Abstract
Purpose:
More than a dozen studies have investigated whether blue-light filtering (BLF) intraocular lens (IOL) implants influence color vision, generally finding they do not. These studies have not tested color vision per se; rather, they have measured color vision deficiencies or chromatic discrimination. Here, we used additive trichromatic colorimetry to assess color appearance in participants with BLF and clear IOL.
Methods:
Seventy-six participants were recruited from two populations: older participants (n = 52) with BLF and clear IOL (n = 98 eyes; M = 67.33 ± 7.48 years; 58.8% female; 25.5% non-White), and young adult control participants (n = 24; M = 21.0 ± 5.13 years; 70.8% female; 41.5% non-White). Participants used a custom-built tricolorimeter to mix three primaries until a perceived perfect neutral white was achieved. Color appearance, expressed as chromaticity coordinates, was measured with a spectral radiometer (ILS950).
Results:
Between subjects, the BLF IOL chromaticity coordinates (x = 0.34, y = 0.35, u′ = 0.21, v′ = 0.48) were not significantly different from the clear IOL (x = 0.34, y = 0.33, u′ = 0.22, v′ = 0.48). BLF and clear IOL were also not different within-contralateral subjects (n = 21; BLF x = 0.34, y = 0.33, u′ = 0.22, v′ = 0.47; clear x = 0.34, y = 0.33, u′ = 0.21, v′ = 0.48). Both IOL groups differed from young adults (v′[0.45; P = 0.001], x[0.31; P = 0.008], and y[ 0.30, P < 0.000], but not u′[0.21]).
Conclusions:
One advantage of geometric representation of color space is the ability to specify the appearance (rather than spectral composition) of any light mixture by specific coordinates. Using this system, only minor differences in color appearance were found between a BLF, clear IOL, and young natural lens.
Translational Relevance:
When color perception is directly measured, the BLF and clear IOL are not meaningfully different.
In 1994, Hoya Surgical Optics, based in Singapore, was the first to introduce a yellow-tinted polymethyl methacrylate intraocular lens (IOL) implant. This was followed (circa 2000) by Alcon (Alcon Laboratories, Fort Worth, TX) , who also developed a blue-light filtering (BLF) IOL that was promoted on a more global scale. These lenses were originally developed based on a simple premise: existing IOLs did not resemble the healthy adult version of the lens they were replacing. Older, more vulnerable eyes, are actually somewhat protected by (or, at least, are adapted to) a more yellowed lens.
1 Hence, it seemed logical that implanting a lens tinted to match a normal adult lens represented a more natural prosthesis. Indeed, there were numerous reports around that time that clear IOLs induced significant perceptual changes to color vision.
2,3 For example, in an article entitled “Colors do look different after a lens implant!”
4 the authors noted:
I received a [clear] implant in my right eye, which I hereafter will call the ‘new eye’. The left, or ‘old eye’, did not receive an implant for a year later. What I was unprepared for were the differences in color and appearance of familiar objects between the two eyes.
With the introduction of naturally tinted IOLs, however, the controversy soon shifted to whether actually adding yellow tinting itself introduced significant changes in color perception. A “change,” of course, is only meaningful relative to its comparison. Cataract, for example, likely changes the perception of short-wave light relative to a younger more transparent lens.
5 The controversy surrounding BLF IOLs was unusual in that medical treatments are often created in an attempt to mimic the natural state as closely as possible. Most BLF IOLs are designed to mimic the lens absorbance of a 30-year-old.
6 This practice means, of course, that even a BLF IOL represents significantly more short-wave light reaching the older retina when compared with a cataractous lens. Simunovic et al.
7 expressed the worry that short-wavelength light absorbing IOLs would result in changed color perception as follows:
Because [short-wavelength light absorbing] IOLs influence the spectral quality of light incident on the retina, one of the anticipated deleterious effects of such lenses is on color vision. Compared with conventional UV-absorbing IOLs, [short-wavelength light absorbing] IOLs would be anticipated to effectively decrease the chromaticity difference between warm and cool colors (ie, they should induce a tritan color vision deficiency).
Why the tinting of a BLF IOL would have “deleterious effects,” but the normal yellow of the adult crystalline lens would not, is unclear. Nonetheless, this type of “back and forth controversy”
8 inspired a wave of studies that assessed whether yellow IOLs (or BLF filters, generally) had negative effects on color vision. These studies
9–26 are shown in the
Supplementary Data. All of these studies, unfortunately, suffer at least one and often two major limitations. The first is that none of them actually measured color perception per se. The second is that most of them used a clear implanted lens as their normal control (again, a completely transparent crystalline lens is not “normal” in an adult eye).
With respect to the first issue, most studies evaluating color differences used measures based on discriminable differences based on metameric color space, chromatic discrimination, or color vision deficiencies. These studies have shown that BLF IOLs (mostly relative to clear IOLs) do not induce clinical deficits (e.g., analogous to missing a cone type or having anomalous opsins). Subjects with BLF or clear IOLs can also generally perform metameric matching. That is somewhat different, however, to saying that colors appear the same to them as they do to adults with an intact natural lens. For example, one cannot infer appearance (that a particular light looks “vivid blue”) from the color equations of a metameric match or an ordering of chromatic plates.
There are well-validated methods for measuring the actual appearance of colors. One well-studied method is by using the achromatic or white point
27,28; that is to say, where the stimuli seem to be totally devoid of chromatic color. The advantage of using the white point is that the zone of normal is very well-documented
29 and can be quantitatively expressed as CIE chromaticity coordinates (
x and
y values). The interpretation is also quite straightforward: the subject's white point is based on their response regarding the appearance of the light (i.e., it looks white and has no chromatic tint); physiologically, that white point is only achieved when the chromatic systems are balanced.
In this study, we used a case control design to assess the effects of a BLF IOL on color appearance. Subjects with the BLF IOL (cases) were compared against two different controls: older subjects with a clear IOL and younger subjects with their natural crystalline lenses. Color appearance was measured using additive trichromatic colorimetry.
A custom-designed tricolorimeter was constructed to determine and specify the locus of perceptual white within the CIE chromaticiy diagram, as schematized in
Figure 1. The optical system was built around two integrating spheres, 1 and 2 as shown in
Figure 1. They were constructed from aluminum hemispheres painted on the hollow side with white paint, which is nonluminescent, highly diffusing, and 98% reflecting (Labsphere, North Sutton, NH). Each hemisphere was drilled for two apertures (1-inch diameter, labelled A1-A4). The finished hemispheres were joined with a rigid adhesive to form the two spheres.
The light source was a 1-inch diameter, chip-on-board array of cool white LEDs (6500° color temperature). The array was located at A1 in Sphere 1. The dashed line, originating at the center of the array, traces the path of the principle ray as it traverses through the entire system ultimately entering the eye. From A1, the principle ray projects to the opposite side of the sphere where it is reflected and diffused in all directions. At that point one can follow a principle ray projecting at a 45° angle that impinges at the point that is directly opposite the center of A2. Again, light is diffusively reflected at all angles from every impingement ad infinitum. Thus, the exit port, A2, becomes a Lambertian emitter where the luminance toward an observer is independent of the viewing position; the perception is one devoid of all texture and perfectly uniform (appearing like a disembodied light).
As shown in the
Figure 1, the principle ray passes through the center of A2 toward the center of A3 passing into sphere 2. Very near A3 is a filter assembly composed of a blue (B), a green (G), and a red (R) Wratten filter (the important characteristic of each is simply that it provide a highly saturated primary color, in our system the λmax was 447, 543, 615 nm; Wratten filters, #47, #40, #26, respectively; Edmund Optics, Barrington, NJ). The assembly is mounted onto a vertical/horizontal micrometer stage allowing various proportions of the RGB filters to be sampled. A wide ratio of the three colored filters can be set to sample a large subset of the CIE chromaticity diagram ranging from clearly red, or green, or blue to a perfect white, and all shades in between. Several opaque shields were positioned such that the cone of light from A2 was blocked except for the rays entering A3, preventing crosstalk between the various components. After passing through the RGB filter the principle ray impinges on the surface opposite from A3 and follows a path similar to sphere 1, the light exiting at A4. Sphere 2 serves to additively mix the R, G, and B components thoroughly so that the emitted light from A4 is Lambertian as for A2 and color appearance is constant across the perceived target for a given R, G, B setting.
After the principle ray passes through the center of A4, it is transmitted through a lens (L1) and then passes into a beam splitter where half the light is reflected onto a lens (L2) and positioned on the detector of a spectral radiometer, which calculates the chromaticity coordinates. The other one-half of the light is directed through an eyecup and into the eye. L1 is initially positioned one focal length (4 inches) from A4. Thus, the rays emerging from A4 are collimated before passing into the eye where they are imaged on the retina, as shown in the diagram. For an emmetrope A4 would be in sharp focus. But for a myope or a hyperope the image would be in front of the retina or behind the retina, respectively. Perfect focus of the image for any observer was achieved by simply increasing or decreasing the distance between A4 and L1 (one example of a class of telecentric lens assemblies). The primary advantage over other focusing procedures is that the magnification of the image is constant in size no matter the distance between A4 and L1. For our application, the eye cup must be fixed, because it is the reference point for eye position. Therefore, we mounted the Spheres on a platform, which can be translated along the z-axis. The observer varies position by turning a dial until perfect focus is achieved.
From each of these three starting points, the experimenter systematically adjusted one axis (e.g., red–green) at a time and instructed the subject to state when the visual field was as close to white as possible along that axis, at which point the experimenter stopped turning that dial. Then, the experimenter moved to the other axis (red–blue in this example) and adjusted that dial until the visual field was as close to white as possible. This continued, back and forth, with small, fine-tuned adjustments, until the subject reported that the visual field was pure white without any tint of color.
Once an approximate white setting was obtained, subjects were asked to look away from the device for approximately 5 seconds, and then back into the eye piece to double check that the visual field did not contain any tint of color at that setting, by asking questions such as, “If you had to call what you are seeing a color other than white, what would you call it?” Also, using that setting, the experimenter turned each knob to bracket the area (meaning, at what point did the subject first perceive a tint). Four trials, one from each starting point, were collected for each test eye (the order of testing was randomly varied). For the purpose of analysis, we recorded the CIE color space coordinates,
x and
y, and u′, v′ (based on the 1976 update to the CIE diagram; the more perceptually Uniform Color Space; for a review see Schanda [2007]
30). We also recorded total illuminance (lux) of the stimulus, and total irradiance (uW/cm
2).
The authors thank Lauren Hacker for her support with medical records review and recruitment, and Colin Gardner for support with data collection.
Supported by an investigator-initiated grant from Alcon Laboratories. Alcon Laboratories had no role in the design or conduct of this research.
Disclosure: B.R. Hammond, Jr, Alcon (C); B.R. Wooten, None; S.E. Saint, None; L. Renzi-Hammond, None