Supplementation of Conventional Therapy With the Novel Grain Salba (Salvia hispanica L.) Improves Major and Emerging Cardiovascular Risk Factors in Type 2 Diabetes: Results of a randomized controlled trial

The study was approved by the St. Michael’s Hospital Ethics Review Board. Eligible participants gave written informed consent. Eligibility criteria included the following: documented type 2 diabetes for at least 6 months’ duration without clinically manifest complications, age 18–75 years, nonpregnant, metabolically stable (A1C 6.0–8.5% and fasting plasma glucose 6.4–8.5 mmol/l), and not taking insulin, dietary fiber supplements, ALA supplements, or fish oil or consuming cold-water fish more than three times per week. Subjects were instructed not to change their lifestyle and level of physical activity during the study. Assuming a 35% attrition rate and a two-tailed α = 0.05 and 1-β = 80% to detect a 0.8% difference in A1C with an SD of 0.8%, sample size calculations indicated that 27 participants needed to be enrolled to yield 19 participants for final analysis.

The supplements were provided both in ground form and in specially formulated breads that were similar in appearance. Both the test and control intervention food were matched for energy and total dietary fiber. Salba was obtained from Salba Nutritional Solutions (Toronto, Canada), and the wheat bran control was from the American Association of Cereal Chemists (St. Paul, MN). The Salba and control breads were prepared in a bakery using a standardized recipe and methodology. The ground supplements (Salba and wheat bran control) were matched in appearance and were provided in opaque containers. Subjects were instructed to keep the bread and supplements in the refrigerator to minimize possible oxidation of the n-3 fats. Supplements were provided at a level of 15 g/1,000 kcal intake and were calculated according to subject’s individual daily energy requirements, as estimated by the Harrison-Benedict equation multiplied by a “very light” activity factor of 1.3 and verified with the initial 3-day dietary record. Control supplements contained wheat bran, which is considered to be lipid neutral and has little effect on glucose tolerance (3).

Salba or control supplements were added to conventional therapy. Dietitians instructed the participants to consume a diet that followed the Canadian Diabetes Association nutrition recommendations (4). The targeted macronutrient profile of carbohydrate:protein:fat was ∼55:15:30%, with an emphasis on low–glycemic index carbohydrate sources, 25–35 g total fiber, ≤10% sugars, and <10% saturated fatty acids. To monitor compliance, dietary data were analyzed at weeks 0 and 12 for energy and macro- and micronutrients. Three-day food records were analyzed by using Food Processor Nutrition Analysis software version 7.1 (ESHA Research, Salem, OR). Participants were also instructed to maintain their usual therapy (type and dose) of oral hypoglycemic, antihypertensive, or lipid-lowering medications. Salba or control supplements were added to this combination of a diet that followed Canadian Diabetes Association nutrition recommendations and usual therapy. By mimicking the macronutrient profile of the background Canadian Diabetes Association diets, the control supplements were designed to act as an extension of the diets.

The study used a randomized, placebo-controlled, single-blind, cross-over design with four distinct periods, in which subjects acted as their own control. Participants began with a run-in phase for at least 2 weeks to adjust to a healthy diet and to stabilize baseline parameters. Half of the participants were then randomized to either the Salba or control treatment for the first of two 12-week treatment arms. This period was followed by a 4- to 6-week washout phase to mitigate carryover effects. For the second treatment phase, participants were crossed-over to the alternate treatment. During each treatment phase, participants attended the clinic every 2 weeks to have anthropometric and clinical measurements taken, submit symptom diary and 3-day dietary records, receive new treatment foods, return unused bread and ground supplements, and have an interview with the dietitian and principal investigator. At weeks 0 and 12, fasting blood samples were collected.

Throughout the trial, an independent research assistant maintained the blinding of packages, labels, and diet randomization. Randomization was achieved using a computer-generated random-number table. Participants were excluded if during the course of the study there were any changes to their regular antihypertensive, lipid-lowering, or oral hypoglycemic medications; if they consumed <50% of the study supplements; or if they had a significant weight change over the course of the study, defined as >3 kg per 3 months.

There were three levels of outcome measures (efficacy, safety, and compliance) for which separate models were constructed. Efficacy measures included glycemic control (A1C, fasting plasma glucose, and fasting plasma insulin), blood pressure (office systolic blood pressure [SBP] and diastolic blood pressure [DBP]), lipids (total, LDL, and HDL cholesterol and triglycerides), and emerging risk factors for CVD (high-sensitivity C-reactive protein [hs-CRP], fibrinogen, von Willebrand factor [vWF], and factor VIII). Safety measures included markers of hepatic (aspartate aminotransferase and alanine aminotransferase), renal (serum urea and creatinine), and hemostatic (prothrombin time, partial thromboplastin time, and international normalized ratio) function. Finally, compliance measures included body weight change, quantity of returned bread and ground Salba supplements at follow-up visits, lifestyle adherence (dietary profiles and body weight), and plasma fatty acids.

Office blood pressure measurements were performed as described previously (5), with the cuff secured around the participant’s nondominant arm. Samples for A1C and hepatic, renal, and hemostatic function were analyzed directly by St. Michael’s Hospital Core laboratory. Plasma A1C was determined using ion-exchange high-performance liquid chromatography (Diamat HPLC; Bio-Rad Laboratories, Mississauga, Canada). Prothrombin time and partial thromboplastin time were determined using appropriate reagents and automatic clot timers. International normalized ratio was derived according to the following formula: (patient prothrombin time/mean normal prothrombin time)^{internal sensitivity index}. Serum alanine aminotransferase and aspartate aminotransferase were determined enzymatically. Serum creatinine and urea were determined by standard methods. Plasma glucose and insulin analyses were performed by the glucose oxidase and double-antibody radioimmunoassay methods, respectively. Plasma total lipids were analyzed by gas liquid chromatography (6).

Statistical analyses were performed using NCSS 2000 (NCSS Statistical Software, Kaysville, UT). Comparisons of within- and between-treatment differences in compliance, efficacy, and safety outcomes were assessed using repeated-measures general linear model ANCOVA adjusted for sex, age, and sequence. Results were expressed as means ± SD, and significance was set at P < 0.05. The Newman-Keuls procedure was used post hoc to adjust for multiple comparisons. Between-treatment results were adjusted for baseline values. Percent changes for each variable are based on the calculation for each individual subject’s percent change. All comparisons were adjusted for age, sex, weight, BMI, and use of cholesterol and hypotensive medications, aspirin, and oral hypoglycemic agents.

Forty-four subjects with type 2 diabetes were screened, and 27 eligible subjects were enrolled in the study. Exclusions during Salba treatment and control were equal. Reasons for dropouts during the protocol included medication changes (n = 3), refusal to continue (n = 3), and increased gastrointestinal side effects (n = 2). Final analysis included 20 patients with type 2 diabetes: 11 men and 9 women (means ± SD), aged 64 ± 8 years, BMI 28 ± 4 kg/m², and A1C 6.8 ± 0.9%. Medication treatment received concomitantly by the participants was single or combined oral hypoglycemic agents (10 subjects were on insulin secretagogues, 9 on metformin, and 1 on pioglitazone). Four subjects were on HMG-CoA (hydroxymethylglutaryl coenzyme A) reductase inhibitors (three on atorvastatin and one on simvastatin), three were on low-dose aspirin, and nine were on antihypertensive agents (four on ACE inhibitors, three on β-blockers, and two on calcium-channel blockers). Participants who had a change in their medications were excluded from the final analysis.

Within- and between-treatment differences in markers of compliance were assessed for both Salba and control treatment. Neither the proportion of supplements (cumulative bread and ground supplement) consumed over the 12 weeks between Salba and the control treatment (82 ± 24 vs. 85 ± 6%) nor body weight change from week 0 to 12 (0.25 ± 0.42 vs. 0.18 ± 0.37 kg) was significantly different between the Salba and control treatment. Further proof of compliance is supported by the results of the plasma total fatty acid analysis assessed at the end of each treatment. After 12 weeks of consuming Salba, the participants had approximately double the plasma level of ALA (2.5 ± 0.3 vs. 1.2 ± 0.1%; P < 0.05) and eicosapentaenoic acid (EPA) (0.18 ± 0.03 vs. 0.09 ± 0.02%; P = 0.006), when compared with the control treatment.

Analysis of the 3-day dietary records showed that participants maintained the diets they were instructed to follow. The addition of Salba to a healthy diet changed the macronutrient intakes between the Salba and control treatment with respect to the percent of calories from carbohydrate and fat. In the Salba group, the dietary profile was 45:21:34 (% calories from carbohydrate:protein:fat), whereas in the control diet group it was 54:19:27. Both total fat and PUFAs were significantly higher in the Salba treatment than control groups (P < 0.001) at the expense of carbohydrate, which was significantly lower in the Salba than control group (P < 0.001). The monounsaturated fatty acid content of the Salba phase was higher at P = 0.04. As expected, the n-3 PUFA intake was also significantly higher and reached 7.4 ± 4.3 g/day compared with 1.1 ± 0.8 g/day on the control treatment (P < 0.001). The mineral intakes were not significantly different except for magnesium, which was significantly higher (P = 0.03) with Salba (612 ± 149 mg) due to the high content of magnesium in Salba compared with the control diet (424 ± 167 mg).

Compared with baseline, SBP dropped on the Salba treatment by 6.3 ± 4.2 mmHg (P < 0.001) to an average of 123 ± 16 mmHg. DBP dropped by 3 ± 1.3 mmHg on the Salba treatment but did not reach statistical significance (Table 1). On the control treatment, the SBP increased by 7 ± 1.17 mmHg and the DBP increased by 3.14 ± 1.12 mmHg; as a result, SBP achieved statistical significance at P < 0.05, with a 20 ± 1% reduction and a nonsignificant 7% reduction for DBP between the Salba and the control treatment. Mean measures of glycemic control at the end of the Salba phase (A1C, fasting blood glucose, and fasting blood insulin) were not statistically different when compared with the end of the control phase (Table 1). No significant difference was observed comparing change from baseline in any parameter on the control phase. Fasting plasma blood glucose and insulin levels were ∼3% lower after treatment with Salba, but this did not reach significance. A1C was significantly reduced from baseline during the Salba treatment (P = 0.02) but not when compared with the control treatment.

Blood lipids were not different between the Salba and control treatments in any measured parameter (triglycerides and LDL, HDL, and total cholesterol) (Table 1). Baseline lipids were within the targets set by the 2006 Clinical Practice Guidelines (7), with the exception of LDL cholesterol, which was marginally elevated (primary target LDL cholesterol ≤2.0 mmol/l).

On the Salba treatment, fibrinogen levels significantly decreased over the 12 weeks, from 3.52 ± 0.76 to 3.37 ± 0.61 g/l (P = 0.03), but was not significantly different from the control treatment (Table 1). At the end of the Salba phase, vWF was significantly lower than in the control treatment (P = 0.03), while factor VIII was reduced by 23%, although this did not reach statistical significance (P < 0.06) when compared with the control treatment. Low-grade body inflammation measured by hs-CRP was not significantly different from baseline in the Salba phase (−7%, P = 0.24) but increased significantly (33%) during the control phase (P = 0.01) and was significantly higher at the end of the control phase compared with the end of the Salba phase (40%, P = 0.04).

Mean international normalized ratio, prothrombin time, and partial thromboplastin time did not change over the 12 weeks of the study period in either the Salba or the control treatments (data not shown). Mean values of blood urea nitrogen and creatinine, both measurements of kidney function, were not significantly different. Aspartate aminotransferase and alanine aminotransferase, measures of liver function, also did not change. Both baseline and postintervention values were within normal ranges as set by St. Michael’s Hospital Core Laboratory.

This study demonstrated that a 12-week dietary supplementation with the novel whole grain Salba (Salvia hispanica L.) was associated with attenuated SBP and the emerging risk factors (hs-CRP and vWF) in people with type 2 diabetes controlled on diet and oral hypoglycemic agents.

An increasing body of evidence from epidemiological observational studies suggests a strong inverse relationship between consumption of whole grain and the risk of diabetes and CVD (8). Many components of whole grains, including complex carbohydrates, vegetable protein, n-3 PUFAs, dietary fiber, minerals, antioxidants, and their combined effects could be responsible for this reduction. However, controlled interventional studies supporting health claims for whole grain are scarce. Therefore, our study was designed to determine whether the novel whole grain Salba could reduce CVD risk factors when added to usual care of individuals with well-controlled type 2 diabetes. To our knowledge, this is one of the first long-term randomized controlled trials to demonstrate a simultaneous reduction of major and emerging CVD risk factors in an intervention using whole grain in well-controlled diabetes.

All subjects in the study consumed >50% of supplements and were compliant with the treatment diet. Twelve-week dietary supplementation with 37 ± 4 g/day of Salba (7 g/day ALA) resulted in a twofold elevation in plasma phospholipid EPA, with no change of docosahexaenoic acid, compared with the control treatment. Similar changes were observed in another study (9) in which dietary supplementation of 9.5 g ALA per day during 3 months resulted in an 87% increase in plasma total lipid EPA compared with the control treatment. A parallel 1.7-g/day EPA plus docosahexaenoic acid supplementation elevated plasma phospholipid EPA similarly, by 79% (9).

This study demonstrated the safety of adding 37 g of Salba to a traditional healthy diet and conventional therapy. Although Salba has been used for many years in the diets of Aztecs in Mexico with no apparent side effects, there was a concern that a high amount of n-3 PUFAs could alter the eicosanoids produced, leading to an adverse effect on clotting factors. High doses of n-3 PUFAs have been implicated in altered bleeding and clotting time (10). However, coagulation, liver enzymes, and kidney function were not significantly affected by the addition of Salba in our study.

Despite the fact that our acute pilot study demonstrated that bread made with Salba reduced postprandial glucose and insulinemia (11), there were no long-term beneficial effects of Salba treatment on fasting blood glucose and insulin. Although a significant change was seen across Salba treatment in A1C, this was not significant when compared with control treatment, presumably due to the already optimal baseline glycemic control (A1C 6.8 ± 0.9%) achieved by subjects’ underlying diabetes therapy. Many factors in Salba may be responsible for this moderate glycemic-lowering effect. Salba contains 36% of its weight as fiber, of which only 4 g are soluble fiber but of a very high viscosity. The viscous mucilage formed when fiber from Salba is exposed to water and human digesta might be one of the main factors potentially thought to affect glycemic control. Factors in Salba that may act against good glycemic control include high content of n-3 PUFAs, which in some studies has been shown to increase A1C and fasting and postprandial glycemia in type 2 diabetes (12).

Dietary n-3 PUFAs have been associated with the reduction of serum triglycerides but at the expense of increasing LDL cholesterol levels. In the Lyon Diet Heart Study (13), the addition of n-3 PUFAs to a high-carbohydrate low-fat Mediterranean diet did not affect triglycerides or LDL, HDL, or total cholesterol, yet it still produced a 65% reduction in coronary heart disease mortality, indicating that changes in traditional risk factors such as blood lipids are not the sole cause of coronary heart disease. Our study provides evidence that 37 g of Salba (7 g ALA) has no detrimental effect on blood lipid profile in a group of individuals with type 2 diabetes, who are typically two to four times more susceptible to heart disease than the nondiabetic population.

Diet has long been implicated in reducing blood pressure. In addition, the micronutrients potassium (K), calcium (Ca), and magnesium (Mg) have also demonstrated blood pressure–lowering effects both individually and in combination (14). With the exception of higher quantity of n-3 PUFAs and magnesium in Salba, the treatments were not significantly different in known factors affecting blood pressure such as protein, fiber, sodium, potassium, and calcium. Although nine subjects were receiving antihypertensive medications, both SBP and DBP dropped on the Salba treatment, although DBP did not achieve statistical significance, indicating that the drug therapy did not obscure the response of blood pressure to n-3 treatment (15). It can be speculated that the mechanism of the reduction seen in SBP, and possibly DBP, could be due a conversion of Salba’s ALA to EPA, leading to the production of less vasoconstrictive prostaglandins through modification of the eicosanoid pathway. Most intervention studies show the benefits of fish oil and EPA on blood pressure (16).

Inflammation plays a major role in CVD, and measurement of inflammatory markers such as hs-CRP may be beneficial for risk stratification (17). Many of the previous prospective studies evaluating the effect of either fish oil or ALA on hs-CRP did not find significant changes in hs-CRP levels with n-3 PUFA consumption (18). However, in several recent randomized controlled trials (19,20) conducted in hypercholesterolemic subjects, consumption of ALA diets significantly decreased serum levels of CRP. Consistent with these findings, our results demonstrate that consumption of Salba, a grain naturally high in ALA, resulted in a significant decrease of hs-CRP, and the changes in serum ALA and EPA were inversely associated with changes in CRP. More recently, it has been suggested that increased levels of EPA alone may be cardioprotective. This is supported by the results of our study and a recent study that demonstrated that supplementation of purified EPA over 3 months to individuals with the metabolic syndrome resulted in reductions in small dense LDL particles and CRP levels (21).

Fibrinogen is negatively associated with the relative risk for CVD and cardiovascular risk factors (22). Although fibrinogen level was not different from the control treatment, it significantly decreased from baseline on the Salba by 0.15 g/l. A systematic review by Wendland et al. (23) indicated that ALA significantly affects fibrinogen concentrations, decreasing them by 0.17 mmol/l. This reduction would be expected to lead to a 6% decrease in coronary heart disease.

Although vWF is only weakly associated with the risk of CVD in the general population, it is more significant in high-risk populations such as people with diabetes (24). Salba treatment decreased the plasma level of this risk factor by 21%, but whether it could be attributed to n-3 PUFA or other components of the grain is unclear. There was a trend toward reduction of factor VIII (by 23%) in our study. Similarly, previous studies demonstrated either no change (25) or decrease (26) in activity of this factor with ALA or fish oil supplementation.

A simple whole-grain dietary intervention with Salba may potentially play an important role in the primary prevention of type 2 diabetes by increasing adherence to the recommended three servings per day, as well as being a therapeutic option that could be effective beyond currently recommended conventional therapies in improving major and emerging cardiovascular risk factors in type 2 diabetes. The amount of Salba consumed daily on this study was equivalent to the two portions of whole grain, as recommended by the Food and Drug Administration.

The study was designed to determine the safety and efficacy of consuming the novel whole grain Salba when added to the diet of individuals with type 2 diabetes while on conventional diabetes therapies. In such a design, no specific functional component from Salba treatment can clearly be implicated in any effect seen, but it is comparable with the cardioprotective effect seen in other whole grains. In addition to the numerous functional components in Salba, the supplementation of Salba produced changes in the dietary macronutrient profile between the two treatments. Therefore, the changes in risk factors observed could possibly be explained by the change in the carbohydrate and fat composition of the different diets (27), and future studies should compare Salba with control supplements that are matched for macronutrient composition. Furthermore, although subjects were instructed not to change their level of physical activity, the physical activity level during the study was not quantified, which might bias the results. Finally, our main objective was to assess the effect of Salba on A1C in subjects with type 2 diabetes, and the power of the study, limited to 20 participants, may not be sufficient for other than A1C parameters. Although the participants who completed the study were stable and well controlled, they may not represent typical candidates for adjunctive therapy. Both of these cavities render the conclusion that larger studies need to be conducted with the need for intent-to-treat analysis instead of per-protocol analysis.

In conclusion, the novel whole grain Salba, which can be simply added to conventional treatment for type 2 diabetes, improves major and emerging CVD risk factors while maintaining good glycemic and lipid control in well-controlled type 2 diabetes. The systolic blood pressure was reduced by 6.3 ± 4 mmHg, despite the fact that 9 of 20 participants were taking antihypertensive medications, which were not changed throughout the course of the study. As Salba contains a high amount of n-3 PUFAs, high dietary fiber, vegetable protein, and magnesium, the combination may have resulted in the pronounced blood pressure–lowering effects seen on the Salba treatment. Emerging CVD risk factors (hs-CRP and vWF) were attenuated beyond usual therapy (three subjects taking low-dose aspirin) on the Salba diet. This effect could be attributed to a high content of ALA in the Salba treatment and possibly the subsequent increase in EPA blood levels. No adverse effects of the conventional diet supplemented with the addition of Salba were documented, and liver function, kidney function, coagulation, and bleeding time were normal. In addition, there were no adverse effects on fasting plasma glucose, A1C, or LDL cholesterol, unlike previous studies with high doses of n-3 PUFAs in individuals with diabetes.

We thank Salba Nutritional Solutions, Inc., Toronto, Canada, for providing Salba grain.

We thank Lawrence Brown for invaluable logistic assistance in relation to study planning and ensuring the quality control of the test products. We also thank Trudy Brown for assistance in planning the study and whose diligence in supplement preparation made this study possible.