Scholarly article on topic 'Identifying Sustainable Foods: The Relationship between Environmental Impact, Nutritional Quality, and Prices of Foods Representative of the French Diet'

Identifying Sustainable Foods: The Relationship between Environmental Impact, Nutritional Quality, and Prices of Foods Representative of the French Diet Academic research paper on "Animal and dairy science"

CC BY-NC-ND
0
0
Share paper
Keywords
{Foods / Sustainable / Cost / "Nutritive value" / "Greenhouse gas emissions"}

Abstract of research paper on Animal and dairy science, author of scientific article — Gabriel Masset, Louis-Georges Soler, Florent Vieux, Nicole Darmon

Abstract Background Sustainable diets, as defined by the Food and Agriculture Organization, need to combine environment, nutrition, and affordability dimensions. However, it is unknown whether these dimensions are compatible, and no guidance is available in the official recommendations. Objective To identify foods with compatible sustainability dimensions. Methods For 363 of the most commonly consumed foods in the Second French Individual and National Study on Food Consumption, environmental impact indicators (ie, greenhouse gas [GHG] emissions, acidification, and eutrophication), and prices were collected. The nutritional quality of the foods was assessed by calculating the score for the nutritional adequacy of individual foods (SAIN) to score for disqualifying nutrients (LIM) ratio. A sustainability score based on the median GHG emissions, price, and SAIN:LIM was calculated for each food; the foods with the best values for all three variables received the highest score. Results The environmental indicators were strongly and positively correlated. Meat, fish, and eggs and dairy products had the strongest influence on the environment; starchy foods, legumes, and fruits and vegetables had the least influence. GHG emissions were inversely correlated with SAIN:LIM (r=–0.37) and positively correlated with price per kilogram (r=0.59); the correlation with price per kilocalorie was null. This showed that foods with a heavy environmental impact tend to have lower nutritional quality and a higher price per kilogram but not a lower price per kilocalorie. Using price per kilogram, 94 foods had a maximum sustainability score, including most plant-based foods and excluding all foods with animal ingredients except milk, yogurt, and soups. Using price per kilocalorie restricted the list to 42 foods, including 52% of all starchy foods and legumes but only 11% of fruits and vegetables (mainly 100% fruit juices). Conclusions Overall, the sustainability dimensions seemed to be compatible when considering price per kilogram of food. However, this conclusion is too simplistic when considering price per kilocalorie, which highlights the need to integrate the data at the diet level.

Academic research paper on topic "Identifying Sustainable Foods: The Relationship between Environmental Impact, Nutritional Quality, and Prices of Foods Representative of the French Diet"

RESEARCH

Original Research

Identifying Sustainable Foods: The Relationship between Environmental Impact, Nutritional Quality, and Prices of Foods Representative of the French Diet*

Gabriel Masset, PhD; Louis-Georges Soler, PhD; Florent Vieux, PhD; Nicole Darmon, PhD

ABSTRACT

Background Sustainable diets, as defined by the Food and Agriculture Organization, need to combine environment, nutrition, and affordability dimensions. However, it is unknown whether these dimensions are compatible, and no guidance is available in the official recommendations.

Objective To identify foods with compatible sustainability dimensions. Methods For 363 of the most commonly consumed foods in the Second French Individual and National Study on Food Consumption, environmental impact indicators (ie, greenhouse gas [GHG] emissions, acidification, and eutrophication), and prices were collected. The nutritional quality of the foods was assessed by calculating the score for the nutritional adequacy of individual foods (SAIN) to score for disqualifying nutrients (LIM) ratio. A sustainability score based on the median GHG emissions, price, and SAIN:LIM was calculated for each food; the foods with the best values for all three variables received the highest score.

Results The environmental indicators were strongly and positively correlated. Meat, fish, and eggs and dairy products had the strongest influence on the environment; starchy foods, legumes, and fruits and vegetables had the least influence. GHG emissions were inversely correlated with SAIN:LIM (r=—0.37) and positively correlated with price per kilogram (r=0.59); the correlation with price per kilocalorie was null. This showed that foods with a heavy environmental impact tend to have lower nutritional quality and a higher price per kilogram but not a lower price per kilocalorie. Using price per kilogram, 94 foods had a maximum sustainability score, including most plant-based foods and excluding all foods with animal ingredients except milk, yogurt, and soups. Using price per kilocalorie restricted the list to 42 foods, including 52% of all starchy foods and legumes but only 11% of fruits and vegetables (mainly 100% fruit juices). Conclusions Overall, the sustainability dimensions seemed to be compatible when considering price per kilogram of food. However, this conclusion is too simplistic when considering price per kilocalorie, which highlights the need to integrate the data at the diet level.

J Acad Nutr Diet. 2014;H — -H.

ARTICLE INFORMATION

Article history:

Accepted 30 January 2014

Keywords:

Sustainable

Nutritive value Greenhouse gas emissions

Copyright © 2014 by the Academy of Nutrition

and Dietetics.

2212-2672/$0.00

http://dx.doi.org/10.1016/jjand.2014.02.002

During 2010, the food and agriculture

Organization introduced a definition of sustainable diets that includes dimensions related to environmental impact, nutritional adequacy, cultural acceptance, affordability, and economic development.1 Improving the supply chain of foods might help to achieve more sustainable diets,2 but meeting all of the sustainability dimensions without major dietary changes may prove challenging.3,4

The production of foods of animal origin, particularly ruminant meat, emit more greenhouse gases (GHGs) than the

production of plant-based foods,5-7 and a reduction in meat intake has been modeled as the main strategy to significantly reduce global GHG emissions arising from our food choices.8 Overconsumption of red meat is also related to the increased incidence of mortality from noncommunicable diseases.9 Therefore, diets containing fewer meat products may have less of an environmental impact and may be healthier,10-12 suggesting that the environmental and public health requirements of sustainable diets might be compatible. However, such studies are based on theoretical dietary patterns with questionable cultural acceptance, and the nutritional adequacy of the proposed diets is seldom analyzed. Foods of animal origin contain high amounts of essential nutrients. Reducing their intake at population level may be challenging, especially in countries in which the population has a documented risk of nutrient deficiencies.13

*This is an open access article under the CC BY-NC-ND license ( :p:// creativecom mons.org/licenses/by-nc-nd/3 ).

ARTICLE IN PRESS

RESEARCH

In addition, meat and fish are expensive food items, but fruits and vegetables (F/V) are also expensive when their energy costs are considered,14 which explains why healthy diets that provide sufficient energy intake are often more expensive than unhealthy ones.15,16 Very few studies estimating the environmental impact of diets also include a cost analysis. Berners-Lee and colleagues17 showed that vegetarian and vegan diets could be cheaper than the higher GHG-emitting observed UK diet, but Macdiarmid and col-leagues11 observed no price difference between theoretical diets with reduced GHG emissions and the observed UK diet.

To identify food combinations that could combine all sus-tainability dimensions, data on individual foods are needed. An analysis of the relationships between the different dimensions of sustainability at the food level would enable the determination of the compatibility or incompatibility of these dimensions and to identify the most "sustainable" foods. The objective of our study was to identify foods that have a low environmental impact, a high nutritional quality, and an affordable price by combining the three dimensions into a single sustainability score.

MATERIALS AND METHODS

Survey Data and Food Selection

The foods that best represent the French diet were selected because assessing the environmental impacts of all foods and drinks would be a very costly process. To account for the diversity of French food consumption patterns, data from the Second Individual and National Study on Food Consumption cross-sectional dietary survey conducted during 2006-2007 on a nationally representative sample of the French adult population were used (N=2,624).18 The sampling method was a three-stage stratified random sampling strategy that has been described elsewhere.19 To ensure the representativeness of the sample, statistical adjustments were made for region, town size, age, sex, occupation of the head of household, household size, and seasonal variables. The study was approved by the French Data Protection Authority (Commission Nationale Informatique et Libertés).

After excluding the energy underreporters using the Goldberg and Black equations,20,21 1,918 healthy adults (776 men and 1,142 women) aged 18 to 79 years were retained. Dietary intake was assessed using a 7-day diet record, and all reported food items (N=1,314 foods and beverages, including water) were aggregated into 16 food groups and 36 food subgroups within a food nutrient composition database associated with the survey.22 For each food in the database, the percentage of individuals who consumed this item (ie, the participants who reported consuming the food at least once in their 7-day diet record) was calculated. Then, within each of the 36 food subgroups, the food items were ranked in decreasing order based on the percentages of consumers, and at least one food item was selected from among the most widely consumed items in each subgroup. This process resulted in 391 representative foods from the 1,314 items initially listed in the food database. These 391 foods covered 71% of the total weight intake and 66% of the total energy intake of the participants in the Second Individual and National Study on Food Consumption.

Environment Impact Indicators

An environmental consulting firm, Greenext Service, assigned values to the 391 foods for three environmental impact indicators: GHG emissions expressed in grams carbon dioxide equivalent units, air acidification (emissions in the atmosphere responsible for acid rains) in grams sulfur dioxide equivalent units, and freshwater eutrophication (the accumulation of ions in water, which is responsible for unwanted algae development) in milligrams phosphate equivalent units. The three indicators were assessed with a life cycle analysis, defined as follows by the ISO14040 and 14044 standards: the "compilation and evaluation of the inputs, outputs and the potential environmental impacts of a product system throughout its life cycle."23,24 Thus, the environmental impact indicator estimates include the results associated with each stage of the production, transformation, packaging, distribution, use, and end-of-life of food products. Using a top-down approach combining French trade and production data25,26 and standard life cycle inventory data (eg, Ecoinvent27), the final values for all three indicators reflected the average food product consumed in the French market.28

Nutritional Quality of Foods

To assess the nutritional quality of each food, the score for the nutritional adequacy of individual foods (SAIN) and score for disqualifying nutrients (LIM) scores were used.29 Both scores are calculated as average (nutrient content/recommendation) ratios, with the SAIN per 100 kcal and the LIM per 100 g. The SAIN uses five basic nutrients (ie, protein, fiber, calcium, vitamin C, and iron), and the LIM includes three nutrients (ie, saturated fatty acids, added sugars, and sodium). Vitamin D is used as an optional nutrient for calculating the SAIN; it replaces one of the five basic nutrients if the content/recommendation ratio for vitamin D is greater than one of the basic nutrients. The SAIN and LIM algorithms were applied to each food in the same manner, with the exception of two food categories29: for sweet drinks, the LIM was multiplied by 2.5, assuming a regular portion size of 250 mL, and for nuts and foods deriving more than 97% of their energy content from fat (eg, oils, margarine, and butter), vitamin E, monounsaturated fatty acids, and a-linoleic acid were used as optional nutrients in the SAIN algorithm to account for the quality of lipids. To integrate the nutritional quality of foods into one dimension, the ratio of SAIN to LIM was used, with the LIM set to one when lower than one. This ratio correlates well with modeled diets that meet a full set of nutrient recommendations: the median SAIN:LIM of the foods included in modeled diets increases with the increasing nutritional quality of the

diets.30

Food Prices

Food prices were obtained from the 2006 Kantar WorldPanel French household consumer panel,31 which gives the annual expenditures and the quantity purchased of each food item available on the market by a representative sample of 12,000 French households. The mean prices were calculated by dividing the annual expenditures by the quantities purchased. Prices in Euros were converted to US dollars using an exchange rate of 1.26 (2006 average).32

ARTICLE IN PRESS

RESEARCH

Analysis

All indicators (environmental indicators, the SAIN:LIM, and the Kantar prices) were calculated per 100 g or per kilogram of edible food (ie, after the changes in weight associated with the trimming or cooking processes were taken into account using the appropriate conversion factors, such as the refuse percentage in the US Department of Agriculture National Nutrient Database for Standard Reference nutrient composition table33).

Because the distribution of food characteristics is generally nonnormal, medians and nonparametric tests were mostly used. The SAIN:LIM could not be calculated for energy-free foods (eg, water and diet soft drinks). As a result, all virtually energy-free drinks were excluded from the analyses, resulting in a final sample of 363 foods. Spearman rank correlations were computed to assess the relationship between the three environmental impact indicators, the SAIN:LIM, and the prices (per kilogram and 100 kcal) for the whole food database and for each food group.

A score combining all three dimensions of sustainability was developed for the purpose of the study. This sustain-ability score was based on the overall medians of the GHG emissions, the SAIN:LIM, and the price of each food. It ranged from 0 to 3, with each food scoring 1 point if its GHG emissions were under the median, 1 point if its price was under the median, and 1 point if its SAIN:LIM was above the median. On this basis, the 363 foods were classified based on the score (0,1,2, or 3), and the most "sustainable" foods (with a score of 3) were identified. Two types of sensitivity analyses were conducted. First, the results obtained with prices expressed in dollars per kilogram were systematically compared with those obtained with prices expressed in dollars per 100 kcal, another way to express the economic dimension.34 Second, the analyses were duplicated using a modified LIM that included free sugars—defined by the World Health Organization35 as all monosaccharides and disaccharides added to foods by the manufacturer, cook, or consumer, plus sugars that are naturally present in honey, syrups, and fruit juices— instead of added sugars in its algorithm.

All analyses were computed using SAS statistical software (version 9.3, 2011, SAS Institute Inc).

RESULTS

Descriptive Statistics

The median values across food groups and subgroups for the three environmental indicators, the SAIN:LIM, and price are presented in Table 1. Food groups and subgroups within food groups are listed in order of decreasing GHG emissions.

Environment Dimension. The environmental impact indicators were highest for animal products (eg, meats, fish, eggs, and dairy products) and, to a lesser extent, foods containing animal ingredients (eg, mixed dishes, sandwiches, and animal fat). The GHG emissions and acidification indicators were highest for the ruminant meat subgroup, and the eutrophication indicator was highest for the pork, poultry, and eggs subgroup. Foods that were high in fat/salt/ sugar had values slightly over the general median for the three indicators. Starchy foods and F/V had the lowest values for the three indicators of environmental impact. Acidification and eutrophication were strongly correlated both with

each other (the Spearman coefficient, r, was 0.75) and with GHG emissions (r=0.90 and 0.85, respectively; data not shown). Due to the very high correlations observed between the three environmental factors, the GHG emissions indicator was retained as the sole environmental impact indicator in the following analyses.

The Figure illustrates the higher GHG emissions arising from animal products. It also shows the high intrasubgroup variability in the environmental impacts of foods. Fish products had the most variable GHG emissions, whereas the highest values for F/V were similar to most of the animal product medians (except ruminant meat).

Nutrition Dimension. According to the SAIN:LIM, the F/V food group had the highest nutritional quality, and the foods that were high in fat/salt/sugar had the lowest (Table 1). Animal products had intermediate SAIN:LIM values, with fish and fish products achieving much higher nutritional quality values than meat and poultry. Deli meats achieved very low SAIN:LIM values. Desserts, sweets, pastries, butter and creams, and soft drinks had the lowest SAIN:LIM values.

Economic Dimension. All meat, fish, and eggs except deli meats were expensive according to the prices expressed either per kilogram or per 100 kcal; of those foods, fish and fish products were the most expensive foods. The rankings of foods that were high in fat/sugar/salt and of F/V depended on the way the prices were calculated (per kilogram or per 100 kilocalories). Starchy foods comprised the cheapest food group, based on prices expressed per kilogram and per 100 kcal.

Correlations between the Environment Dimension and the Nutrition and Economic Dimensions

When calculated for the whole food database, GHG emissions were inversely associated with the SAIN:LIM ratio and positively associated with the price per kilogram, but not with price per 100 kcal (Table 2). These significant correlations observed across all foods were also found in four of seven food groups, namely dairy products, starchy foods, F/V, and mixed dishes and sandwiches. The price per 100 kcal was correlated with GHG emissions only for dairy products.

Identifying Sustainable Foods Using the Sustainability Score

Using the price per kilogram, 94 (26%) foods obtained the maximum score of three for sustainability (Table 3). Most plant-based foods obtained the maximum score; that is, F/V, including 100% juices, vegetable oils, and starchy foods. Some plant-based foods did not obtain the maximum sustainability score either due to higher GHG emissions (eg, dried fruits), or lower nutritional quality (eg, breads with a high salt content), or higher price per kilogram (eg, figs, mango, asparagus, and dried fruits), relative to the respective medians. The only foods containing animal ingredients to obtain the maximum sustainability score were milks, yogurts with no added sugar, and soups containing meat or fish. Most fish products obtained a score of 1, except inexpensive canned sardines and mackerel rich in vitamin D, which obtained scores of 2. Meats did not score >1, and most deli meats scored 0 (data not shown). Foods that were high in fat/salt/sugar and mixed dishes mainly obtained scores of 0 or 1 (data not shown).

RESEARCH

Table 1. Environmental indicators, score for nutritional adequacy of individual foods to score for disqualifying nutrients ratio (SAIN:LIM), and price median values across food groups and subgroupsa

GHG Air Freshwater

emissions acidification eutrophication Price

Group and family_n (gCQ2eq/100 g) (gSQ2eq/100 g) (mgPQ4-eq/100 g) SAIN:LIMc Price $/kgd $/100 kcald

All foods (general median)a 363 224 1.64 124 0.68 6.33 0.40

Meat, fish, and eggs 56 604*** 10.2*** 401*** 1.50 17.0*** 1.01***

Ruminant 10 1587** 33.7** 443** 1.32 16.0** 1.07

Pork, poultry, eggs 9 684** 11.4** 540** 1.63* 11.6* 0.87

Deli meats 12 573*** 10.9*** 405*** 0.12*** 11.2*** 0.33

Fish and fish products 25 471*** 2.19 176*** 4.37*** 20.0*** 1.92***

Dairy products 42 457* 7.42*** 124 0.23** 9.70 0.35

Cheese 28 519*** 8.67*** 151*** 0.17*** 14.2*** 0.43

Yogurt 10 195 2.3** 63** 0.99 2.49** 0.33*

Milk 4 129 1.91 44 3.48 0.89 0.18

Mixed dishes and sandwiches 44 346** 4.20** 182** 0.65 7.93* 0.43

With food of animal origin 35 452*** 4.82*** 200*** 0.56 8.15* 0.42

Vegetarian 9 174 0.86 73 1.63 4.62 0.50

Foods high in fat/salt/sugar 70 225 2.09 145.5 0.12*** 6.30 0.24***

Breakfast cereals 5 266 1.37 230 0.41 6.33 0.16

Salty snacks 6 245 1.15 161* 0.18 10.7 0.23

Desserts, sweets, pastries 53 240 2.56*** 148 0.10*** 6.60 0.24***

Soft drinks 6 53.5* 0.28* 33* 0.01* 1.22* 0.26*

Fats and condiments 30 171 1.03 120 0.42 4.57 0.21

Butter, cream 5 369 6.1 107 0.04 4.99 0.16

Oils, margarine 10 171 0.77 137 0.32 2.90 0.04**

Condiments 15 140 1.01 94 2.58 6.24 0.92

Starchy foods 29 133** 0.65*** 114 1.62 3.14*** 0.15***

Grains 19 133 0.88*** 114 0.64 3.81** 0.16***

Potatoes 5 132 0.49 51 1.62 2.02 0.13

Legumes 5 118 0.30 102 9.10 2.76 0.30

Fruit and vegetables 92 92.6*** 0.69*** 42.4*** 13.0*** 3.52*** 0.83***

Dried fruit and nuts 8 196 1.85 69 0.64 11.0 0.28

Cooked vegetables 34 161*** 0.76*** 98.5 18.4*** 3.65*** 1.89***

Processed fruit and juices 12 100*** 0.6*** 38** 1.46 1.46*** 0.32**

Fresh fruit 24 81.1*** 0.60*** 29.4*** 10.5*** 3.10*** 0.73***

Raw vegetables 14 75.0** 0.61*** 35.3*** 28.9*** 3.88 2.18***

aAllvariables computed for edible foods. bGHG=greenhouse gas.

cSAIN:LIM was used as an indicator of nutritionalquality.29,30

dFood prices derived from 2006 household consumer paneldata,31 an average Euro to US dollar exchange rate was used for the conversion.32 *P<0.05 for sign test for comparison with generalmedian. **P<0.01 for sign test for comparison with generalmedian. ***P<0.001 for sign test for comparison with generalmedian.

Computing the same score using food prices per 100 kcal four foods did not have the maximum score using the price led to a much shorter list of 42 (12%) foods with the per kilogram (ie, muesli cereals, wheat germs, walnut oil, and maximum score of three (Table 3). Among these 42 foods, 38 dates). Most F/V lost their maximum score due to a high price already had a score of three using the price per kilogram, and per 100 kcal. Milks and yogurts with no added sugars, fruit

ARTICLE IN PRESS

RESEARCH

2000 -

1500 -

т: 0) о о о

ет аз см О О

if *■ ^ «Г

о/' ^ -

..er у ^

ЙГ <?v

Figure. Mean, median, interquartile range, and extreme greenhouse gas emissions for all food subgroups, expressed as grams of carbon dioxide equivalents (gCO2eq) per 100 g edible food. (C) Indicates mean. Mixed animals and mixed veg indicates mixed dishes with or without meat, fish, or eggs. Break=breakfast. Dr=dried. Pr=processed (also includes fruit juices).

Table 2. Spearman rank correlations between greenhouse gas emissions per 100 g, score for nutritional adequacy of individual foods to score for disqualifying nutrients ratio (SAIN:LIM), and price per kilogram and per 100 kcala for individual food groups

GHGb emissions Price Price

(gCQ2eq/100 g) n SAIN:LIMc ($/kg)d ($/100 kcal)d

For all foods 363 -0.37*** 0.59*** 0.09

Dairy products 42 -0.55*** 0.73*** 0.40**

Starchy foods 29 -0.54** 0.66*** 0.28

Fruit and vegetables 92 -0.34** 0.22* -0.03

Mixed dishes, 44 -0.30* 0.42** 0.23

sandwiches

Meat, fish, and eggs 56 0.01 0.23 0.22

Foods high in 70 -0.13 0.18 -0.02

fat/salt/sugar

Fats and condiments 30 -0.16 0.20 -0.07

aAllindicators computed for edible food.

bGHG=greenhouse gas (expressed in carbon dioxide equivalents).

cSAIN:LIM was used as an indicator of nutritionalquality.29,30

dFood prices derived from household consumer panel data,31 an average Euro to US

dollar exchange rate was used for the conversion.32

*P<0.05.

**P<0.01.

***P<0.001.

juices, starchy foods (still excluding breads), and vegetable oils all kept the maximum score. Starchy foods, including legumes, stood out as the only food group in which more than half of the foods achieved the maximum score (almost all nonbread items).

Using the SA1N:L1M ratio with the modified LIM subscore (ie, including free sugars instead of added sugars in the algorithm) did not change the conclusions. A total of 92 foods were identified as sustainable using the price per kilogram, and 40 were considered sustainable using the price per kilocalorie. Of these 40 foods, 39 were selected using the original score, with three fruit juices excluded and a sweetened flavored yogurt added to the list (data not shown).

DISCUSSION

The originality of our analysis lies in the food-level assessment of the relationships among three dimensions of sus-tainability in foods: environmental impact, nutritional quality, and price. Our study confirms some previous observations; for example, animal products are the biggest GHG emitters6,7 and the most expensive foods,14 and F/V have the best nutrient profile and are expensive sources of energy.34 1n addition, the three environmental indicators are strongly correlated, indicating that the conclusions obtained for GHG emissions might be generalized to air acidification and freshwater eutrophication. The absence of animal products— apart from dairy products—in the list of sustainable foods identified with the score based on the median GHG emissions, SA1N:L1M, and price per kilogram or price per 100 kcal

RESEARCH

Table 3. Sustainable foods that obtained the maximum sustainability scorea

Food group n

Foods with maximum sustainability score using price/kilogram (n [%])_

Foods with maximum sustainability score using price/100 kcal (n [%])_

Meat, fish,

and eggs Fruit and vegetables

58 (63%)

Canned mushrooms; cooked cauliflower; cooked broccoli; cooked green cabbage; cooked brussels sprouts; cooked frozen green beans; cooked spinach; cooked zucchini; cooked bell pepper;cooked eggplant; canned tomatoes;cooked squash; cooked carrot; cooked onion; cooked turnip;canned salsify; canned carrots;cooked leek; cooked fennel;cooked celery; ratatouille; canned/frozen mixed vegetables; canned tomato paste; red cabbage; white cabbage; chicory; lettuce; tomato; avocado; carrot; beet; radish; canned sweet corn; apple; mandarin; pear; orange; kiwi fruit;white/blackgrapes; peeled/unpeeled peach; grapefruit; nectarine; apricot; pineapple; plum; fruit sauce;banana; 100% orange juice, from concentrate; 100% mixed fruit juice, from concentrate;100% apple juice, from concentrate; 100% mixed fruit juice with added vitamins, from concentrate;100% grapefruit juice, from concentrate;100% pineapple juice, from concentrate; 100% grape juice 2 (2.8%)

Soy-based dairy-like dessert;semolina cake 4 (13%)

Vinegar;sunflower oil;mixed plant oil; rapeseed oil Starchy foods 29 15 (52%)

Cooked couscous; cooked pasta; cooked egg pasta; cooked white rice;cooked wheat;forexample, boulgur; cooked wholemeal pasta; cooked wholemeal rice; cooked chickpeas;cooked kidney beans;cooked peas; cooked lentils; boiled potatoes; frozen diced potatoes; frozen french fries; reconstituted mashed potatoes

Foods high in

fat/salt/sugar Fats and condiments

Mixed dishes, sandwiches

7 (16%)

Homemade vegetable soup; industrial vegetable soup; dehydrated vegetable soup; fish soup; chicken and pasta soup; couscous salad; coleslaw Dairy products 42 8 (19%)

Fat-free milk, reconstituted powder; reduced-fat milk; fat-free milk; whole milk; standard unsweetened yogurt; standard fruit yogurt; bifidus yogurt; Greek-style yogurt 94 (26%)

10 (11%)

Canned sweet corn; banana; 100% orange juice, from concentrate; 100% mixed fruit juice, from concentrate; 100% apple juice, from concentrate; 100% mixed fruit juice with added vitamins, from concentrate; 100% grapefruit juice, from concentrate; 100% pineapple juice, from concentrate; 100% grape juice; dried dates

1 (1.4%)

Swiss style muesli

5 (17%)

Sunflower oil; mixed plant oil; rapeseed oil; walnut oil; wheat germs

15 (52%)

Cooked couscous; cooked pasta; cooked egg pasta; cooked white rice; cooked wheat; for example, boulgur; cooked wholemeal pasta; cooked wholemeal rice; cooked chickpeas; cooked kidney beans; cooked peas; cooked lentils; boiled potatoes;frozen diced potatoes;frozen french fries; reconstituted mashed potatoes

4 (9.1%)

Dehydrated vegetable soup; chicken and pasta soup; couscous salad; coleslaw

7 (17%)

Fat-free milk, reconstituted powder; reduced-fat milk; fat-free milk;whole milk;standard fruit yogurt; bifidus yogurt; Greek-style yogurt

42 (12%)

aTo achieve the maximum score of 3, foods needed to have their greenhouse gas emissions below the overallmedian, their price below the overallmedian, and their nutritionaladequacy of individualfoods (SAIN) to score for disqualifying nutrients (LIM) ratio above the overallmedian; the SAIN:LIM ratio was used as an indicator of nutritional quality;29,30 prices were derived

from household consumer paneldata.

ARTICLE IN PRESS

RESEARCH

strengthens the rationale that reducing animal product consumption could be a major lever to increase the sustainability of diets.3'8'36

Our work provides new insights regarding food sustain-ability by showing that the three dimensions of sustainability were generally compatible with each other when price was expressed per kilogram. Most low-GHG-emission foods had higher nutritional quality and a lower price per kilogram, with 26% of the 363 analyzed foods identified as sustainable. The compatibility of the three sustainability dimensions was less obvious when price was expressed per 100 kcal: only 42 foods (12%) were identified as sustainable. More than half of the starchy foods and almost no fresh F/V could be considered sustainable. Therefore, the results of our analysis indicate that simple messages suggesting a straightforward relationship among environmental impact, healthfulness, and price of foods should be disseminated cautiously. Choosing the best option to identify sustainable foods must depend on the intended application.

Deriving a sustainable diet from the 42 sustainable foods mentioned above is questionable considering the French World Wide Fund Livewell sustainable dietary patterns previously identified using dietary modeling with similar data.37 Indeed, that modeled diet contained high amounts of legumes, potatoes, and dairy products, which are all among the 42 sustainable foods. However, the Livewell diet contained many F/V that did not achieve the maximum score based on price per kilocalorie. The Livewell dietary pattern also contained foods of animal origin, including some meat and fish that were not identified as sustainable in our analysis.

Therefore, restricting the sustainable food list too much when using price per kilocalories may not allow for realistic and culturally acceptable dietary patterns, and the present sustainability score method, which selects foods without taking food groups into account, may not be the most appropriate method. Similar to the food nutrient profiling concept, identifying the most sustainable foods within each food group may be a more sensible solution.38 From a consumer's point of view, identifying the most sustainable foods on a 100-g basis might be more practical insofar as it promotes greater choice during daily shopping. In addition, such a basis for the calculations would also coincide with the food labeling objectives in the European Union because the mandatory existing nutrition labels use a 100-g basis.39

The results of our analysis need to be integrated at the diet level to identify culturally acceptable food combinations that are nutritious, environmentally friendly, and economical, so these practical and achievable sustainable dietary patterns can be listed in official recommendations and communicated to the general public. Some institutions have proposed food-based recommendations that include both nutrition and environment dimensions, such as the double pyramid of the Barilla Center for Food and Nutrition.40 However, these recommendations were not verified in theoretical or observed dietary patterns, and the conclusions were based only on a per weight analysis.

The various steps of the life cycle of food products (eg, production, packaging, transportation, and preservation) can influence their environmental impact.41-43 Our analysis used three different environmental impact indicators calculated using the standardized ISO 14040 and 14044 life cycle analysis method23,24; a strength of our study. However, our

analysis did not capture the full complexity of the environmental impact of individual foods. Degradation of the environment and associated ecosystems has other dimensions (eg, biodiversity, ecotoxicity, land use, and depletion of natural resources such as fish stocks) for which standardized indicators at the food level are under development.

Our results provide useful insights into the relationship between the environmental impact, nutritional quality, and price of individual foods. Overall, the foods that had the greatest environmental impact had lower nutritional quality and a higher price per kilogram, suggesting that these three dimensions of sustainable diets may be generally compatible. However, the role of the energy density of foods and the related price per kilocalorie showed that this compatibility is not entirely straightforward. A diet-level approach is now needed to integrate our results and derive sustainable dietary patterns that could be used by public health practitioners. In addition, the environmental impact was highly variable within food groups, showing that improvements could be achieved within the food supply chain. To facilitate cultural acceptability of proposed changes, regulators need to address the sustainability issue both on the production and consumption sides of the food sector.

References

1. Annex I. International scientific symposium Biodiversity and sustainable diets—Final document. In: Burlingame B, Dernini S, eds. Sustainable Diets and Biodiversity - Directions and Solutions for Policy, Research and Action. Rome, Italy: Food and Agriculture Organization; 2012:294.

2. Gerber P, Steinfeld H, Henderson B, et al. Tackling Climate Change Through Livestock - A Global Assessment of Emissions and Mitigation Opportunities. Rome, Italy: Food and Agriculture Organization; 2013.

3. Audsley E, Brander M, Chatterton J, Murphy-Bokern D, Webster C, Williams A. How Low Can We Go? An Assessment of Greenhouse Gas Emissions from the UK Food System and the Scope for Reduction by 2050. Godalming, UK: Food Climate Research Network and World Wildlife Fund UK; 2010.

4. Macdiarmid JI. Is a healthy diet an environmentally sustainable diet? ProcNutr Soc. 2013;72(1):13-20.

5. Steinfeld H, Gerber P, Wassenaar T, Castel V, Rosales M, de Haan C. Livestock's Long Shadow - Environmental Isssues and Options. Rome, Italy: Food and Agriculture Organization; 2006.

6. Tukker A, Huppes G, Guinee J, et al. Environmental Impact of Products (EIPRO): Analysis of the Life Cycle Environmental Impacts Related to the Total Final Consumption of the EU 25. Seville, Spain: European Commission DG Joint Research Centre; 2006.

7. Carlsson-Kanyama A, González AD. Potential contributions of food consumption patterns to climate change. Am J Clin Nutr. 2009; 89(5 suppl):1704S-1709S.

8. Stehfest E, Bouwman L, van Vuuren DP, et al. Climate benefits of changing diet. Clim Chang. 2009;95(1-2):83-102.

9. Pan A, Sun Q, Bernstein AM, et al. Red meat consumption and mortality: Results from 2 prospective cohort studies. Arch Intern Med. 2012;172(7):555-563.

10. Scarborough P, Allender S, Clarke D, Wickramasinghe K, Rayner M. Modelling the health impact of environmentally sustainable dietary scenarios in the UK. Eur J Clin Nutr. 2012;66(6):710-715.

11. Macdiarmid JI, Kyle J, Horgan GW, et al. Sustainable diets for the future: Can we contribute to reducing greenhouse gas emissions by eating a healthy diet? Am J Clin Nutr. 2012;96(3):632-639.

12. Tukker A, Goldbohm RA, de Koning A, et al. Environmental impacts of changes to healthier diets in Europe. Ecol Econ. 2011;70(10): 1776-1788.

13. Millward DJ, Garnett T. Plenary Lecture 3: Food and the planet: Nutritional dilemmas of greenhouse gas emission reductions through reduced intakes of meat and dairy foods. Proc Nutr Soc. 2010;69(1):103-118.

ARTICLE IN PRESS

RESEARCH

14. Maillot M, Darmon N, Darmon M, Lafay L, Drewnowski A. Nutrient-dense food groups have high energy costs: An econometric approach to nutrient profiling. J Nutr. 2007;137(7):1815-1820.

15. Drewnowski A, Darmon N, Briend A. Replacing fats and sweets with vegetables and fruits—A question of cost. Am J Public Health. 2004;94(9):1555-1559.

16. Maillot M, Darmon N, Vieux F, Drewnowski A. Low energy density and high nutritional quality are each associated with higher diet costs in French adults. Am J Clin Nutr. 2007;86(3): 690-696.

17. Berners-Lee M, Hoolohan C, Cammack H, Hewitt CNN. The relative greenhouse gas impacts of realistic dietary choices. Energy Policy. 2012;43:184-190.

18. French food safety agency (Anses). Summary of the Individual and National Study on Food Consumption 2 (INCA2) 2006-2007. Maison-Alfort, France: Anses; 2009.

19. Dubuisson C, Lioret S, Touvier M, et al. Trends in food and nutritional intakes of French adults from 1999 to 2007: Results from the INCA surveys. Br J Nutr. 2010;103(7):1035-1048.

20. Goldberg GR, Black AE, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur J Clin Nutr. 1991;45(12):569-581.

21. Black AE. Critical evaluation of energy intake using the Goldberg cutoff for energy intake:basal metabolic rate. A practical guide to its calculation, use and limitations. Int J Obes Relat Metab Disord. 2000;24(9):1119-1130.

22. Vieux F, Soler L-G, Touazi D, Darmon N. High nutritional quality is not associated with low greenhouse gas emissions in self-selected diets of French adults. Am J Clin Nutr. 2013;97(3):569-583.

23. International Standard. ISO 14040:2006 Environmental Management -Life Cycle Assessment - Principles and Framework. Geneva, Switzerland: ISO; 2006.

24. International Standard. ISO 14044:2006 Environmental Management -Life Cycle Assessment - Requirements and Guidelines. Geneva, Switzerland: ISO; 2006.

25. National Institute of Statistics and Economic Studies. Definitions and methods - statistical operation: Survey on industrial energy consumption. 2013. http://www.insee.fr/en/methodes/default.asp?page= sources/ope-enq-conso-energie-industrie-eacei.htm. Accessed May 10, 2013.

26. French Department of Ecology, Sustainable Development, and Energy. SitraM database. 2013. http://www.statistiques.developpement-durable.gouv.fr/donnees-ligne/r/flux-marchandises-sitram-i.html?tx_ ttnews[tt_news]=20519&cHash=a891e4085d89a9486f97d0282957 ec1a. Accessed May 10,2013.

27. Althaus H, Doka G, Dones R, et al. Overview and Methodology—Data v2.0. Dubendorf, Switzerland: Ecoinvent; 2007.

28. Greenext. Commitments & methods [in French]. 2012. http://www. greencode-info.fr/. Accessed November 12, 2012.

29. Darmon N, Vieux F, Maillot M, Volatier JL, Martin A. Nutrient profiles discriminate between foods according to their contribution to nutritionally adequate diets: A validation study using linear programming and the SAIN: LIM system. Am J Clin Nutr. 2009;89(4): 1227-1236.

30. Maillot M, Ferguson EL, Drewnowski A, Darmon N. Nutrient profiling can help identify foods of good nutritional quality for their price: A validation study with linear programming. J Nutr. 2008;138(6): 1107-1113.

31. Kantar Worldpanel. French household consumer panel. 2013. http:// www.kantarworldpanel.com/global/Sectors. Accessed May 5, 2013.

32. Federal Reserve. H.10 Release—Foreign Exchange Rates—May 13, 2013. http://www.federalreserve.gov/releases/h10/hist/dat00_eu. htm. Accessed May 17, 2013.

33. US Department of Agriculture. National Nutrient Database for Standard Reference, Release 25.2012. http://ndb.nal.usda.gov/ndb/foods/ list. Accessed November 25, 2012.

34. Darmon N, Darmon M, Maillot M, Drewnowski A. A nutrient density standard for vegetables and fruits: Nutrients per calorie and nutrients per unit cost. J Am Diet Assoc. 2005;105(12):1881-1887.

35. Diet, Nutrition and the Prevention of Chronic Diseases. Report of a Joint WHO/FAO Expert Consultation. Geneva, Switzerland: World Health Organization; 2003.

36. Audsley E, Chatterton J, Graves A, et al. Food, Land and Greenhouse Gases. Cranfield, UK: Cranfield University; 2010.

37. Thompson S, Gower R, Darmon N, Vieux F, Murphy-Bokern D, Maillot M. A balance of healthy and sustainable food choices for France, Spain, and Sweden. Godalming, UK: World Wildlife Fund UK; 2013.

38. Scarborough P, Arambepola C, Kaur A, Bhatnagar P, Rayner M. Should nutrient profile models be "category specific" or "across-the-board"? A comparison of the two systems using diets of British adults. Eur J Clin Nutr. 2010;64(6):553-560.

39. Council Directive of 24 September 1990 on Nutrition Labelling for Foodstuffs (90/496/EEC). Brussels, Belgium: Council of the European Union; 1990.

40. 2011 Double Pyramid: Healhty Food for People, Sustainable for the Planet. Parma, Italy: Barilla Center for Food and Nutrition; 2011.

41. Page G, Ridoutt B, Bellotti B. Carbon and water footprint tradeoffs in fresh tomato production. J Clean Prod. 2012;32:219-226.

42. Pardo G, Zufia J. Life cycle assessment of food-preservation technologies. J Clean Prod. 2012;28:198-207.

43. Noponen MR, Edwards-Jones G, Haggar JP, Soto G, Attarzadeh N, Healey JR. Greenhouse gas emissions in coffee grown with differing input levels under conventional and organic management. Agric EcosystEnv. 2012;151:6-15.

AUTHOR INFORMATION

G. Masset and F. Vieux are research assistants, and N. Darmon is director of research, Unité Mixte de Recherche, Nutrition, Obesity, and Risk of Thrombosis, Institut national de la recherche agronomique, Institut National de la Santé et de la Recherche Médicale, Aix-Marseille Université, Marseille, France. L.-G. Soler is unit head, Unité de Recherche Alimentation et sciences sociales, Institut national de recherche agronomique, Ivry sur Seine, France.

Address correspondence to: Nicole Darmon, PhD, Unité Mixte de Recherche, Nutrition, Obesity, and Risk of Thrombosis, Institut national de la recherche agronomique 1260 INRA, Institut National de la Santé et de la Recherche Médicale 1062 INSERM, Aix-Marseille Université, F-13385, Marseille, France. E-mail: nicole.darmon@univ-amu.fr

STATEMENT OF POTENTIAL CONFLICT OF INTEREST

No potential conflict of interest was reported by the authors.

FUNDING/SUPPORT

This work was supported in part by the French Environment and Energy Management Agency (Ademe), the French National Institute for Agricultural Research (Insitut national de la recherche agronomique), and the French National Research Agency, under the Ocad project (ANR-11-ALID-0002).