Risk Factor Modification Center (DJAJ, CWCK, AM, ALJ, and LSAA) and the Department of Medicine, Division of Endocrinology and Metabolism (DJAJ), St Michael’s Hospital, Toronto; the Department of Nutritional Sciences, Faculty of Medicine, University of Toronto (DJAJ, CWCK, AM, and LSAA); the Department of Medicine, Children’s Hospital, Boston (DSL); the Physicians Committee for Responsible Medicine, Washington, DC (NDB); and the VA Medical Center, Graduate Center for Nutritional Sciences, University of Kentucky, Lexington (JWA).
Based on what is known of the components of plant-based diets and their effects from cohort studies, there is reason to believe that vegetarian diets would have advantages in the treatment of type 2 diabetes. At present there are few data on vegetarian diets in diabetes that do not in addition have weight loss or exercise components. Nevertheless, the use of whole-grain or traditionally processed cereals and legumes has been associated with improved glycemic control in both diabetic and insulin-resistant individuals. Long-term cohort studies have indicated that whole-grain consumption reduces the risk of both type 2 diabetes and cardiovascular disease. In addition, nuts (eg, almonds), viscous fibers (eg, fibers from oats and barley), soy proteins, and plant sterols, which may be part of the vegetarian diet, reduce serum lipids. In combination, these plant food components may have a very significant impact on cardiovascular disease, one of the major complications of diabetes. Furthermore, substituting soy or other vegetable proteins for animal protein may also decrease renal hyperfiltration, proteinuria, and renal acid load and in the long term reduce the risk of developing renal disease in type 2 diabetes. The vegetarian diet, therefore, contains a portfolio of natural products and food forms of benefit for both the carbohydrate and lipid abnormalities in diabetes. It is anticipated that their combined use in vegetarian diets will produce very significant metabolic advantages for the prevention and treatment of diabetes and its complications.
Type 2 diabetes
plant foods
vegan and vegetarian diets
glycemic index
nuts
vegetable proteins
soy
plant sterols
fiber
Previous SectionNext Section
DIABETES AND PLANT FOOD–BASED DIETS
There are few studies assessing the effects of a vegetarian diet in diabetes (1). Most of the studies involving plant foods, plant food components, or diets have been assessed for their ability to reduce blood lipids or other risk factors related to cardiovascular disease (CVD). Nevertheless, these attributes of diet are also very relevant to the treatment of diabetes because diabetes greatly increases the chance that an individual will suffer from CVD, possibly by 3–5-fold (2). Diabetes is a key factor in the predictive equations for CVD (3). It is therefore appropriate that dietary advice determined to be of use in the prevention and treatment of CVD should be considered as part of the advice for the prevention and treatment of diabetes. Thus, although an attempt will be made to discuss the effects of plant foods on glycemia, a large part of this discussion of the diabetic diet will deal with the role of plant foods in prevention of the major complications of diabetes, especially CVD.
Growth of interest in dietary fiber and its possible metabolic benefits in the prevention and treatment of chronic diseases, including diabetes, has been put forward as one of the reasons to include more plant foods in the diet (4). Notable are the early studies of Anderson using high-carbohydrate, high-fiber diets with initial carbohydrate contents of 70% and maintenance intakes of 60% (5). These diets resulted in improved glycemic control in type 2 diabetes, lower serum cholesterol levels, and no rise in serum triacylglycerol. Shortly after these studies, supportive data appeared from the Pritikin Institute (6, 7), where high-carbohydrate plant-based diets were emphasized together with exercise as part of the program for treatment of type 2 diabetic subjects. These studies demonstrated reductions in oral hypoglycemic agent use, together with improved blood glucose, cholesterol, and triacylglycerol levels, the latter 2 by 25% and 27%, respectively (6). These improvements tended to be maintained over the 2–3 y of follow-up (6). However, these studies were also confounded by exercise and weight loss, which has a major effect on all aspects of diabetes control (8). Confounding by weight loss has also existed in the majority of studies targeted more specifically at the use of vegetarian diets.
In assessing the overall impact of these very-low-fat (10% of energy) vegetarian diets, Barnard et al (9) reported their effects on 652 diabetic subjects. They showed that 39% of those treated with insulin (83 out of 212 subjects) could stop insulin and 71% of those on oral hypoglycemic agents (140 out of 197 subjects) could discontinue their use. At the same time, fasting blood glucose fell by 24% in those on diet alone at the start. In the whole group, serum cholesterol fell by over 20% and triacylglycerol by over 30%, while the fall in HDL was only half that seen in total cholesterol. Again, however, weight loss was significant: over 4 kg over the 26 d of the program. In the same year, Crane et al (10), using a vegan diet with daily exercise and weight loss (5 kg/25 d), showed similar metabolic advantages in type 2 diabetes together with complete relief of painful neuropathy in the legs within 2 wk in 81% (17 out of 21 subjects) of cases.
The only study to test the effect of a vegetarian (vegan) diet over a 12-wk period without a weight loss and exercise component showed a significantly greater weight loss on the vegan compared with the control diet of < 3 kg (1). Unexpectedly, despite weight loss, HDL cholesterol levels were reduced more on the vegan diet. On the positive side, fasting glucose levels were significantly lower on the vegan diet and, though not significant, 24-h microalbuminuria was reduced from baseline on the vegan diet but increased on the control diet. The lack of treatment difference in total cholesterol was surprising and contrasts markedly with the effect of plant-based diets on both normal and hypercholesterolemic subjects (11, 12).
In addition to diet trials, cohort studies have tended to support vegetarian diets or increased consumption of plant foods in the prevention of diabetes. In the Seventh-day Adventist Study cohort of 25 698 adults identified in 1960 and followed for 21 y, self-reported diabetes was lower in vegetarians than in nonvegetarians. The association with meat consumption was not confounded by body weight, other dietary factors, or exercise (13). A 20-y follow-up of a US cohort of 9665 adults aged 25–74 in which 1018 developed diabetes indicated that participants consuming 5 or more servings of fruits and vegetables daily compared with none had a relative risk for diabetes of 0.73 (CI: 0.54, 0.98) largely because of the beneficial effect on women, for whom the relative risk was 0.54 (CI: 0.36, 0.81) (14). Supporting data came from the Health Professionals study of 42 504 men aged 40–75 y where consumption of a “prudent” diet (fish, poultry, vegetables) versus a “Western” diet (red and processed meats and French fries) resulted in a relative risk of 0.84 (CI: 0.70, 1.00) in the 12-y follow-up; during those 12 y, 1321 new cases of diabetes were diagnosed (15). Finally, from the same study there is a recent report that supports the contention that processed meats may increase the incidence of diabetes, possibly through their nitrite content, with bacon showing the most significant trend (16).
There is therefore an urgent need for further assessments of the effects of plant-based diets in diabetes, especially in view of the benefits of such diets in nondiabetic subjects and the increasing recognition of the potential benefits of components of plant-based diets in both hyperlipidemia and diabetes. These components include dietary fiber, vegetable proteins, plant sterols, unsaturated vegetable oils, and slow-release carbohydrates (especially of cereal and legume origin).
Previous SectionNext Section
METABOLIC BENEFITS OF PLANT FOOD COMPONENTS
Increased intake of fruits and vegetables has been endorsed as public health policy for a number of reasons. Displacement of saturated fat and increased intake of fiber have been seen as general reasons for increasing fruit and vegetable consumption. Increased fiber intake may improve glycemic control in diabetes (17). Fruits and vegetables in the highly successful Dietary Approaches to Stop Hypertension Trial (DASH) diet, together with increased calcium intake and salt restriction in a diet low in total fat and very low in saturated fat, produced marked effects similar to those of the initial dose of single drug therapy in reducing blood pressure (18). Fruits and vegetables as sources of alkali also reduced urinary calcium loss (19). Recently, a lower proportion of animal protein in the diet was associated with reduced hip fracture risk (20). Substitution of vegetable for animal protein may preserve renal function (21, 22). Nevertheless, it is for the reduction of cardiovascular risk that plant-based diets have attracted the most interest. Large cohort studies of vegetarians have shown that they have a reduced risk of CVD (23). Vegan diets have been associated with very low LDL cholesterol levels (11). Short-term studies have confirmed large reductions in LDL cholesterol of 25–30% in healthy subjects on vegan diets based on fruits, leafy vegetables, and nuts (12). Cereal fiber consumption has consistently been associated with reduction in risk of both CVD and diabetes (24–26). Total fiber intake has been associated with reduced lipid risk factors for coronary artery disease (CAD) in young people (27), and increased whole-grain intake appears not only to reduce the incidence of CAD and diabetes but also to reduce blood glucose, fasting insulin, and evidence of insulin resistance in obese, insulin-resistant subjects (28).
Blood lipids
Related to the reduction in the high CAD mortality seen in diabetes is the need for tight control of established risk factors. Blood lipids are one of the best-established risk factors, and plant foods have clearly recognized effects in the control of blood lipids.
Vegetarian diets and body weight
Perhaps one of the major benefits that a vegetarian diet may have in the treatment of diabetes is its effect in increasing satiety, possibly related to amino acid composition or the sheer bulk of the diet (29). Evidence for this is the “confounding” effect of weight loss in interpreting the results of vegetarian diets on diabetic patients in the previous section. The weight loss, far from being “confounding,” may be one of the advantages of vegetarian diets. The paramount importance of body weight has been illustrated in recent studies where diet and lifestyle changes have prevented the development of diabetes in susceptible individuals (30).
Lipid-lowering plant food components
The US Food and Drug Administration (FDA) has approved health claims for cholesterol lowering for specific components of plant foods, many of which tend to be at higher intake levels in the diets of vegetarians and more particularly in the diets of vegans compared with the general population. These include viscous fibers from oats (β-glucans) (31) and psyllium (32), soy protein (33), and most recently plant sterols (34). Although these so-called functional food components have for the most part been tested in healthy or hyperlipidemic subjects, their use is very relevant to diabetes patients, who have a CAD risk 3–5 times greater than the nondiabetic population and for whom lipid-lowering medications are recommended even with a relatively small elevation in serum lipids—that is, LDL cholesterol above 3.35 mmol/L (> 130 mg/dL) (35).
The dietary components selected for FDA approval have all been well recognized for their cholesterol-lowering properties. Meta-analyses have suggested reductions in serum LDL cholesterol of 12.5% for 45 g soy protein/d (36); 6–7% for 9–10 g psyllium/d (37, 38), with smaller reductions for other viscous fibers (39); and 10% for 1–2 g plant sterol/d (40). These data come from studies often with background diets higher in saturated fat and cholesterol than current National Cholesterol Education Program (NCEP) recommendations (41–43). In studies of soy eaten in diets with lower intakes of saturated fat, in which subjects were in conformity with the NCEP dietary recommendations (44), smaller reductions in cholesterol of 4% for 52 g soy protein/d were seen (45). Similarly, plant sterols have also been shown to be less effective in diets lower in saturated fat and cholesterol (46).
On a theoretical basis, even if viscous fiber, soy protein, and plant sterols were additive only at the 5% level, in combination, a 15% or better reduction in cholesterol would be expected (47), which could be added to the 15–20% reduction reported with strict application of an NCEP Step 2 diet (48). Although again theoretical, this would result in a 30–35% reduction in serum cholesterol, equivalent to a therapeutic dose of the first-generation statins. Preliminary data in hyperlipidemic subjects have suggested that a 25–30% reduction in LDL cholesterol may be achieved by these means (49). Data are required in the treatment of diabetes, but there is reason to suppose that these plant food components may provide advantages for diabetic individuals that are similar to those seen in nondiabetic individuals (49).
Nuts
Nuts have traditionally been significant contributors to protein in the vegetarian diet. Nuts are also valuable sources of unsaturated fat, many (eg, almonds, hazelnuts) are excellent sources of monounsaturated fat, and walnuts contain n−3 fatty acids. Nuts’ protein and associated phytochemicals may provide some of the same advantages as seen with soy, and inclusion of nuts in the diet has been shown to lower serum cholesterol in nondiabetic subjects (50–63).
A number of studies have been published on almonds with reductions of 15% in LDL cholesterol per 100 g/d (1% lowering for 7 g almonds/d) (50), 12% for 100 g/d (1% lowering for 8 g almonds/d) (52), and 10% for 84 g/d (1% lowering for 8 g almonds/d) (53). From data on other published nut studies, 1% reductions in LDL cholesterol for walnuts, pecans, peanuts (an oil seed legume), macadamias, and pistachios would be achieved with daily intakes of 4, 11, 4, 10, and 4 g, respectively (51, 54–61).
Dose-response studies have also been carried out with almonds confirming the 1% reduction in LDL cholesterol for every 7–10-g increase in almond intake (62, 63). It is also possible that nuts such as almonds may be of value as sources of monounsaturated fat, which some reports suggest improves glycemic control when substituted for carbohydrate in the diets of type 2 diabetic subjects (64).
Dietary fiber: glycemic and blood lipid effects
Dietary fiber is another potentially useful component of a vegetarian diet. Early studies with viscous fibers focused on cholesterol reduction (65–68). However, it soon became apparent that viscosity also related to a flattening of the postprandial glycemia (69) when viscous fibers were added to test meals of both nondiabetic (70) and diabetic subjects (68). Long-term addition of the viscous legume fiber guar gum to breads resulted in reduced urinary glucose losses (68). Further studies indicated similar effects for diets enriched with whole legumes (dried beans) (71). In none of these studies had Hb A1c been measured. More recently, the viscous fiber psyllium has been used in the treatment of type 2 diabetes, with reductions in both blood lipids and day-long blood glucose but not Hb A1c (72). The latest study to demonstrate benefits in type 2 diabetes involved a mixed high-fiber diet from cereals, fruits, and vegetables without a specific focus on viscosity. This study did not find a change in Hb A1c over 6 wk but did find improvements in postprandial glycemia and blood lipids (17). However, when fiber-rich, low-glycemic-index foods were fed to type 1 diabetic patients for 24 wk, a fall in Hb A1c was seen in compliant subjects (73).
The picture is much less clear with respect to wheat bran. Cohort studies—from the studies by Morris et al on Whitehall civil servants (24) to the assessments of the Nurses’ Health Study, the Health Professionals Study, and the Iowa Women’s Health study (74–79)—have repeatedly shown that wheat bran appears protective for the development of both diabetes and CVD.
Nevertheless, wheat bran at modest intake levels failed to make an impact on CVD in the Diet and Reinfarction Trial study (80). Furthermore, by and large wheat bran has little effect on serum lipids in healthy or hyperlipidemic volunteers (68), and although beneficial effects have been reported on carbohydrate tolerance in healthy volunteers (81) and diabetic subjects (82), no reductions in glycated proteins have been reported.
More recently, the picture has been clarified with demonstration that simple additions of wheat bran (equal to 20 g dietary fiber) to the diet of type 2 diabetic subjects made no difference over 3 mo to Hb A1c concentrations (83). At the same time, studies of whole-grain cereals fed to insulin-resistant subjects appeared to improve insulin resistance parameters after a 6-wk feeding of whole-grain cereals as opposed to refined-cereal foods in the diet (84).
The issue of fiber and whole grains in the diet of the diabetic subject requires further study to determine whether these aspects of a vegetarian diet require emphasis in the treatment of diabetes>
Replies
Whole grains, legumes, and low-glycemic-index foods
Whole grains and legumes have also been staples in the diets of many vegetarians, both those living in traditional cultures and those who have adopted this lifestyle. These classes of foods processed in minimal or traditional ways may have a low glycemic index. Low-glycemic-index diets have attracted attention in terms of the prevention (85) and treatment of diabetes (86), with possibly beneficial effects on blood lipids and food intake regulation (87). Although the issue is being debated, the benefits of low-glycemic-index diets have also included higher HDL cholesterol levels (88), reduced CAD risk (89), and reduced risk of certain cancers associated with insulin resistance, namely, cancers of the colon (90) and breast (91). The low-glycemic-index components of a vegetarian diet may therefore be another potentially useful facet of foods in the prevention and treatment of type 2 diabetes.
The importance of slowing the rate of carbohydrate absorption has been illustrated recently by the results of the Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM) type 2 diabetes study, in which glucose-intolerant individuals had their risk of developing frank diabetes reduced by administration of acarbose, the α-glucosidase hydrolase inhibitor, which reduces the rate of carbohydrate digestion and absorption (92). There is a clear analogy here with the physiologic effect of viscous fibers and low-glycemic-index foods. These components of the vegetarian diet may therefore have the potential for similar effects of slowing carbohydrate absorption.
Nonlipid CAD risk factors: blood pressure and homocysteine
Plant food–based diets have the potential to reduce nonlipid risk factors for CVD. The success of the DASH diet—with its emphasis on increased fruit and vegetable consumption—in reducing blood pressure has already been mentioned (18). There is now a significant body of evidence to suggest that increased fruit and vegetable consumption in nondiabetic subjects protects from CVD (93–95). There has also been interest in the protein component of the diet and its effect on blood pressure. The DASH diet involves reduced meat consumption. Recent studies have focused on vegetable proteins; there is growing evidence that soy protein lowers blood pressure in both men and women (96–98). Other vegetable proteins remain to be tested. Indeed, studies involving the substitution of soy protein for animal protein in the diets of type 2 diabetic subjects have shown advantages in reducing not only lipids but serum homocysteine levels, possibly because of the lower sulfur amino acid content of the soy protein (99). Other plant proteins with lower sulfur amino acid content may have similar effects. At the same time, plant-based diets with high folate content will also contribute to the reduction in homocysteine levels as a possible risk factor for both CAD and stroke. This action, in addition to blood pressure control, provides another reason for their positive effect on cardiovascular health and therefore their potential importance in the diabetic diet.
Low-glycemic-index foods and glycemic control
The carbohydrate component of the diet is provided almost entirely by plant foods. When low-glycemic-index foods are selected, favorable effects may be seen on the blood lipid profile, as already discussed. The main interest in the glycemic index has been in relation to glycemic control, especially in type 2 diabetes. Selection of low-glycemic-index foods would be expected to allow increased consumption of carbohydrate to replace animal proteins and fat without compromising glycemic control in the postprandial period. Eleven studies where low-glycemic-index diets have been formally tested in type 2 as well as type 1 diabetes have now been reported in the literature. The conclusion of most (100, 101), though not all (102), analyses of these studies’ data has been that there is an overall benefit in diabetes of consuming low-glycemic-index foods. Such foods include traditionally processed starchy foods such as pasta; parboiled rice; cracked wheat; and whole grain (pumpernickel) breads; legumes (peas, beans, and lentils); temperate climate fruits; and nuts. It is of interest that the foods that tend to reduce postprandial glycemia are also those that reduce blood lipids.
PLANT PROTEINS, SOY, AND RENAL DISEASE
Plant-based diets may be superior to traditional animal protein diets for prevention and treatment of diabetic kidney disease (22). Intake of meals rich in animal protein increases renal blood flow and glomerular filtration rates (GFR) in this order: beef, chicken, and fish. However, intake of equivalent amounts of soy protein does not appear to alter these renal parameters (103, 104). Two studies reported beneficial effects on renal function of diabetic individuals with proteinuria with changes from animal protein to vegetable protein diets. However, both studies had limitations in their clinical design and outcome measures and must be considered preliminary (105, 106).
Based on the available evidence, the soy protein hypothesis was developed: substitution of soy protein for animal protein in individuals with diabetic nephropathy would decrease hyperfiltration and glomerular hypertension with resultant protection from diabetic nephropathy (22). A recent study in individuals with type 1 diabetes provides support for this hypothesis. When 13 diabetic subjects with hyperfiltration (glomerular filtration rates > 120 mL • min−1 • 1.73 m−2) incorporated 55 g soy protein daily into their diet and decreased animal protein intake, there was a significant reduction in GFR (107). Three of the subjects had microalbuminuria or albuminuria, and nonsignificant decreases in protein excretion were seen. Further studies are in progress to explore the effects of substitution of soy protein for animal protein in subjects with increased albuminuria.
IMPLICATIONS FOR FUTURE RESEARCH
Much work has been carried out on the effect of plant food components on blood lipids, and much still remains to be done, especially in studies where effective plant components are combined in the same diet to maximize the lipid-lowering outcome and provide an alternative to drug therapy (the “portfolio diet”) (47). This concept provides an opportunity for useful new plant components to be incorporated into the diet as they are identified (108). Apart from glycemic index testing, very few studies, by comparison, have been carried out in diabetes to assess the effects of plant foods on glycemic control in the long term. These are needed before the most useful combination of components can be assembled in the same diet to maximize the effect on glycemic control. Furthermore, there is a need for studies to be carried out in diabetes using vegetarian diets where weight loss is not a study objective. In the case of all but one study so far (1), weight loss has been an objective. However, this has not permitted an accurate assessment of the effect of the diet because of the potent effect of weight loss on glycemic control. Although weight loss in type 2 diabetes may be achieved in studies, it is not part of the natural history of this disease. It is therefore important to test diets in the same situation in which most type 2 diabetes subjects find themselves—at best, weight maintenance.
CONCLUSION
There have been no major studies in the absence of weight loss that have attempted to determine the potential advantages of a vegetarian or vegan diet in the treatment of diabetes. However, there are many facets or components of a plant-based diet that might confer benefits on glycemia and, more specifically, on blood lipids. Traditionally processed cereals and legumes have a low glycemic index, and whole-grain cereals appear to reduce the risk of developing diabetes. Serum lipid abnormalities are an increasing concern and reason for medication use in diabetes and may be improved by viscous fibers, soy, and other vegetable proteins and plant sterols. Nuts are also increasingly seen as useful in improving the blood lipid profile. There is therefore good reason to expect that all these factors combined as a dietary portfolio in the treatment of diabetes will have significant metabolic benefits when formally tested in the future.
Acknowledgments
The authors had the following conflicts of interest: research funding, Loblaws Brands Ltd, Toronto; research funding, travel support, and honoraria, Almond Board of California and the Solae Company, St Louis.
Footnotes
↵2 Presented at the Fourth International Congress on Vegetarian Nutrition, held in Loma Linda, CA, April 8–11, 2002. Published proceedings edited by Joan Sabaté and Sujatha Rajaram, Loma Linda University, Loma Linda, CA.
↵3 Supported by the British Medical Research Council, the British Diabetic Association, the Canadian Diabetes Association, the Natural Sciences and Engineering Research Council, the Canada Research Chairs Endowment of the Federal Government of Canada, Loblaws Brands, and the Almond Board of California. DJAJ is funded by the Federal Government of Canada as a Canada Research Chair in Nutrition and Metabolism. ALJ holds a doctoral research award from the Heart and Stroke Foundation of Canada.
↵4 Address reprint requests to DJA Jenkins, Clinical Nutrition & Risk Factor Modification Center, St Michael’s Hospital, 61 Queen Street East, Toronto, Ontario, Canada M5C 2T2. E-mail: cyril.kendall@utoronto.ca.
REFERENCES