Video color perimetry: impairment in glaucoma suspects

Documenta Ophthalmologica 103: 81–90, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

Video color perimetry: impairment in glaucoma suspects NERI ACCORNERO1, MARCO CAPOZZA1, ALESSIA DE FEO1, STENO RINALDUZZI1, MILENA DE MARINIS1, JOSE PECORIGIRALDI2, ANTONELLA MOLLICONE2 and VERONICA VOLANTE2

Dipartimento di Scienze Neurologiche and 2 I Clinica Oculistica, University of Rome ‘La Sapienza’ Rome, Italy Accepted 9 May 2001

Abstract. Purpose: To detect mild visual field impairment in asymptomatic glaucoma suspect patients. Methods: Color perception within the visual field was tested with customized color video perimetry. The key features of the system were stimuli color desaturation, low-level luminance and equiluminant gray background. Twenty patients with asymptomatic glaucoma were tested and compared with a group of age-matched control subjects. Results: Automated perimetry test findings differed significantly in the two groups, particularly for short-wavelength sensitivity (blue). The severity of color impairment correlated directly with intraocular pressure. Conclusion: Desaturated low-luminance video perimetry will reliably detect and quantify asymptomatic visual field defects. A previous work on multiple sclerosis has detected a mild long-wavelength (red) impairment in asymptomatic patients after an episode of optic neuritis, even in clinically unaffected fellow eyes. Our findings in glaucoma suspect patients indicate that a mild blue impairment could be the initial sign of this disease. Key words: glaucoma, color perimetry, color perception, automatic perimetry, isoluminance

Introduction The initial manifestation of many visual pathway diseases is mild impairment of color perception. The subject is rarely aware of the defect unless it has a sudden, acute onset as in some cases of vascular occipital cortical damage or exogenous intoxication. Most patients with retinal degeneration, glaucoma and multiple sclerosis have long-standing mild color perception impairment before substantial vision loss develops. Because color defects may initially affect perifoveal rather than foveal vision, routine foveal acuity color testing [1–3] (for example Lanthony, Ishihara and Munsell) may miss defects in the early stages of the disease. Conversely, peripheral color analysis has shown high sensitivity [4–23]. Recently we developed a software procedure running on a standard PC that has proven

82 handy and efficacious in routine clinical use and is easily clonable. In a previous study [22] we showed that patients with asymptomatic optic neuritis in multiple sclerosis differed significantly from control subjects, notably in detecting desaturated red. In this paper we present our color perimetry findings in glaucoma suspect (GS) patients. Methods For this study we used a video color perimetry device, developed a few years ago and described in detail elsewhere [22, 24]. During testing, a personal computer software program displayed on the monitor, positioned at a calibrated distance from the patient’s eye (30 cm), a random sequence of round, colored, desaturated patches, 1◦ in diameter with linearly smoothed borders extending 0.2◦ , on a gray equiluminant background. Color patches appeared first at a lower saturation level (level 1); after 300 ms, if the subject failed to detect the stimulus, saturation was increased to a higher level (level 2) and 300 msec later the patches disappeared. Level 1 of saturation was defined as a distance of 0.5 unit of the CIE 1931 (x–y) chromaticity diagram grid from the central gray equiluminant point (x = 0.34 y = 0.34 k = 4500) and level 2 as 1 unit along the two color axes (red and blue) of the monitor used [24]. The orientation of color axes differs slightly according to the type of monitor used but this seemed not to affect test results. The key feature of this visual field testing procedure is the use of equiluminant color contrast stimuli. Neither LED nor fiberoptic projection perimetry systems can provide these features because in these devices the target stimuli are superimposed on the background illumination. Equiluminance was checked with a professional TV photometer, thus avoiding the heterochromatic flicker modality and similar subjective methods. These have some drawbacks. First, they mainly test foveal sensitivity alone, and in a system exploring, as ours does, the central visual field, a foveal impairment could bias the entire examination. Secondly, having to adjust color and saturation settings individually, makes perimetric test results difficult to compare and could also normalize an abnormal perimetric test. Furthermore, the low luminance level (10 cd/m2 ), needed to reduce screen glare, and the low color contrast that we used to test preeminently the parvocellular system, make it difficult for subjects to detect the minimum perceived flickering of a spot swapping cyclically between a desaturated color and the reference gray required to determine equiluminance. With low luminance, subjective equiluminance can be obtained within a wide range of color saturation. Detection of stimuli luminance and color contrast could be tested in the same session because the randomly presented target sequences included a

83 mixture of equiluminant colored patches, typically red or blue at 10 cd/m2 , and achromatic gray patches with the same CIE coordinates of the gray background displayed at two levels of luminance: 12 and 15 cd/m2 . These values were chosen to obtain similar quantitation scores (QPI, see later section) for the three tests in the control population. Because of possible monitor phosphor decay, stimuli luminance and color values were checked monthly with a Minolta xy-1 chroma meter. Subjects were requested to gaze fixedly, with one eye, at a central marker on the monitor and to press a switch as soon as they perceived a spot on the screen, regardless of its position, color or brightness. Central fixation was checked by an infrared television camera directed toward the patient’s eye. Poorly performed tests (those in which the blind spot appeared smeared or absent) were discarded or repeated. Each of the 88 points of a grid on the visual field, spaced 3◦ in the inner portion (10◦ ) of the visual field and 4◦ in the outer field within 24◦ × 40◦ (owing to the shape of the monitor) were randomly sampled, for position, color and luminance. If the three tests for the same point yielded discrepant results, that point was automatically retested. The examination took about 10 min per eye. At the end of the examination, after automatic interpolation and smoothing, three maps (luminance and two colors) of 960 pixels for each eye were displayed on the screen together with quantitative information (Figure 4). This consisted of percentage values for the number of stimuli perceived at level 1, at level 2 and not perceived at all, separately for each chromatic and achromatic test. For each map, a quantitative perimetric index (QPI) was then computed by subtracting the number of the stimuli not perceived from the number of stimuli perceived at level 1. The QPI index was then normalized between 0 and 100 according to the following equation:

L0 − L1 + 960 × 100 960 × 2 This kind of quantitation is useful for statistical analysis and for routine clinical diagnosis. Red–blue contrast (RBC), computed as Red QP − Blue QP 1 × 100 Luminance QP differentiates RBC-positive prevalent blue impairment (glaucoma and retinitis), from RBC-negative prevalent red impairment (multiple sclerosis, optic neuritis), taking into account the individual luminance contrast score.

84 Table 1. Quantitative perimetry indexes (QPI) Luminance QPI

Red QPI

Blue QPI

RBC

Control subjects (n◦ 34; 93.10±3.92 78.7±4.84 78.08±8.09 0.0078±0.094 means + S.D.) 66.35±20.92 37.77±17.91 24.22±16.28 0.18±0.21 Glaucoma suspects (n◦ 20; means + S.D) STATISTICAL DIFFERENCES p