Abstract
purpose. Macular pigment (MP) may protect against age-related macular
degeneration. This study was conducted to determine the extent of
changes in the macular pigment density as a consequence of oral
supplementation with lutein. A second purpose was to compare two
objective measurement techniques.
methods. In the first technique, reflectance maps were made with a scanning
laser ophthalmoscope. Digital subtraction of log reflectance maps and
comparison between the foveal area and a 14° temporal site provided
MP density estimates. In the second technique, spectral fundus
reflectance of the fovea was measured with a fundus reflectometer and
analyzed with a detailed optical model, to arrive at MP density values.
Eight subjects participated in this study. They took 10 mg lutein per
day for 12 weeks. Plasma lutein concentration was measured at 4-week
intervals.
results. After 4 weeks, mean blood level of lutein had increased from 0.18 to
0.90 μM. It stayed at this level throughout the intake period and
declined to 0.28 μM 4 weeks after termination. Measurement of the
density of MP showed a within-subject variation of 10% with MP maps
and 17% with spectral reflectance analysis. MP density showed a mean
linear 4-week increase of 5.3% (P < 0.001) and 4.1%
(P = 0.022), respectively.
conclusions. Supplementation with lutein significantly increased the density of the
MP. Analyzing reflectance maps with a scanning laser ophthalmoscope
provided very reliable estimates of MP.
Macular pigment (MP), concentrated in the central area of the
retina, contains the carotenoids lutein and zeaxanthin. It protects the
macular region by its capability of filtering blue light, thereby
possibly decreasing photochemical light damage.
1 In
addition, MP is capable of scavenging free radicals.
2 Cross-sectional studies have observed an inverse association between a
diet with a high content of the carotenoids lutein and zeaxanthin and
the prevalence of age-related macular degeneration
(AMD).
3 4 Although for a definite proof of this
relationship a follow-up study is needed, it is of interest to know
whether MP density can be modified. Two studies have been published
showing a change in MP density, one by a dietary modification
(consumption of spinach and/or corn) on 11 subjects
5 and
one by supplementing lutein in two subjects.
6 Hammond et
al.
5 had some subjects fail to show a change in MP
density. Both studies used heterochromatic flicker photometry to
determine MP density in experienced subjects. This psychophysical
method has disadvantages. It is time consuming, and its reliability
depends on the understanding by the subject of the task involved. This
makes the test less suitable for a naive target population. Recently,
Bernstein et al.
7 demonstrated in an animal model
resonance Raman scattering as an objective method that may have future
use in studying human MP density. The purpose of the present study was
twofold. First, we wanted to establish the extent of changes in the MP
density as a consequence of oral supplementation with lutein. Second,
we wanted to compare two objective techniques for measuring MP on the
basis of fundus reflectance.
Eight male, nonsmoking volunteers between the ages of 18 and 50
years (mean age, 40.6 years) were recruited for the study. To assure
homogeneity in lutein plasma levels at baseline, a simple
food-frequency questionnaire was used to exclude subjects with a high
intake of lutein–that is, subjects who had eaten vegetables high in
lutein, such as spinach and kale, more than four times in the 4-week
period preceding the beginning of the study. Furthermore, only subjects
with a relatively normal (Dutch) dietary pattern and not taking vitamin
or mineral supplements were selected. The subjects took a daily dose of
10 mg of lutein, in the form of lutein diesters derived from marigolds,
for a period of 12 weeks. The study was conducted according to good
clinical practice, approved by the local medical ethics committee, and
was conducted in accordance with the tenets of the Declaration of
Helsinki. Informed consent was obtained from all subjects before
participation in the study.
Blood was sampled after an overnight fast at baseline and at
4-week intervals during the study. Four weeks after the end of the
lutein supplementation period, a final blood sample was taken. For
lutein analyses 1 ml plasma was mixed with 1 ml ethanol (containing 16
to 32 micromoles tocopheryl acetate per liter as internal standard).
After 10 minutes, 2 ml hexane was added, and the sealed tubes were
vortexed for 4 minutes. After centrifugation for 10 minutes at
3000g at 4°C, the hexane layer was separated and
evaporated under nitrogen at room temperature. The residue was
dissolved in 0.4 ml high-performance liquid chromatography (HPLC)
solvent and transferred into brown HPLC injection vials. Lutein was
quantified by HPLC with a hyperchrome 3-μm column Nucleosil
120. The mobile phase consisted of acetonitrile-methylene
chloride-methanol (70:15:10, vol/vol/vol), and the flow rate was 1
ml/min. An absorbance detector was used at 445 nm for detection of
lutein. Limit of detection of lutein was 3.0 nM plasma. The amount of
lutein present in the plasma sample was quantified by calculating the
ratio of the peak height of lutein to that of the internal standard.
MP density was measured in both eyes at baseline and at 4-week
intervals during the study. Four weeks after the end of the lutein
supplementation period, a final measurement was taken. Two techniques
were used: MP maps from scanning laser ophthalmoscopy (SLO) and
spectral analysis. The methods were always applied successively in the
order of SLO first, spectral analysis second. A mydriatic was used to
dilate the pupil for both setups.
MP Maps from SLO.
Fundus reflectance maps at 488- and 514-nm argon laser wavelengths were
made with a custom-built SLO
(Fig. 1) . The SLO covers a retinal area of 40° × 23°, has a well-defined
exit pupil 2 mm in diameter, and allows reflectance maps at different
wavelengths to be grabbed within a few video frames. Blood and melanin
effectively absorb light that has entered the choroid, and the major
contribution to the reflectance is from the discs in the outer segments
of the cones.
8 This leaves the lens and the MP as
the only relevant absorbers in this wavelength region. As a
consequence, digital subtraction of log reflectance provides density
maps of the sum of both absorbers.
Figure 1C shows a typical example,
calculated from the reflectance maps shown in
Figures 1A and 1B . At
14°, temporal MP density is assumed to be negligible. Thus, with this
site providing an estimate for the lens density, the mean MP density
was calculated in a 1.5° field centered at the fovea.
9 The densities were corrected for the slightly lower difference in the
MP absorbance spectrum at 488 and 514 nm, compared with the peak at 460
nm and null at λ > 540 nm as in a standard MP
spectrum.
10 To avoid possible influence, visual pigments
were bleached with a 6-log-troland (Td) bleaching light (96% bleach)
before reflectance maps were made. A bite board and temple pads were
used to maintain head position.
Spectral Analysis.
Spectral fundus reflectance was measured with the Utrecht
retinal densitometer.
8 Briefly, a rotating wheel (14
revolutions per second) offers a sequence of 14 interference filters in
the range 430 to 740 nm to enable quasisimultaneous measurement of the
reflectance across the visual spectrum. The illumination field was
1.8° centered at the fovea. Light reflected from the fundus was
measured in a detection field of 1.5°, concentric within the
illumination field. To obtain an estimate of the mean MP density in
this area, spectral fundus reflectance was measured in two conditions:
perpendicular, with the instrument’s entry and exit pupils aligned to
the peak of the Stiles–Crawford (SC) function, and oblique, 2 mm
temporal to the SC peak.
11 A detailed optical model of
foveal reflection was used to arrive at individual estimates of
parameters as equivalent thickness of blood layer, and densities of the
lens, MP, and melanin.
11 In short, the incoming light is
assumed to reflect, at the inner limiting membrane (ILM), discs in the
outer segments of the photoreceptors and the sclera. Using known
spectral characteristics of the different absorbers within the eye
(lens, MP, blood, and melanin), the densities of the pigments and the
percentage of reflectance at the interfaces are optimized, to fit the
measured data at all wavelengths. The two spectra measurements,
perpendicular and oblique, were fitted simultaneously with eight free
parameters. None of the parameters was assumed unchanged between
baseline and supplemented status. Only MP showed significant changes in
time and will be discussed. Visual pigments were bleached, and a bite
board and temple pads were used to maintain head position.
To quantify the quality of the measurement techniques, we compared
the within-subject variations. Both relative SD and coefficient of
repeatability were calculated.
12 To estimate the possible
increase in MP density over time, we applied a statistical general
linear model (GLM), with repeated-measurements analyses on MP density
with both time and eye as within-subject factors. Time was included in
the model as linear effect. For one of our subjects, one of the fundus
reflectance measurements failed in one eye. In the GLM analysis, we
used the mean of his other four MP densities (in the same eye) for this
data point. For clarity, in
Figures 2 and 4 , mean values of left and
right eyes are presented.
Individual response curves of plasma lutein concentration are
shown in
Figure 2A . All subjects showed a substantial increase in plasma lutein
concentration at week 4. Mean lutein concentration, shown in
Figure 2B ,
increased from 0.18 ± 0.08 μM at baseline to 0.90 ± 0.18μ
M (
P < 0.001) at 4 weeks, and it remained at this
level throughout the supplementation period. Four weeks after
termination, the lutein level was still elevated (0.28 ± 0.06μ
M;
P = 0.005). For comparison, literature data are
included in
Figure 2B (see the Discussion section).
Figures 3A and 3B show individual response curves for the MP density with both
techniques. Baseline MP density values showed a large variation between
subjects. In the majority of the measurements, MP density showed an
increase with time. For the relative SD there was a within-subject
variation of 10% with MP maps and 17% with spectral analysis. The
coefficients of repeatability were 0.17 and 0.27, respectively.
To emphasize changes in density as a result of lutein supplementation,
we normalized the individual response curves for each subject by
dividing each density by the mean of the five measurements
(Figs. 3C 3D) . Statistical analysis yielded a linear 4-week increase in relative
MP density of 5.3% (
P < 0.001), calculated from MP
maps, and of 4.1% (
P = 0.022), obtained with the
spectral analysis. Mean response curves for both techniques are
compared in
Figure 4 , together with literature data.
At baseline, plasma lutein concentration showed a significant
correlation with MP density,
r = 0.78
(
P < 0.001) determined with MP maps and
r = 0.82 (
P < 0.001) determined with
spectral analysis, respectively
(Fig. 5) .
A daily dose of 10 mg lutein supplementation induced an increase
in mean plasma lutein by a factor of 5 and a linear 4-week increase in
relative MP density of 4% to 5%. To our knowledge, this is the first
study in which the effects of intake of lutein have been assessed with
objective measurement techniques. In particular, the SLO-based
technique provided very reliable results. With this technique all
subjects showed a significant increase in MP density. The spectral
reflectance analysis provided noisier results that did not allow such a
conclusion for all individuals.
Figure 4 shows a comparison between our
results on naive subjects and heterochromatic flicker photometry data
from literature. Landrum et al.
6 published results
of a study of 20 weeks of supplementation of 30 mg lutein per day
derived from marigolds in two skilled subjects who were measured four
to five times week. The noise in their results seems comparable to that
in our technique. They found a 4-week increase of 4.2% in relative MP
density, which also compares well with our results. Hammond et
al.
5 presented results of a dietary intake study for 15
weeks in 11 subjects who consumed spinach and/or corn (10.8 mg of
lutein). Eight subjects showed an increase in MP density, but two
showed a slight decrease. We estimated a 4-week increase in relative
density of 3.5% for all subjects in their study. Their mean results
suggest a considerably higher noise level than that in the other
studies, partly due to a temporary return to baseline MP density at 8
weeks, which they attribute to interaction with other tissues that
accumulate lutein. This was neither observed in this study, nor in
Landrum et al. It could be due to the use by Hammond et al. of dietary
intake of lutein modification (consumption of spinach and/or corn),
compared with supplementing lutein in the form of lutein diesters
derived from marigolds in this study and the study by Landrum et al.
Hammond et al. also studied both men and women, whereas in this study
and that of Landrum et al., only men were involved. In women, there is
no significant correlation between MP density and lutein in the diet,
and the correlation between MP density and lutein in the blood is much
weaker than in men.
13 A division between responders and
nonresponders, as proposed by Hammond et al., also could be seen in the
spectral analysis of the fundus reflectance presented in the current
study. We wonder whether this may be an artifact of the measurement
techniques. With the SLO-based technique, all subjects showed a
significant increase.
Heterochromatic flicker photometry is a rather demanding technique for
subjects—in particular the task of adjusting flicker at a peripheral
location. This may be the cause for noisier results than those obtained
with the present SLO technique. A significant improvement in
heterochromatic flicker photometry, however, was recently described in
a psychophysical setup that avoids Maxwellian view.
14
In Landrum et al.
6 lutein intake was three times as high
and plasma lutein concentration was twice as high as in the present
study
(Fig. 2) . However, their increase in MP density was similar to
ours. Apparently, a dose of 10 mg per day is sufficient to provide a
4-week increase of 4% to 5%. This conclusion is supported by the
still-elevated plasma lutein concentration a month after the end of
supplementation, compared with baseline, accompanied with a
still-increasing MP density between weeks 12 and 16 (
P = 0.03 for the MP maps,
P = 0.42 for the reflectance
analysis). Thus, a high plasma lutein level seems to elevate MP density
gradually over time. This is further corroborated by the high
correlation between plasma lutein concentration and MP density at
baseline in our (all male) subject group
(Fig. 5) . Considering the
study design, we cannot exclude that this effect was due to factors
other than plasma lutein alone. For a definitive answer, a double-blind
randomized controlled trial is needed.
Because we obtained maps of MP distribution we were able to look for
changes in MP distribution with lutein supplementation. An exponential
decay of MP density as a function of eccentricity fitted our data well
up to 4°. No changes were found between the different measurements in
time.
Our results show that fundus reflectance, in particular as obtained
with our custom-built SLO, can be used as a fast and objective test to
obtain reliable estimates of MP density. To reduce the effect of
differences in pupil size, our SLO has a well-defined exit pupil,
whereas commercially available SLOs use the whole pupil plane. Further,
to minimize the influence of head and eye movements on the reflectance
maps, our SLO allows reflectance maps at different wavelengths to be
grabbed within a few video frames, which may be difficult in other
SLOs. However, both adaptations may be of minor influence, because MP
density is determined by a relative comparison of two reflectance maps.
We opted for using a bite board, which is costly and time consuming. It
may be possible to avoid using it in larger scale studies. The Utrecht
retinal densitometer uses a spot with a rather low intensity to measure
spectral fundus reflectance, because it has been optimized for retinal
densitometry. Therefore, to obtain an adequate signal-to-noise ratio,
densitometer outputs were averaged over a 2-minute interval.
Nevertheless, one subject, the fourth in
Figure 5 , had such low fundus
reflectance in the bluish wavelength region, that the MP density,
determined by analyzing this wavelength region, showed an SD of 0.40 OD
and 0.28 OS. All others showed a mean SD of 0.12. This may explain the
discrepancy for this particular subject between the results obtained
with the MP maps and with the spectral analysis. Severe cataract lowers
the intensity and gives rise to similar problems. Increasing the
intensity of the measuring light may shorten the time interval for data
acquisition and improve the performance of the reflectance analysis. At
baseline, mean values for MP density (single pass) were 0.26 for the
SLO technique and 0.47 for the reflectance analysis. Reflectance at the
ILM, anterior to the MP, may lower the apparent MP density measured.
This ILM reflectance is corrected for in the model for spectral
analysis. Other sources of light scatter, such as floaters, could also
introduce an underestimate of MP density, both in the SLO technique and
spectral analysis. However, influence of light scatter is minimized by
the confocal optics and by the spatial separation of entrance and exit.
Determination of MP density by either comparing foveal and peripheral
reflectance with the SLO or by analyzing the spectral content of the
reflectance assumes a retinal structure as in normal subjects. It will
be of interest to measure MP density in various stages of AMD. Drusen,
present in early stages of AMD, act as a neutral reflector at the level
of the retinal pigment epithelium. Their presence is not accounted for
in the model. However, their reflectance shows up as an enhancement of
the reflectance at the discs in the outer segments of the
photoreceptors, which does not affect the determination of MP density.
Only in late stages of AMD, in case of neovascularization or atrophy,
would the methods fail.
The authors thank Hélène G. de Vries,
Wilfred Sieling, Itta M. Minderhout and Anneloes L. Berends
for their assistance and the Henkel Nutrition and Health Group for
providing the lutein.