Introduction

Dietary antioxidants are a complex mixture of hydrophilic and lipophilic substances that come from plant foods and beverages. They exhibit a capacity to protect DNA, proteins and lipids against oxidative damage through different mechanisms and also to modulate gene expression or colonic microbiota selectively [1–3]. Strong cumulative evidence supports the role of dietary antioxidants as a whole in disease prevention and health promotion. Epidemiological and clinical studies show that diets with a high antioxidant capacity (AC), such as diets rich in fruit and vegetables and Mediterranean-type diets (MD), are associated with significant decreases in plasma C-reactive protein [4], ischemic stroke [5] or the overall risk of cardiovascular disease [6] or colorectal cancer [7]. Therefore, the determination of dietary antioxidant capacity, as the AC provided by all the foods and beverages consumed in a diet, is a useful parameter for estimating the potential of a diet to prevent chronic diseases associated with oxidative stress and ageing. Most research to date on food and diets AC has focused on soluble antioxidants (vitamins A, C and E, selenium, carotenoids and phenolic compounds), which are mostly low molecular weight compounds partially available in the small intestine. However, recent findings show that there is another important fraction of dietary antioxidants, the macromolecular antioxidants, which are insoluble in both intestinal fluids and aqueous-organic solvents. They appear as high molecular weight structures, constituted either by polymeric polyphenols or by small molecular weight antioxidants –phenolic compounds and carotenoids- associated with polysaccharides or proteins [8]. Different studies reported the individual constituents of macromolecular antioxidants [9–12]. Overall, the two major constituents of macromolecular antioxidants are polymeric non-extractable proanthocyanidins or condensed tannins (NEPA) and hydrolysable polyphenols (HPP), i.e., different phenolic compounds (hydrolysable tannins, phenolic acids, flavonols, etc.) associated with macromolecules [8].

After ingestion, macromolecular antioxidants cross unaltered the upper gut and once in the colon they are subjected to extensive fermentation, releasing bioactive metabolites that either have an in situ effect- for instance, they lowered polyps in mice with colon cancer and decreased lipid oxidation in the colonic mucose of healthy rats [13]- or are absorbed, exhibiting potential systemic effects in relation to cardiometabolic risk factors [14–16]. However, despite the increasing evidence of the role of macromolecular antioxidants in the health-related properties of dietary antioxidants, the fact that they are not transferred to the supernatants of the aqueous-organic extractions commonly performed for food antioxidant analysis make them to be absent in most food composition databases and in nutrition and health studies. Anyway, when a food is consumed, both soluble and macromolecular antioxidants are ingested.

Therefore, this study aimed to advance in the knowledge of macromolecular antioxidants as dietary bioactive compounds, estimating for the first time their contribution to the dietary AC. For that, the AC of the Spanish diet as example of a MD was evaluated, considering both soluble and macromolecular antioxidants.

Materials and Methods

Food Intake and Samples

Estimates of plant food and beverage consumption in the Spanish diet were based on national consumption data for December 2013 [17]. Such data are obtained annually from 12,000 daily household spending questionnaires. Samples were acquired at different local supermarkets and were representative of common brands, which appear together with the detailed list of the items in Table S1.

Preparation of Soluble Antioxidant and Macromolecular Antioxidants

Samples were subjected to an aqueous-organic extraction previously reported [9], with the supernatants of this extraction corresponding to soluble antioxidants. The residue, corresponding to macromolecular antioxidants, was subjected to two different procedures (in independent replicates) in order to obtain HPP and NEPA [18, 19]. In the case of samples with a fat content higher than 20 %, they were previously defatted to avoid interferences in AC determinations [20] and AC was measured separately in defatted samples, according to this procedure, and in their corresponding oil with a specific extraction [20]. In the case of beverages, measurements carried out directly on the original samples corresponded to total AC; in order to determine the fraction of this total AC that corresponded to macromolecular antioxidants, enzymatic treatments followed by dialysis were carried out to release low molecular weight antioxidants, according to a procedure previously described for isolating soluble dietary fiber with associated macromolecular antioxidants [21].

Analysis of Antioxidant Capacity

Two complementary AC methods, based on different mechanisms of reaction (metal reducing capacity in FRAP assay and radical scavenging ability in ABTS assay) and widely used in foods, intestinal fluids and plasma [22–25] were applied in this study, according to procedures previously reported [23, 26]. FRAP assay was applied to soluble antioxidants and HPP since it cannot be applied in the butanol medium of NEPA solution, while ABTS assay was applied to soluble antioxidants, HPP and NEPA. For both methods, total AC was calculated as the sum of the AC of the different fractions. Values were expressed as μmol Trolox/100 g w or 100 mL and as μmol Trolox/serving- mean serving sizes per food group previously reported were used [27] (Table S2).

Determination of Antioxidant Capacity Intake

We constructed a database of FRAP and ABTS values for each individual food, including both soluble and macromolecular antioxidants. Since these values were obtained for the dried edible fraction of each plant food, they were transformed into values for original fresh matter in order to interpolate them in the food consumption data reported above to calculate daily AC intake. For food groups in which there was a fraction of non-detailed “others” (fruit, vegetables and nuts) a mean value from the other items in the group was obtained and applied to the non-detailed fraction. Similarly, since part of food consumption takes place outside the home, the percentages for eating out in Spain recorded for 2006 [28] were applied to these data in order to consider total food consumption. Values were expressed as mean value ± SD; only SD for experimental AC data was considered, since it was not available for consumption data.

Results and Discussion

Total Antioxidant Capacity of Foods

The AC derived from soluble and macromolecular antioxidants was measured in 54 common foods and beverages, classified in different food groups (fruit, vegetables, cereals, nuts, legumes, cocoa products, beverages and oils). Significant correlations were obtained for AC results determined by FRAP and ABTS assay for soluble antioxidants (R2 = 0.8467, P < 0.001), HPP (R2 = 0.5049, P < 0.001) and the sum of both (R2 = 0.5741, P < 0.001). Since ABTS can be applied to both fractions of macromolecular antioxidants (HPP and NEPA), while FRAP can only be applied to HPP, it was decided to include here only results from ABTS assay, in order to make clearer the discussion. The results obtained by FRAP assay may be found as Supplementary material (Tables S3-S6).

The mean values for each group, corresponding to the two fractions of antioxidants, are shown in Fig. 1, as expressed per g or mL of sample or per serving. Macromolecular antioxidants were present in all the food groups except oils. Indeed, they provide more AC than soluble antioxidants except in the groups of vegetables and beverages, where they were significant but not major contributors to total AC (Fig. 1a). AC was also calculated per food serving (Fig. 1b). In this case, legumes and cocoa products provided the highest AC from macromolecular antioxidants. Regarding soluble antioxidants, a serving of cocoa products provided the highest AC, followed by beverages, legumes and nuts, all of them about 500 μmols of Trolox.

Fig. 1
figure 1

Mean values of antioxidant capacity in the different food groups determined by ABTS assay: a μmol Trolox/g dry matter or mL; b μmolTrolox/serving. ABTS, 2,2′-Azino-bis(3-ethylbenzo-thiazoline-6-sulfonic acid); HPP, hydrolysable polyphenols; NEPA, non-extractable proanthocyanidins

Full size image

The detailed data for each food item are shown in Tables S3-S4. It is remarkable that the AC data provided in this work constitute the most complete list of AC derived from macromolecular antioxidants. Thus, previous comprehensive lists of the AC in foods and beverages commonly consumed in other countries [22] only considered the contribution of soluble antioxidants, and data on AC derived from macromolecular antioxidants were limited to some specific samples [20].

Total Antioxidant Capacity of the Spanish Diet

National food consumption data from 2013 (after converting total amount of food to edible amount of food) were used to estimate the AC of the Spanish MD, including contributions from soluble and macromolecular antioxidants (Table 1). Total AC of this diet was about 8050 μmol Trolox by ABTS assay. The data for dietary AC from soluble antioxidants are similar to those previously reported for the Spanish diet [29, 30] and in the same range as those reported for other countries [5]. In particular, beverages were largely the main contributors to AC from soluble antioxidants (55 %), followed by fruit (16 %).

Table 1 Daily antioxidant capacity per capita intake in the Spanish diet determined by ABTS assay (μmol Trolox), expressed per food group
Full size table

The new aspect that arises from these data, however, is that macromolecular antioxidants, not previously considered in the evaluation of dietary AC of any population, are responsible for more than a half of the daily dietary AC: 61 % vs 39 % from soluble antioxidants. In the case of HPP, cereals, fruit and vegetables were the main contributors; while regarding NEPA intake, the major contributors were cocoa products and legumes. It is remarkable that the main contributors to dietary AC from NEPA were the food groups with the daily lowest consumption (nuts, legumes and coca products), but with a high NEPA content. Overall, these data constitute an updated estimation of the AC of the Spanish MD, which was previously determined including only soluble antioxidants [29, 30].

Details about the dietary AC intake derived from each individual food item can be seen in Table S6. Dietary AC from macromolecular antioxidants ranged from below 2 μmol Trolox for certain juices (orange, pineapple), vegetables (asparagus, carrot, cucumber, pepper) and almond, to more than 200 μmol Trolox for some legumes (lentils, pinto beans), cocoa products, banana, coffee and bread. Consequently with their high AC, some food items with low consumption (< 5 g/day) such as lentils, pinto beans, pistachio and cocoa soluble product had an important contribution to dietary AC (>50 μmol Trolox). As regards to the different food groups, the main individual contributors were: bananas in fruit, artichokes in vegetables, bread in cereals, lentils and pinto beans in legumes, pistachio in nuts, and soluble cocoa in cocoa products.

The intakes estimated here were based on food consumption data of 2013. It should be considered that current Spanish diet, although it still falls within the pattern of the MD [31], has suffered certain modifications in the consumption of different food groups since the 60s when the highest adherence to the MD was followed. In particular, from the sixties there has been a decrease in the consumption of certain antioxidant-rich food groups or items, such as legumes, olive oil and wine, partially attenuated by an increase in fruit consumption [32]. If data reported here for AC of foods and beverages are extrapolated to the consumption data in the sixties, an increase of about 30 % of the dietary AC is observed, with macromolecular antioxidants still providing a half of the total AC. Thus, the decrease in the adherence to the pattern of the MD in Spain also caused a decrease in dietary AC, but did not modify the relative contributions of soluble and macromolecular antioxidants.

Overall, the data obtained in this study show the important contribution of macromolecular antioxidants to AC in foods and diets. Moreover, this contribution is not only quantitative, but also qualitative, due to the specific characteristics that macromolecular antioxidants exhibit as compared to soluble antioxidants. Thus, since they reach the colon associated with macromolecules such as cell wall polysaccharides that promote the activity of the colonic microbiota, this increases the fermentation rate of macromolecular antioxidants that produce bioavailable metabolites [15]. Also, given their macromolecular structure, the colonic fermentation of these dietary antioxidants takes longer than that of soluble antioxidants, increasing the time that the derived beneficial metabolites will be in contact with the colonic tissue as well as their circulation time through the human body once absorbed [33]. Emerging evidence from animal and human studies have shown that the supplementation with macromolecular antioxidants concentrates may have beneficial effects in different parameters related to gastrointestinal health, including colon cancer [13], as well as systemic effects, especially in relation to cardiometabolic alterations [14–16]. Therefore, excluding macromolecular antioxidants from antioxidant studies implies ignoring a fraction of dietary antioxidants with a probably relevant and specific role on the described health effects of dietary antioxidants. The present study aimed to increase the knowledge on that “antioxidant gap” by determining for the first time the AC of a diet including macromolecular antioxidants. It is remarkable that all the methods used in this work to obtain and analyse the macromolecular antioxidant fractions had been previously validated [10], and individual constituents were identified in pools corresponding to macromolecular antioxidants in different food groups by HPLC-MS [34].

Conclusions

The AC associated with soluble antioxidants and macromolecular antioxidants in common foods and beverages was evaluated, showing the important contribution of macromolecular antioxidants to total AC. These data were used to determine the total AC of the Spanish MD including by the first time both soluble and macromolecular antioxidants. Over half of daily dietary AC in the Spanish MD is derived from macromolecular antioxidants, commonly ignored in antioxidant studies. Food AC data provided in this study can be used to estimate the dietary AC in other populations, which may be a useful parameter for estimating the potential of the diet to reduce the risk of certain chronic diseases or associations with their incidence. Including macromolecular antioxidants in studies on dietary antioxidants may contribute to a better understanding of the role of antioxidants in nutrition and health.