The relationship between dietary protein intake and blood ... - Nature

Journal of Human Hypertension (2008) 22, 745–754 & 2008 Macmillan Publishers Limited All rights reserved 0950-9240/08 $32.00 www.nature.com/jhh

ORIGINAL ARTICLE

The relationship between dietary protein intake and blood pressure: results from the PREMIER study YF Wang1, WS Yancy Jr2,3, D Yu4, C Champagne5, LJ Appel6 and P-H Lin2 1

Health and Productivity Management Program, Society of Health Risk Assessment and Control, Chinese Association of Preventive Medicine, 1202 Fortune International Center, Haidian District, Beijing, China; 2 Department of Medicine, Duke University Medical Center, Durham, NC, USA; 3Center for Health Services Research in Primary Care, Department of Veterans’ Affairs Medical Center, Durham, NC, USA; 4Department of Biostatistics and Bioinformatics, Duke University Medical Center, Durham, NC, USA; 5Dietary Assessment and Food Analysis Core, Pennington Biomedical Research Center, Baton Rouge, LA, USA and 6Department of Medicine, Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins University, Baltimore, MD, USA

Observational and clinical studies suggest that high protein intake, particularly protein from plant sources, might reduce blood pressure (BP). To examine the association of dietary protein with BP, we analysed data from PREMIER, an 18-month clinical trial (n ¼ 810) that examined the effects of two multi-component lifestyle modifications on BP. We examined the association of protein intake with BP, and in particular the independent relationship of plant and animal protein with BP. Multivariable linear regression analyses were performed with both cross-sectional and longitudinal data. Dietary plant protein was inversely associated with both systolic and diastolic BP in cross-sectional analyses at the 6-month follow-up (P ¼ 0.0045 and 0.0096, respectively). Fruit and vegetable intake was also inversely associated with both systolic and diastolic BP cross-sectionally at 6 months (P ¼ 0.0003 and 0.0157, respectively). In

longitudinal analyses, a high intake of plant protein at 6 months was marginally associated with a reduction of both systolic and diastolic BP from baseline to 6 months only (P ¼ 0.0797 and 0.0866, respectively), independent of change in body weight and waist circumference. Furthermore, increased intake of plant protein, and fruits and vegetables was significantly associated with a lower risk of hypertension at 6 but not at 18 months. Results of this study indicate that plant protein had a beneficial effect on BP and was associated with a lower risk of hypertension at 6 months. Our data, in conjunction with other research, suggest that an increased intake of plant protein may be useful as a means to prevent and treat hypertension. Journal of Human Hypertension (2008) 22, 745–754; doi:10.1038/jhh.2008.64; published online 26 June 2008

Keywords: plant protein; animal protein; blood pressure; DASH diet

Introduction Many earlier studies, mainly observational, have shown an inverse relationship between total dietary protein intake and blood pressure (BP).1 This inverse association is also found in more recent studies.2–4 However, results have been mixed when animal and plant proteins were examined separately. Cross-sectional studies conducted in rural Japanese5,6 and Chinese7,8 populations only found an inverse relationship between animal protein and Correspondence: Dr YF Wang, Department of Medicine, Duke University Medical Center, Co. Pao-Hwa Lin, Box 3487, Durham, NC 27710, USA. E-mail: [email protected] Received 7 February 2008; revised 17 May 2008; accepted 22 May 2008; published online 26 June 2008

BP, based on ratios of urinary excretions of sulphate and other animal protein-related amino acids. On the contrary, cohort studies9,10 conducted in the United States have shown an inverse association between annual change in systolic BP and baseline intakes of plant protein but not animal protein, as assessed by diet history. Recent evidence from randomized controlled trials suggests that an increased intake of protein, from both plant and animal sources, may lower BP and thus potentially reduce the risk of cardiovascular disease.11–15 However, it is unclear whether plant and animal proteins have different effects on BP. PREMIER was a randomized clinical trial designed to compare the impact of two multi-component lifestyle modification programs on BP. The

The relationship between dietary protein intake and blood pressure YF Wang et al 746

3964 Individual Screened 3154 ineligible 194 BP too high 2103 BP too low 857 other exclusions 810 randomized

273 assigned to advice only group

268 assigned to established Intervention group

269 assigned to established + DASH intervention group

6 month assessment on blood pressure, dietary intakes and other information (n=712)

18 month assessment on blood pressure, dietary intakes and other information (n=740) Figure 1 Study design.

PREMIER study data provided an excellent opportunity to examine, in a clinical trial setting, the association of dietary protein intake with BP, and the independent relationship of animal and plant protein with BP in particular.

Methods Study design

The present study is a secondary analysis using data from the PREMIER trial. The details of the PREMIER study design, eligibility criteria of participants, intervention and primary outcomes have been described elsewhere.16,17 In brief, PREMIER was an 18-month, multi-centre, three-group randomized clinical trial designed to determine the effect of two multi-component lifestyle interventions on BP for individuals who met the Joint National Committee on Prevention, Detection, and Treatment of High Blood Pressure18 criteria for a 6–12 month trial of non-pharmacological therapy (that is, individuals who were not taking antihypertensive medications and had a systolic BP of 120–159 mm Hg and/or diastolic BP of 80–95 mm Hg). The diagnosis of hypertension was made when the average of two or more diastolic BP measurements on at least two separate visits was X90 mm Hg or when the average of multiple systolic BP readings on two or more separate visits was consistently X140 mm Hg, according to the Joint National Committee on Prevention, Detection, and Treatment of High Blood Pressure.18 This study was conducted at four clinical centres (Johns Hopkins University, Pennington Biomedical Journal of Human Hypertension

Research Center, Duke University Medical Center and Kaiser Permanente Center for Health Research) from 2001 to 2003 and included a coordinating centre at Kaiser Permanente Center for Health Research. The National Heart, Lung and Blood Institute Project office also participated. Figure 1 shows a summary of the flow and design of the study. Interventions. The two multi-component 18-month interventions were (1) a behavioral lifestyle intervention (established recommendation, EST) designed to help participants follow long established recommendations on BP control: losing weight if overweight, reducing sodium intake, increasing physical activity and limiting alcohol intake, and (2) a behavioral lifestyle intervention combining the established recommendations plus the DASH (the Dietary Approaches to Stop Hypertension) eating plan (EST þ DASH). The goals of the DASH eating plan included increasing intake of fruits, vegetables and low-fat dairy products, and reducing intakes of saturated and total fats. Both interventions provided a total of 14 group sessions and four individual sessions during the first 6 months of the study and then monthly group sessions for the last 12 months. The third randomized group, termed ‘advice only’, was the control group, in which individuals received a 30-min advice session immediately after randomization and again after the 6-month data collection. Measurements. All data used for the present study were collected during the trial including dietary intake, anthropometric measurements, energy expenditure, laboratory assays and medical


assessments at baseline, 6 and 18 months. Staff members involved in follow-up data collection were blinded to treatment assignment. Intervention staff members were blinded to all outcome data during the entire study. Per cent of participants lost to follow-up was similar across the three treatment groups, approximately 11, 13 and 12% for the control, EST and EST þ DASH groups, respectively. Blood pressure measurements were obtained by centrally trained and certified individuals who used a random zero sphygmomanometer following a standard protocol. At each visit, two BP measurements separated by at least 30 s were obtained. At each assessment time point, BP data were the mean of all available measurements: eight BP measurements across four visits at baseline and six BP measurements across three visits at 6 months and again at 18 months. Dietary intake of food and nutrients was obtained on a different day from biomarker collection, using an unannounced 24-h recall method. Two recalls, one on a weekday and the other on a weekend day, were obtained at baseline, 6 and 18 months by the Diet Assessment Center of Pennsylvania State University. Trained and certified staff conducted recalls over the phone and data were entered directly into a computer. Nutrient and food group intake was then respectively calculated using the Nutrition Data System for Research software (version NDS-R, 1998) developed by the Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN. Biomarkers of dietary intake including 24-h urinary excretion of sodium (reflecting salt intake), potassium (reflecting fruit and vegetable intake), phosphorus (reflecting dairy intake) and urea nitrogen (reflecting protein intake) were also assessed. Information on dietary fat intake was collected in two parts: fat as a macronutrient (‘as nutrient’) from all food intakes that was expressed in g day1 and ‘fat serving’ as a food group including butter and oil. A 7-day interviewer-administered physical activity recall was used to assess energy expenditure from physical activity.19 Alcohol intake was obtained from a questionnaire. Statistical analysis

Statistical analysis for this report was conducted by investigators at Duke University Medical Center. Only participants who had at least one diet recall at each time period were included in the analysis. In addition, missing was treated by the default design of the General Linear Model in SAS, that is, a participant’s data are excluded when it is missing one or more of variables listed in the model. For analyses in this report, the three treatment groups were combined to allow better estimation of and more efficient inferences on parameters of interest. Treatment group indicators, however, were always included in the regression model as independent

variables. Energy intake was also adjusted in all statistical models by calculating protein and fat intake as per cent of total calories (% kcal) and other nutrient intakes as g per 1000 kcal per day (or mg per 1000 kcal per day). Paired t-tests were first performed to examine the differences in protein intake and covariates between baseline and 6 months and between baseline and 18 months of follow-up. Multivariable linear regression was then performed both cross-sectionally at all the three time points and longitudinally from baseline to 6 and 18 months, respectively. Mean systolic and diastolic BP measurements were the outcome variables for cross-sectional analyses, and the changes of BP from baseline to 6 and 18 months of follow-up were the outcome variables for the longitudinal analyses. The independent variables included the mean values (for cross-sectional analyses) and the mean change values (for longitudinal analyses) of dietary intake of protein, including total plant and animal proteins. Three models were examined cross-sectionally at each time point. The first two models included nutrients that might be related to BP control (intake of total protein, total fat, dietary fibre, calcium and potassium in model 1 and then total protein was separated into plant and animal proteins in model 2). The third model included major food groups only. The following covariates were always included in each regression model unless indicated baseline characteristics including age, gender, race, treatment groups, study site, education, income and BP. Further, alcohol intake, physical activity, waist circumference, urinary creatinine and sodium at 6 or 18 months were included in all the 6- or 18-month models, respectively. Longitudinal analyses were also performed by examining the association between the changes of BP and changes of dietary protein intakes between baseline and 6 months. In addition, a separate model was analysed to assess the association between baseline intakes of nutrients, including protein, with 18-month BP. All longitudinal models were controlled for the baseline values of the outcome variable in the respective multivariable regression analyses. Logistic regression was also performed to examine the association of hypertension status (presence/ absence) with various nutrient and food intake variables at each time point separately. All analyses were conducted using SAS statistical software (version 9.0, SAS Institute, Cary, NC, USA). As this secondary analysis is of exploratory nature, no multiple comparison adjustments were made, and a P-value of less than 0.05 was considered statistically significant.

Results Table 1 shows the baseline characteristics for all study participants. A total of 810 individuals Journal of Human Hypertension


Table 1 Characteristics at baseline for all trial participants (N ¼ 810) Variable

Mean (s.d.) or number (%)a

Age (years)

50.0 (8.9)

Gender; no. (%) Male Female

310 (38.3) 500 (61.7)

Race or ethnicity; no. (%) African American Non-Hispanic white All others

279 (34.4) 520 (64.2) 11 (1.4)

Income; no. (%) o$30 000 $30 000–$59 000 X$60 000 Unknown (no answer)

84 256 441 29

Education; no. (%) High school or less College or some college Postgraduate school

74 (9.1) 476 (58.8) 260 (32.1)

Current cigarette smoking; no. (%) BMI (kg m2) Waist circumference (cm) Hypertensionb; no. (%)

48 33.1 107.6 301

(10.4) (31.6) (54.4) (3.6)

(13.4) (5.8) (15.3) (37.5)

Abbreviations: BMI, body mass index; DBP, diastolic BP; SBP, systolic BP. a Mean (s.d.) for a continuous variable and number (%) for a discrete variable. b Criteria of hypertension was SBPX140 or DBPX90 mm Hg.

participated in the study with 62% female participants and a mean age of 50.0 years, ranging from 25 to 79 years. Most of them were non-Hispanic white whereas 34% were African American. Mean body mass index was 33.1 kg m2 (range 21–48 kg m2), and mean waist circumference was 107.6 cm (range 69.9–151.3 cm). Mean systolic and diastolic BPs were 135 (range 116–161) and 85 mm Hg (range 74–98), respectively, at baseline. A majority of the participants had at least a college education (91%) and were non-smokers (87%). More detailed description of the characteristics of the participants have been published elsewhere.17 Table 2 presents descriptive statistics on the outcome variables as well as various food and nutrient variables measured at baseline, and their changes at 6 and 18 months. Systolic and diastolic BP, body weight and waist circumference significantly decreased, and physical activity significantly increased at 6 and 18 months when compared with baseline levels (all Po0.0001). Mean total caloric intake of all participants (overall across all three groups) decreased from 1939 to 1713 kcal day1 at 6 months (Po0.0001) and to 1705 kcal day1 at 18 months (Po0.0001) (data not presented in table). Dietary intake of the following food components increased significantly from baseline to 6 months and from baseline to 18 months (all Po0.0001): total Journal of Human Hypertension

protein (6 months: þ 1.5%, 18 months: þ 1.7%), animal protein ( þ 1.0, þ 1.3%), plant protein ( þ 0.4, þ 0.4%), fibre ( þ 2.3 g per 100 kcal per day, þ 2.2 g per 1000 kcal per day), fruits and vegetables ( þ 1.3 serving, þ 1.1 serving), dairy foods ( þ 0.2 serving, þ 0.1 serving), calcium ( þ 71.3 mg per 1000 kcal per day, þ 79.4 mg per 1000 kcal per day) and potassium ( þ 337.7 mg per 1000 kcal per day, þ 289.1 mg per 1000 kcal per day). Intake of fat serving and total fat reduced at both time points by about 1.1–1.3 servings and 4.2 and 3.4% kcal, respectively, as compared to baseline (both Po0.0001). Alcohol intake appeared to be stable over the course of study whereas meat consumption decreased about 0.3 serving per day but did not reach statistical significance. The urinary sodium reduced at both 6 and 18 months by 643.1 mg per 24 h and 370.4 mg per 24 h, respectively (all Po0.0001). On the basis of the cross-sectional analyses, no association was found between baseline total protein intake and baseline BP (all P40.05), nor between 18-month total protein intake and 18-month BP (6-month results shown in Table 3 model 1). However, when total protein was separated into plant and animal proteins (model 2), plant protein intake showed a strong inverse association with both systolic (P ¼ 0.0045) and diastolic BP (P ¼ 0.0096) after adjusting for all covariates and intakes of fat, fibre, calcium and potassium. Fruit/ vegetable intake was inversely associated with systolic BP (P ¼ 0.0003) and diastolic BP (P ¼ 0.0157) after controlling for all the covariates and intakes of dairy, meat and fat at 6 months (model 3). Total calorie was adjusted in all models by expressing nutrient intake as either % of total calorie (fat and protein) or g per 1000 kcal per day (alcohol, fibre) or mg per 1000 kcal per day (calcium, potassium). At 18 months, none of the above significant associations sustained and results were not presented in Table 3. Change in plant protein intake from baseline to 6 months was inversely associated with change in systolic BP (P ¼ 0.0486) but not with change in diastolic BP (P ¼ 0.0759) (Table 4, model 1). Both associations weakened after including waist circumference in the model (Table 4, model 2; systolic BP: P ¼ 0.0797, diastolic BP: P ¼ 0.0866). At 18 months, the association between change in plant protein intake and change in BP was not statistically significant (Table 4, model 3). Multiple logistic regression analysis was performed to examine the risk of hypertension with dietary intake adjusting for the same covariates as in the cross-sectional regression models. All participants, regardless if they were considered hypertensive or not at baseline, were included in this analysis. Plant protein intake was found to be significantly and inversely associated with the risk of hypertension at 6 months (Table 5, model 1), without adjusting for dietary fibre intake. The odds of having hypertension were 25% lower with every


Table 2 Changes in variables of interest from baseline to 6 and 18 monthsa Variable

Change at 6 months (6 months to baseline)

Baseline value

SBP (mm Hg) DBP (mm Hg) Weight (lb) Waist circumference (cm) Physical activity (estimated energy expenditure, kcal kg1 day1) Total calorie (kcal day1) Total protein (% kcal) Animal protein (% kcal) Plant protein (% kcal) Total fat (as nutrient, % kcal) Saturated fat (% kcal) Carbohydrate (% kcal) Cholesterol (mg per 1000 kcal per day) Alcohol (g per 1000 kcal per day) Dietary fibre (g per 1000 kcal per day) Calcium (mg per 1000 kcal per day) Potassium (mg per 1000 kcal per day) Fruit and vegetable (serving per day)c Dairy (serving per day) Meat (serving per day) Fat (as food, serving per day) Urinary sodium (mg per 24 h) Urinary urea nitrogen (mg per g creatinine)

Change at 18 months (18 months to baseline)

N

Mean±s.d.

N

Mean±s.d.

N

Mean±s.d.

810 810 810 810 808

134.9±9.6 84.8±4.2 209.8±41.5 107.6±15.2 33.7±2.9

810 810 756 737 712

9.4±9.9b 4.8±6.7b 8.4±12.1b 4.1±6.5b 0.4±2.7b

810 810 761 750 722

8.5±11.1b 5.3±7.6b 6.9±13.6b 3.2±7.1b 0.5±3.2b

807 807 807 807 807 807 807 807 807 807 807 807 807 807 807 807 798 798

1939.4±629.5 15.8±3.9 10.7±4.0 4.9±1.5 29.7±6.8 10.9±3.2 51.3±9.6 143.2±81.3 2.1±4.5 9.0±3.6 386.6±163.6 1369.5±407.1 1.8±1.5 1.7±1.3 2.5±1.4 5.7±3.8 4014.2±1672.7 7792±14026

712 712 712 712 710 710 710 710 710 710 710 710 710 710 710 710 653 653

1713.9±572.1 1.5±4.9b 1.0±5.2b 0.4±1.7b 4.2±8.8b 1.7±4.1 4.0±11.5 8.3±96.9 0.02±4.2 2.3±4.5b 71.3±207.1b 337.7±553.1b 1.3±3.23b 0.2±1.6b 0.3±1.6 1.1±4.4b 643.1±1733.4b 601±15 538

740 740 740 740 737 737 737 737 737 737 737 737 737 737 737 737 679 679

1705.7±565.3 1.7±4.9b 1.3±5.3b 0.4±1.8b 3.4±8.5b 1.6±3.8 2.8±11.3 6.6±102.6 0.02±4.3 2.2±4.5b 79.4±216.7b 289.1±519.3b 1.1±3.2b 0.1±1.4b 0.3±1.7 1.3±4.2b 370.4±1983.6b 641±14 535

Abbreviations: DBP, diastolic BP; SBP, systolic BP. a Paired t-test was used to compare the baseline to 6 or 18 months. b Po0.0001 when compared to baseline value, changes in alcohol and meat intake did not reach significance level of Po0.05. c A serving of each of the food group is defined as: 12 cup cut-up fruits, six ounces fruit juice, 12 cup cooked vegetables, one cup raw leafy vegetables, six ounces vegetable juice, 1.5 ounces cheese, six ounces yogurt, three ounces cooked meats, one tablespoon mayonnaise, two tablespoons salad dressing and one teaspoon oil.

one-unit (% of kcal) increase in dietary plant protein intake (odds ratio ¼ 0.75; 95% confidence interval: 0.60, 0.95). However, when dietary fibre intake was included in the model (Table 5 model 2), the association of plant protein with risk of hypertension became insignificant (P ¼ 0.0790), even though the association of fibre with risk of hypertension was not statistically significant. In model 3, only fruit/vegetable intake was significantly and inversely associated with risk of hypertension. For every extra serving of fruit/vegetable, the odds of having hypertension decreased by about 23% (odds ratio ¼ 0.77; 95% confidence interval: 0.79, 0.97).

Discussion In most but not all of our analyses, a higher intake of plant protein was associated with lower BP in individuals with pre-hypertension or stage 1 hypertension. We discovered a significant inverse association between plant protein intake and both systolic and diastolic BP at 6 months cross-sectionally. In addition, change in plant protein intake was inversely associated with change in systolic BP at 6 months in one of the models examined but the association reduced to a nonsignificant level at

18 months. In separate regression models, fruit and vegetable intake was inversely associated with systolic BP at 6 months only. No significant association was found between total or animal protein and BP at any time point. Our findings are consistent with previous studies that showed an inverse relationship between plant protein and BP. The Chicago Western Electric Study,9 a longitudinal epidemiological study with a 9-year follow-up, reported that baseline dietary plant protein was inversely associated with annual change in systolic BP in men. A recent crosssectional study with 4680 participants also reported that plant protein intake, but not animal protein intake, was inversely associated with hypertension.4 This study also indicated that subjects with the highest vegetable protein/lowest animal protein consumption had mean reductions in systolic and diastolic BP that were 4.15 and 2.15 mm Hg, respectively, lower than those for subjects in the lowest vegetable/highest animal protein consumption group. It is unclear why our results did not support an inverse relationship between total protein intake and BP as reported by many previous studies.20–22 Intake of total protein and fibre was reported to have significant additive effects in lowering 24-h and Journal of Human Hypertension


Table 3 Adjusted cross-sectional association between blood pressure and dietary intake at 6 months of follow-up Outcome ¼ SBP

Analysis model

Outcome ¼ DBP

Effect

s.e.

P

Effect

s.e.

P

Model 1 (nutrient intake: 6 months) a Total protein intake (% kcal) Total fat (% kcal) Dietary fibre (g per 1000 kcal per day) R2

0.08 0.05 0.12 0.4795

0.09 0.05 0.12

0.4114 0.3236 0.3368

0.03 0.06 0.09 0.2157

0.08 0.04 0.11

0.7359 0.1594 0.4068

Model 2 (nutrient intake: 6 months) a Plant protein (% kcal) Animal protein (% kcal) Total fat (% kcal) Dietary fibre (g per 1000 kcal per day) R2

0.98 0.08 0.02 0.17 0.4884

0.34 0.10 0.05 0.15

0.0045 0.4249 0.6685 0.2649

0.70 0.03 0.04 0.15 0.2280

0.27 0.08 0.04 0.12

0.0096 0.7137 0.3662 0.2103

Model 3 (food intake: 6 months) b Fruit and vegetable (serving per day) Dairy food (serving per day) Meat (serving per day) Fat (serving per day) R2

0.48 0.28 0.35 0.12 0.4801

0.13 0.32 0.32 0.12

0.0003 0.3920 0.2656 0.3039

0.25 0.28 0.41 0.001 0.2127

0.10 0.25 0.25 0.09

0.0157 0.2784 0.1026 0.9948

Abbreviations: DBP, diastolic BP; SBP, systolic BP. All dietary intake variables are either as per cent of calories (protein and fat intake), g per 1000 kcal per day (alcohol, fibre, calcium and potassium) or serving per day (model 2), where calorie was controlled in the model. a Models 1 and 2: controlled for age, gender, race, treatment, study site, education, income, baseline blood pressure, dietary intake of alcohol, calcium and potassium, body weight, waist circumference, physical activity, urinary creatinine and sodium excretion, and dietary intake of calcium and potassium at 6 months. b Model 3: controlled for age, gender, race, treatment, study site, education, income, alcohol intake, body weight, waist circumference, physical activity, urinary creatinine and sodium excretion, and dietary calorie intake at 6 months.

Table 4 Adjusted longitudinal association of blood pressure and dietary intake Analysis model a

Change of SBP (mm Hg) Effect

s.e.

Change of DBP (mm Hg) P

Effect

s.e.

P

Model 1 (changes from baseline to 6 months): without change of waist circumference Change in plant protein (% kcal) 0.59 0.30 0.0486 Change in animal protein (% kcal) 0.03 0.09 0.7258 Change in total fat (% kcal) 0.00 0.05 0.9943 Change in dietary fibre (g per 1000 kcal per day) 0.13 0.14 0.3391 0.3348 R2

0.38 0.03 0.01 0.02 0.2794

0.22 0.06 0.04 0.10

0.0759 0.9663 0.8617 0.8464

Model 2 (changes from baseline to 6 months): with change of waist circumference Change in plant protein (% kcal) 0.53 0.30 Change in animal protein (% kcal) 0.02 0.09 Change in total fat (% kcal) 0.01 0.05 Change in dietary fibre (g per 1000 kcal per day) 0.13 0.14 0.3357 R2

0.37 0.01 0.006 0.02 0.2793

0.22 0.06 0.04 0.10

0.0866 0.9663 0.8656 0.8249

0.10 0.05 0.02 0.13 0.2130

0.22 0.07 0.04 0.11

0.6449 0.4181 0.6275 0.2499

0.0797 0.8431 0.8471 0.3509

Model 3 (changes from baseline to 18 months): with change of waist circumference Change in plant protein (% kcal) 0.30 0.31 0.3449 Change in animal protein (% kcal) 0.07 0.09 0.4751 Change in total fat (% kcal) 0.05 0.06 0.3814 Change in dietary fibre (g per 1000 kcal per day) 0.07 0.15 0.6377 0.2218 R2

Abbreviations: DBP, diastolic BP; SBP, systolic BP. a All results are controlled for age, gender, race, treatment, study site, education, income, baseline blood pressure and changes of alcohol intake, physical activity, waist circumference, urinary creatinine, sodium excretion, body weight and dietary intake of calcium, and potassium at 6 or 18 months, respectively. All dietary intake variables are either as per cent of calories (protein and fat intake) or as g per 1000 kcal per day (alcohol, fibre), mg per 1000 kcal per day (calcium and potassium). Journal of Human Hypertension


Table 5 Odds ratio and 95% confidence interval for the presence of hypertension in relation to dietary nutrient and food intake at 6 months of follow-up Analysis modela

ORb

95% CIb

P-value

Model 1 (without fibre) Plant protein (% kcal) Animal protein (% kcal) Total fat (% kcal)

0.75 1.00 0.95

0.60, 0.95 0.93, 1.07 0.91, 1.00

0.015 0.9309 0.0520

Model 2 (with fibre) Plant protein (% kcal) Animal protein (% kcal) Total fat (% kcal) Fiber (g per 1000 kcal per day)

0.79 0.99 0.95 0.96

0.60, 0.93, 0.91, 0.84,

1.02 1.07 0.99 1.09

0.0790 0.9032 0.0465 0.5013

Model 3 Fruit and vegetable (serving per day) Meat (serving per day) Dairy (serving per day) Fat (serving per day)

0.85 1.01 0.99 0.97

0.77, 0.80, 0.78, 0.89,

0.95 1.29 1.26 1.07

0.0041 0.9083 0.9396 0.5803

Abbreviations: CI, confidence interval; OR, odds ratio. a All results are controlled for baseline characteristics including age, gender, race, treatment, study site, education, income and blood pressure; and for 6-month covariates of body weight, waist circumference, physical activity level, intakes of calcium, potassium, fibre and alcohol, urinary creatinine and sodium excretion. All dietary intake variables are either as per cent of calories (protein and fat intake), g or mg per 1000 kcal per day (alcohol, fibre, calcium and potassium) or serving per day (in model 3). b OR and its 95% CI are for increment of one unit change in each corresponding predictor while fixing the remaining predictors at a constant.

awake systolic BP in hypertensive subjects.13 The Intersalt Study, a large cross-sectional international study with 10 020 men and women aged 20–59 years from 54 population-based samples in 32 countries worldwide, also demonstrated that total dietary protein intake had an inverse relationship with BP.23 Similarly, Liu et al.22 concluded in a metaanalysis report that the inverse association between dietary protein intake and BP was quite convincing in nine population-based cross-sectional studies. In addition, the inverse relationship between total protein intake and BP was further supported by two large controlled feeding trials.14 In the original DASH study, a dietary pattern that was higher in total protein content than the control diet (17.9 vs 13.8% of energy) and rich in vegetable, fruit and dairy content significantly lowered BP.11 The OmniHeart Randomized Trial14 showed that a diet that substituted carbohydrate with protein, mostly from plant sources, or with monounsaturated fat can further lower BP, improve lipid levels and reduce estimated cardiovascular risk. Although the exact mechanism(s) linking plant or total protein to BP is still unclear, possible explanations exist. First, an increase in protein intake may increase plasma amino acids that may directly influence proximal sodium reabsorption or alter cell permeability and subsequently increase renal plasma flow, renal size and glomerular filtration rate.24–27 Second, the amino acid arginine may act as a vasodilator through nitric oxide and contribute to BP lowering.28 Plant foods including soy products, nuts, seeds and certain whole grains like oats are all good sources of arginine but so are animal foods such as animal meats, seafood, dairy products and

eggs. Third, plant protein may serve as a surrogate marker for intake of other nutrients or food components that were not identified in our study and that may benefit BP control. Soy protein and isoflavone have been suggested to be beneficial for BP control. Several randomized trials have shown that supplementation of soy protein significantly reduced BP.12,29–31 As the soy protein or isoflavone contents in the diets of the current study participants were unavailable, we cannot determine if these dietary factors accounted for our observed relationship between plant protein intake and BP. The lack of a significant relationship between animal protein intake and BP in our study is not consistent with some studies;5,7,8,25 however, this is not totally unexpected. Previous studies that have reported an inverse relationship between animal protein intake and BP have mainly been conducted in Japan and China where animal protein intake was relatively low.5,7,8 Zhou et al.7 found an inverse association of dietary animal protein intake with BP in a rural population of China in 1989. The followup study with a sample of 705 men and women in 1994 confirmed their previous finding.8 He et al.25 also reported an inverse association of animal protein, but not plant protein, with BP in another sample of Chinese population. The populations involved in these studies were living in underdeveloped rural areas of China and were relatively poor compared with the general Chinese population; thus, their dietary patterns differed significantly from those of the populations in the Western countries. Compared with the general Chinese population, these study populations had a relatively Journal of Human Hypertension


lower intake of total protein, averaging 11% of total energy intake according to the 1992 China National Nutrition Survey;32 thus intake of animal protein is likely to be even lower as compared with the US population.33 In addition, the socio-economic factors that might affect BP were not considered in these analyses. In the present study, the study population had relatively moderate intake of total protein (15.8%) and animal protein (10.7%). Even though the meat consumption decreased from baseline to 6 and 18 months, consumption of dairy, poultry and seafood increased and thus contributed to the small increase in animal protein at both time points. In addition, the type of animal protein consumed by the various study populations in previous studies may also differ in other nutritional components that may complicate the examination of their impact on BP. For example, even though both fish and meats are rich sources of animal protein, the former is rich in omega 3 fatty acids, which may benefit BP control, whereas the latter is usually rich in saturated fat that may have a negative impact on BP control.34 Similarly, dairy products are rich in animal protein and calcium and can be high in saturated fat; each of these components may impact BP differently. Although dietary fibre has been suggested to benefit BP control,35 intake of dietary fibre was not significantly associated with presence of hypertension at 6 months in the present study. However, adding fibre into the analysis model weakened the association between plant protein and BP. We also found that intake of fruits and vegetables, the major sources of fibre, was inversely associated with BP at 6 months. Our findings were consistent with many previous studies that suggested that fibre-rich food, such as fruits and vegetables, may exert a protective effect in the cardiovascular system by lowering BP. Potential mechanisms of the BP-lowering effect of fibre may include enhancing insulin sensitivity and improving vascular endothelial function,36 and improving magnesium absorption in the gastrointestinal system, which then indirectly benefits BP control.37,38 Thus increasing fibre intake in Western populations, where current US intake is far below recommended levels, may help prevent and control hypertension.14,30,35,39 Even though fruit and vegetable intake and plant protein intake each had a significant association with BP in our results, the potential mechanisms for their influence on BP may differ. The benefit of fruit and vegetable intake on BP may be related to their high contents of antioxidant vitamins, potassium or magnesium. Plant protein, on the other hand, may affect BP via mechanisms related to amino acid composition such as content of the vasodilator arginine, as mentioned previously, or through other unidentified factors. Although fruits and vegetables also provide plant protein, the contribution is far less than that from grains. According to a report Journal of Human Hypertension

analysing the National Health and Nutrition Examination Surveys III data,40 grains contributed 54% of total plant protein intake whereas fruits and vegetables combined contributed 29.5%. Legumes, nuts and seeds also contributed about 12% to the total plant protein intake. In this study, the increases in plant protein occurred mainly through increased consumption of whole grains, vegetables and fruits. On average, approximately 33–36% of the daily total protein intake was from plant sources. The contributions of refined grains, whole grains, vegetables, fruits, nuts/seeds and legumes (such as soy products) to the total plant protein intakes were approximately 29–41, 12–14, 18–26, 6–15, 3–6 and 0.3–4%, respectively. The loss of statistical significance in multivariable regression analyses when additional nutrients were entered into the regression models (such as fibre added in model 2 of Table 5) is consistent with colinearity, that is, nutrients are often interrelated in the diet of free-living persons. Similar findings have been reported in numerous studies of nutrition and BP.41 In the setting of observational studies, it is often difficult to attribute biological effects to individual nutrients or foods. However, at least for protein, prospective clinical trials indicate that increased plant protein intake can help lower BP.14 In this study, the total intake of plant protein was only 4.9% of total calories at baseline, and it increased by 0.4 (% of calories) at 6 and 18 months of follow-up with relatively small variation. This might have minimized the true association between BP and protein intake over the study period. The other limitation of this study includes using only the two 24-h diet recalls to assess dietary intake at each time point, which might not have reflected usual dietary intake of the study participants. However, collecting more than two recalls per time point in such a large-scale clinical trial is not practical and often not feasible. Furthermore, the large sample size should minimize the effect of variation due to occasional capture of non-usual dietary intake. Overall, the PREMIER participants made dietary changes that were consistent with the current recommendations for BP control.42 These changes included reducing fats and meats, and increasing fruits, vegetables and dairy products, which then resulted in lowered intakes of total fat, saturated fat, sodium and increased intakes of dietary fibre, calcium and potassium. In addition, these participants increased physical activity levels and reduced body weight. Amid all the dietary changes, plant protein intake was related to BP quite consistently. A higher intake of plant protein was associated with lower BP in most but not all of our analyses. Neither total nor animal protein was associated with BP. Our results confirm other studies showing that regular consumption of fruits and vegetables, nuts/seeds, whole grains and soy products is likely to benefit BP. Although the exact mechanism by which plant


protein affects BP is uncertain, there exist several viable hypotheses. Increasing plant protein sources is consistent with the current guideline of adopting an eating pattern like DASH for BP management.43 Further research is needed to confirm the association and to elucidate the exact mechanisms.

8

9 What is known about this topic K Many earlier studies, mainly observational, have shown an inverse relationship between total dietary protein intake and blood pressure.1 5,6 K Cross-sectional studies conducted in rural Japanese and Chinese7,8 populations only found an inverse relationship between animal protein and blood pressure. However, cohort studies9,10 conducted in the United States have shown an inverse association between baseline intakes of plant protein, but not animal protein, and annual change in systolic blood pressure. K It is unclear if plant and animal proteins have similar impact on blood pressure. What this study adds K This study of secondary data analysis found that a higher intake of plant protein, not total and animal protein, was associated with lower blood pressure. K This study finding suggests that increasing plant protein may benefit blood pressure control.

10

11

12

13

14

Acknowledgements Dr Yancy is supported by a Veterans’ Affairs Health Services Research Career Development Award (RCD 02-183-1).

15

References

16

1 Obarzanek E, Velletri PA, Cutler JA. Dietary protein and blood pressure. JAMA 1996; 275: 1598–1603. 2 Iseki K, Iseki C, Itoh K, Sanefuji M, Uezono K, Ikemiya Y et al. Estimated protein intake and blood pressure in a screened cohort in Okinawa, Japan. Hypertens Res 2003; 26: 289–294. 3 Jenkins DJ, Kendall CW, Vuksan V, Vidgen E, Parker T, Faulkner D et al. Soluble fiber intake at a dose approved by the US Food and Drug Administration for a claim of health benefits: serum lipid risk factors for cardiovascular disease assessed in a randomized controlled crossover trial. Am J Clin Nutr 2002; 75: 834–839. 4 Elliott P, Stamler J, Dyer AR, Appel L, Dennis B, Kesteloot H et al. Association between protein intake and blood pressure: the INTERMAP Study. Arch Intern Med 2006; 166: 79–87. 5 Yamori Y, Kihara M, Nara Y, Ohtaka M, Horie R, Tsunematsu T et al. Hypertension and diet: multiple regression analysis in a Japanese farming community. Lancet 1981; 1: 1204–1205. 6 Kihara M, Fujikawa J, Ohtaka M, Mano M, Nara Y, Horie R et al. Interrelationships between blood pressure, sodium, potassium, serum cholesterol, and protein intake in Japanese. Hypertension 1984; 6: 736–742. 7 Zhou BF, Wu XG, Tao SQ, Yang J, Cao TX, Zheng RP et al. Dietary patterns in 10 groups and the relationship

17

18

19

20

21

with blood pressure. Collaborative study group for cardiovascular diseases and their risk factors. Chin Med J (Engl) 1989; 102: 257–261. Zhou B, Zhang X, Zhu A, Zhao L, Zhu S, Ruan L et al. The relationship of dietary animal protein and electrolytes to blood pressure: a study on three Chinese populations. Int J Epidemiol 1994; 23: 716–722. Stamler J, Liu K, Ruth KJ, Pryer J, Greenland P. Eight-year blood pressure change in middle-aged men: relationship to multiple nutrients. Hypertension 2002; 39: 1000–1006. Stamler J, Caggiula AW, Grandits GA. Relation of body mass and alcohol, nutrient, fiber, and caffeine intakes to blood pressure in the special intervention and usual care groups in the Multiple Risk Factor Intervention Trial. Am J Clin Nutr 1997; 65: 338S–365S. Appel L, Moore T, Obarzanek E, Vollmer W, Svetkey L, Sacks F et al. A clinical trial of the effects of dietary patterns on blood pressure. N Engl J Med 1997; 336: 1117–1124. Washburn S, Burke G, Morgan T, Anthony M. Effect of soy protein supplementation on serum lipoproteins, blood pressure, and menopausal symptoms in perimenopausal women. Menopause 1999; 6: 7–13. Burke V, Hodgson JM, Beilin LJ, Giangiulioi N, Rogers P, Puddey IB. Dietary protein and soluble fiber reduce ambulatory blood pressure in treated hypertensives. Hypertension 2001; 38: 821–826. Appel LJ, Sacks FM, Carey VJ, Obarzanek E, Swain JF, Miller III ER et al. Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: results of the OmniHeart randomized trial. J Am Med Assoc 2005; 294: 2455–2464. Hodgson JM, Burke V, Beilin LJ, Puddey IB. Partial substitution of carbohydrate intake with protein intake from lean red meat lowers blood pressure in hypertensive persons. Am J Clin Nutr 2006; 83: 780–787. Funk KL, Elmer PJ, Stevens VJ, Harsha DW, Craddick SR, Lin PH et al. PREMIER—a trial of lifestyle interventions for blood pressure control: intervention design and rationale. Health Promot Pract 2008; 9: 271–280. Appel LJ, Champagne CM, Harsha DW, Cooper LS, Obarzanek E, Elmer PJ et al. Effects of comprehensive lifestyle modification on blood pressure control: main results of the PREMIER clinical trial. [comment]. JAMA 2003; 289: 2083–2093. The sixth report of the Joint National Committee on prevention, detection, evaluation and treatment of high blood pressure. Arch Intern Med 1997; 157: 2413–2446. Washburn RA, Jacobsen DJ, Sonko BJ, Hill JO, Donnelly JE. The validity of the stanford seven-day physical activity recall in young adults. Med Sci Sports Exerc 2003; 35: 1374–1380. Stamler J, Caggiula A, Grandits GA, Kjelsberg M, Cutler JA. Relationship to blood pressure of combinations of dietary macronutrients. Findings of the Multiple Risk Factor Intervention Trial (MRFIT). Circulation 1996; 94: 2417–2423. Stamler J, Elliott P, Kesteloot H, Nichols R, Claeys G, Dyer AR et al. Inverse relation of dietary protein markers with blood pressure. Findings for 10 020 men and women in the INTERSALT Study. INTERSALT Cooperative Research Group. INTERnational study of SALT and blood pressure. Circulation 1996; 94: 1629–1634. Journal of Human Hypertension


22 Liu L, Ikeda K, Sullivan DH, Ling W, Yamori Y. Epidemiological evidence of the association between dietary protein intake and blood pressure: a metaanalysis of published data. Hypertens Res 2002; 25: 689–695. 23 Stamler J, Elliott P, Kesteloot H, Nichols R, Claeys G, Dyer A et al. Inverse relation of dietary protein markers with blood pressure. Findings for 10 020 men and women in the INTERSALT Study. Circulation 1996; 94: 1629–1634. 24 Yamori Y, Horie R, Tanase H, Fujiwara K, Nara Y, Lovenberg W. Possible role of nutritional factors in the incidence of cerebral lesions in stroke-prone spontaneously hypertensive rats. Hypertension 1984; 6: 49–53. 25 He J, Klag M, Whelton P, Chen J, Qian M, He G. Dietary macronutrients and blood pressure in southwestern China. J Hypertens 1995; 13: 1267–1274. 26 Cho MM, Yi MM. Variability of daily creatinine excretion in healthy adults. Hum Nutr Clin Nutr 1986; 40: 469–472. 27 Woods LL. Mechanisms of renal hemodynamic regulation in response to protein feeding. Kidney Int 1993; 44: 659–675. 28 Palloshi A, Fragasso G, Piatti P, Monti LD, Setola E, Valsecchi G et al. Effect of oral L-arginine on blood pressure and symptoms and endothelial function in patients with systemic hypertension, positive exercise tests, and normal coronary arteries. Am J Cardiol 2004; 93: 933–935. 29 Teede HJ, Dalais FS, Kotsopoulos D, Liang YL, Davis S, McGrath BP. Dietary soy has both beneficial and potentially adverse cardiovascular effects: a placebo-controlled study in men and postmenopausal women. J Clin Endocrinol Metab 2001; 86: 3053–3060. 30 Jenkins DJ, Kendall CW, Connelly PW, Jackson CJ, Parker T, Faulkner D et al. Effects of high- and lowisoflavone (phytoestrogen) soy foods on inflammatory biomarkers and proinflammatory cytokines in middle-aged men and women. Metabolism 2002; 51: 919–924. 31 Sagara M, Kanda T, M NJ, Teramoto T, Armitage L, Birt N et al. Effects of dietary intake of soy protein and isoflavones on cardiovascular disease risk factors in high risk, middle-aged men in Scotland. J Am Coll Nutr 2004; 23: 85–91.

Journal of Human Hypertension

32 Ge K. The Dietary and Nutritional Status of Chinese Population: 1992 National Nutrition Survey. People’s Medical Publishing House: Beijing, China, 1996. 33 Bialostosky K, Wright JD, Kennedy-Stephenson J, McDowell M, Johnson CL. Dietary intake of macronutrients, micronutrients, and other dietary constituents: United States 1988–94. Vital Health Stat 2002; 11: 1–158. 34 Lin PH, Batch B, Svetkey LP. Nutrition, lifestyle and hypertension. In: Coulston AM, Boushey CJ (eds). ‘Nutrition in the Prevention and Treatment of Disease, 2nd edn. Academic Press: Baltimore, MA, USA, 2008, pp 551–574. 35 Streppel MT, Arends LR, Van’T Veer P, Grobbee DE, Geleijnse JM. Dietary fiber and blood pressure: a metaanalysis of randomized placebo-controlled trials. Arch Intern Med 2005; 165: 150–156. 36 Cleland SJ, Petrie JR, Ueda S, Elliott HL, Connell JM. Insulin as a vascular hormone: implications for the pathophysiology of cardiovascular disease. Clin Exp Pharmacol Physiol 1998; 25: 175–184. 37 Coudray C, Demigne C, Rayssiguier Y. Effects of dietary fibers on magnesium absorption in animals and humans. J Nutr 2003; 133: 1–4. 38 Greger JL. Nondigestible carbohydrates and mineral bioavailability. J Nutr 1999; 129(7 Suppl): 1434S–1435S. 39 Keenan JM, Pins JJ, Frazel C, Moran A, Turnquist L. Oat ingestion reduces systolic and diastolic blood pressure in patients with mild or borderline hypertension: a pilot trial. J Fam Pract 2002; 51: 369. 40 Smit E, Nieto FJ, Crespo CJ, Mitchell P. Estimates of animal and plant protein intake in US adults: results from the Third National Health and Nutrition Examination Survey, 1988–1991. J Am Diet Assoc 1999; 99: 813–820. 41 Beilin LJ. Vegetarian and other complex diets, fats, fiber, and hypertension. Am J Clin Nutr 1994; 59(5 Suppl): 1130S–1135S. 42 The seventh report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure. National Heart Lung and Blood Institute, NIH: Bethesda, 2004. 04–5230. 43 Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo Jr JL et al. The seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure: the JNC 7 report. JAMA 2003; 289: 2560–2572.