Changing the diet from animal based to plant based protein has several positive health effects [6]. However, since the DIAAS score for plant proteins is generally lower than the DIAAS score for animal proteins [8], care must be taken to meet the recommendation for dietary protein and amino acid intake. In this study, the actual food intake from 40 vegans over three days was measured using a three-day dietary record and the protein- and amino acid intake calculated.

An individual’s energy requirement is dependent on his/her sex, body weight and activity level, and this was calculated individually for each participant. In contrast, the protein recommendation was calculated as g/kg BW and was independent of sex and activity level. Some studies include only one sex, often men, [32] while others do not break down the results according to sex [33]. In studies that break down the results according to sex, the difference between the dietary groups (vegans, omnivores) is the same in the two sexes [34].

The energy intake was lower than recommended for most of the participants (55%) in our study (Table 3). However, with an average of 8.2 MJ (Table 2), it was at the same level as that of other studies reporting vegan diets ranging from 8.14 MJ/day [22] to 9.97 MJ/day [5]. Despite the low energy intake, the protein intake was to a larger degree sufficient and at 0.98 g/kg body weight it was on average above the level recommended by WHO (Table 2) which is in accordance with other studies which reported 0.94 g/ kg body weight [35], 1.0 g/kg body weight [36] and 1.01 g/kg body weight [37] and higher than that of another study which reported 0.64 g/kg body weight [38]. For the individual days, the protein intake was within the level recommended by the WHO on all three days for 60% of the participants (Table 3). This also means that, for 40% of the participants, their protein intake was below the recommended level on one or more days, and, for 25% of the participants, the protein intake was below the recommended level every day or on two out of three days. In comparison, Allès et al. [39] report that 27.3% of 789 vegan participants had a protein intake below the acceptable level while Waldmann et al. [35] report the same for 31.3% of the vegan males and 41.4% of the vegan females. In comparison, only 4% of the meat eaters and 15.3% of the vegetarians had a protein intake below the acceptable level in the study by Allès et al. [39]. Other studies state that the average protein intake is sufficient in a vegan diet [5, 36, 40]. For example, a study with 269 vegan males found that the average intake was 0.91 g/kg BW [24], although it did not investigate single days. To maintain body function, growth, and for females reproduction and lactation, protein intake needs to meet the recommendations each day [29]. It is therefore not sufficient just to look at the average consumption over several days or the average consumption for a population, as most studies do, but, as in our study, go into detail each day.

One concern in changing from animal to plant proteins in the diet is the protein dietary quality [3]. The quality of plant proteins is lower due to an unbalanced amino acid composition and a lower digestibility [11], and the circulating blood level of the EAAs in subjects who follow a vegan diet compared with subjects who follow an omnivore diet has been shown to be lower for Lys while different results have been seen for, Cys and Met [21]. Interestingly, the correlation between circulating plasma amino acids and the intake is low, as long as the intake is adequate [36, 38]. This confirms the results of other studies that show that plant protein can be as efficient as animal protein in muscle anabolism, as long as the amino acid composition is within the recommendation [17, 41]. In our study, Lys, followed by SAA, Leu and Val were the EAAs most often below the recommended level which corresponds to other studies [16, 21, 36].

Even though several studies state that a vegan diet overall can be healthier than an omnivore diet [5, 32], some studies point out that there can be drawbacks, for example in the immune system, to changing to a vegan diet [42] and also stress that the long-term effect of a vegan diet remains unknown [43]. We have shown that, even though almost half of the vegans in this study met the recommendations for protein and EAAs on all three days (47.5%), a significant number of them (35%) did not meet the protein and EAA requirements on most of the days, giving rise to nutritional concerns.

To ensure a balanced amino acid composition [19, 20] it is necessary to combine several different protein sources that complement each other [10]. The lower digestibility of plant proteins can be attributed to both the protein chemical structure and the interaction with other macronutrients, and the digestibility can be further impaired by the presence of antinutritional compounds [10, 11]. One way to compensate for the lower digestibility of plant proteins is to increase protein intake. However, half of the participants in our study did not meet the recommendations for protein intake on all three days. Another strategy is to enhance the digestibility by applying technological processing steps that can contribute to changes in the protein chemical structure and reduce the antinutritional components [3, 10, 11]. This might present a challenge, since focus groups have pointed out that there is a low level of trust among consumers regarding industrial products that they perceive as highly processed [44] and even though these consumers were omnivores, the same could be true for vegan consumers. It has also been stated that, unless the vegan diet is unrealistically uniform, the amino acid supply will be sufficient [45]. When considering the choice of food items, this study shows that the diet for many of the vegan participants was uniform. Cereals provided one of the main protein sources, with oats alone eaten on 60.8% of the days and other cereals eaten on 71.7% of the days (Table 4). Cereals are known to be low in Lys [46] in particular, and Lys is also the limiting amino acid in most of the diets (Table 3). This is further strengthened by the intake of different seeds in which Lys is also the limiting amino acid [11]. Peanut butter and nuts also tended to be popular among vegan consumers in this study, and they were eaten on almost half of the days. Peanut butter is high in energy, but, as with the cereals and seeds, Lys is the limiting amino acid [11]. Lys helps the body absorb calcium and also has an impact on the synthesis of collagen [47]. Lys deficiency has been demonstrated to impair amino acid metabolism and induce cellular imbalance [48]. This underlines the need to supplement the diet with other protein sources that are richer in Lys.

Rice is a cereal that is often eaten as the carbohydrate part of the meal, though the intake of rice was very low among the vegans in this study (Table 4). This might be because climate change is one of the main reasons many vegans avoid animal products [2], and the production of rice has a high climate impact [49].

Pulses were the second most frequently consumed protein source, followed by vegetables (Table 4). Other studies have reported the average intake per day measured in weight. In these studies vegetables and fruits are the most abundant food items with an intake having approximately ten times higher intake than that of legumes and three to five times higher than that of cereals and starchy food [31, 39]. However, this is problably because these food items are heavier than legumes and cereals and it does not indicate how often they are eaten. Chickpeas were the single most abundant food item, consumed as either hummus or falafel. Hummus is known to be a nutritious food item rich in protein, dietary fibres, micronutrients and different bioactive compounds, but, as with the other pulses, chickpeas need to be supplemented with other protein sources to achieve a complete amino acid profile [50]. In addition to chickpeas, different soy products such as yoghurt and soy drink made up most of the intake of pulses, although other types of beans were also included in the diet. Protein from pulses is known to have a low content of the SAAs [11], in particular, which is also reflected in the intake, since the intake of SAAs was the second most insufficient amino acid, achieving recommended levels on all days for only two thirds of the participants (Table 3). SAAs are involved in the synthesis of several key metabolites [47], and there are indications that a deficient intake of SAA might lead to neural disorders [51, 52]. Met and Cys are most abundant in the proteins of pseudocereals such as quinoa and hemp [11], and, as seen in Table 4, even though some of the participants supplemented their diet with hempseeds, neither of these pseudocereals was widespread in the diet.

Even though it is argued that it is possible to achieve a balanced amino acid intake by eating a varied diet containing different plant protein sources, Table 5 shows that the vegan participants’ diet in this study is mostly made up of three, four or five protein sources. Furthermore, the fact that some of the protein sources have the same limiting amino acids, in particular Lys, but also the SAAs, shows that the combination of different protein sources required in order to include all of the EAAs in a vegan diet in sufficient amounts, are not necessarily present today, since less than half of the participants met the recommended protein intake and all the amino acids on all three days (Table 3). The reason for the low variation in protein sources is unknown, but could be a lack of awareness, poor cooking skills or others.

Some studies have pointed out that many vegan consumers reject highly processed foods due to distrust [3, 44]. This was confirmed in our study, since convenience food made up only part of the diet on 14.2% of the days, often as some very well-known dishes such as pizza and burgers. Instead, the vegans made their own food, which for some of them resulted in uniform meals containing few protein sources, while others made very complex food from scratch containing several protein sources. This distrust is something of a dilemma, in that some kinds of processing, such as extrusion, can actually enhance protein digestibility [53]. Furthermore, the production of plant-based convenience foods also offers the possibility of combining protein sources to provide an adequate amino acid composition. The high intake of hummus and peanut butter further points to a need for greater variation in the processed food available on the market. A supply of different convenience food items with high nutritional quality and a high protein quality could help improve the quality of vegan diets. This should be accompanied with information regarding the health effect of these food items, to encourage the more sceptical vegans to include them in their diet rather than only benefiting those who are not wary of processed food.

According to Berrazaa et al. [10], a discussion has been ongoing on whether to increase the recommended protein intake in a vegan diet, since the protein dietary quality of plant protein is lower than that of animal protein. This idea has been rejected, since it is possible to combine different protein sources to achieve the required amino acid composition. Our study shows that a significant number of the participants did not achieve an adequate protein and amino acid intake in their diet every day and that complementary protein sources were not combined systematically. Furthermore, when taking into account lower protein bioavailability, it is necessary to provide more guidance on how to design an adequate vegan diet. Focus should also be aimed at developing mildly processed convenience food items that take into account improved protein quality.

The study has certain limitations. Nutritional value of a diet can be investigated by several methods. A three-day dietary record is a prospective open-ended assessment method used to record all food items consumed during the assessment period. The participant weighs all food items consumed on two weekdays and one weekend day and records detailed information such as recipes, preparation methods and brands [25]. Although this method is time-consuming and difficult for the participant to manage, it does have several advantages. First of all, the actual food intake is recorded, and the quantity of the intake is measured. This can subsequently be related to different databases of nutritional content, and the intake of macro- and micronutrients can be calculated. However, one drawback to using this method could be that the participant has an unusual dietary pattern during the three-day period due to the specific focus on the food. Another drawback is that the data are self-reported, and the intake might therefore be underestimated. Other methods used to estimate food intake include a 24-hour food interview in which the participant is interviewed about his or her consumption retrospectively and a food frequency questionnaire in which food items and portion sizes are systematically recorded. Since both the 24-hour food interview and the food frequency questionnaire are retrospective, they are challenged by the participant’s ability to recall his/hers intake, and, furthermore, the exact quantity of the food intake is not measured. In our study, we therefore decided to use the three-day dietary record to collect quantitative data on the actual intake on these days despite its limitations.

The amino acid content of many of the food items was not available in the databases. For some of the food items, the ingredients were present in the database, and we therefore calculated the amino acid content on the background of a recipe. For other food items, we searched the scientific literature to determine the amino acid composition. This introduced a degree of uncertainty regarding the intake of amino acids, but it was necessary to obtain enough information and it was more precise than only including the food items that were already present in the database.

Another limitation of the study is that the participants’ weight and activity level were self-reported. This means that their weights might have been underreported, which is a well-known problem [54], and therefore the calculated dietary needs might have been underestimated. In addition, more females than males participated in the study, which indicates that more females than males want to reduce their meat consumption [55] or that more females were willing to participate in this kind of a study. The recommended intake of protein, amino acids and energy was calculated on a personal level, and it is therefore not necessary to distinguish between sexes in the analysis. The age distribution was broad, which can also be seen as an advantage since the participants were representative of a broad section of the population.

During the recruitment, we did not ask the participants how long had followed a vegan diet, or about their motivations for this and whether they were trying to lose weight. These factors might have influenced their eating habits.

The sample number in this study was 40 people. They were not necessarily representative of all vegans in Denmark, as they were sampled using social media such as Facebook and the snowballing effect. Furthermore, the incentives of providing nutritional advice and offering tickets to the cinema were mentioned during recruitment and this might have further biased the sampling toward recruiting vegans who were not confident in their knowledge of what they should eat. When taking into account the time consumption and complexity of the three-day dietary record method, it must be expected that the participants represent the vegans who are most interested in diet and nutrition, since they are the most motivated. Although the number of people who follow a vegan diet in Denmark is unknown, a recent survey sample (n = 1005) using a sampling strategy to include as many vegan and vegetarian participants as possible, found 1% followed a vegan diet [56]. Relating this to the Danish population as a whole would mean that approximately 59,000 people in Denmark follow a vegan diet. However, since the survey was specifically aimed at this group of people, the actual number can be expected to be lower. According to the Danish Vegetarian Society, 45,000 Danes follow a vegan diet [57], which means that the 40 participants in this study represent between 0.09% and 0.07% of all Danes who follow a vegan diet.

In this study we focused on a vegan population and did not include other dietary groups. Therefore, part of the conclusion might also apply to other dietary groups though this was not within the scope of the study.