Figures 1 (left) and 2b-d show the age-corrected pointwise mean normal sensitivities standardized for a 45-year-old person at the test locations in Fig 2a. There was a gradual decrease in sensitivity (Fig 1a left) with eccentricity for the Goldmann size III results for the 60° surrounding fixation except for some test locations in the far peripheral edge (in red in Fig 2) and locations known to be susceptible to lid artifacts (in yellow), for which sensitivity dropped substantially. Figures 1b and c, and 2 c and d show data for the 30° to 60° test locations of the 60–4 test. Although similar findings were observed for the larger stimulus sizes, there were fewer edge locations for which sensitivity dropped substantially. The right side of Figure 1 shows the point-wise age-effect slopes (in dB per decade) for the three stimulus sizes. Note the change in slope is greatest for Goldmann stimulus size III and least for size VI.
Mean sensitivities were smallest for Goldmann size III at 22.2 dB. They were, on average, 7.1 dB and 9,2 dB greater for size V and VI, respectively. The average age effect for Goldmann sizes III, V, and VI with the 60–4 test was –0.97, –0. 76, and –0. 56 dB per decade, respectively. After correcting for age, the sensitivities decreased by about –0.37, –0.34, and –0.31 dB per degree of visual angle for Goldmann sizes III, V, and VI, respectively. All differences between Goldmann size III and size V, size III and size VI, and size V and size VI were all significantly different from zero. The average age effect for SITA Standard size III of the 24–2 test was –0.59 dB / decade.
Figure 3 shows the pointwise within- (left) and between-subject (right) standard deviations after age correction for the Goldmann size III with the 30–2 SITA standard and the 60–4 full-threshold tests (top), and for Goldmann sizes V (middle) and VI (bottom) for the 60–4 full-threshold test. The mean estimated within-subject standard deviations were 3.1, 2.8, and 2.4 dB, for Goldmann sizes III, V, and VI, respectively. Differences between Goldmann size III and size V, size III and size VI, and size V and size VI were all significantly different from zero. The mean estimated between-subject deviations were similar at 3.3, 2.8, and 2.5 dB. Differences between Goldmann size III and size V, size III and size VI were significantly different from zero, but the difference between size V and size VI was not after Bonferroni correction (p = 0.004). It is clear that the outer superior and nasal test locations have greater within- and between-subject variability than test locations elsewhere in the visual field for all stimulus sizes.
The mean increase in sensitivity on retest across test locations (learning effect) was less than 0.1 dB for all stimulus sizes. When the most eccentric zone was analyzed, there was less than 0.2 dB improvement.
Figure 4 shows the conditional test–retest distribution and limits established from the empiric 5th and 95th percentiles. For all three stimulus sizes, retest variability is small above about 25 dB and worsens below that level. The highest sensitivities for size III are in the low 30 dB, for size V in the mid 30 dB, and for size VI in the high 30 dB.
A typical example of 60–4 visual field examinations from a subject is shown in Figure 5. Notice the common inferior nasal depression or “dent” that does not respect the horizontal meridian, most prominent with the size III stimulus. We had two examples of marked nasal contractions, one of which is shown in Figure 6. Both subjects with these contractions had deep set orbits with prominent brows.
Average of examination times ± standard deviation for full-threshold testing of the 60–4 tests were 12.3 ± 1.1 minutes for Goldmann size III, 11.7 ± 1.0 for size V and 11.4 ± 0.9 for size VI testing. SITA Standard testing of the central 24° took 4.7 ± 0.5 minutes per test. The average false positive rate for the 60–4 size III testing was 1.6 ± 3.3 %, size V 1.4 ± 3.3 % and for size VI, 1.2 ± 3.1 %; the false positive rate for SITA Standard was 1.2 ± 1.6 %. The corresponding false negative rates were 3.2 ± 6.2 %, 2.4 ± 5.2 % and 1.2 ± 3.3 %; the rate for SITA Standard testing was 0.7 ± 1.9 %.