Analysis of Eight Types of Plant-based Milk Alternatives from the United States Market for Target Minerals and Trace Elements

Abstract

A wide variety of commercial plant-based foods that are marketed and sold as alternatives for milk (plant-based milk alternatives or PBMAs) are available to consumers. In this study, PBMAs from the United States (n=85) were subjected to analysis for target minerals (magnesium, phosphorus, selenium, and zinc) to compare their variability across PBMA types, brands, and production lots. Samples were also screened for the environmental contaminant elements arsenic, cadmium, and lead. The eight PBMA types sampled were produced from almond, cashew, coconut, hemp, oat, pea, rice, and soy. Elemental analysis was conducted using microwave-assisted acid digestion followed by inductively coupled plasma-mass spectrometry. The results showed that pea PBMAs contained the highest mean amounts of phosphorus, selenium, and zinc, while soy PBMAs were highest in magnesium. Mean amounts of minerals were lower than those found in milk for the majority of PBMA types. There was significant variation (P<0.05) in amounts of minerals across the majority of product brands. The amounts of phosphorus and magnesium varied across production lots (P<0.05), but the absolute value of these differences was low. Total arsenic was highest in rice PBMAs; amounts of cadmium and lead across PBMAs were generally found at low or non-quantifiable amounts. These results underscore the importance of generating analytical data on the elemental composition of products within the rapidly growing category of PBMA.

1. INTRODUCTION

The consumption of plant-based foods that are marketed and sold as alternatives for milk (plant-based milk alternatives or PBMAs) is increasing among Americans, and there are a wide variety of product types and brands available to meet the consumer demand (Stewart, Kuchler, Cessna, & Hahn, 2020; Watson, 2022). In contrast, fluid bovine milk (“milk”) intake has been declining over the past several decades, but consumer substitution of milk with PBMAs does not appear to be a major driver of this decrease (Slade, 2023; Stewart & Kuchler, 2022). There are multiple factors for why consumers choose to include PBMAs in their diets rather than milk; some factors include taste, allergy to milk protein, lactose intolerance, religious reasons, and adherence to a vegan or certain types of vegetarian diets (Auclair, Han, & Burgos, 2019; Chalupa-Krebzdak, Long, & Bohrer, 2018; Haas, Schnepps, Pichler, & Meixner, 2019; Paul, Kumar, Kumar, & Sharma, 2020; Rizzo, Harwood, & Drake, 2020). Because milk is generally associated with being a source of the micronutrients vitamin D and calcium, some manufacturers fortify PBMAs with these micronutrients to match amounts found in milk (Hodges, Cao, Cladis, & Weaver, 2019; Rizzo et al., 2020; Singhal, Baker, & Baker, 2017). Besides vitamin D and calcium, other key nutrients provided by milk and other dairy foods include the minerals magnesium, phosphorus, selenium, and zinc (US Department of Agriculture and US Department of Health and Human Services, 2015; US Food and Drug Administration, 2023).

The minerals magnesium, phosphorus, selenium, and zinc are micronutrients required for a vast array of biological functions, including maintenance of bone tissue, immune function, and nucleic acid synthesis (Institute of Medicine, 1997, 2000, 2001). The National Academy of Medicine has established Dietary Reference Intakes (DRIs) for all four of these minerals (Institute of Medicine, 1997, 2000, 2001). Besides the dairy group, there are a variety of foods and beverages from both plant- and animal-derived sources can serve as a source of these micronutrients.

Information on key minerals in PBMAs is needed to compare the nutritional content of PBMAs to milk, but publicly available analytical data on the amounts of minerals in PBMAs are limited. A number of studies examining the nutritional content of PMBAs rely on the values declared on the product’s Nutrition Facts label or on the US Department of Agriculture’s (USDA) FoodData Central nutrient database (Chalupa-Krebzdak et al., 2018; Craig & Fresán, 2021; Drewnowski, 2021, 2022). These sources have been previously noted to be limited because FoodData Central currently does not contain information on the increasingly large variety of PBMAs available on the market. Additionally, the U.S. Food and Drug Administration’s (FDA’s) food labeling regulations do not require magnesium, phosphorus, selenium, and zinc to be declared on the Nutrition Facts label unless they are added as a nutrient supplement to the food or when a claim is made about the nutrient (such as a nutrient content claim) (US Food and Drug Administration, 2018). In addition to minerals present in PBMAs, some of these products have been found to contain trace amounts of environmental contaminants, including the elements arsenic and cadmium (Kosečková, Zvěřina, Pruša, Coufalík, & Hrežová, 2020; Shannon & Rodriguez, 2014). It is crucial to build data on both nutrients and elements of concern in a range of PBMA types to understand how increased consumption of these products affects overall dietary intake of these elements.

The present study analyzed commercially available PBMAs and milk for the minerals phosphorus, magnesium, selenium, and zinc, as well as the environmental contaminants arsenic, cadmium, and lead. The eight different types of PBMAs selected for the study were almond, cashew, coconut, hemp, oat, pea, rice, and soy. Elemental analysis was accomplished using microwave-assisted acid digestion followed by inductively coupled plasma-mass spectrometry (ICP-MS), an established analytical method used for measuring minerals and trace elements in foods and beverages (Gray & Cunningham, 2019; Redan, 2020; Soares, Moraes, Rocha, & Virgilio, 2023). The analytical results were used to compare differences in the amounts of target elements across product types and brands, and, for almond PBMAs, within and across production lots. The results were also compared to any value declared on the product’s Nutrition Facts label and to the amounts measured in milk.

2. MATERIALS AND METHODS

2.1. Chemicals, standards, and reference materials

Optima grade concentrated nitric acid and electronic grade 2-propanol were obtained from Thermo Fisher Scientific (Waltham, MA, USA). Stock standard containing ionic arsenic (As), cadmium (Cd), magnesium (Mg), phosphorus (P), lead (Pb), selenium (Se), and zinc (Zn) was from Inorganic Ventures (Blacksburg, VA, USA). Internal standard mix for ICP-MS was purchased from Agilent Technologies (Santa Clara, CA, USA).

Deionized (DI) water (18.2 MΩ cm at 25 °C) was obtained from a Milli-Q system (Millipore-Sigma; Burlington, MA, USA). Standard Reference Material (SRM) 1549a (Whole Milk Powder) and SRM 1643f (Trace Elements in Water) were procured from the National Institute of Standards and Technology (NIST; Gaithersburg, MD, USA).

2.2. Sampling and storage of PBMAs and milk

Samples used for this analysis were from a previous study conducted by our group (Sevillano Pires et al., 2023). The eight different types of PBMAs in this study were made from almond, cashew, coconut, hemp, oat, rice, pea protein, and soy, as described by the product packaging labels. These product types were selected based on the most commonly consumed PBMAs in the US as of September 2022 (Watson, 2022). Milk samples were bovine fluid whole milk (3.25% milkfat). Locations of sample purchase were randomly selected from a comprehensive list of retail food markets within a 50-mile radius of Bedford Park, IL, USA. Samples were obtained from 10 different markets and an online retailer. Two to three brands of each beverage type, and at least three replicate product units within the same production lot were obtained. Overall, samples were from 19 different brands. Almond PBMAs were selected, as they are the most commonly consumed type in the US, for additional analysis to capture the variability of minerals across lots. Three different lots of one brand of an almond PBMA were used for analysis. A summary of the samples is shown in Table 1.

Table 1.

Summary of the 8 types of PBMAs and number of product units for elemental analysis.

PBMA Type/Code	Brands	Product Units
Almond/A	3	18
Cashew/CW	2	9
Coconut/CT	2	9
Hemp/H	2	8
Oat/O	3	12
Pea/P	2	8
Rice/R	2	8
Soy/S	3	13
Total	19	85

Each product was mixed thoroughly before being dispensed into 15 mL amber polypropylene tubes (Chemglass Life Sciences; Vineland, NJ, USA). Samples were stored at −20 °C until analysis.

2.3. Elemental analysis using ICP-MS

Samples were analyzed for total As, Cd, Mg, P, Pb, Se, and Zn content using microwave-assisted digestion followed by ICP-MS, according to a method adapted from the FDA’s Elemental Analysis Manual 4.7 (Gray et al., 2019). Briefly, each sample was subjected to microwave-assisted digestion with concentrated nitric acid using a Discover SP-D microwave digestion system (CEM Corporation; Matthew, NC, USA). After digestion, each sample was diluted gravimetrically to 100 g with DI water in trace metal quality plastic tubes (SCP Sciences; Champlain, NY). Samples for Mg and P analysis were further diluted 20-fold with 5% (v:v) nitric acid because of the naturally high concentrations of these elements in the samples. An Agilent 8800 ICP-MS (Santa Clara, CA, USA) was set in single quadrupole mode with research grade helium (99.999%) from Airgas (Radnor, PA) as the collision gas. Internal standard mix (Agilent Technologies; Santa Clara, CA, USA) was prepared in aqueous nitric acid:2-propanol (5:20; v:v). The internal standard was set to scandium (⁴⁵Sc) for Mg; germanium (⁷²Ge) for P, As, and Se; rhodium (¹⁰³Rh) for Zn and Cd; and bismuth (²⁰⁹Bi) for Pb. Data for total Mg, P, Zn, As, Se, and Cd are reported as ²⁴Mg, ³¹P, ⁶⁶Zn, ⁷⁵As, ⁷⁸Se, and ¹¹¹Cd, respectively. Total Pb is reported as the sum of isotopes 206, 207, and 208.

To assess method performance for the mineral micronutrients, SRM 1549a was analyzed and the resulting data was compared to the certified values provided by NIST. Because certified values for the trace elements arsenic, cadmium, and lead were not available for SRM 1549a, SRM 1643f was used for these analytes. Quality control measures for analysis included validating the analytical sequence using the reference materials and by analyzing calibration verification standard every 10 samples. Samples were analyzed in randomized order.

2.4. Data processing and statistical analyses

ICP-MS data processing was conducted using MassHunter Workstation software version 4.6 (Agilent Technologies). Statistical analysis was performed using JMP 15 (SAS Institute, Cary, NC, USA). One-way ANOVA followed by pairwise mean comparisons using the Tukey-Kramer post-hoc correction (Tukey’s HSD) was used to determine significant differences (P < 0.05) between groups. Data presented as box-and-whisker plots were generated using Excel (Microsoft Co.; Redmond, WA, USA). For statistical purposes and for generation of the figures, data below the LOQ was imputed a value of the analyte’s LOQ/2. The percent coefficient of variation (%CV) was calculated as SD/mean*100%.

3. RESULTS

3.1. ICP-MS method performance

The ICP-MS detector response was highly linear (R²>0.995) for each analyte (Supplementary Table 1). The recovery of mineral micronutrients from SRM 1549a was found to be within 90-110% of the NIST certified values. Recovery of trace elements from SRM 1643f was between 95-105% of the NIST certified values. The LOQs for minerals were 0.35 mg/100 g for magnesium, 4.0 mg/100 g for phosphorus, 0.94 μg/100 g for selenium, and 14 μg/100 g for zinc. The LOQs for the environmental contaminant elements were found to be as follows: 9.9 μg/kg for arsenic, 2.3 μg/kg for cadmium, and 6.0 μg/kg for lead.

3.2. Distribution and variability of target minerals in PBMAs and milk

The box plots in Figure 1 show the distribution and variability of magnesium, phosphorus, selenium, and zinc for the eight PBMA types and for milk. Figure 1A displays the range of magnesium concentrations across PBMA types; the highest mean amount was in soy PBMAs (16.1 mg/100 g) and the lowest in oat PBMAs (4.9 mg/100 g). Soy PBMAs contained significantly greater (P<0.05) amounts of magnesium compared to the other PBMA types, except for hemp. Magnesium amounts in milk (10.7 mg/100 g) were not statistically different (P>0.05) compared to the mean amount in any PBMA type.

Figure 1.

Amounts of target minerals in eight different types of PBMAs. Box plots show the distribution of (A) magnesium, (B) phosphorous, (C) selenium, and (D) zinc in PBMAs and milk. Values not sharing the same letter are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

The distribution of phosphorus amounts in PBMAs is shown in Figure 1B. The highest mean value for phosphorus was in pea PBMAs (164 mg/100 g) and the lowest in almond PBMAs (10.4 mg/100 g). Phosphorus amounts in milk (91 mg/100 g) were statistically higher (P<0.05) than those in almond and coconut PBMAs, but they were not different compared to the other PBMA types (P>0.05).

Selenium content was found in quantifiable amounts in only three PBMA types (Figure 1C). The highest mean value was in pea PBMAs (4.4 μg/100 g), and the lowest quantifiable amount was in hemp PBMAs (1.5 μg/100 g). Five PBMA types (almond, cashew, coconut, rice, and soy) contained selenium concentrations less than the LOQ. Selenium amounts in milk (3.2 μg/100 g) were not statistically different (P>0.05) compared to the three PBMA types with values above the LOQ (hemp, oat, and pea).

Figure 1D displays the distribution and variability of zinc across PBMA types. The highest amount of zinc was in pea PBMAs (mean: 360 μg/100 g) and the lowest in coconut (mean: 39.5 μg/100 g). The amount of zinc in milk (mean: 392 μg/100 g) was not statistically different (P>0.05) compared to the mean value in any PBMA type, except for coconut PBMAs.

3.3. Occurrence of arsenic, cadmium, and lead in PBMAs and milk

The box plots in Figure 2 show the distribution and variability of total arsenic, lead, and cadmium for the eight PBMA types and for milk. Total arsenic was found in quantifiable amounts in only three of the eight PBMA types (Figure 2A). Mean total arsenic concentrations in rice PBMAs (15.8 μg/kg) were significantly (P<0.05) higher compared to the total arsenic in hemp and coconut PBMAs. All other PBMA types and milk contained insufficient concentrations of arsenic (<LOQ) to use for statistical comparisons. For lead, only five samples contained quantifiable amounts (Figure 2B). The mean amount of lead across all PBMA types was less than 4.2 μg/kg. Statistical comparisons were not performed for lead due to the low number of samples with a value above the LOQ. For cadmium, soy PBMAs contained the highest mean value (3.9 μg/kg); all other PBMA types and milk contained less than 2.3 μg/kg (Figure 2C).

Figure 2.

Amounts of target environmental contaminant elements in eight types of PBMAs. Box plots display the distribution of (A) total arsenic, (B) lead, and (C) cadmium in PBMAs and milk. Statistical comparisons for arsenic values were only done for coconut, hemp, and rice PBMAs. No statistical comparisons were done for lead values. Values not sharing the same letter are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

3.4. Variability of magnesium across PBMA brands and comparison to milk

Figure 3 shows the magnesium amounts for each PBMA type according to brand. Out of the eight PBMA types, seven had at least one brand significantly (P<0.05) different compared to the other brands sampled. Hemp PBMAs showed the greatest absolute difference in magnesium amounts between brands (~26 mg/100 g).

Figure 3.

Magnesium variability across different brands of PBMAs. Information on the number of samples per PBMA type are found in Table 1. Values not sharing the same letter within PBMA type are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

The variability (%CV) of magnesium within a PBMA brand is shown in Table 2. This table also displays the magnesium amount declared on the Nutrition Facts label (if declared), the percent difference between the measured and declared value, and a comparison to amounts in milk. Magnesium variability within a lot was less than 15% for all PBMA brands. Of the six brands with a magnesium amount declared on the Nutrition Facts label, four were within 20% of the declared value. The two brands with variation greater than 20% compared to the declared value were brand H2 (−93%) and brand P1 (+900%). In comparison to milk, only six PBMA brands contained higher amounts of magnesium, with soy PBMAs being the only type where all brands contained higher magnesium amounts compared to milk. Hemp PBMA brand H1 contained the highest magnesium, with amounts 157% higher than milk. Although there was a statistically significant (P<0.05) difference in the magnesium content across lots 1 and 2 of almond PBMA brand A1, the absolute value was only ~2 mg/100 g (Figure 4A).

Table 2.

Mean concentration of magnesium and associated variability in a market basket of eight types of PBMAs and milk.^a

PBMA type/brand code	Mean concentration (mg/100 g portion)	%CV	Nutrition Facts label value (mg/100 g portion)	% Difference From Nutrition Facts label value	% Difference from milk
Milk	10.7	1.6
Almond
Almond/A1	7.3	13.9	6.3	+15.9	−31.8
Almond/A2	7.5	3.6	6.3	+19.0	−29.9
Almond/A3	3.8	10.5			−64.5
Cashew
Cashew/CW1	8.9	9.2			−16.8
Cashew/CW2	6.4	4.7			−40.2
Coconut
Coconut/CT1	3.8	2.1			−64.5
Coconut/CT2	6.8	5.3			−36.4
Hemp
Hemp/H1	27.5	0.9	27.1	+1.5	+157
Hemp/H2	1.2	6.9	17.5	−93.1	−88.8
Oat
Oat/O1	2.8	13.6			−73.8
Oat/O2	4.6	9.0			−57.0
Oat/O3	10.8	1.4			+0.9
Pea
Pea/P1	13.0	2.5	1.3	+900	+21.5
Pea/P2	1.8	4.7			−83.2
Rice
Rice/R1	5.3	8.2			−50.5
Rice/R2	4.8	6.3			−55.1
Soy
Soy/S1	15.3	5.3			+43.0
Soy/S2	16.2	7.1			+51.4
Soy/S3	17.5	3.3	20.8	−15.9	+63.6

^aNo magnesium ingredients were declared on the Nutrition Facts label for any brand.

Figure 4.

Magnesium, phosphorous, and zinc variability across three production lots of brand A almond beverage. Values not sharing the same letter are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

3.5. Variability of phosphorus across PBMA brands and comparison to milk

For phosphorus, all PBMA types contained at least one brand with a significant (P<0.05) difference in the amounts of this mineral compared to the other brands within the same type (Figure 5). The PBMA brands that did not contain significantly (P>0.05) different amounts of phosphorus were almond brands A1 and A3, and soy brands S1 and S2. Hemp PBMAs showed the greatest absolute difference in phosphorus amounts (130 mg/100 g) between the two brands analyzed.

Figure 5.

Phosphorous variability across different brands of PBMAs. Information on the number of samples per PBMA type are found in Table 1. Values not sharing the same letter within PBMA type are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

Table 3 displays the variability of phosphorus within each PBMA brand. This table shows the phosphorus amount declared on the product’s Nutrition Facts label (if declared), the percent difference of the measured compared to the declared value, and any phosphorus-containing ingredient listed on the label. It also lists the percent difference of the phosphorus content in a PBMA brand compared to milk.

Table 3.

Mean concentration of phosphorous and associated variability in a market basket of eight types of PBMAs and milk.

PBMA type/brand code	Mean concentration (mg/100 g portion)	%CV	Nutrition Facts label value (mg/100 g portion)	% Difference from Nutrition Facts label value	% Difference from Milk	Phosphorous ingredient(s) on Nutrition Facts label
Milk	91.0	1.8
Almond
Almond/A1	10.2	6.0	8.3	+22.9	−88.8
Almond/A2	14.7	29.8			−83.8
Almond/A3	8.3	6.8			−90.9
Cashew
Cashew/CW1	25.5	22.1			−72.0	Tricalcium phosphate
Cashew/CW2	10.2	14.1			−88.8
Coconut
Coconut/CT1	24.8	5.0			−72.7	Dipotassium phosphate
Coconut/CT2	6.6	2.9			−92.7
Hemp
Hemp/H1	136	0.9			+49.5	Disodium phosphate, tricalcium phosphate
Hemp/H2	3.6	11.9	104	−96.5	−96.0	Tricalcium phosphate
Oat
Oat/O1	65.6	3.8			−27.9	Dipotassium phosphate
Oat/O2	93.3	7.3			+2.5	Tricalcium phosphate, dipotassium phosphate
Oat/O3	149	11.0	113	+31.9	+63.7	Tricalcium phosphate, potassium phosphate
Pea
Pea/P1	205	2.2			+125	Tricalcium phosphate, dipotassium phosphate
Pea/P2	124	6.8			+36.3	Dipotassium phosphate
Rice
Rice/R1	51.6	32.7	78.1	−33.9	−43.3	Tricalcium phosphate
Rice/R2	20.4	4.3			−77.6	Sodium phosphate
Soy
Soy/S1	29.8	15.2			−67.3
Soy/S2	36.0	9.8			−60.4
Soy/S3	94.8	1.5	91.7	+3.4	+4.2	Tricalcium phosphate

The majority of brands (12/19) formulated their products with at least one phosphorus-containing ingredient, with tricalcium phosphate being the most common. Almost all PBMA brands (16/19 brands) had phosphorus variability of less than 20% within a lot. The three brands that exceeded this phosphorus variability ranged from approximately 22-33%. Of the five brands with a phosphorus amount declared on the Nutrition Facts label, only one brand (S3) had less than a 20% difference. The brand with the greatest difference was hemp brand H2, with a value ~97% less than the declared amount. Only six PBMA brands contained phosphorus amounts higher than milk. The pea PBMA was the only type where both brands contained phosphorus in amounts higher than milk.

As shown in Figure 4B, there was significantly lower (P<0.05) phosphorus in lot 2 compared to lots 1 and 3 of almond PBMA brand A1. Although this difference reached statistical significance, the absolute difference was small (~1 mg/100 g).

3.6. Variability of selenium across PBMA brands and comparison to milk

Selenium was found in quantifiable amounts in a limited number of PBMA samples, so comparisons across brands could be made only for oat, hemp, and pea PBMAs (Figure 6). Oat PBMAs were the only type where selenium was found in quantifiable amounts for all brands. A small but statistically significant (P<0.05) difference in selenium content was found between oat PBMA brands O1/O2 and brand O3. For PBMA types hemp and pea, only one out of the two brands sampled contained quantifiable amounts of selenium. Because selenium in almond PBMAs was not above the LOQ, a comparison across lots could not be performed for this product.

Figure 6.

Selenium variability across different brands of PBMAs. Information on the number of samples per PBMA type are found in Table 1. Selenium values for other PBMA types sampled were <LOQ and are not shown. Values not sharing the same letter within PBMA type are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

Table 4 displays the variability of selenium content within a PBMA brand. It also lists the percent difference between the selenium content of a PBMA brand compared to milk. Selenium was not declared on the Nutrition Facts label for any PBMA type.

Table 4.

Mean concentration of selenium and associated variability in a market basket of eight types of PBMAs and milk.^a

PBMA Type/Brand Code	Mean Concentration (μg/100 g portion)	%CV	% Difference from Milk
Milk	3.2	10.1
Almond
Almond/A1	<LOQ		-
Almond/A2	<LOQ		-
Almond/A3	<LOQ		-
Cashew
Cashew/CW1	<LOQ		-
Cashew/CW2	<LOQ		-
Coconut
Coconut/CT1	<LOQ		-
Coconut/CT2	<LOQ		-
Hemp
Hemp/H1	3.2	30.6	0
Hemp/H2	<LOQ		-
Oat
Oat/O1	1.3	21.4	−59.4
Oat/O2	1.4	15.1	−56.3
Oat/O3	2.2	17.9	−31.3
Pea
Pea/P1	8.3	3.5	+159
Pea/P2	<LOQ		-
Rice
Rice/R1	<LOQ		-
Rice/R2	<LOQ		-
Soy
Soy/S1	<LOQ		-
Soy/S2	<LOQ		-
Soy/S3	<LOQ		-

^aNo selenium content claim was made on the Nutrition Facts label for any brand. No selenium ingredients were declared on the Nutrition Facts label for any brand.

Of the five PBMA brands with quantifiable amounts of selenium, three had variability of less than 20%. The other two PBMA brands had selenium variability of approximately 31% (H1) and 21% (O1). Only one PBMA brand (P1) contained selenium amounts higher than milk. One PBMA brand contained equal amounts of selenium (H1), while the three oat PBMA brands were lower in comparison to milk.

3.7. Variability of zinc across PBMA brands and comparison to milk

Except for rice PBMAs, all other PBMA types contained at least one brand with a significant (P<0.05) difference in zinc amounts compared to others brands within the same type (Figure 7). Almond PBMA samples had the greatest absolute difference in zinc amounts (430 μg/100 g) between brands.

Figure 7.

Zinc variability across different brands of PBMAs. Information on the number of samples per PBMA type are found in Table 1. Values not sharing the same letter within PBMA type are significantly different (P<0.05) as determined by Tukey’s HSD post hoc test.

The variability of zinc within a PBMA brand is displayed in Table 5. It also lists the percent difference of the zinc amount in a PBMA brand compared to milk. Only almond PBMA brand A3 declared zinc on the Nutrition Facts label (417 μg/100 g portion), which was formulated with zinc gluconate. The percent difference between the declared and measured value for brand A3 was +15.8%. No other PBMA brand declared zinc on the Nutrition Facts label. Of the samples with quantifiable amounts of zinc, the vast majority (16/18 brands) had a variability of less than 20%. The brands exceeding this variability were CT1 (23%) and O1 (37%).

Table 5.

Mean concentration of zinc and associated variability in a market basket of eight types of PBMAs and milk.^a

PBMA Type/Brand Code	Mean Concentration (μg/100 g portion)	%CV	% Difference from Milk
Milk	392	1.2	-
Almond
Almond/A1	53	12.5	−86.5
Almond/A2	89	4.2	−77.3
Almond/A3b	483	12.4	+23.2
Cashew
Cashew/CW1	159	3.4	−59.4
Cashew/CW2	87	14.9	−77.8
Coconut
Coconut/CT1	26	22.9	−93.4
Coconut/CT2	50	10.1	−87.2
Hemp
Hemp/H1	432	7.4	+10.2
Hemp/H2	<LOQ	-
Oat
Oat/O1	58	36.6	−85.2
Oat/O2	85	11.2	−78.3
Oat/O3	193	9.0	−50.8
Pea
Pea/P1	541	1.7	+38.0
Pea/P2	179	8.3	−54.3
Rice
Rice/R1	102	15.0	−74.0
Rice/R2	114	5.3	−70.9
Soy
Soy/S1	215	13.2	−45.2
Soy/S2	231	9.2	−41.1
Soy/S3	289	11.3	−26.3

^aOnly brand A3 made a zinc content claim on the Nutrition Facts label (417 µg/100 g portion). The percent difference between the declared and measured value was +15.8%.

^bZinc gluconate was declared in the ingredients.

Three PBMA brands contained zinc amounts higher than milk (A3, H1, and P1). Almond PBMA brand A3 was the only brand formulated with added zinc; it contained amounts ~23% higher than milk. As shown in Figure 4C, there was no significant (P>0.05) difference in zinc amounts across three lots of almond PBMA brand A1.

4. DISCUSSION

Multiple studies that have sought to evaluate nutritional differences between PBMAs and milk have been limited by the relative lack of data on the amount of key minerals in PBMAs (Chalupa-Krebzdak et al., 2018; Drewnowski, Henry, & Dwyer, 2021). As previously mentioned, it is critical to have data on minerals in PBMAs (including magnesium, phosphorus, selenium and zinc) to better assess how the increased consumption of these products affects the nutrient intake of a population; however, obtaining these values may be challenging since they are voluntarily declared in the Nutrition Facts label unless the nutrients are added as a nutrient supplement to the food or when a claim is made about the nutrient (such as a nutrient content claim). Because our data show significant differences in the amounts of minerals across PBMA types and brands, it is crucial to build available data on a wide variety of PBMA products.

One source of publicly available data for the mineral content of PBMAs that we were able to use for comparison was from the USDA’s FoodData Central Foundation Foods database (US Department of Agriculture, 2019). In addition to milk, this database included nutritional data for almond, oat, and soy PBMAs (all annotated as plain, unsweetened products). Besides this database, there have been several relatively recent studies from Brazil, New Zealand, and a number of countries in Europe that have performed elemental analysis on select PBMA types, including those made from almond, cashew, coconut, hemp, oat, rice, and soy (Astolfi, Marconi, Protano, & Canepari, 2020; Karasakal, 2020; Manousi & Zachariadis, 2021; Marquès et al., 2022; Pointke, Albrecht, Geburt, Gerken, Traulsen, & Pawelzik, 2022; Silva, Rebellato, dos Santos Carames, Greiner, & Pallone, 2020; Smith, Dave, Hill, & McNabb, 2022; Walther et al., 2022).

The mean values for magnesium content from FoodData Central for almond (6.8 mg/100 g), oat (5.9 mg/100 g), and soy (21.5 mg/100 g) PBMAs were comparable to the current analysis for almond (6.4 mg/100 g), oat (4.9 mg/100 g), and soy (16.1 mg/100 g) PBMAs. Astolfi et al. (2020) reported similar amounts of magnesium for the six PBMA types they analyzed, with values ranging from 6-17 mg/100 g. Several other researchers have found magnesium in PBMAs at amounts generally less than 20 mg/100 g (Manousi et al., 2021; Marquès et al., 2022) .

The phosphorus data from FoodData Central for almond (30 mg/100 g), oat (89 mg/100 g) and soy (69 mg/100 g) PBMAs were all largely similar to the present analysis. Although some variation in phosphorus amounts are present across studies, other authors have found phosphorus in PBMAs as low as ~2 mg/100 g for rice PBMAs up to 80 mg/100 g for soy PBMAs (Karasakal, 2020; Manousi et al., 2021; Marquès et al., 2022; Smith et al., 2022). There were phosphate-containing ingredients on the label of some PBMAs, including dipotassium phosphate. These phosphate salts are widely used as functional ingredients by the food industry (Cooke, 2017; Weiner, Salminen, Larson, Barter, Kranetz, & Simon, 2001). In the case of beverages, they can act as a pH buffer or stabilizer (Cooke, 2017), and therefore were not likely added for nutritional purposes.

A large number of PBMA samples were below the LOQ for selenium in both the FoodData Central database and in the current analysis. In both datasets, almond PBMA samples contained selenium content below the LOQ. Although all oat PBMAs samples in FoodData Central were below the LOQ (<2.5 μg/100 g), the mean value for the PBMAs in our study (1.5 μg/100 g) would also be under the quantitation threshold for the method used by USDA. Selenium content in soy PBMAs from FoodData Central contained a mean of 1.9 μg/100 g, whereas our samples were all below the current method’s LOQ. Still, the data from FoodData Central had a range of selenium concentrations (0-8.9 μg/100 g), and a median of 0 μg/100 g, indicating that a majority of soy PBMA samples had selenium content below the LOQ. Selenium has been found in non-detectable or non-quantifiable amounts in various PBMA types by several researchers (Astolfi et al., 2020; de Paiva et al., 2023; Godebo et al., 2023; Karasakal, 2020; Smith et al., 2022). Astolfi et al. (2020) found low amounts of selenium in coconut PBMAs (1.9 μg/100 g); these levels would be in amounts under the current method’s LOQ.

Mean zinc amounts in almond PBMAs from data reported by FoodData Central were 170 μg/100 g, similar to our measured value of 179 μg/100 g. However, it is important to bring attention to the wide range of zinc content in FoodData Central (range: 50-850 μg/100 g) and in the products we sampled (range: 39-537 μg/100 g). The values on the high end of the range for our samples were driven by the product brand that was fortified with zinc. For both soy and oat PBMAs, mean zinc content was similar across both the dataset from FoodData Central and the present analysis. There were limited additional data from the literature, but in an analysis by Astolfi et al. (2020) and Smith et al. (2022) of multiple types of PBMAs, soy PBMAs contained the highest zinc amounts (140-200 μg/100 g).

One advantage of our study is that we sampled at least two brands of each PBMA for analysis. Even within the same PBMA type, there were often significant differences in the mineral content across different brands. With the exception of one product that was fortified with zinc (brand A3), no other brands appeared to contain ingredients to fortify the product with magnesium, selenium, or zinc. This suggests that variation in mineral content across brands was related to factors other than formulation of the product with a nutrient supplement, perhaps differences in the type and amount of the raw ingredients used in manufacturing.

The occurrence of arsenic in rice has been widely studied, and several governmental agencies have established regulatory limits or provided guidance to industry for arsenic amounts in rice products (da Rosa et al., 2019; European Commission, 2023; Shannon et al., 2014; US Food and Drug Administration, 2020). In our analysis, rice PBMAs contained the highest concentrations of total arsenic, with hemp PBMAs being the second highest. Interestingly, both brands of hemp PBMAs were formulated with a rice-based sweetener (brown rice syrup), which may explain why these products were relatively similar in arsenic content to the rice PBMAs. One study found total arsenic concentrations in rice and hemp PBMAs of 8.5 μg/kg and 16 μg/kg, respectively, with no PBMA type containing arsenic at amounts higher than 16 μg/kg (Astolfi et al., 2020). A limitation of our analysis was that we only measured total arsenic and did not differentiate between the organic and inorganic species. Because inorganic arsenic is generally regarded as having greater toxicity compared to the major organic forms found in rice and rice products, speciation is especially critical when ascertaining toxic effects of samples with elevated amounts of arsenic (Jackson, Taylor, Karagas, Punshon, & Cottingham, 2012). There are currently no specific regulatory limits for arsenic in PBMAs. However, one value that can be used for comparison is the 100 μg/L total arsenic limit by Health Canada for ready-to-serve beverages (Health Canada, 2017). All of the products we sampled had arsenic amounts below this limit.

Amounts of the environmental contaminants cadmium and lead were generally found at low or non-quantifiable levels in PBMA samples. Cadmium occurrence in soy products has been previously reported (Kosečková et al., 2020), and the cadmium amounts we measured were typically found at relatively low concentrations. The mean value for cadmium in the soy PBMAs was 3.9 μg/kg, which is lower than the FDA’s limit for cadmium in bottled water (US Food and Drug Administration, 2022). Similarly, the mean amount of lead across PBMA types was below the FDA’s 5 μg/L limit for lead in bottled water, although a small number of individual samples (n=5) were above that level.

A limitation of our study is that our analysis was based on samples obtained within the greater Chicago, Illinois region. A follow-up study that samples a larger number of products from various regions and across multiple timepoints is warranted to better capture the variation of minerals and other elements in PBMA available on the US market. These results would also likely be different if sampled from another country with specific nutritional regulations for PBMA. For example, Health Canada policy allows fortification of PBMA with zinc (0.4 mg/100 mL), phosphorus (100 mg/100 mL), and magnesium (12 mg/100 mL) (Health Canada, 2022). Selenium fortification is not addressed in the Health Canada policy for PBMA fortification.

The purpose of this study was only to assess amounts of target minerals in these beverages, so we did not consider their bioavailability. The bioavailability of minerals are known to be affected by the compounds present in the food matrix (Weaver & Kannan, 2001). This is an important aspect of PBMAs to follow up with continued research, as certain anti-nutrients present in PBMAs (e.g., phytates) have the ability to bind to zinc and other minerals and reduce their bioavailability. It is also crucial to consider interaction between the minerals and other nutrients present in PBMAs. As an example, preliminary data using in vitro models has indicated the possibility that increasing amounts of calcium in a PBMA can in turn decrease the bioavailability of vitamin D, a fat soluble vitamin (Zhou, Zheng, Zhang, Zhang, He, & McClements, 2021). Follow up studies using animal models or a clinical trial will need to be conducted in order to assess whether the in vitro studies are biologically relevant.

5. CONCLUSION

In this study, we performed an analysis of a market basket of eight types of PBMAs for the minerals magnesium, phosphorus, selenium, and zinc, along with the environmental contaminants arsenic, cadmium, and lead. The results indicated that amounts of the minerals in the PBMAs largely depended on the product type and brand. The vast majority of PBMA types had differences in the amount of minerals across brands. There was statistically significant variation of magnesium and phosphorus across lots for one almond PBMA brand, but the absolute difference was low and likely not nutritionally relevant. Total arsenic was found in the highest amounts in rice and hemp PBMAs, compared to the other PBMA types. Cadmium and lead were found in quantifiable amounts in some PBMA samples; however, the mean concentrations were lower than the FDA’s limit for cadmium and lead in bottled water.

Although PBMAs generally contained lower amounts of target minerals compared to milk, there were a few exceptions. Pea PBMAs stood out as containing mean phosphorus, selenium, and zinc amounts comparable to or higher than milk. Oat PBMAs contained comparable amounts of phosphorus. Hemp and soy PBMAs contained mean magnesium amounts higher than milk. Only one PBMA brand (an almond PBMA) was formulated with zinc, which allowed for the zinc amounts to be comparable to milk. These data are significant from a nutritional standpoint because researchers can use such data to assess the impact of how incorporating PBMAs into the diet affects dietary intake of minerals.

Highlights.

• Elemental analysis was performed on 85 samples of plant-based milk alternatives (PBMAs).

• The minerals measured were magnesium, phosphorus, zinc, and selenium.

• Minerals varied across PBMA types and brands.

• The majority of PBMA types were lower in minerals compared to milk.

• Mean total arsenic amounts were highest in rice PBMAs.