Protein content and amino acid profile
The coefficient of variation (CV) was in general large and varied among beverage types (Table 2). Soya beverages contained the highest content of protein (3.47%), followed thereafter by cow (3.42%) and goat milk (3.25%). Indeed, legumes statistically differed from the cereal group, which constituted rice- and oat-based beverages, in terms of total protein content (Table 3). The fruit and crop groups did not significantly differ in total protein content, although an advantage was found in fruit-based beverages. Considering the single amino acids, both types of animal milk had higher content per unit volume of all essential amino acids than the soya-based beverages, except for Phe (Table 4), while soya had higher contents of Arg, Ser, Gly, Asp, Ala, and Cys. As expected, the PBB groups significantly differed between them in regard to the single amino acids, except for the fruit and crop group which except for His, Met, Cys, and Try, were similar in amino acid content (Table 4). It is evident that milk protein content and amino acid profile is much less variable than in the individual PBB, with particularly cow milk showing a low CV (Table 2). For example, the CV for Cys ranged between 19 and 46% for PBB, but only between 7 and 10% for milk. This demonstrates a lack of standardisation across PBB production in comparison to milk, inferring potential shortcomings in consumer assurance and in meeting consumer expectations in terms of nutrient content.
However, considering the mere quantity of each amino acid may be misleading and should be done with caution. Indeed, a joint report by FAO and WHO18 drew global attention in 1991 to the quality of protein present in foodstuffs, hereunder referring specifically to the bioavailability of amino acids and the human gut’s ability to absorb said amino acids. They suggested the use of the so-called protein digestibility-corrected amino acid score (PDCAAS). Undoubtedly, the nature of proteins relates to a vast array of factors such as their tertiary structure, but also non-protein factors with which they may interact such as antinutrients, tannins, and fibre18. In particular, Kunitz trypsin inhibitors and Bowman-Birk inhibitors in soya require a heat treatment at very high temperatures for their deactivation, which is an additional processing step often implemented in commercial settings to optimise the nutritional benefit of the soya product at the expense of higher costs and resource-use19,20. What is more, the abundance of tannins in cereals can reduce protein digestibility21.
Based on the measurements of ileal protein matter in pigs, a quality scoring referred to as digestible indispensable amino acid score (DIAAS) was suggested in the 2011 FAO expert consultation report as the universal mean by which foodstuff protein quality should be evaluated, substituting the outdated PDCAAS22. It recognises the presence of each indispensable amino acid as an individual nutrient and evaluates its digestibility which varies among proteins23. The higher the score the greater the quality of the protein material in the food is said to be. In a recent study, soya, coconut, oat, rice, and almond-based beverages were given a score of 1.08, 0.72, 0.59, 0.43, and 0.39, respectively, when using an adult’s reference pattern. Cow milk, on the other hand, was given a DIAAS of 1.45. The older, but to date still more commonly referenced score in literature, PDCAAS, has also been widely applied to evaluate popular protein sources18. Similarly, cow’s milk has consistently scored a PDCAAS value of 1.00, soya 0.84–1.00, coconut 0.89–0.94, oat 0.54–0.57, almond 0.40, and rice 0.37–0.5924,25,26,27. Therefore, although almond-based beverages ranked the second PBB in terms of protein content, its poor DIAAS and PDCAAS scores may translate into consumers benefiting less from its protein than oat- or coconut-based beverages. It is thus evident that, with the generally low protein content and poor DIAAS and PDCAAS of PBB, these beverages do not represent substitutes for protein sources of animal milk. The exception is given by soya-based beverages, which beside having a greater protein content, also generally have high-quality protein, although the essential amino acid Met is the least abundant. It must be considered however, that 14% of people suffering from cow milk allergy, one of the main reasons for the PBB market growth, also experience allergies towards soya protein, and these beverages can thus not act as valid substitutes for cow milk28.
Although the protein intake is generally above adequate in the western world, certain demographic groups must be more attentive as their needs surpass the normal adult requirements. These include infants, children, and elderly consumers10. Not meeting the dietary requirements of essential amino acids could have negative physiological and biochemical consequences29. In particular, consumers following vegan diets should plan meticulously to meet said requirements as the poor DIAAS of plant products extends beyond beverages but also to any plant-based foodstuff. Indeed, animal-based protein sources are so well-adapt to human requirements in terms of essential amino acid composition that their ecological footprints, i.e. their land and water use, and their full life cycle CO2 emissions, are in general equal to, or lower than vegetal-based sources of proteins when considering amounts of foodstuff required to meet human requirement30,31,32 although soya-based beverages may be the exception33.
Fat content and fatty acid profile
As expected, both cow and goat whole milk demonstrated the highest values of lipid content at 3.55% and 3.72%, respectively (Table 3), but also the lowest CV of 4% and 17%, respectively (Table 2). The plant group had statistically different lipid content from the animal group with the lowest found in oat-based beverages (0.37%) and the highest in almond-based beverages (1.99%). The plant-based lipids almost exclusively consisted of LCFA, aside from coconut which had high MCFA, while the animal milk consisted primarily of LCFA and MCFA (Table 5). In regard to the individual FA and their saturation groups, it is evident that a much larger fraction of the lipid content of animal milk comprised SFA compared to PBB, about 72 vs. 9.59–17.92% (Table 5). Again, the exception was given by coconut, which consisted of about 93% SFA. On the contrary, the USFA fraction was significantly larger in PBB than in animal milk, of which a significant difference within the PBB was evident between dicots (high) and monocots (low), probably due to the coconut-based beverages in the monocot group.
Indeed, the generally lower lipid content, and consequently lower caloric content, is one aspect that has drawn health-orientated consumers to PBB in the Western world, along with their very low levels of SFA8. In this case, mirroring the lipid content and profile of animal milk in PBB alternatives may not be desirable, at least not in the West. Partial or full skimmed animal milk remain options for consumers seeking products low in lipids and free from the negative health impacts of SFA.
Saturated fatty acids can be hypercholesteraemic causing increases in low-density lipoprotein (LDL) cholesterol34,35 and evidence suggest that their consumption increases the risk of developing cardiovascular diseases29,36 although this remains controversial37. This has led to dairy products becoming entangled in consumer beliefs of lowered health due to increased cardiovascular risk. Nonetheless, studies investigating dairy consumption and human cardiovascular health have found no negative correlations between the two26,38. Some studies even present evidence for reduced cardiovascular health risks with milk consumption39 which has been, in part, attributed to other substances found in dairy products offsetting the otherwise potentially harmful effects of elevated dietary SFA consumption26.
Nevertheless, the high USFA fraction, and particularly PUFA fraction in PBB, which, aside from coconut, was roughly 3 factors greater than in animal milk (Table 5), represents a great source of healthy lipids. Additionally, by adding external crop oils to PBB, these beneficial fatty acid fractions can be further elaborated, while simultaneously improving taste and mouthfeel40. Indeed, substituting SFA with PUFA have shown to reduce the level of LDL, while concurrently increasing high-density lipoprotein (HDL) cholesterol35. In fact, one of the two essential fatty acids, namely C18:2n6, constituted a significantly greater fraction in PBB lipids, hereunder particularly soya-based beverages and to a lesser extend oat-based beverages. This omega-6 fatty acid has been associated with reduced cardiovascular risk, improved long-term glycaemic control, and insulin resistance41. C18:2n6 also acts as a precursor to the pro-inflammatory FA C20:4n6 however, and it is therefore commonly believed that its intake promotes inflammation, although results are controversial42. The high content of C18:2n6 in the PBB may thus not be as beneficial to consumers as believed. On the other hand, the omega-3 essential fatty acid C18:3n3, which has been shown to possess neuroprotective properties43 was not statistically different between PBB and animal milk, but instead significantly different between groups of PBB, likely due to very high contents in soya-based beverages.
Furthermore, it is recognised that oat derivatives have additional positive health effects due to the abundance of soluble fibres and antioxidants capable of reducing triglycerides and LDL cholesterol44. Oat is rich in avenanthramides (about 300 mg/kg in kernels), phenolic compounds reported to exhibit antioxidant and anti-inflammatory activities45,46,47, while rice oil provide γ-oryzanol with similar effects48. However, the content of avenathramides is highly correlated with the whole grain oat content49. Since the actual oat content in the PBB ranged from just 10 to 16%, it is highly likely that the avenanthramide content is very limited, while the content of γ-oryzanol was not detectable in milled rice beverages in a recent study50.
Although the lipid content of milk provides essential caloric input important to many countries particularly in Africa, this focus has shifted in the western market where low-calorie, low-SFA products are sought after51. Coconut-based beverages frequently presented values of FA content dissimilar to other PBB (Table 5). This has been widely reported and discussed in literature2,9,10. Particularly, coconut-based beverages consisted of a much larger SFA and MCFA fraction than its plant counterparts, while simultaneously exhibiting smaller fractions of USFA, LCFA, and VLCFA (Table 5). On these grounds, coconut-based beverages appear to approximate the lipid composition of animal milk.
Although this could be interpreted as coconut-based beverages being less healthy than other PBB, one must look at the more detailed composition of the FA, and indeed, links between coconut oil consumption and cardiovascular diseases remain scarce and highly controversial52. The detailed lipid composition of the coconut-based PBB highlight C12:0 (47% of the total) and C14:0 (17.8% of the total) as the most abundant FA, both of which have been reported to lower LDL content53, while C12:0 has also been reported to possess anti-cancerous and antiviral effects52. Medium chain fatty acids have been shown to favour weight control as they are preferentially oxidised and transported to the liver similarly as carbohydrates, to be used as a rapid source of energy. This is contrary to LCFA which, due to their large size, preferentially become stored in the adipose tissue unless their energy is required immediately54,55. Therefore, although the SFA content of coconut-based beverages is higher than that of animal milk, the SFA profile in coconut-based beverages advantageously differ from those in animal milk.
Total carbohydrates, sugar, and additives
Rice-based beverages contained the greatest amount of carbohydrates (Table 3), which was also reflected in their high dry matter (DM) (12.31%). Although glucose content was the highest in rice-based beverages (3.12%), this high carbohydrate content was probably due to the extremely high starch content in rice, similarly as to the other cereal-based beverage oat. The main energy storage form in cereals is starch, constituting up to 60–75% of the grain’s weight56, indicating that consumers wishing a low-carb diet should avoid cereal-based beverages. Oat-based beverages indeed contained the third-most carbohydrate (8%), immediately after almond (8.2%). The high carbohydrate content of the almond-based beverages was unexpected as the innate carbohydrate content of the almond nut is very low. However, this may be due to two coconut brands containing the polysaccharide xanthan gum and carbohydrate-binding sunflower lectin (Table 1), thereby inflating the total carbohydrate content. Indeed, almond had the highest CV for carbohydrate content (77%; Table 2). As expected, only cow and goat milk contained lactose (4.72% and 4.31%, respectively). Glucose and fructose content were significantly higher in PBB with almond-based beverages containing the most fructose (Table 3). The difference between PBB groups was more statistically significant for fructose than for glucose, likely due to exploitation of the higher sweetening power of fructose in naturally less sweet products. The difference between monocots vs. dicots and cereal vs. legume was highly significant in fructose, but only moderately significant in glucose (Table 3). Glucose was not significantly different between the fruit and crop group, but fructose was weakly significant (Table 3). Taste, colour, and consistency of PBB are often improved by the use of additives such as sucrose to fit consumer expectations of milk alternatives7. Indeed, 73% of the tested PBB brands contained added salt, 23% of them added sugar, and 67% contained at least one additive other than salt, sugar, and water (e.g. xanthum gum, sunflower oil, aromas; Table 1). With addition of sugar, the glycaemic-index of the PBB is increased, which impairs the consumer’s ability to control blood glucose levels, similarly with products containing high levels of starch, which the rice and oat-based beverages appear to do. A low glycaemic load diet has been linked to reduced risks of cardiovascular disease, obesity, and diabetes28. Furthermore, high sodium diets are known to cause increased blood pressure and risk of kidney diseases57 and the addition of salt may therefore also impair the nutritional quality of the PBB.
The perceived advantageous health effects of PBB in terms of low SFA content and increased PUFA fractions may thus be counteracted by the use of these additives and the high carbohydrate content in certain PBB groups (cereals as well as in almond). In fact, studies suggest that substituting SFA with carbohydrate, and in particular highly refined carbohydrates such as sugar, results in elevated levels of triglycerides and LDL particles while concurrently reducing HDL cholesterol54. Such effects are of particular concern in Western settings where the prevalence of obesity and insulin resistance is an ever-growing problem.
Moreover, PBB are often referred to as “highly processed” products with oat-based beverages requiring on average 14 production steps, almond 15, and soya 13. These processes involve steps such as soaking, grinding, blanching, cooking, and filtering31. This does not include steps for improving the nutritive profile of said PBB. In comparison, fresh milk requires only two major steps: pasteurisation and homogenisation31. These additional processing steps add to the PBBs’ ecological footprint33, which due to their poor nutritional quality, has been estimated to be equal to or greater than those of dairy products’ ecological footprint considering their nutrient density and not kg32. It must therefore also be carefully considered if the PBB really is a low-impact alternative to the sustainable-oriented consumer.
Iodine was exclusively found in cow and goat milk (242 and 377 µg/kg; Table 6). This mirrors data obtained by other groups in similar studies, undergone in other countries. Walther et al. analysed a set of PBB, taken from the Swiss market and found significant I content only in the subset of PBB to which red algae had been added, an additive used as a source of I fortification10. As opposed to Walther et al., fortified products were excluded from the current study, and no product containing a significant content of I was thus obtained. Escobar et al. carried out a descriptive study, analysing the composition of 179 commercial PBB available on the Spanish market, as declared from labelling. No I-fortified PBB are available on the Spanish market, thus none of the products analysed reported a significant I content58. Since dairy products have a central role in contributing to I adequacy in many countries14,59, the progressive reduction of their consumption and the shift towards a more plant-based diet risks to affect I adequacy. Indeed, increases in mild I-deficiencies currently observed in Australia and continental Europe has been linked to the reduction in the use of sanitising iodophors in the dairy industry which leave I residues in milk60. This demonstrates the importance of dairy products in meeting I requirements in these geographical areas. Therefore, to cover the recommended intakes, especially during infancy, pregnancy, and breastfeeding, the industry should ensure efforts are made to promote I fortification in PBB.
All minerals were in general present in higher concentrations in animal milk than PBB, except for Mg whose content was statistically similar between animal milk and PBB (Table 6). Rice-based beverages were the most mineral-poor beverages, lacking Mg, K, and S and demonstrating low to moderate contents of Ca, P, and Na (124, 84 and 228 mg/kg, respectively). This is also mirrored in rice-beverages low ash content at 0.09% (Table 3). On the contrary, soya-based beverages demonstrated the greatest mineral content among the PBB, which is further evident in its high ash content (0.45%). Na was the only mineral found in greater amounts in other PBB than soya (325 mg/kg), namely in coconut- and oat-based beverages (367 and 381 mg/kg; Table 6). This could be due to salt being added in higher concentrations in coconut- and oat-based beverages than in soya (Table 1). Calcium content was about 10-times greater in cow milk than in the PBB (about 1000 mg/L), with soya-based beverages demonstrating the largest calcium content of the PBB (260 mg/L). Legumes, whole-grains, and nuts are known to be great sources of certain minerals, i.e. K, P, and Mg, although this is not reflected in the mineral content of the PBB, suggesting that flours of dehulled cereal grains were probably used for their preparation. This low mineral content could also be explained by the plant material constituting a small percentage of the total volume of the PBB. Indeed, the content of the main plant origin of each PBB ranged from just 3% in one almond brand to 40% in one coconut brand (Table 1).
Due to the important role Ca plays in bone development31, milk has been campaigned as a great nutritive source for healthy bones, particularly in growing children. For this reason, many PBB products are fortified with Ca to mirror concentrations found in cow milk. According to the EFSA (2017)61, an adult over the age of 25 requires 950 mg of Ca per day, which can be met with 890 mL of cow milk or 1077 mL goat milk. On the contrary, despite being the PBB with the highest Ca content, 3654 mL of unfortified soya-based beverage are needed to obtain the required daily intake of Ca. Similarly, 330 mL of cow milk will provide five-times the required daily P intake, and 13% and 22% of the daily requirements of K and Mg, respectively, while with 330 mL soya-based beverage these values are 28%, 13%, and 17%61. As for the poorest of the PBB in terms of mineral content, namely rice-based beverages, 7600 mL of beverage are required to obtain the 950 mg Ca requirement, while 330 mL will not provide any substantial amount of neither P, K, nor Mg61.
Overall nutritional value
Beyond the mere content of individual macro and micronutrients, one must always consider the overall content of each beverage and their bioavailability. The DM confers information on percentage of solids present in foodstuff, and therefore indirectly on amount of total nutrients. Rice-based beverages were found to have the highest DM at 12.31%, followed by cow and goat milk at 12.22% and 11.84% (Table 3), respectively. However, these high DM values constitute highly diverse nutrient profiles. As regards rice, carbohydrates represented the largest fraction of the DM at 12.60%, most of which was likely starch56. On the other hand, the DM of animal milk constituted in large parts of carbohydrate, hereunder mainly lactose, as well as lipids and protein. The generally high carbohydrate content in PBB, except for coconut and soya, with concurrently low mineral, protein, and lipid contents make these products poor sources of nutrients in relation to the energy provided. Although the low SFA and high PUFA contents in PBB may represent healthy FA profiles associated with reduced risk of cardiovascular disease onset, the amount of FA per 100 mL is still very limited. What is more, the lack of evidence on the presence of the often-referenced health-promoting avenanthramides and γ-oryzanol after oat and rice PBB processing, does not strengthen the nutritional profile of these two PBB49,50.
Milk therefore represents a matrix with a more nutritive profile than PBB, with the exception given by soya. Soya-based beverages sustain similar albeit lower mineral contents in comparison to milk, however, and can therefore not act as a complete substitute. This is of particular concern to I, Ca, and P. However, even mineral fortification may not be a mean by which to overcome this issue. Craig and Fresán62 found that the amount of Ca absorbed from Ca3(PO4)2, a common compound used as a Ca fortifier, was significantly less than Ca absorbed from milk. Concurrently, the amount absorbed from CaCO3, another popular Ca fortifying compound, was comparable to that of milk62. However, issues regarding sedimentation of these Ca fortifiers, which occurs during the product’s shelf life, persist31.
Research into the bioavailability of minerals in the matrices of PBB is still lacking. Nonetheless, it is evident that some phyto-compounds actively block the absorption of certain minerals exacerbating the issue of low mineral contents in PBB. Phytates, used by plants as storage compounds for P and inositol, are grouped as so-called antinutrients. They are found in cereals, nuts, and seeds, with particularly oat having high concentrations at 2618 mg/100 g dry matter63. They form soluble complexes with divalent cations under acidic environments, like those found in the digestive tract. Phytates can thus sequester dietary Z, Fe, and Ca, and since monogastrics such as humans do not produce the phytase enzyme64, the minerals remain unavailable for absorption65. Phytates are heat stable66 and are therefore not removed or denatured during the common thermal processing of PBB. Raghavendra et al. explored their content in soya-based beverages and found it to be 0.106 g/100 g beverage. With fermentation using the phytate-degrading Pediococcus pentosaceus, the concentration of phytates decreased to 0.093 g/100 g, a reduction of 12%, and the Ca bioavailability concurrently increased by 68%67. This, however, required a 12 h fermentation step not compatible with commercial production.
Thus, although PBB may be marketed as nutritive dairy alternatives low in cholesterol and cardiovascular disease-inducing FA, their nutrient content is highly limited, and their carbohydrate and sugar contents elevated. Due to the presence of anti-nutritive phyto-compunds such as phytates, and their low endogenous mineral content and high carbohydrate content, if implemented in the diet to fully substitute dairy products, issues regarding mineral deficiencies and blood glycaemic control may arise. This is of special concern for vulnerable demographic groups such as infants, elderly, and lactating women if implemented without regards to nutritional requirements as may be the case if proper information is not correctly communicated. Indeed, the high CV of most traits in comparison to milk demonstrates the lack of standardisation across PBB manufacturing, making informed decision-making by consumers difficult.
Discriminant ability of composition traits
The PCA on the common traits among all beverage types revealed PC1 and PC2 to explain 50.6% and 37.0% of the variance, respectively (Figs. 1 and 2). The five most influential traits of each principal component included Leu, Ser, Val, Pro, and Met for PC1, and Glu, Asp, Tyr, Leu, and Lys for PC2 (Fig. 1; Table 7). The beverages formed clear clusters when plotted on a 2D-plot, with particularly cow and goat milk clearly segregating from the PBB (Fig. 2). The soya-based beverages were the PBB most dissimilar from the other PBB, while almond, oat, coconut, and rice-based beverages formed an overlapping cluster (Fig. 2). The PCA therefore demonstrates that the nutrient composition is indeed different not just between the animal and PBB but also between soya-based beverages and all other PBB, and that despite the high variability between brands of PBB, there is a somewhat accordance between them. This corroborates the fact that PBB differ according to the manufacturing process, recipe, and basic material. The supervised LD analysis (Fig. 3; Table 7) showed excellent discrimination ability; 100% of the beverages were correctly classified, with 14.9% being of rice and coconut origin each, 16.2% of soya, oat, and almond origin each, and 10.8% of cow and goat origin each. Indeed, with LD1 and LD2, the LD analysis was able to maximise the class separability, explaining 67.2% and 23.9% of the variance, respectively (Fig. 3). This study used detailed FA and amino acid composition, as well as mineral content and gross composition to identify the best discriminatory traits between different PBB and milk. These were determined as the different FA saturation groups, and the two individual FA C18:1n9 and C16:0, as well as Glu (Table 7). The differing contents of the various FA groups is indeed one of the main factors driving health-orientated consumers from animal milk to PBB, in which it is evident that even within PBB, differences exist (Table 5; Fig. 3). Consumers must therefore consider which type of PBB best accommodates their dietary aspirations and needs. Analyses highlight that the almond-, coconut-, rice- and, oat-based PBB are similar in nutritional content, while soya-based PBB differ highly from other PBB. The LD demonstrated that cow milk is more similar to PBB other than soya in regard to LD2, while goat milk is more similar to PBB in regard to LD1.
It can thus be concluded that the nutrition profile of the investigated PBB is not of comparative quality to those of both cow and goat milk, with the exception given by soya. The I, Ca, and P content, however, remained greatly elevated in milk in comparison to the soya-based beverages, while also the FA profile differed. Oat, rice, almond, and coconut-based beverages all had significantly lower protein and essential amino acid content than animal milk. Soya, however, had consistently comparable content of protein and essential amino acids to animal milk, but due to lower bioavailability and poorer DIAAS, consumers may not benefit from these nutrients to the same extended were they sourced from animal milk. Coconut FA composition mirrored to a certain extent that of the two milk types, but the USFA present in coconut-based beverages are considered to possess beneficial health effects and should therefore not be considered a negative aspect of this PBB. Both cow and goat milk represent great sources of most minerals, unlike PBB which were generally mineral-poor and implicated in issues regarding bioavailability due to endogenous antinutrients. The variability between brands of the same PBB was extremely high, unlike for animal milk. This reflects a necessity for industry standardisation concerning PBB production so to gain consumer assurance. For this reason, further analysis with more PBB brands is required to fully elucidate the nutrient profile of PBB currently available on the market. What is more, although milk is not regarded as a vital source of vitamins, they remain a highly important part of the human diet and should also be taken into consideration in future studies. The combination of poor nutritional profile and need for a high level of processing nullifies what consumers often believe to be a nutritious alternative to traditional animal milk.