Assessment of vitamin A, vitamin B₂, vitamin B₁₂, vitamin K, folate, and choline status following 4 months of multinutrient supplementation in healthy vegans: a randomised, double-blind, placebo-controlled trial

Abstract

Purpose

The aim of the MultiVeg study, a double-blind, randomised controlled trial (RCT), was to investigate the nutritional status of healthy vegans following 4 months of multinutrient supplementation.

Methods

A double-blind, RCT was conducted with 72 vegan adults (19–57 years) in Germany. Data on anthropometric parameters, dietary nutrient intake, and nutritional status were collected. The nutritional status of the participants was assessed at baseline and after 4 months. The results were compared between groups using ANCOVA. The results for vitamins and choline are presented here.

Results

After adjustment for baseline values, age, sex, and multiple testing, no significant between-group differences in biomarker concentration changes from baseline to 4 months were observed for vitamin A, retinol-binding protein, transthyretin, beta-carotene, methylmalonic acid, homocysteine, choline, total osteocalcin, carboxylated and undercarboxylated osteocalcin, and folate. In contrast, significant between-group differences in changes were observed for flavin adenine dinucleotide (FAD), serum vitamin B₁₂, holotranscobalamin, and the combined vitamin B₁₂ status indicator (cB₁₂) after adjustment.

Conclusion

A multinutrient supplement containing 82 µg of vitamin B₁₂ per day significantly positively affected vitamin B₁₂ blood biomarkers in healthy vegans.

Registration

This study was registered in the German Clinical Trials Register (DRKS00028151).

Supplementary Information

The online version contains supplementary material available at 10.1007/s00394-025-03814-7.

Introduction

Vegan diets have gained increasing popularity in many countries around the world. A recent review estimated the percentage of adults calling themselves vegan in Germany to be approximately 3% [1]. For vegan diets, several nutrients may be considered potentially critical such as vitamin B₁₂, vitamin D, vitamin B₂, iron, zinc, iodine, and selenium [2]. Furthermore, since conversion rates from α-linolenic acid to docosahexaenoic acid (DHA) are estimated to typically be low, beneficial effects of a direct intake of the long-chain omega-3 fatty acids eicosapentaenoic acid (EPA) and DHA for vegans are being discussed [3]. In addition, vitamin A is currently under discussion as a potentially critical nutrient in vegan diets [4]. Humans can obtain vitamin A in two ways: as preformed vitamin A (retinol or other retinoids) from animal products or supplements and as provitamin A (beta-carotene and other provitamin A carotenoids) from plants or other non-animal foods [5]. However, genetic polymorphisms in the beta-carotene 15,15’-monooxygenase (BCMO1) gene, the key enzyme responsible for converting beta-carotene into retinal, may potentially lead to an insufficient vitamin A status [6, 7], especially in vegans. In a small study with young, non-obese women, approximately 27–45% exhibited such genetic dispositions resulting in a 32–69% decreased beta-carotene conversion [8]. Affected individuals may hardly convert beta-carotene to vitamin A, leading to a vitamin A deficiency and hypercarotenaemia, if they adhere to a retinol-poor diet such as a vegan diet [9]. These low-converters, if following a vegan diet, might benefit from supplementation with retinol at doses of 500–1000 µg per day without exceeding the recommended dosage [10]. Moreover, the bioavailability of vitamin A depends on several factors, including intake of retinol and beta-carotene, concomitant fat ingestion, genetic factors, and food preparation techniques [11]. For vitamin B₁₂, there is a consensus among nutrition scientists that this vitamin should be supplemented when adhering to a vegan or near-vegan diet unless an adequate intake of this vitamin is ensured via fortified foods [12]. Apart from vitamin B₁₂, vegans may benefit from supplementing other nutrients like vitamin B₂, vitamin D, calcium, iron, iodine and selenium, depending on which vegan foods are habitually consumed, whether fortified products are chosen, as well as other circumstances such as sun exposure (for endogenous vitamin D synthesis) [2]. The primary source of vitamin B₂ (riboflavin) in mixed diets are dairy products [13]. It is therefore considered a potentially critical nutrient in vegan diets, a view supported by the evidence that vegans often had the lowest vitamin B₂ status in studies, compared to vegetarians and omnivores [2, 14, 15]. Finally, choline is also discussed as a potentially critical nutrient in vegan diets as its primary sources are animal-based foods like eggs, poultry, and meat [16, 17].

Currently, a variety of nutrient supplements specifically designed for vegans are available on the market. While a small multinutrient intervention study with vegans was conducted in the 1960s, to date [18], no randomised controlled trial (RCT) appears to have assessed the respective nutrient status of vegans before and after 4 months of multinutrient supplementation.

Therefore, it was the aim of the present study to assess the status of vitamin A and vitamin B₁₂ as well as vitamin B₂, vitamin K, folate, and choline in healthy vegans before and after 4 months of supplementation of a multinutrient supplement. The supplement included three distinct products: a multinutrient capsule containing vitamin A, vitamin B₂, vitamin B₁₂, vitamin D₃, vitamin K₂, calcium, iodine, iron, selenium, and zinc; an omega-3 supplement (providing EPA, DHA, vitamin D₃, and vitamin E); and a soya lecithin powder rich in choline. Additionally, we did a subgroup analysis (e.g., single vs. double dosage in the omega-3 supplement and separate reference values for ferritin levels in men and women). Examination of mineral constituents within the multinutrient supplement as well as the omega-3 supplement (containing vitamins D₃ and E) is outside of the scope of this analysis and will be published elsewhere.

Subjects and methods

Study design

The MultiVeg study is a single-centre (Bad Homburg), double-blind, RCT collecting data on anthropometric parameters, dietary nutrient intake, and nutritional status of vegan adults in Germany from October 2022 (baseline, t0) to February 2023 (4-month follow-up, t2).

Participants in the intervention group (INT) were instructed to take a multinutrient supplement (1 capsule daily), an omega-3 supplement (2 capsules daily, also containing vitamin D₃ and vitamin E) as well as a supplement in powder form (choline: ~320 mg/day) for 4 months, preferably with a meal. The control group (CON) received placebo supplements, which had the same organoleptic properties. The objective of this analysis is to assess the effect of the multinutrient supplement on blood biomarkers for vitamins and choline. Both groups were instructed to maintain their regular dietary habits throughout the intervention. Additionally, all supplements underwent laboratory analysis performed by an external laboratory (Gesellschaft für Bioanalytik mbH, Hamburg, Germany) to determine their nutrient content (active compounds) or lack thereof (placebo) (Tables 1, 2 and 3). Details regarding randomisation and group allocation can be found in Supplementary Fig. 1.

Table 1.

Declared and actual composition of the vegan multinutrient supplement (vitamins only, daily dose: 1 capsule)

Nutrient	Declared nutrient content (per capsule)	Actual nutrient contenta (per capsule)
Vitamin A (retinyl acetate and retinyl palmitate, 1:1 ratio)	500 µg	743 µg
Vitamin B2 (sodium-riboflavin-5-phosphate)	1.40 mg	1.03 mg
Vitamin B12 (methyl-, hydroxo-, and adenosylcobalamin, 1:1:1 ratio)	100 µg	82 µg
Vitamin D3 (cholecalciferol)	25 µg	26 µg
Vitamin K2 (all-trans menaquinone-7)	50 µg	45 µg
Calcium (calcium bisglycinate)	120 mg	129 mg
Iodine (potassium iodide)	150 µg	154 µg
Iron (iron bisglycinate)	6.0 mg	7.2 mg
Selenium (sodium selenite)	55 µg	70 µg
Zinc (zinc bisglycinate)	6.5 mg	6.4 mg

^a According to laboratory analysis (GBA Gesellschaft für Bioanalytik mbH, Hamburg)

Table 2.

Declared and actual composition of the vegan omega-3 supplement (daily dose: 2 capsules)–not objective of this Article

Nutrient	Declared nutrient content (per capsule)	Actual nutrient contenta (per capsule)
EPA	At least 75.0 mg	98.7 mg
DHA	At least 150 mg	171 mg
Vitamin D3 (cholecalciferol)	25 µg	36 µg
Vitamin E (d-α, d-β, d-γ- & d-δ tocopherol)	3.0 mg	3.7 mg

EPA: eicosapentaenoic acid; DHA: docosahexaenoic acid

^a According to laboratory analysis (GBA Gesellschaft für Bioanalytik mbH, Hamburg)

Table 3.

Declared and actual composition of the vegan choline supplement (daily dose: 11.5 g, equivalent to one level measuring spoon)

Nutrient	Declared nutrient content	Actual nutrient contenta
Soya lecithin	11.5 g	–
of which phosphatidylcholine	2.30 g	2.07 g
of which choline	300.0 mg	321.5 mg

^a According to laboratory analysis (GBA Gesellschaft für Bioanalytik mbH, Hamburg)

Compliance, defined as taking the study supplements exactly as instructed, was assessed based on self-reported data at t2.

The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the ethics committee of the Faculty of Medicine of the University of Giessen on June 28, 2022 (reference: 57/22).

Recruitment

Vegan subjects were recruited in August 2022 through various online channels including YouTube, the Research Institute for Plant-Based Nutrition participant database, newsletters, and social media platforms. Interested individuals completed an online recruitment questionnaire and were informed of the inclusion and exclusion criteria prior to their participation. Eligible participants were then selected based on inclusion and exclusion criteria (Fig. 1).

Fig. 1.

Flow chart of participants through the study with randomisation and group allocation in detail CON: control group; INT: intervention group; INT1: intervention group standard dosage; INT2: intervention group double dosage (the only vitamins affected by this subgroup intervention were vitamins D and E). Cells marked in grey indicate differences in the intake of the omega-3 supplement, which is the primary focus of another paper.

Inclusion criteria: Participants were required to be ≥ 18 years of age and to have followed a vegan diet (self-reported and verified by questions about frequency of consumption of animal foods), defined as the exclusion of meat, fish/seafood, dairy products, and eggs for ≥ 1 year. However, it was allowed and accepted if participants reported consuming meat and/or meat products and fish and/or fish products less than once a month, and dairy products and/or eggs and/or egg products (including those in bakery products, pasta or processed foods) no more than 3 times a month.

Exclusion criteria: Not accepted were smoking, pregnancy, lactation, diagnosis of any vitamin or mineral deficiency by a physician, a chronic medical condition (e.g., chronic gastritis, chronic gastrointestinal disease, thyroid disorders, or other absorption-related disorders), or regular use of medication (except contraceptive drugs). Additionally, participants were excluded if they had engaged in any of the following within three months prior to the study commencement:

• supplementation of vitamin A, beta-carotene, or selenium (any amount)

• average intake of > 2 Brazil nuts per day

• vitamin B₁₂ supplementation of > 1000 µg once per month, or > 200 µg once a week, or > 50 µg once per day, or use of vitamin B₁₂-fortified toothpaste more than once per day

• supplementation of vitamin B₂, vitamin D, vitamin E, zinc, iron, choline, or omega-3 fatty acids more than once a week (any amount)

• regular supplementation of iodine (any amount; iodised salt was permitted) or consumption of iodine-rich seaweed more than once a week (permitted: up to 1 teaspoon of nori flakes per day or up to 1 nori sheet per day).

To prevent any interference with the investigated multinutrient supplement, preference was given to participants taking few or no dietary supplements.

Study schedule

If study inclusion criteria were met, participants received detailed written information about the study. 180 individuals met the inclusion criteria. The individuals were contacted for scheduling appointments at the study center (t0, t2). To initially allocate 96 appointment slots, a randomised approach with a lottery method (51 were excluded) was adopted, aiming for a balanced distribution (1:1) of male and female participants. Due to scheduling difficulties, 41 could not participate (Fig. 1).

All participants gave their written informed consent prior to inclusion in the study. Participants (n = 88) were allocated to either the INT or the CON using the Clinical Trial Randomisation Tool by the National Cancer Institute (USA; https://ctrandomization.cancer.gov/tool/). Disclosure of group allocation to the investigator who conducted the primary statistical analysis (T.Z.) as well as to study participants was performed only after statistical analysis had been conducted.

For full study participation, the study subjects received an incentive of 100 Euros each (at t2) and were provided with their laboratory results.

Anthropometric measurements

Anthropometric measurements were performed in the fasted state and by trained staff. Body weight was assessed with the participants in underwear, without shoes, using scales (Seca 799, Hamburg, Germany). Height was assessed using a stadiometer (Seca 222).

Biomarker selection

As one of the main outcomes, for vitamin A four biomarkers were measured: Firstly, serum retinol as primary marker of vitamin A status reflecting recent intake [19]. Secondly, retinol binding protein (RBP) as a sensitive indicator because it transports retinol and decreases in vitamin A deficiency states. Thirdly, transthyretin (TTR) assists in retinol transport and reflects overall vitamin A status [20, 21]. Finally, beta-carotene, a precursor of vitamin A, is influenced by diet and conversion efficiency, making it a variable marker [19]. As a second main outcome, for vitamin B₁₂ also four parameters were investigated. Serum vitamin B₁₂ is the most commonly used marker to assess vitamin B₁₂ status, but it may lack sensitivity, particularly in cases of subclinical deficiency. In contrast, holotranscobalamin (holoTC) is considered a more sensitive marker as it represents the fraction of vitamin B₁₂ available for cellular uptake. Additionally, methylmalonic acid (MMA) is a functional marker that specifically increases in vitamin B₁₂ deficiency, reflecting impaired enzyme activity in vitamin B₁₂-dependent pathways. Finally, homocysteine (Hcy) levels also increase with vitamin B₁₂ deficiency, although they are less specific as they can be influenced by folate and other factors [22]. In combination, these four parameters allow a more accurate determination of vitamin B₁₂status, particularly when using the combined indicator cB₁₂. This approach integrates serum vitamin B₁₂, holoTC, MMA, and Hcy into a single diagnostic value, expressed as cB₁₂ = log10[(holoTC × B₁₂) / (MMA × tHcy)] – (age factor). This formula enhances the accuracy of deficiency diagnosis by accounting for the interactions between these biomarkers [23]. For vitamin B₂, flavin adenine dinucleotide (FAD), a coenzyme form of vitamin B₂ that reflects the functional status of this vitamin at the cellular level, was determined [24]. To evaluate the interaction of folate and vitamin B₁₂ with homocysteine metabolism, whole blood folate was analysed. Whole blood folate levels provide a stable and accurate measure of long-term folate status, reflecting folate stores in tissues over several months, making it a superior indicator compared to serum or plasma folate, which only reflects recent dietary intake [25, 26]. For choline, there is currently a lack of highly reliable biomarkers for assessing its status. Plasma choline is the most commonly used marker, but has significant limitations in that its levels do not consistently correlate with dietary intake or tissue stores [27]. Despite these limitations, plasma choline is still employed due to the absence of superior alternatives [28]. Furthermore, vitamin K-related bone health markers, such as osteocalcin (OC), were also measured. OC is a vitamin K-dependent calcium-binding protein that strongly binds to hydroxyapatite. OC is metabolically activated to its functional, carboxylated form (cOC). Under conditions of vitamin K deficiency and/or increased bone turnover, undercarboxylated osteocalcin (ucOC) is elevated in blood [29].

Blood samples

Venous blood samples were drawn in the morning after an overnight fast. To verify fasting, participants were asked about the last time they had eaten or consumed caloric beverages. With the exception of carboxylated and undercarboxylated osteocalcin, all blood parameters were analysed at MVZ Medical Laboratory Bremen GmbH. Serum samples were obtained by centrifugation at 2500 x g for 10 min within 1 h after blood sampling. Whole blood and serum samples were stored light-protected and frozen at − 20 °C for a maximum of 3 months and shipped to the laboratory on dry ice. Vitamin B₂ and folate were determined in EDTA whole blood, choline in EDTA plasma. Vitamin B₁₂, holoTC, MMA, Hcy, vitamin A, RBP, TTR, beta-carotene, and total OC were determined in serum. cOC and ucOC in serum samples were assessed using enzyme immunoassay (EIA) kits (Takara Bio Inc, Shiga, Japan) at BioTeSys–Nutritional CRO & Testing Laboratory GmbH, Esslingen. For the measurement of FAD, an isocratic HPLC in-house method with fluorimetric detection was applied. Vitamin B₁₂, OC, and folate were assessed with electrochemiluminescence immunoassays (Cobas, Roche Diagnostics, Mannheim, Germany). An enzymatic immunoassay (IBL International, Hamburg, Germany) was used for the measurement of holoTC, and a gas chromatographic in-house method (non-polar capillary separation column with single-ion monitoring detection) was used for MMA. Hcy and choline were determined by means of an in-house HPLC-tandem mass spectrometry method. For the measurement of the primary marker, vitamin A and beta-carotene, isocratic in-house HPLC-methods with UV (313 nm) and spectrophotometric detection (460 nm), respectively, were used. Turbidimetric immunoassays on an Optilite (The Binding Site, Schwetzingen, Germany) were used for the measurement of RBP (Trimero Diagnostics, Barcelona, Spain) and TTR (The Binding Site). The day-to-day coefficients of the analytical methods varied from 1–10%, depending on the method. The laboratory reference values used to evaluate nutrient status are given in Table 4.

Table 4.

Examined nutrient biomarkers and the laboratory’s reference ranges

Nutrient	Nutrient biomarkers [unit]	Reference ranges
Vitamin A	Vitamin A [mg/L]	0.2–1.2 [31]
	RBP [mg/dL]	2.08–4.25a
	TTR [g/L]	0.20–0.40 [32]
	Beta-carotene [µg/L]	150–1250b
Vitamin B2	FAD [µg/L]	199–382 [24]
Vitamin B12	Vitamin B12 [pg/mL]	197–771 [33]
	HoloTC [pmol/L]	21–123c [33]
	MMA [µg/L]	9–32 [33]
	Hcy [µmol/L]	< 13 [33]
	cB12	− 0.5–1.5d [23]
Choline	Choline [µg/L]	728–1287 [34]
Folate	Folate [µg/L]	212–534 [35]
Vitamin K	OC [ng/mL]	11–70e
	cOC [ng/mL]	no reference range available
	ucOC [ng/mL]	no reference range available

RBP: retinol-binding protein; TTR: transthyretin; FAD: flavin adenine dinucleotide; HoloTC: holotranscobalamin; MMA: methylmalonic acid; Hcy: homocysteine; cB₁₂: combined indicator for vitamin B₁₂ (calculated); OC: total osteocalcin; cOC: carboxylated osteocalcin; ucOC: undercarboxylated osteocalcin

^a CE-certified test, according to test manufacturer (Trimero Diagnostics, Barcelona, Spain)

^b From measurements in healthy volunteers, n = 30, Medical Laboratory Bremen

^c <35: vitamin B₁₂ deficiency is likely, 35–50: vitamin B₁₂ deficiency is possible, >50: vitamin B₁₂ deficiency is unlikely [33]

^d >1.5: elevated vitamin B₁₂, − 0.5–1.5: vitamin B₁₂ adequacy, − 1.5– − 0.5: low vitamin B₁₂, − 2.5– − 0.5: possible vitamin B₁₂ deficiency, <− 2.5: probable vitamin B₁₂ deficiency [23]

^e CE-certified test, according to test manufacturer (Roche Diagnostics, Mannheim, Germany)

Dietary assessment

Participants completed a 3-day weighed dietary record in the period of one week before to one week after the respective examination time points (t0, t2). These two dietary protocols were used to adjust for the mean intake values of each nutrient, measured as %DACH (individually achieved percentage of the dietary reference values for Germany, Austria and Switzerland [D-A-CH]) [30]. Additionally, a 3-day weighed dietary record at 2 months (t1) during the midpoint of the intervention was also completed by the participants. This dietary food record at t1 was primarily used to monitor any changes in dietary habits during the 4-month intervention period and to identify any potential deviations in nutrient intake for subsequent discussion.

All foods and beverages consumed as well as leftovers were weighed and recorded over 3 days using electronic kitchen scales. Participants also collected the packaging of convenience food products which they had consumed. When exact weighing was not possible (for example, in the case of eating outside or regarding table salt and spices), semi-quantitative household measurements (e.g., tablespoons, teaspoons, or cups) or photos were permitted for determination of portion sizes. For each time point, energy and nutrient intakes were calculated as the mean intake of the 3 days reported. The analysis of dietary nutrient intake data was conducted using the nutritional analysis software Optidiet Basic (Version 6.0.0.001, GOE GmbH, Linden, Germany). The nutrient composition of foods was based on the German Nutrient Database (Bundeslebensmittelschlüssel, BLS, version 3.02). Energy and nutrient content of convenience food products (ready-to-eat meals or snack foods) were estimated by recipe. Simulation was conducted using labelled ingredients and nutrient contents, including those added through fortification.

Choline intake was not assessed as the German Nutrient Database (BLS 3.02) does not contain data regarding choline content of foods.

Statistics

The sample size was calculated based on the primary parameter of the study (vitamin A) using the program G*Power (version 3.1.9.2). To our knowledge, no other study has compared vegans who supplemented vitamin A to vegans who did not supplement vitamin A. Therefore, the calculation was based on a study from Switzerland (n = 206) which investigated serum retinol levels in vegans and omnivores. For the calculation, a difference in serum retinol levels between omnivores (1869 ± 436 nmol/L) and vegans (1562 ± 408 nmol/L) was found to be 307 nmol/L [15]. Based on this assumption, a strong effect size (Cohen’s d) of approximately 0.7 was anticipated, and the required number of participants was determined to be 31 per study group, resulting in a power of 0.8 and a two-sided α of 0.05 (a priori two-sided t-test). Overall, a dropout rate of 30–40% (based on a previous study [36]) was expected. Thus, a total sample size of 44 per group was aimed for.

For continuous variables, baseline characteristics are presented as means ± standard deviation (SD) or median and interquartile range (IQR). Categorical variables are presented as numbers and percentages.

Shapiro-Wilk test was employed to test for non-normality in the data. Fisher’s exact test was utilised for comparing categorical variables, while independent t-test was applied to normally distributed, continuous variables and Mann-Whitney U test for those that were non-normally distributed (two-sided tests). To evaluate within-group changes in blood parameters, paired t-test was used for normally distributed and Wilcoxon test was used for non-normally distributed data (two-sided tests).

To assess differences in blood parameters between the two groups, independent t-test was employed for normally distributed data while Mann-Whitney U test was utilised for non-normally distributed data (two-sided tests). To assess intragroup differences, a paired samples t-test was employed for data that were normally distributed, while the Wilcoxon test was utilised for non-normally distributed data (two-sided tests). In addition, one-way analysis of covariance (ANCOVA) was performed, adjusting for baseline values [37]. ANCOVA analyses were repeated adjusting for other potential confounders (age, sex, and each nutrient’s intake). For dietary intake adjustments, the mean intake was calculated from the available records at both time points (t0 and t2). If one of these records was missing (n = 6), the mean intake was based on a single time point (t0 or t2).

Statistical significance was set at the 0.05 level. Bonferroni-Holm correction was applied to control for multiple testing [38]. All analyses were conducted using IBM SPSS Statistics (Version 28.0, Armonk, NY).

Results

For 72 participants blood values were available for both time points (t0, t2). These participants were included in the analysis. Of these, only 66 provided both 3-day weighed dietary records, while 67 participants submitted a mid-term dietary record at t1 (Table 6). Figure 1 shows trial details.

Table 6.

Median (IQR) nutrient intake (including fortification and supplements, excluding the supplements that were part of the study) of the participants of the MultiVeg study at baseline (t0), after 2 months (t1), and after 4 months (t2), measured as %DACH (individually achieved proportion of the DRV for Germany, Austria and Switzerland) [30]

Nutrients (%DACH)	CON (n = 33)a			INT			p-valuesb
	t0	t1	t2	t0 (n = 33)a	t1 (n = 34)a	t2 (n = 33)a	t0	t1	t2
Vitamin A (RE)c	82.9 (73.8)	77.9 (93.9)	74.2 (69.4)	75.7 (89.1)	57.0 (54.9)	64.8 (71.9)	1	1	1
Vitamin B2c	99.6 (62.3)	86.7 (49.6)	92.6 (42.0)	77.3 (50.5)	64.5 (38.3)	76.1 (39.0)	1	0.582	0.855
Vitamin B12c	12.9 (26.3)	11.3 (23.7)	8.9 (19.2)	10.1 (28.1)	2.3 (13.5)	6.7 (13.5)	1	1	1
Choline	NA	NA	NA	NA	NA	NA	NA	NA	NA
Folate (DFE)c	126.9 (88.9)	118.1 (58.0)	110.1 (58.4)	113.4 (70.1)	102.8 (62.0)	97.6 (44.6)	1	1	1
Vitamin Kc	318.1 (345.0)	357.4 (379.3)	291.7 (297.1)	311.4 (272.7)	243.6 (343.7)	228.6 (220.5)	1	1	1

^a Missing at t0 and t2: n = 4 (CON), n = 2 (INT). Missing at t1: n = 4 (CON), n = 1 (INT) ^b Mann-Whitney U test (non-normally distributed continuous variables) for between-group differences, adjusted by Bonferroni-Holm correction. ^c Median (IQR). NA = Choline intake calculation could not be performed, as there are no food composition data available for choline in the German Nutrient Database (Bundeslebensmittelschlüssel BLS 3.02). RE = retinol equivalent. DFE = dietary folate equivalent

Baseline characteristics

No significant differences were observed between the two groups in terms of baseline characteristics such as sex, age, body weight, height, BMI, and duration of adherence to a vegan diet (Table 5).

Table 5.

Baseline characteristics of the MultiVeg study participants

Variables [unit]	CON (n = 37)	INT (n = 35)	p-valuea
Men [n (%)]	18 (48.6)	19 (54.3)	1
Age at baseline [years]b	26.0 (10.0)	25.0 (7.0)	1
Body weight [kg]c	68.8 ± 12.0	68.0 ± 11.3	1
Height [cm]c	175.1 ± 9.2	175.9 ± 9.5	1
BMI [kg/m2]c	22.4 ± 3.7	21.9 ± 2.9	1
Vegan diet duration [years]b	3.5 (2.4)	2.5 (3.5)	1

BMI: body mass index

^a Fisher’s exact test (categorical variables), independent t-test (normally distributed continuous variables), or Mann-Whitney U test (non-normally distributed continuous variables). P-Value are adjusted for multiple testing by Bonferroni-Holm correction

^b Median (IQR)

^c Mean ± SD

Some participants reported occasional consumption of animal products or supplements in the recruitment questionnaire (Supplementary Table S2).

Only a small percentage reported taking any supplements within the study’s allowed limits (CON 18.9%, INT 20.0%), with most of them supplementing vitamin B₁₂ (CON 13.5%, INT 14.3%), vitamin D₃ (CON 8.1%, INT 5.7%), vitamin K₂ (CON 2.7%, INT 5.7%), iron (CON 2.7%, INT 0.0%), and omega-3 (CON 2.7%, INT 0.0%).

Compliance

In the INT, compliance was relatively high (71.4% for daily intake, according to [39]) for the multinutrient supplement. The remaining subjects followed the intake instructions six (22.9%) or five (5.7%) times a week. Compliance regarding the choline supplement (powder) was moderate: only 19 of 35 participants (54.3%) adhered to the daily intake instructions, eleven (31.4%) took it six times a week, three (8.6%) five times a week, and the remaining two participants (5.7%) only three times a week.

In the CON, compliance was comparable: 27 of 37 participants (73.0%) consumed the placebo daily, as instructed. The remaining subjects took it six (24.3%) or five (2.7%) times a week. For choline, 22 out of 35 participants (59.5%) adhered to the daily intake instructions, five (13.5%) adhered six times a week, eight (21.6%) five times a week, one (2.7%) four times a week, and one participant (2.7%) adhered three times a week.

Dietary vitamin intake (including fortified foods and supplements, excluding the supplements that were part of the study) at baseline (t0), after 2 months (t1) and after 4 months (t2).

Median dietary vitamin A intake (calculated as retinol equivalent, RE) was between 64.8 and 82.9% of the dietary reference values (DRV), expressed as %DACH (individually achieved proportion of the DRV for Germany, Austria and Switzerland). Median dietary vitamin B₂ intake was between 64.5–99.6%DACH in both groups at all three time points. In both groups and at all three time points, median dietary vitamin B₁₂ intake was ≤ 12.9%DACH.

No significant differences in dietary vitamin intake between the INT and CON were observed at any of the time points. Median dietary folate intake, calculated as dietary folate equivalent (DFE), was only slightly below the DRV in the INT at t2 (97.6%DACH) while the INT and CON met the DRV at all other time points.

Median dietary nutrient intakes of vitamin K exceeded the dietary reference values in both the INT and CON at all three time points (t0, t1, and t2) (Table 6).

Percentage of participants with circulating biomarkers outside of the reference ranges

At baseline, most of the participants (> 75%) of each group were within or above the reference ranges (for Hcy and MMA below or within the recommended value[s]) for most nutrient biomarkers, except for RBP (40% [CON], 46% [INT]), and for choline (57% [CON], 54% [INT]). At t2, most participants (> 75%) in each group still fell within or exceeded the reference ranges for most nutrients. However, a proportion of participants had values below the reference ranges, notably for RBP (76% [CON], 60% [INT]) and for TTR (30% [CON]). Details are shown in Supplementary Table S1.

Vitamin status at baseline (t0) and after 4 months (t2)

For all biomarkers of vitamin status, no significant differences between the two groups (intergroup differences) were observed at t0 and t2 (Table 7). Except for RBP and FAD, mean or median values for each of the other nutrient biomarkers assessed were within the reference ranges at t0 and t2 (Table 4). For RBP, the median levels in both groups were below the reference range (2.08–4.25 mg/dL) at both time points (indicating low vitamin A status). For FAD at t2, the mean value was below the reference range (199–382 µg/L) only in the CON.

Table 7.

Average nutrient biomarkers concentration at baseline (t0) and after 4 months (t2) and changes in biomarkers concentration of the MultiVeg study participants from baseline to 4 months (t2-t0)

Nutrient biomarker [unit]		CON (n = 37)	INT (n = 35)	p-values a	p-values (baseline-adjusted)b	p-values (adjusted for baseline, age, and sex)c	p-values (multi-adjusted)d
Vitamin A [mg/L]	t0e	0.47 (0.20)	0.48 (0.14)	1
	t2e	0.68 (0.23)	0.69 (0.20)	1
	t2-t0e	0.17 (0.15)	0.20 (0.17)	1	1	1	1
RBP [mg/dL]	t0e	1.90 (0.7)	2.00 (0.5)	1
	t2e	1.72 (0.53)	1.99 (0.53)	1
	t2-t0e	− 0.29 (0.30)	− 0.11 (0.29)	1	1	1	0.940
TTR [g/L]	t0f	0.26 ± 0.06	0.25 ± 0.04	1
	t2f	0.23 ± 0.05	0.23 ± 0.05	1
	t2-t0f	− 0.03 ± 0.03	− 0.02 ± 0.05	1	1	1	1
Beta-carotene [µg/L]	t0e	331.0 (427.0)	409.0 (347.0)	1
	t2e	416.0 (257.0)	369.0 (307.0)	1
	t2-t0e	23.0 (285.0)	1.0 (219.0)	1	1	1	1
FAD [µg/L]	t0e	246.0 (43.0)	232.0 (31.0)	1
	t2f	194.1 ± 24.6	212.8 ± 24.1	0.204
	t2-t0e	− 46.0 (24.0)	− 18.0 (32.0)	0.008	< 0.001	< 0.001	< 0.001
Vitamin B12 [pg/mL]	t0e	369.0 (266.0)	330.0 (204)	1
	t2e	293.0 (187.0)	376.0 (159.0)	1
	t2-t0e	− 73.0 (83.5)	38.0 (158.0)	0.029	0.011	0.016	0.013
HoloTC [pmol/L]	t0e	58.0 (37.0)	50.0 (25.0)	1
	t2e	72.0 (28.0)	83.0 (20.0)	1
	t2-t0e	11.0 (16.0)	34.0 (24.0)	0.007	< 0.001	< 0.001	< 0.001
MMA [µg/L]	t0e	18.0 (12.0)	20.0 (7.0)	1
	t2e	14.0 (6.0)	11.0 (5.0)	1
	t2-t0e	− 6.0 (7.0)	-8.0 (5.0)	1	0.048	0.051	0.069
Hcy [µmol/L]	t0e	10.0 (4.0)	10.1 (5.0)	1
	t2e	10.4 (4.3)	9.2 (4.8)	1
	t2-t0e	1.3 (2.5)	− 0.3 (4.5)	0.5	0.202	0.103	0.104
cB12	t0f	0.29 ± 0.58	0.17 ± 0.52	1
	t2f	0.34 ± 0.53	0.68 ± 0.45	0.495
	t2-t0f	0.06 ± 0.25	0.51 ± 0.44	< 0.001	< 0.001	< 0.001	< 0.001
Choline [µg/L]	t0e	746.0 (169.0)	751.0 (182.0)	1
	t2e	901.0 (238.0)	940.0 (242.0)	1
	t2-t0f	126.6 ± 175.7	224.5 ± 209.8	1	1	1	NAg
Folate [µg/L]	t0e	389.0 (82.0)	376.0 (117.0)	1
	t2e	332.0 (79.0)	342.0 (123.0)	1
	t2-t0e	− 61.0 (55.0)	-45.0 (63.0)	1	1	1	1
OC [ng/mL]	t0f	29.6 ± 9.8	32.0 ± 9.9	1
	t2f	28.2 ± 7.5	27.4 ± 8.6	1
	t2-t0f	− 1.4 ± 5.5	− 4.6 ± 6.3	1	1	1	1
cOC [ng/mL]	t0e	11.4 (6.6)	12.5 (5.1)	1
	t2e	11.9 (6.2)	11.7 (6.7)	1
	t2-t0e	− 0.19 (3.2)	0.10 (3.2)	1	1	1	1
ucOC [ng/mL]	t0e	10.6 (11.4)	11.2 (8.4)	1
	t2e	8.5 (6.1)	7.6 (6.5)	1
	t2-t0f	− 1.6 ± 5.4	− 3.4 ± 3.9	1	1	1	1

RBP: retinol-binding protein; TTR: transthyretin; FAD: flavin adenine dinucleotide; HoloTC: holotranscobalamin; MMA: methylmalonic acid; Hcy: homocysteine; cB₁₂: combined indicator; NA: not available; OC: total osteocalcin; cOC: carboxylated osteocalcin; ucOC: undercarboxylated osteocalcin

^a Independent t-test (normally distributed continuous variables) or Mann-Whitney U test (non-normally distributed continuous variables). ^b ANCOVA, adjusted for baseline blood values. ^c ANCOVA, adjusted for baseline blood values, age, and sex. ^d ANCOVA, adjusted for baseline blood values, age, sex, and mean dietary nutrient intake from t0 and t2 (vitamin A, RBP, TTR, and beta-carotene were adjusted for vitamin A intake [as retinol equivalent], FAD for vitamin B₂ intake, vitamin B₁₂, holoTC, MMA, Hcy and cB₁₂ for vitamin B₁₂ intake, folate for folate intake [as dietary folate equivalent] and OC, cOC, and ucOC for the intake of vitamin K, vitamin D, and calcium [the last two nutrients are covered in detail in other articles])

All p-values are adjusted for multiple testing by Bonferroni-Holm correction. Bold values indicate statistically significant results (p < 0.05)

^e Median (IQR). ^f Mean ± SD. ^g Choline cannot be adjusted for choline intake, as there are no food composition data available for choline in the German Nutrient Database (Bundeslebensmittelschlüssel BLS 3.02)

Additionally, significant intragroup differences (t0 vs. t2) were observed for nutrient biomarkers in both groups except for beta-carotene, holoTC, Hcy, cB₁₂, and cOC in the CON, and RBP, TTR, beta-carotene, vitamin B₁₂, Hcy, folate, and cOC in the INT (Supplementary Table S3).

Changes in vitamin status from baseline (t0) to 4 months (t2)

No significant between-group differences in biomarker concentration changes from baseline to 4 months were observed for vitamin A, RBP, TTR, beta-carotene, MMA, Hcy, choline, OC, cOC, ucOC, and folate. In contrast, significant between-group differences in changes were detected for all other nutrient biomarkers analysed (Table 7) after adjusting for baseline, age, and sex: FAD − 21 [95% confidence interval: − 29; − 13] µg/L, vitamin B₁₂ − 90 [− 134; − 45] pg/mL, holoTC − 14 [− 18; − 9] pmol/L, cB₁₂ − 0.42 [− 0.57; − 0.27]. For MMA, the between-group difference was 5 [2; 7] µg/L. The difference in mean changes in MMA between the groups was statistically significant only in the baseline-adjusted analysis, with a p-value of 0.048.

Discussion

In the present study, the multinutrient supplement improved the vitamin B₁₂ status of healthy vegans. Although it did not increase it, it also had an influence on vitamin B₂ status (FAD) as the decline in FAD levels was less pronounced in the INT compared to the CON. Despite these observations, the supplement did not significantly affect the blood levels of vitamin A, RBP, TTR, beta-carotene, choline, folate, OC, cOC, and ucOC.

After 4 months of multinutrient supplementation, there were no significant changes in the relevant parameters of vitamin A status between the groups. Both groups experienced a non-significant increase in serum retinol, but a decrease in RBP. However, TTR changes were similar in both groups. Overall, RBP and TTR parameters were consistently low before and after the intervention in comparison to the reference values (Tables 4 and 7), suggesting an overall poor vitamin A status.

Nevertheless, beta-carotene increased more, but non-significantly in the CON than in the INT. This can be explained by a higher dietary intake of beta-carotene: At all three time points median dietary nutrient intakes of RE (mostly beta-carotene [19]) were slightly higher in the CON than in the INT. Administering a multinutrient supplement with 743 µg of vitamin A (measured, declared 500 µg) may slightly preserve the vitamin A status, but it might not significantly improve vitamin A status within 4 months.

For vitamin B₂, there are several plant sources, e.g., nuts, mushrooms, legumes, yeast flakes, or fortified plant-based dairy alternatives [2]. Despite conflicting study results, vitamin B₂ may be considered a critical nutrient in vegan diets. In a cross-sectional study from Switzerland, there was no significant difference in vitamin B₂ plasma levels between vegans, lacto-ovo-vegetarians (LOV), and omnivores [15]. In contrast, a cross-sectional study from Germany demonstrated lower plasma vitamin B₂ levels among vegans in comparison to omnivores. However, the median concentration in both groups exceeded the established reference value [40]. Conversely, in a cross-sectional study from Austria from 2006, approximately 30% of vegan participants had vitamin B₂ levels below the reference value. This was in contrast to around 10% observed among both LOV and omnivorous participants [41]. Comparably, in the present study, the CON showed a decrease in FAD levels, falling just below the reference range (199–382 µg/L). Moreover, a decline in vitamin B₂ status from t0 to t2 was observed in both the INT and CON. However, the reduction in status was significantly less pronounced in the INT, suggesting that the multinutrient supplement may have attenuated the decrease in vitamin B₂ status compared to placebo. One explanation might be a lower intake as both groups reduced their vitamin B₂ intake during the study from t0 to t1. However, it rebounded in both groups at t2. Changes (possibly related to infection, genetics or exercise) in FAD levels are another hypothetical explanation, but these have not been well studied [42]. Furthermore, the supplement’s dosage (1.03 mg vitamin B₂) might not be sufficient to maintain the vitamin B₂ status in the long term, even though the INT was still within the reference range due to a lower decrease compared to the CON. Thus, monitoring over an extended period is recommended.

For vitamin B₁₂, the intergroup difference of the serum vitamin B₁₂ concentration, a less specific and sensitive marker of vitamin B₁₂ status [33], was significant (INT: +38 pg/mL, CON − 73 pg/mL). Similarly, holoTC, an early indicator of availability of vitamin B₁₂ to body cells, increased significantly more in the INT compared to the CON (+ 34.0 pmol/L, CON + 11.0 pmol/L). Although the CON did not receive vitamin B₁₂, the rise in holoTC levels in the CON may be due to various factors. The dietary intake of vitamin B₁₂ on average was 9–12%DACH (CON) and 2–10%DACH (INT) (Table 6), thus, might not explain a rise in holoTC. However, it also cannot be ruled out that vitamin B₁₂-enriched foods were consumed beyond those recorded in the dietary protocols, leading to slight increases in holoTC levels. Both groups maintained holoTC levels within the reference range throughout the study, indicating sufficient vitamin B₁₂ stores. Serum vitamin B₁₂ levels may decrease before holoTC changes due to enterohepatic recycling maintaining holoTC concentrations or may even slightly increase them until stores are considerably depleted [22, 43, 44]. Additionally, a negative vitamin B₁₂ balance in the CON may affect holoTC levels differently than an outright vitamin B₁₂ deficiency. An abrupt cessation of dietary vitamin B₁₂ intake can occur when transitioning from an omnivorous to an unsupplemented vegan diet. In the context of the current study, some participants may have stopped taking their usual supplements before the recruitment phase in order to fulfil the inclusion criteria and participate in the study. This scenario is possible, given that data suggest 82–92% of vegans in Germany consume vitamin B₁₂ supplements or vitamin-B₁₂-fortified foods [40, 45]. Moreover, MMA levels decreased (only significant in the baseline-adjusted model) and Hcy decreased more in the INT compared to the CON (but not statistically significant) and over the 4-month period. Both folate and vitamin B₁₂ are crucial for the regulation of homocysteine levels [46]. Two outliers in the INT exhibited notably elevated Hcy levels at t2 and had one of the lowest serum folate levels of all participants. Overall, as the multinutrient supplement contained no folate, no significant differences were found for the vitamin between the groups at either time point (t0 and t2) and in terms of changes (t0 to t2), as expected. The median Hcy levels in both groups were within the reference range before and after the intervention. For both groups, values for cB₁₂ remained within the reference range indicating vitamin B₁₂ adequacy (-0.5 to 1.5 [23]), with the INT demonstrating a statistically significant improvement over time in comparison to the CON. Vitamin B₁₂ is regarded as the most critical nutrient in vegan diets [47], so the observed change in vitamin B₁₂ status with vitamin B₁₂ supplementation compared to placebo was to be expected. It should be noted that a small percentage of the participants reported taking vitamin B₁₂ (CON 13.5%, INT 14.3%) within the study’s allowed limits (Recruitment Section, Supplementary Table S2). However, this intake was not confirmed in the 3-day weighed dietary records, possibly because the subjects consumed the supplement irregularly or just not during the recorded days.

In vegan diets, vitamin K is not considered a critical nutrient [2] as there are various plant sources of this vitamin, e.g., dark green leafy vegetables or soya milk [48]. The results of the present study demonstrate that the OC levels in the INT declined more compared to the CON. However, this difference was not significant, especially when adjusting for both baseline blood values, intake of vitamin K, vitamin D, and calcium, age, and sex. Because of their roles in bone health [49–51], vitamin D and calcium were considered as confounders and future articles will focus on both nutrients. Nevertheless, vitamin D supplementation had a significant effect on 25-OH-D3 compared with placebo (data not shown). Besides, no significant difference was observed for cOC or cOC changes over time between the INT and CON. In the INT, ucOC levels decreased more than in the CON, but not significantly. However, these findings are consistent with the observation that vitamin K₂ supplementation decreases serum levels of ucOC, which indicates an increase in cOC [29]. Concerning the within-group changes of OC, cOC, and ucOC levels, no significant differences were observed between t0 and t2. Both groups were within the laboratory’s reference range for OC levels, both at t0 and t2, while there are none for cOC and ucOC. This suggests that the influence of the supplement was not clearly demonstrable or may have been influenced by other factors. For instance, there was no need for vitamin K₂ supplementation as the dietary vitamin K intake was already sufficient.

Choline is currently discussed as a potentially critical nutrient in vegan diets [16] because animal foods such as eggs and meat are rich dietary sources [17]. The human body can synthesize choline, but there are various genetic polymorphisms that can limit its endogenous synthesis capability [52]. The increase in serum choline levels in the INT was greater, but not significantly. One reason for the increase in serum choline levels in the CON could be a higher dietary intake which we were unfortunately unable to assess. It should be noted that some participants in both groups reported occasional consumption of animal products in the recruitment questionnaire (Supplementary Table S2), particularly eggs (31.4% CON, 20.0% INT < 1/month; 2.7% CON, 17.1% INT 1–3/month), which may explain the observed choline levels. However, this intake was not confirmed in the 3-day weighed dietary records. Another reason might be a greater endogenous synthesis capability in the CON. In addition, compliance was only moderate in both groups with 54–60% following the study protocol. This could have caused the lack of significant differences between the groups. A sufficient folate status is linked to higher plasma choline levels [53]. As there was no significant difference in the folate status, this cannot explain the lack of difference. Furthermore, both groups were within the reference range for plasma choline before and after the intervention, underlining that neither plasma choline levels nor other parameters are reliable biomarkers for assessing choline status [27, 54]. Alternatively, this may suggest that a vegan diet provides sufficient dietary choline to meet dietary requirements. However, further explanation or investigation is needed to fully understand this finding.

Strengths and limitations

The MultiVeg study makes an important contribution to the field of vegan nutrition because, to our knowledge, it is the first RCT testing a multinutrient supplement in healthy vegans. This is important as there is a broad variety of multinutrient supplements for vegans on the market. Another strength is the similarity of both study groups with no significant differences in baseline characteristics. A further strength is the detailed dietary assessment.

A limitation lies in the recruitment process. Participants were mainly recruited through posts on social media (convenience sample) and via a YouTube live event which included a speaker of the funding company which also markets the multinutrient supplements. This should be kept in mind when interpreting the results as it appears to limit the generalizability of the results.

Furthermore, most participants appear to have an adequate nutrient status at baseline, with over 75% of individuals in each group within or above the reference ranges for most nutrients (below or within the recommended value[s] for Hcy and MMA). This pre-existing adequate nutrient status likely contributed to the observation that, after the intervention, still over 75% of participants in both groups remained within the reference ranges. However, in the INT, fewer participants fell outside these ranges, with many shifting further up within the reference intervals, while in the CON, more participants fell outside the reference ranges or shifted downward in regards of their nutrient status, although many remained within the acceptable limits (Supplementary Table S1).

Another limitation is the study protocol which required a daily intake of three capsules and a powder, which may have had a detrimental impact on participants’ compliance and adherence to the intervention regimen. This is evidenced by the moderate compliance with the choline supplement. This moderate adherence highlights the potential challenges associated with following a vegan diet, which may require multiple supplements to meet nutrient needs, and may have influenced the study’s findings on serum choline levels. Given that it is not yet fully established that choline is a critical nutrient for vegans, it is possible that the inclusion of this nutrient in the intervention could be reconsidered. Without the choline powder, participants would have only needed to take two supplements, which could have potentially improved compliance. In contrast, compliance in capsule form was high (> 94% of participants in both groups taking it 6–7 times a week), highlighting that the capsule format may be more feasible for regular use than a powdered supplement. Also, compliance was only assessed at the end of the study and not at the midpoint, which may limit our ability to fully evaluate adherence throughout the entire intervention period.

A further limitation of our study could be the 16-week duration, which may not have been long enough to observe the potential long-term impact of the supplement on every analysed nutrient. However, for key parameters like vitamin B₁₂ and vitamin A, previous studies have demonstrated that significant changes can also be observed even within 12–16 weeks [36, 55–57], suggesting that the study duration was sufficient to assess the primary outcomes.

Furthermore, the inclusion of participants without diagnosed deficiencies may have constrained the capacity to evaluate the supplements’ efficacy in addressing deficiencies. To address these limitations, we mainly included interested individuals taking no or few supplements prior to the study. Nevertheless, the participants mostly had a normal baseline status. Therefore, it was anticipated that there might be little to no change in many nutrient levels, which aligns with the findings for several nutrients for which no significant impact was observed.

In our study, 3-day weighed dietary records were utilised, with participants using their own kitchen scales. This could introduce measurement bias due to potential inaccuracies in reported portion sizes, influencing the reliability of reported dietary nutrient intake data. Furthermore, currently there is no database that includes choline concentrations in foods in Germany. Therefore, no statements could be made regarding dietary choline intake in this study. Additionally, a considerable number of processed vegan products consumed by the participants are not yet included in the German Nutrient Database (BLS). Consequently, we were obliged to simulate these foods based on their declared ingredients and provided information regarding their energy and macronutrient content. This method permitted the estimation of micronutrient content, which is often absent from product labels. However, this approach is susceptible to inaccuracies. Nevertheless, it represents the most viable option within the context of dietary surveys [58].

Conclusions

In conclusion, the daily intake of the investigated multinutrient supplement with a dosage of 82 µg vitamin B₁₂ appears to constitute a reliable approach to increase vitamin B₁₂ status among healthy vegans. As there are currently no standardised reference values for sufficient vitamin B₁₂ intake via supplements, these results might help to establish vitamin B₁₂ intake recommendations for vegans. In contrast, no clear positive impact of the multinutrient supplement on nutrient status could be demonstrated for all other examined vitamins and choline, leading to the assumption that the participants were already adequately supplied with these nutrients. The overall poor vitamin A status of the participants supports the suggestions that vitamin A should be included in the group of potentially critical nutrients in a vegan diet and the dose of a vitamin A supplement should be higher than in the present supplement (i.e., >  750 µg per day). The lack of improvement in vitamin A and B₂ status might be due to interactions with other nutrients that may decrease their bioavailability in a multinutrient supplement and/or insufficient amounts in the supplement. Although our data show that well-nourished vegans do not benefit from additional multivitamin supplementation, it may still be worth considering as a prophylactic measure. However, the feasibility of administering three different supplements per day should be reconsidered, as compliance may be a challenge in practical settings.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

T.Z. drafted the manuscript. S.W. and M.K. conceptualized the study. C.K. and T.Z. were responsible for the implementation of the study. A.S. performed the laboratory analyses. Statistical analyses were conducted by T.Z. and C.K. S.W., M.K. and G.P.E. provided scientific supervision and critically revised the manuscript. All authors reviewed, edited, and approved the final version of the manuscript.

Funding

Open Access funding enabled and organized by Projekt DEAL. This research was mainly funded by Watson Nutrition, Berlin. Neither Watson Nutrition nor any advisor of Watson Nutrition had any role in the data collection, data analysis or writing of this article.

Data availability

The dataset is available upon reasonable request from the corresponding author.

Declarations

Conflict of interest

T.Z., S.W., C.K., A.S., and G.P.E. declare no conflict of interest. M.K. is the managing director of the Research Institute of Plant-Based Nutrition (IFPE) which received funding from Watson Nutrition for conducting the study.

Ethical approval

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the ethics committee of the Faculty of Medicine of the University of Giessen (57/22).

Consent to participate

Written consent was obtained from all subjects involved in the study.