Do vegans need to combine proteins at every meal?

The objection

“Plant proteins are incomplete. Grains are short of lysine and beans of methionine, so unless a vegan pairs them at every meal, rice with beans or toast with peanut butter, the body can’t use the protein properly. Studies find vegans eat less of every essential amino acid than meat-eaters, and even the new US dietary guidelines tell vegans to watch their amino acid balance. It’s a faff that proves the diet is unnatural.”

The answer

Combining proteins at every meal is a zombie rule, long dead in the science and still walking round dinner tables. It was popularised by Frances Moore Lappé’s 1971 book Diet for a Small Planet, and by the 1981 edition Lappé had changed her mind [1]. She wrote that she had stressed combining because she assumed the only way to get enough protein was to create a protein as usable by the body as animal protein, and that in fighting the myth that meat is the only source of good protein she had reinforced another, the idea that eating enough protein without meat takes great care. With a healthy, varied diet, she concluded, most people need not worry about it [1].

The rule fails mainly because a single plant protein is used better than its amino acid make-up predicts, and what counts is intake across the day, as a 2019 review by François Mariotti of AgroParisTech and Christopher Gardner of Stanford concluded [4]. A meal is also never the only source of amino acids: after a meal, the amino acids reaching the blood come from the breakdown of the body’s own tissue as well as from the food [2], and digestive enzymes, mucus proteins and proteins from cells shed by the gut lining are digested in the same way as food protein [3]. The body’s ability to smooth out an uneven meal has limits, since it keeps no dedicated store of any single amino acid [4], and Mariotti and Gardner (2019) found no evidence yet that plant proteins need a partner in the same meal [4]. A 2024 trial at the University of Texas Medical Branch tested the question directly. Nine healthy women, aged 56 on average, each ate meals with 23 g of protein from lean beef, from beans with wholemeal bread, or from either beans or bread alone, which together covered every amino acid over the day, and their muscle protein synthesis after breakfast and over 24 hours did not differ between the three (Arentson-Lantz et al., 2024) [15]. The trial was small, and after breakfast only the beef and the combined meals clearly beat a separate group given a breakfast with 5 g of protein [15].

Measured diets mostly bear this out, though newer surveys leave less room to spare. Lysine is the amino acid that plant proteins are most often low in [4], and in the UK’s EPIC-Oxford study vegan men aged 30 to 49 ate an estimated 44% less of it than meat-eaters, 2.82 g a day against 5.01 g once differences in calorie intake, age, BMI, smoking and alcohol were allowed for, and had 13% less of it in their blood (Schmidt et al., 2016) [5]. From those figures, assuming a body weight of 65 kg, Mariotti and Gardner (2019) calculated that the vegans’ intake still came to 43 mg per kilo of body weight a day, and the vegetarians’ to 58 mg, against an average requirement of 30 mg [4]. The EPIC-Oxford men were recruited between 1993 and 2001 and reported their diets on a questionnaire, a method the study’s authors say finds protein particularly hard to measure, and they add that it is not known whether the gaps they found affect health [5].

Food-diary studies from 2023 and 2025 put vegans closer to the line. Among 193 New Zealand vegans, in a study published in 2025, lysine intake averaged 39.3 mg per kilo of body weight a day, or 31.0 mg once the digestibility of each food’s protein was allowed for, against the same 30 mg requirement. On that adjusted basis only about half met their lysine and leucine needs, and 78.8% of men and 73.0% of women met the estimated average requirement for protein; legumes and pulses supplied most of their protein and lysine (Soh et al., 2025) [16]. In a 2023 Danish study of 40 vegans’ three-day food records, half met the average requirement for lysine on all three days, which the authors put down to the small number of protein foods eaten in a day (Aaslyng et al., 2023) [17]. Neither study tested pairing foods at a meal, and the remedies their authors suggest, more legumes and pulses in New Zealand and a wider range of protein foods in Denmark, are about the whole day’s eating [16][17].

Where vegans do fall short, it is on protein as a whole. In EPIC-Oxford figures gathered by Mariotti and Gardner (2019), 16.5% of vegan men and 8.1% of vegan women ate less protein than their estimated requirement, though the reviewers note that diet surveys tend to overstate shortfalls like these [4]. Their conclusion is that a vegan’s protein question is mostly one of total intake, which legumes, nuts and seeds answer [4]. The US Dietary Guidelines for 2025 to 2030 also list protein among the shortfalls of vegan diets, and their advice to vegans is to “diversify plant protein sources for amino acid balance” [6], which asks for variety and says nothing about pairing foods at a meal.

MedlinePlus, the US National Library of Medicine’s medical encyclopaedia, says you do not need to combine protein foods at the same meal to get enough protein, though it does suggest a protein source with every meal [7]. The British Dietetic Association’s fact sheet (2021, now under review) says protein needs are easily met by eating protein-rich foods such as lentils, beans, chickpeas, nuts, seeds, tofu and tempeh [8]. In the research literature, Young and Pellett (1994) concluded that mixtures of plant proteins can serve as a complete and well-balanced source of amino acids [9]. How much protein to aim for is covered in how much protein you need.

How far apart can complementary proteins be eaten?

Complementary proteins can be eaten anywhere in the same day on most readings of the evidence, though one review sets a more cautious window of about three hours. Gardner and colleagues at Stanford (2019), citing Young and Pellett’s 1994 review, write that complementary foods need not be eaten in the same hour or even the same meal, as long as they are eaten within the same 24 hours [10]. Mariotti and Gardner (2019), citing evidence that lysine is partly conserved between meals, some of it from studies in rats, conclude that a reasonable variety of protein sources across the day is enough, while noting that evidence for buffering beyond 24 hours is scarce [4]. A 2020 review by scientists employed by the nutrition company Abbott reads the same paper more narrowly: within three hours of a protein-rich meal, it says, muscle can store about 60% of an adult’s daily lysine requirement, so foods eaten within about three hours of each other need not share a plate [11].

Which plant foods are complete proteins?

In a 2020 review of plant protein quality, soya protein scored as complete on PDCAAS, the longer-established test, as did protein isolates made from potato and canola, while quinoa and buckwheat are better balanced than most grains but still limited by one or more amino acids. Every plant protein contains all nine essential amino acids, so “incomplete” only means that one of them is present in a lower proportion than people need, which is why Mariotti and Gardner (2019) call the labels misleading [4]. Grains are low in lysine, and pulses very slightly low in methionine and cysteine [4]. PDCAAS, the longer-established score for a single food, gives 1.00 to a protein that meets every requirement once digestion is allowed for, and caps scores there [11]. The 2020 Abbott review gives these scores against the pattern of needs for older children and adults [11]:

Protein	PDCAAS
Milk protein	1.00
Soya protein	1.00
Potato protein	1.00
Canola protein	1.00
Pea protein	0.91
Quinoa	0.84
Lentils	0.80
Rice protein	0.64
Oats	0.62
Wheat	0.54

Milk, soya, potato, canola, pea and rice are scored mainly as extracted protein, and the others mostly as whole foods [11]. The same review notes that blends of pea and rice protein containing 40 to 90% pea reach 1.00 on the adult pattern, a case of two proteins making up for each other [11]. On DIAAS, a newer score introduced by the FAO in 2011 that allows for the digestion of each amino acid and is not capped at 1.00, the review puts soya protein at 0.92 and milk at 1.08, measured against the needs of children aged six months to three years [11]. Lysine is what holds quinoa back on every reference pattern the review used [11]. Buckwheat protein is unusually rich in lysine, 5.1% against 2.6% for wheat in measurements published in 1980, but compounds in buckwheat lower how well its protein is digested, and threonine and the sulphur amino acids limit it (Luthar et al., 2020) [12].

Seitan is made of wheat gluten [7][8], and wheat is the clearest case of low lysine: commercial wheat protein isolates analysed by researchers in Maastricht in 2018 were 1.4% lysine, against an adult requirement of 4.5% [13]. The BDA still counts seitan among high-quality protein sources [8], and on a mixed diet the lysine in beans, lentils and soya eaten during the day makes up what wheat lacks [11][13]. Because seitan is made of gluten, anyone with coeliac disease, who has to cut out gluten for life, should leave it out [14].

When does combining at a meal still make sense?

Combining at a meal makes sense when total protein is low or the diet rests on one staple. Gardner and colleagues (2019) say a conscious effort to combine plant proteins matters only when three things hold at once: total protein close to the requirement, little variety in the diet, and no animal protein at all [10]. Vegans meet the last condition by definition, so the two to watch are a low intake and a narrow diet. Mariotti and Gardner (2019) add that mixing complementary proteins within a meal may be a practical safeguard when total intake is low, and that lysine could only run short in vegans eating little protein from a monotonous diet, grains alone for example [4].

Mariotti and Gardner single out older people, among whom small, inactive vegan women may eat only just enough protein for their needs, or less [4]. For them the review’s concern is enough protein and leucine at each meal, which it does not expect to require mixing proteins [4]. Trials since then show why total protein matters at this age. In a 2026 trial at Wageningen University in the Netherlands, 72 healthy people aged 65 and over who usually ate an omnivorous diet were randomly assigned for 12 weeks to keep to it, to switch to a vegan diet of their own choosing, or to switch and also do resistance exercise. On the vegan diet alone, protein intake fell by 27.7 g a day in the first week and was still 19.3 g lower at week 12, lean mass in the arms and legs fell by 1.15 kg, against a 0.05 kg gain on the usual diet, and muscle protein synthesis was lower; with resistance exercise the loss was 0.62 kg (Domić et al., 2026) [18]. In an earlier 10-day trial by the same group, a planned vegan diet that matched an omnivorous one for calories and protein, using a wide variety of plant proteins, made no difference to muscle protein synthesis in 34 active people averaging 72 (Domić et al., 2025) [19]. Older people eating vegan need to plan for enough protein, and strength training helps [18]; neither trial tested pairing foods at a meal.

MedlinePlus advises teenagers and pregnant women to work with a registered dietitian to make sure they get enough protein and other key nutrients [7]. Babies are a separate case: Mariotti and Gardner (2019) stress that they must have breast milk or infant formula, without which they will be malnourished, and that between six and twelve months vegetarian babies keep their protein up by continuing it alongside solid food [4]. For children and teenagers, the nutrition committee of the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) concluded in a 2025 position paper, drawing on 10 studies covering about 1,500 vegan children, that the evidence is inconclusive on whether a strictly vegan diet supports normal growth, though no significant differences in height or BMI scores were found against omnivorous children. It recommends regular checks on growth, diet and nutritional status, attention to protein, omega-3, calcium and iron, and vitamin B12 supplements [20]. Planning a vegan diet for a child is covered in vegan diets for children.