Integrating Environmental and Nutritional Health Impacts Using Disability-Adjusted Life Years: Study Using the Ajinomoto Group Nutrient Profiling System Toward Healthy and Sustainable Japanese Dishes

Abstract

This study integrates the health impacts of environmental burdens and dietary intake using disability-adjusted life years (DALYs) to inform a healthier, more sustainable Japanese diet. Climate change, air pollution, ozone depletion, photochemical oxidants, and water consumption were quantified with Life cycle Impact assessment Method based on Endpoint modeling (LIME), while eleven dietary risks were converted to DALYs using dietary risk factors. Recipes collected online on a per-serving basis were classified into staple, main, side, and soup dishes and stratified into quartiles based on a nutrient profiling system (NPS) tailored to Japanese well-consumed dishes—the Ajinomoto Group NPS (ANPS) for dishes. ANPS—a culturally adapted NPS emphasizing protein, vegetables, sodium, and saturated fatty acids—was regressed against total DALYs to test whether higher ANPS scores correspond to lower combined health impacts of environment and diet. The analysis identified dish groups and high-scoring quartiles that minimized environmental and nutrition-related DALYs, revealing practical dish combinations that balance reduced sodium and red meat with increased vegetables, seafood, and nuts. These findings demonstrate the utility of coupling nutrient profiling with life cycle assessment (LCA) and provide a scientific basis for dietary guidelines that jointly advance human and planetary health within the emerging nutritional LCA framework.

1. Introduction

The Sustainable Development Goals, which were adopted in 2015 by the United Nations, urge the international community to act quickly, with numerous goals at the intersection of food, health, and environment, including eradicating hunger, ensuring healthy lives, promoting sustainable production and consumption patterns, and combating climate change [1]. In the background where these goals were established, a strong sense of urgency exists that global population growth and dietary changes are seriously impacting both the global environment and human health [2]. Food production, distribution, and consumption are major contributors to a variety of environmental problems, including global warming, water scarcity, and land degradation [3], while at the same time, undernutrition, overconsumption, and unbalanced diets increase the risk of lifestyle-related diseases such as heart disease, diabetes, and cancer, threatening the health of many people worldwide [2,3,4,5]. In short, what we “eat” is a critical issue that affects both the future of the Earth and our own health [6,7].

In light of this situation, disability-adjusted life years (DALYs) have attracted attention as an indicator for comprehensively assessing environmental and health impacts. This indicator is used to qualify disease burden by combining the Years of Life Lost (YLLs) due to early death and the Years Lived with Disability (YLDs), representing living with diseases and disabilities [8]. Therefore, DALYs are widely employed in various public health studies and environmental impact assessments because of their ability to comprehensively capture the magnitude of health damage on a year-by-year basis, which cannot be captured by the mere number of deaths or morbidity rate [5,9]. The Global Burden of Disease (GBD) study by the Institute for Health Metrics and Evaluation provides the most comprehensive data on the relationship between nutrient intake and health [4,5]. A comprehensive reanalysis of the GBD2017 showed that 15 dietary-related risk factors were estimated to have caused 11 million deaths (22% of adult deaths) and 255 million DALYs (15% of adult DALYs) in 2017, indicating that the overall dietary risk contributed to more deaths than any other risk factor, including smoking [4]. In particular, high sodium intake is the most influential risk factor in East Asia, including Japan, where the average daily sodium intake among adults is 3812 mg/day, approximately twice the World Health Organization (WHO) recommended amount of 2000 mg/day [10,11]. In 2024, the Life Cycle Initiative released dietary risk factors (DRFs) for detailed analysis of the relationship between food and nutrient intake and DALYs, which was developed based on the framework of Stylianou et al. [12] However, the GBD and DRFs are used with no consideration of environmental impacts to only analyze the relationship between nutrition and health.

To fill these gaps, a framework has recently been proposed to simultaneously assess both nutrition and environment by integrating a nutrient profiling system (NPS) and life cycle assessment (LCA) [13]. An NPS reflects community-specific dietary habits and health challenges, which are used to determine the nutritional value of foods based on the nutrients to be encouraged or limited. LCA, however, is a method used to quantify environmental burden from raw material procurement to disposal. This integration is expected to objectively support the transition to a healthy and sustainable diet. An NPS is primarily employed as an information disclosure tool for consumers in policy, with the EU’s Nutri-Score [14,15] and Australia’s and New Zealand’s Health Star Rating (HSR) [16] presenting the score per 100 g or 100 mL to consumers as a Front-of-Package (FoP) nutrition label. The Nutrient-Rich Food (NRF) index [17], developed in the United States, is also used to calculate continuous nutrient densities based on the amount of encouraged and limited nutrients per 100 kcal contained in a food. The NPSs developed in Japan include the Ajinomoto Group Nutrient Profiling System (ANPS) for dishes, which is tailored to Japan’s unique dietary culture and health challenges [18]. The ANPS evaluates the nutritional value of Japanese cuisine by encouraging the consumption of protein and vegetables and focusing on the limits of sodium and saturated fatty acids (SFAs). The advantages of this system are that it is designed to allow the sodium content, a significant health issue in Japan, to be highly detected and the nutritional value per dish to be assessed. In particular, Japanese traditional food culture, known as “Washoku (Japanese cuisine)”, was registered as a United Nations Educational, Scientific and Cultural Organization Intangible Cultural Heritage in 2013, and its basic composition is a combination of rice (the staple food), main dishes, side dishes, and soup. Therefore, it is a very important perspective for consumers who cook at home to evaluate Japanese food by the nutrients per dish rather than on a score per reference unit.

The objective of this study was to propose a healthy and sustainable Japanese diet by comprehensively assessing the impact of environmental burden and nutritional intake on health using DALYs and utilizing an NPS specific to Japanese food culture. Through this analysis, the study also aimed to verify whether improving the NPS score contributes to reducing the health impact of environmental factors to build a foundation for proposing healthy and sustainable diets adapted to Japanese food culture.

2. Materials and Methods

2.1. Data Collection and Preprocessing

2.1.1. Target of Study

In this study, we collected information on 12,438 recipes from the food database on the Ajinomoto Park Recipe Encyclopedia, a recipe website operated by Ajinomoto [19], using web scraping with the Python library BeautifulSoup4 (version 4.12.3). The data included metadata such as ingredients and cooking procedures, nutrient information per serving, cooking genre (e.g., Japanese, Western, and Chinese), number of favorite entries, and hashtags. Each dish was evaluated using a typical serving size; thus, the functional unit was one dish (one serving). Recipes that could not be standardized to a single serving (e.g., those with “20 servings” or “easy-to-make portions” in the ingredient column) were excluded, and a total of 9429 recipes were included in the final analysis.

The system boundary was defined as encompassing the production of raw materials, food preparation, and the disposal of food waste generated during preparation. Cooking utensils and tableware were excluded from the assessment as their contribution per dish was difficult to quantify. Figure 1 illustrates the workflow for the environmental and nutritional health impact assessments conducted in this study. The calculation methods for each health impact are explained in the following sections.

2.1.2. Data Preprocessing

The collected recipe data were standardized and preprocessed prior to analysis. First, ingredient names were mapped to the 2538 items in the 8th revised and supplemented edition of the Standard Tables of Food Composition in Japan 2023 (STFC2023) [20]. These ingredients were assigned to their food group category name (category number), as shown in the major (category No. 18), medium (category No. 34), and minor (category No. 99) categories (

Table A1. Food group classification table based on food number [11].

Major Category	No.	Medium Category	No.	Minor Category	No.
Grains	1	Rice and processed products	1	Rise	1
				Processed rice products	2
		Wheat and processed products	2	Wheat flour	3
				Bread (excluding sweet bread)	4
				Sweet bread	5
				Udon, Chinese noodles	6
				Instant Chinese noodles	7
				Pasta	8
				Other wheat processed products	9
		Other grains Processed products	3	Soba and processed foods	10
				Corn and processed products	11
				Other grains	12
Potatoes and starches	2	Potatoes and processed products	4	Sweet potatoes and processed products	13
				Potatoes and processed products	14
				Other potatoes and processed products	15
		Starch and processed products	5	Starch and processed products	16
Sugar and sweeteners	3	Sugar and sweeteners	6	Sugar and sweeteners	17
Beans	4	Soybeans and processed products	7	Soybeans (whole), processed	18
				Tofu	19
				Fried tofu	20
				Natto (fermented soybeans)	21
				Other processed soybean products	22
		Other beans and processed products	8	Other processed bean products	23
Nuts and bolts	5	Nuts and bolts	9	Nuts and bolts	24
Vegetables	6	Green and yellow vegetables	10	Tomato	25
				Carrot	26
				Spinach	27
				Green pepper	28
				Other green and yellow vegetables	29
		Other Vegetables	11	Cabbage	30
				Cucumber	31
				Japanese radish	32
				Onion	33
				Chinese cabbage	34
				Other light-colored vegetables	35
		Vegetable juice	12	Vegetable juice	36
		Tsukemono	13	Leafy vegetable pickles	37
				Yellow pickled radish and other pickled vegetables	38
Fruits	7	Fruit	14	Strawberry	39
				Citrus fruits	40
				Banana	41
				Apple	42
				Other fresh fruits	43
		Jam	15	Jam	44
		Fruit juice and fruit juice beverages	16	Fruit juice and fruit juice beverages	45
Mushrooms	8	Mushrooms	17	Mushrooms	46
Seaweed	9	Seaweed	18	Seaweed	47
Seafood	10	Raw or fresh fish	19	Aji sardines	48
				Salmon, trout	49
				Sea bream, flounder	50
				Tuna, marlin	51
				Other fishes	52
				Shellfish	53
				Squid, octopus	54
				Shrimps, crabs	55
		Processed Seafood Products	20	Seafood (salted, dried)	56
				Seafood (canned)	57
				Seafood (tsukudani)	58
				Seafood (fish paste products)	59
				Fish, ham, sausage	60
Meat	11	Animal flesh	21	Beef	61
				Pork	62
				Ham, sausages	63
				Other meat	64
		Chicken meat	22	Chicken meat	65
				Other poultry meat	66
		Meat (offal)	23	Meat (offal)	67
		Other meats	24	Whale meat	68
				Other processed meat products	69
Egg	12	Eggs	25	Eggs	70
Milk	13	Milk and dairy products	26	(Cow’s) milk	71
				Cheese	72
				Fermented milk and lactic acid beverages	73
				Other dairy products	74
		Other dairy products	27	Other dairy products	75
Fats and oils	14	Fats and oils	28	Butter	76
				Margarine	77
				Vegetable fats	78
				Animal fat	79
				Other fats and oils	80
Confectioneries	15	Confectioneries	29	Japanese confectioneries	81
				Cakes and pastries	82
				Cookies	83
				Candies	84
				Other confectionery	85
Luxury beverages	16	Alcoholic beverage	30	Japanese rice wine	86
				Beer	87
				Western wine and other	88
		Other beverages of choice	31	Tea	89
				Coffee, cocoa	90
				Other beverages of choice	91
Seasonings and spices	17	Seasoning	32	Source	92
				Soy sauce	93
				Salt	94
				Mayonnaise	95
				Miso	96
				Other seasonings	97
		Spices and others	33	Spices and others	98
Cooked foods	18	Cooked foods	34	Cooked foods	99

).

Since ingredient quantities are described differently for each recipe, even for the same ingredients, units were standardized to allow them to be calculated. Specifically, for ingredient-specific expressions such as “2 tomatoes”, “1 horse mackerel”, and “1 tablespoon sugar”, we referred to the Ajinomoto website for standard amounts of ingredients and books for basic data to convert them to grams [21,22,23,24]. For milliliters (mL) of water, seasonings, etc., the specific gravity was considered to be 1 to convert it to grams (g). In the case of expressions with ambiguous amounts, such as “ moderate amount”, “a little”, and “a pinch”, the amount was treated uniformly as 0.5 g. Next, to systematically analyze recipes by ingredient, the ingredient with the highest weight among those used, excluding water and seasonings, was extracted as the main ingredient and classified into the corresponding food group.

2.2. Environmental Health Impact Assessment of Food Production and Consumption

2.2.1. Life Cycle Impact Assessment

In this study, we used the Inventory Database for Environmental Assessment version 3.4 (IDEAv3.4) [21], which was developed by the National Institute of Advanced Industrial Science and Technology in Japan, specifically for LCA, to estimate the life cycle inventory for environmental impact assessment of the entire life cycle from raw material production to cooking and disposal. Then, the environmental impact by life stage for each recipe was calculated using Equation (1).

$$I_{r,c}=I_{ingredient,r,c}+I_{cooking,r,c}+I_{waste,r,c}$$

(1)

where I_r,c is the environmental impact assessment result of impact category c in recipe r, ingredient is the raw material procurement stage, cooking is the cooking stage, and waste is the disposal stage. The calculation method of environmental impact by life stage is provided in Appendix B (Equations (A1)–(A21)).

2.2.2. Environmental Health Impact Assessment

The Life cycle Impact assessment Method based on Endpoint modeling 3 (LIME3), which was developed by Itsubo et al., is a representative method for assessing the health impacts of environmental burden in DALYs [22]. This study quantified the human health impact for five major environmental impact categories: climate change, ozone depletion, air pollution, photochemical oxidants, and water consumption, using existing assessment methods and previously reported health impact coefficients. Human health impacts for each impact category were converted to DALYs by multiplying the environmental burden for each impact category, as obtained in 2.2.1, by the damage coefficient for each impact category using LIME3 [22] (Equation (2)). LIME3 is capable of integrating multiple environmental impacts and assessing human health impacts associated with climate change, air pollution, water consumption, etc. The damage coefficients used in this study were based on the assumption that the climate scenario is the SSP2 scenario and the consuming country is Japan.

$$\mathrm{Health} {\mathrm{Impact}}_{\mathrm{environment},\mathrm{r}}=\sum _{\mathrm{r}}(I_{r,c}\times D{F}_c)$$

(2)

where DF_c is the damage assessment factor for impact category c. The details of the impact pathways for human health in each impact category are as follows:

• Climate change

Greenhouse gases (GHGs) (CO₂, CH₄, N₂O, etc.) emitted from fossil fuel combustion and agricultural activities accumulate in the atmosphere, causing increases in global average temperatures and frequent extreme weather events. Such climate changes have been identified as having the potential to increase the number of patients with heat stroke due to heat waves, increase the spread of infectious diseases (especially mosquito-borne diseases, etc.) due to changes in rainfall patterns, and increase the risk of malnutrition due to reduced crop yields.

• Ozone depletion

Ozone-depleting substances, such as fluorocarbons, halons, and N₂O, increase the amount of ultraviolet B at the Earth’s surface by reducing stratospheric ozone. Excessive ultraviolet B exposure increases the risk of developing skin cancer (melanoma and non-melanoma), cataracts, etc.

• Air pollution (fine particulate matter formation)

Air pollution is caused by pollutants, such as PM2.5 (fine particulate matter with particle size 2.5 µm or less), SO₂, and NOx, emitted from energy use, industrial processes, and transportation. PM2.5 greatly affects the respiratory and cardiovascular systems, and it is a known factor that increases the risk of lung cancer, chronic obstructive pulmonary disease, and myocardial infarction.

• Photochemical oxidants

Photochemical oxidants, such as ozone (O₃), produced near the Earth’s surface as a result of atmospheric reactions of NOx and non-methane volatile organic compounds under sunlight, can stimulate the respiratory system, thereby causing chronic airway inflammation and leading to a reduction in lung function and worsening of respiratory disease.

• Water consumption

Excessive consumption of freshwater for agricultural and domestic use leads to water shortages and deterioration of water quality, affecting the health of local residents. Water scarcity increases the risk of malnutrition due to reduced crop production and restrictions on livestock rearing, and increases hygiene-related diseases such as diarrheal diseases and parasitic infections when safe drinking water is not available.

2.3. Nutritional Health Impact Assessment by Food Intake

2.3.1. Estimation of Recipe-Specific Nutrient Content

In calculating the health impacts attributable to dietary intake, it was necessary to obtain the nutrient content for each recipe. Information on energy, sodium, protein, vegetable intake, carbohydrate, cholesterol, and fiber was available on the recipe website and thus known at the time of data acquisition. By contrast, fatty acids and calcium were not publicly available; these were estimated using the ingredient weights for each recipe converted in Section 2.1.2 and the nutrient composition per 100 g of edible portion for each food item reported in STFC2023 [20]. As STFC2023 provides, for some items, nutrient values before and after cooking, we used those cooking-condition–specific values when such data and cooking methods were available; otherwise, estimates were generally based on pre-cooking nutrient values.

2.3.2. Nutritional Health Impact Assessment

In this study, we used DRFs developed under the Life Cycle Initiative’s Global Guidance for Life Cycle Impact Assessment Indicators and Methods (GLAM) project to calculate health impacts by nutrient in DALYs [12,23]. GLAM DRFs were prepared based on the GBD2019 study [5] by expanding and generalizing the framework of Stylianou et al. [24], which provides DRFs for 15 major dietary risks in approximately 200 countries. For 11 of these 15 items, foods and nutrients (

Table 1. Correspondence table between dietary risk in GLAM and Japan’s NHNS.

Dietary Risk	Definition in Stylianou et al. [24]	Food Group Number and Nutrient Name in NHNS [11]	Associated Health Outcomes (Effect)	Effective Intake [4]	DRFs (DALYs/g-Intake)
Diet low in vegetables	Fresh, frozen, cooked, canned, or dried vegetables, excluding legumes, salted or pickled vegetables, juices, and starchy vegetables	25–35, 46, 47	Hemorrhagic stroke, IHD, IS	<360 g/day	−3.7 × 10−8
Diet low in calcium	Calcium from all sources	Calcium	Colorectal cancer	<1.25 g/day	−1.0 × 10−5
Diet low in fiber	Fiber from all sources	Fiber	IHD, colorectal cancer	<23.5 g/day	−1.4 × 10−6
Diet low in fruits	Fresh, frozen, cooked, canned, or dried, excluding fruit juices and salted or pickled fruits	39–43, 44	10 health outcomes 1	<250 g/day	−1.5 × 10−7
Diet low in legumes	Fresh, frozen, cooked, canned, or dried legumes	18, 23	IHD	<60 g/day	−1.1 × 10−7
Diet low in milk	Milk (including non-fat, low-fat, and full-fat milk) but excluding plant derivatives	71	Colorectal cancer	<435 g/day	−1.3 × 10−8
Diet high in processed meat	Meat preserved by smoking, curing, salting, or the addition of chemical preservatives	63	T2DM, IHD, colorectal cancer	>2 g/day	3.4 × 10−7
Diet high in red meat	Beef, pork, lamb, and goat, but excluding poultry, fish, eggs, and all processed meats	61, 62, 64	T2DM, colorectal cancer	>22.5 g/day	3.8 × 10−7
Diet low in nuts	Nut and seed foods	24	T2DM, IHD	<20.5 g/day	−1.8 × 10−6
Diet low in PUFA	Omega-6 fatty acids from all sources	n-6 fatty acids	IHD	<11% of total daily energy	−5.9 × 10−7
Diet high in sodium	Dietary sodium from all sources	Sodium	15 health outcomes 2	>3.5 g/day	1.2 × 10−5

IHD = ischemic heart disease, IS = ischemic stroke, T2DM = type 2 diabetes mellitus, other CVD = other cardiovascular and circulatory disease. ¹ The 10 health outcomes associated with fruits are T2DM; esophageal cancer; hemorrhagic stroke; IHD; IS; larynx cancer; lip and oral cavity cancer; nasopharynx cancer; other pharynx cancer; tracheal, bronchus, and lung cancer. ² The 15 health outcomes associated with sodium are aortic aneurysm, atrial fibrillation and flutter, cardiomyopathy and myocarditis, chronic kidney disease due to T2DM, chronic kidney disease due to glomerulonephritis, chronic kidney disease due to hypertension, chronic kidney disease due to other causes, endocarditis, hemorrhagic stroke, hypertensive heart disease, IHD, IS, other CVD, peripheral artery disease, rheumatic heart disease, stomach cancer.

) that were consistent in terms of their definition according to Japan’s National Health and Nutrition Survey (NHNS) were assessed. [11]. A positive coefficient indicates an increase in DALYs per gram of food/nutrient intake, i.e., a loss in healthy life expectancy, while a negative coefficient indicates a decrease in DALYs per gram of food/nutrient intake, i.e., a benefit from an increase in healthy life expectancy.

These coefficients were multiplied by the edible weight of nutrients per serving, which is the nutrient intake per recipe (excluding discards, as estimated in Section 2.1.2), to evaluate the nutritional health impact associated with nutrient intake per serving (Equation (3)).

$$\mathrm{Health} {\mathrm{Impact}}_{\mathrm{nutrient},\mathrm{r}}=\sum _r((Q_{r,i}-({Q_{r,i}{ \times d}_i))\times N}_{i,n}\times DR{F}_n)$$

(3)

where N_i,n is the content of nutrient n in food item i [g] and DRF_n is the dietary risk factor [DALYs/g-intake] in nutrient n.

2.4. Integrated Health Impact Assessment

Integrated health impact assessment was based on the sum of the results of the environmental health impact assessment and the nutritional health impact assessment, which is obtained by Equation (4).

$$\mathrm{Health} {\mathrm{Impact}}_{\mathrm{integrated},\mathrm{r}}=\mathrm{Health} {\mathrm{Impact}}_{\mathrm{environment},\mathrm{r}}+\mathrm{Health} {\mathrm{Impact}}_{\mathrm{nutrient},\mathrm{r}}$$

(4)

where Health Impact_{environment,r} represents the environmental health impact in recipe r calculated by Equation (5) and Health Impact_nutrient,r represents the nutritional health impact in recipe r calculated by Equation (6).

2.5. Dish Scoring Based on Nutrient Profiling System

In this study, the ANPS for dishes was adopted as a nutritional evaluation index for each dish [18]. The ANPS for dishes, which is based on the assumption of Japanese food culture, is designed to comprise 4 major categories (staple food, main dish, side dish, and soup) and further subdivide them into 13 food categories, with a target nutritional value (target intake/upper limit) per dish set for each category. Target values were established for four nutritional indices, with protein and vegetable intake recognized as “encouraged” and SFAs and sodium recognized as “limited”, by comparing the Dietary Reference Intakes (DRIs) for Japanese (2020 edition) and actual measurements from the National Health and Nutrition Survey [11,25].

Category-specific targets were grounded in the following daily reference values: protein, 66 g/day; vegetables, 350 g/day; SFAs, 22.2 g/day; and sodium, 2756 mg/day (approximately 7 g salt). The sodium target is higher than the WHO recommendation of 2000 mg/day (approximately 5 g salt) because it was set as a realistic and achievable goal within the current Japanese dietary culture. Moreover, because sodium intake in Japan is comparatively high and about 62% of intake derives from seasonings and condiments, sodium was assigned higher scoring sensitivity within ANPS. The scoring was based on the following steps:

• The contents of four indicators were extracted for each dish.
• Scores for encouragement items were 10 points above the target value, with 1 point deducted for each 10% shortfall thereafter (up to 0 points).
• Scores for limitation items were 10 points below the target value, with 1 point deducted for each 10% excess thereafter (up to 0 points). However, for sodium, for which the deviation between the Japanese standard recipe and the target value was large, points were deducted by 0.5 points per 10% excess to increase the sensitivity.
• The scores for the four indicators were totaled, and the total was converted to an ANPS score of 0–100 by multiplying the full score of 40 by 2.5.

This symmetrical design (+2 items/−2 items, equally weighted) and the adoption of the one-dish criterion make the ANPS suitable for the integrated evaluation of this study because it can simultaneously assess the excess salt and vegetable deficiency characteristic of Japanese food and can be easily connected to the health impact assessment at the dish level.

2.6. Statistical Analysis

For statistical analysis in this study, Python 3.11.9 was run on Visual Studio Code (version 1.96.2), Pandas 2.2.3 and Numpy 2.1.3 were used for data processing, and the Statsmodels 0.14.4 library was used for statistical analysis. First, each recipe was classified into four food categories based on

Table 2. Category classification based on ingredients and nutrient composition of dishes.

Major Dish Group	Subcategory Number	Number of Dishes	Characteristics	Dish Examples
Staple dish	1	84	Staple foods with simple seasonings only, OR energy < 400 kcal, AND protein < 6 g, AND vegetables < 50 g	Plain rice, plain bread
	2	559	Energy < 400 kcal AND with other ingredients	Steamed rice (mixed) with red beans, hamburger
	3	144	Energy < 400 kcal AND soup AND with other ingredients	Udon noodles with soup, soba noodles with soup
	4	1334	Energy < 400 kcal AND with other ingredients	Curry rice, chicken-and-egg bowl
	5	131	Energy ≥ 400 kcal AND soup AND with other ingredients	Ramen noodles
Main dish	6	1301	Total ingredients ≥ 120 g AND protein ≥ 6 g	Grilled fish, beef steak, and grilled chicken
	7	1901	Total ingredients ≥ 120 g AND protein ≥ 6 g AND vegetables ≥ 50 g	Vegetable stir fry
	8	570	Total ingredients ≥ 120 g AND soup AND protein ≥ 6 g AND vegetables ≥ 50 g	Japanese hot pot
Soup	9	338	Soup AND protein < 6 g AND vegetables < 50 g	Tofu miso soup
	10	285	Soup AND protein < 6 g AND vegetables ≥ 50 g	Minestrone soup
	11	646	Soup AND protein ≥ 6 g	Pork and vegetable miso soup
Side dish	12	1816	Dishes that do not fall under category 1–11 AND (protein ≥ 6 g OR vegetables ≥ 50 g OR energy ≥ 100 kcal)	Boiled spinach seasoned with soy sauce
	13	325	Dishes that do not fall under category 1–12	Pickles

Partially revised table according to [18].

: staple, main dish, side dish, and soup. Then, Pearson’s product-moment correlation coefficient r was calculated for all recipes, excluding missing values, in order to determine the relationship between environmental health impacts and nutritional health indicators for each major category classified in the data preprocessing. The significance of the correlations was assessed by a two-tailed test with a theoretical t-distribution (α = 0.05), and the significance level was determined as p < 0.05. The sample was also divided into quartiles (Q1–Q4) based on the ANPS score value for each cuisine category. A single regression model was then constructed using the quartiles as explanatory variables and each nutrient and food intake, as well as the inventory and environmental impact assessment results by impact category as objective variables, and the presence or absence of a linear trend was tested using the regression coefficient and p-value (significance level of p < 0.05) for quartile numbers (1–4).

In the following sections, the results are presented for the 3772 recipes representing the “main dish” category. This is because “main dish” is the most nutritionally balanced and substitutable category in the “menu”, a traditional Japanese food culture form. The analysis is limited to the main ingredient category that accounts for more than 5% of the total number of recipes in each quartile. Meat was evaluated by subdividing it into three subcategories: beef, pork, and chicken.

3. Results

3.1. Comparison of the Health Impact of Environmental Burden and Nutritional Intake by Food Group

Analysis of the relationship between environmental health impact (DALYs/day) on the horizontal axis and nutritional health impact (DALYs/day) on the vertical axis in 3772 main dishes (Figure 2) revealed that food-group-specific clusters were formed, with recipes with beef as the main ingredient showing particularly high values for both indices. However, recipes with chicken and seafood as the main ingredients were concentrated near the origin, confirming that both environmental and nutritional impacts are small. The correlation coefficients’ explanatory power values (R²) by food group were as follows: vegetables, r = 0.37 (R² = 0.14); seafood, r = 0.05 (R² = 0); chicken, r = −0.12 (R² = 0.02); beef, r = 0.73 (R² = 0.53); and pork, r = 0.32 (R² = 0.1). A strong correlation was found between increased environmental-impact-related DALYs and increased nutritional risk for beef, while the other food groups showed weak or no correlation. Most environmental-impact-related DALYs for beef main recipes were greater than 75% of the total, which is consistent with reports that the environmental impacts associated with beef production are significantly greater than for other foods [26,27]. The study also showed that beef and pork main recipes commonly tended to have a higher nutritional-impact-related DALYs. In contrast, chicken and seafood main recipes had relatively low environmental-impact-related DALYs. Average environmental impacts of industrially produced chicken and seafood are reportedly similar; however, the range of environmental-impact-related DALYs for seafood may be underestimated because wild or farmed fish and fishing methods are not taken into account, necessitating careful interpretation [28]. Vegetable-based recipes showed the widest distribution of both environmental- and nutritional-impact-related DALYs, which may be explained by the fact that although the classification based on weight used is vegetable-based, small amounts of meat are often included.

3.2. Results of the Quantile Analysis Based on NPS Scores

Based on the ANPS score in the main dish category, the recipe dataset was divided into four quartiles (Q1: 15–66.25 points, Q2: 67.5–75 points, Q3: 76.25–87.5 points, Q4: 88.75–100 points) for analysis. These quartile groups included 1039, 855, 995, and 883 main dish recipes in Q1, Q2, Q3, and Q4, respectively, with mean ANPS scores distributed from 52.3 ± 8.3 in Q1 to 89.8 ± 4.5 in Q4 (

Table 3. Results of quartile analysis of environmental and health impacts based on NPS score in main dish.

	Q1			Q2			Q3			Q4
	(n = 1039)			(n = 855)			(n = 995)			(n = 883)
	Mean	±	SD	Mean	±	SD	Mean	±	SD	Mean	±	SD	p for Trend
ANPS score	52.3	±	8.3	69	±	3	77.9	±	2.7	89.8	±	4.5
Environmental Impact
Climate Change (kg CO2-eq)	7.3 × 10−1	±	7.1 × 10−1	5.6 × 10−1	±	5.0 × 10−1	5.8 × 10−1	±	5.6 × 10−1	5.0 × 10−1	±	3.9 × 10−1	<0.001
Ozone Depletion (kg-CFC-11eq)	7.8 × 10−8	±	1.2 × 10−7	5.3 × 10−8	±	5.7 × 10−8	4.7 × 10−8	±	4.9 × 10−8	4.2 × 10−8	±	5.1 × 10−8	<0.001
Air Pollution (kg-SO2eq)	2.4 × 10−4	±	1.7 × 10−4	2.1 × 10−4	±	1.4 × 10−4	2.2 × 10−4	±	1.5 × 10−4	2.1 × 10−4	±	1.4 × 10−4	0.003
Photochemical Oxidants (kg-C2H4eq)	3.2 × 10−5	±	7.7 × 10−5	3.8 × 10−5	±	8.3 × 10−5	4.3 × 10−5	±	9.2 × 10−5	4.4 × 10−5	±	9.7 × 10−5	0.001
Water Consumption (m3)	1.9 × 10−1	±	1.4 × 10−1	1.5 × 10−1	±	1.1 × 10−1	1.5 × 10−1	±	1.2 × 10−1	1.4 × 10−1	±	9.8 × 10−1	<0.001
Health Impact (DALYs)
Vegetables	−2.2 × 10−6	±	2.4 × 10−6	−2.4 × 10−6	±	2.2 × 10−6	−2.8 × 10−6	±	2.2 × 10−6	−2.8 × 10−6	±	2.2 × 10−6	<0.001
Fruits	−2.0 × 10−7	±	1.1 × 10−6	−2.4 × 10−7	±	1.4 × 10−6	−1.5 × 10−7	±	9.0 × 10−7	−1.5 × 10−7	±	7.2 × 10−7	0.137
Legumes	−1.0 × 10−7	±	5.8 × 10−7	−1.5 × 10−7	±	7.4 × 10−7	−1.9 × 10−7	±	8.7 × 10−7	−2.6 × 10−7	±	1.2 × 10−6	<0.001
Milk	−9.0 × 10−7	±	8.0 × 10−7	−6.6 × 10−7	±	5.2 × 10−7	−5.1 × 10−7	±	6.0 × 10−7	−4.9 × 10−7	±	4.5 × 10−7
Processed Meat	1.2 × 10−5	±	1.0 × 10−5	8.2 × 10−6	±	4.4 × 10−6	8.2 × 10−6	±	5.4 × 10−6	8.2 × 10−6	±	7.5 × 10−6
Red Meat	1.5 × 10−5	±	1.9 × 10−5	9.1 × 10−6	±	1.4 × 10−5	8.5 × 10−6	±	1.4 × 10−5	7.1 × 10−6	±	1.2 × 10−5	<0.001
Nuts	−8.7 × 10−6	±	1.2 × 10−5	−1.2 × 10−5	±	1.4 × 10−5	−8.9 × 10−6	±	1.1 × 10−5	−9.7 × 10−6	±	1.6 × 10−5
Calcium	−8.6 × 10−6	±	1.1 × 10−6	−8.5 × 10−7	±	1.2 × 10−6	−8.4 × 10−7	±	1.1 × 10−6	−7.8 × 10−7	±	8.5 × 10−7	0.118
Fiber	−4.9 × 10−6	±	6.4 × 10−6	−4.9 × 10−6	±	5.2 × 10−6	−5.1 × 10−6	±	5.5 × 10−6	−4.6 × 10−6	±	4.2 × 10−6	0.336
PUFAs	−1.9 × 10−6	±	1.4 × 10−6	−1.7 × 10−6	±	1.5 × 10−6	−1.6 × 10−6	±	1.4 × 10−6	−1.5 × 10−6	±	1.4 × 10−6	<0.001
Sodium	1.5 × 10−5	±	4.3 × 10−5	1.2 × 10−5	±	7.3 × 10−6	1.1 × 10−5	±	9.1 × 10−6	7.1 × 10−6	±	5.3 × 10−6	<0.001
Total Nutritional Health Impact	2.0 × 10−5	±	4.8 × 10−5	9.8 × 10−6	±	1.7 × 10−5	8.6 × 10−6	±	1.7 × 10−5	3.6 × 10−6	±	1.5 × 10−5	<0.001
Climate Change	1.1 × 10−6	±	1.0 × 10−6	8.2 × 10−7	±	7.3 × 10−7	8.6 × 10−7	±	8.3 × 10−7	7.4 × 10−7	±	5.8 × 10−7	<0.001
Ozone Depletion	7.2 × 10−11	±	1.1 × 10−10	4.9 × 10−11	±	5.3 × 10−11	4.4 × 10−11	±	4.5 × 10−11	3.8 × 10−11	±	4.7 × 10−11	<0.001
Air Pollution	1.6 × 10−8	±	1.2 × 10−8	1.4 × 10−8	±	9.8 × 10−9	1.5 × 10−8	±	1.0 × 10−8	1.4 × 10−8	±	9.7 × 10−9	0.003
Photochemical Oxidants	5.9 × 10−11	±	1.4 × 10−10	7.0 × 10−11	±	1.5 × 10−10	7.9 × 10−11	±	1.7 × 10−10	8.1 × 10−11	±	1.8 × 10−10	0.001
Water Consumption	9.0 × 10−8	±	6.5 × 10−8	7.4 × 10−8	±	5.4 × 10−8	7.4 × 10−8	±	5.8 × 10−8	6.5 × 10−8	±	4.7 × 10−8	<0.001
Total Environmental Health Impact	1.2 × 10−6	±	1.1 × 10−6	9.1 × 10−7	±	7.8 × 10−7	9.5 × 10−7	±	8.8 × 10−7	8.2 × 10−7	±	6.1 × 10−7	<0.001
Integrated Health Impact	2.1 × 10−5	±	4.8 × 10−5	1.1 × 10−5	±	1.8 × 10−5	9.6 × 10−6	±	1.8 × 10−5	4.5 × 10−6	±	1.6 × 10−5	<0.001

). A trend test showed that as the quartile of recipes increased from 52.3 ± 8.3 points in Q1 to 89.8 ± 4.5 points in Q4, climate-change impacts decreased by 31%, and ozone depletion and water consumption impacts also decreased significantly (−47% and −27%, both p < 0.001). However, photochemical oxidant impacts increased significantly by 30% with an increase in seafood and plant foods (both p < 0.05). Nutritional health impacts showed a significant increase in health benefits from vegetable and legume consumption (+27% and +154%, respectively), while benefits from polyunsaturated fatty acid (PUFA) consumption decreased significantly by 22%. However, health damages from red meat and salt intake decreased significantly (−51% and −54%, both p < 0.001). Taking environmental and nutritional health impacts together, they decreased by 30% and 82%, respectively, and the integrated environmental and nutritional health impacts also decreased significantly by 79% (from 2.1 × 10⁻⁵ DALYs/plate to 4.5 × 10⁻⁶ DALYs/plate).

As shown in

Table 4. Results of the quartile analysis of the amount of main ingredients and nutrients based on the NPS score in the main dish.

	Q1			Q2			Q3			Q4
	(n = 1039)			(n = 855)			(n = 995)			(n = 883)
	Mean	±	SD	Mean	±	SD	Mean	±	SD	Mean	±	SD	p for Trend
ANPS score	52.3	±	8.3	69	±	3	77.9	±	2.7	89.8	±	4.5
Ingredient Information
Grains (g) *	3.4	±	12.9	4.1	±	17.3	4.8	±	19.0	3.3	±	13.4	0.725
Legumes (g) *	10.0	±	30.0	11.1	±	31.0	11.7	±	31.9	12.9	±	33.6	0.04
Meats (g)	63.5	±	55.6	44.0	±	45.9	41.9	±	44.3	38.4	±	38.9	<0.001
Vegetables (g) *	26.1	±	30.0	29.2	±	28.2	31.5	±	25.4	31.5	±	25.2	<0.001
Potatoes (g)	17.3	±	43.9	13.0	±	32.1	12.7	±	35.5	8.0	±	24.2	<0.001
Eggs (g)	7.4	±	17.7	5.8	±	14.7	5.0	±	13.2	4.8	±	13.1	<0.001
Milks (g) *	7.1	±	23.5	3.1	±	14.4	2.4	±	12.4	2.4	±	11.4	<0.001
Seafood (g)	10.2	±	26.2	14.8	±	29.0	19.4	±	32.9	21.7	±	34.2	<0.001
Mushrooms (g)	4.5	±	11.0	4.2	±	9.9	4.3	±	10.3	3.8	±	9.4	0.202
Fruits (g) *	1.3	±	7.2	1.4	±	8.0	0.9	±	5.6	1.0	±	4.6	0.116
Seaweeds (g)	0.4	±	3.5	0.5	±	4.3	0.2	±	2.1	0.1	±	1.4	0.02
Seeds (g)	0.3	±	2.0	0.5	±	2.8	0.3	±	2.0	0.4	±	2.9	0.631
Sweeteners (g)	1.4	±	3.4	1.2	±	2.9	0.8	±	2.1	0.4	±	1.5	<0.001
Vegetables (g) **	59.2	±	66.0	65.2	±	58.9	75.5	±	60.1	74.9	±	59.3	<0.001
Fruits (g) **	1.3	±	7.2	1.6	±	9.1	1.0	±	6.0	1.0	±	4.8	0.137
Legumes (g) **	0.9	±	5.3	1.4	±	6.7	1.7	±	7.9	2.3	±	10.7	<0.001
Milk (g) **	69.6	±	61.5	50.8	±	40.0	39.2	±	46.5	37.4	±	34.5
Processed Meat (g)	36.6	±	29.4	24.1	±	12.9	24.2	±	15.9	24.0	±	22.0
Red Meat (g)	38.4	±	49.4	24.0	±	38.0	22.3	±	36.0	18.8	±	31.1	<0.001
Nuts (g)	4.8	±	6.6	6.7	±	7.8	4.9	±	6.4	5.4	±	8.8
Nutrient Information
Energy (kcal)	360.7	±	176.2	267.7	±	124.2	257.1	±	113.2	233.4	±	94.3	<0.001
Protein (g)	18.5	±	9.2	16.7	±	7.7	17.1	±	7.3	17.0	±	5.8	<0.001
Salt (g)	2.4	±	1.1	2.0	±	1.1	1.9	±	1.0	1.2	±	0.6	<0.001
Saturated fat (g)	7.0	±	5.9	3.2	±	3.1	2.1	±	1.9	1.8	±	1.4	<0.001
Calcium (g)	0.1	±	0.1	0.1	±	0.1	0.1	±	0.1	0.1	±	0.1	0.118
Fiber (g)	3.5	±	4.6	3.5	±	3.7	3.6	±	3.9	3.3	±	3.0	0.336
Polyunsaturated Fatty Acids (PUFAs) (g)	3.2	±	2.4	2.9	±	2.6	2.7	±	2.4	2.5	±	2.4	<0.001
Sodium (g)	1.3	±	3.6	1.0	±	0.6	0.9	±	0.8	0.6	±	0.4	<0.001

* Content based on food group classification. ** Content based on GLAM’s definition of dietary risk.

, meat consumption decreased significantly by approximately 40% (51% for red meat) across quartiles, while vegetable intake, legumes, and seafood increased significantly (+27%, +155%, and +113%, respectively), accompanied by significant decreases in energy intake, SFAs, and salt intake (−35%, −75%, and −54%, respectively). These trends are consistent with previous studies reporting lower GHG emissions for higher scoring foods in other NPS such as Nutri-Score and NRF9.3, indicating that a higher nutritional profile can simultaneously achieve a lower environmental impact [27,29,30]. However, the present study was characterized by an increased impact of photochemical oxidants in response to an increase in legumes and seafood. Thus, attention needs to be paid to the rebound phenomenon of some environmental indicators associated with improved nutrition, as reported by Conrad et al. [31]

3.3. Comparative Analysis of Q1 and Q4 Groups by Main Ingredient

Figure 3 shows the distribution of (a) environmental health impacts (environmental DALYs), (b) nutritional health impacts (nutritional DALYs), and (c) integrated health impacts (integrated DALYs) for groups Q1 and Q4 by major food ingredient category, with recipes classified into quartiles based on ANPS scores.

• (a) Environmental DALYs

The median environmental health impact was significantly lower in the Q4 group than in the Q1 group for recipes based on vegetables, chicken, beef, and pork, especially for beef-based recipes (t = 5.28, p < 0.001) and pork-based recipes (t = 7.22, p < 0.001). This may have been due to the reduced amount of animal ingredients in the Q4 group recipes and the complementary use of vegetables and seeds, thereby reducing the environmental impact. However, the difference between the Q1 and Q4 groups was small for the fish- and chicken-based recipes, and there was no significant difference between the Q1 and Q4 groups for the fish-based recipes (p = 0.19).

• (b) Nutritional DALYs

The nutritional health impact was significantly lower in the Q4 group than in the other groups for vegetable-, beef-, pork-, and chicken-based recipes. In particular, the difference between the Q1 and Q4 groups was particularly large for vegetable-based recipes (t = 12.38, p < 0.001), indicating a significant benefit from vegetable intake as a result of reduced salt and red meat use while maintaining or increasing vegetable intake in the Q4 recipes. Beef-, pork-, and chicken-based recipes also showed a significant decrease in the Q4 group, with a significant decrease in nutritional DALYs. Seafood-based recipes were not significantly different (p = 0.10), although the median health impact value in the Q4 group tended to be lower.

• (c) Integrated DALYs

In the integrated health impact, in which environmental and nutritional DALYs were taken together, the Q4 group had a statistically significant lower overall health impact than the Q1 group for recipes based on beef (t = 4.50, p < 0.001), pork (t = 5.19, p < 0.001), chicken (t = 5.12, p < 0.001), and vegetables (t = 12.29, p < 0.001). This supported the idea that the Q4 recipe had a lower overall health impact from both environmental and nutritional perspectives. The difference between the Q1 and Q4 groups was non-significant for the seafood-based recipes; however, the median value for the Q4 group was smaller than that for the Q1 group, suggesting potential room for improvement through better ingredient composition.

3.4. Feature Analysis of Representative Dishes in Q1 and Q4 Groups by Main Ingredient

Figure 4 visualizes the actual composition of the food groups used by extracting representative recipes, by main ingredient, that are located near the median of the Q1 and Q4 groups. For the vegetable-based recipe, Q1 contained the same combined amount of pork and eggs as vegetables, but Q4 showed a good protein balance by adding chicken. Overall, both Q1 and Q4 groups commonly showed a higher ratio of animal protein to the total amount. For the seafood main recipe, both Q1 and Q4 contained a combination of a small amount of vegetables and seafood, but the amount of seafood used differed. For the beef main recipe, Q1 contained small amounts of vegetables, and beef made up the majority of the ingredients, whereas in Q4, the proportion of meat was reduced by partially adding eggs, fruits, and seeds. For the pork main recipe, both Q1 and Q4 had a simple composition of only vegetables and pork, with Q4 showing a decrease in the total amount of ingredients and the amount of pork used. For the chicken main recipe, Q1 was a simple dish of only chicken, while Q4 added a small amount of vegetables to these ingredients. These recipes strongly suggest that the selection and distribution of sub-ingredients can vary widely among recipes, even in the same main ingredient category. This confirms that the variation observed in Figure 2 is derived from the overall diet variation.

These differences in food composition were also reflected in the integrated health impacts, as shown in Figure 5. For vegetable-, seafood-, and chicken-based recipes, an increase in plant foods like vegetables and legumes and a decrease in red meat and sodium were observed, especially in the Q4 group. These changes in nutrient intake offset the health damage, thereby reducing the health impact. For example, for Q4, the seafood main recipe, which is composed of potatoes and small amounts of vegetables, in addition to seafood, as also shown in Figure 4, the health benefits of dietary fiber and calcium intake work, while the health impact of salt intake can be significantly avoided by reducing salt intake. In addition, in the chicken-based recipe, the combination of potatoes, vegetables, and mushrooms in Q4 increases the benefits, and the intake of red meat can be avoided, making it easier to control the negative nutritional factors compared to beef and pork. Even for Q4 recipes, although the health impact is somewhat improved by reducing the amount of beef or pork use and adjusting the sub food ingredients, this recipe category had a greater health impact compared to the other categories. However, the effect of seed reduction in Q4 for the beef main recipe is very distinctive, with the use of about 2 g of almonds reducing the health impact by approximately 12%.

In summary, the careful selection of sub-ingredients such as those shown in Q4, especially the active use of plant foods (vegetables, potatoes, legumes, mushrooms, etc.) and switching to chicken and seafood as animal food sources, is expected to reduce negative nutritional factors such as SFAs, sodium, and red meat, as well as to reduce environmental impact. In terms of environmental impact, it is expected to reduce the negative nutritional factors such as SFAs, sodium, and red meat and is likely to contribute to the extension of healthy life expectancy. However, recipes using large amounts of beef or pork often carry predominant health risks, suggesting that effective inclusion of dietary fiber and seeds may reduce health impacts.

4. Discussion

This study evaluated the integrated health impacts attributable to environmental burden and nutritional intake for 9429 recipes in Japan and characterized them by food group and NPS score quartile. The results showed that for the 3772 recipes falling into the main dish category, (i) recipes containing beef had significantly higher environmental and nutritional impacts; (ii) the higher the ANPS score was (Q4), the more predominantly reduced both environmental and nutritional health impacts were; and (iii) a decrease in red meat and sodium and an increase in plant foods such as vegetables and nuts contributed to the decrease in health impact. However, the methodology used in this study, to combine environmental and nutritional health impacts into a single integrated indicator, still requires careful consideration [32,33]. Approaches to quantifying both environmental and nutritional health impacts should be further discussed. Many methodologies for assessing nutritional-health-impact-related DALYs have been adopted using GBD dietary risks [9,24,34]. Here, we discuss environmental impact assessment methodologies and nutrition profiling methodologies that may substantially affect the direction of the results.

4.1. Sensitivity Analysis Using Different LCIA Methods

The environmental and nutritional contributions to the integrated DALYs obtained in this study showed that nutritional impacts were 10 to 20 times greater than environmental impacts; however, this balance is highly dependent on the DALY conversion factors employed. In this study, the health impacts were calculated using the damage assessment coefficients of LIME3 [35], as shown in

Table 5. Comparison of damage assessment coefficients for human health of different LCIA methods.

Impact Category (DALYs/Each Midpoint Category Impact)	LIME3	IMPACT World+	ReCiPe2016 (I)	ReCiPe2016 (H)	ReCiPe2016 (E)
Global warming (kg-CO2eq−1)	1.47 × 10−6	8.18 × 10−7	8.12 × 10−8	9.28 × 10−7	1.25 × 10−5
Ozone depletion (kg-CFC11 eq−1 )	9.22 × 10−4	1.76 × 10−3	2.37 × 10−4	5.31 × 10−4	1.34 × 10−3
Fine particulate matter formation (kg-PM2.5 eq−1)	5.22 × 10−4	1.20 × 10−3	6.29 × 10−4	6.29 × 10−4	6.29 × 10−4
Photochemical ozone formation (kg-NMVOC−1)	1.85 × 10−6	3.90 × 10−8	1.64 × 10−7	1.64 × 10−7	1.64 × 10−7
Water consumption (m3 −1 )	4.78 × 10−7	6.35 × 10−5 *	3.10 × 10−6	2.22 × 10−6	2.22 × 10−6
Ionizing radiation (kBq-Co-60 eq−1)	-	1.65 × 10−8	6.80 × 10−9	8.50 × 10−9	1.40 × 10−8
Toxicity (cancer) (kg-1,4-DCB emitted to urban air eq−1)	-	2.08 × 10−6	3.32 × 10−6	3.32 × 10−6	3.32 × 10−6
Toxicity (non-cancer) (kg-1,4-DCB emitted to urban air eq−1)	-	1.46 × 10−7	6.65 × 10−9	6.65 × 10−9	6.65 × 10−9

* Direct comparison with water consumption is unrealistic since it indicates water scarcity.

. However, ReCiPe2016 and IMPACT World+ are typically used for European and American populations, respectively, for the integrated assessment of environmental and nutritional impacts using DALYs [9,24,36,37]. As found from the comparison of damage coefficients (Table 5), the environmental DALYs derived in LIME3 can vary greatly depending on the approaches with different coefficients and populations. Walker et al. reported similar integrated environmental and nutritional impacts in their analysis using the Egalitarian scenario coefficients from ReCiPe2016 (hereafter referred to as ReCiPe2016(E)), suggesting that differences in coefficients directly affect the overall magnitude of health impacts [9].

First, regarding the climate-change coefficient, ReCiPe2016(E) is 1.25 × 10⁻⁵ DALYs/kg CO₂eq, which is approximately 8.5 times higher than LIME3. If ReCiPe2016(E) were applied in this study, environmental DALYs would have increased significantly in the main red meat recipe, thereby reducing the differences in integrated DALYs between Q1 and Q4. This suggests that the environmental impacts are highly likely to exceed the nutritional impact. As noted in group Q4 (Table 3), climate-change-derived DALYs account for approximately 16% of the total integrated DALYs. Thus, the contribution rate is expected to exceed 60% if the ReCipe2016(E) coefficient is applied here. On the contrary, if the Individualist scenario coefficient from IMPACT World+ or ReCiPe2016 (hereafter ReCiPe2016(I)) were used, the climate-change coefficient would be much smaller. Thus, in group Q4 in Table 3, the contribution rate of climate change drops to approximately 1% to 3%, and the nutritional DALY-dominant relationship is expected to be further emphasized.

For fine particulate matter production, LIME3 and ReCiPe2016 are relatively close at 5.22 × 10⁻⁴ DALYs/kg PM2.5eq and 6.29 × 10⁻⁴, respectively, while IMPACT World+ is about twice as large at 1.20 × 10⁻³. However, the proportion of integrated DALYs represented by air-pollution-derived DALYs is low at 0.1% to 0.3% in each quartile group, and the increase in environmental DALYs is expected to be approximately 1% to 3% even for regions with high air pollution.

When water consumption values, which are used as an indicator in LIME3 and ReCiPe2016, were compared, the coefficient in ReCiPe2016 was approximately 4.6 to 6.5 times larger than that in LIME3. Thus, when using the ReCiPe2016(I) coefficient, the contribution of climate change and water consumption to integrated DALYs may be reversed. Although strict comparisons are difficult because IMPACT World+ uses water scarcity as its indicator, the coefficient in IMPACT World+ is approximately 130 times greater than that in LIME3, which could change the evaluation of menus utilizing a lot of water-intensive ingredients such as rice and vegetables.

Furthermore, IMPACT World+ and ReCiPe2016 provide damage coefficients for ionizing radiation and toxicity (carcinogenicity and non-carcinogenicity), while LIME3 does not include these coefficients. Because of this, environmental DALYs may be higher for crops with high quantities of pesticides and chemical substances, processed foods, etc., and the ranking of integrated DALYs may change. Walker et al. reported that toxicity is the most dominant environmental health impact in the integrated health impact assessment [9]. If these impact categories were included, the environmental DALYs and integrated DALYs in Q4 would significantly increase. Therefore, both the choice of method selection and coefficient setting are critical parameters in determining not only the absolute value of the integrated DALYs but also the relative importance of environmental and nutritional impacts and policy prioritization. It is considered necessary to reduce the influence of methodological bias in policymaking and menu optimization; therefore, the benefits of dietary patterns to be proposed should be consistently judged by multiple methods (including sensitivity analysis), rather than using the results of a single LCIA method.

4.2. Sensitivity Analysis Using Different NPSs for Quantile Analysis

As shown in Table 4, the quartile analysis using ANPS scores showed that red meat, SFAs, and sodium intake significantly decreased in the higher-scoring group. At the same time, the proportion of fish, seafood, legumes, and vegetables increased. In line with this, environmental, nutritional, and integrated DALYs decreased by 30–82% from Q1 to Q4, indicating a consistent “simultaneous improvement in health and environment” for the main dish recipe. A similar recipe set has been reported as having criterion validity evidence [38]. Thus, based on three continuously updated NPSs (Nutri-Score [14,15], HSR [16], and NRF9.3 [17]), we discuss trends for nutrients in recipes in quartile analysis in terms of NPS based on differences in NPS design, as summarized in

Table 6. NPS comparison.

NPS	Major Countries of Use/Institutional Positioning	Encouraged Nutrients (Addition Factor)	Limited Nutrients (Reduction Factor)	Standard Unit	Calculation Logic	Output Format/Remarks
ANPS for dishes	Japan/company development	Protein and vegetable intake	Sodium, SFAs	Per dish	Encouraged items: add 10 points to achieve the target value; reduce 1 point for each 10% shortfall. Limited items: reduce 10 points for not achieving the target value; reduce 1 point for each 10% excess (for sodium only, 10% excess = 0.5 point reduction for high sensitivity). For each item, full 40 points × 2.5 = converted to 0–100 points	Consecutive scores on a scale of 0–100 (the higher the better). Conforms to Japanese food form (one soup and three dishes)
Nutri-Score	France and other EU countries/ FoP Label	FVNL 1, dietary fiber, protein	Energy, sodium, SFAs, total sugars	Per 100 g or 100 ml	Subtract addition points based on encouraged items (0–15) from reduction points based on limited items (0–40); rating on a 5-point scale (A-E) based on threshold values	5-color labels for A (dark green) to E (red); reduction-point-dominant design
HSR	Australia, New Zealand/ FoP Label	FVNL, dietary fiber, protein	Energy, sodium, SFAs, total sugars	Per 100 g or 100 mL	Subtract modifying points (P, V, F) based on limited items from baseline points based on encouraged items; rating on a 10-point scale with categorical thresholds	0.5 −5 (0.5 increments). Addition-point-dominant design
NRF9.3	US/research	Protein, dietary fiber, vitamins A/C/E, Ca, Fe, Mg, K	Sodium, SFAs, added sugars	Per 100 kcal	Subtract the sum of the percentages of 3 limited nutrients from %MRV4 from the sum of the percentages of 9 encouraged nutrients from %DV3 (up to 100% of each)	Continuous value (the higher the value, the better the nutrient density)

¹ FVNL: fruit, vegetable, nut, and legume content.

. Nonetheless, because the ANPS for dishes is culturally tailored to Japanese cuisine and operates on a per-dish basis, direct comparability with international front-of-pack profiling systems is limited. Accordingly, our cross-system discussion is presented as a qualitative, design-based benchmarking (Table 6) rather than a numerical re-scoring.

In interpreting quartile trends, differences in unit of analysis (per dish vs. per 100 g/100 kcal), the symmetry or dominance of penalty/bonus terms, and the definition of FVNL across systems should be considered. The quartile analysis using ANPS scores showed a tendency for lower ANPS scores in recipes utilizing red meats such as beef and pork. This may be explained by the fact that protein is included among the nutrients considered in the ANPS evaluation algorithm, but the SFAs in red meat and salt for seasoning are also contained as factors that reduce the points. However, all of the Nutri-Score, HSR, and NRF9.3 include protein as a scoring factor, placing meat foods in the middle to high ranks for these indicators [27,39,40,41]. In the fundamental algorithms of the Nutri-Score and HSR, limited nutrients and encouraged nutrients are treated symmetrically in the algorithm. Therefore, energy and sugars, which are point-reduction factors, decreased significantly from Q1 to Q4 (−35% and −70%, respectively), and fruit, vegetable, nuts, and legumes, which are point-addition factors, increased significantly (+27% and +155%). Thus, as these nutrients were additionally evaluated, they were anticipated to have a tendency for an increase in the interquartile range at the same level as or higher than the ANPS. Additionally, the Nutri-Score has a point-reduction-dominant structure, with a maximum point reduction of 40 points for limited nutrients and a maximum point addition of 15 points for encouraged nutrients; thus, the 54% reduction in sodium and the 75% reduction in SFAs (from Q1 to Q4 in Table 4) should be strongly reflected in the score. However, the HSR has a point-addition-dominant structure where the upper limit of modifying points based on the encouraged nutrients (10–38 points) is larger than the upper limit of baseline points based on the limited nutrients (10–30 points), although these points vary by category. Therefore, the scores of beef- and pork-based main recipes may improve because of the high protein content. Additionally, the quartile trend may become moderate. For the NRF9.3 evaluation algorithm, because “high nutrient density” is added on a 100-kcal basis, beef- and pork-based main recipes high in protein, iron, magnesium, and other micronutrients are expected to score relatively higher. Additionally, even if red meat is regarded as the main ingredient in the per-dish evaluation, plant-based ingredients may be incorporated simultaneously. Accordingly, when reanalyzing in the HSR and NRF9.3 quartiles, beef- and pork-based main recipes remained in the upper quartile, raising the possibility that the gradient between environmental and nutritional health impacts and integrated health impacts may be attenuated from Q1 to Q4, or these gradients may be reversed. Taken together, these findings indicate that the ANPS most effectively explains the relationship between score improvements and concurrent improvements in environmental and nutritional outcomes within Japanese dietary patterns.

4.3. Limitations and Future Prospects

It should be noted that the recipe data were obtained from a single web platform; therefore, the dataset may be subject to selection bias and might not fully represent the diversity of Japanese dietary patterns. Potential sources of bias include (i) editing choices and the use of brand-linked seasonings that may influence ingredient proportions, particularly sodium content, and (ii) discrepancies between posted portions and actual household serving sizes and consumption frequency. Accordingly, generalization to the broader population should be performed with caution. As a partial mitigation, analyses were stratified by main ingredient category and ANPS quartiles, and the interpretation emphasized the consistency of directional trends rather than absolute values. Additionally, because ANPS is culturally specific and portion-based, international comparability is constrained; future work should harmonize units (per 100 g/100 kcal) and component definitions to enable formal cross-system scoring.

Ingredients origin was not specified, and environmental impacts were evaluated assuming average production and distribution within Japan. However, Japan’s food self-sufficiency rate is about 38% on a calorie basis and 61% on a production value basis, and except for rice, most of its food ingredients are imported [42]. Because the environmental impact of foods varies markedly by production method [27,28], future impact assessments should consider product-specific origins and import destinations. Additionally, nutrient composition varies with cultivation method and seasonality [43,44,45]; however, this study did not consider the differences in nutrients contained by these factors.

As a prospect, it is expected that more detailed food intake data and analysis that takes into account the production areas of food ingredients will lead to more practical and accurate analysis, as well as the establishment of a sustainable dietary model in line with the Japanese food culture represented by “one soup, three dishes”. Furthermore, the ecological impact of land use for food production is a significant concern in the current food production system [2,3]. Future analysis should consider the trade-off between human health and ecological impact.

5. Conclusions

This study evaluated the integrated health impact of environmental burden and nutritional intake by examining the potential for healthy and sustainable dietary recommendations using ANPS, an NPS specific to the Japanese food culture. The findings clarified that, in pursuing a “healthy and sustainable diet”, it is important to avoid the excessive use of red and processed meats and salt seasonings, to replace animal protein sources with chicken and seafood, and to combine a variety of foods, mainly vegetables and legumes, whenever possible. This proposal partially supports the planetary health diet recommended by the EAT-LANCET committee, in which the intake of plant-based foods should be increased and the intake of red and processed meats should be minimized. The findings of this study are significant in that they demonstrate the possibility of using animal proteins such as chicken and seafood as protein sources by quantitatively evaluating environmental and nutritional health impacts using DALYs. Utilizing the knowledge obtained in this study could realize a diet that maximizes nutritional benefits while reducing environmental impact, thereby contributing to the extension of healthy life expectancy in total.