1. IntroductionThe Mediterranean diet was firstly introduced for its health promoting benefits nearly three decades ago [1] and we recently celebrated a decade of it being recognized as an intangible world heritage monument [2]. In this period, the Mediterranean diet has been promoted across the globe as a healthy dietary pattern and incorporated into multiple policy actions in issues relating equally to health and agriculture/food promotion [3,4,5,6,7].As research data continue to emerge confirming the role of Mediterranean diet adherence to longevity [8,9], new concepts and metrics are being developed in order to translate dietary guidelines into actionable food and nutrition policy interventions [10]. One such approach is the use of nutrient profiling to develop front-of-pack labeling (FOP) schemes that can effectively guide food choices in line with the existing dietary guidelines [10]. Nutri-Score is one such scheme that converts the nutritional content of foods into a five-tier score ranging from A to E (green to red) from healthier to less healthy choices within food groups [11].The Nutri-Score scheme has been so far validated through research as a system that could successfully promote longevity and reduce mortality from non-communicable diseases [12,13]. However, more research is needed to understand how this system would operate in the food market and among consumers and, thus its maturity as a tool of public health policy [14].A great volume of the research is focused on the consumer understanding of Nutri-Score [15] and its ability to discriminate foods successfully based on their nutritional composition [16,17,18]. However, an important element to be studied is whether Nutri-Score grades foods in accordance with the local food-based dietary guidelines in order to assess whether its launch would create confusion among consumers [17,19,20]. Recently, we published an analysis showing that in the Greek marketplace, Nutri-Score showed a good discrimination capacity among foods in the same food group and overall agreement with the national food-based dietary guidelines [21].However, if Nutri-Score is to be launched regionally it is important to understand how the algorithm performs against not just national guidelines but with the overarching dietary pattern that serves as their inspiration, the Mediterranean Diet itself, and the foods that compose it [22,23,24].

In this study, we aim to assess whether the Nutri-Score algorithm aligns with the Mediterranean Diet guidelines and to what extent. To achieve that, foods currently sold in the Greek marketplace, as curated in the HelTH Branded Food Composition Database (BFCD), were (a) classified as being part of the Mediterranean diet guidelines or not and then (b) their Nutri-Score grading was compared to their positioning in the Mediterranean diet pyramid.

4. DiscussionThis study is the first, to our best knowledge, to directly investigate the agreement of Nutri-Score with the Mediterranean diet pyramid. It expands upon previous work that aimed to measure the alignment of Nutri-Score with national food based dietary guidelines in Mediterranean countries [17,21].

Of the foods analyzed only 25% fit in the traditional Mediterranean Diet Pyramid, and only 57.8% of all foods fit in the updated sustainable Mediterranean Diet Pyramid. Overall, for both pyramids, Nutri-Score was successful in clearly separating the foods at the top and the bottom of the pyramids as the majority of foods at the top were scored as “D” or “E” while those at the bottom received mainly “A”, thus suggesting a good alignment between the two. In the mid of the pyramids, Nutri-Score showed less of a capacity to granularly follow the pyramids as foods in the middle received primarily a score of “B”.

Nutri-Score grades were similar between the two pyramids. As the tMDP represents a more prudent version of the Mediterranean diet the lower average FSAm-NPS score found with this definition is in line with the spirit of the guidelines. When comparisons between the respective tiers are considered the Mediterranean diet definition used did not influence the average FSAm-NPS score in the same or similar tier.

The biggest source of disagreement between the two pyramids was the definitions of tiers and particularly the fact that the tMDP has more tiers/food groups as compared to the sMDP and they are not all positioned in the same level of the pyramid. The most striking differences are those of potatoes and eggs which are mentioned separately from vegetables and white meat, respectively, in the tMDP and positioned higher in the pyramid as opposed to the sMDP. This is linked with an anomaly in the granularity of Nutri-Score grading along the tMDP.

It is of interest to note that all Nutri-Score grades were seen in all tiers, a sign of a good discriminatory capacity as defined in previous research as it can guide towards improvements within the same food group/tier [17,19,21,24]. This became apparent in the nested analysis for dairy products in which it became clear that low-fat, low-sodium dairy such as milk and yogurt receive a substantially better Nutri-Score compared to cheese in line with the Mediterranean Diet guidelines. However, Nutri-Score was still capable of identifying cheese products that were significantly different from their peers, and even within the cheese subcategory, all Nutri-Score grades from “A” to “E” were documented. These results are in line with previous research studying the agreement of Nutri-Score with national food-based guidelines from Mediterranean countries, which reflect more or less the basic principle of the Mediterranean diet pyramid [11,16,17,19,21]. In the case of our analysis, the dataset used did not include any data on vegetable oils [25], as such the results presented herein do not include any data on vegetable oils including olive oil. As per the latest Nutri-Score algorithm [26], olive oil is automatically graded as (C) with discussions underway for an updated automatic “B” grading. Previous research indicates that olive oil is in fact graded better than other fats and oils and that consumers are capable of identifying olive oil as the optimal choice among all fats and oils under the Nutri-Score scheme [21,24].The current analysis highlighted a challenge linked to the Mediterranean diet pyramid itself and how it affects the mapping of foods as eligible or not. The biggest issue faced in the current analysis was the definition of which foods are considered traditional and local as those are integral to the definition of the Mediterranean diet. For the definition of traditional, a rule of identifying whether a food was part of the culinary tradition in the 1960s was employed for the tMDP, while for the sMDP an extended definition of whether a food would be considered a part of the Mediterranean culinary heritage would be considered [1,27,28]. In both cases, foods produced with modern processing techniques (extrusion, etc.), or including modern ingredients (esp. sweeteners) as well as the low-fat versions of traditional cheeses that were introduced in the market the past two decades (e.g., low-fat feta cheese) were considered as non-eligible. The second element of bias was introduced in the classification of composite foods as part of the Mediterranean diet pyramid. For example, composite traditional dishes such as stuffed vegetables, spinach pie, pulses cooked vegetables, and/or meat, although clearly a part of the culinary heritage of the Mediterranean, was not explicitly mentioned as examples in the Mediterranean diet pyramid guidelines. The classification of such foods within the pyramid remains unclear and difficult, especially for foods combining ingredients from different tiers. To overcome such bias, a systematic approach was developed for the current analysis, and a separate category of traditional composite foods was developed. These foods represented ~5% of the whole dataset and were currently excluded from the analysis. Although they do represent a small proportion of the dataset, it is important to highlight this gap in the Mediterranean diet pyramid guidelines as employing traditional cooking techniques is mentioned as part of the dietary pattern but no consensus is given on how composite foods should be studied and promoted [1,2,28].At this point, it is important to highlight some limitations of the current study. The primary limitations of this analysis are either linked to the nature of the HelTH BFCD and BFCDs in general or the nature of the Mediterranean diet pyramid guidelines. As a BFCD, HelTH only includes products that are sold as packaged food and those that are required by the legislation to carry a nutritional declaration. Although the use of branded foods data is an improvement in the relevance of the results for the consumer and the food industry as it is a direct reflection of the marketplace as compared to analyses performed on generic food composition data [29,30], for the case of the Mediterranean diet pyramid it also introduces specific hurdles.In its nature, the Mediterranean diet promotes the consumption of minimally processed, seasonal, and local produce often sold as fresh and non-packaged. A BFCD such as HelTh would not be able to map those products which translate to an underrepresentation of the foods currently available in the marketplace for categories such as fruits, vegetables, meat, and fish. It is safe to stipulate that the majority of packaged foods from these categories will be more extensively processed than the fresh equivalent and that could be linked to an overestimation of the Nutri-Score grade of the respective food category [31].Despite its limitations, this research is important in starting the discussion around the interplay between Nutri-Score and the Mediterranean diet. As the Mediterranean diet is a building block both for the health and agriculture agenda in the respective countries, ensuring that the two policies are aligned, and act synergistically is key. The current research provides a structured framework to start defining which foods can be considered as part of the Mediterranean diet and whether the available policy options are capable of guiding consumers towards its optimal integration into a healthy diet. An important finding of this work is that depending on the MD definition numerous foods could be excluded from the MD and that highlights the complexity of the MD dietary pattern and the need for in-depth consumer knowledge. It also highlights areas of improvement and future gaps which are necessary to guide future food development. For example, the case of how composite foods are positioned in the Mediterranean diet pyramid and whether the use of modern ingredients and processing techniques should be employed to improve the nutritional quality of traditional products are important considerations for the future. In this light, the current research builds upon previous findings which suggest good agreement between Nutri-Score and the national food-based dietary guidelines [21] by showing good alignment with the Mediterranean diet pyramid as well.