点击显示 收起
1 From the University of Western Australia, School of Medicine and Pharmacology at Royal Peth Hospital and the Western Australian Institute for Medical Research, Perth, Western Australia, Australia
2 Supported by a Red Meat and Human Nutrition grant provided by Meat and Livestock Australia Limited. 3 Reprints not available. Address correspondence to J Hodgson, School of Medicine and Pharmacology, GPO Box X2213, Perth, WA 6001, Australia. E-mail: jonathan{at}cyllene.uwa.edu.au.
ABSTRACT
Background: Compared with carbohydrate intake, dietary intake of plant protein can lower blood pressure in humans, but the effects of animal protein intake on blood pressure have yet to be investigated.
Objective: We aimed to determine whether partial substitution of carbohydrate intake with animal protein intake from lean red meat changes blood pressure and other markers of cardiovascular disease risk in hypertensive persons.
Design: Hypertensive persons (n = 60) were recruited to an 8-wk parallel-design study. The participants were randomly assigned either to maintain their usual diet (control group) or to partially replace energy intake from carbohydrate-rich foods with protein from lean red meat (protein group). Measurements were performed at baseline and at the end of the intervention.
Results: Compared with the control group, the protein group had a significantly higher protein intake [
Key Words: Diet • protein • red meat • carbohydrates • blood pressure • cardiovascular disease
INTRODUCTION
The relation between protein intake and blood pressure (BP) has been extensively investigated in population studies (1-3). Overall, the results of these studies support the proposal that higher protein intake can lower BP. Estimated total (4, 5), plant (6-8), and animal (9-12) protein intakes have been inversely associated with BP. However, some studies have found that plant, but not animal, protein is associated with a reduced BP (7, 8). Total carbohydrate intake has not been consistently associated with either an increased or reduced BP (4, 8). Differences in the types and sources of carbohydrate ingested in the diet may differentially influence the relation between carbohydrate intake and BP (13).
Several randomized controlled trials have assessed the effect of dietary protein intake compared with carbohydrate intake on BP with the use of isoenergetic diets (14-17). We previously showed that supplementation with 66 g soy protein/d resulted in a 5.9 mm Hg reduction in the 24-h ambulatory systolic BP of hypertensive persons (14). Washburn et al (15) found a lower clinic BP in women who ingested a supplement of 20 g soy protein/d than in women who ingested a supplement of 20 g complex carbohydrates/d, and He et al (16) showed that a group who ingested a supplement of 26 g soy protein/d had significantly lower clinic systolic and diastolic BPs (by 4.3 and 2.8 mm Hg, respectively) than did a control group. Recently, Appel et al (17) showed that partial substitution of carbohydrate intake in a modified Dietary Approaches to Stop Hypertension (DASH) diet with 55 g protein/d from various sources resulted in an additional 1.4 mm Hg reduction in systolic BP compared with the control diet.
The BP changes observed in these trials (14-17) may have been caused by a BP-lowering effect of protein intake or a BP-raising effect of carbohydrate intake, or both. Whether soy protein differs from other sources of protein in its effects on BP, with the isoflavones in soy potentially mediating a vasodilator effect, remains uncertain (18). The results of randomized controlled trials that have compared soy protein intake to either dairy protein intake (19-22) or wheat protein intake (23) have not clarified this question. Any role that refined and high-glycemic-index carbohydrates may have in influencing BP remains uncertain.
Ultimately, any advice to increase protein intake to reduce BP within populations will be linked to increasing the intake of high-protein foods. Within most populations, an increase in protein intake is likely to result in a reduction in the intake of refined and high-glycemic-index carbohydrates, rather than in a reduction of fat intake. In Western populations, much of the total protein intake is derived from animal sources, including red meat. Therefore, our objective was to determine whether partial substitution of carbohydrate intake from refined and high-glycemic-index carbohydrate-rich foods with animal protein intake from lean red meat changed BP in hypertensive persons. We simultaneously investigated the effect of this substitution on other markers of cardiovascular disease risk.
SUBJECTS AND METHODS
Participants
Nonsmoking men and women aged >20 y were recruited from the general population with the use of newspaper advertisements. After measuring systolic and diastolic BPs in 2 screening visits 1 wk apart, volunteers were included if their systolic BP was between 130 and 160 mm Hg and diastolic BP was <100 mm Hg. Persons who were receiving drug therapy for hypertension were targeted with the recruitment strategies, but persons who were not receiving drug therapy for hypertension were also included if they met the BP criteria. Exclusion criteria included the following: use of >3 antihypertensive agents; a change in drug therapy within the previous 3 mo; diabetes; symptomatic heart disease; history of renal disease, liver disease, or gout; major psychiatric illness; other major illnesses, such as cancer; women who were currently pregnant or intended to become pregnant; and alcohol intake >200 g alcohol/wk for women and >300 g alcohol/wk for men. In addition, persons with no history of diabetes but with fasting plasma glucose concentrations 6.0 mmol/L were excluded. The study was approved by the Royal Perth Hospital Human Ethics Committee, and all participants gave written informed consent before inclusion in the study. All procedures followed were in accordance with institutional guidelines.
Study design
For the present parallel-designed study, the eligible participants were randomly assigned by using computer-generated random numbers to either maintain their usual diet (control group) or partially replace energy intake from carbohydrate with protein from lean red meat (protein group) for 8 wk. The participants maintained their usual diet and lifestyle during a baseline period of 2 wk, which preceded randomization. A dietitian counseled the participants in both groups every 2 wk throughout the 8-wk intervention to ensure the maintenance of usual lifestyle and weight (all participants), maintenance of usual diet (control group), or achievement of set dietary changes (protein group). In addition, directly after randomization, a dietitian provided the participants in the protein group with detailed instructions on how to achieve the required dietary changes. The participants in the protein group were supplied with lean red meat at 1 of 2 quantities, depending on their usual energy intake. The participants whose usual energy intake was <8500 kJ/d received 180 g raw weight/d (36 g protein/d) of lean red meat, and those whose energy intake was >8500 kJ/d received 250 g raw weight/d (50 g protein/d) lean red meat. The participants were instructed to consume the meat in place of carbohydrate-rich foods, including bread, pasta, rice, potatoes, and breakfast cereals. An estimated mean increase in animal protein intake of 43 g/d in the protein group was offset by a small decrease in plant protein intake as a result of reducing the intake of carbohydrate-rich foods. The objective in the protein group was to achieve 35–40 g/d higher protein intake (7-8% of total energy intake) than that of the control group. A 2-wk supply of lean red meat was provided to the participants in the protein group at each of the 4 visits during the 8-wk intervention.
Dietary and lifestyle assessments
A 3-d weighed food diary was completed by all participants at baseline, 1 wk before commencement of the 8-wk intervention, and during the last week of the intervention. The food diaries were completed on 2 weekdays and 1 weekend day. Food intake data were analyzed with FoodWorks Software (Xyris, Brisbane, Australia), which is based on the Australian Food Composition Database. A single measurement of weight was recorded at baseline, 4 wk, and 8 wk while the participants wore light clothing and no footwear. Food and alcohol intake, physical activity, health status, and medication use were monitored by interview at each of the 4 biweekly visits. Height was measured at baseline with a wall-mounted stadiometer.
Blood pressure screening
BP was measured at 2 screening visits 1 wk apart with a standard mercury sphygmomanometer. The participants rested for 10 min in a sitting position; BP was then measured 3 times at 2-min intervals. The mean of all 6 BP measurements was used as the screening value.
Ambulatory blood pressure
During the study, 24-h ambulatory BP was assessed at baseline and at the end of intervention (8 wk). The use of antihypertensive medication remained as prescribed and did not change throughout the study. The timing of antihypertensive medication use in relation to measurement of clinic BP was also unchanged. Staff who were not involved in the measurement of clinic BP carried out the group allocation. Staff who were involved in the measurement of clinic BP were blinded to group allocation.
Ambulatory BP was assessed by using a previously described method (24). Briefly, a trained nurse fitted a Spacelabs monitor (Spacelabs Medical Inc, Redmond, WA) onto the participants and explained its use to the participants. BP and heart rate (HR) were measured every 20 min during the day and every 30 min overnight. The participants were instructed to continue their usual daily activities and to avoid any vigorous exercise. They filled out an activity diary, which included waking and sleeping times. Measurements showing an error code or participants with a pulse pressure of <20 mm Hg were excluded from the analysis. BP traces that were missing >4 hourly means over the 24 h were excluded from the analysis.
Clinic blood pressure and arterial compliance
Clinic BP and arterial compliance were measured at baseline and at the end of intervention. The participants rested for 10 min in a supine position. BP and arterial compliance were then assessed on 3 occasions by measuring radial artery pulse waveforms, which were recorded for 30 s with the use of a calibrated tonometer (model CR-2000, Hypertension Diagnostics Inc-Pulse Wave; CR-2000 Research Cardiovascular Profiling System, Minneapolis, MN). The CR-2000 uses a piezoelectric-based, direct contact acoustical sensor placed over the radial artery to collect and amplify the BP waveform signal. The device was placed over the participant’s radial artery at its most superficial location in the wrist of the right hand. A wrist stabilizer was used to support the arm to ensure the best possible positioning of the tonometer and to minimize movement. Waveforms were calibrated by the oscillometric method with the use of systolic and diastolic BPs obtained from the left arm. The large artery elasticity index, or capacitive arterial compliance, and small artery elasticity index, or oscillatory-reflective arterial compliance, were calculated from a Windkessel-based model of the circulation. The mean of all 3 BPs and arterial elasticity measurements were calculated. The mean intraobserver CVs of repeated measures (1 wk apart) for systolic BP, diastolic BP, large artery elasticity index, and small artery elasticity index were 5.0%, 5.4%, 15%, and 28%, respectively. This method was previously described in depth (25).
Biochemical analysis
Plasma glucose concentrations were measured at screening. Plasma glucose and serum insulin, total cholesterol, HDL-cholesterol, and triacylglycerol concentrations were measured at baseline and at the end of the intervention. The Friedewald formula (converted to SI units) was used to calculate LDL-cholesterol concentrations. The homeostasis model assessment (HOMA) score was calculated with the following formula: (plasma glucose x serum insulin)/22.5. A 24-h urine sample was collected from each participant during baseline and the last week of the intervention for the measurement of urinary sodium, potassium, and creatinine excretion. The Department of Clinical Biochemistry at Royal Perth Hospital performed all routine biochemical measurements.
Statistics
Statistical analyses were performed with either SPSS version 11.5 software (Chicago, IL) or SAS version 8.2 software (SAS Institute, Cary, NC). Results are presented as means (±SDs) or means (95% CIs), and P < 0.05 was the level of significance. Log transformation was performed on variables that were not normally distributed. At baseline, the characteristics of the participants in the 2 groups were compared with the independent-samples t test. General linear models (analysis of covariance) were used to assess the baseline-adjusted differences in body weight, energy and nutrient intakes, excretion of urinary analytes, glucose and insulin concentrations, blood lipids, arterial compliance, and clinic BP and HR between the groups at the end of the intervention. General linear models (analysis of covariance) were also used to assess baseline-adjusted differences between the groups at the end of the intervention after adjustment for potential confounders. Potential confounding factors that were included as covariates were age, sex, change in weight, alcohol intake, and change in urinary sodium and potassium excretion. The baseline-adjusted differences in ambulatory BP and HR between the groups at the end of the intervention were analyzed with random effects models in SAS by using the PROC MIXED procedure. In the random-effects models, the participant was treated as the random effect, which accounts for correlated error structures, and treatment (control or protein group) was treated as the fixed effect. Between-group differences were also adjusted for potential confounding factors, including age, sex, and changes in weight, alcohol intake, and urinary sodium and potassium excretion, which were included as covariates.
RESULTS
Recruitment
The study was performed between April 2003 and October 2004. The trial profile, with the number of persons screened, excluded, and randomly assigned and the number of persons that withdrew from the study, is shown in Figure 1. A total of 460 persons were screened, and 389 who did not meet the entry criteria or were unwilling to increase their red meat intake were excluded. Seventy-one participants were randomly assigned, 36 to the control group and 35 to the intervention (protein) group. In the control group, 4 participants withdrew due to an inability to commit to the time requirements and 1 withdrew due to an unrelated medical problem. In the intervention group, 3 participants withdrew due to a high mean 24-h ambulatory BP at baseline, which required a change in medication, 2 withdrew due to unrelated medical problems, and 1 withdrew due to an inability to commit to the time requirements.
FIGURE 1.. Trial profile showing the numbers of participants at each stage of the trial.
Baseline characteristics
A total of 60 participants completed the study. Thirty-one participants (18 men and 13 women) were assigned to the control group; their mean (±SD) age and body mass index (in kg/m2) were 60.0 ± 9.9 y and 27.9 ± 4.0, respectively. Twenty-nine participants (20 men and 9 women) were assigned to the protein group; their mean age (±SD) and body mass index were 57.2 ± 7.3 y and 27.5 ± 3.1, respectively. The groups were also well matched in their use of medication. In the control group, 7 participants were taking a statin, 5 were taking aspirin, and 25 were taking antihypertensive medication (11 were taking an angiotensin-converting enzyme inhibitor, 8 were taking an angiotensin II receptor blocker, 9 were taking a calcium channel entry blocker, 2 were taking a ß-blocker, and 9 were taking a diuretic). In the protein group, 9 participants were taking a statin, 6 were taking aspirin, and 26 were taking antihypertensive medication (7 were taking an angiotensin-converting enzyme inhibitor, 15 were taking an angiotensin II receptor blocker, 4 were taking a calcium channel entry blocker, 6 were taking a ß-blocker, and 8 were taking a diuretic). No significant differences in any of these variables were observed between the groups.
Body weight, nutrient intake, and urinary analytes
Body weight, energy and nutrient intakes, and urinary analytes at baseline and the baseline-adjusted, end of intervention differences in these variables between the groups are presented in Table 1. No significant differences were observed between groups at baseline. Body weight at the end of the intervention was not significantly different between the groups. An analysis of the nutrient intakes confirmed compliance with the set dietary changes. Compared with the control group, there was a significantly higher protein intake and a matching significantly lower carbohydrate intake in the protein group. The lower carbohydrate intake was accounted for by a lower starch intake, with no significant difference in sugar intake. The mean (95% CI) protein intake was 36 g/d (26, 46 g/d) higher in the protein group than in the control group. Intake of iron and zinc were significantly higher in the protein group that in the control group (P < 0.01), and the protein group had a borderline significantly lower 24-h urinary sodium excretion than did the control group. Fiber, fat, and alcohol intakes at the end of the intervention did not differ significantly between the groups.
View this table:
TABLE 1. Body weight, energy and nutrient intakes, and urinary analytes of the participants at baseline and at the end of intervention1
Blood pressure
BPs and HRs at baseline and the baseline-adjusted, end of the intervention differences in BP and HR between the groups are presented in Table 2. No significant differences were observed between the groups at baseline. The 24-h, awake, and asleep ambulatory systolic BPs and clinic systolic BPs were significantly lower in the protein group than in the control group. These differences in systolic BPs were independent of age, sex, and changes in weight, alcohol intake, and urinary sodium or potassium excretion. The diastolic BPs and HRs were not significantly different between the groups.
View this table:
TABLE 2. Mean 24-h, awake, asleep, and clinic systolic and diastolic blood pressures (BPs) and heart rates of the participants at baseline and at the end of intervention1
Cardiovascular disease risk markers
Markers of cardiovascular disease risk at baseline, and the baseline-adjusted end of the intervention differences in these variables between the groups are presented in Table 3. The participants who were randomly assigned to the protein group had a significantly higher large artery elasticity index at baseline than did the participants in the control group. Other variables were not significantly different between the groups at baseline. Fasting plasma glucose concentrations were significantly higher in the protein group than in the control group at the end of the intervention. A similar trend was observed for the HOMA score.
View this table:
TABLE 3. Mean values of markers of cardiovascular disease risk of the participants at baseline and at the end of intervention1
DISCUSSION
Hypertension has been identified as the leading risk factor for mortality worldwide. It now affects approximately one-fourth of the world’s population and is projected to affect one-third of the world’s population within 25 y (26). Currently, two-thirds of all persons with hypertension live in developing countries, which generally have low protein and high carbohydrate intakes. Most of the projected rise in hypertension can be attributed to a projected increase in the prevalence of hypertension in developing countries, primarily due to changes in diet and lifestyle. A change in diet and lifestyle is also the most effective population-based approach for the prevention of hypertension. Results of population (1-12) and intervention (14-17) studies suggest that a higher protein intake can lower BP.
It was previously thought that animal protein specifically, and high total protein intakes more generally, would raise BP. This idea derived from studies showing that vegetarian diets, which are lower in total protein and lack meat protein, can lower BP (27, 28). However, similar BP reductions are found with increased intakes of fruit, vegetables, fish and nuts, and low-fat dairy products and decreased intakes of saturated fats and sugar (29, 30), and the absence of meat protein per se does not explain the reduction in BP observed with vegetarian diets (31, 32). Increasing the intake of nondigestible carbohydrate (ie, dietary fiber) can lower BP, particularly in hypertensive persons (33, 34). Some support exists for the proposal that fiber-rich, low-glycemic-index, whole-grain foods may contribute to the reduction in BP (35-37). The effects of refined or high-glycemic-index starch on BP in humans remains uncertain.
To investigate the effect of increased protein intake on BP, changes in protein intake need to be balanced by changes in carbohydrate intake, fat intake, or both. This is necessary to prevent confounding due to altered energy intake and changes in body weight. Low-fat diets are already encouraged, and it can be more difficult to substantially reduce fat intake than carbohydrate intake. Carbohydrate intake is the easiest nutrient to change while increasing protein intake to maintain an isoenergetic diet. The effect of increased animal protein intake, at the expense of carbohydrate intake, on BP has not been previously examined. In the present study, we investigated the effects on BP of a modest substitution of carbohydrate intake from carbohydrate-rich foods, including bread, pasta, rice, potatoes, and breakfast cereals, with protein intake from lean red meat. Compared with the control group, there was a significantly higher protein intake and a matching lower carbohydrate (starch) intake in the protein group. This resulted in a 4-mm Hg reduction in the 24-h ambulatory systolic BP, with no significant effect on diastolic BP. Arterial compliance and lipids were largely unchanged, but there was a significant increase in the fasting plasma glucose concentration in hypertensive persons.
In population studies, total and plant protein intakes are consistently and inversely associated with BP (1-3). Findings from randomized controlled trials that show lower BP with plant protein (soy) intake than with carbohydrate intake (14-16) are consistent with the population data. The relation of animal protein intake with BP in population studies is less consistent. Results of several studies suggest an inverse association (9-12), whereas others suggest either no association (6, 7) or a positive association (8). However, within certain populations the intake of animal protein and foods such as red meat that contribute to animal protein intake may be related to dietary patterns and lifestyle factors that can contribute to elevated BPs (38, 39). Thus, controlled trials are needed to specifically assess the effects of animal protein on BP.
In the present study, we found that an animal protein intake of 36 g/d, which replaced starch intake, resulted in a 4-mm Hg reduction in systolic BP. Previously, trials found that supplementation with 66 g/d (14), 20 g/d (15), and 26 g/d (16) of plant (soy) protein resulted in lower systolic BPs than the control groups by 6 mm Hg, 2 mm Hg, and 4 mm Hg, respectively. In addition, a recent study found that a 10% increase in energy intake from protein (55 g/d), which was derived from varied sources and replaced carbohydrate intake in a modified DASH diet, resulted in a further reduction in systolic BP of 1.4 mm Hg compared with the control diet (17). The BP-lowering effects of dietary protein, compared with carbohydrate, are more clearly evident in hypertensive persons (16, 17). In subgroup analyses of hypertensive persons, He et al (16) found decreases in systolic and diastolic BPs of 8 and 5 mm Hg, respectively, in persons who consumed a soy-protein supplement compared with those who consumed the carbohydrate control, and Appel et al (17) found decreases in systolic and diastolic BPs of 3.5 and 2.4 mm Hg, respectively, in persons who partially replaced carbohydrate intake with protein intake. Estimates from the International Study of Salt and Blood Pressure (INTERSALT Study) indicate that an increase of 37 g protein/d would lead a 3 mm Hg reduction in the population mean systolic BP (5). This estimate would be consistent with observed effects in intervention studies. A population reduction in systolic BP of this magnitude would be sufficient to substantially reduce the prevalence of cardiovascular disease within populations.
Several possible explanations exist for the observed differences in BP. Intake of amino acids, including arginine and taurine, may be important in determining BP. Red meat is a good source of both arginine and taurine. Arginine, which is a substrate for nitric oxide, can improve vasodilation and endothelial function (40) and lower BP (41). Taurine was inversely related to BP in population studies (10) and can lower BP in rats and in hypertensive humans (42). A nonspecific, dose-related diuretic effect of amino acids may also contribute to a reduction in BP (43). Another possible explanation is that the observed differences in BP relate to differences in carbohydrate intake. In the present study and in previous studies (14-17), the observed differences in BP could be equally influenced by the differences in carbohydrate and protein intakes. Consistent with this explanation, a recent study showed that partial substitution of carbohydrate intake with either protein or monounsaturated fat intake resulted in lower BPs of a similar magnitude (17). Although there is little evidence from human studies that intake of refined carbohydrates can raise BP, results of animal studies support this proposal (44, 45). Furthermore, our results also suggest a small decrease in sodium intake (25 mmol/d) in the protein group compared with the control group. This difference could be explained by a reduction in bread and breakfast cereal intake in the intervention group. Data from the INTERSALT Study (46) and the DASH-Sodium Trial (47) suggest that a change in sodium intake of this magnitude may account for 1 mm Hg difference in BP. However, the BP differences we observed were not significantly changed after adjustment for changes in either urinary sodium or potassium excretion.
Another interesting finding of our study was the significant increase in fasting plasma glucose concentrations and a similar trend for the HOMA score with increased protein intake. The physiologic importance of these changes is uncertain, given that mean plasma glucose concentrations and HOMA scores remained well within the normal range. However, these changes are in the opposite direction to those predicted given the dietary changes. A reduction in the intake of refined starch replaced with protein would be expected to reduce the plasma glucose response because of a small reduction in the glycemic load of the diet (48). Sustained changes in the glycemic load of the diet are suggested to provide long-term benefits in glucose and insulin metabolisms (49). A previous trial involving weight loss in overweight women investigated the effect of a high-protein compared with a high-carbohydrate diet on glycemic control and found no difference between the diets (50). We are not able to offer an explanation for our findings.
We showed that modest substitution of carbohydrate intake from carbohydrate-rich foods, including bread, pasta, rice, potatoes, and breakfast cereals, with protein intake from lean red meat results in a reduction in systolic BP in hypertensive persons. Within the context of other studies, these results suggest that modest replacement of carbohydrate-rich foods with protein-rich foods may lower BP in hypertensive persons. Given that the prevalence of hypertension worldwide is high and is increasing, and the importance of hypertension as a risk factor for cardiovascular disease, these findings may have important public health implications. The findings may be especially relevant to populations with relatively low total protein intakes.
ACKNOWLEDGMENTS
We thank Jackie Ritchie and Kitiya Dufall for their help in carrying out the study and Nella Giangiulio for dietetic assistance during the study. Meat and livestock Australia had no input into the design, conduct, or analysis of the study.
All authors were responsible for the conception, design, and conduct of the study and for data interpretation and writing of the manuscript. JMH and VB were responsible for statistical analyses of the data. The authors have no conflicts of interest.
REFERENCES