (RNFL) thicknesses measured by OCT have been shown to be useful for glaucoma diagnosis and progression detection. The purpose of this study was to determine the disease severity threshold (tipping point) at which the longitudinal rate of change for RNFL and GCIPL thickness slows down; allowing to determine the preferred location to follow structural damage along the spectrum of the disease severity.

Methods : Subjects with comprehensive ophthalmic examination and ≥ 5 visits with qualified visual field (VF; Humphrey Field Analyzer; Zeiss, Dublin, CA) and OCT (Cirrus HD-OCT; Zeiss) optic nerve head and macular scans were enrolled. Piece-wise linear mixed effects model was used to identify the tipping points of RNFL and GCIPL vs. VF mean deviation (MD), respectively, and model with quadratic term for RNFL vs. GCIPL. To account for the difference in dynamic ranges, RNFL and GCIPL rate of change were standardized (normal distribution, mean=0 sd=1).

Results : 177 eyes (125 open angle glaucoma, 45 glaucoma suspect, and 7 healthy eyes) of 114 subjects were analyzed. Subjects’ mean age: 70 ± 11 years, median VF MD: -1.78 dB ([Q1, Q3]; -5.8, -0.2), baseline average RNFL and GCIPL thicknesses: 73.2 ± 14.9µm and 68.3 ± 10.7µm, respectively, average follow-up time: 3.4 ± 1 years, mean visits per subject: 5.6, with a grand total of 1,010 follow-up visits. Tipping point for RNFL occur at earlier stage of the disease and the rate of change is faster than for GCIPL (Table). However, after standardization, the slopes for RNFL and GCIPL were similar. When comparing the rate of change for RNFL and GCIPL, no tipping point was detected, but as RNFL and GCIPL decreased, the rate of change in RNFL became significantly smaller.

Conclusions : Rate of RNFL thinning slows down at an earlier stage of functional damage than GCIPL. Different dynamic ranges give the impression that RNFL decreases faster, but accounting for the larger dynamic range, a similar rate of change to GCIPL is observed across the entire disease severity spectrum. Our results do not indicate whether RNFL or GCIPL is better for detecting progression except for very advanced stages of the disease where RNFL progression rate stalls.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.