Laminar disinsertions and optic disc hemorrhages in glaucoma

Abstract

 

Purpose

 

<br /> Recent studies suggest a causal relationship between optic disc hemorrhages (ODH) and peripheral lamina cribrosa defects. We objectively determined the frequency of laminar disinsertions (LD) in glaucoma patients with and without a history of ODH, as well as the spatial correspondence between LD and ODH. The strength of the current study design was that patients were selected on the basis of a history of ODH without prior review of OCT images, minimizing bias introduced by retrospectively determining ODH history in eyes with LD.

 

Methods

 

With stereo optic disc photographs supplemented with clinical notes from 2 prospective studies, we identified 52 eyes of 46 patients with a history of ODH (ODH group). We then identified 52 control eyes of 46 patients with no documented history of ODH (non-ODH group). The date and position of the ODH was noted. Spectral domain OCT (high resolution, 24 radial B-scans centred on the optic nerve head in enhanced depth imaging mode; Spectralis, Heidelberg Engineering) images were de-identified. A trained observer, masked to patient group and position of ODH determined the presence of LD in each of the 48 positions (2 per B-scan) in each image with a confidence score of 1 (most confident) to 5 (least confident). For this analysis only LDs with a score of ≤ 3 were included. The frequency and spatial location of LD in the two groups as well as the spatial correspondence of LD to ODH was examined.

 

Results

 

The ODH and non-ODH groups matched for age [70.6 (sd, 9.6) and 74.0 (sd, 11.4) years, respectively (P = 0.11)] and visual field damage [Mean Deviation = -6.11 (sd, 5.56) and -6.52 (sd, 5.01) dB, respectively (P = 0.70)]. Eyes in the ODH group were almost twice as likely to have LD compared to eyes in the non-ODH group (P < 0.01). Fifty (96%) eyes in the ODH group had LD, with 32 (62%) eyes having multiple LD, while 27 (52%) eyes in the non-ODH had LD, with 7 (13%) eyes having multiple LD. The width of LD in the ODH and non-ODH was similar [28.8° (sd, 16.3°) and 28.2° (sd, 18.9°) respectively (P = 0.92)]. Of the total of 84 ODH, only 33 (39%) corresponded spatially with LD, while the remaining ODH occurred in areas with no LD (Figure).

 

Conclusions

 

Our results confirm the strong association between LD and ODH. However, a significant number of LD also occur in patients with no documented history of ODH. Finally, there was a relatively poor spatial correspondence between LD and ODH.