Blood pressure

Fruits and vegetables of DASH provide ample quantities of K. Epidemiological, observational and interventional studies have proposed a protective effect of K on blood pressure and the prevention of hypertension⁽ Reference Houston ^{61 )}. Epidemiological studies show that blood pressure is lower in populations with higher fruit and vegetable consumption⁽ Reference Khaw and Rose ^{62 )}. In the INTERSALT study^{( 63 )}, a worldwide epidemiological study of large sample size from thirty-two countries, K intake (as measured by 24 h urinary K excretion) was an important determinant of population blood pressure, independent of Na^{( 63 )}. A meta-analysis of RCT showed that K supplementation lowered blood pressure in hypertensive patients and those who did not use antihypertensive medications⁽ Reference Binia, Jaeger and Hu ^{64 )}. The American Heart Association⁽ Reference Appel, Brands and Daniels ^{65 )} and American Society of Hypertension⁽ Reference Appel, Giles and Black ^{66 )} suggest 4·7 g dietary K per d. This amount is equal to the amount of K provided in the DASH diet⁽ Reference Appel, Moore and Obarzanek ^{13 )}.

In addition to K intake, the ratio of dietary Na to K is also important for blood pressure control. In high K consumptions, elevated dietary Na may not lead to increased blood pressure⁽ Reference Rodrigues, Baldo and Machado ^{67 )}. A large cross-sectional study estimated adjusted OR for hypertension as 1·40 (95 % CI 1·07, 1·83), 0·72 (95 % CI 0·53, 0·97) and 1·30 (95 % CI 1·05, 1·61), respectively, for the highest v. the lowest quartiles of intake of Na, K or Na:K ratio⁽ Reference Zhang, Cogswell and Gillespie ^{68 )}. A high dietary Na:K ratio was also positively associated with cardiovascular risk and mortality⁽ Reference Cook, Obarzanek and Cutler ^{69 )}. Genetics may also affect blood pressure response to dietary Na and K. A family-based study in China showed that a genetic polymorphism in the adiponectin gene may play a role in blood pressure change in response to dietary Na and K although mechanisms of this contribution have not been discovered⁽ Reference Chu, Wang and Ren ^{70 )}.

K causes vasodilation through hyperpolarisation of vascular smooth muscle cells through stimulation of Na⁺/K⁺ ATPase pumps, increased activity of the Na–Ca exchanger, and subsequently decreased cytosolic Ca (Fig. 3)⁽ Reference Haddy, Vanhoutte and Feletou ^{71 )}. Other possible mechanisms by which K affects blood pressure include increased Na excretion, reduced sensitivity to vasoconstrictive activity of noradrenaline and angiotensin II, modulation of baroreceptor sensitivity, and improved insulin sensitivity⁽ Reference Houston ^{61 ,} Reference Stone, Martyn and Weaver ^{72 )}. However, K also benefits the cardiovascular system and prevents mortality independent of its effect on blood pressure through mechanisms such as inhibition of vascular smooth muscle cell proliferation, thrombosis and macrophage adhesion to the vascular wall⁽ Reference He and MacGregor ^{73 )}.

Fig. 3. Mechanisms by which potassium prevents hypertension. VSMC, vascular smooth muscle cells. For a colour figure, see the online version of the paper.

Glucose tolerance

K has beneficial effects other than blood pressure reduction. A large cohort study with 9 years of in-person follow-up and 17 years of telephone follow-up indicated that serum K level is an independent predictor of type 2 diabetes mellitus incidence⁽ Reference Chatterjee, Yeh and Shafi ^{74 )}. Compared with individuals with normal to high serum K levels (5·0–5·5 mEq/l), those with serum K levels < 4·0 mEq/l, 4·0 to < 4·5 mEq/l, and 4·5 to < 5·0 mEq/l had an adjusted hazard ratio of incident diabetes mellitus of 1·64, 1·64 and 1·39, respectively. This type of association was also observed for dietary and urinary K in African Americans but not whites⁽ Reference Chatterjee, Colangelo and Yeh ^{75 )}. A meta-analysis of prospective cohort studies confirmed the inverse association between serum K level and diabetes risk but no association between dietary or urinary K and the risk of diabetes was observed⁽ Reference Peng, Zhong and Mi ^{76 )}. K may be involved in glucose-dependent insulin secretion from pancreatic β-cells. In the resting state, β-cell membranes are polarised by K efflux from ATP-sensitive K channels. Upon a glucose load, uptake of glucose by β-cells leads to production of ATP and inactivation of ATP-sensitive K channels. Subsequently, depolarisation of β-cell membranes leads to opening voltage-gated Ca²⁺ channels, Ca influx, a rise in intracellular Ca concentrations and insulin secretion⁽ Reference Ashcroft ^{77 )}. Besides the proposed insulin secretagogic effect of K, a protective effect of K against salt-induced insulin resistance has been reported through suppression of IL-17A, an inflammatory cytokine involved in the metabolic syndrome, especially in salt loading states⁽ Reference Wen, Wan and Zhou ^{78 )}. In addition to the direct effect of K on insulin secretion, an inverse association between serum K and fasting insulin levels was observed⁽ Reference Chatterjee, Yeh and Shafi ^{74 )}. This inverse association is proposed to be due to stimulation of K entry to cells by insulin, leading to reduced serum K levels⁽ Reference Sterns, Grieff and Bernstein ^{79 )}.

Blood pressure

Due to containing green leafy vegetables, nuts, legumes, seeds and whole grains, the DASH diet has a high Mg content. A large cross-sectional study showed that a higher DASH score is associated with higher urine concentrations of K and Mg⁽ Reference Taylor, Stampfer and Mount ^{80 )}. Many health benefits have been attributed to Mg, but Mg is famous for its beneficial effects on the cardiovascular system, particularly hypertension. Atherosclerosis, hypertension, arrhythmias, dyslipidaemia, impaired glucose tolerance, insulin resistance and increased risk of the metabolic syndrome are reported in Mg deficiency⁽ Reference Gröber, Schmidt and Kisters ^{81 )}. A meta-analysis of prospective cohort studies supported the inverse dose–response relationship between dietary Mg intake and the risk of hypertension⁽ Reference Han, Fang and Wei ^{82 )}. Meta-analyses of RCT have also reported benefits of Mg supplementation on blood pressure⁽ Reference Kass, Weekes and Carpenter ^{83 ,} Reference Zhang, Li and Del Gobbo ^{84 )}.

Various mechanisms have been suggested for the antihypertensive effect of Mg. Mg inhibits the renin–angiotensin–aldosterone system, by counteracting angiotensin II and inhibiting aldosterone production. This inhibition may be exerted through activation of Na⁺/K⁺ ATPase and decreasing intracellular Na and Ca levels⁽ Reference Schutten, Joosten and de Borst ^{85 )}. Mg also acts as a Ca antagonist by raising the excitation threshold of voltage-gated Ca²⁺ channels, thus decreasing vascular smooth muscle cell contractility⁽ Reference Kostov and Halacheva ^{86 ,} Reference Sontia and Touyz ^{87 )}. Moreover, the production of vasoactive agents, such as endothelin-1, in vascular cell membranes is decreased in the presence of Mg. Furthermore, along with ATP, Mg mitigates catecholamine release from the adrenal gland in response to Ca. Mg may also reduce vascular stiffness⁽ Reference Joris, Plat and Bakker ^{88 )}, probably through regulating the synthesis of structural molecules in the vascular extracellular matrix⁽ Reference Kostov and Halacheva ^{86 )}. A part of the vasodilatory potential of Mg may be exerted through inhibiting systemic inflammation⁽ Reference Song, Li and van Dam ^{89 )}. Interestingly, in salt-induced hypertension a decrease in intracellular Mg along with accumulation of intracellular Ca is observed⁽ Reference Barbagallo, Dominguez and Resnick ^{90 )}.

Glucose tolerance

Prospective cohort studies and meta-analyses have also shown Mg to be advantageous in other metabolic disorders. For instance, in a prospective cohort study with 15 years follow-up, individuals in the highest quartile of Mg had lower risk (hazard ratio 0·69) for the development of the metabolic syndrome compared with those in the lowest quartile⁽ Reference He, Liu and Daviglus ^{91 )}. Also, a longitudinal study with 15·6 years of follow-up on a non-diabetic Japanese population showed decreased incidence of type 2 diabetes with a hazard ratio of 0·63 in the highest quartile of Mg intake compared with the lowest⁽ Reference Hata, Doi and Ninomiya ^{92 )}. Moreover, a meta-analysis of cohort studies indicated a significant inverse association between Mg intake and diabetes risk⁽ Reference Schulze, Schulz and Heidemann ^{93 )}. In another meta-analysis of prospective cohort studies an inverse correlation between plasma Mg levels and incidence of hypertension, CHD and type 2 diabetes was observed⁽ Reference Wu, Xun and Tang ^{94 )}.

Mg is also involved in insulin function. Mg is necessary for all reactions in which high-energy phosphate bonds are transferred. It functions as a cofactor for phosphorylation of tyrosine kinase at the insulin receptor⁽ Reference Kostov and Halacheva ^{86 )}. Therefore, decreased Mg may impair insulin signalling. Mg may also improve insulin secretion as reported by cross-sectional⁽ Reference Rodríguez-Morán and Guerrero-Romero ^{95 )} and interventional⁽ Reference Guerrero-Romero and Rodríguez-Morán ^{96 )} studies. Individuals with the metabolic syndrome generally have inadequate dietary Mg intake, and thus Mg supplementation may help in correction of their metabolic disturbances⁽ Reference Wang, Persuitte and Olendzki ^{97 ,} Reference Morais, Severo and Santos ^{98 )}. Meta-analyses of controlled trials have supported the effectiveness of Mg supplementation in reducing fasting glucose⁽ Reference Song, He and Levitan ^{99 ,} Reference Veronese, Watutantrige-Fernando and Luchini ^{100 )} and raising HDL-cholesterol levels in patients with type 2 diabetes⁽ Reference Song, He and Levitan ^{99 )}. Another meta-analysis revealed that Mg supplementation for more than 4 months improved insulin resistance and fasting glucose in both diabetic and non-diabetic individuals⁽ Reference Simental-Mendía, Sahebkar and Rodríguez-Morán ^{101 )}.

It is worth noting that it is not only Mg which affects glucose tolerance but hyperglycaemia, even in transient form as occurs in non-diabetic subjects following consumption of meals, also decreases intracellular Mg and increases intracellular Ca, indicating that Mg depletion can be both a cause and consequence of hyperglycaemia⁽ Reference Barbagallo, Dominguez and Resnick ^{90 )}. In addition, insulin resistance may impair Mg reabsorption in the kidney resulting in urinary Mg excretion⁽ Reference Gommers, Hoenderop and Bindels ^{102 )}.

Weight

One of the important ingredients of DASH is Ca. In the DASH diet, Ca is mainly supplied by dairy products and to a less extent by green leafy vegetables, soya, nuts, particularly almonds, and fish when the bones are eaten. Ca and dairy products protect against the metabolic syndrome through a variety of mechanisms. Reduced adiposity is one of the mechanisms of Ca against the metabolic syndrome. A meta-analysis of cross-sectional studies showed the inverse association between dairy product consumption and decreased risk of obesity⁽ Reference Wang, Wu and Zhang ^{103 )}. However, meta-analyses of randomised clinical trials did not show a beneficial effect of dairy products on the weight of overweight/obese adults but advantages were reported when dairy products were consumed in combination with energy-restricted diets⁽ Reference Chen, Pan and Malik ^{104 –} Reference Abargouei, Janghorbani and Salehi-Marzijarani ^{106 )}. The evidence suggests that individuals with a low Ca intake may benefit from dairy products for decreasing waist circumference and sagittal abdominal diameter⁽ Reference Wennersberg, Smedman and Turpeinen ^{107 )}. The reductions in weight and waist circumference may eventually lead to lower blood pressure and control of hypertension⁽ Reference Park, Lee and Kim ^{108 )}.

There are mechanisms by which Ca may engage in weight control. Dietary Ca may decrease fat absorption through formation of insoluble fatty acid soaps in the intestine⁽ Reference Christensen, Lorenzen and Svith ^{109 )}. On the other hand, stimulation of parathyroid hormone in response to low serum Ca concentrations results in the elevation of intracellular Ca concentration and the triggering of adiposity through stimulation of enzymes involved in lipogenesis and inhibition of lipolysis⁽ Reference Zemel, Teegarden and Loan ^{110 )}. Sufficient ingestion of dietary Ca can hinder this process through optimising serum Ca levels and inhibiting stimulation of parathormone, leading to reduced adipogenesis and body fat mobilisation and oxidation.

Compared with supplemental Ca, dairy Ca has a greater effect on weight, suggesting that dairy components other than Ca may be effective⁽ Reference Zemel, Thompson and Milstead ^{111 ,} Reference Zemel ^{112 )}. For instance, dairy proteins may help in the suppression of appetite. Whey is supposed to have a stronger effect than casein on appetite control but each of whey and casein executes distinct mechanisms. While the casein fraction postpones gastric emptying following coagulation in the stomach, whey induces satiety due to rapid digestion, absorption and elevation of amino acids in the blood⁽ Reference Bendtsen, Lorenzen and Bendsen ^{113 )}. Bioactive compounds, low-glycaemic index carbohydrates and Ca content of dairy foods may also be implicated in appetite control⁽ Reference Dougkas, Reynolds and Givens ^{114 )}. It is worthwhile to note that the beneficial effect of dairy products on weight occurs when low-fat dairy products are consumed, but whole-fat dairy products may promote weight gain⁽ Reference Alonso, Zozaya and Vázquez ^{115 )}.

Glucose tolerance

Decreased weight has positive metabolic consequences such as improved insulin sensitivity. A meta-analysis of observational studies indicates a relatively consistent association between low Ca or dairy product intake and prevalent type 2 diabetes or the metabolic syndrome⁽ Reference Pittas, Lau and Hu ^{116 )}. Similarly, a meta-analysis of cohort studies suggested an inverse association between the intake of dairy products, low-fat dairy products and cheese and the risk of type 2 diabetes⁽ Reference Aune, Norat and Romundstad ^{117 )}. A prospective cohort study showed an association of higher consumption of dairy products and Ca with lower 9-year incidence of the metabolic syndrome and impaired fasting glucose or type 2 diabetes⁽ Reference Fumeron, Lamri and Abi Khalil ^{118 )}. Also, a prospective cohort study with 20 years of follow-up revealed the inverse association of milk and dairy product consumption and prevalence of the metabolic syndrome⁽ Reference Elwood, Pickering and Fehily ^{119 )}. A possible mechanism for dairy product protection against type 2 diabetes is the low-glycaemic index of milk and other dairy foods which suppresses postprandial hyperglycaemia⁽ Reference Ballard and Bruno ^{120 )}. The beneficial effect of dairy products on weight may also contribute to the prevention of type 2 diabetes by dairy foods. Moreover, milk proteins, in particular whey, have insulinotropic activity when consumed with meals⁽ Reference Nilsson, Stenberg and Frid ^{121 )}. Ca is also essential for insulin secretion by pancreatic β-cells (Fig. 4)⁽ Reference Pittas, Lau and Hu ^{116 )}.

Fig. 4. The role of calcium in secretion of insulin by pancreatic β-cells. For a colour figure, see the online version of the paper.

Blood pressure

Blood pressure is also affected by Ca and dairy product intake. Meta-analyses of prospective cohort studies support the inverse association between low-fat dairy products and risk of elevated blood pressure⁽ Reference Soedamah-Muthu, Verberne and Ding ^{122 ,} Reference Ralston, Lee and Truby ^{123 )}. The pooled relative risk (RR) per 200 g/d was 0·97 (95 % CI 0·95, 0·99) for total dairy products, 0·96 (95 % CI 0·93, 0·99) for low-fat dairy products and 0·96 (95 % CI 0·94, 0·98) for milk⁽ Reference Soedamah-Muthu, Verberne and Ding ^{122 )}. The protective effect may be greater in an at-risk population⁽ Reference Toledo, Delgado-Rodríguez and Estruch ^{124 )}. On the other hand, high-fat dairy products may have an adverse effect on blood pressure due to the increasing effect on weight⁽ Reference Alonso, Zozaya and Vázquez ^{115 )}. Both casein and whey constituents of milk protein have been effective in reducing blood pressure and arterial stiffness⁽ Reference Fekete, Givens and Lovegrove ^{125 –} Reference Boelsma and Kloek ^{127 )}. The effect is supposed to be exerted mainly by bioactive tripeptides which are produced in the gut following digestion of milk proteins. A part of the beneficial effect of milk on blood pressure may result from its K. Dairy products have a relatively high K content which contributes largely to daily K intake⁽ Reference Ballard and Bruno ^{120 )}. Furthermore, through prevention of transient postprandial hyperglycaemia, dairy products as a low-glycaemic index food prevent hyperglycaemia-induced oxidative stress, thus improving NO bioavailability and vascular function⁽ Reference Ballard and Bruno ^{120 )}.

Blood pressure

Although reducing Na intake was not initially among the principles of the DASH diet⁽ Reference Appel, Moore and Obarzanek ^{13 )}, Na restriction is recommended along with DASH in order to augment its blood pressure-lowering effect⁽ Reference Sacks, Svetkey and Vollmer ^{38 )}. However, reduced blood pressure following consumption of DASH does not occur due to Na restriction only, but other DASH components are also effective in the regulation of blood pressure. In fact, in the absence of salt restriction, a reduction in blood pressure is still observed following DASH consumption⁽ Reference Appel, Moore and Obarzanek ^{13 )}. Reducing Na intake from a high-Na ordinary diet (about 3·5 g/d Na) to a low-Na DASH diet (1·8 g/d Na) resulted in mean systolic blood pressure decreases of 7·1 and 11·5 mmHg in normotensive and hypertensive individuals, respectively⁽ Reference Sacks, Svetkey and Vollmer ^{38 )}. The American Heart Association and American College of Cardiology recommend no more than 2·4 g Na per d (equal to 6 g salt per d)⁽ Reference Eckel, Jakicic and Ard ^{128 )}. A further decrease to 1·5 g/d may be required for high-risk patients, for instance, for those at risk of stroke⁽ Reference Van Horn ^{129 )}. On the other hand, very intensive Na restriction is also not recommended because a low Na intake is associated with increased cardiovascular mortality, indicating that a U-shaped relationship may exist between Na intake and health outcomes⁽ Reference Graudal, Jürgens and Baslund ^{130 )}.

Mechanisms by which Na increases blood pressure include expansion of extracellular fluid volume and increased cardiac output⁽ Reference Feng, Dell’Italia and Sanders ^{131 )}, impaired renin–angiotensin–aldosterone system⁽ Reference Shimosawa ^{132 )}, activation of the sympathetic nervous system⁽ Reference Gavras and Gavras ^{133 )}, augmented vascular smooth muscle cell proliferation⁽ Reference Liu, Hitomi and Rahman ^{134 )} and reduced NO bioavailability, which decreases endothelium-dependent vasodilation⁽ Reference Boegehold ^{135 )}. Kidneys and the vascular system play pivotal roles in modifying the haemodynamic changes caused by Na, but their function is impaired in salt sensitivity⁽ Reference Feng, Dell’Italia and Sanders ^{131 )}.

Glucose tolerance

There are other benefits with Na restriction which may help against the metabolic syndrome. For instance, low dietary Na reduces insulin secretion without effecting insulin sensitivity, probably through interruption of the renin–angiotensin–aldosterone system⁽ Reference Luther, Byrne and Yu ^{136 )}. High aldosterone concentrations may impair β-cell function⁽ Reference Mosso, Carvajal and Maiz ^{137 )}. Also, a direct relationship between aldosterone, insulin resistance and hyperinsulinaemia⁽ Reference Colussi, Catena and Lapenna ^{138 )} has been suggested.

Weight

Prospective studies on cereal fibre and whole grains reported small but significant reductions in weight gain⁽ Reference Cho, Qi and Fahey ^{139 )}. A meta-analysis of RCT also showed reductions in weight, BMI and body fat in overweight and obese adults consuming soluble fibre supplements⁽ Reference Thompson, Hannon and An ^{140 )}. Due to resistance to gastrointestinal enzymes, dietary fibre is devoid of energy, and fibre-containing foods such as whole grains, vegetables and fruit have low energy density. In addition, because of their bulking properties, fibre-containing foods stimulate satiety signals without adding much into daily energy intake⁽ Reference Papathanasopoulos and Camilleri ^{141 )}. As a result of their viscosity and formation of gel in the stomach, soluble fibres delay gastric emptying, thus slowing down food transition through the small intestine which results in decelerating glucose absorption⁽ Reference Weickert and Pfeiffer ^{142 )}. Slow glucose absorption suppresses the insulin response and prevents episodes of hypoglycaemia, thus minimising hunger sensations. Moreover, SCFA produced as a result of colonic fermentation of soluble fibres have shown potential in regulating appetite through suppressing hunger hormones and stimulating satiety hormones although large doses of fibre are required to exhibit such effects⁽ Reference Nilsson, Johansson and Ekström ^{143 )}. SCFA may also increase energy expenditure through increased thermogenesis and fat oxidation although this effect has not been investigated in human studies⁽ Reference Byrne, Chambers and Morrison ^{144 )}.

Glucose tolerance

A meta-analysis of prospective cohort studies indicated that a two servings/d increase in whole grain consumption was associated with a 21 % decrease in the risk of type 2 diabetes⁽ Reference de Munter, Hu and Spiegelman ^{145 )}. Another meta-analysis of cohort studies showed a reduced risk of diabetes with higher intake of cereal fibre (RR for extreme categories 0·67), but no significant association for fruit was observed⁽ Reference Schulze, Schulz and Heidemann ^{93 )}. Later meta-analyses showed the inverse association between intake of total dietary fibre and cereal fibre as well as insoluble fibre and fruit and risk of type 2 diabetes⁽ Reference Yao, Fang and Xu ^{146 )}. In a large case–control study, lower risk of diabetes was associated with the intake of cereal and vegetable fibre, but not fruit fibre^{( 147 )}. Meta-analyses of RCT have also shown benefits of low-glycaemic index foods in the prevention of diabetes⁽ Reference Barclay, Petocz and McMillan-Price ^{148 )} or improvement of glycaemic control⁽ Reference Thompson, Hannon and An ^{140 ,} Reference Brand-Miller, Hayne and Petocz ^{149 –} Reference Bajorek and Morello ^{153 )}.

The anti-diabetic effect of fibre may be partly due to its beneficial effect on weight. In a meta-analysis of cohort studies, the inverse association of fibre intake with diabetes risk diminished after adjustment for BMI^{( 147 )}. Fibre also has a direct impact on gastric emptying which increases intestinal transit time. SCFA that are produced in the colon as a result of colonic bacteria fermentation can also delay gastric emptying and decelerate glucose absorption through stimulating secretion of anorectic hormones, peptide YY and glucagon-like peptide-1 (GLP-1)⁽ Reference Scheithauer, Dallinga-Thie and de Vos ^{154 ,} Reference Canfora, Jocken and Blaak ^{155 )}. There are also additional mechanisms by which SCFA may contribute to the prevention of diabetes. These include anti-inflammatory effect of SCFA, balancing composition and activity of gut microbiota, and reducing hepatic glucose production by suppressing gene expression of the gluconeogenic enzymes glucose 6-phosphatase and phosphoenolpyruvate carboxykinase⁽ Reference den Besten, van Eunen and Groen ^{156 –} Reference Kuo ^{158 )}. The above-mentioned mechanisms are suggested for soluble fibre. Nevertheless, benefits of fibre on diabetes are related to both types of soluble and insoluble fibre. In fact, data for the anti-diabetic potential of insoluble fibre are stronger and more consistent than soluble fibre although the mechanisms of the protection are less recognised⁽ Reference Davison and Temple ^{159 )}. Insoluble fibre may exert its effect through accelerating secretion of glucose-dependent insulinotropic polypeptide (GIP) and insulin response following meals, thus reducing the postprandial glucose rise⁽ Reference Weickert, Mohlig and Koebnick ^{160 )}. In addition, the consumption of high-cereal fibre diets may prevent high-protein diet-induced insulin resistance by interfering with protein absorption⁽ Reference Weickert, Roden and Isken ^{161 )}. High-protein diets are generally applied in weight loss programmes and impair insulin signalling through phosphorylation of serine kinase-6-1⁽ Reference Weickert, Roden and Isken ^{161 )}.

Lipids

In an umbrella review, thirty-one meta-analyses reported reductions in the RR of CVD mortality (RR 0·77–0·83), the incidence of CVD (RR 0·72–0·91), CHD (RR 0·76–0·93) and stroke (RR 0·83–0·93) in the highest v. the lowest dietary fibre intake⁽ Reference McRae ^{162 )}. The evidence also suggests effectiveness of fibre on risk factors of CVD. Meta-analyses on supplementation studies using viscous soluble fibres, β-glucan, psyllium or konjac glucomannan also reported statistically significant reductions in both total and LDL-cholesterol concentrations⁽ Reference McRae ^{162 –} Reference Ho, Jovanovski and Zurbau ^{164 )}. Similarly, a systematic review revealed that a breakfast based on oats, barley or psyllium may lower cholesterol concentrations⁽ Reference Williams ^{165 )}. In another meta-analysis, TAG levels did not change but HDL-cholesterol concentrations were increased slightly by fibre⁽ Reference Hartley, May and Loveman ^{166 )}. The mechanisms involved include: (1) interfering with enterohepatic circulation through prevention re-absorption of bile acids in the ileum, thus enhancing utilisation of blood cholesterol for de novo synthesis of bile acids; (2) slowing glucose absorption and subsequently suppressing insulin response, thereby reducing stimulation of hepatic cholesterol synthesis by insulin; (3) increasing hepatic LDL receptors; and (4) diminution of intestinal cholesterol absorption and its hepatic synthesis by SCFA⁽ Reference Papathanasopoulos and Camilleri ^{141 ,} Reference Gunness and Gidley ^{167 ,} Reference Chen, Xu and Huang ^{168 )}. Compared with soluble fibres, insoluble fibres exhibit a smaller reduction on blood cholesterol, but binding bile acids has been mentioned as their mechanism of action against blood cholesterol⁽ Reference van Bennekum, Nguyen and Schulthess ^{169 )}.

Blood pressure

A meta-analysis of RCT showed a lowering effect of viscous soluble fibre on systolic and diastolic blood pressure⁽ Reference Khan, Jovanovski and Ho ^{170 )}. Psyllium had a stronger effect than β-glucan, guar gum, konjac and pectin. Another meta-analysis indicated the effectiveness of dietary fibre in trials conducted on hypertensive patients and in trials with an intervention duration of ≥ 8 weeks but reduction in normotensive subjects was less conclusive⁽ Reference Whelton, Hyre and Pedersen ^{171 )}. However, compared with other components of the metabolic syndrome, the mechanisms of the effect of fibre on blood pressure are less recognised. The beneficial effect of fibre on blood pressure may have resulted from of its impact on weight. Also, some of the antihypertensive effect of fibre may be exerted through amelioration of insulin resistance and reducing insulin concentrations⁽ Reference Aleixandre and Miguel ^{59 ,} Reference Sarafidis and Bakris ^{172 )}. Insulin has an antinatriuretic potential by which it stimulates renal Na reabsorption⁽ Reference Sarafidis and Bakris ^{172 )}. Therefore, reduction of insulin concentrations could be beneficial for blood pressure. Furthermore, fermentation of soluble fibre in the distal intestine and colon produces acidic metabolites which may improve the absorption of minerals that are advantageous for blood pressure⁽ Reference Aljuraiban, Griep and Chan ^{173 )}. Since hypercholesterolaemia and hypertension are closely related, fibre may hinder hypertension by preventing hypercholesterolaemia⁽ Reference Tuñón, Martín-Ventura and Blanco-Colio ^{174 )}.

Weight

A systematic review of observational studies showed that obese individuals in any age group possess lower concentrations of antioxidants⁽ Reference Hosseini, Saedisomeolia and Allman-Farinelli ^{179 )}. Plasma levels of carotenoids, vitamins E and C, as well as Zn, Mg and Se were inversely correlated with obesity and body fat mass⁽ Reference Hosseini, Saedisomeolia and Allman-Farinelli ^{179 )}. In addition to antioxidant vitamins and minerals, polyphenols present in fruits and vegetables can prevent obesity. For instance, meta-analyses of RCT have shown the anti-obesity effect of flavanols⁽ Reference Akhlaghi, Ghobadi and Mohammad Hosseini ^{180 )} and isoflavones⁽ Reference Akhlaghi, Zare and Nouripour ^{181 )}. The anti-obesity effect of polyphenols may be exerted through increasing β-oxidation of fatty acids, induction of satiety, stimulating thermogenesis in brown adipose tissue, increasing lipolysis, control of adipocyte differentiation, down-regulation of fatty acid synthase gene expression, and functioning as a prebiotic for gut microbiota (Fig. 5)⁽ Reference Castro-Barquero, Lamuela-Raventós and Doménech ^{182 –} Reference Lin and Lin-Shiau ^{186 )}. Similarly, carotenoids may deter adiposity by enhancing fat oxidation and increasing energy waste in brown and white adipocytes⁽ Reference Bonet, Canas and Ribot ^{187 )}. Likewise, vitamin C may inhibit adipocyte differentiation, increase lipolysis and prevent glucose uptake by adipocytes⁽ Reference Garcia-Diaz, Lopez-Legarrea and Quintero ^{188 )}.

Fig. 5. Mechanisms of polyphenols against hyperglycaemia and obesity. GLP-1, glucagon-like peptide 1. For a colour figure, see the online version of the paper.

Glucose tolerance

The association between antioxidants and risk of diabetes has also been reported. A large retrospective cohort study with 5 years of follow-up on twenty-five communities across Japan showed an inverse association between consumption of green tea, coffee and total caffeine and the risk for type 2 diabetes⁽ Reference Iso, Date and Wakai ^{189 )}. Also, in a prospective cohort study with 10 years of follow-up in Japan, the hazard ratio for development of type 2 diabetes in the highest v. the lowest tertiles of serum α-carotene, β-cryptoxanthin and total carotenoids was 0·35, 0·43 and 0·41, respectively⁽ Reference Sugiura, Nakamura and Ogawa ^{190 )}. Various mechanisms have been suggested for the effect of polyphenols on glucose tolerance and insulin sensitivity (Fig. 5 ⁾. Polyphenols may increase insulin secretion through protection of β-cell integrity⁽ Reference Dragan, Andrica and Serban ^{191 )}. They may also stimulate secretion of GLP-1, a hormone involved in quick postprandial insulin response, increase GLP-1 half-life, stimulate β-cells to secrete insulin, and increase insulin sensitivity in peripheral tissues⁽ Reference Domínguez Avila, Rodrigo García and González Aguilar ^{192 )}. In peripheral tissues, polyphenols may activate PPARγ, thus inducing adiponectin production and improving insulin resistance⁽ Reference Umeno, Horie and Murotomi ^{193 )}. In addition, cell culture studies indicate that green tea polyphenols inhibit gluconeogenesis in hepatocytes and stimulate glucose uptake in rat skeletal muscle cells by using a phosphatidylinositol 3-kinase-dependent mechanism that mimics metabolic actions of insulin⁽ Reference Munir, Chandrasekaran and Gao ^{194 )}. Polyphenols have also potential to decrease starch digestion by inhibiting α-amylase activity⁽ Reference Xu, Shao and Xiao ^{195 )}.

No meta-analysis has ever examined the effect of vitamin C or vitamin E on the risk of type 2 diabetes but meta-analyses on the effect of these antioxidants on patients with type 2 diabetes have produced conflicting results. Two meta-analyses showed no beneficial effect of vitamin E supplementation on glycaemic control of patients with type 2 diabetes but results were more promising for individuals with poor glycaemic control or low serum vitamin E levels⁽ Reference Suksomboon, Poolsup and Sinprasert ^{196 ,} Reference Xu, Zhang and Tao ^{197 )}. Another meta-analysis indicated that supplementation with vitamin C, vitamin E or their combination did not improve insulin resistance of type 2 diabetes patients⁽ Reference Khodaeian, Tabatabaei-Malazy and Qorbani ^{198 )}, but a limited number of RCT showed that a single dose of vitamin C may be beneficial in reducing fasting blood glucose of these patients⁽ Reference Tabatabaei-Malazy, Nikfar and Larijani ^{199 )}. Another meta-analysis revealed that vitamin C did not modify glucose, HbA1c and insulin concentrations in a population containing both diabetic and non-diabetic participants but subgroup analyses indicated that vitamin C significantly reduced glucose concentrations in patients with type 2 diabetes and in interventions longer than 30 d; also vitamin C administration had a greater effect on fasting compared with postprandial insulin concentration⁽ Reference Ashor, Werner and Lara ^{200 )}.

Mechanisms of the possible effect of vitamin C and vitamin E in the management of diabetes are largely unknown but the beneficial effect of these antioxidants in establishing glycaemic control may be exerted through a direct effect on pancreatic β-cells by protecting them from oxidative stress-induced cell damage⁽ Reference Suksomboon, Poolsup and Sinprasert ^{196 )}. Hyperglycaemia-associated oxidative stress is also suggested to be involved in the development of insulin resistance⁽ Reference Henriksen, Diamond-Stanic and Marchionne ^{201 ,} Reference Ceriello and Motz ^{202 )}. Thus, a part of antioxidants’ protection against type 2 diabetes may be delivered by suppression of oxidative stress. Carotenoids may reduce insulin resistance by induction of PPARγ as well as inhibiting c-Jun NH₂-terminal kinase (JNK) and inhibitor κB kinase β (IKKβ) which induce insulin resistance through phosphorylation of insulin receptor substrate-1 (IRS-1)⁽ Reference Roohbakhsh, Karimi and Iranshahi ^{203 )}.

Lipids

There is limited evidence for the effect of antioxidants on blood lipids. A meta-analysis of RCT showed that vitamin C supplements may decrease TAG and LDL-cholesterol but the increase in HDL-cholesterol was not significant⁽ Reference McRae ^{204 )}. Similarly, meta-analyses of RCT showed that consumption of green tea catechins was associated with a significant reduction in total and LDL-cholesterol levels without causing significant changes in HDL-cholesterol or TAG levels⁽ Reference Kim, Chiu and Barone ^{205 ,} Reference Zheng, Xu and Li ^{206 )}. Another meta-analysis revealed beneficial effects of dark chocolate and cocoa products on total and LDL-cholesterol with no major effect on HDL-cholesterol and TAG⁽ Reference Tokede, Gaziano and Djoussé ^{207 )}. The cholesterol-lowering effect of polyphenols may be due to interfering with cholesterol absorption⁽ Reference Bursill, Abbey and Roach ^{208 )}, inhibiting cholesterol synthesis⁽ Reference Kobayashi, Nishizawa and Inoue ^{209 )}, and inducing expression and activity of LDL receptors⁽ Reference Dávalos, Fernández-Hernando and Cerrato ^{210 )}. Moreover, by preventing LDL oxidation, antioxidants such as vitamin C improve recognition of LDL particles by hepatic LDL receptors and thus expedite their removal from blood⁽ Reference McRae ^{204 )}. In addition, by preventing oxidation of HDL, antioxidants may improve reverse cholesterol transport which is a process that exchanges cholesterol from peripheral tissues and circulating lipoproteins with TAG of HDL, thereby facilitating elimination of cholesterol from blood⁽ Reference Millar, Duclos and Blesso ^{211 )}.

Blood pressure

Meta-analyses of RCT have shown benefits of polyphenols⁽ Reference Serban, Sahebkar and Zanchetti ^{212 –} Reference Ras, Zock and Draijer ^{215 )} and vitamin C on blood pressure⁽ Reference Juraschek, Guallar and Appel ^{216 )}. Dietary polyphenols have shown a vasoprotective effect by augmentation of endothelial synthesis of NO and endothelium-derived hyperpolarising factor, inhibition of angiotensin-converting enzyme, suppression of endothelin-1 synthesis and increased bioavailability of NO by scavenging free radicals⁽ Reference Medina-Remón, Estruch and Tresserra-Rimbau ^{217 ,} Reference Hügel, Jackson and May ^{218 )}. There is also evidence that polyphenols promote vasodilation through an endothelium-independent mechanism by a direct effect on vascular smooth muscle cells via blocking Ca channels⁽ Reference Larson, Symons and Jalili ^{219 )}.

Weight

Although less important than total energy content, the macronutrient composition of the diet is also believed to affect weight. Low-fat diets may have potential in weight control attempts⁽ Reference Peters ^{220 )}; however, meta-analyses of controlled trials suggest that the effect of low-fat diets on weight depends on the diet of the control group⁽ Reference Tobias, Chen and Manson ^{221 )}. As an example, a meta-analysis of RCT indicated that low-fat diets lead to weight reduction only when compared with usual diet, but not in comparison with other dietary compositions such as low-carbohydrate or high-fat interventions⁽ Reference Tobias, Chen and Manson ^{221 )}.

Glucose tolerance

Large-scale cohort studies on healthy populations have shown a positive association between total and saturated fat intake and the development of type 2 diabetes; however, the association disappeared after adjustment for BMI⁽ Reference van Dam, Willett and Rimm ^{222 )}. Nonetheless, individuals in the prediabetic state or those who are genetically vulnerable to metabolic disorders may develop type 2 diabetes following the consumption of high-fat diets⁽ Reference Nagao, Asai and Sugihara ^{223 )}. In fact, the susceptibility of individuals to insulin resistance is affected by genes, and dietary factors can alter this susceptibility⁽ Reference López-Miranda, Pérez-Martínez and Marin ^{224 )}. High-fat diets increase concentration of TAG constituents such as diacylglycerols and ceramides in muscle and adipose tissue. Such compounds phosphorylate serine residues of insulin receptor substrate through activation of serine kinases, leading to an impaired insulin function, decreased translocation of insulin-dependent GLUT-4, and eventually decreased glucose uptake⁽ Reference Galgani, Uauy and Aguirre ^{225 )}. On the other hand, as high-fat diets have low carbohydrate content, they may benefit patients with type 2 diabetes. In this regard, a meta-analysis of RCT suggested that high-fat diets reduce fasting blood glucose in type 2 diabetes patients, but not prediabetic individuals⁽ Reference Schwingshackl and Hoffmann ^{226 )}.

Blood pressure

The quantity of dietary fat has less impact on blood pressure than fatty acid composition of the diet⁽ Reference Hall ^{227 )}. However, an increased concentration of NEFA in blood as occurs following consumption of high-fat meals can impair endothelial function. This postprandial acute effect may potentiate stronger chronic impacts on endothelium if high-fat meals are consumed persistently⁽ Reference Hall ^{227 )}. An increased serum concentration of angiotensin-converting enzyme is proposed as a mechanism of a high-fat diet (45 %) on blood pressure⁽ Reference Schüler, Osterhoff and Frahnow ^{57 )}.

Weight

The ratio of saturated:unsaturated fats also affects weight. Saturated fats are probably more obesogenic than unsaturated fats⁽ Reference Krishnan and Cooper ^{228 )}. Compared with saturated fats, unsaturated fatty acids induce greater energy expenditure, diet-induced thermogenesis and fat oxidation⁽ Reference Krishnan and Cooper ^{228 )}. Gene studies revealed that polyunsaturated fats up-regulate expression of PPARα, a transcription factor involved in lipid oxidation, and down-regulate expression of PPARγ, a transcription factor involved in lipogenesis⁽ Reference Georgiadi and Kersten ^{229 )}. Saturated fats have the opposite effect on these genes⁽ Reference Staiger, Staiger and Haas ^{230 )}.

Glucose tolerance

The quality of consumed fats is more important than their quantity for the incidence of diabetes and insulin sensitivity. Saturated fats are associated with an increased risk of type 2 diabetes while unsaturated fats are associated with insulin sensitivity⁽ Reference Galgani, Uauy and Aguirre ^{225 )}. In a multinational study, newly diagnosed diabetes was observed more frequently in patients with higher consumption of total and animal fat and lower plant to animal fat ratio⁽ Reference Thanopoulou, Karamanos and Angelico ^{231 )}. It is worthwhile to note that high quantities of saturated fat (for instance > 15 %) probably increase the risk of insulin resistance and type 2 diabetes, but moderate quantities may not have such an effect⁽ Reference Morio, Fardet and Legrand ^{232 )}.

In diabetes patients, serum cholesteryl esters are mainly composed of SFA. In contrast, a number of studies reported the association of high concentration of unsaturated fatty acids in serum and muscle of healthy individuals and insulin sensitivity⁽ Reference Riccardi, Giacco and Rivellese ^{233 )}. Fatty acids which are received through the diet incorporate into the cell membrane and affect its activity. The ratio of unsaturated fatty acids to SFA present in membrane phospholipids affects membrane fluidity and has a direct relevance to insulin function and glucose transport efficiency⁽ Reference Weijers ^{234 ,} Reference Risérus, Willett and Hu ^{235 )}. Apart from the direct effect on insulin sensitivity of peripheral tissues, PUFA also improve hepatic insulin sensitivity through the suppression of lipogenesis and stimulation of fat oxidation⁽ Reference Risérus, Willett and Hu ^{235 ,} Reference Georgiadi and Kersten ^{229 )}.

Lipids

Despite the general notion for the atherogenicity of saturated fats, cohort studies do not consistently support the link between saturated fats and CVD. A meta-analysis of prospective studies with follow-up periods of 5 to 23 years showed no significant relationship between dietary saturated fat and risk of CHD or CVD⁽ Reference Siri-Tarino, Sun and Hu ^{236 )}. Another meta-analysis of prospective cohort studies supported a causal relationship between CHD and intake of high-glycaemic index foods and trans-fatty acids and an inverse association between CHD and consumption of vegetables, nuts and MUFA, and healthy dietary patterns, but no association was observed between saturated fats and CHD⁽ Reference Mente, de Koning and Shannon ^{237 )}. The lack of the association of saturated fats with CHD in cohort studies is probably due to substitution of dietary carbohydrates for saturated fats which reduces concentrations of total, LDL- and HDL-cholesterol in a way that total to HDL-cholesterol ratio, an indicator of CVD risk, does not change⁽ Reference Siri-Tarino, Sun and Hu ^{238 )}. In contrast, replacement of saturated fats with unsaturated fats lowers total and LDL-cholesterol, resulting in a decreased total:HDL-cholesterol ratio⁽ Reference Siri-Tarino, Sun and Hu ^{238 ,} Reference Schwab, Lauritzen and Tholstrup ^{239 )}. Such substitution has reduced the incidence of cardiovascular events when embedded in an eating pattern for a period of more than 2 years⁽ Reference Hooper, Abdelhamid and Moore ^{240 )}. In this regard, a prospective cohort study with 14 years of follow-up indicated that replacement of 5 % energy from saturated fats with unsaturated fats would reduce the risk of CHD by 42 %⁽ Reference Hu, Stampfer and Manson ^{241 )}. Also, a meta-analysis of RCT suggested that each 5 % energy increase from polyunsaturated fats in place of saturated fats led to a 10 % reduction in CHD⁽ Reference Mozaffarian, Micha and Wallace ^{242 )}. Nevertheless, evidence on the effect of substitution of polyunsaturated fats for saturated fats on risk factors of CVD is still lacking. For instance, insufficient available data did not allow a meta-analysis of RCT to find significant reductions in total cholesterol, LDL-cholesterol and TAG following replacement of unsaturated fats for saturated fats⁽ Reference Hannon, Thompson and An ^{243 )}.

Blood pressure

Studies on the relationship between blood pressure and dietary fats are rather scarce and no meta-analysis has been performed based on the current evidence. The available evidence suggests that saturated fats have unfavourable and unsaturated fats have beneficial impact on blood pressure and vascular function⁽ Reference Hall ^{227 )}. A cross-sectional study revealed that saturated fat intake was independently and strongly associated with hypertension⁽ Reference Beegom and Singh ^{244 )}. Also, in a RCT, substitution of 10 % saturated fats with either mono- or polyunsaturated fats decreased blood pressure and E-selectin without affecting flow-mediated dilation and other measures of vascular function⁽ Reference Vafeiadou, Weech and Altowaijri ^{56 )}. n-3 Fatty acids are the most advantageous fatty acids for blood pressure control. These fatty acids inhibit angiotensin-converting enzyme activity and increase NO bioavailability through augmentation of endothelial NO synthase activity, suppression of pro-inflammatory cytokines and inhibition of cyclo-oxygenase activity⁽ Reference Das ^{245 )}.