Macular Pigment Density Is Reduced in Obese Subjects

Abstract

purpose. Because of the potential protective function of lutein (L) and zeaxanthin (Z) within the retina and lens, a better understanding of factors influencing tissue deposition is needed. The largest fractions of L and Z are stored in adipose tissue. Thus, higher body fat content and body mass index (BMI) may be expected to influence the quantities of L and Z in the retina (measured as macular pigment optical density, MPOD).

methods. Six hundred eighty subjects were tested. Information on MPOD, body mass index (BMI), body fat percentage (n = 400, using bioelectric impedance), dietary intake (n = 280, using a food frequency questionnaire), and serum carotenoid content (n= 280, using reversed phase high-performance liquid chromatography) was obtained.

results. There was an inverse relationship between MPOD and BMI (n= 680, r = −0.12, P < 0.0008) and between MPOD and body fat percentage (n = 400, r = −0.12, P < 0.01). These relationships were largely driven by data from the subjects with higher BMI (more than 29, 21% less MP) and higher body fat percentage (more than 27%, 16% less MP). Dietary carotenoid intake and serum carotenoid levels were also lower in subjects with higher BMI (n= 280).

conclusions. Obese subjects tend to have lower retinal L and Z. This reduction may be due to decreased dietary intake of L and Z and/or competition between retina and adipose tissue for uptake of L and Z.

Carotenoids are found in many tissues of the human body. For example, the dihydroxy-carotenoids lutein (L) and zeaxanthin (Z) are found in the liver, ovary, pancreas, kidney, spleen, testes, and adrenals ¹ and in many tissues of the eye (e.g., retina, lens, iris, choroid, RPE). ^{2 3} Although the functions of different carotenoids at the different sites have not been fully determined, some sites may store carotenoids for later use. For instance, it has been estimated that more than 80% of the total carotenoids in the body are found in adipose tissue, which could serve as a store. ⁴ Thus, variation in the quantity of body fat may be expected to influence carotenoid levels in serum and in less voluminous tissues that also take up carotenoids, such as the retina. A potential effect on serum carotenoids is consistent with the observation that anorexic persons, who have low body fat, have higher than normal levels of serum carotenoids, not directly related to carotenoid intake. ⁵ Nonetheless, the relationship between measures of body composition and carotenoid levels in the blood of normal persons has been inconsistent (e.g., body mass index[ BMI] ^{6 7 8} and body fat percentage ^{8 9 10 11} ). Furthermore, there are few data relating body composition to carotenoid concentrations in nonadipose tissue. 

In the present study, we tested the hypothesis that body composition is related to tissue concentrations of carotenoids in the central retina by comparing body fat percentage and BMI with macular pigment optical density (MPOD), which is a measure of L and Z in the retina. ¹² In our initial study, ¹³ we found no relationship between body composition and MPOD. This null finding, however, may have been due to a lack of statistical power because of the small sample size (N = 13). Thus, in the present study we extended the analysis to a larger sample (N = 680), tested at two different geographic locations, and found that individuals considered obese have lower MPOD. 

Methods

Subjects

A total of 682 subjects, 294 males and 388 females (mean age, 29.6 ± 13.1) were tested. Of those, 139 were current smokers, 109 were past smokers, and 434 reported that they had never smoked. Four hundred of the subjects were recruited from Athens, Georgia, and environs and were tested at the Vision Science Laboratory at the University of Georgia (300 of these subjects were undergraduates at the university). Two hundred eighty of the subjects were recruited from Indianapolis and bordering counties and tested at Indiana-Purdue University at Indianapolis, IN (see Ciulla et al. ¹⁴ for details regarding recruitment). The average BMI of the Georgia sample was similar (mean, 24 ± 4.9) to that of the sample tested in Indianapolis (mean, 26 ± 6.0). 

Two female subjects, one from each sample, with BMIs of 68 and 73 were excluded from further statistical analysis due to extreme values. These subjects had MPODs (460 nm, 1° test stimulus) of 0.0 and 0.08, respectively. The study followed the tenets of the Declaration of Helsinki. All procedures were approved by the institutional review board at the University of Georgia and the Indiana University School of Medicine, and informed consent was obtained from each subject before testing. 

Anthropometric Assessment

Information on BMI (body weight in kilograms divided by height in meters squared) was obtained from all subjects. Height and weight were determined at the time of the session using a calibrated, dual-reading, heavy-duty die-cast beam scale. Body fat percentage was assessed only in the subjects tested at the Georgia site. This assessment was performed using bioelectric impedance (Omiron analyzer, HBF 300; Carolina Biological Supply, Burlington, NC). Evidence has indicated that bioelectrical impedance analysis is reliable ¹⁵ and correlates highly with other methods of estimating body fat percentage, such as hydrodensitometry and skinfold calipers (Pearson’s r = 0.81–0.86). To assess the reliability of the specific instrument at the Georgia site, we measured four subjects (two males, two females) on 20 separate occasions over a period of 2 months, under various conditions (before and after meals, at various levels of hydration, for example). The average body fat percentage of each of the subjects was 8.0% ± 0.6%, 13.2% ± 0.41%, 15.4% ± 0.4%, and 21.0% ± 0.55% with an absolute range of approximately 2%, indicating that our measurement device was reliable. 

Dietary and Serum Carotenoid Analyses

Dietary and serum carotenoid data were collected only on the 280 subjects who were tested at the Indianapolis site. The equipment and procedures used in these assessments are detailed in Ciulla et al. ¹⁴ and are described briefly in this report. 

For dietary assessment, a 1-year food frequency questionnaire (FFQ) was used that assesses usual intake of 122 food items. ¹⁶ The analysis of these FFQs was performed at the Fred Hutchinson Cancer Research Center (Seattle, WA). The serum carotenoid analysis was conducted using a gradient reversed phase HPLC system with a photograph diode array detector set (HP 1100; Hewlett Packard, Burlington, MA) at 425 nm and a C18 guard column. ¹⁴  

Assessment of MPOD

MPOD was measured in the right eye only of each subject, by using a common psychophysical method based on flicker photometry. ¹⁷ This method is reliable, ¹⁷ correlates well with physical estimates of the pigments, ¹⁸ and yields optical densities that correspond to chemical concentrations of the pigments. ¹² The apparatus and stimuli used for the measurements are fully described and schematized in Wooten et al. ¹⁹ This same design and procedure was used at both the Georgia and the Indianapolis sites. 

MPOD was measured with a 1° test stimulus. Test stimuli were presented in natural view and near the center of a 6°, 10.5-cd/m², 470-nm circular background. The test stimulus was alternately composed of a 458-nm measuring field (peak MP absorbance) and a 570-nm, 16.7-cd/m² reference field (minimal MP absorbance). Light for the measuring and reference fields and the background was produced by 40-nm band-pass LEDs with peak energy at 458, 570, and 470 nm (Nichia Corp., Mountville, PA). This measurement was obtained in the fovea (where MP is the most dense) and 4° in the parafovea (where light absorption by MP is negligible). A tiny (5-minute) opaque fixation point was located on the left edge of the background, and subjects fixated this point when making the parafoveal measurement. The measuring and reference fields were superposed and presented out of phase at an alternation rate of 11 to 12 Hz in the foveal condition and 6 to 7 Hz in the parafoveal condition. Subjects adjusted the radiance of the 458-nm measuring field to achieve minimal flicker with the 570-nm reference. Subtracting the foveal from the parafoveal sensitivity measurement yields an optical density measure of MP. ¹⁷ Subjects were given brief instructions on the method and a practice trial before five foveal and five parafoveal measurements were made. The foveal and parafoveal values were calculated from the average of the final five readings, and these averages were then used to calculate MPOD. 

Results

As shown in Figure 1 , there was an inverse relationship between BMI and MPOD (n = 680, y = 0.31–0.003x, r =− 0.12, P < 0.0008). This relationship was essentially the same when the males and females were analyzed separately and was not influenced by smoking status. As is evident from the graph, the relationship was largely driven by those subjects whose BMI exceeded 29. These subjects had significantly (P < 0.001) lower average MPOD (mean, 0.19 ± 0.13 [SD]) than subjects with BMI under 29 (mean, 0.24 ± 0.13). When only subjects with a BMI under 29 were analyzed, there was no relationship between MPOD and BMI (r = 0.00). 

As shown in Figure 2a similar inverse relationship was found between MP and body fat percentage (n = 400, y = 0.29–0.002x, r = −0.12, P < 0.01). As with BMI, the inverse relationship between MPOD and body fat percentage seems largely driven by the subjects with the highest body fat percentage. When data from only those subjects with less than 27% body fat were analyzed (which is equivalent to a BMI of 29 for this sample, n = 310), there was no relationship between MPOD and body fat percentage (r = 0.03). MPOD was significantly (P < 0.0005) lower in subjects (n= 90, mean, 0.218 ± 0.13) with higher body fat (>27%) than in subjects with lower (<27%) body fat (average MP = 0.258± 0.13). Although the females had significantly (P < 0.0001) higher mean body fat than the males (females, 23.6%; males, 15.6%), the effects of high body fat on MPOD appeared to be similar. Males and females with high body fat (>27%) had 28% and 24% less MPOD, respectively, than did males and females with lower body fat. 

;T1>Dietary and serum data were also collected from those subjects who were tested in Indianapolis (Table 1) . As noted in Ciulla et al., ¹⁴ there were no sex differences in the relationship between serum L and Z and MPOD in this sample. When all the subjects were considered together, serum L and Z and dietary intake of L and Z were significantly correlated and significantly related to variations in MPOD (r = 0.21, P < 0.001 and r = 0.25, P < 0.001, respectively). As shown in Table 1 , however, dietary intake of L and Z and serum levels of total carotenoids were significantly lower in those subjects with a high BMI (>29), despite similar overall caloric intake. Serum levels of L and Z were slightly lower in subjects with a high BMI, but the difference was only marginally significant (P < 0.07). The relationship between MP and serum L and Z was similar in subjects whether BMI was high or low (>29, y = 0.095 + 0.26x, r = +0.34, P < 0.002; BMI <29, y = 0.16 + 0.15x, r = 0.19, P < 0.003). When one outlier is removed from the group with higher BMI, the relationship is improved (y = 0.07 + 0.33x, r = +0.40, P < 0.00001). 

Discussion

We found that subjects traditionally considered obese had lower MPOD than subjects of normal weight. For example, the MPOD of subjects with higher BMI (>29) was 21% less than subjects with lower BMI. A BMI higher than 28 to 30 is often used as a clinical standard for obesity. ^{20 21} Similarly, the MPOD of subjects with the highest body fat (>27%) was 16% less than that in subjects with lower body fat. BMI and body fat percentage were strongly related (R ² = 0.76 ²⁰ ), which probably accounts for the similarity in the results. 

There was no relationship between MP and adiposity when only subjects with a BMI below 29 and body fat below 27% were considered. Our analysis further suggests that the relationship between higher body fat and BMI and MPOD is not related to gender. Both males and females with higher body fat and BMI tended to have lower MPOD. Nonetheless, because females also tend to have higher average body fat, an effect of obesity on MPOD would affect a larger number of females. 

There are at least two possible nonexclusive explanations for our results. First, adipose tissue could compete with the retina for uptake of L and Z, resulting in less incorporation in the retina and lower MPOD. If competition takes place, then the effect is clearly not linear. An effect of adiposity is only seen when obese subjects are included in the sample. Moreover, the relationship between serum L and Z and MPOD was stronger in the subjects with higher BMI (>29). 

The second factor influencing the relationship between adiposity and MPOD is probably the subjects’ dietary patterns. Past studies have shown that both MPOD ^{22 23} and BMI ⁷ are related to dietary intake of L and Z. Higher body fat percentage has also been associated with poor dietary habits. ²⁴ Our present analysis also showed that those subjects with higher BMI had lower MPOD and decreased L and Z intake. Thus, a poor diet could promote both obesity and lower MPOD. 

Nonetheless, an analysis of the Indianapolis sample (see Table 1 ) suggests that the small differences in blood L and Z concentrations related to differences in diet are not sufficient to account for the 18% difference in MPOD found between subjects with low versus high BMI. For example, the regression line relating serum L and Z to MPOD predicts that the serum values would have to double to produce the 18% change seen in MPOD (see Fig. 3 in Ciulla et al. ¹⁴ ). Thus, some factor associated with higher adiposity (e.g., competition for L and Z uptake), in addition to diet, may have contributed to the observed MP differences. 

Past epidemiologic data have linked obesity to risk of age-related macular degeneration (AMD) ^{25 26} and age-related cataract. ²⁷ Recent studies suggest that reduced MP is associated with greater risk for AMD ²⁸ and cataract. ^{29 30} Thus, it is important to consider MPOD as one of multiple linked variables that may contribute to risk for eye disease in obese individuals. 

 Supported by Roche Vitamins Inc., Parsippany, New Jersey, and a faculty Research Grant from the University of Georgia. TAC is a recipient of a career development award from Research to Prevent Blindness, Inc.

 Submitted for publication June 1, 2001; revised August 27, 2001; accepted September 14, 2001.

 Commercial relationships policy: F.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

The relationship between BMI and MPOD (N = 680, y = 0.31 – 0.003x, r = −0.12, P < 0.0008).

Figure 2.

 

The relationship between body fat percentage and MPOD (n= 400, y = 0.29 – 0.002x, r = −0.12, P < 0.01).

Table 1.

 

Dietary, Serum, and MP Data According to BMI

Variable	BMI (range, 17.2–28.9; n = 211)	BMI (range, 29.1–68.1; n = 67)
Age	36 ± 7.7	36 ± 8.6
* BMI	23.7 ± 2.8	34.9 ± 6.6
Serum lutein and zeaxanthin (μmol/L)	0.38 ± 0.17	0.346 ± 0.158
* Serum beta-carotene (μmol/L)	0.304 ± 0.307	0.210 ± 0.200
* Serum carotenoid total (μmol/L)	1.55 ± 0.63	1.37 ± 0.51
Calories	2079 ± 1247	2107 ± 889
* Dietary lutein and zeaxanthin (μg/day)	1198 ± 904	957 ± 565
* Dietary BC (μg/day)	3138 ± 2977	2296 ± 1350
* Macular pigment OD	0.219 ± 0.135	0.18 ± 0.12

 Data are means ± SD from the sample tested at the Indianapolis site. ¹⁴ The dietary data of two subjects were not included in this analysis, because the data were considered too extreme (>10 SDs above the mean) to be plausible.

*  P < 0.05.