Using adaptive bracketing “normal” (N) strategy it has been shown that cluster trend analysis is useful for early detection of glaucomatous progression with Octopus perimetry. However, in clinical practice “tendency-oriented perimetry” (TOP) is mostly used. In TOP, responses given at neighboring points modify the threshold values, and this process does not respect cluster borders. In the current investigation, we evaluated the influence of test strategy on the cluster mean defect (cluster MD) values.Materials and Methods:
Sixty-nine eyes with normal visual field and 63 glaucoma eyes with predominantly mild to moderate glaucomatous visual field defect underwent Octopus perimetry with N and TOP strategy on the same day. Cluster MD values and global parameters were compared between test strategies in each group.Results:
In the normal group, all global parameters (P≤0.0003) and cluster MD for 6 uncorrected and 5 corrected clusters differed significantly between the strategies (P≤0.0052). In the glaucoma group, no significant difference was seen for global mean sensitivity and global MD, but square loss variance reflecting inhomogeneity was 1.0 dB lower with TOP (P<0.0001). Two uncorrected and 3 corrected clusters differed significantly (P≤0.0026) and several other clusters differed almost significantly between the strategies. The differences spread up to 1.9 dB without any systematic direction.Conclusions:
Using TOP, cluster MD determined for one cluster is statistically and clinically significantly influenced by responses given at points outside the cluster. This may also influence cluster trend analysis for glaucomatous progression. Therefore, TOP, although fast and easily performed, is not the optimal procedure for cluster and corrected cluster analyses.