Dietary Protein and Soluble Fiber Reduce Ambulatory Blood Pressure in Treated Hypertensives
In population studies, higher blood pressure has been associated with lower intake of protein and, possibly, lower fiber consumption. In the present randomized controlled trial, we sought to determine whether dietary protein and fiber had additive effects on blood pressure reduction in hypertensives. Treated hypertensive patients changed for 4 weeks (familiarization) to a diet low in protein (12.5% energy) and fiber (15 g/d). Patients (n=41) were then randomized to 1 of 4 groups in an 8-week factorial study of parallel design in which they continued the low-protein, low-fiber diet alone or had supplements of soy protein to increase protein intake to 25% energy, of psyllium to provide an additional 12 g soluble fiber/d, or of both protein and fiber. The 24-hour ambulatory blood pressure was compared from the end of familiarization to the end of intervention. In the 36 subjects who provided complete data, protein and fiber had significant additive effects to lower 24-hour and awake systolic blood pressure. Relative to control subjects, the net reduction in 24-hour systolic blood pressure was 5.9 mm Hg with fiber and with protein. Findings were independent of age, gender, and change in weight, alcohol intake, or urinary sodium and potassium. Relative to reduced fiber and protein intake, dietary protein and soluble fiber supplements lower blood pressure additively in hypertensives. These findings have important implications for the prevention and management of hypertension, particularly in populations in which high blood pressure is prevalent in association with diets low in protein, fiber, or both.
In population studies, estimated protein intake has been inversely related to blood pressure (BP).1–7 Controlled trials that focused on dietary protein and BP have usually compared vegetable and animal proteins, investigating whether vegetarians have a lower BP,8 and have in general proved to be negative. However, increased dietary protein attenuated the pressor effect of salt among young volunteers with a family history of hypertension.9 Recently, soy protein supplements of 40 g/d were found to lower systolic BP (SBP) by 3 mm Hg and diastolic BP (DBP) by ≈2 mm Hg, relative to control subjects,10 in 35- to 65-year-old Chinese subjects with untreated high-normal BP.
Dietary fiber may also be related inversely to BP.2,11–13 A recent meta-analysis14 concluded that fiber probably has a small BP-lowering effect. In Chinese populations, BP has been associated inversely with both protein and fiber intake,2 consistent with additive effects. However, associations with specific nutrients may arise through other dietary or lifestyle differences that affect BP.
Given the worldwide prevalence of hypertension, it is important to resolve whether dietary protein and fiber can influence BP alone or additively. We therefore conducted a randomized controlled trial in treated hypertensives in whom protein or fiber intake, or both, was increased against a background of a standardized diet low in both constituents.
Nonsmoking men or women who were ≥20 years old and were receiving drug therapy for hypertension were recruited via advertisement. Inclusion criteria included drug therapy for ≥6 months with ≤2 antihypertensive agents, SBP between 130 and 160 mm Hg, and alcohol intake of ≤210 g alcohol/wk. Exclusion criteria included diabetes, renal disease (creatinine >130 mmol/L), symptomatic heart disease, regular use of nonsteroidal anti-inflammatory drugs, psychiatric illness, or body mass index of >33 kg/m2 for men and >37 kg/m2 for women. The study was approved by the University of Western Australia Committee for Human Rights, and all participants gave written consent.
During the initial 4-week period (baseline), subjects followed their usual diet, and nutrient intake was assessed individually from 3-day weighed-food records to establish isoenergetic, weight-maintenance diets for the subsequent intervention. In the next 4 weeks (familiarization period), subjects were supplied home-delivered meals that provided 12.5% of energy from protein and 15 g fiber/d. Participants then entered into an 8-week-long study of factorial design in which they were randomized to continue the 12.5% protein–15 g fiber/d diet (low fiber, low protein) or to consume a protein additive to provide a total of 25% energy as protein. Within each arm, subjects were further randomized to continue the low fiber intake or to consume an additional 15 g psyllium/d, providing 12 g soluble fiber/d. The low-protein diet provided an intake of protein equal to the 10th percentile for the average Australian diet, and the supplements equaled a protein intake of the 90th percentile.15
Protein was provided as 66 g soy protein/d (Archer Daniels Midland Co). The equivalent energy intake was provided as a supplement of 66 g maltodextrin powder/d (Starch Australasia Ltd) in groups not consuming the soy supplement. Soluble fiber was provided as psyllium husks from a single batch (DHA Pty Ltd Designer Wise Foods). Supplements were provided as weighed daily amounts to be taken as a drink mixed with juice or water throughout the day; these consisted of (1) maltodextrin only in the low-fiber, low-protein group; (2) soy protein only in the low-fiber, high-protein group; (3) psyllium with maltodextrin in the high-fiber, low-protein group; and (4) soy protein with psyllium in the high-fiber, high-protein group. Because the soy protein was higher in potassium than the maltodextrin or psyllium supplements, an additional 786 mg potassium/d, as potassium chloride, was added to the nonprotein supplements.
Dietary meals (lunch, dinner, snack, and fruit drinks) with specified nutrient content were obtained from Home Chef. Participants provided breakfasts, based on a list of permissible low-fiber cereals or low-fiber white bread. Meals were based on 2 serving sizes, supplying ≈8 and ≈10 MJ/d, respectively. Participants were instructed on appropriate deviations from their supplied meals to ensure that energy requirements were met and not exceeded. Individual adjustments were made as needed based on food diaries, weight monitoring, and weekly reviews with the same dietician. Compliance with the supplements was monitored by weekly issue of counted sachets and a final check on the number of sachets returned at the end of the study. Urinary urea provided a further check on compliance in the groups receiving protein supplements. Alcohol intake was assessed using 7-day retrospective diaries at the end of baseline, familiarization, and intervention periods.
Block randomization was carried out at the end of familiarization using computer-generated randomly assigned numbers provided by the statistician. Differences in taste and texture of the supplements prevented blinding of the participants and the staff carrying out the intervention, who also needed to provide specific dietary advice to participants who had difficulty in compliance with the intervention. Outcome assessment depended on electronically recorded ambulatory BP data and was independent of assessment by trial staff.
Anthropometry, Biochemical Variables, and Lifestyle Monitoring
Weight was recorded weekly; height was measured with a stadiometer. During the weekly visits, any changes in alcohol intake, physical activity, health status, and medication use were monitored by interview. Laboratory measurements, carried out at baseline and at the end of each study period, included serum electrolytes, urea, creatinine, uric acid, and 24-hour urinary creatinine, sodium, and potassium.
Ambulatory BP Monitoring
BP was recorded over 24 hours at the end of baseline, familiarization, and intervention with an Accutracker II (model 104; Suntech). BP was recorded every 30 minutes during waking hours and every hour during sleep. A nurse fitted the monitor and explained its use to the volunteers, who completed diaries describing their activity at the time that ambulatory BP (ABP) was recorded. If the pulse pressure was <20 mm Hg or readings were associated with a test code, they were excluded from the analysis.
Xyris software, based on the 1995 Australian NUTTAB database, was used to assess nutrient intake. Characteristics of groups at baseline were compared by ANOVA. Effects of dietary change on change in 24-hour, awake, and asleep SBP, DBP, and heart rate were examined using pooled time series random effects models (PROC MIXED; SAS Institute) with change in mean 24-hour ABP as the primary outcome measure. Age, gender, change in weight, change in urinary sodium, and change in alcohol intake were included in subsequent models. P<0.05 was considered significant.
Subjects and Recruitment
Recruitment was performed over 1 year. The Figure shows the trial profile with the numbers of participants who were randomized and those who completed the study. On screening, 190 subjects were excluded because of smoking, medication, diabetes, illness, dietary restrictions (including the need for a high-fiber diet), participation in a weight-loss program or vegetarian diet, body mass index above the exclusion limits, distance from the trial center, or inability to attend the department because of time commitments. Before baseline, 9 volunteers decided to withdraw, 2 men were excluded because of high creatinine levels, and 2 women were excluded because they were being treated with >2 antihypertensive drugs. Four withdrew before randomization because of dietary intolerance, illness, or holidays, and 2 men were excluded because non–insulin-dependent diabetes mellitus was detected at the screening. The main reasons for dietary intolerance were distaste for the meals provided and difficulty in drinking the maltodextrin supplement. Forty-one volunteers were randomized, and 36 provided complete data. There were 3 withdrawals during the intervention related to dietary intolerance and 1 withdrawal due to a change in antihypertensive medication. One subject in the high-fiber, high-protein group completed the intervention, but acute illness prevented recording of the postintervention ABP. Table 1 shows that baseline characteristics were similar between groups.
Analysis of nutrient intake confirmed compliance with the diets, showing a significant increase in protein, fiber, or carbohydrate intake appropriate to group allocation (Table 2). There were no other significant changes in dietary variables with no significant between-group differences in sodium or potassium intake, confirming that potassium added to the low-protein supplements achieved intakes equivalent to those from the soy protein supplements.
Body Weight, Urinary Analytes, and Antihypertensive Drugs
There were no significant differences in body weight related to the intervention (Table 3). Urine urea increased significantly with higher protein intake (mean change relative to nonprotein groups, 228.4 mmol/L; 95% confidence limits [CL], 156.1 and 300.7 mmol/L), but there was no significant change in the group with fiber only (net mean change, −121.3 mmol/L; 95% CL, −375.0 to 132.5 mmol/L). Changes in serum creatinine and urea and urinary potassium, sodium, sodium/potassium ratio, and creatinine did not differ significantly between treatment groups. Except for the 1 patient who withdrew from the study, there was no change in the dose or type of antihypertensive medication during the trial.
ABP and Heart Rate
Table 3 shows the mean 24-hour SBP, DBP, and heart rate at the end of familiarization and at the end of intervention, with the greatest fall in BP in the high-fiber, high-protein group. Net changes in mean 24-hour, asleep, and awake SBP, DBP, and heart rate for main effects are shown in Table 4. Random effects models showed significant independent effects of fiber and protein on 24-hour SBP and awake SBP and a significant effect of protein on asleep SBP. Interactions between protein and fiber were not significant. Change in 24-hour and awake DBP was significantly related to protein but not to fiber. Changes in BP were independent of age, gender, and changes in weight, alcohol intake, and urinary sodium or potassium excretion. Protein significantly reduced 24-hour and awake heart rates.
Relative to control subjects, 24-hour mean SBP fell by 2.4 mm Hg (95% CL, −10.0 and 5.2 mm Hg) in the high-fiber, low-protein group; by 2.9 mm Hg (95% CL, −14.5 and 2.8 mm Hg) in the low-fiber, high-protein group; and by 10.5 mm Hg (95% CL, −20.4 and −0.6 mm Hg) in the high-fiber, high-protein group. Net decreases for DBP were 1.9 mm Hg (95% CL, −7.8 and 4.0 mm Hg), 2.5 mm Hg (95% CL, −6.6 and 1.6 mm Hg), and 3.6 mm Hg (95% CL, −8.3 and 1.1 mm Hg), respectively, whereas heart rate increased by 3.5 bpm (95% CL, −5.7 and 12.7 bpm) in the high-fiber, low-protein group and fell by 0.5 bpm (95% CL, −5.4 and 4.4 bpm) in the low-fiber, high-protein group and by 1.7 bpm (95% CL, −5.8 and 3.1 bpm) in the high-fiber, high-protein group.
This study is the first to examine the independent and combined effects on 24-hour BP of dietary protein and soluble fiber. Relative to a low-protein, low-fiber diet, SBP fell by ≈6 mm Hg both with soy protein sufficient to increase protein intake from the 10th to the 90th percentile for Australian diets15 and with psyllium supplements that doubled the fiber intake. Additive effects of protein and fiber produced the greatest fall in BP in the group that received both supplements. Our findings are consistent with population studies that link lower BP with higher intake of protein and fiber and indicate a causal association.
Population studies suggest an inverse association between SBP and protein intake of either animal or vegetable origin. Among rural Japanese, a higher ratio of urinary sulfate to nitrogen, reflecting the intake of animal protein, was related to lower SBP in men.16 In a 7-year prospective study in the United States, in which animal products are the main source of dietary protein, a change in SBP related inversely to protein intake at baseline.5 Similar inverse relationships have been reported for the intake of vegetable protein2,4 and animal protein.6,7
Amino acid content of proteins may be relevant to effects on BP.17 l-Arginine, acting via NO,18 lowers BP and improves endothelial function. Taurine lowers BP in animals19 and is inversely related to BP in population studies.20 A nonspecific dose-dependent diuretic effect of amino acids may also contribute to the BP-lowering effect.21
There are no satisfactory controlled trials that compare the effects of both the quantity and the type of protein on BP (see Beilin8 for a review). We used a supplement of soy protein rather than of animal protein to minimize changes in other nutrients, but soy proteins contain isoflavones, which may themselves have cardiovascular effects.22 Soy protein supplements that provide 118 mg isoflavone/d reduced SBP by 7.5 mm Hg in men and postmenopausal women.23 In a randomized controlled trial, 55 mg isoflavonoids/d, on a background of a usual diet, had no significant effect on ABP in men or women,24 consistent with the report that BP in women was unchanged by 80 mg isoflavones/d.25 In the present study, soy protein provided 23 mg isoflavones/d, and it is unlikely that the isoflavone content alone accounts for the findings.
Several epidemiological studies have shown an inverse association between dietary fiber and BP2,11–13 or the development of hypertension,12 and clinical trials suggest a small antihypertensive effect.14 Inconsistent results of fiber supplement studies may have arisen from variations in types of fibers and background diets. In stroke-prone spontaneously hypertensive rats, psyllium attenuated salt-accelerated hypertension,26 an effect that the authors suggested may be explained by increased fecal excretion of sodium bound to the soluble fiber. Soluble fiber seems more likely to induce cardiovascular effects.
Estimates from the International Study of Salt and Blood Pressure (INTERSALT) study1 suggest that an increase of 37 g dietary protein/d, less than the 66 g/d used in the present study, would lead to falls in population mean SBP of ≈3 mm Hg, sufficient to substantially reduce population morbidity and mortality rates from cardiovascular diseases.27 These findings therefore have important implications for the prevention and management of hypertension. In developing populations, a high prevalence of hypertension occurs against a background of a low intake of protein and fiber and high dietary salt consumption. However, even in industrialized countries, there are wide discrepancies in dietary fiber and protein intakes, in large part related to socioeconomic status. Our findings suggest that adequate intake of protein and fiber, particularly with fruits and vegetables as sources of soluble fiber, should be considered in recommendations of an optimal diet for reduction of cardiovascular risk in subjects with normal renal function. Epidemiological studies on the whole suggest that low-fat animal products may be equally effective protein sources for lowering BP, whereas fish, fruit, and vegetables have a variety of other constituents thought to be of benefit for cardiovascular health.
This work was supported by the National Health and Medical Research Council of Australia. We thank Home Chef, Perth, who carefully prepared meals with specified nutrient contents, and Fabien Dalais, Monash University, for isoflavone assays on the soy supplement.
- Received November 9, 2000.
- Revision received December 13, 2000.
- Accepted April 4, 2001.
Stamler J, Elliott P, Kesteloot H, Nichols R, Claeys G, Dyer A, Stamler R. Inverse relation of dietary protein markers with blood pressure. Circulation. 1996; 94: 1629–1634.
Stamler J, Caggiula A, Granditz GA. Relationships of dietary variables to blood pressure (BP): findings of the Multiple Risk Factor Intervention Trial. Circulation. 1992; 85: 867.
Liu K, Ruth K, Flack J, Jones-Webb R, Burke G, Savage PJ, Hulley SB. Blood pressure in young blacks and whites: relevance of obesity and lifestyle factors in determining differences: the CARDIA Study. Circulation. 1996; 93: 60–66.
Liu K, Ruth KJ, Shekelle RB, Stamler J. Macronutrients and long-term change in systolic blood pressure. Circulation. 1993; 87: 679.
Zhou B, Zhang X, Zhu A, Zhao L, Zhu S, Ruan L, Zhu L, Liang S. The relationship of dietary animal protein and electrolytes to blood pressure: a study on three Chinese populations. Int J Epidemiol. 1994; 23: 716–722.
Beilin LJ. Vegetarian and other complex diets, fats, fiber, and hypertension. Am J Clin Nutr. 1994; 59 (5 Suppl): 1130S–1135S.
Yamori Y, Horie R, Nara Y, Ikeda K, Kihara M, Ooshima A, Fukase M. Genetics of hypertensive diseases: experimental studies on pathogenesis, detection of predisposition and prevention. Adv Nephrol. 1981; 10: 51–74.
He J, Wu X, Gu D, Duan X, Whelton PK. Soybean protein supplementation and blood pressure: a randomized controlled clinical trial. Presented at the 40th Annual Conference on Cardiovascular Disease Epidemiology and Prevention, American Heart Association, La Jolla, Calif, 2000.
Ascherio A, Rimm EB, Giovanucci EL, Colditz GA, Rosner B, Willett WC, Sacks F, Stampfer MJ. A prospective study of nutritional factors and hypertension among US men. Circulation. 1992; 86: 1475–1484.
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: 338–365.
He J, Whelton PK, Klag MJ. Dietary fiber supplementation and BP reduction: a meta-analysis of controlled trials. Presented at the 16th Scientific Meeting of the International Society of Hypertension, Glasgow, UK, 1996.
McLennan W, Podger A,eds. National Nutrition Survey: Nutrient Intakes and Physical Measurements. Australia 1995. Canberra, Australia: Australian Bureau of Statistics; 1998: 85.
Kihara M, Fujikawa J, Ohtaka M, Mano M, Nara Y, Horie R, Tsunematsu T, Note S, Fukase M, Yamori Y. Interrelationships between blood pressure, sodium, potassium, serum cholesterol and protein intake in Japanese. Hypertension. 1984; 6: 736–742.
Fujita T, Sato Y. Hypotensive effect of taurine: possible involvement of the sympathetic nervous system and endogenous opiates. J Clin Invest. 1988; 82: 993–997.
Cernadas MR, Lopez-Farre A, Riesco A, Gallego MJ, Espinosa G, Diguini E, Hernado L, Casado S, Caramelo C. Renal and systemic effects of amino acids administered separately: comparison between L-arginine and non-nitric oxide donor amino acids. J Pharmacol Exp Ther. 1992; 263: 1023–1029.
Adlercreutz H. Western diet and western diseases: some hormonal and biochemical mechanisms and associations. Scand J Clin Lab Invest. 1990; 50 (Suppl 201): 3–23.
Teede HJ, Dalais FS, Kotsopoulos D, Liang YL, Davis SR, McGrath BP. Phytoestrogen dietary supplementation improves lipid profiles and blood pressure: a double blind, randomised, placebo-controlled study in men and postmenopausal women. Climacteric. 1999; 2 (Suppl 1): 127.
Nestel PJ, Yamshita T, Sashra T, Pomeroy S, Dart A, Komesaroff P, Owen A, Abbey M. Soy isoflavones improve systemic arterial compliance but not plasma lipids in menopausal and perimenopausal women. Arterioscler Thromb Vasc Biol. 1997; 17: 3392–3398.