Influence of learning effect on blue-on-yellow perimetry.

摘要

PURPOSE. Some studies have found that a significant blue-on-yellow perimetry (B/YP) learning effect exists in patients with ocular hypertension (OHT) or open-angle glaucoma who were experienced in standard automated perimetry. However, very little is known about the B/YP learning effect in normal subjects and patients without previous white-on-white perimetry (W/WP) experience. Meanwhile, it is unclear whether the B/YP learning effect is influenced by age and refraction. METHODS. Twenty healthy subjects, 26 OHT and 14 primary open angle glaucoma (POAG) patients, underwent three full-threshold B/YP tests at intervals of 7 to 21 days. Of the 60 subjects, 38 had no previous W/WP experiences, 22 had previous W/WP experiences for at least two times. The parameters investigated to detect a learning effect were the perimetric indices and the test duration (TD). RESULTS. Learning effects were demonstrated for mean deviation (MD), pattern standard deviation (PSD), short-term fluctuation (SF), and TD. Significant differences were found between the MD, PSD, SF, and TD of the first test and those of the second and third tests (p0.05). However, no difference was found between those parameters of the second and third tests. No statistically significant differences were noted in terms of MD1st-2nd, MD1st-3rd, PSD1st-2nd, PSD1st-3rd, TD1st-2nd, and TD1st-3rd between the group with W/WP experience and the group without previous W/WP experience (p0.05). No statistically significant differences were noted in terms of those parameters among the normal subjects and the patients with OHT or POAG (p0.05). No significant difference was found in the B/YP learning effect among various age groups (p0.05) and among various refraction groups (p0.05). CONCLUSIONS. A significant learning effect was observed between the first and the second or third tests and the perimetric indices appeared improved at full-threshold B/YP. The previous W/WP experience and the subject age and refraction did not influence the B/YP learning effect. This above all should be taken into account when considering the clinical use of this test to avoid erroneous diagnostic conclusions.