Lutein status is routinely assessed together with zeaxanthin, using dietary or biochemical methods, both of which have advantages and limitations, although serum concentration has been considered the best method for establishing their nutritional status in humans . Different factors linked either to the subject or to diet can influence their concentrations in serum and their bioavailability [3, 7, 10]. For this report, we selected a sample of apparently healthy subjects with similar serum lipids, body weight and dietary habits (varied diet and no dietary supplements) from two different age groups having the same proportion of men and women. To our knowledge, this is the first study to simultaneously assess dietary and status markers of lutein and zeaxanthin (diet, serum and MPOD) in Spanish subjects taking into account several confounding factors.
Serum lutein and zeaxanthin concentrations are similar to those reported in other groups of Spanish subjects  and somewhat lower than [4, 10] or similar to [1, 19, 20] those of comparable age in several European and North American countries. Serum lutein concentrations in the present study, and in practically all those mentioned above, are much lower than the >34 μg/dL (0.6 μmol/L) that seem to be consistently associated with lower risk in epidemiological studies (including lower risk for age-related macular degeneration and cataracts) and higher MPOD .
We observed no sex differences in serum concentrations and dietary intake of lutein and zeaxanthin or in MPOD, findings reported in a number of studies [4, 20–22], although other authors observed higher intake in women . In contrast, although cholesterol was within normal range, there were differences in serum lipids, and higher triglyceride and lower HDL-cholesterol levels were observed in men. However, serum lutein and zeaxanthin concentrations differed according to the age group, being higher in the older subjects. Their intake, both crude and energy-adjusted, is also greater, as reported elsewhere , and this group consumes more fruits and vegetables, as well. These differences disappear when serum concentrations are related to levels of circulating lipids (lutein + zeaxanthin/cholesterol + triglycerides), transporters of these xanthophylls, which are being considered for a better interpretation of the antioxidant/nutritional status [11, 25] or are expressed as 1000 kcal intake. The MPOD in the total sample was 0.35 du, similar to that of other studies [21, 22, 26], and differed according to age .
However, the amount of lutein plus zeaxanthin intake (median value) is much lower than that found by our group more than a decade ago in European subjects (n = 400) . In that study, the intake in the Spanish subgroup (n = 80) was 3.25 mg/day (range: 1.8-4.3 mg/d), which contrasts with the 0.45 mg/d of this study, which, in turn, is similar to the mean intake in the Spanish population obtained in national surveys (0.5 mg and 0.1 mg/person/day of lutein and zeaxanthin, respectively) . It is also similar to the levels reported in other populations . If we compare the mean intake usually found in the literature, rather than the median value (preferable since the data usually do not follow a normal distribution), the mean intake of 1.2 mg is similar or slightly below the range described in other studies [6, 8, 21, 24].
The different intake levels found in the literature are largely attributable to the differences among databases for carotenoid composition in foods and the types of dietary questionnaires employed . Regarding the aforementioned European study, the food composition data utilized are very similar to those used in this report  since, in both cases, much of the data resulted from HPLC analysis of Spanish foods; however, the European study was based on a semiquantitative food frequency questionnaire (FFQ), whereas in the present study, 3-day food records were employed. FFQ have been reported to overestimate carotenoid intake , especially that of lutein and zeaxanthin when comparing these two methods . FFQ were used in those studies in which intake concentrations were higher and 2 or 3-day recalls were employed in others in which the concentrations were comparable. On the other hand, the data from population-based studies in the USA and Spain indicate that levels of lutein and zeaxanthin intake have declined, particularly from dark-green leafy vegetables . Another aspect to be considered regarding lutein intake is food seasonality; however, in the Spanish population, the mean intake of lutein has been reported to be relatively constant throughout the year .
Concerning the sex differences in lutein and zeaxanthin intake, the results in the literature are inconsistent, indicating a higher intake among women [22, 23], in whom a lower energy intake  or no differences  are likewise reported. There are also discrepancies in the findings according to age group, with some studies that show no differences  and others that do; in the later, as in our study, a higher intake is reported for older individuals .
Lutein intake is higher than that of zeaxanthin in all populations [6, 20]; in a typical western diet, a ratio of 7:1 (lutein:zeaxanthin) has been reported . This is lower than that found in our study, which is 11:1 in the total sample and 13:1 in the older subjects, who consume more fruit and vegetables (Table 2) than the younger participants (10:1). The mean intake of fruits and vegetables is higher than that recommended by the WHO to decrease risk of chronic diseases (400 g/day).
The degree of correlation between lutein and zeaxanthin concentrations in serum and diet is significant, being higher if determined by dietary intake of fruit and vegetables, rather than by measuring the concentrations of these xanthophylls as provided by the diet (Table 3). The Pearson correlation coefficient between their serum concentrations and dietary intake is >0.2 and is >0.3 when serum concentrations correlate with food consumed, as in other studies [8, 21, 22, 26]. Aside from the amount injested, the degrees of correlation between biological markers, apparently low, are influenced by many other subject-related factors (i.e. sex, age, BMI, eating habits, sample size), as well as metabolic factors [2, 8], and in our study, the degrees of correlation were found to vary according to age group, but not sex. Thus, in younger subjects, there is a correlation between serum levels and dietary intake (in terms of both the concentrations of these compounds intake density) whereas, in older subjects correlation was observed only for zeaxanthin, despite the fact that these individuals have higher levels both in their intake and in serum. Therefore, other metabolic factors must influence metabolism of these xanthophylls, such as the uptake of lutein and zeaxanthin from plasma and their transport to the retina (e.g. serum lipid levels, or binding protein in optical tissue with binding capacity for zeaxanthin isomers, but not for lutein) . Correlation in young subjects but not in their elders has been reported by other authors . However, the correlations between serum concentrations and fruit and vegetable intake are significant in both age groups, especially for fruit intake.
The concentration of lutein and zeaxanthin in the retina, assessed by MPOD, was lower in the older subjects, despite the fact that their intake was higher than that of younger individuals (also expressed in relation to 1000 kcal). These higher intakes are reflected in the higher serum concentrations, although the differences disappear when the levels are expressed in relation to serum lipids. Correlations between MPOD and serum lutein and zeaxanthin and dietary intake reached the highest significance level for zeaxanthin and for lutein + zeaxanthin in relation to lipids (cholesterol + triglycerides, LDL-cholesterol and HDL-cholesterol) and when related to diet (lutein density and fruits and vegetables) (Table 4). Lutein and zeaxanthin are transported in LDL and HDL in similar amounts and it is widely accepted that oxidative modifications of LDL and HDL affect lipoprotein metabolism and modified LDL may have an effect on retinal pigment epithelial cells .
There are significant correlations in the elderly group, but not in the younger group (except the correlation with the lutein intake expressed as concentration per 1000 kcal, [ρ = −0.214, p = 0.026]). Although, MPOD is often positively associated with dietary and serum lutein and zeaxanthin concentrations , conflicting results have been reported on the influence of sex , age , BMI [2, 31], and other postprandial or environmental factors (e.g. smoking) . The highest correlations between MPOD and lutein and zeaxanthin in serum and dietary intake correspond to fruit and vegetables (ρ = 0.35) and to lutein + zeaxanthin/HDL-cholesterol (ρ = 0.301); the latter is consistent with the identification of HDL as the specific transporter of lutein and zeaxanthin to retina in chicks . The strong correlation between MPOD and fruit and vegetable intake, also reported elsewhere , indicates that not only the amounts of lutein and zeaxanthin are important for ocular tissue; there are other micronutrients and bioactive compounds in these foods that are also beneficial (e.g. fiber and polyunsaturated fatty acid intake are also directly related to MPOD) . On the other hand, few studies have assessed lutein and zeaxanthin status simultaneously using the three possible markers (diet, serum and MPOD) and these studies are not very homogeneous in terms of factors that can exert an influence such as sex, age, serum lipids and BMI, among others [21, 22, 26]. Nevertheless, although consistent associations among these three markers have not always been found, most studies point to the protective role of a diet rich in fruits and vegetables. However, in multivariate regression analysis, performed to assess the predictive value of lutein and zeaxanthin concentrations in serum and dietary intake, sex, age and the concentrations of different serum lipids for MPOD, only the serum lutein concentration and lutein + zeaxanthin in relation to cholesterol + triglycerides, but not fruit and vegetable intake, correlate with MPOD. Moreover, age is an important predictor in the total sample, as is sex in the older group, although we observed no sex differences in the MPOD values in this study, a fact that could be explained if the availability of lutein and zeaxanthin to the retina could be influenced by physiological sex differences  (i.e. if binding protein in optical tissue specific for zeaxanthin responds differently in men and women ).