Modeling undertaken in 112 pregnant women (65.5%) consuming >1 serve of beef per week. AI, Adequate Intake; EAR, Estimated Average Requirement; NRV, Nutrient Reference Value.

### **4. Discussion**

The Food and Agriculture Organization of the United Nations and others has identified an urgent need to shift to environmentally sustainable diets on a global scale [1,2]. Concerns over nutrient adequacy in populations with high nutrient demands, such as pregnant women, are potential challenges to the broad implementation of environmentally sustainable diets. Our findings indicate these concerns are likely misplaced within the context of a mixed diet. Modeled replacement of animal-derived food with more environmentally sustainable plant-based alternatives has only a small effect on overall nutrient intake but a considerable positive effect on environmental sustainability.

There remain concerns among many practitioners and community members regarding the potential risk of nutrition inadequacy of pregnant women consuming plant-based diets [21,22]. Our focus was not plant-based diets but rather environmentally sustainable foods within the context of mixed diets. Focusing on mixed diets enables our findings to be relevant to a large proportion of the population for whom consuming a purely vegetarian or plant-based diet is neither practicable nor desirable [9,10].

In general, our results support animal-derived foods as a rich source of zinc, and plantbased foods as being rich in calcium, folate, and dietary fiber. A specific swap replacing one serve per week of beef with firm tofu reduces zinc and protein levels, while calcium, folate, and dietary fiber increases. The absolute differences in the nutrient intake of this swap were small. In this modeling, the largest differences were for calcium (raised by about 13% of a standard deviation), zinc (reduced by about 10% of a standard deviation), and fiber (raised by about 5% of a standard deviation), resulting in an increase in the proportion of pregnant women who meet NRVs of calcium, dietary fiber, and iron, but a decrease for zinc. Maternal zinc deficiency during pregnancy may increase the risk of low birth weight and small for gestational age infants [23], although severe zinc deficiency is

rare. Indeed, in our population, the majority of women met the NRVs for zinc, based on both actual and modeled intakes. Furthermore, zinc is a common ingredient in pregnancy multivitamins, which are used by approximately 70–80% of women in the USA, Europe, and Australia [7,24–26]. The amount of zinc in such multivitamins (typically 11 mg per day) [27] exceeds the NRV for zinc.

Alongside folate, the public is perhaps most aware of concerns regarding sufficient iron intake during pregnancy [3,7]. Plant-based foods are a good source of overall dietary iron, but this does not account for differences in the bioavailability of heme and non-heme iron. Heme iron is only found in animal-derived meat products. Heme iron constitutes approximately 40% of total iron from animal-derived meat products and it is more readily absorbed by humans than non-heme iron [3]. To account for this, we estimated the amount of absorbed iron. This estimation did not account for the increased absorption of non-heme iron during pregnancy [28], and as such is a conservative estimate of absorbed iron from plant-based foods. Nonetheless, both total dietary iron intake and estimated absorbed iron were slightly higher after replacing a serve of beef with an isoenergetic serve of firm tofu (Table S2), although the magnitude did not appear to be clinically meaningful on an individual basis. It is notable that the proportion of women meeting NRVs for iron intake by diet alone was low in our population (about 5%). This is consistent with other studies of pregnant women in Australia [7], and with dietary modeling undertaken as part of the development of the Australian Guide to Healthy Eating, in which no dietary models could provide sufficient iron to meet the needs of pregnant women [29].

The low prevalence of participants meeting the NRVs for folate, iron, and fiber, for both actual and modeled intakes, highlights the necessity of appropriately planned diets, by health professionals, such as dietitians or individuals with nutrition training, to fulfill the nutritional needs of women during pregnancy. The use of dietary supplements during pregnancy may at least partially alleviate these deficiencies, irrespective of the background diet. There is currently limited publicly available information concerning the environmental sustainability of pregnancy supplements.

Iron and zinc absorption can be affected by other factors. Within the context of a mixed diet, non-heme iron and zinc absorption can be enhanced by other components of the diet, including meat, poultry, fish, and other seafood [30,31], alongside vitamin C-rich foods, e.g., citrus fruits and green leafy vegetables [32]. Therefore, one way to implement a one-serve per week replacement within the context of a mixed diet whilst maintaining the effect of iron and zinc absorption enhancers (e.g., meat products, green leafy vegetables) would be to replace half the portion of a less environmentally sustainable animal-derived meat product with a more sustainable plant-based food twice per week.

The most environmentally sustainable alternatives produced approximately 98% less GHG emissions than one serve of beef when matched for energy or protein content. To facilitate a broader understanding of the impact of incorporating more environmentally sustainable foods, we compared GHG emissions generated in food production to those produced by typical passenger vehicle usage. Using the example above, replacing one serve of beef with an isoenergetic serve of firm tofu per week during pregnancy could reduce GHG emissions by the equivalent to those produced by a typical passenger vehicle driven for 1498 km. Similarly, most of the proposed environmentally sustainable alternatives have a lesser environmental impact when assessed by other measures of environmental sustainability (including land use, acidifying emissions, and eutrophying emissions), although water use in the production of some plant-based alternatives (e.g., legumes and beans) appears to be similar to that of some animal-derived foods.

When matching the energy and protein content to reference foods, the portion size of plant-based alternatives is two to four times heavier than animal-derived foods, consistent with the density of energy and protein being notably higher in animal-derived foods. Plantbased alternatives are rich in dietary fiber with lower energy density, which increase satiety, helping to maintain a healthy weight by limiting calorie intake [33,34], and optimize weight

gain during pregnancy [35,36]. Excessive protein intake from animal sources, primarily meat products, may also increase the risk of overweight and obesity in offspring [37].

Our study has a number of limitations. We used an FFQ administered in the immediate postpartum period. Women were asked to recall their habitual diet during pregnancy, which we validated against dietary biomarkers in a subgroup [38]. FFQs are well described as tools for assessing habitual diet over 6–12-month periods, although we cannot rule out that there may have been a greater emphasis on third trimester intake due to recency bias. Furthermore, detailed information of dietary intake (e.g., ingredients of mixed dishes) is difficult to assess but will include meat or other food items as an ingredient. As such, the values for meat and other food items will have been underestimated. Nonetheless, previous research has shown dietary patterns during pregnancy remain relatively stable when compared to pre-pregnancy intake [39–41] and are not significantly different to those of nonpregnant women of reproductive age [42]. Our study population from which we identified the commonly consumed food items during pregnancy has a relatively low prevalence of overweight and obesity (22%). This is less than in the general population in Western countries, where up to 50% of women have overweight or obesity before pregnancy [43,44], and is likely due at least in part to the exclusion of women with diabetes and preeclampsia from our analyses. Future research may seek to determine the environmental sustainability of foods consumed by representative samples of pregnant women. We mainly focused on GHG emissions given their contribution to global warming and did not describe the impact of environmentally sustainable foods on the economy and society. A range of indicators of economic and societal aspects [45] (e.g., affordability, employment, and food insecurity) can be used to assess the effects of improving environmental sustainability and should be the topic of future research. We acquired measures of environmental sustainability from a global dataset by Poore and Nemecek [13], consisting of data derived from 570 studies in 119 countries to ensure that our findings can be broadly generalizable. Future studies could employ country-specific measures of environmental sustainability to enable a more geographically accurate indication of the environmental impact of these food swaps. The role of food–food interactions that influence absorption was beyond the scope of our current study; however, future research should look to model these interactions within the context of dietary changes to promote environmental sustainability. Finally, to translate our findings into practice, the acceptability and popularity of proposed environmentally sustainable options need to be taken into consideration. Future research should identify whether there are unique challenges or opportunities for promoting environmentally sustainable foods during pregnancy. Nutrition communicators, dietitians, and practitioners may need to focus on the promotion of health benefits of environmentally sustainable plant-based foods, and provide practical advice (e.g., design recipes) in incorporating these replacements into their individual diets.

### **5. Conclusions**

Our research highlights simple dietary substitutions that can substantially reduce environmental impact without compromising essential nutrient intake during pregnancy. Moving forward, environmentally sustainable food replacements should be the focus of applied clinical research and inform nutrition practice and policy development.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/nu13103355/s1, Figure S1: Flow diagram of participant selection; Table S1: Dietary intake analysis of pregnant women in the Newborn Body Fatness study; Table S2: Food swaps—beef. Differences in nutrients and measures of environmental sustainability between beef and more sustainable isoenergetic options; Table S3: Food swaps—chicken. Differences in nutrients and measures of environmental sustainability between chicken and more environmentally sustainable isoenergetic options; Table S4: Food swaps—white fish. Differences in nutrients and measures of environmental sustainability between white fish and more environmentally sustainable isoenergetic options; Table S5: Food swaps—milk. Differences in nutrients and measures of environmental sustainability between milk and more environmentally sustainable isoenergetic options; Table S6: Nutrient analysis and measures

of environmental sustainability of the reference food (beef) and more environmentally sustainable protein-matched options; Table S7: Food swaps—beef. Differences in nutrients and measures of environmental sustainability between beef and more environmentally sustainable protein-matched options; Table S8: Food swaps—chicken. Differences in nutrients and measures of environmental sustainability between chicken and more environmentally sustainable protein-matched options; Table S9: Food swaps—white fish. Differences in nutrients and measures of environmental sustainability between white fish and more environmentally sustainable protein-matched options; Table S10: Food swaps—milk. Differences in nutrients and measures of environmental sustainability between milk and more environmentally sustainable protein-matched options.

**Author Contributions:** Conceptualization, M.R.S., T.W., S.B. and A.G.; Data curation, T.W.; Formal analysis, T.W.; Investigation, T.W. and H.U.D.; Methodology, M.R.S., T.W., S.B. and A.G.; Project administration, M.R.S. and T.W.; Resources, H.U.D. and M.R.S.; Software, T.W.; Supervision, M.R.S.; Validation, T.W., M.R.S. and A.G.; Visualization, T.W.; Writing—original draft preparation, T.W.; Writing—review and editing, T.W., M.R.S., A.G., S.B. and H.U.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was approved by the Ethics Committee of the Sydney Local Health District (HREC/14/RPAH/478, date of approval: 20 February 2015), with written informed consent provided by the participating mothers.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available for ethical reasons.

**Acknowledgments:** The authors thank Graham Giles, Cancer Epidemiology Centre, Cancer Council Victoria for permission to use the Dietary Questionnaire for Epidemiological Studies (Version 2).

**Conflicts of Interest:** Skilton is employed by The University of Sydney as the Maurice Blackmore Principal Research Fellow in Integrative Medicine. This position was established through a gift from the Blackmores Institute. Skilton receives no research funding or in kind support from Blackmores Limited. The authors have no other conflicts of interest to declare.

### **References**


### *Article* **Comparison of Plate Waste between Vegetarian and Meat-Containing Meals in a Hospital Setting: Environmental and Nutritional Considerations**

**Andrew Berardy 1,\*, Brianna Egan 1, Natasha Birchfield 2, Joan Sabaté <sup>1</sup> and Heidi Lynch <sup>3</sup>**


**Abstract:** Vegetarian diets can satisfy nutritional requirements and have lower environmental impacts than those containing meat. However, fruits and vegetables are wasted at higher rates than meat. Reducing both food waste (FW) and the environmental impacts associated with food production is an important sustainability goal. Therefore, the aim of this study was to examine potential tradeoffs between vegetarian meals' lower impacts but potentially higher FW compared to meat-containing meals. To examine this, seven consecutive days of plate FW data from Loma Linda University Medical Center (LLUMC) patients were collected and recorded from 471 meals. Mean total FW and associated greenhouse gas emissions (GHGE) were higher among meat-containing meals (293 g/plate, 604 g CO2-eq/plate) than vegetarian meals (259 g/plate, 357 g CO2-eq/plate) by 34 g (*p* = 0.05) and 240 g CO2-eq (*p* < 0.001), respectively. Statistically significant differences were observed in both FW and associated GHGE across major food categories, except fruit, when comparing vegetarian and meat-containing meals. Overall, vegetarian meals were preferable to meat-containing meals served at LLUMC both in terms of minimizing FW and lowering environmental impacts. Other institutions serving vegetarian meal options could expect similar advantages, especially in reduced GHGE due to the high CO2 embodied in meat.

**Keywords:** food waste; global warming; vegetarian meals; hospital setting; plant based; sustainability; public health

### **1. Introduction**

Human activities cause global environmental changes that threaten to disrupt the stability of the Earth's systems, leading to potentially disastrous consequences [1]. This recognition has prompted a widespread call for emergency action to limit global temperature increases, restore biodiversity, and protect health [2]. Food systems are responsible for between 19 and 37% of global anthropogenic greenhouse gas emissions (GHGE), depending on what is included in the estimate [3,4]. A recent estimate attributed 34% of global GHGE to food systems, with 71% coming from agriculture and land use, and the rest from downstream supply chain activities [5].

Yet, current practices of food production and distribution are insufficient, as there are 815 million people globally, or one in nine, who are undernourished [6]. In order to end hunger, different scenarios predict that between 3 and 20% more food production will be necessary, depending upon the approach, increasing the associated environmental impacts [7]. This challenge will only become more difficult as the global population continues to expand to approximately 9 billion people [8].

**Citation:** Berardy, A.; Egan, B.; Birchfield, N.; Sabaté, J.; Lynch, H. Comparison of Plate Waste between Vegetarian and Meat-Containing Meals in a Hospital Setting: Environmental and Nutritional Considerations. *Nutrients* **2022**, *14*, 1174. https://doi.org/10.3390/ nu14061174

Academic Editor: Lindsay Brown

Received: 5 January 2022 Accepted: 7 March 2022 Published: 11 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Reducing the consumption of animal-based foods is a possible measure to reduce environmental impacts while improving health outcomes, with the potential to reduce diet-related GHGE by between 33 and 51% in the United States [9,10]. A systematic review found that vegan diets could reduce GHGE by up to 70%, land use by up to 86%, and water use by up to 70% [11]. Another review found that along with improved health, shifting from current omnivorous dietary patterns to vegetarian or vegan diets increases environmental sustainability while also improving health [12].

The Academy of Nutrition and Dietetics considers appropriately planned vegetarian diets to be healthful and nutritionally adequate for all stages of the life cycle [13]. Consuming vegetarian or vegan diets has been shown to lower risk for developing obesity [14], cardiovascular diseases [15], hypertension [16], type 2 diabetes [17], and metabolic syndrome [18]. These health-protective effects may be due to the higher nutrient quality typical of plant-based diets [19]. Notably, vegetarian and vegan diets tend to be lower in total fat, saturated fat, monounsaturated fat, dietary cholesterol, protein, alcohol, and sodium, and higher in polyunsaturated fat, fiber, and iron [19]. This is likely because plant-based diets tend to be higher in fruits, greens, and pulses; subcategories of vegetables [19]. In addition, plant-based diets have been found to sufficiently support athletic performance while also contributing to better overall health and reducing environmental impacts [20–22].

Nonetheless, certain nutrients are less bioavailable or less frequently consumed on a vegetarian or vegan diet. For example, as non-heme iron (found in plants) is less bioavailable compared to heme iron (from animals), the Recommended Dietary Allowance for iron for vegetarians and vegans is 1.8-fold greater than that for omnivores [23]. Additionally, vegans (who exclude all animal products) must be mindful to consume foods fortified with vitamin B12 or take a vitamin B12 supplement as this vitamin is not present in plant foods [13]. Lacto-ovo vegetarians typically consume at least the recommended intake for calcium, while vegans may risk insufficiency. Furthermore, vitamin D is not abundant in food and is a nutrient for which the use of supplements is frequently advised, regardless of dietary pattern [24].

Higher diet quality, as measured by the Healthy Eating Index, is associated with higher food waste (FW), primarily in the form of fruits and vegetables [25]. FW is a significant challenge, as 32% of all food produced in the world by weight or 24% by kilocalories (kcal) is wasted [26]. If global FW were treated as its own country, it would be the third largest emitter of GHGE, behind China and the United States, occupy 30% of the world's agricultural land area, and use the equivalent water of the annual discharge of the Volga river in Russia (i.e., 250 km3) [27]. These FW statistics represent wasted resources and wasted opportunities to eat health-promoting foods, which comprise a large portion of total waste. In fact, the average global FW per capita per year could fulfill a person's dietary recommended intake (DRI) of 25 nutrients for 18 days [28]. Based on the types of food wasted, that amount of FW contains between 25 and 50% of the DRI for vitamin C, K, zinc, copper, manganese, and selenium for a person [28].

It is important to understand possible tradeoffs when promoting a solution to one problem to ensure it does not exacerbate another. For example, given the relatively high proportion of fruit and vegetable waste compared to meat waste, and the small proportion of vegetarians in the general population, could there be higher FW as a result of reducing meat-containing meals? Moreover, would the environmental impacts associated with that FW be substantial enough to negate the benefits of serving vegetarian meals as the default in large institutional settings? Although there are publications that assess hospital FW, its environmental impacts, and techniques for FW reduction, no literature has previously examined these questions [29–32].

Loma Linda University Medical Center (LLUMC) provides a unique setting to examine the potential tradeoffs associated with serving lower environmental impact foods with potentially higher FW compared to higher environmental impact foods with lower FW. Unlike many hospitals, LLUMC serves lacto-ovo vegetarian meals to patients by default for the first 24 h upon admission. However, patients have the option to reject the default

and can choose their preferred meal items from standard menus, which include meat, after 24 h. As such, the aim of this case study was to examine the differences in FW and GHGE between vegetarian meals and meat-containing meals served in a hospital setting.

### **2. Materials and Methods**

A plate waste audit was performed by Loma Linda University dietetics graduate students across seven consecutive days, from September 6 to 12, 2020 at LLUMC. Plates audited included those served at breakfast, lunch, and dinner and were provided by meal services on three hospital floors, which housed patients with the fewest special dietary orders (e.g., liquid diets or "nil per os" (NPO, nothing by mouth)). At least 20 plates were audited, upon tray return and prior to disposal, after each meal service. Each tray was assigned a de-identifiable number and meal type (i.e., "meat-containing" or "vegetarian") based on the food items listed on the tray ticket, which reflected the patient's menu order. Trays returned without tray tickets and no remaining meat items were categorized as unknown meal type. Floor number and diet order (regular or therapeutic) were also noted for each tray. Institutional review board approval was not needed since no patientidentifying information was collected. For each tray, all remaining individual food items were removed and individually weighed in grams before being discarded. Liquid diet trays were excluded from measurement due to the high proportion of total weight from liquid.

Data processing included removing container weight values from FW measured in containers. LLUMC also provided recipes for composite foods such as cooked entrees and soups, which were used to determine the proportional weights for individual ingredients (e.g., spinach, cheese, and egg white for spinach quiche). In addition to FW weight, GHGE were estimated using a combination of SimaPro life cycle assessment software and published literature used to fill any gaps where SimaPro did not have appropriate data [33–37]. The life cycle assessment (LCA) studies used for GHGE estimates all had cradle to farm or manufacturer gate system boundaries and reported results using a weightbased functional unit. The parameters of LCA studies included were system boundaries from cradle to farm or manufacturer gate or distributor, excluding retail, consumption, and disposal, and used a weight-based functional unit and attributional assumptions.

Descriptive and statistical analyses were conducted using IBM SPSS Statistics for Windows, version 28 (IBM Corp., Armonk, NY, USA). Tests for assumptions of normality and homogeneity of variance were performed using Kolmogorov–Smirnov and Levene's test, respectively. To examine between-group differences in total FW and total GHGE, independent *t*-tests were performed. Values that were ±2.5 standard deviations from the mean were considered outliers. Visual assessment using boxplots indicated that there were no outliers, defined as values that were ±2.5 standard deviations from the mean. Post hoc exploratory analyses were also conducted using independent t-tests to compare between-group differences for FW and GHGE by primary food categories. The exploratory analyses were considered secondary analyses, which were not driven by hypothesis testing; therefore, the significance level was not adjusted for multiple comparisons. In addition, effect size was calculated using Hedges' g for all primary and secondary outcomes. Data are reported as the mean ± standard deviation and the level of statistical significance was set at *p* = 0.05.

### **3. Results**

Plate data were analyzed for 447 patient trays of the 471 that were collected. Twentyfour patient trays were excluded from analysis due to unknown meal type and absence of identifying characteristics (e.g., leftover meat or a tray ticket). Key findings of this study were that mean total plate waste was higher among meat-containing meals, and that the associated GHGE was lower among vegetarian meals.

### *3.1. Food Waste*

3.1.1. Descriptive Statistics

The corresponding means and standard error of mean are presented graphically in Figure 1. The data for total FW were not normally distributed for either group (*p* < 0.05). Skewness of variables prevented transformation to a normal distribution. Non-parametric tests, such as the Mann–Whitney U test, resulted in unacceptable values (U > 10,000) and are most appropriate for analyzing ordinal data. Thus, non-parametric testing was excluded from the analytical approach. However, based on the central limit theorem, with adequate sample sizes (*n* ≥ 30), violation of the normality assumption is unlikely to affect statistical findings. Therefore, parametric tests were acceptable due to the large sample size. Homogeneity of variances was observed (*p* = 0.64). Descriptive statistics for FW (g) by each food category and meal type are provided in Table A1 in Appendix A.

**Figure 1.** Categories of foods and their respective amounts of waste differentiated by meal type. Error bars represent the standard error of mean.

Total mean FW was greater among meat-containing meals (292.51 ± 180.77 g/plate) compared to vegetarian meals (258.46 ± 186.09 g/plate), with a mean difference of 34.05 g/plate, t(445) = 1.96, *p* = 0.05, g = 0.19 (Figure 2). The largest FW source for meat-containing meals was vegetables and fruit, while vegetarian meals had the most FW from grains and vegetables.

**Figure 2.** Plate waste from meat-containing and vegetarian meals. Vegetarian meals had less FW than meat-containing meals. Error bars represent the standard error of mean.

### 3.1.2. Exploratory Analyses

Exploratory analyses revealed significant differences in FW and GHGE between groups for analyzed food categories except fruit (Table 1). Vegetable and dessert waste were significantly greater among the meat-containing meals, while grains, dairy, egg, and plant protein waste were significantly greater among the vegetarian meals.

**Table 1.** Exploratory comparison analyses for food waste (g/plate) and GHGE (g CO2 eq/plate) between meat-containing and vegetarian meal types by food category.


<sup>1</sup> Includes vegetables and starchy vegetables; <sup>2</sup> Plant protein items consist of peanut butter, tofu, black beans, brown lentils, and hummus.

There were statistically significant differences between meat-containing meals and vegetarian meals for every major food category shared by both meal types except fruit.

### *3.2. Global Warming Potential*

Descriptive statistics for GHGE by food category and meal type are provided in Table A2. The difference in total GHGE was also compared between meal types. The data were not normally distributed (*p* < 0.001) and homogeneity of variance was not observed (*p* < 0.001). The ratio of the meat-containing meals to the vegetarian meals is 1.1; thus, this violation is unlikely to affect statistical findings. Total GHGE was significantly greater for meat-containing meals (604.20 ± 643.45 g CO2 eq) compared to vegetarian meals (356.66 ± 376.98 g CO2 eq), t(445) = 4.995, *p* < 0.001, g = 0.47 (Figure 3).

Total GHGE were significantly higher for FW from meat-containing meals than for vegetarian. The highest contributor to GHGE was animal protein, followed by dessert. The highest contributor to FW from vegetarian meals was dairy, followed by dessert.

GHGE from both meat-containing and vegetarian meals' waste had a high standard error of means. GHGE from vegetarian meals' waste was much lower than that from meat-containing meals' waste.

GHGE was significantly greater among meat-containing meals for the vegetable and dessert food categories compared to vegetarian meals. GHGE was significantly greater among vegetarian meals for grains, dairy, egg, and plant protein.

GHGE associated with plate waste showed statistically significant differences across all food categories except fruit when comparing plate waste from meat-containing meals to plate waste from vegetarian meals.

### **4. Discussion**

The objective of this study was to examine the differences in FW and GHGE between vegetarian meals and meat-containing meals to determine if greater FW among vegetarian meals offset the associated environmental benefits when compared to meat-containing meals. Analysis of plate FW failed to demonstrate evidence that vegetarian meals are associated with more FW or corresponding GHGE. Therefore, there does not appear to be a tradeoff or downside to providing vegetarian meals to patients by default for the first 24 h following their admission to a hospital setting from this perspective.

Previous work has not investigated the possibility that extra FW would be generated by providing vegetarian meals by default, which could potentially negate the environmental benefit of doing so, when compared to serving meat-containing meals by default. Only a couple of studies have reported actual FW in hospitals at the item level [29,32]. Change in meal service style from traditional foodservice to room service can reduce FW by approximately one-third [30]. GHGE from meals in a hospital setting were estimated to be approximately 5 kg CO2-eq per day for a 2000 kcal diet, with a range between approximately 0.5 and 8 kg CO2-eq for liquid diets and high protein diets, respectively [31]. GHGE from plate waste itself amounted to an average of approximately 1 kg waste per patient per day, which was associated with approximately 1.8 kg CO2-eq [32]. Plate waste refers to food that was served to a patient but not consumed, as opposed to tray waste, which includes other non-food waste, such as packaging [38]. Numerous studies indicate that the GHGE from animal-based foods are higher than those from plant-based foods [9,10,33,35].

However, FW from vegetarian meals in this study was approximately 11% lower than that from meat-containing meals, which represents a difference that is approximately half the reduction in FW observed in another study that examined FW reduction from a transition to room service rather than traditional foodservice [30]. Additionally, the average GHGE from daily plate waste per patient reported here for meat-containing and vegetarian meals was approximately 36% and 21%, respectively, of the average GHGE per day for a 2000 kcal diet in a hospital setting reported in another study [31]. In addition, the GHGE per patient per day in this study of approximately 1.8 kg CO2 eq for meat-containing meals matches the value reported in another study of hospital FW and emissions of 1.8 kg CO2 eq per patient per day [32].

Meal provision is considered an "environmental hot spot" in hospitals [39]. To address this, it has been proposed to list vegetarian meal choices first on menus and to offer more vegetarian meal options in hospitals [39]. The European Society for Clinical Nutrition and Metabolism (ESPEN) affirms the importance of providing vegetarian meals and other specialized dietary patterns to be respective of religious and dietary preferences to patients as well, noting the increased demand for vegetarian meals by patients [40]. Providing vegetarian meals in hospital settings may have synergistic benefits beyond reducing FW and environmental impacts by also promoting health.

California licensed health care facilities and state prisons are required by law to make available "wholesome, plant-based meal options" to meet patient needs and follow

physicians' diet orders according to CA Senate Bill No. 1138 [41]. Additional California law (Senate Bill No. 1383) sets targets for statewide organics recycling to reduce short-lived climate pollutants, such as methane from food waste sent to landfill [42]. The American Medical Association passed a resolution in 2017 (H-150.949) calling on US hospitals to "improve the health of patients, staff, and visitors by providing a variety of healthy food, including plant-based meals" [43]. As hospitals work to comply with such laws and resolutions, this study demonstrates that serving plant-based or vegetarian meals may provide overall reductions in FW and GHGE generated from meal service.

United States federal regulations require that hospitals provide "a nourishing, palatable, well-balanced diet that meets the daily nutritional and special dietary needs" of patients (42 Code of Federal Regulations 483.35), informed by the recommendations of a qualified registered dietitian, and that menus meet nutritional needs as recommended by the Food and Nutrition Board of the National Research Council, National Academy of Sciences [44]. While maintaining compliance with such regulations, as well as specific state regulations, there may be particular advantages conveyed by providing vegetarian meals. For example, there is a clear connection between proper nutrition and a healthy immune system to protect against infections [45]. Of particular relevance currently, healthy diets as measured by the Plant-Based Diet Score are associated with lower risk and severity of COVID-19 [46]. Health care workers (who often eat meals provided by the hospital cafeteria) who reported following plant-based diets and low-meat diets also had lower odds of moderate to severe COVID-19 [47].

There were some limitations to this study. Some food categories were excluded from exploratory statistical analysis due to inherent differences between meal types (e.g., vegetarian meals contained no animal protein). Some additional food categories were excluded due to having near negligible mean values. The food categories excluded were meat analogues, animal protein, sugars, condiments, and sauces. The larger amount of plant protein waste from vegetarian meals was expected, as these trays were more likely to contain higher amounts of plant proteins including peanut butter, tofu, black beans, brown lentils, and hummus.

Future research should include measurements of initial food weights to understand the proportion of each meal wasted and facilitate comparison across meals with different starting weights. It may also be useful to explore differences when correcting for kcal content of meals. Additional research could also examine correlations between meal type (e.g., liquid, dysphagia, cardiac, and low sodium), patient ward (e.g., surgery and intensive care), and outcomes (e.g., length of stay), as well as explore differences based on demographic factors such as sex or age.

Generalizability of the findings from this research is likely most applicable to other hospitals and similar settings where food is provided, but from fairly limited options and with few if any alternatives. In a hospital setting, there are often limited choices and the consumer may be feeling unwell, both of which increase the likelihood of them wasting food. In contrast, consumers are normally able to choose from a wide array of foods in a variety of settings, reducing the likelihood that they will waste the food they choose to consume. Therefore, it is unlikely that similar levels of food waste would be observed outside a hospital setting. It is unclear whether or not a proportional difference in food waste between vegetarian and meat-containing meals would be maintained outside a hospital setting. However, it is well known that the environmental impacts associated with meat are greater than those associated with most vegetarian foods, so it is reasonable to expect that food waste from meat-containing meals would still have higher GHGE for a similar amount of food wasted.

### **5. Conclusions**

It is important both to reduce the GHGE associated with food provision and reduce the proportion of food that goes to waste as part of efforts to limit the negative environmental consequences of food systems. Fortunately, the case study examined here provides an example where one choice—serving vegetarian meals to patients by default for their first 24 h in a hospital setting—improves both outcomes. Food waste from vegetarian meals was lower in both total weight and associated GHGE than food waste from meat-containing meals.

**Author Contributions:** Conceptualization, A.B., J.S. and H.L.; methodology A.B., B.E. and H.L.; validation, A.B.; formal analysis, N.B.; investigation, B.E.; data curation, B.E. and N.B.; writing original draft preparation, A.B., B.E. and N.B.; writing—review and editing, A.B., B.E., J.S. and H.L.; visualization, A.B. and N.B.; supervision, H.L. and J.S.; project administration, H.L.; funding acquisition, H.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Point Loma Nazarene University's Wesleyan Center (internal) grant.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data supporting results are available in Appendix A.

**Acknowledgments:** The authors thank Jean Sellars and Kalvin Lazcano for their cooperation and allowing the collection of food waste at LLUMC, as well as Anna Salisbury for her role in data collection and analysis.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

### **Appendix A**

**Table A1.** Descriptive statistics for plate waste (g/plate) by meal type and food category, presented as M ± SD.


**Table A2.** Descriptive statistics for GHGE (g CO2 eq/plate) by meal type and food category, presented as M ± SD.




### **References**


### *Article* **Pea Proteins Have Anabolic Effects Comparable to Milk Proteins on Whole Body Protein Retention and Muscle Protein Metabolism in Old Rats**

**Jérôme Salles 1, Christelle Guillet 1, Olivier Le Bacquer 1, Carmen Malnero-Fernandez 2, Christophe Giraudet 1, Véronique Patrac 1, Alexandre Berry 1, Philippe Denis 1, Corinne Pouyet 1, Marine Gueugneau 1, Yves Boirie 1,3, Heidi Jacobs <sup>2</sup> and Stéphane Walrand 1,3,\***


**Abstract:** Plant proteins are attracting rising interest due to their pro-health benefits and environmental sustainability. However, little is known about the nutritional value of pea proteins when consumed by older people. Herein, we evaluated the digestibility and nutritional efficiency of pea proteins compared to casein and whey proteins in old rats. Thirty 20-month-old male Wistar rats were assigned to an isoproteic and isocaloric diet containing either casein (CAS), soluble milk protein (WHEY) or Pisane™ pea protein isolate for 16 weeks. The three proteins had a similar effect on nitrogen balance, true digestibility and net protein utilization in old rats, which means that different protein sources did not alter body composition, tissue weight, skeletal muscle protein synthesis or degradation. Muscle mitochondrial activity, inflammation status and insulin resistance were similar between the three groups. In conclusion, old rats used pea protein with the same efficiency as casein or whey proteins, due to its high digestibility and amino acid composition. Using these plant-based proteins could help older people diversify their protein sources and more easily achieve nutritional intake recommendations.

**Keywords:** pea proteins; plant proteins; sarcopenia; skeletal muscle; protein digestibility; muscle protein metabolism

### **1. Introduction**

Alongside animal proteins, plant proteins are a critical part of the equation to help meet future protein demand and achieve worldwide food security. In the US, demand for plant proteins grew by 20% in both 2018 and 2019 [1]. This growing interest in plant proteins is driven by multiple factors, such as food safety concerns, rising food intolerances, increased accessibility of vegetarian and vegan foods, environmental concerns, sustainability imperatives, and consumer adoption of proactive approaches to health and wellbeing. The nutritional benefits of these new protein sources are still under investigation, with studies looking into their health benefits while also exploring their limits, such as allergenicity or anti-nutritional substance content [2]. Consumer acceptability needs to be carefully defined, as it remains the final bottleneck for developing new protein sources.

Grain legumes are a valuable source of plant food proteins, and so rising protein demand is expected to increase the dietary importance of grain legumes. Pulses generally have a higher nutritional value than other crops, especially since the onset of domestication

**Citation:** Salles, J.; Guillet, C.; Le Bacquer, O.; Malnero-Fernandez, C.; Giraudet, C.; Patrac, V.; Berry, A.; Denis, P.; Pouyet, C.; Gueugneau, M.; et al. Pea Proteins Have Anabolic Effects Comparable to Milk Proteins on Whole Body Protein Retention and Muscle Protein Metabolism in Old Rats. *Nutrients* **2021**, *13*, 4234. https://doi.org/10.3390/nu13124234

Academic Editors: Winston Craig and Ujué Fresán

Received: 13 October 2021 Accepted: 22 November 2021 Published: 25 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

and genetic selection processes operated by humans. Pea proteins have enough essential amino acid (EAA) content (30%) to meet WHO/FAO/UNU-recommended requirements [3]. Note that EAA requirement is based on a recommended adult protein intake of 0.8 g/kg body weight/day. Note also that peas provide well above the recommended leucine requirements [4]. In addition to providing proteins with suitable EAA profiles, legumes contain digestible carbohydrates, and some of them also contain fat.

There is considerable interest in the potential of using plant-based proteins to support muscle mass maintenance and/or growth, as demonstrated by the number of recent papers studying the impact of intakes of plant-based protein, e.g., pea proteins, on skeletal muscle anabolic response in athletes [2,5]., Dairy whey protein is a shared choice for protein supplementation in athletes because of its leucine content, its digestibility, and its ability to activate muscle protein synthesis. Most extant research on plant proteins in athletes has set out to compare and evaluate the effects of dietary supplementations with whey and pea proteins in conjunction with resistance training on muscle anabolism and strength. Taken together, the data revealed that whey and pea protein treatments led to similar responses to resistance exercise. Whey and pea proteins promote comparable muscle strength, physical performance, and body composition following resistance training [6], especially in beginners or people returning to weight training [7].

These same plant proteins could be equally valuable in other populations, such as older people, to help maintain muscle mass and slow down the aging-related process of sarcopenia. However, despite their reported efficacy in athletes, the effects of pea and other plant proteins in older people suffering from sarcopenia have not yet been disclosed. The fact that pea protein provides well above the recommended leucine requirements points to it playing a potentially valuable role in combating the loss of skeletal muscle mass and function in older subjects. Leucine is an anabolic amino acid with proven effectiveness for the maintenance of muscle mass during aging [8]. Meeting the body's quantitative daily demand for EAA is vitally important; the quality of protein consumed by older people is an equally important factor, and is generally determined by its digestibility and utilizability by the body. Among milk proteins, whey protein digests quickly, while casein digests slowly as it clots at acidic pH in the stomach. Numerous experiments have set out to determine whether fast or slow digestion was better for muscle protein synthesis and muscle building. The bottom line is that rapid digestion is best for stimulating muscle protein synthesis and increasing muscle mass, even in older people [9]. Interestingly, a previous study has shown that pea protein transiently aggregates in the stomach and has an intermediately-fast intestinal bioavailability midway between those of whey and casein [10].

When new sources of dietary proteins are tested for nutritional quality, the first studies are carried out using animal models, as advised by FAO. The second step in such studies is often to evaluate the interest of the protein in some pathophysiological situations characterized by a reduced capacity to assimilate and metabolize proteins, as is the case in older subjects. These animal studies make it possible to precisely assess protein metabolism in certain key tissues such as skeletal muscle. Such a study is difficult to perform in humans. For pea proteins, although its digestibility is high in young rats, there is little data on the nutritional value of pea proteins in old rats as compared to dairy proteins, and particularly in terms of protein digestibility and metabolism. To address this gap, this study used old rats to evaluate the efficiency of pea proteins as compared to dairy proteins, i.e., casein and whey proteins, in terms of protein digestibility, body protein retention, muscle protein synthesis and degradation and muscle protein accretion.

### **2. Materials and Methods**

### *2.1. Animal Experiment*

All animal procedures were approved by the local institutional animal care and use committee (Comité d'Ethique en Matière d'Expérimentation Animale Auvergne: C2EA-02) and conducted in accordance with the European guidelines for the care and use of laboratory animals (2010-63UE) (Authorization number: APAFIS#5329-2016051115541284 v2). Animals were housed in the INRAE's Human Nutrition Research animal facility (Agreement No. D6334515).

A total of thirty 20-month-old male Wistar rats were obtained from Janvier Labs (Le Genest-St-Isle, France). All animals came from the same batch and were bred under the same conditions throughout their lives. The rats were housed in individual cages under controlled environment conditions (12-h light/12-h dark cycle, temperature 22 ◦C) with free access to water. All of the rats were fed a maintenance diet (A04, Safe, Augy, France) ad libitum for a 2-week acclimatization period. Rats were then randomized into three groups according to body weight, fat mass and lean mass. Animals were assigned (*n* = 10 per group) to a diet containing either 14% casein (Armorprotéines, Saint-Briceen-Cogles, France) (CAS rats), 14% soluble milk protein, i.e., Protarmor™ 80, a Whey protein concentrate (Armorprotéines, Saint-Brice-en-Cogles, France) (WHEY rats) or 14% pea proteins, i.e., Pisane™ (Cosucra, Warcoing, Belgium) (PEA rats) for 16 weeks. The three experimental diets were isoproteic and isocaloric (Tables 1 and 2). Different protein to nitrogen conversion factors were used depending on the protein source used. Specifically, the conversion factors used were: 6.15 for casein, 6.08 for whey and 5.36 for pea protein. Dietary AA levels were analyzed by the ABioC laboratory (Arzacq, France) according to EN ISO 13903:2005 standard method (Table 1). Body weight and food intake were measured weekly. At the end of the experiment and after an overnight fast, the remaining CAS (*n* = 6), WHEY (*n* = 6) and PEA (*n* = 8) rats were anesthetized. Blood samples were collected from the abdominal aorta and drawn into precooled ethylenediaminetetraacetic acid (EDTA) tubes. After centrifugation, plasma was removed and frozen at −80 ◦C until analysis. Liver, heart, adipose tissues and hindlimb skeletal muscles were weighed, snap-frozen in liquid nitrogen, and stored at −80 ◦C for later analysis.

**Table 1.** Experimental diet: composition and amino acid content.



**Table 2.** Composition of the protein sources.

Compositions were obtained from technical data sheets provided by suppliers.

### *2.2. Whole Body Composition*

At the beginning, middle (after 8 weeks) and end (after 16 weeks) of the experiment, fat and lean body mass (g) were measured in non-anesthetized living animals placed in an EchoMRI-100 body composition analyzer (Echo Medical Systems LLC, Houston, TX, USA).

### *2.3. Protein Quality Evaluation*

To collect total urine and feces, rats were placed in metabolic cages (Tecniplast France, Decines-Charpieu, France) for 4 days in the last week of the experimental protocol. Total excreted nitrogen was then determined by the Dumas method at Institut UniLaSalle (Beauvais, France) [11]. Dietary protein quality was evaluated by calculating nitrogen balance (NB), apparent protein digestibility (AD), true protein digestibility (TD), net protein utilization (NPU) and biological value (BV) using the following equations [12]:

$$\text{NB(g)} = \text{NI} - \text{(FN} + \text{UN)}$$

$$\text{AD (\%)} = \frac{\text{NI} - \text{FN}}{\text{NI}} \times 100$$

$$\text{TD (\%)} = \frac{\text{NI} - \text{(FN} - \text{EFN})}{\text{NI}} \times 100$$

$$\text{NPU (\%)} = \frac{\text{NI} - \text{(FN} + \text{UN} - \text{EFN} - \text{EUN)}}{\text{NI}} \times 100$$

$$\text{BV (\%)} = \frac{\text{NPU}}{\text{TD}} \times 100$$

where NI is nitrogen intake, FN is fecal nitrogen, UN is urinary nitrogen, EFN is endogenous fecal nitrogen, and EUN is endogenous urinary nitrogen. A group of old rats that received a nitrogen-free diet during the metabolic cage period was used to deduce fecal and urinary endogenous nitrogen excretions.

### *2.4. Plasma Analyses*

Plasma levels of fasting glucose, triglycerides, and total cholesterol were determined using a Konelab 20 analyzer (Thermo-Electron Corporation, Waltham, MA, USA). ELISA kits were used to determine insulin (Alpco Diagnostics, Salem, NH, USA), leptin, (Biovendor, Bmo, Czech Republic), adiponectin (AssayPro, St Charles, MO, USA), TNFα (Millipore, Molsheim, France) and IL-10 (Diaclone, Besançon, France). Homeostatic model assessment of insulin resistance (HOMA-IR) was calculated to assess insulin sensitivity in old rats, using the formula:

$$\text{HOMA} - \text{IR} = \frac{(\text{fastig glucose} \times \text{fastig insulin})}{22.5}$$

with fasting glucose level expressed as mmol/L and fasting insulin level expressed as mIU/L.

### *2.5. Protein Synthesis Measurement*

To study muscle protein synthesis, we measured rate of incorporation of a stable isotope, i.e., an AA L-[13C6]-labeled phenylalanine (Eurisotop Saint-Aubin, France), into muscle proteins using the flooding dose method. Fasting rats were injected subcutaneously with a large dose of L-[13C6] phenylalanine (50% mol excess, 150 μmol/100 g) to flood the precursor pool of protein synthesis. Incorporation time of labeled phenylalanine was 50 min. A 50-mg piece of plantaris muscle was used to isolate and hydrolyze total mixed proteins as previously described [13]. After derivatization, L-[13C6] phenylalanine enrichments in hydrolyzed proteins and in tissue fluid were assessed using gas chromatography– mass spectrometry (Hewlett-Packard 5971A; Hewlett-Packard Co., Palo Alto, CA, USA). Fractional synthesis rates (FSR) of proteins were calculated using the equation:

$$\text{FSR} = \frac{\text{Ei}}{\text{Ep} \times \text{t}} \times 100 \tag{1}$$

where Ei is enrichment as atom percent excess of L-[13C6] phenylalanine derived from phenylalanine from proteins at time t (minus basal enrichment), Ep is mean enrichment in the precursor pool (tissue fluid L-[13C6] phenylalanine), and t is incorporation time in hours.

### *2.6. Western-Blot Analysis*

Homogenates of frozen plantaris muscles were prepared as previously described [14]. Denatured proteins were separated on a polyacrylamide gel and electrotransferred to a polyvinylidene difluoride membrane (Millipore, Molsheim, France). After blocking with 5% skimmed dry milk in Tris-buffered saline (TBS) + 0.1% Tween-20, membranes were incubated with primary antibodies: p70 S6 kinase (Thr389) and anti-total p70 S6 kinase (Cell Signaling Technology, Ozyme distributor, Saint-Quentin-en-Yvelines, France). After washing with TBS + 0.1% Tween-20, immunoblots were exposed to swine anti-rabbit immunoglobulins conjugated with horseradish peroxidase (HRP) (DAKO, Trappes, France). The antigen/primary antibody/secondary antibody/HRP complexes were visualized by luminescence using ECL Western Blotting Substrate (Pierce, Thermo Fisher Scientific, Courtaboeuf, France) and a Fusion Fx imaging system (Vilber Lourmat, Collegien, France). Quantification of band density was done using MultiGauge 3.2 software (Fujifilm Corporation, Tokyo, Japan). The values represented the ratio of the phosphorylated protein levels to total protein levels, and were expressed in arbitrary units.

### *2.7. mRNA Analysis*

The protocol for total RNA extraction and mRNA analysis has been previously described [14]. Briefly, a piece of plantaris muscle was homogenized in Tri-Reagent (Euromedex, Mundolsheim, France) and total RNA was isolated according to manufacturer's instructions. RNA amount was measured by spectrophotometry at 260 nm. Total RNA was reverse-transcribed using SuperScript III reverse transcriptase and a random hexamer and oligo dT primer combination (Invitrogen, Life Technologies, Saint-Aubin, France). PCR amplification was performed using a Rotor-Gene Q system and 2 × Rotor-Gene SYBR Green PCR master mix (Qiagen, Courtaboeuf, France). Relative concentrations of mRNA corresponding to genes of interest were quantified using Rotor-Gene software and the standard curve method. The primers used for real-time PCR analysis were listed in Table 3. Hypoxanthine-guanine phosphoribosyltransferase (HPRT) was used as housekeeping gene. Data were expressed in arbitrary units.


**Table 3.** Primer sequences used for quantitative analysis of gene expression.

### *2.8. Mitochondrial Enzymatic Assays*

First, 50 mg of frozen rat plantaris muscle was homogenized in homogenization buffer (225 mM mannitol, 75 mM sucrose, 10 mM Tris-HCl, 10 mM EDTA, pH 7.2) and then centrifugated at 650× *g* for 20 min at 4 ◦C. The supernatant was kept and the pellet was suspended in homogenization buffer and resubmitted to the same procedure. Both supernatants were pooled and used for activity measurements [14–16]. Complex I and 3-hydroxyacyl-CoA dehydrogenase (HAD) activities were spectrophotometrically assayed in the supernatant fraction by following the oxidation of nicotinamide adenine dinucleotide, reduced (NADH). Citrate synthase (CS) activity was measured by following the reduction of 5,5-dithiobis (2-nitrobenzoic acid) (DTNB) [14–17]. Activities were expressed in nmol/min/mg of proteins.

### *2.9. Statistics*

To calculate the sample size, we used published and unpublished data of net protein utilization (NPU) [18]. A difference of 20–25% and a mean variance of 10% were expected for this parameter between CAS group and WHEY group. Based on these data, the setting of type I error (α) at 5% and a power of 90%, a total of 6 rats per group was required. To anticipate potential rat death for the 16-week experimental period, 10 rats were assigned to each diet. All results were presented as means ± SEM. Animals that died or developed tumors during the experiment were excluded from the analysis. In detail, while we had 10 rats per group at baseline, the number of rats remaining at the end of the experiment was 6 CAS rats, 6 WHEY rats, and 8 PEA rats. The data were analyzed for homogeneity of variance and normality. Homogeneous data were analyzed by a one-way analysis of variance (ANOVA) followed by a Tukey-Kramer test to evaluate the significance of intergroup differences. Heterogeneous data were analyzed using Kruskal-Wallis test and the significance of inter-group differences was assessed using a Steel–Dwass test. Differences were considered significant at *p* < 0.05. Statistical analysis was performed using NCSS 2020 software (NCSS LLC., Kaysville, UT, USA).

### **3. Results**

### *3.1. Caloric Intake, Body Composition Evolution, and Final Tissue Weights*

No significant difference in calculated daily caloric intake was observed between experimental groups throughout the study period (86.0 ± 4.6 kcal/day, 92.0 ± 3.2 kcal/day and 94.8 ± 5.9 kcal/day for CAS, WHEY and PEA rats, respectively). Rat groups were purpose-defined at the beginning of the experiment to ensure no significant between-group differences in body weight, fat mass and lean mass. Thereafter, body weight, fat mass and lean mass remained not significantly different between CAS, WHEY and PEA rats at each timepoint (i.e., the middle (week 8) and the end (week 16) of the experiment) (Table 4). In accordance with the body composition measurements, the weights of several lean tissues, (i.e., skeletal muscle, liver and heart) and two different fat tissues (i.e., perirenal adipose tissue and subcutaneous adipose tissue) presented no significant between-group differences at the end of the experiment (Table 5).


**Table 4.** Body weight, fat mass and lean mass variations over the course of the experimental study.

Week 0, week 8 and week 16 mark the beginning, the middle and the end of the experiment, respectively. Data are expressed as means ± SEM.

**Table 5.** Tissue weights in CAS, WHEY and PEA old rats after 16 weeks of different diets.


Results are given as means ± SEM. Hindlimb muscle mass is the sum of plantaris, soleus, gastrocnemius, quadriceps and tibialis muscle weights.

### *3.2. Protein Quality Evaluation*

Nitrogen intake and fecal and urinary nitrogen contents were evaluated during the metabolic cage period (Table 6). None of these parameters were significantly different between rat groups. Nitrogen balance, which is the difference between nitrogen intake and nitrogen loss by both fecal and urinary routes, was similar between CAS, WHEY and PEA rats (Table 6). There were no significant between-group differences in apparent digestibility, which considers all of the digestive processes involving protein digestion, including endogenous nitrogen losses, or in true digestibility, which considers the specific digestion of dietary protein by subtracting endogenous nitrogen losses. Finally, net protein utilization, which is the ratio of retained nitrogen to ingested nitrogen, and biological value, which is the ratio of retained nitrogen to absorbed nitrogen, were similar between CAS, WHEY and PEA rats (Table 6).

**Table 6.** Evaluation of the protein quality of the different experimental diets during the 4-day period in metabolic cages.


Results are given as means ± SEM.

### *3.3. Plasma Metabolic Parameters and Cytokines*

Fasting levels of lipid metabolic markers, i.e., triglycerides and total cholesterol, were not significantly different between CAS, WHEY and PEA rats (Table 7). There were no significant dietary source-protein effects on parameters related to insulin sensitivity, i.e., fasting glucose and insulin concentrations and calculated HOMA-IR. Circulating leptin concentrations were similar between experimental groups, while adiponectin levels tended to be higher in PEA rats compared to CAS rats and WHEY rats (*p* = 0.07). After 16 weeks of feeding with dietary treatment, rats showed similar plasma concentrations of pro-inflammatory cytokines such as IL-1β and TNFα, and the anti-inflammatory cytokine IL-10 (Table 7). To evaluate inflammatory status, we calculated the ratios of the inflammatory markers TNF-α and IL-1β to the anti-inflammatory marker IL-10. TNFα/IL-10 and IL-1β/IL-10 ratios did not differ between groups (Table 7).

**Table 7.** Fasting metabolic parameters in plasma of old rats after the 16 weeks of different diets.


Results are given as means ± SEM.

### *3.4. Markers of Muscle Protein Anabolism and Catabolism*

Fractional synthesis rates (FSR) were measured in plantaris muscles of old rats (Figure 1A). According to skeletal muscle mass measurements, muscle FSR was similar between CAS, WHEY and PEA rats. Associated with these data, protein quality did not affect the phosphorylation rates of p70 S6 kinase (an intermediate of the translation initiation step) in plantaris muscles of old rats (Figure 1B). The involvement of the ubiquitinproteasome pathway in the regulation of skeletal muscle mass in the three experimental groups was assessed by measuring mRNA expressions of MuRF1 and MAFbx. Gene expressions of both E3 ubiquitin ligases were also unchanged by experimental diets in rat skeletal muscles (Figure 1C,D).

### *3.5. Muscle Mitochondrial Activity*

To explore the effect of protein quality on muscle mitochondrial function in old rats, we measured the maximal activity of citrate synthase, which is a mitochondrial matrix enzyme often used as a marker of mitochondrial density. CAS, WHET and PEA rats showed similar citrate synthase activities in plantaris muscles (Figure 2A). Likewise, the activities of muscle complex 1 and 3-hydroxyacyl-CoA dehydrogenase (HAD), i.e., one of the electron transport chain complexes and a key enzyme of the mitochondrial β-oxidation cycle, respectively, were not affected by the different experimental diets (Figure 2B,C).

**Figure 1.** Effects of different experimental diets on protein synthesis and expression of ubiquitin-proteasome pathway markers in plantaris muscles of old rats. Fractional synthesis rate (**A**) was measured by tracer enrichment in plantaris muscles after a 50-min incubation with L-[13C6] phenylalanine. In the same muscles, the phosphorylation states of p70 S6 kinase (**B**) were determined by Western-blotting, and the gene expressions of the two ubiquitin E3 ligases MuRF1 (**C**) and MAFbx (**D**) were analyzed by quantitative RT-PCR analysis. Statistical significance was assessed by ANOVA, followed by a Tukey-Kramer test or a Kruskal-Wallis test followed by a Steel–Dwass test depending on homogeneity of variance and normality. Data are expressed as means ± SEM. A.U.: Arbitrary units.

**Figure 2.** Mitochondrial enzyme activity in skeletal muscles of old rats after 16 weeks of different diets. Mitochondrial function was assessed by measuring citrate synthase (**A**), complex 1 (**B**) and 3-hydroxyacyl-CoA dehydrogenase (**C**) activities in plantaris muscles. Statistical significance was assessed by ANOVA, followed by a Tukey-Kramer test or a Kruskal-Wallis test followed by a Steel–Dwass test, depending on homogeneity of variance and normality. Data are expressed as means ± SEM.

### **4. Discussion**

Protein quality is an important component of protein intake to support growth, development, and maintenance of essential body tissues and functions [19]. The nutritional value of a protein depends on how its AA balance matches to needs, in particular EAA, and on its digestibility, i.e., on the release of AA and small peptides ready for intestinal absorption [20]. Proteins from alternative sources, such as plant proteins, are often described as having less balanced EAA profiles and lower digestibility than animal-sourced proteins [21]. However, there is a lack of data directly comparing the nutritional values of animal and plant proteins under the same experimental conditions, especially in older subjects. Here, we examined the effects of a 16-week pea protein diet on protein digestibility, body weight and composition, tissue weight, metabolic indexes, and muscle protein turnover and metabolism in old rats. Pea protein was compared to two dairy proteins, i.e., whey protein and casein, that are considered to be among the best-quality proteins, especially for maintaining body composition and muscle mass and function during aging [22]. Overall, we clearly showed that in old rats, a 16-week ingestion of milk proteins or pea protein did not influence protein assimilation and nitrogen retention, particularly in skeletal muscle. It should therefore be possible to use such plant-based protein sources for older people, which would make it possible to diversify intake and more easily attain the nutritional recommendations for this population.

### *4.1. Nitrogen Balance, Digestibility and Rate of Utilization*

When studies set out to compare the nutritional quality of several dietary proteins, the first issue to consider is usually how effectively the proteins are assimilated by the body. In particular, it is important to measure nitrogen balance, digestibility and rate of utilization to get a picture of the capacity of the protein to get digested and absorbed and to get assimilated in the tissues. Overall, the data on nitrogen balance, true digestibility and net protein utilization showed that the three proteins tested in this work had a similar effect in old rats. First, the apparent and true digestibilities of pea proteins were in the same range of values of the other proteins. Recent studies have reported that pea protein is highly digestible in rats [18,23]. However, this work represents one of the first studies to show that pea protein is also highly digestible in old rats. It has been suggested that the digestibility of plant proteins is impaired due to the presence of both anti-nutritional factors and indigestible fractions in their sequence [23]. However, the pea protein used here was a protein isolate, and protein isolates are generally well-digested [24]. In addition, protein isolates are particularly low in anti-nutritional factors, due to the manufacturing process used to extract the protein [25]. High protein digestibility induces a high quantity of AA available for intestinal absorption and, thus, improves the nutritional value of the protein source [26]. Hence, net protein utilization was equivalent between old rats fed pea protein, casein or whey protein. Urinary and fecal nitrogen excretion in old rats did not differ between the three groups, leading to an equivalent whole-body nitrogen retention. This observation contrasts with other studies done in pigs that reported increased urinary nitrogen excretion and plasma urea levels in response to soybean protein compared to casein [27]. We previously showed in young rats that protein utilization increased after feeding animals with wheat pasta enriched with fava bean flour as compared to an isoproteic wheat pasta enriched with gluten. However, in this work, protein utilization still remained lower than that measured in rats fed casein [28]. However, when the same study was carried out in old rats, there was no difference between the group fed wheat pasta enriched with fava bean and the group fed casein [18].

Evaluation of the nutritional quality of dietary proteins relies not only on protein digestibility but also on its AA composition, notably its EAA content. The EAA composition of the pea protein used in this study was close to casein and to the needs of rats, according to National Research Council [29]. The AA composition of pea protein is characterized by a limiting content of methionine (Met) [30], but the total sulfur AA content is adequate [29]. Consequently, the net protein utilization and biological value measured in old rats were

equivalent regardless of the protein used in the diet. Note that this result could be explained not only by EAA composition, in particular a high leucine content, but also by the high digestibility of the pea protein. To sum up, we showed that the biological value of ingested nitrogen, in particular nitrogen retention, did not differ in old rats, regardless of whether the protein in the diet was casein, whey, or pea protein.

### *4.2. Body Composition and Skeletal Muscle Mass*

In the present study, although we observed an age-related physiological trend towards increased body fat and reduced lean mass between the first and last month of the study, the protein source in the diet did not significantly change body composition in old rats. This result was also confirmed by the tissue weights at the end of the 16-week period. In accordance with the whole-body composition measurements, the weight of tissues constituting the lean mass, i.e., skeletal muscles, liver and heart, and of tissues resulting from the fat mass did not differ between different dietary protein groups. Few studies have focused on comparing the effects of animal versus plant proteins on body composition in old rats. We previously evaluated (also in old rats) the nutritional value of pasta made from a mix of wheat semolina and legume flours, i.e., fava bean, lentil, or pea flour [18]. Two groups were fed diets with casein or whey protein as protein source, and three groups were fed diets made with fava bean pasta, lentil pasta or pea pasta as protein source. The study found that body weight and composition, i.e., fat mass and lean mass were not significantly different between groups at each timepoint, i.e., the beginning, the middle, and the end of the experiment [18]. The effect of dietary protein sources on body composition and tissue weight has been evaluated in other works, but these studies were generally done in young rats. A lower lean mass gain was observed in young rats given soy protein for 28 days than in young rats fed whey protein [31]. At the muscular level, other studies found that, compared to casein, 16 to 20 days of ad libitum consumption of proteins from legumes, i.e., beans or lentils provoked lower muscle weights in young rats [32–34]. In addition, Alonso et al. found that muscle mass and muscle protein content were lower in young rats fed seed peas than in young rats receiving casein. In this latter study, peas were extruded and cooked to reduce the antinutritional factor content [35]. The change in lean mass or skeletal muscle mass after long-term consumption of plant-based meals has not been thoroughly assessed in older people. The rare studies available have shown that the consumption of plant proteins, when provided at sufficient amounts in each meal (i.e., >30 g/meal), should be able to maintain lean and muscle mass, and therefore increase the potential to mitigate sarcopenia in older subjects [5,36,37]. Taken together, the data presented here showed that some plant proteins, e.g., pea proteins, promoted a similar effect on body composition and muscle mass to casein and even whey protein in old rats, and could therefore be tested in the elderly as an intervention to counteract sarcopenia.

### *4.3. Mechanisms*

Several mechanisms may explain the similar action of milk proteins and pea protein on body composition and muscle mass in old rats. First, analysis of the AA content of each protein showed equivalent leucine contents between pea protein and casein. There is clear evidence that during aging, the leucine content of dietary proteins is an important parameter impacting its anabolic effect on lean mass, and specifically skeletal muscle mass [38]. It is now well recognized that leucine acts as an anabolic signal by stimulating protein synthesis and inhibiting protein breakdown at muscle level. For instance, leucine supplementation for 10 days attenuated the decrease in expression of eukaryotic translation initiation factors in young and old rat muscles [39]. In addition, this supplementation decreased the levels of ubiquitinated proteins and inhibited proteasome activity in old rats [40]. The leucine content of pea protein could thus explain its effectiveness on muscle protein turnover and therefore on muscle mass and lean body mass in old rats. Nevertheless, we did not measure the effects of pea protein under postprandial conditions and therefore we cannot draw conclusions on the role of the leucine content on protein anabolism in old rats. Note that a

second mechanisms may be involved, as we did not observe any difference between the three dietary proteins in terms of their effect on muscle protein synthesis and degradation, although we measured the rate of muscle protein turnover in postabsorptive condition. The changes observed for plantaris muscle protein synthesis in old rats were relatively in line with the changes that were observed in muscle mass. Although muscle mass tended to be higher in the whey-protein group than the pea protein group, we suggest that pea protein intake could enhance postprandial muscle protein anabolism (although we did not measure it) in old rats, which would translate into muscle protein accumulation and increased skeletal muscle mass. The influence of plant-based proteins and animal-based proteins on muscle protein synthesis has been investigated in several studies. The rate of protein synthesis in gastrocnemius muscle was lower in young rats fed raw fava bean intake than in young rats fed milk protein [41]. In addition, a lower muscle protein synthesis rate was observed in young rats when fed beans and lentils than when fed casein [34]. However, to our knowledge, the long-term effects of plant protein intake on muscle protein synthesis rate in old rats has never before been investigated.

Mitochondrial abnormalities have also been singled out as key factors in muscle changes during aging. Research on the mitochondrial electron transport chain (ETC) in skeletal muscle clearly demonstrated deficient ETC activity in muscles exhibiting the greatest loss of muscle mass with age [42]. Here, citrate synthase activity, complex 1 activity and HAD activity did not differ between dietary protein sources in old rats. Additionally, once more in old rats, we previously demonstrated that maintained mitochondrial function in skeletal muscle was associated with maintained muscle protein synthesis and muscle mass as animals aged [13]. This previous study also demonstrated that one of the mechanisms behind this action was the ability of protein intake to maintain protein turnover at the mitochondrial level [13]. This makes is tempting to postulate that pea protein, like milk proteins, could potentially help to prevent the age-related alteration of mitochondrial functional capacities in skeletal muscle, thus helping to maintain muscle mass.

### *4.4. Metabolic Parameters*

We also measured metabolic parameters related to aging-related changes in muscle mass, in particular plasma pro-inflammatory and anti-inflammatory cytokine levels [43]. The increase in blood pro-inflammatory factors and the decrease in blood antiinflammatory factors during aging causes inflammatory conditions conducive to muscle protein catabolism [44]. Here too, we showed that pea protein consumption by old rats did not modify some of the markers of the inflammatory system compared to milk proteins. It has been reported that milk protein has anti-inflammatory properties that might be effective in reducing the circulation of pro-inflammatory cytokines, such as interleukin-6 (IL-6) and tumor necrosis factor (TNF-α) [45]. A recent study on pea protein reported that a tripeptide, LRW (Leu-Arg-Trp), characterized from the pea protein legumin, and its previously studied isomer IRW (Ile-Arg-Trp) exerted strong anti-inflammatory effects by modulating the nuclear factor-κB pathway [46]. Hence, the consumption of such proteins could help keep inflammation at a level that prevents muscle protein catabolism in old rats. In addition to inflammation, insulin resistance has been described as another cause of decline in muscle protein anabolism and muscle mass in older people [47]. Here we found no between-group differences in HOMA-IR except a trend towards a reduction in insulin resistance in the PEA group compared to the CAS group. Recent studies have shown that pea glycoproteins and peptides have antidiabetic activities, in particular by reducing insulin resistance [48,49]. Therefore, it may be possible that long-term pea protein consumption could improve age-related insulin resistance in old rats. However, further studies are needed to bridge the gap between age-related inflammation and insulin resistance and pea protein intake.

### **5. Conclusions**

This study, carried out in old rats, showed that, under our experimental conditions, e.g., use of protein isolates, the body uses nitrogen with the same efficiency regardless of whether it is provided by pea protein, casein or whey. This result is partly due to the high digestibility of the pea protein, together with its EAA composition, which is close to that found in milk proteins. The divergence between our results and studies using growing rats or young rats, however, has posed unresolved questions. Here, we found evidence that plant proteins would be more effective in very old animals than in young animals. Further research is warranted to find out whether this is due to an increase in the metabolic efficiency of plant proteins or a decrease in the metabolic efficiency of milk proteins with age. In addition, clinical studies should be set up to assess the quality of plant proteins in humans, in particular the elderly, taking into consideration their pathophysiological situation and their nutritional status.

**Author Contributions:** Conceptualization: J.S., C.M.-F., Y.B., H.J. and S.W.; Data curation: J.S. and S.W.; Formal analysis: J.S., C.M.-F., H.J. and S.W.; Funding acquisition: H.J. and S.W.; Investigation: J.S., C.G. (Christelle Guillet), O.L.B., C.M.-F., C.G. (Christophe Giraudet), V.P., A.B., P.D., C.P., M.G., Y.B., H.J. and S.W.; Methodology: J.S., C.M.-F., Y.B., H.J. and S.W.; Project administration: J.S., C.M.-F., Y.B., H.J. and S.W.; Resources: H.J. and S.W.; Supervision: H.J. and S.W.; Validation: J.S., C.M.-F., H.J. and S.W.; Visualization: J.S., C.M.-F., H.J. and S.W.; Writing—original draft: J.S. and S.W.; Writing—review & editing: J.S., C.M.-F., V.P. and S.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was partly supported by a grant from COSUCRA SA, Warcoing, Belgium.

**Institutional Review Board Statement:** The study was approved by the local institutional animal care and use committee (Comité d'Ethique en Matière d'Expérimentation Animale Auvergne: C2EA-02) (protocol code: APAFIS#5329-2016051115541284 v2 approved on 23 january 2017).

**Acknowledgments:** The authors thank all the staff at the INRAE–Theix animal facilities for their valuable care and cooperation.

**Conflicts of Interest:** C.M.F. and H.J. are employees of COSUCRA-Groupe SA (Belgium).

### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Nutrients* Editorial Office E-mail: nutrients@mdpi.com www.mdpi.com/journal/nutrients

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-7451-6