The role of nuts in the prevention and management of type 2 diabetes and metabolic syndrome

Transcription

1 The role of nuts in the prevention and management of type 2 diabetes and metabolic syndrome Systematic Review Completed by: Dr Kate Marsh BSc, MNutrDiet, PhD, Grad Cert Diab Edn & Mgt Advanced Accredited Practising Dietitian Credentialled Diabetes Educator On behalf of Nuts For Life October 203

2 Table of contents Executive Summary 3 Background 5 Description of the food- health relationship 6 Methodology for systematic review 9 Assessment of the body of evidence Conclusions 26 Appendix : Abbreviations 27 Appendix 2: Health Canada Quality Appraisal Tool 3 Appendix 3: Study Tables 34 Appendix 4: Quality Appraisal Tables 87 Quality Appraisals for Observational Studies 88 Quality Appraisals for Acute Feeding Studies 00 Quality Appraisals for Experimental Studies 09 References 42 October 203 Page 2

3 Executive Summary Nuts for Life is a nutrition communications initiative established for the Australian tree nut industry to provide information about the nutrition and health benefits of tree nuts. They aim to provide evidencebased and up-to-date information on the nutritional importance of tree nuts in the diet, including reports, research, newsletters, fact sheets and nutrient composition tables. In order to develop evidence-based resources for health professionals and consumers in relation to the benefits of nuts for diabetes and metabolic syndrome, and to determine whether any health claims can be made in regards to nuts and type 2 diabetes (T2DM) or metabolic syndrome (MetS), Nuts for Life require a systematic evaluation of currently available scientific research looking at the role of nuts in the prevention and management of these conditions. This systemic review was carried out using the process recommended in the FSANZ document Guidance on establishing food-health relationships for general level health claims (Version.; September 203), which is described in Schedule 6 of Standard.2.7 Nutrition, Health and Related Claims[]. The review aimed to answer the following questions:. Does eating nuts regularly reduce the risk of developing type 2 diabetes (T2DM) and Metabolic Syndrome (MetS)? 2. Does eating nuts improve measures of glycemic control in individuals with or without T2DM? 3. Does eating nuts regularly provide health benefits other than effects on glycemic control for those diagnosed with T2DM and/or MetS? The main findings of the review were: Large prospective cohort studies are strongly suggestive of an association between nut intake and reduced risk of developing T2DM and MetS, although two studies have failed to show a relationship. Nuts are low in available carbohydrate and have a low glycemic index and glycemic load. Acute, single-meal feeding studies have consistently demonstrated the benefits of nuts in blunting the post-prandial glycemic response when consumed with carbohydrate-rich meals. Longer-term intervention studies have demonstrated mixed findings with respect to the benefits of regular nut consumption on glycemic control. Of the 26 studies reviewed, 0 have demonstrated improvements in at least one measure of glycemic control (HbAc, fasting glucose, fasting insulin and/or HOMA [a measure of insulin resistance]) with diets containing nuts (28-75g/day) compared to nut-free control diets. A further two studies have demonstrated a reduction in risk of T2DM and MetS with a Mediterranean diet incorporating 30g/day of nuts. Of those which didn t find a benefit, several showed a trend towards improvements in the nut versus control groups but the changes didn t reach statistical significance. In addition, in many of these studies glycemic control was not the primary outcome measure and the studies may not have been designed to properly assess the effect of nuts on glycemic control. October 203 Page 3

4 Of the longer-term intervention studies in individuals with T2DM and MetS investigating the effect of nut consumption (ranging from 28 to 08g/day) on cardiovascular risk factors, 7 of 23 studies demonstrated a positive effect for at least one marker of cardiovascular risk including blood lipids, blood pressure, lipid oxidation, inflammatory markers and endothelial function, while a further 6 studies failed to show a benefit. However there is some inconsistency between the findings of different studies with respect to the various measures of cardiovascular risk. Longer-term studies in individuals with T2DM and MetS show no evidence of weight gain with the incorporation of 28-08g/day of nuts into the diet, either as part of an energy-controlled or ad libitum diet. Two studies showed a greater reduction in BMI, one a greater reduction in waist circumference, and another a greater reduction in body fat with diets containing nuts. A greater reversion of abdominal obesity in subjects with MetS was also found with a nut intervention compared to control diet. Taken together, the evidence is supportive of a likely benefit, and no harm, from the regular consumption of nuts for the prevention and management of T2DM and MetS, but further, longterm, well controlled studies are needed findings before a definitive conclusion can be made. October 203 Page 4

5 Background Diabetes is a condition where there is too much glucose in the bloodstream, and occurs when the body either doesn t produce insulin or when the insulin that is produced no longer works effectively. Insulin is the hormone produced by the pancreas which facilitates the movement of glucose from the bloodstream into the muscles and cells, where it is used for energy. There are two main types of diabetes. Type diabetes (TDM) is an autoimmune disorder which occurs when the body s immune system destroys the insulin-producing cells in the pancreas and is treated with insulin injections. TDM represents 0-5% of all cases of diabetes and is more commonly diagnosed in children and young adults, but can occur at any age. Much more common is type 2 diabetes (T2DM), which makes up around 85-90% of all cases of diabetes. T2DM results from insulin resistance, where the body s insulin is unable to work properly. Initially extra insulin is produced to overcome this resistance, but over time, the body is unable to produce enough insulin to maintain normal blood glucose levels. The risk is higher in individuals who are overweight and inactive, and diet and exercise are the first line of treatment, but most people with type 2 diabetes will need oral medications and possibly insulin over time. Diabetes is Australia s fastest growing chronic disease and currently affects around million Australians. According to the recently released results of the Australian Health Survey 20-2 around 5% of Australian adults aged 8 years or over have diabetes, with % being unaware of their condition[2]. A further 3% are estimated to have pre-diabetes. Diabetes was the sixth leading cause of death in 2009[3]. It was the underlying cause of death for 4,70 people (3.0% of all registered deaths), and contributed to almost 4,300 deaths in 2009 (0% of total deaths) as either an underlying or associated cause of death[3]. Diet plays an important role in the prevention of type 2 diabetes, in managing existing diabetes (all types of diabetes) and in preventing or reducing the progression of diabetes-related complications. Current recommendations include limiting saturated and trans fatty acids, choosing high fibre, wholegrain and low glycemic index (GI) carbohydrate foods, and monitoring carbohydrate intake to assist with glycemic control in those with existing diabetes[4]. Moderate weight loss achieved through caloric restriction and regular physical activity is also important for those who are overweight[4]. Metabolic syndrome (MetS) is a clustering of cardiovascular risk factors including elevated blood glucose levels, dyslipidaemia, hypertension and abdominal obesity[5]. Individuals with metabolic syndrome have a significantly greater risk of developing cardiovascular disease and type 2 diabetes[5]. The prevalence of the metabolic syndrome is increasing globally and while there are no official statistics on the number of Australian s with the conditions, studies have estimated a prevalence of around 5-30% depending on the diagnostic criteria used. Using the International Diabetes Federation criteria, the AusDiab Study reported a prevalence of 30.7%[6]. Each individual component of the metabolic syndrome confers an increased risk of cardiovascular disease but this increases when the syndrome itself is present and the more components of the metabolic syndrome an individual has, the higher their risk of cardiovascular mortality. Insulin resistance and central obesity are underlying features of both the MetS and T2DM. Lifestyle changes including dietary modification are recommended as the first line of treatment to target insulin resistance and obesity and reduce the risk of T2DM, MetS and associated CVD. October 203 Page 5

6 Description of the food- health relationship Description of the food For the purpose of this review, nuts were considered to be tree nuts (almonds, Brazil nuts, cashew nuts, chestnuts, hazelnuts, macadamia nuts, pecans, pine nuts, pistachios and walnuts) and peanuts. Description of the health effect It is now widely accepted that diet can not only help to achieve optimal health and development, but also plays an important role in reducing the risk of disease. There is also evidence that by modulating specific physiological targets, diet may have additional beneficial physiological and psychological effects beyond the widely accepted nutritional effects. This understanding has led to a move towards health claims on foods[7]. In , ILSI Europe co-ordinated The European Commission Concerted Action on Functional Food Science in Europe (FUFOSE), aimed at establishing a science-based approach to health claims on foods[7]. The aim was to critically assess the science base required to provide evidence that specific nutrients and food components positively affect target functions in the body, to examine the available science from a function-driven perspective rather than a product-driven one, and to reach consensus on targeted modifications of food and food constituents, and options for their application. FUFOSE recommended that claims for enhanced function" and "reduced risk of disease" are only justifiable when they are based on appropriate, validated markers of exposure, enhanced function or reduction of disease risk[7]. Building on the work of FUFOSE, the PASSCLAIM (Process for the Assessment of Scientific Support for Claims on Foods) project, also supported by ILSI Europe, aimed to identify criteria to assess the scientific support of claims[8]. The main objectives of this project were to produce a generic tool with principles for assessing the scientific support for health-related claims for foods and food components and to select common criteria for how markers should be identified, validated and used in well-designed studies to explore the links between diet and health. Criteria were developed for key health claim areas including diet-related cardiovascular disease, bone health, physical performance and fitness, body-weight regulation, insulin sensitivity and diabetes risk, diet-related cancer, mental state and performance and gut health and immunity. Expert groups reviewed the availability of indicators of health and disease states within their respective areas of expertise, identifying possible markers with potential for health claims, and discussing the limitations of existing markers. The PASSCLAIM report on body weight regulation, insulin sensitivity and diabetes risk was published in 2004 and discusses the link between insulin sensitivity and obesity, metabolic syndrome and diabetes[9]. The guidelines identify a number of potential claims for health effects based on modification of the target function for both metabolic syndrome and type 2 diabetes, including both improved function and reduce risk claims. These are outlined below. October 203 Page 6

7 Diabetes Improves blood glucose control Reduces plasma glucose levels Limits the postprandial glucose rise Improves glucose tolerance Reduces the risk of type 2 diabetes Helps to reduce the risk of long-term complications of hyperglycemia Metabolic Syndrome Improves insulin sensitivity Increases insulin-mediated glucose disposal Helps to reduce the risk for metabolic syndrome The literature search was therefore based on these markers i.e. reduced risk of T2DM or MetS, and improved glycemic control and/or insulin sensitivity. Description of the proposed food-health relationship Nuts have a unique nutrition profile, being a rich source of healthy fats, dietary fibre, plant protein, magnesium and antioxidant, vitamins and minerals (Figure ). It is the combination of these beneficial nutrients that is most likely responsible for their proposed benefits in the prevention and management of T2DM and MetS. Several review papers have outlined possible explanations and mechanisms by which nuts may benefit those with T2DM and MetS, and may reduce the risk of these conditions[0-3]. Based on their nutritional profile (discussed below) and current research findings it is proposed that the regular consumption of nuts (including tree nuts and peanuts) may: Reduce the risk of developing T2DM and MetS Reduce the postprandial glycemic response when consumed with a high carbohydrate meal Improve longer-term glycemic control in individuals with and without diabetes Improve markers of cardiovascular risk in individuals with T2DM and MetS Fatty acid profile Nuts have a low proportion of saturated, have no trans fats, and are rich sources of monounsaturated and polyunsaturated fats. While more research is needed to determine the optimal fatty acid composition of the diet to manage and prevent T2DM and MetS, evidence suggests that replacing saturated and trans fats with unsaturated fats improves insulin sensitivity and reduces T2DM risk[4-6]. Low available carbohydrate and glycemic load Nuts are relatively low in carbohydrate (apart from chestnuts and cashews). Cashews and chestnut meal have been glycemic index (GI) tested and both are reanked low GI. Nuts also have a low glycemic load (GL). Consuming nuts with a high carbohydrate meal has also been shown to lower the glycemic response of the meal which can help to reduce post-prandial hyperglycemia, now a well-recognized cardiovascular disease risk[7]. There is good evidence that diets with a low GI and/or GL can reduce the risk of type 2 diabetes and cardiovascular disease[8], and can help in the management of existing diabetes [9, 20]. October 203 Page 7

8 High in dietary fibre Nuts are a good source of dietary fibre, providing around 5-0g per 00g. Diets high in fibre have been shown to help in managing diabetes and metabolic syndrome and can reduce the risk of developing type 2 diabetes[2-24]. Good source of plant protein Apart from chestnuts, which are relatively low in protein, tree nuts are a good source of plant protein, ranging from 9 to 20g/00g. Studies have shown that higher intakes of animal protein are associated with an increased risk of diabetes, while vegetable protein shows no significant association with diabetes risk[25, 26]. In particular, intake of red meat is positively associated with diabetes risk[27, 28] and replacing red meat with nuts has been shown to reduce the risk of T2DM and GDM[28, 29]. There is also good evidence that vegetarian diets, which are more likely to feature nuts as a protein source, are associated with a significantly lower risk of type 2 diabetes[30-32]. Good source of magnesium Most nuts are a rich source of magnesium, with a 30g serve of some varieties of nuts providing around 25-30% of the RDI for adults. Magnesium intake has been inversely associated with type 2 diabetes risk[33, 34] and supplementing magnesium intake has been shown to improve fasting blood glucose and HDL levels in those with type 2 diabetes[35]. Figure : Nuts for Life 203 Nutrient composition of tree nuts ( October 203 Page 8

9 Methodology for systematic review. A search was conducted using the Medline database ( National Library of Medicine, Bethesda, MD) and Cochrane Reviews ( The Cochrane Collaboration) through to September 203. The search strategy is outlined below. Additional studies were obtained from reviewing the reference lists of other original studies and review papers. All studies included for evaluation were limited to those published in English. Search strategy: ((("diabetes mellitus"[mesh Terms] OR ("diabetes"[all Fields] AND "mellitus"[all Fields]) OR "diabetes mellitus"[all Fields]) OR (metabolic[all Fields] AND ("syndrome"[mesh Terms] OR "syndrome"[all Fields])) OR ("glucose"[mesh Terms] OR "glucose"[all Fields]) OR ("insulin"[mesh Terms] OR "insulin"[all Fields])) AND (("nuts"[mesh Terms] OR "nuts"[all Fields] OR "nut"[all Fields]) OR ("prunus"[mesh Terms] OR "prunus"[all Fields] OR "almond"[all Fields]) OR ("bertholletia"[mesh Terms] OR "bertholletia"[all Fields] OR ("brazil"[all Fields] AND "nut"[all Fields]) OR "brazil nut"[all Fields]) OR ("anacardium"[mesh Terms] OR "anacardium"[all Fields] OR "cashew"[all Fields]) OR chesnut[all Fields] OR ("corylus"[mesh Terms] OR "corylus"[all Fields] OR "hazelnut"[all Fields]) OR ("macadamia"[mesh Terms] OR "macadamia"[all Fields]) OR ("carya"[mesh Terms] OR "carya"[all Fields] OR "pecan"[all Fields]) OR ("arachis hypogaea"[mesh Terms] OR ("arachis"[all Fields] AND "hypogaea"[all Fields]) OR "arachis hypogaea"[all Fields] OR "peanut"[all Fields]) OR ("pistacia"[mesh Terms] OR "pistacia"[all Fields] OR "pistachio"[all Fields]) OR (pine[all Fields] AND ("nuts"[mesh Terms] OR "nuts"[all Fields] OR "nut"[all Fields])) OR ("juglans"[mesh Terms] OR "juglans"[all Fields] OR "walnut"[all Fields]))) AND English[Language] 2. The literature search retrieved 954 citations. Of these, 849 were excluded based on the title. The remaining 05 abstracts were reviewed and of these, 46 were included in the final review. An additional 4 studies were identified from reviewing the reference lists of other relevant review papers. Therefore 50 studies were available for review, including observational studies, 8 acute, single-meal feeding studies and 3 randomised intervention studies (24 parallel group and 9 crossover design). Inclusion criteria were human studies of edible nuts, published in English including: observational studies looking at the association between nut intake and T2DM or MetS risk, acute feeding studies investigating the effect of eating nuts on measures of postprandial glycemic control, randomised intervention studies investigating the effect of nuts (compared to a control diet) on glycemic control in all populations randomised intervention studies investigating the effect of nuts (compared to a control diet) on other health outcomes (eg cardiovascular risk factors) in those with diagnosed T2DM or MetS. Exclusion criteria included animal studies and non-randomised, uncontrolled trials. October 203 Page 9

10 3. Each study was reviewed and the key details included in a table format (see Appendix 3) including the following: a. The study reference b. The study design c. The study objectives d. The sample size and loss to follow-up e. The participant characteristics f. The method used to measure the food including the amount consumed g. Confounders measured h. The method used to measure the health effect i. The study results, including effect size and statistical significance j. Any adverse effects 4. The quality of each study was assessed using the Health Canada Quality Appraisal Tool, recommended in their Guidance document for preparing a submission for food health claims (Appendix 4)[36]. This tool appraises studies based on the following: For experimental studies: documented inclusion and exclusion criteria, appropriate randomisation, blinding of subjects and research personnel, reporting of attrition, documentation of the intervention/exposure including the amount of the food, documentation of the methodology used to measure the health effect, appropriate statistical analysis and consideration of potential confounders. Studies have a maximum rating of 5, with scores of 5-8 reflecting higher quality studies, and scores of 0-7 reflecting lower quality studies. For observational studies: documented inclusion and exclusion criteria, reporting of attrition, documentation of the methodology to measure the exposure, repeated assessment of exposure, documentation of the methodology used to measure the health effect. Studies have a maximum rating of 2, with scores of 7-2 reflecting higher quality studies and scores of 0-6 reflecting lower quality studies. A sample of these tools with explanatory notes is provided in Appendix Each category of studies (observational, acute-feeding and intervention) were summarised into a single table (Tables to 4) to allow for an assessment of the results of the studies as a whole. October 203 Page 0

11 Assessment of the body of evidence Evidence that nuts may play a beneficial role in the prevention and treatment of T2DM and MetS comes from observational studies, acute feeding studies and longer-term intervention studies looking at both measures of glycemic control, or metabolic risk factors in those diagnosed with T2DM or MetS. Observational studies The first observational study to show an association between nuts and diabetes risk was the Nurses Health Study published in 2002[37]. In a cohort of close to 84,000 women, both total nut intake and peanut butter intake were associated with a reduced risk of T2DM, with a 27% and 2% lower risk, respectively, for the highest versus lowest intakes[37]. A further analysis from the same cohort, looking only at walnuts was recently published showing similar findings[38]. In the Shanghai Women s Health Study, peanut intake was found to be inversely associated with T2DM risk, with a 20% lower risk for the highest versus lowest intakes, but the intake of other types of nuts wasn t assessed[39]. Finally the recent PREDIMED study found 3% reduction in risk of T2DM for those consuming more than 3 servings/week ( serve = 28g) of nuts compared to those consuming less than serve/week[40]. However two studies, one in older postmenopausal women and one in male health professionals have failed to show an association between nut intake and diabetes risk[4, 42]. Study limitations may explain these findings in the latter two studies dietary intake was only assessed at baseline, so changes in nut intake over time (2-9 years of follow-up) were not accounted for. Both also relied on self-report of diabetes diagnosis with no validation[4] or validation showing poor reliability (a tendency to over-report a diagnosis)[42]. Differences in study populations may explain the inconsistent findings. Nuts may also reduce the risk of gestational diabetes (GDM). In the Nurses Health Study there was a significant positive association between animal protein intake and GDM risk and a significant inverse association between vegetable protein intake and GDM risk. It was estimated that substituting one serving per day of red meat with nuts (28g serve) could reduce the risk of GDM by 5%[29]. Since GDM results from the same underlying pathology as T2DM, and predisposes to T2DM risk, this study is relevant to the current review. One cohort and three cross-sectional studies have demonstrated an inverse association between nut intake and metabolic syndrome[40, 43-45]. The SUN (Seguimiento Universidad de Navarra, University of Navarra Follow-up) study followed close to 0,000 healthy adults free of diabetes and CVD for 6 years and found that compared to those with the lowest intakes, those with the highest intakes of nuts had a 27% lower risk of MetS. A cross-sectional analysis from NHANES, where participants were divided into nut consumers or non-consumers found that those who reported eating tree nuts had a significantly lower risk of MetS compared to non-consumers (2.2% vs 26.6%)[45]. Consumers of all nuts and tree nuts also had a lower prevalence of a number of individual components of MetS. In a baseline analysis from the PREDIMED study, the prevalence of MetS was 26% lower in those consuming more than 3 servings/week of nuts ( serve = 28gg) compared to those consuming less than serve/week[40]. Finally, a baseline analysis from the Dose Responses to Exercise Training (DR s EXTRA) Study reported a significant reduction in risk of MetS in men with the highest versus lowest thirds of nut intake (OR 0.56), but no association was found for women[44]. Table outlines the key characteristics of these studies. October 203 Page

12 Acute feeding studies Seven acute feeding studies have been conducted investigating the short term effects of nut consumption on glycemic control (Table 2). [46-53]. Four studies have shown that consuming almonds (ranging from 28g to 90g ) can lower postprandial glucose and insulin responses when eaten alone or with higher carbohydrate foods [46-48, 5]. In one of these studies, which compared the effects of a 30g, 60g and 90g dose of almonds added to white bread, the GI of the meal was reduced in a dose-dependent manner [48]. A study of pistachios, comparing a 28g, 56g and 84g serve found a similar dose-dependent response when consumed with white bread, but little impact on post-prandial glycemia when eaten alone [50]. In a study of mixed nuts, comparing 30g, 60g and 90g doses, the glycemic response was lowered for all three doses compared to a white bread control in both normoglycemic and T2DM subjects, but no clear dose response was observed[49]. However there was a linear association between the dose of mixed nuts and 2 hour postprandial glucose response, which was significant in normoglycemic subjects and approached significance in subjects with T2DM. Two studies of peanuts have similarly shown reduced glucose and insulin responses when consumed with a meal, with varying responses depending on the form of the peanuts[52, 53]. Longer-term intervention studies glycemic control Twenty-seven papers reporting on the findings of twenty-six studies were found which have investigated various measures of glycemic control, in both healthy subjects and those with T2DM or MetS (Table 3). All but one study were of higher quality. Of these, 0 studies showed improvements in at least one measure of glycemic control (fasting glucose, fasting insulin, HOMA (a measure of insulin resistance) and/or HbAc levels) with diets containing nuts compared to nut-free control diets[46, 54-62]. A further two studies, both part of the PREDIMED trial, demonstrated a reduction in risk of T2DM and MetS with a Mediterranean diet incorporating 30g/day of mixed nuts[63, 64]. The prevalence of MetS was reduced by 3.7%, versus 2.0% for the control group after year, largely due to a greater reversion of MetS with the nut intervention[64]. After 4 years, the incidence of T2DM was reduced by 52% in the MedDiet + Nuts group (and similarly in the MedDiet supplemented with olive oil) compared to a lower fat control diet[63]. A further fifteen studies, however, failed to show a significant benefits of nut consumption on measures of glycemic control[65-79] and two of these studies surprisingly found an increase in glucose or insulin with the nut intervention[70, 72]. In several of these studies, positive changes in outcome measures in the direction of the nut intervention were seen, however these changes were not statistically different from the control diet. The inconsistencies in the findings may be explained by the considerable heterogeneity between studies with respect to subjects, study duration, type and dose of nuts and outcome measures and control diets. The studies have ranged from 4 days to 2 years with the majority being between 4 and 2 weeks in length. Most studies have used walnuts or almonds, but pistachios, hazelnuts and cashews have also been used, with doses ranging from 28g to 08g/day or 0-30% of energy. The study populations have included both healthy subjects and those with T2DM, MetS or hyperlipidaemia. Control diets have ranged from the subject s usual diet, to prudent/healthy diets, low fat or modified fat diets, Mediterranean diets and the NCEP Step 2 diet. Some have been ad libitum, with or without dietary education, while others have been tightly controlled with all meals provided to the study subjects. In all studies, nuts were generally provided free in pre-packaged portions, or incorporated into meals, to the intervention groups. October 203 Page 2

13 There were also a number of limitations among the studies including short duration and small sample sizes, being underpowered for the outcome in question, differing intensities of intervention between the groups, being unable to isolate the effects of nuts from other dietary changes made as part of the intervention, and changes in medications during the study. Longer-term intervention studies cardiovascular risk While insulin resistance and abnormal glucose tolerance are the key underlying problems in T2DM and MetS, it is well established that individuals with these conditions are more likely to have other cardiovascular risk factors including dyslipidaemia (particularly raised triglycerides and low HDL cholesterol) and hypertension. Managing these risk factors is considered equally as important as glycemic control for reducing the long-term complications of diabetes. Frequent nut consumption has been found to protect against coronary heart disease and reduce the risk of death from coronary heart disease[80]. There are a number of reasons nuts may protect against heart disease. Research has consistently shown that regular nut consumption can improve blood lipids. A recent pooled analysis of 25 trials found that consuming an average of 67g of nuts per day could reduce total cholesterol (TC) by 5.%, LDL by 7.4%, LDL:HDL ratio by 8.3% and TC:HDL ratio by 5.6%[8]. In those with raised triglycerides (TG), these were also reduced by 0.2%. There is also some evidence that nuts may reduce inflammation, improve endothelial function, reduce lipid oxidation and lower blood pressure[82-85]. To determine whether similar benefits are apparent for individuals with T2DM and MetS, 28 papers were found, reporting on the findings from 23 intervention studies, which have investigated the effects of eating nuts on various measures of cardiovascular risk in subjects with T2DM or MetS (Table 4). All but two studies were assessed as being of higher quality. As for the studies of glycemic control, the findings of these interventions have been mixed. Seventeen studies demonstrated positive findings for at least one marker of cardiovascular risk with the nut intervention (29-08g/day) compared to a control diet, including blood lipids, blood pressure, lipid oxidation, inflammatory markers and endothelial function[56-62, 66, 68, 70, 7, 77-79, 86-89]. However a further six studies, including one low quality study, failed to show such a benefit[46, 54, 69, 74-76] and one found a greater reduction in HDL and another an increase in CPR with the nut interventions[69, 72]. Again, variation in the study design and duration, subjects, type and dose of nuts, outcome measures and control diets are likely to explain the inconsistent findings. One observational study in women with type 2 diabetes found that eating at least 5 serves of nuts (including peanut butter) per week ( serve was 28g or tablespoon of peanut butter) reduced cardiovascular disease risk by 44% compared to rarely or never eating nuts[90]. Longer-term intervention studies weight management It is well established that overweight and obesity, particularly abdominal obesity, plays a major role in the development of T2DM and MetS. Weight loss of 5-0% through lifestyle changes can reduce the risk of developing T2DM by up to 60% in those at risk[9-94], and modest weight loss in those diagnosed with diabetes who are overweight, can improve glycemic control and reduce cardiovascular risk factors[4]. October 203 Page 3

14 Despite their high fat and energy intake, evidence does not suggest that regular nut consumption leads to weight gain, and in fact nuts may actually help with weight management. Intervention studies have found that nuts either don t contribute to weight gain or assist with weight loss as part of energy controlled diets[95-97]. This is supported by epidemiological studies which have found that regular nut eaters are less likely to be overweight than those who don t eat nuts [95, 97]. Of the studies of individuals with T2DM and MetS included in the current review, there was no evidence of weight gain with the incorporation of nuts (ranging from 28g to 08g/day) into the diet, either in energy controlled or ad libitum diets. Furthermore, two studies demonstrated a greater reduction in BMI[46, 56] and one a reduction in waist circumference[56] with the nut interventions. Another found a reduction in body fat percentage despite no significant changes in weight[59], while a study of Chinese subjects with MetS found that the addition of 30g/day of walnuts to a lifestyle counselling intervention resulted in a greater reversion of abdominal obesity[79]. October 203 Page 4

15 Table - Summary of observational studies looking at the relationship between nut intake and risk of T2DM and MetS Study Study Type Subjects Nut type & diet Jiang et al, Prospective 2002 cohort Kochar et al, 200. Pan et al, 203. Parker et al, 2003 Villegas et al, 2008 Bao et al, 203. Fernandez- Montero et al, 202. Ibarrola- Jurado et al, 203. Prospective cohort Prospective cohort Prospective cohort Prospective cohort Prospective cohort Prospective cohort Crosssectional N = 8388 Women aged years who were free of diabetes, CVD and cancer N = Male health professionals aged years. N = Women aged y in NHS and y in NHS II. N = Postmenopausal women who were free of diabetes, heart disease and cancer at baseline. N = Women with no history of type 2 DM, cancer, or CVD at baseline N = 5294 Women in the NHS II who reported at least one singleton pregnancy between 99 and 200. N = 9887 Healthy middle-aged adults without DM or CVD. N = 720 Older men and women at high risk of CVD All nuts; peanuts assessed separately Dose 28g serve never, -3 times/month, once/week, 2-4 times/week, 5-6 times/week, once/day, 2-3 times/day, 4-6/day and more than 6 times/day. All nuts 28g serve; rarely/never, 3 servings/, serving/week, 2 4 servings/week, 5 6 servings/week, daily, and 2 servings/day Walnuts 28g serve, never/almost never, 3 servings/month, serving/wk, and 2 servings/wk All nuts Peanut butter assessed separately Peanuts All nuts 28g serve nuts, TBsp peanut butter; less than /month, less than /week, -4 times/week and 5 times/week. Average daily intake of individual food items (g/d) was combined to compute the intake of peanuts. Divided into quintiles with median intakes of 0.-3.g/day. 28g serve never, -3 times/month, once/week, 2-4 times/week, 5-6 times/week, once/day, 2-3 times/day, 4-6/day and more than 6 times/day. All nuts Serving size not specified; categories: never/almost never, -3 times/month, once/week or at least two times/week. All nuts 28g serve; less than one serving/wk, 3 servings/wk, and more than 3 servings/wk Duration of Main findings Follow-Up 6 years Nut consumption was inversely associated with risk of type 2 diabetes. RR 0.73 for highest versus lowest intakes. Consumption of peanut butter was also inversely associated with type 2 diabetes. RR 0.79 for highest versus lowest intakes. Mean 9.2 No significant association between nut intake and T2DM years risk after adjusting for confounding factors 0 years Walnut consumption was associated with a lower risk of type 2 diabetes. HR 0.76 for highest versus lowest intakes. 2 years No significant association between nut intake and diabetes risk with multivariate adjustment Average 4.6 years Peanut consumption was inversely associated with risk of T2DM. OR 0.80 for highest versus lowest intakes. 0 years There was a significant positive association between animal protein intake and GDM risk and a significant inverse association between vegetable protein intake and GDM risk. Substituting one serving per day of total red meat with nuts was associated with a 5% lower risk of GDM. 6 years Nut consumption was inversely associated with risk of developing MetS. OR 0.73 for highest versus lowest intakes in multivariate adjusted model. n/a Compared to participants consuming < serving/wk of nuts, those consuming >3 servings/wk had lower adjusted OR for MetS (0.74) and diabetes (0.87). Direction of relationship Positive relationship No relationship Positive relationship No relationship Positive relationship Positive relationship Positive relationship Positive relationship Study Quality Higher quality (0/2) Higher quality (8/2) Higher quality (/2) Higher quality (0/2) Higher quality (0/2) Higher quality (9/2) Higher quality (9/2) Higher quality (0/2) Kouki et al, 20. O Neill et al, 20 Crosssectional Crosssectional N = 334 Men and women aged 57-78yrs taking part in population-based RCT. n = 3292 Men and women aged 9+ participating in the NHANES All nuts (grouped with legumes) All nuts Tree nuts Total intake estimated from food diary. Associations calculated per 0g/day serve of nuts & legumes Total intake estimated from 24 hour dietary recall. Subjects categorised as nut consumers ( 7.09g/day) or nonconsumers (< 7.09g/day) n/a n/a Nut consumption was inversely associated with risk of MetS in men. OR 0.56 for highest versus lowest third of intakes. No association in women. Tree nut consumers had a significantly lower prevalence of MetS. Nut consumers had a lower prevalence of overweight/ obesity, HT and low HDL. Tree nut consumers had a lower prevalence of overweight/obesity, abdominal obesity, HT, low HDL and high FBG. Positive relationship Positive relationship Higher quality (8/2) Higher quality (0/2) October 203 Page 5

16 Table 2: Summary of acute feeding studies looking at the effect of nuts on markers of glycemic control Study Study Type Subjects (male/female) Cohen et al, 20. Randomized n = 9 (NR) cross-over singlemeal 2 healthy feeding subjects and 7 study. subjects with T2DM. Jenkins et al, 2006 Randomized cross-over singlemeal feeding study. n = 5 Healthy subjects Nut type & diet Dose Duration of Follow-Up Almonds 28g Single meal feeding study Almonds 60g Single meal feeding study. Main findings One serving of almonds reduced postprandial glycemia by 30% in subjects with T2DM with no significant difference in those without diabetes. The maximum stimulated glucose concentration was reduced by 0% by almond ingestion in individuals with T2DM but not in healthy subjects. There were no significant differences in the time to maximum glucose concentration or GLP- concentrations at 30 minutes postmeal in either group. The GI and II of the almond meal were significantly less than the potato meal. The postprandial glucose peak height and insulin peak height for the almond meal were significantly less than for the potato and control meals. Direction of effect Positive effect in T2DM Positive effect Study Quality Higher (8/5) Higher (9/5) Josse et al, Kendall et al, 20. Kendall et al, 20. Randomised cross-over singlemeal feeding study. Randomised controlled crossover single-meal feeding study. Randomized cross-over singlemeal feeding study. n = 9 Healthy subjects n = 24 4 healthy subjects, 0 subject with T2DM n = 0 Healthy subjects Almonds 30g, 60g and 90g Single meal feeding study. Mixed nuts 30g. 60g. 90g Single meal feeding study. Pistachios Study : 28, 56 and 84g Study 2: 56g Single meal feeding studies The addition of almonds to white bread resulted in a progressive decrease in the glycemic index of the composite meal in a dose dependent manner. The relative glycemic response was significantly lower with all three doses of mixed nuts in comparison to the white bread control in both normoglycemic and T2DM subjects. However there was no clear dose response effect. The 2 h postprandial blood glucose response was also significantly lower There was a significant linear association between the dose of mixed nuts and relative glycemic response in normoglycemic subjects; this approached significance in subjects with T2DM. Pistachios have little impact on post-prandial glycemia when eaten alone. Pistachios reduce the glycemic response when eaten with high carbohydrate foods despite increasing the total carbohydrate intake of the meal. When consumed with white bread, pistachios reduce the glycemic response in a dose dependant manner Positive effect (dose related) Positive effect Positive effect (dose related) Higher (8/5) Higher (9/5) Higher (8/5) Mori et al, 20. Reis et al, 203. Randomized cross-over singlemeal feeding study. Randomised cross-over singlemeal feeding study. n = 4 Subjects with IGT n = 5 Subjects at high risk of T2DM 3 completed the study Almonds (as whole almonds, almond butter, defatted almond flour, & almond oil) Peanuts (as whole peanuts and peanut butter) 42.5g Single meal feeding study. 42.5g Single meal feeding study. Whole almonds significantly attenuated second meal and daylong glucose iauc., Almond butter and almond oil decreased blood glucose iauc in the morning. Daylong glucose iauc was attenuated with almond oil. Whole almond and almond oil produced a greater second meal insulin response The inclusion of peanuts reduced the GI of the breakfast meals from 60 8 (control) to 56 2 (PB)and 58 4 units (WP) First meal iauc responses for all peanut meals were significantly lower (p=0 03) than the second-meal responses. Positive effect Positive effect Higher (0/5) Higher (8/5) October 203 Page 6

17 Table 3: Summary of experimental studies looking at the effect of nuts on markers of glycemic control Author Study design Subjects (M/F) Study Length Control/ background diet Brennan et al, Crossover N = 5 (9/6) 4 days Isocaloric diet MetS placebo breakfast shake Casas-Agustench et al, 2009 Claesson et al, Parallel N = 50 (28/22) MetS Parallel N = 25 (/4) Healthy lean Cohen et al, 20. Parallel N = 3 (2/) T2DM Damavandi et al, 203 Parallel N = 50 (6/34) T2DM 2 weeks Healthy/prudent dietary advice Treatment Diet & Dose Isocaloric diet + breakfast shake containing 48g walnuts Healthy diet advice + 30g/day of raw nuts (5 g walnuts, 7.5 g almonds and 7.5 g hazelnuts). 2 weeks Usual diet + candy Usual diet + peanuts (roasted & salted, supplying 84kJ/kg body weight) 2 weeks Usual diet + 2 cheese sticks Usual diet + 28g almonds 8 weeks Usual diet Usual diet with 0% E from hazelnuts (mean 29g/day) Outcome measures Fasting glucose Fasting insulin Glucose AUC Insulin AUC HOMA-IR Fasting glucose 2 hour glucose Fasting insulin HOMA Fasting glucose Fasting insulin C-peptide Fasting glucose Fasting insulin HbAc FBG Results Fasting glucose and insulin levels, mean glucose AUC and insulin AUC, and insulin resistance (HOMA-IR) were not significantly different between the groups. Fasting insulin and HOMA significantly reduced in nut group compared to control. No differences in fasting or 2 hour glucose Insulin and c-peptide levels increased significantly in the candy but not the peanut group, but between-group differences not significant No differences in glucose. HbAc levels were reduced by 4% in the almond group vs a % increase in the control group (p=0.045). Fasting glucose was reduced by 5% in the almond group and increased slightly in the control group, while fasting insulin levels increased in both groups and to a greater extent in the control group. However there were no significant differences between the groups. There were no significant between- group or within- group changes for FBG. Direction of effect No effect Positive effect for FI and HOMA; no effect for glucose Positive effect for insulin; no effect for glucose Positive effect for HbAc; no effect for glucose and insulin No effect Study Quality Higher (0/5) Higher (9/5) Higher (8/5) Higher (0/5) Higher (2/5) Estruch et al, 2006 Esposito et al, Parallel N = 772 (339/433) High risk of CVD (including T2DM and MetS) Parallel N = 80(99/8) MetS Jenkins et al, 20 Parallel N = 7 (78/39) T2DM 3 months Low fat diet MedDiet + VOO 2 years Healthy/prudent diet 3 months Usual diet with muffin (with similar protein content as the nuts and no added sugar) MedDiet + Nuts 30g/day; (5g walnuts, 7.5g hazelnuts, 7.5g almonds) Mediterranean diet including 25-50g walnuts/day Usual diet with 75g/day mixed nuts OR 37.5g nuts and half dose muffin Fasting glucose Fasting insulin* HOMA* * only in subjects without diabetes Fasting glucose Fasting insulin HOMA-IR Fasting glucose HbAc Fasting glucose levels were reduced significantly in both MedDiet groups compared to the low fat control group In those without diabetes, FI and HOMA were reduced significantly in both MedDIet groups compared to the control group. The intervention group had significantly greater reductions in HOMA (-.; p < 0.00), FBG (- 6mmol/L; p < 0.00) and insulin (-3.5 U/ml; p = 0.0) HbAc levels fell significantly on the full nut dose but not in the other groups, and the change in HbAc in the full nut dose group was significantly greater than for the other groups.. No differences in fasting glucose between the groups. Positive effect Positive effect for FI and HOMA Positive effect for HbAc with full nut dose; no effect for glucose Higher (3/5) Higher (2/5) Higher (9/5) October 203 Page 7

18 Jenkins et al, Crossover N = 27 (5/2) Hyperlipidaemia, non-diabetic Katz et al, 202. Crossover N = 46 (8/28) MetS Li et al, 20 Crossover N = 20 (9/) T2DM Lovejoy et al, Crossover N = 30 (3/7) T2DM Ma et al, 200 Crossover N = 24 (0/4) T2DM Mitjavila et al, 203. Mukuddem- Petersen et al, Pieters et al, 2005 (Note: Same study as Mukuddem- Peterson) Parallel N = 0 (NR) MetS Parallel N = 64 (29/35) MetS Parallel N = 64 (29/35) MetS month x 3 NCEP Step 2 diet Full-dose almond (73+/-3g) AND Half-dose almond (37 +/- 2 g) plus half-dose muffin. 8 weeks x 2 Ad libitum diet Ad libitum diet + 56g walnuts/day 4 weeks x 2 NCEP Step 2 diet NCEP Step 2 diet + almonds to replace 20% of energy (~60g/day) 4 weeks x 4 High fat control (HFC, 37% fat, 0% from olive/canola oil) Low fat control (LFC, 25% fat, 0% from olive/canola oil). High fat high almond (HFHA = HFC with 0%E from almonds; 57-3g/day) Low fat high almond (LFHA = LFC with 0% E from almonds; 57-3g/day) 8 weeks x 2 Ad libitum diet Ad libitum diet + 56g walnuts/day I year Low fat diet Mediterranean diet with virgin olive oil (Med+VOO) 8 weeks Control diet 20% protein, 47% CHO, 33% fat 8 weeks Control diet 20% protein, 47% CHO, 33% fat Mediterranean diet with nuts (Med+Nuts) Walnut diet (20% E, 63-08g/day) Cashew diet (20% E, 63-08g/day) Walnut diet (20% E, 63-08g/day) Cashew diet (20% E, 63-08g/day) Fasting glucose Fasting insulin HOMA-IR Fasting glucose Fasting insulin HOMA-IR Fasting glucose Fasting insulin HOMA-IR Fasting glucose Fasting insulin 2 hour glucose 2 hour insulin Fasting glucose Fasting insulin HOMA-IR HbAc Fasting glucose Fasting glucose 2 hour glucose Fructosamine Fasting insulin HOMA-IR No significant between-treatment differences No effect Higher (8/5) No significant between-treatment differences No effect Higher (0/5) Fasting glucose, insulin and HOMA-IR were reduced significantly more on the almondenriched diet compared to the control diet. Positive effect Higher (9/5) No significant between-treatment differences No effect Higher (/5) Fasting glucose increased significantly on the walnut diet but not on the control diet, however there was no significant difference between the diets. Fasting insulin increased on the walnut diet and decreased on the control diet and while these changes were not significant from baseline there was a significant difference between the groups. There were no significant differences within or between groups for HbAC and HOMA-IR. No significant changes in glucose between the groups Fasting glucose increased significantly in the cashew group vs control (p=0.04). There was no significant difference in the fructosamine between groups before or after the intervention but frucotosamine levels increased significantly in the control group compared to baseline (p = 0.04). There were no significant differences between the group for the 2 hour glucose. Fasting insulin and HOMA increased significantly in all three groups from baseline to end, with no differences between the three groups. Negative effect for insulin; no effect for FPG, HOMA or HbAc No effect Negative effect on glucose for cashews only No effect for other measures No effect Higher (/5) Higher (3/5) Higher (9/5) Higher (8/5) Richmond et al, 203. Crossover N = 22(0/22) T2DM 3 weeks x 2 Standard diabetic diet + sunflower kernels (30g/d) - Standard diabetic diet + almonds (30g/d) - AD Fasting glucose Fasting insulin HbAc Fasting glucose and HbAc levels were reduced significantly in both groups, with no betweentreatment differences. No effect Higher (0/5) October 203 Page 8

19 Salas- Salvaro, 2008 Salas- Salvaro, 20 Parallel N =224 (566/658) High CVD Risk Parallel N = 48 (74/244) High CVD Risk Scott et al, 2003 Parallel N = 35 (NR) Met S & T2DM Tan & Mattes, 203 Parallel N = 37 (48/39) High risk of T2DM Tapsell et al, 2004 Parallel N = 58 (34/24) T2DM Tapsell et al, 2009 Parallel N = 50 (NR) T2DM Wang et al, 202 Parallel N = 86 (NR) MetS SKD year Low fat diet MedDiet + VOO 4 years Low fat diet MedDiet + VOO 42 weeks American Heart Association (AHA) diet (5% of calories from protein, 30% fat, and 5% MUFAs) MedDiet + Nuts 30g/day; (5g walnuts, 7.5g hazelnuts, 7.5g almonds) MedDiet + Nuts 30g/day; ( 5g walnuts, 7.5g hazelnuts, 7.5g almonds) HiPro-HiMono diet (25% prot, 40% fat, and 22% MUFA). Almonds provided in last 24 weeks (dose not provided) 4 weeks Ad libitum diet 4 intervention diets: Ad libitum diet + 43g almonds eaten with either breakfast, lunch, morning snack or afternoon snack. 6 months Low fat diet (<30% E) Modified low fat diet (using exchange lists inclusive of fatty acid considerations) 2 months Low fat diet (20% P, 30% F, 5% MUFA, 5% PUFA) Walnut diet (modified lowfat diet with 30g/day walnuts) Control diet + 30g/day walnuts (0% MUFA, 0% PUFA) 2 weeks AHA Step diet Control diet + 42g pistachios/day (RSG) OR Control diet + 70g pistachios/day (HSG) Prevalence of MetS Incidence of T2DM Fasting glucose Fasting glucose Fasting insulin Fasting insulin levels were reduced in both groups but this was only significant in the SKD, however there was no difference between the groups. The prevalence of MetS was reduced by 6.7% on Med Diet + VOO, 3.7% on Med Diet + Nuts, and 2% on the control diet after year. The prevalence was significantly lower in the MedDiet + Nuts versus control group. The MedDiet + Nuts group also had a significantly greater reversion of MetS compared to the control group. Following adjustment for confounding factors, diabetes incidence was reduced by 5% in the MedDiet + VOO and 52% in the MedDiet + Nuts, compared to the control group. Increased adherence to Med Diet was inversely associated with diabetes incidence in all study groups No difference in fasting glucose between treatment groups. The majority of study participants improved glycemic control. FBG was normalized in all 0 patients with impaired fasting glucose and was also normalized in 2 and reduced to impaired fasting glucose in 3 of 7 patients with diabetes. Positive effect Positive effect No effect Higher (3/5) Higher (4/5) Lower (6/5) No significant between- treatment differences No effect Higher (/5) HbAc No significant between- treatment differences No effect Higher (8/5) Fasting glucose Fasting insulin HbAc Fasting glucose Fasting insulin 2 hour glucose 2 hour insulin Fasting insulin levels were reduced significantly in the walnut group compared to the control group (p = 0.046). There were no significant between- group differences for fasting glucose or HbAc No significant between- treatment differences however FPG levels increased significantly in the control group and not the nut groups, and 2 hr post-load glucose was reduced significantly in the HSG and approached significance in the RSG, Positive effect for insulin No effect for FPG or HbAc No effect Higher (3/5) Higher (8/5) October 203 Page 9