Histological basis for the variability of visual field defects in early glaucoma.

Abstract

Glaucoma is characterized by pressure-induced damage to retinal ganglion cells (RGC) leading to progressively irreversible vision loss. Progression of visual field changes and retinal nerve fibre layer (RNFL) thinning are monitored with standard automated perimetry (SAP) and optical coherence tomography (OCT), respectively. In early glaucoma, it is difficult to interpret these tests because up to 20–50% of RGCs can be lost before characteristic visual field defects manifest. As such, there is a nonlinear relationship between anatomy and function, that is, the number of functioning RGCs and visual field (Harwerth et al. 2010). A commonly proposed explanation for this nonlinear relationship is the redundancy among RGC axons. RGCs have overlapping receptive fields, and therefore, sub-threshold cell losses could be compensated for by other RGCs that share the same functional retinal area (Kerrigan-Baumrind et al. 2000). Here, we discuss histological evidence that supports RGC redundancy as this may impact interpretation of visual field defects in early glaucoma. Optic nerve head injections of horseradish peroxidase (HRP) demonstrate a consistent topographic organization of RGC axonal bundles (Fig. 1A, Minckler 1980). This organization explains the arcuate nerve fibre bundle defects and corresponding Bjerrum scotomas seen in moderate and severe glaucoma. However, the linear arrangement of nerve fibre bundles may not provide sufficient anatomic resolution to explain visual field changes in early glaucoma. Additionally, in serial sections of HRP-labelled retina, Minckler consistently found unlabelled RGCs intermixed with labelled RGCs. There are immense overlap and variability in RGCs’ receptive fields. A single retinal laser burn of 100 μm width resulted in axonal degeneration in over 75% of one quadrant of the lamina cribrosa (Ryu & Minckler 1983). Furthermore, a fraction of axonal fibres actually pass through neighbouring fibre bundles, thus arborizing the topographic organization of the individual nerve bundles (Fig. 1B). In other words, there is a sub-population of RGC axons that take a meandering course to the optic disc and adhere only loosely to the expected retinotopic organization (Ogden 1983). These findings explain how a focal defect in the retina can become much more ramified downstream. Any spatial point in the visual field may be widely distributed and represented by multiple points at the optic nerve head. If one point of the nerve head sustains damage, the corresponding area in visual field can still be intact due to other ‘collaterals’. This explains the heterogeneous spectrum of visual field changes seen in early glaucoma. Only when a wider area of damage occurs at the optic nerve head will the visual field defects appear more definite and better correspond to the traditionally illustrated topography of the retinal nerve fibre array. Early stage glaucoma is difficult to detect and correlate with visual field testing. Non-specific visual field defects may be dismissed as non-glaucomatous for not fitting the classic patterns, and up to two-thirds of visual field defects disappear upon retesting. Given these histological findings and clinical variability, we encourage careful consideration when evaluating visual field defects that may at first seem to be non-glaucomatous.