- Click here for explanation of classification scheme.

• The present investigation was designed to examine the potential effects of four weeks of daily consumption of 1,000mg of rebaudioside A on blood pressure and heart rate (HR) in healthy men and women with normal baseline blood pressure [below 120mmHg systolic blood pressure (SBP) and below 80mmHg diastolic blood pressure (DBP), as defined in the Seventh Report on Prevention, Detection, Evaluation and Treatment of High Blood Pressure]
• The study was conducted as part of a clinical program designed to address questions raised by the Joint FAO-WHO Expert Committee on Food Additives (JECFA) pertaining to the potential for pharmacological effects of steviol glycosides.

• Normal blood pressure (defined as below 120mmHg SBP and below 80mmHg DBP)
• Not taking antihypertensive medications.

• Body mass index of at least 35.0kg/m²
• History or clinical evidence of significant cardiovascular, gastrointestinal, renal, pulmonary, hepatic or biliary disease
• Fasting blood glucose of at least 126mg per dL or diagnosed with diabetes mellitus (type 1 or 2)
• Women who were pregnant, lactating, planning to be pregnant during the study or those of childbearing potential who were not using an approved method of contraception
• Subjects who were smokers could not have been planning to change smoking habits during the study period.

Recruitment

Information not given.

Design

• Randomized, double-blind placebo-controlled trial
• A two-week single-blind placebo lead-in period was utilized to evaluate subject compliance. To be eligible for randomization, subjects were required to be at least 80% compliant with placebo capsules during the lead-in.

Blinding Used

• Initial single-blind placebo lead-in period
• Treatment period was double-blinded.

Intervention

• Placebo: Microcrystalline cellulose
• Rebaudioside A: 97% purity (rebiana, the common name for rebaudioside A)
• Four 250-mg capsules per day of either placebo or rebaudioside A for four weeks.

Statistical Analysis

• The trial was designed to have at least 80% power to detect a 4.5-mmHg difference in resting seated SBP response for rebaudioside A vs. placebo, following four weeks of treatment, assuming a pooled standard deviation of 7.5mmHg
• SAS version 9.1.3 was used
• All tests of statistical significance were completed at an alpha level of 0.05, two-tailed
• Assumptions of normality of residuals were investigated for each response variable using the Shapiro-Wilk test. It was determined that the distribution could not be approximated by a normal curve (P-value for the Shapiro-Wilk test).
• Baseline and safety data were evaluated for all subjects enrolled in the study who received at least one dose of study product. Blood pressure and HR data were analyzed for all subjects who received at least one dose of double-blind study product and provided at least one post-randomization blood pressure data point. In cases where an intermediate data point was missing, the average of the two surrounding values was used in its place. The value at the previous non-baseline visit was carried forward to the subsequent visit; if not, subsequent data points were available.
• Changes from baseline in blood pressure and HR were analyzed by analysis of covariance using SAS PROC GLM. The initial model for each variable included terms for baseline (pre-treatment) value, treatment, site and treatment-by-site interaction. The model was reduced in a stepwise manner, eliminating terms with P-values over -0.10 until only terms with P<0.10 or treatment remained in the model.
• A similar procedure was followed for the blood pressure and heart rate changes from the pre-meal value during each meal test, except that the pre-meal value for that day was used as the covariate rather than the pre-treatment value
• Data for resting, seated and 24-hour blood pressure responses were also analyzed for pre-specified sub-groups split at the sex-specific median values for SBP in order to assess the effects of rebaudioside A in subjects with low-normal blood pressure.

Timing of Measurements

• Time of year not given
• Two-week placebo lead-in
• Four-week treatment period.

Dependent Variables

• Resting, seated blood pressure: Automated blood pressure device; measurements taken at zero, two, four, six, eight and 10 minutes. The first two measurements were discarded and the four remaining measurements were averaged. Each subject was assigned to a specific cuff and monitor for all measurements.
• Mean arterial pressure was calcualted as DBP + 1/3 (SBP-DBP)
• Supine and standing blood pressure and heart rate: Measured at Weeks Zero and Four; supine after lying for five minutes and standing after standing for two minutes; measured before and for two hours after consumption of a standard breakfast meal with two 250-mg capsules of rebaudioside A or placebo. Post-meal measurements were taken at 30, 60, 90 and 120 minutes and the values were averaged.
• Ambulatory blood pressure: At the completion of the meal tests at Weeks Zero and Four, each subject was fitted with an ambulatory blood pressure monitor which automatically collected data every 30 minutes for 24 hours
• Serum chemistry, hematology and urinalysis
• Diet records (unclear information)
• Physcial activity was assessed with the Stanford Seven-Day Physical Activity Questionnaire administered at Weeks Zero and Four.

Independent Variables

• Placebo: Microcrystalline cellulose
• Rebaudioside A: 97% purity; 1,000mg per day, which is more than 10 times the mean predicted intake of rebaudioside A for consumers of the sweetener and approximately four times the mean predicted daily intake for high-intake consumers
• Four 250-mg capsules per day: Two capsules were taken with first meal of the day and two with the evening meal.

Control Variables

None listed.

Initial N

Total of 100 subjects: 50 per group.

Attrition (Final N)

Two subjects (both in the placebo group) did not complete the study. One was lost to follow-up in Week Three and the other discontinued the study during the final week of the treatment period due to a serious adverse event (gastroenteritis).

Age

• Ranged from 18 years to 73 years of age
• Rebaudioside: 42.1±1.9
• Placebo: 41.0±2.3.

Ethnicity

• Non-Hispanic white: Rebaudioside, 80%; placebo, 76%
• African-American: Rebaudioside, 2%; placebo, 8%
• Hispanic: Rebaudioside, 14%; placebo, 10%
• Asian: Rebaudioside, 2%; placebo, 2%
• Other: Rebaudioside, 2%; plabeo, 4%.

Other Relevant Demographics

• Male: Rebaudioside, 24%; placebo, 18%
• Female: Rebaudioside, 76%; placebo, 82%.

Anthropometrics

BMI (kg/m²): Rebaudioside, 25.7±0.5; placebo, 24.3±0.5.

Location

Six clinical research sites in the United States.

Table 4: Supine and Standing Blood Pressure and Heart Rate Measurements During Meal Tests at Weeks Zero and Four.
[Only significant differences are shown; mean ±SEM] 

Parameters		Week Zero			Week Four
		Rebaudioside A (N=50)	Placebo (N=50)	P-Value	Rebaudioside A (N=50)	Placebo (N=48)	P-Value
Change Post-Meal	DBP (mmHg)	-1.3±0.6	-1.2±0.7	0.811	0.6±0.6	-1.9±0.8	0.045
	MAP (mmHg)	-0.9±0.6	-0.5±0.6	0.955	0.9±0.7	-1.5±0.8	0.043
Standing Pre-Meal	DBP (mmHg)	75.0±1.1	74.9±1.2	0.931	72.4±1.2	74.7±1.5	0.016
	MAP (mmHg)	88.4±1.1	88.0±1.1	0.773	85.9±1.1	87.9±1.1	0.036
	HR (bpm)	74.7±1.5	78.9±1.4	0.045	74.6±1.7	79.9±1.5	0.020
Change Post-Meal	DBP (mmHg)	-0.5±0.7	-0.5±0.9	0.907	1.4±0.8	-1.3±1.2	<0.001
	MAP (mmHg)	-0.1±0.7	0.2±0.8	0.867	1.3±0.7	-0.7±0.8	0.020

Other Findings

• Baseline and demographic characteristics were not significantly different between treatment groups
• Mean study product compliance was 99% in both groups and all subjects were at least 80% compliant. Compliance was assessed by capsule count and subject interview. Compliance percentage was calculated as 100x the number of doses consumed divided by the expected number of doses.
• Values for resting, seated SBP and the primary outcome variable indicated that there were no significant differences between the rebaudioside A and placebo groups at baseline or during the treatment period (data is shown in Table Two of the article; P-values ranged from 0.154 to 0.683)
• Values for resting, seated DBP and MAP were also not significantly different between rebaudioside A and placebo groups at baseline or in the changes from baseline to treatment
• In addition, there were no significant differences between the rebaudioside A and placebo groups for SBP, DBP or MAP, when the changes from baseline to each visit (Weeks One, Two, Three and Four) were considered individually (data was not shown)
• 24-hour ambulatory blood pressure monitoring of SBP and DBP revealed no significant differences between rebaudioside A and placebo treatment groups at Week Zero or in the changes from Week Zero to Week Four for morning, daytime, nighttime and overall continuous 24-hour readings (data is shown in Table Three of article). [Reviewers' note: The DBP change over 24 hours (-0.7±0.9 vs. 1.6±0.7, rebaudioside A vs. placebo, respectively) may have showed a trend at a p-value of 0.072.]
• At Week Zero, there were no significant differences between groups in pre-meal values or the changes from pre-meal to post-meal values for supine SBP, DBP, MAP and HR. This was also true for the pre-meal measurements of these parameters at Week Four.
• Pre-specified analyses were conducted on data for resting, seated and 24-hour blood pressures for sub-groups with baseline SBP split at the sex-specific median (less than and at least 108mmHg for females and less than and at least 117mmHg for males). In the lower baseline SBP sub-group, there was a small relative reduction from baseline to treatment for rebaudioside A vs. placebo observed for resting, seated MAP (-0.3 mmHg vs. 1.5mmHg; P=0.036). Otherwise, 24-hour blood pressure responses in the two SBP sub-groups did not differ significantly between rebaudioside A and placebo treatment groups.
• No changes in body weight, physical activity or dietary measures were observed
• 16 subjects (32%) reported adverse events in the Rebaudioside A Group, compared to 36% in the Placebo Group (P=0.833). None of the adverse events were believed by the investigators to be related to the study products. No signs or symptoms of hypotension were reported in association with rebaudioside A use.
• Laboratory test results (chemistry, hematology and urinalysis) indicated no clinically meaningful or statistically significant differences between the rebaudioside A and placebo groups.

• Rebaudioside A was well tolerated
• These results indicate that consumption of as much as 1,000mg per day of rebaudioside A produced no clinically important changes in blood pressure in healthy adults with normal and low-normal blood pressure.

Other:

not given

This reviewer has some concerns about this research study.

• No disclosure of funding sources
• Two authors disclosed receiving financial support from Cargill for consulting services
• Two authors are from either a food ingredient company or beverage company
• There were six clinical sites, but it is not indicated how many subjects were at each site. The statistical section indicated that the model included site and treatment by site interaction. However, no discussion of these values is given, so it is not clear if there was a site effect.
• Why was intent-to-treat analysis not used?
• Smokers were left in the study
• The age range was very great: 18 years to 73 years of age
• The authors state rebaudioside A was well tolerated. However, in both groups, more than 30% of subjects reported adverse events. This seems high, especially for a placebo. No statement is given as to what these adverse events were except for the one severe adverse event, gastroenteritis which caused a subject to withdraw from the study. This raises, in this reviewers opinion, even more concern regarding conflict of interest.