**3. Results**

#### *3.1. Demographics and Baseline Characteristics*

A total of 702 participants had MPOD measurements among the 1153 from the Montrachet study (Figure 1).

**Figure 1.** Eyes (553 eyes without optic neuropathy and 48 eyes with POAG). The baseline demographics and clinical characteristics of participants and non-participants are shown in Table 1.

**Table 1.** Baseline characteristics between participants and non-participants in the Montrachet study.



**Table 1.** *Cont.*

*p*-value was calculated between participants and non-participants. The results are displayed as *n* (%) for categorical variables and as mean ± standard deviation or as median (interquartile range (IQR)) depending on their distribution for continuous variables. AMD = age-related macular degeneration; L = lutein; Z = zeaxanthin; PUFA = polyunsaturated fatty acids; ALA = alpha-linoleic acid; EPA = eicosapentaenoic acid; DHA = docosahexaenoic acid; LDL = low-density lipoprotein; and HDL = high-density lipoprotein. Missing data for smoking status (current or past, *n* = 20), alcohol consumption (current or past *n* = 136), central retinal thickness (*n* = 8), cup-to-disc ratio (*n* = 105), spherical equivalent (*n* = 103), lens status (*n* = 3), sun protection (*n* = 5), AMD stages (*n* = 85), L/Z (*n* = 358), plasma n-3 PUFAs (*n* = 339), LDL cholesterol (*n* = 14), HDL cholesterol (*n* = 14), and triglycerides (*n* = 15).

Non-participants were more likely to be phakic and to present late AMD. Participants' mean age was 82.11 ± 3.70 years.

The characteristics of the study population are displayed in Table 2.

**Table 2.** Baseline characteristics of eyes without optic neuropathy and eyes with primary open-angle glaucoma in the Montrachet study, *n* = 601.



**Table 2.** *Cont.*

*p*-value was calculated between participants and non-participants. The results are displayed as *n* (%) for categorical variables and as mean ± standard deviation or as median (interquartile range (IQR)) depending on their distribution for continuous variables. POAG = primary open-angle glaucoma; RNFL = retinal nerve fiber layer; AMD = age-related macular degeneration; L = lutein; Z = zeaxanthin; NA = not applicable; PUFA = polyunsaturated fatty acids; ALA = alpha-linoleic acid; EPA = eicosapentaenoic acid; DHA = docosahexaenoic acid; LDL = low-density lipoprotein; and HDL = high-density lipoprotein. Missing data for smoking status (*n* = 11), alcohol consumption (*n* = 73), sun protection (*n* = 2), central retinal thickness (*n* = 1), cup-to-disc ratio (*n* = 7), macular pigment distribution (*n* = 1), AMD stages (*n* = 15), plasma L/Z (*n* = 164), plasma n-3 PUFAs (*n* = 154), LDL cholesterol (*n* = 6), HDL cholesterol (*n* = 6), and triglycerides (*n* = 6).

The POAG patients were more likely to be men and they were significantly older. The eyes with POAG were more likely to be pseudophakic, to present with a larger cup-to-disc ratio and a thinner global RNFL thickness, and with a negative spherical equivalent.

#### *3.2. MPOD and Primary Open-Angle Glaucoma*

The concordance between the two eyes for MPOD among 50 randomly chosen subjects found an agreement with kappa 0.79 (confidence interval = 0.64–0.95). The MPOD means at the 0◦, 0.5◦, 1◦, 2◦, and 6◦ eccentricities for the two groups are presented in Table 3.

**Table 3.** Macular pigment optic density at several degree eccentricities from the fovea in control and primary open-angle glaucoma eyes in the Montrachet study.


*p*-value was calculated between control and primary open-angle glaucoma eyes. The results are displayed as mean ± standard deviation. POAG = primary open-angle glaucoma and MPOD = macular pigment optical density. MPOD was measured in optical density units.

In this univariate analysis, there was no statistically significant difference at each eccentricity degree between the two groups. After adjustment for age, gender, smoking and lens status, and plasma PUFAs, we compared the MPOD at 0.5◦ eccentricity between eyes without optic neuropathy and eyes with POAG. We could not find a statistically significant difference (*p* = 0.336; Table 4).

**Table 4.** Multivariable analysis of association between MPOD at 0.5◦ and the presence of primary open-angle glaucoma in the Montrachet study.


PUFA = polyunsaturated fatty acids; ALA = alpha-linoleic acid; EPA = eicosapentaenoic acid; MPOD = macular pigment optical density; OR = odds ratio; and CI = Confidence interval. In total, 162 observations were deleted due to missing values from smoking status, lens status, and plasma PUFAs. MPOD was measured in optical density units.

Similarly, we did not find a statistical difference for MPOD at 1◦, 2◦, and 6◦ eccentricity (data not shown). In the multivariable analysis, older age (*p* = 0.041), male sex (*p* = 0.017), pseudophakic status (*p* = 0.044), and the elevated plasma level of ALA (*p* = 0.017) were significantly associated with the presence of POAG. The agreement for MP spatial distribution profiles between the two investigators was 0.71 (confidence interval = 0.48–0.93). MP spatial distribution was classified into three patterns, namely ring-like (10.64%), intermediate (8.51%), and no ring-like (80.85%) in eyes with POAG as well as 17.90%, 10.67%, and 71.43% in the control eyes. There was no significant difference regarding the MP spatial distribution between eyes with POAG and eyes without optic neuropathy (*p* = 0.408; Table 2).

#### **4. Discussion**

Our study was designed to compare MPOD in an elderly population between POAG and control eyes using the two-wavelength autofluorescence method. Additionally, we aimed to investigate the MP spatial distribution between these two groups. We found no difference of MPOD and MP spatial distribution between eyes with POAG and eyes without optic neuropathy. These results are not in agreement with previous studies [16,29,30]. Igras E. et al. found a statistically significant difference for MPOD between glaucoma patients and controls with a MPOD median of 0.23 (−0.19; 0.65) and 0.36 (−0.08; 0.80) at 0.5◦ eccentricity, respectively [16]. Ji Y. et al. found MPOD significantly reduced in the glaucoma group compared to the control group, with 0.116 ± 0.033 UD and 0.137 ± 0.026 UD, respectively [30]. Recently, Siah WF. et al. [29] found that MPOD was lower in glaucomatous eyes with foveal ganglion cell complex involvement. We should note that these three case-control studies focused on smaller samples of selected patients (*n* = 40, *n* = 30, and *n* = 88, respectively). Nevertheless, our results could be difficult to compare to other studies as the MP measurement procedures were different. We measured MP with a twowavelength autofluorescence method whereas Siah WF. et al. [29], Igras E. et al. [16], and Ji Y. et al. [30] used customized heterochromatic flicker photometry (cHFP) and a fundus reflectance method. On one hand, the reflectance method has some limitations because it uses only one wavelength, which is problematic in the case of lens opacification [31]. On the other hand, cHFP results should be interpreted with caution as it is influenced by the operator's execution and participants' cooperation [32]. Moreover, cHFP is a time-consuming technique (approximately 30 min), which is difficult to realize in elderly participants. The two-wavelength autofluorescence method for MPOD measurement is faster (2 to 3 min for each participant) and objective, and can be performed by trained technicians. Nevertheless, MP values obtained by means of the Heidelberg Spectralis (HRA) with the two-wavelength autofluorescence method were comparable to MP values obtained using the densitometer (cHFP) [33].

In contrast to the three previous research groups, our results are in line with Daga FB and Bruns Y et al. [19,34]. Daga FB et al. found that patients diagnosed with glaucoma (mean age of 72.5 years) had comparable MP levels to control subjects (mean age of 70.0 years) with the same two-wavelength autofluorescence method. Moreover, they also demonstrated no significant relationship between MP, standard automated perimetry, and retinal nerve fiber layer thickness measurements [19]. Bruns Y et al. presented that there was no evidence for lower MPOD in their 43 glaucomatous-patient (mean age of 70.0 years) case-control study [34]. These conflicting results could also be found in epidemiology studies. Jae H. Kang et al. presented in a large population-based study (Nurses' Health Study and Health Professionals Follow-up Study) that increased dietary levels of carotenoid were associated with a lower risk of POAG [35]. On the other hand, in the Rotterdam study, no protective effect was found between carotenoids and POAG [36].

A negative correlation has been demonstrated between BMI and MPOD but in our study, we did not show any significant difference between the controls and POAG participants regarding their BMI [37].

In our study, the eyes with POAG were more likely to be pseudophakic, which could have influenced the MPOD measurement as mentioned by Sasamoto et al. [38]. Nevertheless, the difference between the phakic and pseudophakic eyes found in the literature is mainly due to the absorption of blue light by the cataractous lens. Thus, we took into account the lens status in the multivariable analysis in order to limit misinterpretation of the MPOD measurement secondary to blurred media due to cataract.

Carotenoids supplementation is a confounding factor for MPOD because it has been reported that supplementation increased the concentration of carotenoids in plasmatic serum and MPOD [39,40]. Unfortunately, we were not able to evaluate this point as there was no lutein/zeaxanthin supplementation in the POAG group. We used a semiquantitative food frequency questionnaire to quantify the dietary intake in terms of carotenoid among participants in order to control for any disparity in MPOD caused by diet. In our study, there was no significant difference between the two groups in term of carotenoids intake (data not shown).

One school of thought suggests from preclinical studies of the glaucoma model that MP could operate as a neuroprotective agent. MP could regulate the production of prooxidant stressors in the early ischemic retinal injury and protect the inner retina from the neurodegenerative process [41,42]. Another school of thought strengthened by the results of the present study found no association between MP and glaucoma [19,34,36]. Hence, positive results could be explained by many confounders (healthy diet and lifetime exposure) and not independently by carotenoids and MP.

We found a significant difference between glaucomatous and control participants for age, gender, and lens status. These results are in agreement with previous studies which showed that POAG prevalence increased with age and men were more likely to present with POAG compared to women [43–45]. Regarding lens status and POAG prevalence, results in the literature are not clear as in some situations, cataract surgery is part of the treatment of POAG. Significative association between an elevated plasma level of ALA and POAG is conflicting. It was previously presented that there was no significant difference regarding isolated plasma FAs between participants with POAG and participants without optic neuropathy in elderly [46]. Considering the dispersion of the results for elevated plasma levels of ALA and POAG in this study, this association should be confirmed by further clinical studies.

We acknowledge several limitations of this study. First, we analyzed only 611 eyes of 1153 participants in the Montrachet study due to imaging quality and availability of MP evaluation. Thus, missing data could decrease the power of our analysis even though there was no difference between participants and non-participants. Second, these findings are based on a Caucasian European population and cannot be extrapolated to other parts of the world and other ethnicities. Third, this exploratory cross-sectional study only enhanced a potential absence of association between MPOD, MP spatial distribution, and glaucomatous optic neuropathy. This should be confirmed by a longitudinal study to validate our findings. Fourth, it is also important to acknowledge that our two groups are unequal (*n* = 48 versus *n* = 553) with a small sample size of glaucomatous patients, which could lead to statistical bias and lower statistical power. Fifth, we did not collect ganglion cell layer thickness measurements, which could have given us the opportunity to focus the analysis on glaucomatous eyes with foveal ganglion cell complex loss.

We thought that the strengths of a population-based study could help to clarify the debate on the relationship of MP and glaucoma. In contrast to control-case studies, participants were enrolled in the Montrachet study regardless of their glaucomatous status.
