Relative Contributions of the Neurosensory Retina and Retinal Pigment Epithelium to Macular Hypofluorescence

Objective  To determine quantitatively the relative contributions of the neurosensory retina (NR) and retinal pigment epithelium (RPE) to the macular hypofluorescence observed during routine fundus fluorescein angiography.

Methods  Macular and peripheral buttons of neurosensory retina and retinal pigment epithelium were obtained from 10 postmortem human eyes. A well was created to simulate a fluorescein-filled choroid. The fluorescence of each tissue and combinations of tissue atop the well was determined using a fluorescence microscope. The percent reduction in the fluorescence of each, relative to the baseline fluorescence of the well alone, was calculated.

Results  Macular RPE demonstrated substantially lower fluorescence than peripheral RPE in all subjects. Macular NR demonstrated lower fluorescence than peripheral NR in all but one subject. The addition of macular NR to macular RPE caused significantly less fluorescence in all cases. Macular RPE caused a much greater percent reduction in fluorescence than macular NR in all subjects.

Conclusions  Hypofluorescence of the macula relative to the peripheral retina is a well-known feature of fluorescein angiography. This phenomenon is predominantly owing to the RPE and minimally to the NR, which cause 90.6 and 13.6 mean percent reductions in fluorescence, respectively.

HYPOFLUORESCENCE of the choroidal circulation in the macula relative to the peripheral retina is a well-known feature of fluorescein angiography. The anatomic basis for this phenomenon has been a source of controversy. Some authors have concluded that this phenomenon is primarily due to xanthophyll pigment in the neurosensory retina (NR) of the macula, which acts as an optical filter of the exciting blue light.^1,2 Most, however, attribute this phenomenon predominantly to the retinal pigment epithelium (RPE), which is more densely pigmented in the macula and believed to act as a barrier of choroidal fluorescence.^3-10 The absence of retinal vessels in the fovea (the foveal capillary-free zone) is believed by some to be a minor contributor as well.^4,6,7,9,10

These conclusions are based largely on clinical and histological observation. We are aware of only 2 studies^11,12 that actually addressed the issue of macular hypofluorescence in a quantitative manner. These, however, examined only spectral light transmission by the NR and RPE. They did not simulate the in vivo situation in which exciting blue light passes through the NR and RPE, sodium fluorescein is stimulated in the choroid, and emitted green light returns through the RPE and NR.

Using donated human tissue, we developed a model of clinical fundus fluorescein angiography to determine quantitatively the relative contributions of the NR and RPE to macular hypofluorescence. (For the purposes of this study, macula is defined as the central 1 mm², ie, a circle centered on the fovea with a radius of 0.564 mm.)

We selected 10 eyes from 10 deceased human donors (5 women, 5 men; 5 white, 5 of unknown race) ranging in age from 5 to 67 years. Donors with a history of or dissecting microscopic evidence for diabetic retinopathy, macular degeneration, vascular occlusion, or retinal detachment were excluded. The eyes were fixed in 4% paraformaldehyde for 1 to 2 days and then stored in 0.1-mol/L phosphate buffer at 4°C. (Attempts at dissecting RPE from fresh, unfixed eyes were unsuccessful owing to tissue fragility.)

After coronal transection at the level of the equator, the vitreous was removed from each globe using a cotton-tipped applicator and Westcott scissors. A 5.0-mm-diameter corneal trephine was used to cut 2 buttons of NR, RPE, and choroid—1 centered on the fovea and 1 positioned about 2 disc diameters nasal to the optic nerve. The NR was easily separated with jeweler's forceps, leaving behind the RPE and choroid. The RPE and choroid were removed from the globe as a unit by creating a flap of tissue and peeling it from the sclera. Care was taken to lyse the short posterior ciliary arteries.

The RPE/choroidal button was then glued to a glass coverslip, RPE-side down, with cyanoacrylate adhesive. The choroid was peeled away, leaving behind a monocellular layer of RPE (Figure 1 and Figure 2). The NR and RPE from each button were placed individually on glass slides with aqueous nonfluorescent mounting medium (Immuno-Mount; Shandon Inc, Pittsburgh, Pa).

A fluorescence microscope (Olympus BH-2; Olympus Corporation, Lake Success, NY) with a 455- to 490-nm bandpass excitation filter and a 515-nm low-pass (ie, passing longer wavelengths) barrier filter was used. A Sony 3-chip charge-coupled device (CCD) digital camera (Sony Medical Systems, Montvale, NJ) was employed to capture images at ×8.35 magnification.

To simulate the human choroid, a well was designed to serve as a fluorescein reservoir. The well consisted of a 2.0-cm rubber washer glued to a glass slide and covered with a glass coverslip. The well was filled with 0.1% sodium fluorescein (Figure 3). This arrangement of lighting, filters, tissue, and fluorescein is analogous to that used in clinical fundus fluorescein angiography, ie, both the exciting and fluorescing light each must pass through and be partially absorbed by the retina (NR and RPE) (Figure 4).

The emitted light from the well alone (including the filtering effect of an overlying glass slide, coverslip, and a thin layer of cyanoacrylate adhesive) was measured with a broad-range, photographic light meter placed in the trinocular tube of the fluorescence photomicroscope. (The digital camera was removed from the trinocular tube during these measurements.) The sensitivity range of the digital camera was limited; consequently, it was necessary to employ 2 levels of exciting light (produced by using neutral density filters): a high intensity for measuring the RPE and a low intensity for the NR. The emitted light levels of the well alone (no tissue—only a slide, glue, mounting medium, and a coverslip) were 433.4 and 33.4 lux, respectively, at these 2 exciting intensities.

Fluorescence images of each tissue type (macular RPE, macular NR, peripheral RPE, peripheral NR, and selected combinations) were obtained. The average gray-scale luminance value of the central 1.0 mm² of each tissue and combinations of tissues was determined using Optimas image analysis software (Media Cybernetics LP, Bothell, Wash) (Figure 5).

A calibration curve was generated using light within the sensitivity range of the digital camera so that its arbitrary luminance values could be converted to lux. Subsequently, the filtering effect, as a percent reduction of the emitted light, by any piece or combination of tissues could be determined. Since the measured values of luminance were dependent on the strength of the light source, they had only relative significance. The results were therefore expressed as a percent reduction of light reaching the digital camera owing to a specific tissue or tissues compared with the well alone.

The macular RPE showed significantly greater reduction in fluorescence than peripheral RPE (Figure 6 and Table 1). Likewise, the macular NR reduced fluorescence more than peripheral NR (Figure 7 and Table 1). Adding macular NR to macular RPE, thus simulating the natural superposition of the 2 tissues, resulted in only a 3.l % further reduction in fluorescence compared with macular RPE alone (Figure 8 and Table 1). Finally, when macular RPE and macular NR were compared, the former had a markedly greater effect on fluorescence than the latter (90.6% vs 13.7%) (Figure 9 and Table 1).

Note that filters combine in a multiplicative (not additive) manner. Therefore, a priori, the combined effect of RPE and NR should be the light transmitted by the RPE (1 − 0.906 = 0.094) times the light transmitted by the NR (1 − 0.136 = 0.864), which equals 0.0812. So 8.12% of the light should be transmitted by both tissues together. Thus, the percent reduction in fluorescence compared with no tissue should be 100 − 8.12 = 91.9%, which agrees, within experimental error, with the measured value of 93.7%. No clearly defined age-related effects were evident (Figure 6, Figure 7, Figure 8, and Figure 9).

Our model for testing the filtering effects of retinal tissue shows significantly greater reduction of fluorescence by both the macular RPE and the macular NR compared with their peripheral counterparts. These data also indicate, regardless of location, a markedly greater reducing effect produced by the RPE than by the NR. In the macula, the RPE causes a 90.6 mean percent reduction in fluorescence, while the NR produces only a 13.6 mean percent reduction. Together the RPE and NR are responsible for a 93.7 mean percent reduction in fluorescence. Combining these 2 tissues reduces choroidal fluorescence by only 3.1% more, compared with the reduction seen by the RPE alone. We therefore conclude that the RPE is responsible for most of the macular hypofluorescence so commonly seen with fluorescein angiography. This study did not address the effects of the lack of retinal vascular fluorescence seen in vivo. We would expect this to contribute to macular hypofluorescence as well.

The choroid has melanocytes that partially block fluorescence from the deeper vessels. In principal, this could affect the relative fluorescence of the macula compared with its surrounds if the pigmentation were greater centrally. However, we are unaware of major differences in choroidal pigmentation between these 2 areas. As with any in vitro experiment, we could not directly assess if there were significant postmortem changes in the tissues with respect to their optical filtering characteristics. Nevertheless, our eyes had a range of times between death and fixation of from 4 to 23 hours. We did not see a trend that would suggest a functional relationship between prefixation time (ie, autolysis) and filtering, at least for these intervals.

Although not a direct part of this study, it is interesting to speculate on the specific origin of this hypofluorescence due to the RPE. The 2 most likely candidate pigments are lipofuscin and melanin. Lipofuscin accumulates throughout life and is probably a by-product of photoreceptor metabolism. Either it, or a similar material,¹³ can produce a strongly hypofluorescent effect as evidenced by the "silent choroid" seen in fundus flavimaculatus (or Stargardt disease).^14-16

It is unlikely, however, that lipofuscin is responsible for macular hypofluorescence in normal eyes for at least 2 reasons. First, this pigment accumulates with age. Accordingly, we should expect to see a greater reduction in fluorescence with increasing age, which was not the case in our study (Figure 6, Figure 7, Figure 8, and Figure 9). Second, the distribution of lipofuscin is not consistent with the observed pattern of hypofluorescence. Although it is more concentrated posteriorly, it reaches its greatest concentration at about 15 arc degrees of eccentricity to the fovea.¹⁷ There is actually a relative decrease in lipofuscin in the central macula, where hypofluorescence is greatest.

We believe melanin pigment is a better candidate for the decline in fluorescence due to the RPE. It has been shown that melanin has a high extinction coefficient in vitro—especially for the short wavelengths of visible light.^2,12 Furthermore, the density of melanin pigmentation in the RPE is greatest in the macula and centered on the fovea.^18-20 Finally, the RPE is completely melanized at birth with no further pigment granules formed thereafter,^19,21 which is consistent with our finding of no increase in the reduction of hypofluorescence with age.

The RPE's predominant contribution to macular hypofluorescence can be appreciated clinically in the case of macular dragging due to an epiretinal membrane. Here, the foveal avascular zone is no longer centered on the zone of hypofluorescence (Figure 10).

Accepted for publication June 29, 1999.

This study was supported in part by grant EY08724 from the National Eye Institute, National Institutes of Health, Bethesda, Md (Dr Nork); an unrestricted gift from Research to Prevent Blindness Inc, New York, NY; the Wisconsin Lions Eye Research Fund, Stevens Point; and the Steve J. Miller Foundation, Marshfield, Wis.

We thank James N. Ver Hoeve, PhD, for his assistance with the statistical analysis.

Reprints: T. Michael Nork, MD, MS, Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School, 600 Highland Ave, Dept F4/336, Madison, WI 53792-3220 (e-mail: tmnork@facstaff.wisc.edu).