Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial

Clinical Nutrition xxx (2014) 1e8

Contents lists available at ScienceDirect

Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu

Randomized control trials

Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial Melanie Bähr a, Anita Fechner a, Michael Kiehntopf b, Gerhard Jahreis a, * a b

Friedrich Schiller University Jena, Institute of Nutrition, Department of Nutritional Physiology, Dornburger Str. 24, 07743 Jena, Germany Jena University Hospital, Institute of Clinical Chemistry and Laboratory Medicine, Erlanger Allee 101, 07747 Jena, Germany

a r t i c l e i n f o

s u m m a r y

Article history: Received 9 December 2013 Accepted 24 March 2014

Background & aims: The objectives of this study were to assess whether 25 g/d lupin protein, integrated into a mixed diet, might affect cardiovascular risk factors and whether L-arginine was responsible for these effects. Methods: Seventy-two hypercholesterolemic subjects participated in the randomized, controlled, double-blind three-phase crossover study. They were assigned to three diets with 25 g/d lupin protein (LP), milk protein (MP) or milk protein plus 1.6 g/d arginine (MPA) each for 28 d in a random order interrupted by 6-week washout periods. Lupin protein and the comparator milk protein were incorporated into complex food products (bread, roll, sausage, and vegetarian spread). Arginine was administered via capsules. Sixty-eight subjects were included in final analyses. Results: Compared with MP, LDL cholesterol was significantly lower after LP. Compared with MP and MPA, homocysteine was significantly lower after LP. Compared with baseline, concentrations of total, LDL, and HDL cholesterol significantly decreased after LP and MPA. Triacylglycerols and uric acid significantly decreased after LP. The relative changes in total and LDL cholesterol were significantly greater for subjects with severe hypercholesterolemia (>6.6 mmol/L) than those with moderate hypercholesterolemia (5.2e6.6 mmol/L). Conclusions: The present study showed for the first time that incorporation of 25 g/d of lupin protein into a variety of complex food products lowers total and LDL cholesterol, triacylglycerols, homocysteine, and uric acid in hypercholesterolemic subjects. The hypocholesterolemic effect is stronger in subjects with severe hypercholesterolemia. Arginine might be responsible for some, but not all of the beneficial effects of lupin protein. This trial was registered at http://clinicaltrials.gov (study ID number NCT01598649). Ó 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Keywords: Lupin protein Milk protein Arginine Plasma lipids Hypercholesterolemic subjects Serum amino acids

1. Introduction The potential of lupin protein to beneficially affect blood lipids and blood pressure has been investigated only in a few human intervention studies.1e4 In all these studies, lupin protein was provided additionally to the normal diet in the form of a protein Abbreviations: LP intervention, intervention with lupin protein; MP intervention, intervention with milk protein; MPA intervention, intervention with milk protein plus arginine; FFP, 3-d Food Frequency Protocol; hs-CRP, high-sensitivity Creactive protein. * Corresponding author. Tel.: þ49 3641 949610; fax: þ49 3641 949612. E-mail addresses: [email protected] (M. Bähr), anita.fechner@uni-jena. de (A. Fechner), [email protected] (M. Kiehntopf), [email protected] (G. Jahreis).

drink1,4 or a dietary bar.2,3 In all but one study, high daily doses of approximately 35 g lupin protein were given. So far, lupin protein has not been tested within a mixed diet by incorporating a modest amount (25 g/d) into a variety of well-designed food products with a complex composition and improved sensory properties. Moreover, the underlying mechanisms that are responsible for the lipidand blood pressure-lowering activities of lupin protein have not yet been elucidated.5 Accounting for more than 10% of total amino acids, L-arginine is abundant in lupin protein and thus suspected to be one of its bioactive compounds. Administration of additional arginine in amounts of 3 g/d up to 17 g/d was found to lower blood pressure6 and total7,8 as well as LDL cholesterol8 in humans. Thus, the objective of the present study was twofold: (1) to investigate the effect of 25 g/d lupin protein on blood lipids, the

http://dx.doi.org/10.1016/j.clnu.2014.03.008 0261-5614/Ó 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

2

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

amino acid profile, homocysteine, uric acid, high-sensitivity Creactive protein (hs-CRP), as well as on blood pressure, incorporated into a variety of complex food products (bread, roll, sausage, and vegetarian spread), as part of a mixed diet; and (2) to test the effects of an arginine-supplemented milk protein diet in comparison with two diets enriched with lupin protein and milk protein. Therefore, a randomized, double-blind crossover intervention study was conducted in hypercholesterolemic subjects consuming three different intervention diets for 28 d each: (i) a diet with 25 g/ d of lupin protein or (ii) milk protein, or (iii) the milk protein diet supplemented with 1.6 g/d arginine. 2. Materials and methods 2.1. Subjects In and around Jena, 132 volunteers between 18 and 80 years of age were recruited. Eligibility criterion was a total cholesterol concentration of 5.2 mmol/L at screening. Exclusion criteria were the intake of lipid-lowering drugs or nutritional supplements that might potentially influence lipid metabolism and intolerance, an allergy, or strong aversion to any food ingredient used in the study products. In addition, breast-feeding mothers and pregnant females were excluded. Thus, 72 eligible participants (41 females, 31 males) were invited to an in-person meeting. Here, participants were offered essential study-relevant information in oral and written forms. Written informed consent was obtained from all subjects before the start of the study. The study was registered at ClinicalTrials.gov as NCT01598649 (National Institutes of Health)

and approved by the Ethics Committee of the Medical Faculty of the Friedrich Schiller University, Jena (no.: 2607-07/09). 2.2. Study design The randomized, double-blind crossover study consisted of three intervention periods of 28 d each, separated by 6-week washout periods. During the intervention periods, subjects received either: (i) lupin protein (LP intervention); (ii) milk protein (MP intervention); or (iii) milk protein plus arginine (MPA intervention) in a random order. Before commencement, subjects were randomly assigned to one of the three randomization groups, A, B, or C using computer-generated random numbers (Fig. 1). The research assistants who performed the randomization did not have access to any information regarding demographic or laboratory characteristics of the subjects. Moreover, all study products were labeled with numeric codes and all research assistants as well as the participants were blinded to group assignments. The study was conducted between June and December 2012 at the Friedrich Schiller University Jena (Institute of Nutrition, Department of Nutritional Physiology, Jena, Germany). 2.3. Study products and capsules In the present study, 25 g lupin protein isolate was incorporated into complex study products and administered as part of a mixed diet for 28 d. The amount of protein and intervention time were based on the results of a previous study of our work group,4 where

Assessed for eligibility (n = 132) Excluded (n = 60) - Not meeting inclusion criteria (n = 19) - Declined to participate (n = 22) - Loss of contact (n = 19) Randomized (n = 72)

Allocated to randomization group A (n = 24)

Allocated to randomization group B (n = 24)

Allocated to randomization group C (n = 24)

Received LP intervention (n = 24)

Received MP intervention (n = 24)

Received MPA intervention (n = 24)

Drop Out (n = 1) Medical reasons Received MP intervention (n = 24)

Received MPA intervention (n = 23)

Received LP intervention (n = 24)

Received MPA intervention (n = 24)

Received LP intervention (n = 23)

Received MP intervention (n = 24)

Analyzed (n = 22) - Excluded from analysis (n = 1)

Analyzed (n = 24) - Excluded from analysis (n = 0)

Drop Out (n = 1) Medical reasons Analyzed (n = 22) - Excluded from analysis (n = 1)

Fig. 1. CONSORT 2010 flow diagram of the study participants. LP intervention, intervention with lupin protein; MP intervention, intervention with milk protein; MPA intervention, intervention with milk protein plus arginine.

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

25 g/d lupin protein isolate in the form of protein drinks positively affected blood lipids after 28 d. The lupin protein was produced from the seeds of Lupinus angustifolius cv. Boregine as described by D’Agostina et al.9 and provided by the Fraunhofer Institute for Process Engineering and Packaging (Fh-IVV, Freising, Germany) in the form of its isolate type E. This protein isolate contained 81.6  1.3% protein (nitrogen  5.8) and low quantities of water (5.7  0.0%), ash (4.4  0.0%), fiber (6.0  0.4%), and fat (1.4  0.1%) in fresh matter. A milk protein isolate consisting of a mixture of 75%:25% (wt %:wt%) sodium caseinate (EM7, DMV International, Veghel, The Netherlands) and whey protein Megglosat HP (ME, Meggle, Wasserburg, Germany) was chosen as the comparator. Sodium caseinate was made up of 88.9  1.0% protein (nitrogen  6.38), water (5.9  0.0%), ash (4.4  0.0%), fiber (2.8  0.1%), and fat (0.6  0.0%) in fresh matter. Megglosat HP consisted of 84.8  2.6% protein (nitrogen  6.38), water (6.6  0.1%), ash (4.9  0.0%), fiber (3.6  0.0%), and fat (1.3  0.2%) in fresh matter. The amino acid composition is shown in Supplemental Table 1. The test proteins were incorporated into four study products (Table 1), which had to be consumed daily: 70 g of bread, 160 g of rolls, 40 g of scalded sausage, and 30 g of a vegetarian spread, totaling circa 25 g test protein isolate/d. Subjects were instructed to consume these study products in place of other food items of their habitual diet. The four study products were provided frozen in single 3-d packages. The bread was baked at the Ludwig Stocker Hofpfisterei GmbH (Munich, Germany). The rolls were produced at the Backstube Wünsche GmbH (Gaimersheim, Germany). Both the scalded sausage and the vegetarian spread were produced at the Metzgerei Boneberger GmbH (Schongau, Germany) with the collaboration of the Fraunhofer Institute for Process Engineering and Packaging (Fh-IVV, Freising, Germany). Color, taste, texture, and nutrient composition of the control products with milk protein were kept as similar as possible to those of the study products with lupin protein in order to maintain blinding. The nutrient composition of the study products and protein isolates was analyzed by applying standard methods with reference to the Association of Analytical Communities10 and the European Community Directive.11 During the LP intervention, the study products with lupin protein provided 2.5 g/d arginine. In contrast, in the MP intervention, the study products with milk protein provided 0.9 g/d arginine. During the MPA intervention, capsules containing 1.6 g/d of arginine (L-arginine hydrochloride) were administered additionally to the milk protein products. The MPA intervention also provided 2.5 g/d arginine. To maintain blinding, in the LP and MP interventions, subjects were supplied with placebo capsules containing mannitol. Subjects had to take twice daily two capsules that each contained 400 mg arginine or mannitol.

3

2.4. Data collection Dietary intake was assessed before and during the last 3 d of each intervention period using a 3-d Food Frequency Protocol (FFP) originating from PRODIÒ 5.9 (Nutri-Science GmbH, Freiburg, Germany). Primary (total, LDL, HDL as well as oxidized LDL cholesterol, and triacylglycerols) and secondary outcome measures (hs-CRP, urea, uric acid, homocysteine, serum amino acids, body composition, and blood pressure) were determined at d0 and d28 of each intervention period. Blood samples were collected by venipuncture following a 12-h overnight fast into serum gel tubes and plasma gel tubes containing either EDTA or lithium heparin (Sarstedt AG & Co., Nümbrecht, Germany). Serum gel tubes and plasma gel tubes with EDTA were centrifuged at 20  C, 2500  g for 10 min. Plasma gel tubes with lithium heparin were centrifuged at 15  C, 4302 g for 7 min. The weight of fasting participants was determined in light clothes and without shoes using a digital scale. Waist circumference was measured at the navel level of the shirtless participants. Body composition was determined in a lying position using bioelectrical impedance analysis (BIA 2000-S, Data Input GmbH, Darmstadt, Germany). Blood pressure was measured in a sitting position in duplicate after 10 min of rest on the left arm using an automatic blood pressure monitor (boso-medicus uno, Bosch þ Sohn GmbH u. Co. KG, Jungingen, Germany). At the end of each intervention period, possible side effects of the interventions were recorded. Palatability of the study products was rated by an evaluation scale from best (1.0) to worst (6.0). 2.5. Analytical methods Plasma was analyzed for total, LDL, and HDL cholesterol as well as for triacylglycerols, hs-CRP, urea, and uric acid according to the protocols of the Institute of Clinical Chemistry and Laboratory Medicine, Jena University Hospital, and quantified using the autoanalyzer ARCHITECT C16000 (Abbott, Illinois, USA). Homocysteine was determined in plasma by HPLC (Shimadzu, Kyoto, Japan) with fluorescence detection after the reductive release from protein and disulfide bonds, thus enabling the analysis of total homocysteine according to the protocols of the Institute of Clinical Chemistry and Laboratory Medicine, Jena University Hospital. Oxidized LDL cholesterol was analyzed in serum by a competitive ELISA (Mercodia Oxidized LDL ELISA, Mercodia AB, Uppsala, Sweden). For the analysis of free amino acids in serum, the method based on the European Community Directive11 was applied as described previously.12 2.6. Statistical analyses Statistical analyses were conducted using PASS 6.0 (NCSS Statistical Software, Kaysville, UT, USA) and SPSS 19.0 (SPSS Inc., Chicago, USA). In all statistical analyses, P was considered significant

Table 1 Dietary nutrients of the study products. Nutrient (per 100 g fresh matter)

Energya Protein Fat Carbohydratesb Dietary fiber a b

Lupin protein isolate

MJ g g g g

Milk protein isolate

Bread

Roll

Scalded sausage

Vegetarian spread

Bread

Roll

Scalded sausage

Vegetarian spread

1.0 13.0 1.7 43.5 12.0

1.4 14.1 4.8 55.9 6.2

1.0 16.6 17.1 0.4 1.7

0.6 13.3 8.5 1.5 3.1

1.0 11.6 1.7 43.5 9.3

1.4 14.3 4.2 55.9 6.4

0.9 15.9 15.7 0.4 2.6

0.5 11.1 8.0 1.5 2.9

Energy was calculated on the basis of macronutrients. Carbohydrates were calculated on the basis of comparable food products in PRODIÒ 5.9 (Nutri-Science GmbH, Freiburg, Germany).

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

4

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

when 0.05. A power analysis revealed >80% power for the present study to detect a 10% difference in the primary outcome measure LDL cholesterol. All collected data were tested for normal distribution and for homogeneity of variances by applying the KolmogoroveSmirnov test and the Levene’s test, respectively. Data that were not normally distributed and/or had heterogeneous variances were transformed to their natural logarithms; otherwise, nonparametric tests were used for statistical analyses. Initial characteristics among the three randomization groups as well as differences between the relative changes following the protein interventions dependent on the severity of the cholesterol- and triacylglycerolemia were tested with t-test for independent samples or ManneWhitney U test. To compare the values after d28 with that of d0 within one protein intervention, a repeated measures ANOVA (general linear model) was used. To compare the values at d28 among the three intervention periods, repeated measures ANCOVA (general linear model) with the initial values as covariate was applied. In both ANOVA and ANCOVA, the within-subject factors ‘randomization group’ and ‘gender’ were used. For data that were not normally distributed and/or had heterogeneous variances after transformation to their natural logarithms, linear mixed model analyses with comparable conditions to those of ANOVA and ANCOVA were applied. Correlation coefficients were determined from pairwise correlation between the absolute changes (value at d28 e value at d0) of different parameters or between the absolute changes and the initial values of one parameter using Pearson’s or Spearman’s (S) correlation for normally distributed or skewed parameters, respectively. 3. Results

evenly distributed across the randomization groups, except for a few parameters in randomization group B, mainly due to a higher proportion of females in this group (Supplemental Table 2). Considering the whole study population, the initial lipid concentrations were 6.31  1.03 mmol/L for total cholesterol, 4.06  0.94 mmol/L for LDL cholesterol, 1.40  0.37 mmol/L for HDL cholesterol, and 1.78  1.52 mmol/L for triacylglycerols. In all three interventions, study products (roll, bread, sausage, and spread) were rated similarly (range: 1.7e2.6, average: 2.2) except for the bread with lupin protein in the LP intervention period, which received better evaluation compared with the bread with milk protein in the MP and MPA interventions (P  0.007, data not shown). No adverse effects of the protein interventions were observed. However, a few subjects reported flatulence with comparable frequency in the LP (19%), MP (24%), and MPA interventions (18%). 3.2. Dietary intake Dietary intakes at d0 and d28 as well as differences between the protein interventions are shown in Table 2. As expected, intake of arginine was significantly higher during the LP and MPA interventions compared with the MP intervention. Comparing dietary intakes during and prior to the protein interventions, a significant increase in energy, protein, carbohydrate, dietary fiber, and cystine intake within all three interventions was observed (P  0.0003). Fat intake increased during all three interventions, but the increase was significant only for the LP and MP interventions (P  0.023). Arginine intake significantly increased during the LP and MPA interventions (P < 0.0001).

3.1. Initial characteristics and evaluation of the study products 3.3. Anthropometric data and blood pressure Sixty-eight (94%) of the 72 participants who started the study completed all three intervention periods (Fig. 1). Two persons withdraw due to medical reasons independent of the study conditions and two had to be excluded from analysis because of a lack of sample material. Initial characteristics of the subjects were

There were no significant differences among the protein interventions in anthropometric data and blood pressure, except for a significantly higher waist circumference after the MP compared with the LP intervention (Table 3). This was due to a slight,

Table 2 Dietary intakes before and during the protein interventions. LP intervention Energyb (MJ/d) Protein (g/d) Fat (g/d) Carbohydrates (g/d) Dietary fiber (g/d) Arginine (mg/d) Cystine (mg/d) Lysine (mg/d) Methionine (mg/d)

d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28

10.0 11.2 90.7 109.3 96.2 104.2 269.2 318.5 26.1 37.4 5018 7441 1242 1663 5960 6465 2013 1893

                 

2.6 2.5*** 27. 9 28.6*** 32.5 28.1* 72.7 73.1*** 10.3 8.9*** 1804 1818*** 382 374*** 2077 2186 668 692

MP intervention 10.0 11.3 95.9 111.3 94.1 105.0 265.9 317.5 28.0 36.6 5375 5673 1293 1572 6415 7475 2147 2448

                 

2.8 3.1*** 29.7 34.3*** 34.0 37.6* 75.1 83.8*** 10.9 9.3*** 1897 2156 415 446*** 2138 2499*** 704 785**

MPA intervention 10.0 11.2 94.8 109.0 97.7 103.8 260.1 314.9 26.0 36.5 5383 7082 1277 1554 6381 7364 2126 2427

                 

2.9 2.4*** 31.4 28.9*** 35.9 32.7 83.0 60.3*** 9.4 7.9*** 1954 1813*** 395 376*** 2512 2304*** 800 742***

LP vs. MP Pa

LP vs. MPA Pa

MP vs. MPA Pa

0.82

0.98

0.79

0.54

0.96

0.59

0.85

0.92

0.78

0.76

0.73

0.95

0.39

0.27

0.98

<0.0001

0.18

<0.0001

0.08

0.041

0.79

0.001

0.002

0.68

<0.0001

<0.0001

0.86

Values are presented as mean  standard deviation, n ¼ 68. LP intervention, intervention with lupin protein; MP intervention, intervention with milk protein; MPA intervention, intervention with milk protein plus arginine. *,**,*** Significant differences comparing d28 with d0 within one protein intervention determined by repeated measures ANOVA with gender and sequence as within-subject factors (*P  0.05, **P  0.01, ***P  0.001). a P-value is for differences between the values of the LP, MP, and MPA interventions at d28 determined by repeated measures ANCOVA with gender and sequence as withinsubject factors and the initial values of the respective parameter serving as covariate. b Parameter was normally distributed after transforming it to its natural logarithm.

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

5

Table 3 Anthropometric data, blood pressure, and concentrations of blood lipids, hs-CRP, urea, uric acid, and homocysteine at d0 and d28 of the protein interventions. LP intervention Body weight (kg) BMI (kg/m2) Body fat (kg) Waist circumference (cm) Systolic BP (mm Hg) Diastolic BP (mm Hg) Total cholesterol (mmol/L) LDL cholesterol (mmol/L) HDL cholesterolc (mmol/L) LDL:HDL cholesterol ratioc Oxidized LDL cholesterol (U/L) Triacylglycerols (mmol/L) hs-CRPb,c (mg/L) Urea (mmol/L) Uric acid (mmol/L) Homocysteine (mmol/L)

d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28 d0 d28

76.8 76.9 26.4 26.5 21.4 21.5 91.6 91.3 139.6 142.2 86.7 87.0 6.41 6.13 4.16 4.01 1.40 1.35 3.21 3.19 58.9 58.3 1.85 1.69 2.30 2.15 5.33 5.83 335.0 321.8 10.5 10.1

                               

16.7 16.5 4.9 5.0 9.6 9.6 15.0 14.5 20.8 20.8 9.3 9.9 1.07 0.95*** 0.98 0.87** 0.39 0.37*** 1.20 1.17 21.4 17.3 1.44 1.29* 3.62 3.22 1.22 1.14*** 87.5 87.4** 2.8 2.5

MP intervention 76.7 76.9 26.4 26.5 21.3 21.5 90.9 92.1 141.5 140.3 86.1 86.8 6.28 6.23 4.06 4.08 1.38 1.36 3.15 3.20 58.1 60.0 1.82 1.77 2.22 2.51 5.46 5.84 328.3 321.1 10.2 10.8

                               

16.4 16.5 5.0 4.9 9.6 9.5 14.0 14.3*** 20.7 19.2 9.7 9.8 0.96 0.97 0.96 0.95 0.36 0.35 1.12 1.16 20.6 20.6 1.59 1.42 3.01 3.88 1.23 1.22** 86.7 90.0 2.6 2.6*

MPA intervention 76.5 76.6 26.3 26.4 21.2 21.2 91.7 91.6 140.5 140.4 85.9 86.0 6.39 6.06 4.15 3.99 1.43 1.33 3.11 3.22 56.1 59.9 1.74 1.64 2.46 2.29 5.38 6.31 333.6 320.4 10.5 10.6

                               

16.7 16.5 4.9 4.9 9.3 9.3 14.4 14.5 20.3 18.6 9.4 9.5 1.08 0.95*** 1.05 0.92** 0.38 0.34*** 1.12 1.16 18.4 16.7* 1.21 1.06 3.99 3.24 1.20 1.34*** 86.3 84.7 2.7 2.5

LP vs. MP Pa

LP vs. MPA Pa

MP vs. MPA Pa

0.85

0.13

0.12

0.87

0.06

0.07

0.88

0.07

0.08

0.002

0.31

0.09

0.35

0.45

0.82

0.84

0.25

0.34

0.07

0.43

0.018

0.044

0.92

0.10

0.37

0.16

0.07

0.40

0.38

0.88

0.53

0.18

0.65

0.49

0.83

0.62

0.15

0.45

0.42

0.65

0.0004

0.001

0.87

0.61

0.79

0.001

0.004

0.42

Values are presented as mean  standard deviation, n ¼ 68. LP intervention, intervention with lupin protein; MP intervention, intervention with milk protein; MPA intervention, intervention with milk protein plus arginine; BMI, body mass index; BP, blood pressure; hs-CRP, high-sensitivity C-reactive protein. *,**,*** Significant differences comparing d28 with d0 within one protein intervention determined by repeated measures ANOVA with gender and sequence as within-subject factors (*P  0.05, **P  0.01, ***P  0.001). a P-value is for differences between the values of the LP, MP, and MPA interventions at d28 determined by repeated measures ANCOVA with gender and sequence as withinsubject factors and the initial values of the respective parameter serving as covariate. b n ¼ 52, hs-CRP values of some participants had to be excluded due to a temporary inflammatory status at time of measurement. c Parameter was normally distributed after transforming it to its natural logarithm.

3.4. Blood parameters Compared with the MP intervention, total cholesterol was significantly lower after the MPA intervention and tended to be lower after the LP intervention (Table 3). However, LDL cholesterol was significantly lower after the LP intervention and tended to be lower after the MPA intervention compared with the MP intervention. Compared with d0, concentrations of total (P < 0.0001), LDL (P  0.009), and HDL cholesterol (P  0.001) significantly decreased after the LP and MPA interventions, whereas triacylglycerols only decreased after the LP intervention (P ¼ 0.034). Oxidized LDL cholesterol significantly increased after the MPA intervention (P ¼ 0.049). Based on the initial total cholesterol concentration, the study population was split into two groups: subjects with moderate hypercholesterolemia (5.2e6.6 mmol/L) and subjects with severe hypercholesterolemia (>6.6 mmol/L). Following the LP intervention, the relative changes in total and LDL cholesterol from d0 to d28 were significantly greater for the group with severe hypercholesterolemia compared with the group with moderate hypercholesterolemia (Fig. 2). Such significant differences in relative changes of plasma lipids were not observed for the MP and MPA interventions (Supplemental Figs. 1 and 2).

Concentrations of uric acid and hs-CRP did not differ among the protein interventions, whereas urea was significantly higher after the MPA intervention compared with the LP and MP interventions. Compared with the MP and MPA interventions, homocysteine was significantly lower after the LP intervention (Table 3). Compared

2

Total cholesterol

LDL cholesterol

HDL cholesterol

Triacylglycerols

0

Revitale change from d0 (%)

significant increase in the waist circumference after the MP intervention compared with d0 (P ¼ 0.0005).

-2 -4 -6 -8

P = 0.038

P = 0.045

Moderate Hypercholesterolemia Severe Hypercholesterolemia

-10

Fig. 2. Relative changes (%) in plasma lipid concentrations from d0 to d28 after the LP intervention in subjects with moderate (6.6 mmol/L, n ¼ 37) or severe hypercholesterolemia (>6.6 mmol/L, n ¼ 31). LP intervention, intervention with lupin protein. a P-value is for differences between the relative changes determined by t-test for independent samples.

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

1.00 0.45*** 1.00 0.69*** 0.08 1.00 0.75*** 0.68*** 0.17 Values are the Pearson’s correlation coefficients. LP intervention, intervention with lupin protein; hs-CRP, high-sensitivity C-reactive protein. *,**,*** Changes (value at d28 e value at d0) after the LP intervention significantly correlated (*P  0.05, **P  0.01, ***P  0.001). t Changes tended to significantly correlate (0.05 > P < 0.1). a n ¼ 52, hs-CRP values of some participants had to be excluded due to a temporary inflammatory status at time of measurement.

1.00 0.16 0.06 0.15 0.06 1.00 0.01 0.43*** 0.37** 0.29* 0.05 1.00 0.39*** 0.11 0.21t 0.13 0.10 0.08 1.00 0.23t 0.73*** 0.05 0.27* 0.30* 0.22t 0.04 1.00 0.01 0.31** 0.29* 0.17 0.14 0.05 0.00 0.07 1.00 0.36** 0.86*** 0.24t 0.64*** 0.05 0.20 0.16 0.16 0.02 1.00 0.21** 0.25* 0.08 0.01 0.09 0.36** 0.07 0.08 0.15 0.22** 1.00 0.21 0.15 0.02 0.15 0.07 0.11 0.06 0.09 0.00 0.13 0.20 hs-CRPa Uric acid Total cholesterol Triacylglycerols LDL cholesterol HDL cholesterol LDL:HDL cholesterol ratio Homocysteine Cystine Methionine Arginine Lysine:arginine ratio

HDL cholesterol LDL cholesterol Triacylglycerols Total cholesterol Uric acid

In the present study, consumption of 25 g/d lupin protein, which was incorporated into a variety of complex food products, reduced total (4%) and LDL cholesterol (4%) as well as triacylglycerols (9%) in subjects with hypercholesterolemia after 4 weeks. These results are in line with previous studies in hypercholesterolemic subjects, which found significant decreases in total1e3 and LDL cholesterol2,4 after lupin protein supplementation, with the amount ranging between 25 and 36 g/d. In the studies by Bähr et al.4 and Weiße et al.,2 triacylglycerols did not decrease after consumption of 25 and 35 g/d lupin protein. However, Sirtori et al.3 observed a decrease in triacylglycerols (13%) after intervention with 35 g/d lupin protein, although this reduction was not significant. In a previous study of our work group, the LDL:HDL cholesterol ratio decreased by 9% due to the addition of 25 g/d lupin protein.4 By contrast, in the present study, we did not observe a reduction in the LDL:HDL cholesterol ratio. This was mainly due to a slight but significant decrease in HDL cholesterol (3%) after the LP intervention. Although not significant, similar decreases in HDL cholesterol were observed by Sirtori et al.3 (2%) and Weiße et al.2 (5%) after supplementation with 35 g/d lupin protein. Moreover, in the present study, neither diastolic nor systolic blood pressure was influenced by the protein interventions. Initially, we hypothesized that lupin protein would decrease systolic blood pressure by about 5e6%, as observed in two intervention studies.1,4 However, in these studies, lupin protein was provided additionally to the normal diet, whereas in present study, lupin protein replaced other dietary proteins. This might be an explanation for the lack of changes in blood pressure. Considering the subjects’ initial total cholesterol concentration, in those with higher initial cholesterol (>6.6 mmol/L), compared with those with moderate hypercholesterolemia (5.2e6.6 mmol/L, Fig. 2), the LP intervention was associated with more pronounced effects on total (6% vs. 2%, respectively) and LDL cholesterol (6% vs. 0%, respectively). This observation is supported by a significant correlation of the initial total cholesterol with the absolute changes in total (r ¼ 0.35, P ¼ 0.004, n ¼ 68) as well as in LDL cholesterol (r ¼ 0.28, P ¼ 0.021, n ¼ 68; Fig. 3) after the LP intervention. Likewise, in a previous study of our work group, the beneficial impact of supplementation with 25 g/d lupin protein on total (5%) and LDL cholesterol (8%) after 4 weeks was restricted to subjects with total cholesterol >6.6 mmol/L.4 As reported by Sirtori et al.13 and Anderson et al.,14 similar relationships between

Table 4 Correlations between the absolute changes from d0 to d28 of several plasma and serum parameters within the LP intervention (n ¼ 68).

4. Discussion

LDL:HDL cholesterol ratio

Homocysteine

Cystine

Methionine

Arginine

Lysine:arginine ratio

with d0, concentrations of urea significantly increased after all three protein interventions (P  0.007). Uric acid significantly decreased after the LP intervention (P ¼ 0.003). Homocysteine tended to decrease after the LP intervention (P ¼ 0.068) and significantly increased after the MP intervention (P ¼ 0.017). Serum amino acids at d0 and d28 as well as differences between the protein interventions are shown in Supplemental Table 3. Concentrations of methionine and lysine as well as the sum of the essential amino acids were significantly lower after the LP intervention compared with the MP and MPA interventions. Cystine was significantly lower after the LP intervention compared with the MPA intervention. Compared with d0, concentrations of all single amino acids significantly decreased after the LP intervention (P  0.010), except for cystine, tyrosine and arginine. After the MP intervention, total amino acids significantly increased (P ¼ 0.021). The ratio of lysine:arginine significantly decreased after the LP intervention (P ¼ 0.037), increased after the MP intervention (P ¼ 0.027), and remained unchanged after the MPA intervention. Correlations between changes in several blood parameters after the LP intervention are shown in Table 4.

1.00

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

hs-CRP

6

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

Fig. 3. Correlations between the initial total cholesterol concentrations and the absolute changes in total (eeee, r ¼ 0.35, P ¼ 0.004, n ¼ 68) and LDL cholesterol (e e e, r ¼ 0.28, P ¼ 0.021, n ¼ 68) from d0 to d28 after the LP intervention. LP intervention, intervention with lupin protein.

efficacy of the intervention and the initial cholesterol concentrations could be observed in studies with soy protein. Likewise, in the present study, the decrease in triacylglycerols after the LP intervention was restricted to subjects with initial triacylglycerols 1.7 mmol/L in comparison with those with triacylglycerols <1.7 mmol/L, which was indicated by significant differences in the relative triacylglycerol changes (14% vs. 7%, respectively, P ¼ 0.004). Accordingly, there was a significant correlation between the absolute changes in triacylglycerols after the LP intervention and the initial triacylglycerol concentrations (rS ¼ 0.44, P ¼ 0.0002, n ¼ 68). This is in line with the findings of Nowicka et al.,1 who observed a strong decrease in triacylglycerols (25%) for subjects with triacylglycerol concentrations >1.7 mmol/L after consumption of 36 g/d lupin protein for 3 months. Another objective of the present study was to evaluate the impact of arginine on several metabolic markers, since this amino acid is abundant in lupin protein. Similarly to the LP intervention, the MPA intervention led to a decrease in total (5%) and LDL cholesterol (4%). These two protein interventions provided 2.5 g arginine daily. In contrast, the MP intervention did not significantly change blood lipids compared with d0. These observations strengthen the hypothesis that the high proportion of arginine in lupin protein might be responsible for its lipid-lowering effects. This is also supported by Jobgen et al.,15 who reported a probable hypocholesterolemic activity of nitric oxide donors such as arginine, mediated by modulation of lipoprotein metabolism. Therefore, it might be assumed that in the present study, the higher arginine intake during the LP and MPA interventions could have been accompanied by increased serum arginine, thereby exerting the hypocholesterolemic effects. However, serum arginine did not increase after either the LP or MPA intervention. Correspondingly, the absolute changes of serum arginine did not correlate with those of total or LDL cholesterol, with neither the LP intervention (Table 4) nor the MPA intervention (total cholesterol: r ¼ 0.14, P ¼ 0.27; LDL cholesterol: r ¼ 0.15 P ¼ 0.23, n ¼ 68, data not shown). Thus, the lipid-lowering effects of the LP and MPA interventions might have been mediated by mechanisms independent of an increase in serum arginine. In a study by Evans et al.,16 arginine supplementation of 9 g/d raised circulating arginine concentrations, whereas 3 g/d did not. Hence, since in the present study, the initial concentrations of serum arginine were high and the amount of arginine administered through the LP and MPA interventions

7

was only 2.5 g/d, a further increase in serum arginine was unlikely from the outset. However, the serum lysine:arginine ratio significantly decreased after the LP intervention (4%), whereas it increased after the MP intervention (5%, Supplemental Table 3). Comparable changes were observed by Bähr et al.4 after intervention with lupin protein and milk protein and can be explained by the higher proportion of arginine and lower proportion of lysine in the lupin protein compared with the milk protein isolate. After the MPA intervention, this ratio remained constant, since here the milk protein products were supplemented with 1.6 g/d arginine. Almost all the other serum amino acids increased after the MP intervention compared with d0, whereas they mostly remained unchanged after the MPA intervention and largely decreased after the LP intervention. As with serum arginine, the initial concentrations of these amino acids were relatively high. Thus, serum concentrations of the amino acids at d28 of the LP intervention can still be considered as physiologically safe according to reference values of plasma amino acids of the A.D.A.M. Medical Encyclopedia.17 A novel finding in the present study was that plasma homocysteine was lower after the LP intervention compared with the MP and MPA interventions and tended to decrease after the LP intervention compared with d0 (4%, P ¼ 0.068). Furthermore, uric acid significantly decreased after the LP intervention (13%, P ¼ 0.003). Within the LP intervention, we observed a significant correlation between the absolute changes in both parameters from d0 to d28 (r ¼ 0.37, P ¼ 0.002, n ¼ 68, Table 4), as well as between their initial values (r ¼ 0.27, P ¼ 0.029, n ¼ 68). These observations are consistent with the findings of Annanmaki et al.,18 who found plasma uric acid concentrations to be associated with those of homocysteine (r ¼ 0.40, P ¼ 0.01) in patients with Parkinson’s disease. Both homocysteine19,20 and uric acid21 have been associated with cardiovascular diseases. Thus, individuals at high cardiovascular risk would especially benefit from a reduction in these parameters by lupin protein consumption. In the present study, the higher absolute protein intake during the interventions compared to that prior to the interventions was accompanied by significantly increased plasma urea following the LP, MP, and MPA interventions (9%, 7%, and 17%, respectively). Despite the comparable amount of arginine of 2.5 g/d provided by the LP and MPA interventions, plasma urea was significantly higher after the MPA intervention compared with the LP intervention. The arginine administration in the form of capsules during the MPA intervention possibly led to a higher bioavailability of arginine in the small intestine and thus, might have resulted in a higher degradation of arginine to urea, since the major part of dietary arginine is utilized in the hepatic urea cycle.22 A clear limitation of the present study was the significantly higher fiber intake during the protein interventions. As reported by Kendall et al.,23 an increased consumption of viscous fiber can lower LDL cholesterol. However, whereas the intake of dietary fiber increased within all three interventions, the total and LDL cholesterol-lowering effects were only observed for the LP and MP interventions. It has to be emphasized that the present study was the first that realized consumption of 25 g/d lupin protein by incorporating it into a variety of complex food products. In contrast, previous human studies on lupin protein supplementation administered the protein in the form of simple food products such as protein drinks or bars in addition to the normal diet.1e4 Thus, in our study, the mode of application and the thermal load of the proteins during scalding of sausages and spread as well as baking of bread and rolls might have resulted in a different bioavailability and efficacy of the protein’s bioactive components due to interaction with the respective food matrix and protein degradation, respectively. Furthermore, the study products replaced common food products within the mixed diet of the participants. Thus, the test

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008

8

M. Bähr et al. / Clinical Nutrition xxx (2014) 1e8

proteins were not consumed additionally, but as a substitute for other dietary proteins. These conditions could have led to attenuated effects of lupin protein on blood lipids and to the lack of changes in blood pressure.

Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.clnu.2014.03.008.

5. Conclusion References This study showed for the first time that incorporation of modest amounts of lupin protein into variety of complex food products is capable of beneficially altering several cardiovascular risk factors such as total and LDL cholesterol, triacylglycerols, homocysteine, and uric acid. The observation of more pronounced effects in subjects with higher initial total cholesterol and triacylglycerol concentrations indicates that subjects with more severe hyperlipidemia might particularly benefit from a diet enriched with lupin protein, either alone or in combination with a pharmacological treatment. However, the present study cannot confirm a positive impact of lupin protein on the ratio of LDL:HDL cholesterol, on HDL or oxidized LDL cholesterol, or on blood pressure. The study contributes to our understanding of the role of the high proportion of arginine in lupin protein with respect to a lipidlowering action. We have shown that consumption of milk protein products with additional arginine positively affected blood lipids, as did lupin protein consumption. Altogether, lupin protein can be considered a valuable source of protein with desirable technofunctional, sensory, and physiological properties. Thus, lupin protein is suitable for use in daily human nutrition, particularly with regard to the ecological advantage that is associated with proteins from plant sources. Statement of authorship MB, AF, and GJ designed the research; MB was responsible for supervising the study, sample handling, coordination, and conduction of the analyses; MB and MK analyzed data; MB performed statistical analyses; MB and GJ were responsible for data interpretation and had primary responsibility for final content; MB wrote the paper. All authors read and approved the final manuscript. Source of funding The study was supported by the Federal Ministry of Education and Research (grant no. 01EA1338C). Study sponsors were not involved in the study design, or collection, analysis, and interpretation of data, in writing the manuscript or in the decision to submit the manuscript for publication. Conflict of interest statement None of the authors had any personal or financial conflicts of interest. Acknowledgments We thank the Federal Ministry of Education and Research for financial support and the Fraunhofer Institute for Process Engineering and Packaging (Fh-IVV) for supplying the protein isolates. Further, we thank the Ludwig Stocker Hofpfisterei GmbH, the Backstube Wünsche GmbH, and the Metzgerei Boneberger GmbH for producing the study products. We express thanks to U. Helms, A. Lochner, and K. Weiske for technical assistance.

1. Nowicka G, Klosiewicz-Latoszek L, Sirtori CR, Arnoldi A, Naruszewicz M. Lupin proteins in the treatment of hypercholesterolemia. Atheroscler Suppl 2006;7(3):477. 2. Weisse K, Brandsch C, Zernsdorf B, Nembongwe GSN, Hofmann K, Eder K, et al. Lupin protein compared to casein lowers the LDL cholesterol:HDL cholesterolratio of hypercholesterolemic adults. Eur J Nutr 2010;49(2):65e71. 3. Sirtori CR, Triolo M, Bosisio R, Bondioli A, Calabresi L, De Vergori V, et al. Hypocholesterolaemic effects of lupin protein and pea protein/fibre combinations in moderately hypercholesterolaemic individuals. Br J Nutr 2012;107(8): 1176e83. 4. Bähr M, Fechner A, Krämer J, Kiehntopf M, Jahreis G. Lupin protein positively affects plasma LDL cholesterol and LDL: HDL cholesterol ratio in hypercholesterolemic adults after four weeks of supplementation: a randomized, controlled crossover study. Nutr J 2013;12(1):107. 5. Duranti M, Morazzoni P. Nutraceutical properties of lupin seed proteins: a great potential still waiting for full exploitation. Agro Food Ind Hi Tec 2011;22(1):20e3. 6. Ast J, Jablecka A, Bogdanski P, Smolarek I, Krauss H, Chmara E. Evaluation of the antihypertensive effect of L-arginine supplementation in patients with mild hypertension assessed with ambulatory blood pressure monitoring. Med Sci Monit 2010;16(5). CR266-CR71. 7. Tripathi P, Chandra M, Misra MK. Oral administration of L-arginine in patients with angina or following myocardial infarction may be protective by increasing plasma superoxide dismutase and total thiols with reduction in serum cholesterol and xanthine oxidase. Oxid Med Cell Longev 2009;2(4): 231e7. 8. Hurson M, Regan MC, Kirk SJ, Wasserkrug HL, Barbul A. Metabolic effects of arginine in a healthy elderly population. J Parenter Enter Nutr 1995;19(3): 227e30. 9. D’Agostina A, Antonioni C, Resta D, Arnoldi A, Bez J, Knauf U, et al. Optimization of a pilot-scale process for producing lupin protein isolates with valuable technological properties and minimum thermal damage. J Agric Food Chem 2006;54(1):92e8. 10. Association of Official Analytical Chemists. Official methods of analysis of the association of official analytical chemists. 15th ed. Virgina, USA: AOAC International; 1995. 11. European Community. COMMISSION DIRECTIVE 98/64/EC of 3 September 1998 establishing Community methods of analysis for the determination of aminoacids, crude oils and fats, and olaquindox in feeding stuffs and amending Directive 71/393/EEC. Off J Eur Commun; 1998:14e28. Brussels. 12. Fleddermann M, Fechner A, Rössler A, Bähr M, Pastor A, Liebert F, et al. Nutritional evaluation of rapeseed protein compared to soy protein for quality, plasma amino acids, and nitrogen balance e a randomized cross-over intervention study in humans. Clin Nutr 2013;32(4):519e26. 13. Sirtori CR, Eberini I, Arnoldi A. Hypocholesterolaemic effects of soya proteins: results of recent studies are predictable from the Anderson meta-analysis data. Br J Nutr 2007;97(5):816e22. 14. Anderson JW, Johnstone BM, Cook-Newell ME. Meta-analysis of the effects of soy protein intake on serum lipids. N Engl J Med 1995;333(5):276e82. 15. Jobgen WS, Fried SK, Fu WJ, Meininger CJ, Wu GY. Regulatory role for the arginine-nitric oxide pathway in metabolism of energy substrates. J Nutr Biochem 2006;17(9):571e88. 16. Evans RW, Fernstrom JD, Thompson J, Morris Jr SM, Kuller LH. Biochemical responses of healthy subjects during dietary supplementation with L-arginine. J Nutr Biochem 2004;15(9):534e9. 17. Plasma amino acids: A.D.A.M. medical encyclopedia [Updated December 5, 2011; accessed 06.11.13]. Available from: http://www.ncbi.nlm.nih.gov/ pubmedhealth/PMH0003841/; 2011. 18. Annanmaki T, Pessala-Driver A, Hokkanen L, Murros K. Uric acid associates with cognition in Parkinson’s disease. Parkinsonism Relat Disord 2008;14(7): 576e8. 19. Mangoni AA, Jackson SHD. Homocysteine and cardiovascular disease: current evidence and future prospects. Am J Med 2002;112(7):556e65. 20. Fruchart JC, Nierman MC, Stroes ESG, Kastelein JJP, Duriez P. New risk factors for atherosclerosis and patient risk assessment. Circulation 2004;109(23):15e9. 21. Kawai T, Ohishi M, Takeya Y, Onishi M, Ito N, Yamamoto K, et al. Serum uric acid is an independent risk factor for cardiovascular disease and mortality in hypertensive patients. Hypertens Res 2012;35(11):1087e92. 22. Böger RH. L-arginine therapy in cardiovascular pathologies: beneficial or dangerous? Curr Opin Clin Nutr Metab Care 2008;11(1):55e61. 23. Kendall CWC, Esfahani A, Jenkins DJA. The link between dietary fibre and human health. Food Hydrocoll 2010;24(1):42e8.

Please cite this article in press as: Bähr M, et al., Consuming a mixed diet enriched with lupin protein beneficially affects plasma lipids in hypercholesterolemic subjects: A randomized controlled trial, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.03.008