The detection of glaucoma progression with retinal nerve fiber layer thickness is not improved by mean sensitivity information

Abstract

Purpose : We previously reported that out of all possible combinations of four tests (two structural and two functional), retinal nerve fiber thickness (RNFLT) alone and in combination with mean sensitivity on static automated perimetry (SAP MS) had the greatest hit rate (percentage of eyes identified as progressing) (Abu et al., North American Perimetric Society Meeting, 2017). In this study, we used a larger sample to assess the sectorial hit rate for different length of follow-up with these tests.

Methods : We selected 254 eyes of 153 patients with either ocular hypertension or primary open-angle glaucoma enrolled in the Diagnostic Innovations in Glaucoma Study or the African Descent and Glaucoma Evaluation Study. Each patient had 10 visits separated by 3 to 24 months. For every visit, RNFLT and SAP MS were measured within a 30-day window. After converting SAP MS into linear units, both measurements were expressed as percent of mean normal. With specificity fixed at 95%, hit rate was obtained for RNFLT alone and for RNFLT and SAP MS combined from 4 to 10 visits in the following sectors: infero-temporal, supero-temporal and central 10 degrees (central 16 test points on the 24-2 test pattern). Hit rates and their 95% confidence interval were compared between the two criteria and between different sectors.

Results : Hit rates increased from visits 4 to 10 for both criteria across all sectors (Figure 1). From visits 4 to 6, hit rates for RNFLT alone were greater than RNFLT and SAP MS combined, however, similar rates were observed for both criteria after 7 visits. Across all visits, the smallest hit rates were obtained in the supero-temporal sector. Significant agreement (k = 0.632-0.637, p<0.01) was found between the two criteria globally and in all sectors (Figure 2).

Conclusions : In short follow-up series, RNFLT alone detected more eyes as progressing. Although combining RNFLT and SAP MS did not increase the hit rate, it identified significant change in some eyes which were undetected by RNFLT alone.

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

 

Figure 1: The global and sectorial hit rates and 95% CI obtained with RNFLT alone (dark green squares) and RNFLT and SAP MS combined (light green squares) are shown for visits 4 to 10.

 

Figure 2: Venn diagrams showing the proportion of eyes flagged as progressing by RNFLT alone and by RNFLT and SAP MS combined for 10 visits.