Dietary fats

Over the last four decades, dietary guidelines have recommended consuming a low fat diet (<30% energy intake), with a reduction in SFA and TFA intake to reduce CVD risk. Dietary fatty acids are known to influence the rate of atherosclerosis, largely through modification of blood lipid levels. However, emerging findings from large intervention studies and meta-analyses have challenged dietary recommendations advocating low fat diets and reduced SFA intake for CVD prevention.Reference Hoenselaar³ Furthermore, when SFA/TFA fat is replaced in the diet, other dietary components are also likely to exert an influence on vascular disease risk, possibly mediated by changes in overall lipid fractions and other mechanisms. Therefore, it is important to evaluate what dietary fat changes are currently effective to reduce CVD risk.Reference Hooper, Summerbell and Thompson⁴

Low-fat diets

There is no convincing evidence demonstrating that low-fat diets reduce CVD risk.Reference Hooper, Summerbell and Thompson^4, Reference Howard, Van Horn and Hsia⁵ The Women's Health Initiative randomized controlled trial (RCT), involving 48,835 women over a mean 8 years follow-up, reported no significant difference in the incidence of CVD between a low-fat intervention and a control group. Total fat intake decreased by 8.2% with reductions in SFA 2.9%, MUFA 3.3% and PUFA 1.5% in the intervention group. There was a significant reduction in low-density lipoprotein (LDL) in the intervention group but no improvement in HDL or other lipid fractions.Reference Howard, Van Horn and Hsia⁵

Saturated fatty acids

It is well established that high SFA consumption increases total cholesterol (TC) and LDL levels, which in turn accelerate atherosclerosis and increase CVD risk. Despite wide acceptance of this classical dietary fat hypothesis, a causal link between SFA intake and CVD has not yet been established. A meta-analysis of 21 prospective studies involving >350,000 adults followed over 5–23 years, found no significant association between SFA intake and CHD, stroke or total CVD risk.Reference Siri-Tarino, Sun, Hu and Krauss⁶ Moreover, findings from a large systematic review of prospective cohort and randomized intervention studies, found only weak evidence to support a causal relationship between total SFA intake and CHD.Reference Mente, de Koning and Shannon² Further large-scale intervention studies are required to fully evaluate the relationship between dietary SFA and CVD risk. A recent meta-analysis suggested that reducing SFA intake by reducing and/or modifying dietary fat reduced risk of CVD events by 14% (RR: 0.86; CI: 0.77–0.96) with no effect demonstrated on CVD mortality or overall mortality risk.Reference Hooper, Summerbell and Thompson⁴ Subgroup analyses indicated that reduction in CVD events was seen in studies of fat modification (not fat reduction), of at least two years duration, and in studies of men only.Reference Hooper, Summerbell and Thompson⁴ Therefore, whilst it remains prudent to advocate a reduced SFA diet, it is unclear as to what type of dietary component should replace SFA to optimally reduce CVD risk.

Modified low-SFA diets

It has been suggested that SFA replacement by n-6 PUFA may be associated with greatest CHD risk reduction.Reference Jakobsen, O'Reilly and Heitmann^7, Reference Mozaffarian, Micha and Wallace⁸ However, the most recent meta-analysis has indicated an adverse effect of replacement with n-6 PUFA on CVD events and mortality, but a significantly reduced CVD risk when SFA was replaced by a combination of both n-3 and n-6 PUFA.Reference Ramsden, Zamora and Leelarthaepin⁹ As with any meta-analysis, differences in the results observed may be partly explained by heterogeneity in study designs, dietary intake analysis, uncontrolled confounding factors and differences between study end-points.

Replacing SFA with carbohydrate has been shown to increase triglyceride (TG) concentrations,Reference Mensink, Zock, Kester and Katan¹⁰ which may adversely affect vascular health, but this will depend on the type of carbohydrate consumed. It has been argued that there is little evidence to suggest that replacement of SFA by MUFA has any benefit for vascular health.Reference Astrup, Dyerberg and Elwood¹¹ However, Mente et al. Reference Mente, de Koning and Shannon² reported strong evidence of an inverse causal link between MUFA and CHD. Furthermore, the MD characterized by increased MUFA is thought to offer greatest protection against CHDReference Mente, de Koning and Shannon² (see discussion on MD).

Trans fatty acids

Ratio of TC to HDL (TC:HDL) appears to be the strongest predictor for ischaemic heart disease (IHD) mortality (twice as informative as TC alone).Reference Lewington and Whitlock¹² Dietary TFA have greater TC-raising effects and HDL-lowering effects than dietary SFA. A meta-analysis evaluating dietary fatty acids and carbohydrates on blood lipids in 60 controlled trials, reported that TFA increased TC:HDL to almost twice that of SFA at comparable doses.Reference Mensink, Zock, Kester and Katan¹⁰ Further small RCTs, investigating blood lipid level responses to changes in TFA intakes, have reported significant reductions in HDL levels and significant increases in TC:HDL ratio, in response to high TFA (6.7–9% total energy intake) diet interventions.Reference Lichtenstein, Ausman, Jalbert and Schaefer^13, Reference de Roos, Schouten, Scheek, van Tol and Katan¹⁴ Overall, the evidence suggests a strong, positive association between dietary TFA intake and CHD risk.Reference Mente, de Koning and Shannon² Therefore, dietary advice should continue to advocate low TFA intake for CVD risk reduction.

n-3 PUFA intake

In contrast to other fatty acids such as SFA and TFA, n-3 PUFA are essential long-chain fatty acids found in oily fish (eicosapentaenioc acid and docosahexaenoic acid) and vegetable oils (α-linolenic acid). n-3 PUFA significantly reduces TG level in a dose-dependent manner and may increase LDL and HDL levels.Reference Van Horn, Mc Coin and Kris-Etherton¹⁵ Mechanistic effects of n-3 PUFA are also believed to include reductions in thrombotic tendency, inflammation, arrhythmias and overall improvement in endothelial function.

Epidemiological studies provide evidence of an inverse association between fish intake and CHD death in both men and women.Reference He, Song and Daviglus^16, Reference Whelton, He, Whelton and Muntner¹⁷ Consumption of fish appears to provide cardioprotection in a dose–response manner in adults, with or without pre-existing coronary disease, although the cardioprotective benefits of fish may be greater in populations with low habitual fish consumption and those at greater risk of developing CHD. Fish consumption has also been inversely related to stroke risk in adults in a dose-dependent manner. In a meta-analysis of 26 prospective cohort studies, comparing categories of fish intake, the pooled RR for cerebrovascular disease for 2–4 servings/week versus ≤1 serving/week was 0.94 (95% CI: 0.90–0.98) and for ≥5 servings/week versus 1 serving/week was 0.88 (95% CI: 0.81–0.96).Reference Chowdhury, Stevens and Gorman¹⁸

In the past decade, several large-scale intervention studies and meta-analyses have evaluated the effect of n-3 fish oil supplements in primary and secondary CVD prevention. A recent meta-analysis of 20 RCTs of 68,680 adults (including those with or without existing CVD) found no effect of n-3 PUFA on all-cause mortality, cardiac death, sudden death, myocardial infarction (MI) or stroke.Reference Rizos, Ntzani, Bika, Kostapanos and Elisaf¹⁹ In a further pooled analysis of 14 randomized, double-blind, placebo-controlled trials involving 20,485 patients with existing CVD, n-3 supplementation did not reduce the risk of overall CVD events (RR: 0.99; CI: 0.89–1.09), all-cause mortality, sudden cardiac death, MI, congestive heart failure, or transient ischaemic attack and stroke.Reference Kwak, Myung, Lee and Seo²⁰ Therefore, the current evidence does not support routine n-3 fish oil supplementation for primary or secondary prevention of CVD.

Phytosterols

Phytosterols include β-sitosterol, campesterol and stigmasterol, and are structurally similar to cholesterol. Plant stanols are simply the ‘saturated’ derivative of the sterol, the most common being sitostanol.Reference Van Horn, Mc Coin and Kris-Etherton¹⁵ Esterifying plant stanols – typically adding oleic acid – makes it easier to incorporate this ingredient into everyday foods.

Both esterified and non-esterified plant sterols and stanols lower TC and LDL in both normocholesterolaemic and hypercholesterolaemic adults.Reference Miettinen, Puska, Gylling, Vanhanen and Vartiainen^21–Reference Volpe, Niittynen, Korpela, Sirtori, Bucci, Fraone and Pazzucconi²⁵ Meta-analyses evaluating cholesterol-lowering effects of phytosterols/stanols suggest an average 10% reduction in LDL when 2 g of plant stanol is included in the diet.Reference Moruisi, Oosthuizen and Opperman^26, Reference Law²⁷ Additive benefits in reducing LDL concentrations are observed when phytosterols/stanols are incorporated into low-fat diets.Reference Maki, Davidson and Umporowicz^23, Reference Volpe, Niittynen, Korpela, Sirtori, Bucci, Fraone and Pazzucconi²⁵ Importantly, further significant benefits in cholesterol lowering have been demonstrated in patients taking statin drug treatments in combination with phytosterols/stanols, with an average net benefit of a further 10% reduction in LDL concentration when 2–3 g/day sitostanol is consumed.Reference Blair, Capuzzi, Gottlieb, Nguyen, Morgan and Cater^22, Reference Gylling, Radhakrishnan and Miettinen²⁸ However, as yet, there are no clinical trials of phytosterols with hard clinical end-points.Reference Silbernagel, Genser, Nestel and März²⁹

Glycaemic index and glycaemic load

It is accepted that carbohydrate-containing foods elicit different postprandial glucose responses. This has been quantified as the ‘glycaemic index’ (GI), defined as the relative area under the curve for blood glucose concentrations during a 2-hour period after consuming a test food compared with a standard food (glucose or white bread) containing the same amount of digestible carbohydrate (50 g).Reference Jenkins, Wolever and Taylor³⁰ A second concept, that of glycaemic load (GL), is calculated by adjusting the GI value of the food to the serving size as quantified by carbohydrate content.Reference Kristo, Matthan and Lichtenstein³¹ Such a classification system has been suggested as useful in guiding food choices to reduce chronic disease risk. However, whilst there is some evidence, mainly from observational studies, that low GI and GL diets are associated with reduced risk of CVD, recent meta-analyses have not been consistent regarding whether both GI and GL are important, whether the association persists in both genders and whether the association is equally strong for both CHD and stroke.Reference Dong, Zhang, Wang and Qin^32–Reference Ma, Liu and Song³⁴ There is some suggestion that the association may depend on initial BMI,Reference Dong, Zhang, Wang and Qin³² while a meta-analysis of intervention studies altering GI/GL suggests inconsistent effects on CVD risk factors.Reference Kristo, Matthan and Lichtenstein³¹

Alcohol

Many epidemiological studies have shown that alcohol consumed in light to moderate amounts reduces mortality from vascular eventsReference Heuberger³⁵ and reduces CHD risk by 20–40%.Reference Mukamal, Jensen and Grønbaek^36–Reference Danaei, Ding and Mozaffarian³⁸ While a moderate consumption of alcohol is protective against CHD, alcohol manifests a more complex relationship with other types of CVD. There appears to be a J-shaped relationship between alcohol intake and hypertension: the lowest risk is seen among light and moderate drinkers,Reference Thadhani, Camargo, Stampfer, Curhan, Willett and Rimm³⁹ while those who consume relatively large amounts of alcohol (>4 drinks/day) frequently have elevated blood pressure.Reference Corrao, Bagnardi, Zambon and La Vecchia^40, Reference Patra, Taylor and Irving⁴¹ A similar pattern is seen with risk of stroke. While the findings from large epidemiological studies are inconsistent, it appears that moderate drinkers experience a reduced risk of stroke while the risk rapidly escalates in heavy drinkers.Reference Danaei, Ding and Mozaffarian^38, Reference Patra, Taylor and Irving⁴¹ The protective effect of moderate consumption of alcohol is seen only for ischaemic stroke, but not for haemorrhagic stroke.Reference Danaei, Ding and Mozaffarian³⁸ A similar J-shaped relationship has been observed between alcohol consumption and several other conditions.Reference Temple, Wilson T and Jacobs⁴²

The cardioprotective effect of alcohol may also depend on the type of alcoholic beverage consumed. While it has consistently been shown that wine confers a cardioprotective effect,Reference Di Castelnuovo, Rotondo, Iacoviello, Donati and De Gaetano⁴³ there is also recent epidemiological evidence that shows a J-shaped relationship between beer consumption and vascular risk, but no association with spirit consumption.Reference Costanzo, Di Castelnuovo, Donati, locoviello and de Gaetano⁴⁴

The mechanistic effect of alcohol is thought to be related to increased HDL concentrations. Alcohol may also prolong blood clot formation by lowering levels of fibrinogen and plasminogen activator, reducing HbA1c and increasing adiponectin, although an increase in TG has also been noted.Reference Mukamal, Jensen and Grønbaek³⁶

There are currently no guidelines regarding alcohol intake for older adults.Reference Heuberger³⁵ A sensible policy would appear to be one that informs people that alcohol in moderation will probably have several health benefits for people in middle age and above, but also stresses that binge drinking should be avoided, and stresses the problems of abuse.

Salt

The estimated daily requirement for dietary sodium in the UK is 1500 mg up to age 50, and 1300 mg between ages 50 and 70 (equivalent to a salt intake of 3.8 and 3.3 g, respectively). However, the intake in most countries is far higher, around 3500 mg/day.Reference Webster, Dunford, Hawkes and Neal⁴⁵

Excess sodium intake is believed to play an important role in the aetiology of hypertension.Reference He and MacGregor⁴⁶ A recent meta-analysis involving 34 RCTs revealed that a decrease in sodium intake of 1700 mg/day (4.4 g of salt) lowered blood pressure by 5.4/2.8 mmHg (systolic/diastolic) in subjects with hypertension and by 2.4/1.0 mmHg in those who were normotensive.Reference He, Li and Macgregor⁴⁷ They also reported that a greater reduction in sodium intake leads to a larger decrease in blood pressure and that, ideally, sodium intake should be reduced to 1180 mg/day (3 g of salt).

As hypertension is a major risk factor for CHD (and other forms of CVD), it can be predicted that sodium intake will be associated with risk of CHD. Supporting evidence for this comes from another systematic review and meta-analysis by Aburto et al.,Reference Aburto, Ziolkovska, Hooper, Elliott, Cappuccio and Meerpohl⁴⁸ who reported that a higher sodium intake was significantly associated with increased risk of CHD mortality, stroke and stroke mortality. A separate analysis concluded that a population-wide reduction in dietary sodium of 1200 mg/day (3 g of salt) would prevent between 44,000 and 92,000 deaths/year in the USA.Reference Bibbins-Domingo, Chertow and Coxson⁴⁹ It would also be expected to decrease the annual incidence of CHD cases by between 60,000 and 120,000 and stroke cases by between 32,000 and 66,000.

The overall evidence suggests that a major reduction in the average population intake of salt should be a priority. Approximately 75–80% of dietary salt intake comes from processed foods,⁵⁰ therefore intake of these foods should be limited. In the UK, voluntary efforts are already being undertaken by the Food Industry to lower the salt content of processed food. This strategy has achieved some success: between 2001 and 2011 mean sodium intakes of adults in the UK fell by about 15% (from about 3700 to 3200 mg/day).Reference Brinsden, He, Jenner and MacGregor⁵¹ However, in order to achieve country specific sodium targets within a reasonable timespan it will be necessary to adopt a legislative strategy, which could be hugely beneficial for public health.Reference Temple⁵²

Sugar-sweetened beverages

There has been speculation for many years that a high sugar intake (i.e. highly processed sugar added to food) may increase the risk of CHD.Reference Yudkin^53, Reference Yudkin⁵⁴ Most attention has focused on sugar-sweetened beverages (SSBs), and recent evidence supports this. The Nurses’ Health Study reported that risk is associated with an excessive consumption of SSBs. In this cohort study, the diets of 88,000 nurses were assessed several times over 24 years. The findings revealed that risk of CHD was 35% higher among women who consumed two or more servings per day of SSBs compared with those consuming SSBs less than once a month.Reference Fung, Malik, Rexrode, Manson, Willett and Hu⁵⁵

There has also been research into the relationship between consumption of SSBs and risk factors for CHD. Large cohort studies have demonstrated that people who regularly consume SSBs have increased risk of developing hypertension. This was shown in the Nurses’ Health Study I and II, where women who consumed four or more servings per day of SSBs had a 44 and 28%, respectively, higher risk of hypertension compared with women who consumed less than 1 serving per day.Reference Winkelmayer, Stampfer, Willett and Curhan⁵⁶ However, these results were not clear cut, as consumers of diet fizzy drinks also had increased risk of hypertension. The Framingham Offspring Study reported that those who drank at least one serving of SSB per day had a roughly 20% higher risk of hypertension, but this increase was not statistically significant.Reference Dhingra, Sullivan and Jacques⁵⁷ This study also observed a significantly higher risk of developing both hypertriglyceridaemia and low HDL-cholesterol in regular consumers of SSBs. On the whole, these findings suggest that frequent consumption of SSBs are associated with detrimental changes in risk factors for CHD.

Whole grains and fibre

A strong body of evidence has emerged from cohort studies showing that people who have relatively high intakes of whole grains or cereal fibre have lower risk of CHD.Reference Crowe, Key and Appleby^58, Reference Mozaffarian, Appel and Van Horn⁵⁹ As cereal fibre is usually eaten as part of whole grains, studies that show a protective association between intake of cereal fibre and risk of CHD are best interpreted as indicating that whole grains are protective; it is unwarranted to conclude without further evidence that cereal fibre specifically is protective.

There has been some investigation into possible mechanisms by which whole grains prevent CHD. One mechanism with reasonably solid supporting evidence, based on findings from several RCTs, is that whole grains lower LDL-cholesterol.Reference Ye, Chacko, Chou, Kugizaki and Liu⁶⁰ Most of the RCTs that have reported this effect have been carried out using oats.Reference Kelly, Summerbell, Brynes, Whittaker and Frost⁶¹ However, this still requires further investigation.

Folic acid and B vitamins

Over the past 20 years, a wealth of observational evidence has linked low concentrations of folate and/or high homocysteine concentrations with increased CVD risk in middle-aged and older adults, and these studies have been subjected to meta-analysis.Reference Wald, Law and Morris^{62, 63} The homocysteine-lowering effect of a combination of B-group vitamins has been demonstrated in a number of studies, and has also been confirmed by meta-analysis.^{64, 65} However, a number of RCTs have examined the effects of B vitamin supplementation on CVD risk, and meta-analyses have shown that such lowering has not led to a reduction in CVD risk,Reference Bazzano, Reynolds, Holder and He⁶⁶ although an effect on stroke has been suggested in some analyses.Reference Wang, Qin and Demirtas^67–Reference Yang, Lee, Hong, Ovbiagele and Saver⁷⁰ For stroke, a beneficial effect was notably stronger among trials with longer duration of follow-up, in those showing a greater reduction in total homocysteine (tHcy), in those without a history of stroke, and in geographical areas where grain is not fortified with folic acid.Reference Wang, Qin and Demirtas^67, Reference Lee, Hong and Chang⁶⁸ However, a comprehensive meta-analysis, which pooled together individual patient data, rather than summary results, showed that folic acid supplementation had no effect on the incidence of cardiovascular events (RR: 1.01; CI: 0.97–1.05) for major vascular events, 1.03 (0.97–1.10) for major coronary events and 0.96 (0.87–1.07) for stroke). There was also no evidence of a beneficial effect among the trials where treatment was continued for >5 years, or in patients presenting with low folate status or high homocysteine levels,Reference Clarke, Halsey and Lewington⁷¹ and these findings appear conclusive.

Antioxidants

Many observational studies have demonstrated an association between vitamin E and β-carotene intake and CVD risk, although the evidence is less consistent for vitamin C.Reference Moats and Rimm⁷² In the past 15 years, a number of RCTs have tested the effect of vitamin E and other antioxidant vitamins or their combinations on clinical CVD end-points. These trials have consistently shown that commonly used antioxidant regimens (vitamins E, C, β-carotene, singly or in combination) do not significantly reduce overall cardiovascular events, and this has been confirmed by meta-analyses.Reference Vivekananthan, Penn, Sapp, Hsu and Topol^73, Reference Bin, Hu, Cao and Gao⁷⁴ The HOPE-TOO study, an extended follow-up of the HOPE study, even suggested an increased risk of heart failure in the vitamin E-supplemented group.⁷⁵ Indeed, a meta-analysis of 19 clinical trials involving 135,967 participants found a significant dose-dependent adverse mortality effect of vitamin E when taken at 400 IU daily and above.Reference Miller, Pastor-Barriuso, Dalal and Riemersma⁷⁶ A beneficial effect of vitamin E supplementation in end-stage renal failure subjects cannot be excluded.Reference Boaz, Smetana and Weinstein⁷⁷

β-Carotene supplementation has also been evaluated in a number of trials examining both cancer and CVD end-points, with no reduction in CVD morbidity or mortality being demonstrated, alongside high-dose β-carotene supplementation increasing the incidence of both lung cancer and all-cause mortality in smokers.^78, Reference Goodman, Thornquist and Balmes⁷⁹ This lack of effect of β-carotene supplementation on CVD end-points has also been confirmed by meta-analysis.Reference Vivekananthan, Penn, Sapp, Hsu and Topol⁷³

A Cochrane review and meta-analysis of 68 RCTs including 232,606 participants has shown that supplementing with antioxidant nutrients (vitamins A, E and β-carotene) did not affect all-cause mortality.Reference Bjelakovic, Nikolova, Gluud, Simonetti and Gluud^80, Reference Bjelakovic, Nikolova, Gluud, Simonetti and Gluud⁸¹ However, when poorer quality trials were omitted, the results from the better quality studies revealed a positive effect of antioxidant supplement use on mortality (RR: 1.05; 95% CI: 1.02–1.08). Significant increases in mortality were also shown for use of vitamins A, E or β-carotene, either singly or in combination.Reference Bjelakovic, Nikolova, Gluud, Simonetti and Gluud^80, Reference Bjelakovic, Nikolova, Gluud, Simonetti and Gluud⁸¹

In 2007, the National Institutes of Health State-of-the-Science Conference on Multivitamin/Mineral Supplements and Chronic Disease Prevention concluded that there was insufficient evidence to recommend either for or against the use of multivitamin/mineral supplements to prevent chronic disease.⁸² It can be concluded that, for CVD specifically, antioxidant vitamin supplements should not be recommended to reduce primary or secondary CVD risk.

Polyphenols

In recent years, there has been growing interest in the role of polyphenols, such as flavonoids, for CVD prevention. Flavonoids are bioactive, polyphenolic metabolites that are primarily found in FV, nuts, cocoa and beverages such as tea and red wine. The link between flavonoids and CVD first emerged in studies conducted in Asian populations.Reference Kokubo, Iso and Ishihara⁸³ A number of prospective studies conducted in Western populations have also shown that high intakes of flavonoids are associated with lower CVD incidence and mortality.Reference Mink, Scrafford and Barraj^84–Reference McCullough, Peterson, Patel, Jacques, Shah and Dwyer⁸⁶ The association between flavonoids and stroke risk or mortality is less clear. While reduced risk of stroke was reported in one study,Reference Mursu, Voutilainen, Numi, Tuomainen, Kurl and Salonen⁸⁵ another study found no association between dietary flavonoids and stroke mortality.Reference Mink, Scrafford and Barraj⁸⁴ The most comprehensive evidence to date comes from the Cancer Prevention Study II Nutrition Cohort. This large prospective study of 38,180 men and 60,289 women showed that those in the highest quintile for total flavonoid intakes had decreased risk of CVD mortality compared with those in the lowest quintile (P for trend = 0.01). The inverse associations between flavonoid intake and CVD mortality were also observed with intermediate intakes of flavonoids, which led them to suggest that even small amounts of flavonoid-rich foods may confer beneficial effects.Reference McCullough, Peterson, Patel, Jacques, Shah and Dwyer⁸⁶ Although clinical trial evidence has yet to appear to support these findings, a few RCTs have shown beneficial effects of polyphenols found in soy and cocoa on CVD risk factors including blood pressure, flow-mediated dilatation and lipids.Reference Hooper, Kroon and Rimm^87, Reference Hooper, Kay and Abdelhomid⁸⁸ However, this area is still largely under-investigated and long-term clinical trials are needed to clarify the effect of flavonoids on clinical CVD end-points.

Vitamin D

Vitamin D, which is often deficient in older adults, may be associated with cardiovascular health. Evidence for a role of vitamin D in CVD mainly comes from epidemiological studies that examined the association between 25(OH)D concentrations and subsequent CVD events. The majority, although not all, of these studies have shown that low levels of 25(OH)D are associated with increased cardiovascular incidents and CVD mortalityReference Schottker, Haug and Schomburg^89–Reference Szulc, Claustrat and Delmas⁹⁵and this has also been confirmed by meta-analyses.Reference Wang, Song and Manson^96–Reference Sokol, Tsang and Aggarwal⁹⁹ The most recent meta-analysis by Wang and colleaguesReference Wang, Song and Manson⁹⁶ reported that individuals with low 25(OH)D concentrations had 48% greater risk of total CVD (95% CI: 1.30–1.77), 58% greater risk of CVD mortality (95% CI: 1.19–1.71), 62% greater risk of CHD (95% CI: 1.21–1.57) and 36% greater risk of stroke compared with those in the highest 25(OH)D categories. Moreover, they observed a general linear dose–response association between 25(OH)D (at a range of 20–60 nmol/l) and CVD risk, with a RR of 1.03 (95% CI: 1.00–1.06) per 25 nmol/l decrease in 25(OH)D.Reference Wang, Song and Manson⁹⁶ Two other meta-analyses also reported a modest association between low levels of 25(OH)D and increased risk of stroke.Reference Sun, Pan, Hu, Manson and Rexrode^100, Reference Brondum-Jacobsen, Nordestgaard, Schnohr and Benn¹⁰¹ However, despite the abundance of epidemiological evidence, numerous shortcomings in the literature have been identified, including variation in vitamin D status, seasonal variation, differences in CV outcomes and baseline vitamin D status.Reference Sokol, Tsang, Aggarwal, Melamed and Srinivas¹⁰² Furthermore, there is still insufficient clinical trial evidence to support a causal role, and meta-analyses of clinical trials conducted thus far have shown no significant effect of vitamin D supplementation on CVD end-points, including CVD mortality, MI, stroke and cardiovascular risk factors.Reference Elamin, Elnour and Elamin^103, Reference Wang, Manson, Song and Sesso¹⁰⁴ However, it is worth noting that the majority of the studies included in these meta-analyses did not evaluate cardiovascular outcomes as a primary end-point and many were using a combined calcium plus vitamin D supplement. A recent report by the Institute of Medicine¹⁰⁵ concluded that there was still insufficient evidence to demonstrate a protective role of vitamin D against CVD, and highlighted a need for more research to be undertaken. At least one trial is ongoing,Reference Manson, Bassuk and Lee¹⁰⁶ and additional large-scale and robustly designed RCTs such as VITAL are needed before vitamin D supplements can be routinely recommended for the prevention and treatment of CVD.

Obesity

Obesity (BMI ≥ 30 kg/m²) is an independent risk factor for CVD,Reference Rashid, Fuentes, Touchon and Wehner^107, Reference Whitlock and Lewington¹⁰⁸ but is also accompanied by a number of CVD risk factors, including hypertension, hyperlipidaemia and insulin resistance.Reference Denke, Sempos and Grundy^109–Reference Haslam, Sattar and Lean¹¹¹ BMI is the most widely used index of obesity, but waist circumference and waist-to-hip ratio (WHR) both measure abdominal adiposity, and each of these have been shown to be associated with increased CHD incidence and mortality (as reviewed by van Horn et al.Reference Van Horn, Mc Coin and Kris-Etherton¹⁵). BMI is poorly associated with total and CVD mortality in those older than 65 years,Reference Folsom, Stevens, Schreiner and McGovern^112, Reference Ellekjaer, Holmen and Vatten¹¹³ but even in those less than 65 years of age, waist circumference and WHR would appear to be better predictors of CVD risk than BMI (as reviewed by van Horn et al.Reference Van Horn, Mc Coin and Kris-Etherton¹⁵). However, whilst cut-points for BMI for overweight and obesity are well-established and accepted, further research is required to determine similar cut-points for WHR and waist circumference in different sex, age and ethnic groups.Reference Van Horn, Mc Coin and Kris-Etherton¹⁵

Physical activity

One of the factors limiting research into PA and primary and secondary CVD risk is the difficulty in objectively measuring PA.Reference Van Horn, Mc Coin and Kris-Etherton¹⁵ Nevertheless, PA has consistently been associated with a reduction both in CVD risk factors and CVD events in both primary and secondary prevention, and this has been shown in both observational epidemiological studies and in RCTs.Reference Tanasescu, Leitzmann, Rimm, Willett, Stampfer and Hu^114–Reference Heath, Parra and Sarmiento¹¹⁹ PA favourably affects blood pressure, lipids, insulin sensitivity, glucose tolerance and body weight.Reference Thompson and Lim¹²⁰ There is some debate over how much activity and what intensity of activity is required to produce a beneficial effect on CVD risk,Reference Blair and Connelly^121, Reference Bauman, Reis, Sallis, Wells, Loos and Martin¹²² but the main health-promoting bodies worldwide (e.g. American Heart Association, Centers for Disease Control and Prevention, British Heart Foundation and the UK government) all recommend individuals to engage in regular, moderate-intensity PA. A major challenge in this area is to determine what behavioural strategies will motivate individuals to become and remain physically active, as PA rates are declining, for example in the United States.Reference Brownson, Boehmer and Luke^123, Reference Hallal, Andersen, Bull, Guthold, Haskell and Ekelund¹²⁴

Whole diet/food approaches

While the traditional approach in nutrition epidemiology is to study the effect of individual nutrients, a growing number of studies are using a whole diet/food approach in the study of CVD prevention and therapy.

Fruit and vegetable intake

FV are complex foods that contain a variety of bioactive compounds,Reference Manach, Williamson, Morand, Scalbert and Rémésy¹²⁵ and the complex mixture of micronutrients found in a FV-rich diet may be more effective than large supplemental doses of a small number of micronutrients at reducing CVD risk.Reference Woodside, McCall, McGartland and Young¹²⁶ The evidence to date regarding the association between FV consumption and CVD risk is largely confined to epidemiological studies.Reference Woodside, Young and McKinley¹²⁷ These studies have consistently shown a positive association between increased FV consumption and reduced risk of both CHD and stroke, and this has been confirmed by meta-analyses.Reference Dauchet, Amouyel and Dallongeville^128–Reference He, Nowson, Lucas and MacGregor¹³¹ This dose-dependent association has largely been confirmed in more recent prospective cohort studies,Reference Takachi, Inoue and Ishihara^132–Reference Bendinelli, Masala and Saieva¹³⁵ including an analysis of the large EPIC study (n = 313,074), where a one-portion increment in FV was associated with a 4% lower risk of fatal IHD (95% CI: 0.92–1.00; P for trend = 0.033).Reference Crowe, Roddam and Key¹³⁶

While the epidemiological evidence is strong, no intervention studies with hard clinical end-points have yet been carried out, which show a causal effect of FV on CVD risk. Evidence has tended to focus on intermediate end-points, such as markers of cardiovascular risk. An effect of increased FV intake on blood pressure has been demonstrated in some studies,Reference John, Ziebland, Yudkin, Roe and Neil^137, Reference Appel, Moore and Obarzanek¹³⁸ although not all,Reference Broeksmans, Klopping-Ketelaars and Kluft^139–Reference Berry, Mulla and Chowienczyk¹⁴¹ while an effect of increased FV intake on vascular function has also been demonstrated in some,Reference McCall, McGartland and McKinley¹⁴² but not all studies.Reference Berry, Mulla and Chowienczyk¹⁴¹ A trend (P = 0.06 for between-group change) towards an effect of FV puree-based drinks (200 ml/day for 6 weeks) on endothelium-dependent vasodilatation in healthy volunteersReference George, Paterson, Waroonphan, Gordon and Lovegrove¹⁴³ has been demonstrated. However, comparing results from these studies is difficult due to variation in participant characteristics and methods used to assess vascular function.

The DASH diet

The Dietary Approaches to Stop Hypertension (DASH) diet is characterized by high intakes of FV, whole grains, low-fat dairy products and low intakes of meat. Interest in the DASH diet began in 1997 when an RCT by Appel and colleaguesReference Appel, Moore and Obarzanek¹³⁸ demonstrated that the DASH diet had a blood pressure-lowering effect that resulted in a reduction in estimated 10-year CHD risk.Reference Chen, Maruthur and Appel¹⁴⁴ This effect was improved further with the accompanying restriction of sodium intake,Reference Sacks, Svetkey and Vollmer¹⁴⁵ and the effectiveness of the DASH diet has also been shown in free-living populations.Reference Appel, Champagne and Harsha^146, Reference Blumenthal, Babyak and Hinderliter¹⁴⁷ Building on this evidence, a number of observational studies have examined the association between the DASH diet and cardiovascular outcomes including fatal and non-fatal CHDReference Fung, Chiuve, McCullough, Rexrode, Logroscino and Hu¹⁴⁸ and stroke.Reference Fung, Chiuve, McCullough, Rexrode, Logroscino and Hu^148, Reference Parikh, Lipsitz and Natarajan¹⁴⁹ While some of these studies have shown that adherence to the DASH diet is independently associated with fatal and non-fatal CHDReference Blumenthal, Babyak and Hinderliter¹⁴⁷ and stroke,Reference Fung, Chiuve, McCullough, Rexrode, Logroscino and Hu^148, Reference Parikh, Lipsitz and Natarajan¹⁴⁹ others have shown no association with fatal CVD,Reference Parikh, Lipsitz and Natarajan^149, Reference Folsom, Parker and Harnack¹⁵⁰ IHD,Reference Parikh, Lipsitz and Natarajan¹⁴⁹ CHDReference Parikh, Lipsitz and Natarajan¹⁴⁹ or stroke.Reference Folsom, Parker and Harnack¹⁵⁰ Consistent with previous findings, a recent study that examined 34,827 women from the Women's Health Study showed that higher DASH dietary scores were associated with reduced CVD and CHD risk, but significance was lost after adjusting for other cardiovascular risk factors.Reference Fitzgerald, Chiuve, Buring, Ridker and Glynn¹⁵¹

Mediterranean diet

It has been recognized for some time that there is a lower mortality from CHD in Mediterranean countries compared with Northern European countries.Reference Tunstall-Pedoe, Kuulasmaa, Mahonen, Tolonen, Ruokokoski and Amouyel¹⁵² High intakes of FV, bread and other forms of cereals, beans, nuts and seeds are key characteristics of the MD pattern. Olive oil is considered an important fat source of the MD, while dairy products, fish and poultry are consumed in low to moderate amounts; eggs are consumed <4 times/week; and little red meat is consumed. In addition, wine is consumed in low to moderate amounts with meals.

A recent meta-analysis by Sofi and colleaguesReference Sofi, Abbate, Gensini and Casini¹⁵³ examined the association between adherence to the MD and health status. The results were pooled from 18 prospective studies (including non-Mediterranean populations) involving more than 2 million individuals, and showed that a two-unit increment in adherence to the MD was associated with a significant reduction in cardiovascular incidence or mortality (RR: 0.90; 95% CI: 0.87–0.93) over a follow-up period of 4–20 years, and this has been confirmed in more recent analyses.Reference Gardener, Wright and Gu^154–Reference Hoevenaar-Blom, Nooyens and Kromhout¹⁵⁷

The MD has been rated as the most likely dietary pattern to offer protection against CHD,Reference Mente, de Koning and Shannon² as, in addition to the observational evidence base, a number of studies have examined the effect of a MD on clinical end-points. The Lyon Diet Heart Study, a secondary prevention trial in MI patients, showed that a MD resulted in a 50–70% reduction of cardiac end-points. de Lorgeril et al.Reference de Lorgeril, Salen, Martin, Monjaud, Delaye and Mamelle¹⁵⁸ reported significant reductions in three composite outcomes (CO): CO1 (cardiac death and non-fatal MI), CO2 (CO1 plus unstable angina, stroke, heart failure, pulmonary or peripheral embolism), and CO3 (CO2 plus minor events requiring hospital admission), with adjusted risk ratios ranging from 0.28 to 0.53.

More recent evidence comes from the primary prevention PREDIMED study.Reference Estruch, Ros and Salas-Salvado¹⁵⁹ This RCT tested the efficacy of two different types of MD (one supplemented with extra-virgin olive oil and the other with nuts) on the primary prevention of CVD in 7447 older men and women. Both diets were shown to reduce the risk of major cardiovascular events among high-risk individuals by approximately 30% (multivariable adjusted hazard ratio 0.70 (95% CI 0.54–0.92) for MD supplemented with extra-virgin olive oil, 0.72 (95% CI 0.54–0.96) for MD supplemented with mixed nuts).

While the evidence from both these studiesReference de Lorgeril, Salen, Martin, Monjaud, Delaye and Mamelle^158, Reference Estruch, Ros and Salas-Salvado¹⁵⁹ is convincing, more well-designed studies are needed to confirm the effectiveness of the MD in protecting against CVD. Further studies also need to be undertaken in non-Mediterranean populations to determine how best an increase in adherence to the MD can be achieved.

A posteriori dietary patterns

Alongside the research examining a priori dietary patternsReference Hu^160–Reference Hu and Willett¹⁶² where the level of adherence to a specific diet, for example, the MD or the DASH diet, or an index based on national dietary guidelines, e.g. the Healthy Eating Index is examined,Reference Agnoli, Krogh and Grioni¹⁶³ a number of observational studies have used an exploratory approach utilizing statistical techniques such as cluster analysis or principal component analysis/factor analysis to define empirical overall dietary patterns and examine their association with CVD risk. Studies using such a posteriori analyses have generally identified two main dietary patterns: the ‘Prudent or Healthy’ pattern which is characterized by high intakes of FV, legumes, fish and whole grains, and the ‘Western’ pattern which is characterized by high intakes of processed meat, red meat, refined grains and sweets and desserts. In general, the ‘Prudent’ pattern has been associated with reduced risk of CHD,Reference Fung, Willett, Stampfer, Manson and Hu^164, Reference Stampfer, Hu, Manson, Rimm and Willett¹⁶⁵ strokeReference Fung, Stampfer, Manson, Rexrode and Willett¹⁶⁶ and CVD mortality,Reference Heidemann, Schulze, Franco, van Dam, Mantzoros and Hu¹⁶⁷ whereas the ‘Western’ pattern has been associated with increased risk of CHDReference Fung, Willett, Stampfer, Manson and Hu^164, Reference Stampfer, Hu, Manson, Rimm and Willett¹⁶⁵ and stroke,Reference Fung, Stampfer, Manson, Rexrode and Willett¹⁶⁶ and a recent analysis of the EPIC-NL cohort showed similar associations.Reference Stricker, Onland-Moret and Boer¹⁶⁸ Overall, dietary pattern analysis has become a complementary approach in the study of diet–disease relationships and has been recommended as a strategy for future cohort studies and RCTs investigating relationships with chronic diseases such as CHD.Reference Mente, de Koning and Shannon²

Diet and vascular health – specific considerations for older people

The evidence base for an effect of diet on CVD risk has been reviewed for all ages, as most studies of diet for chronic disease prevention, although not conducted exclusively in older adults, have included older adults, and it seems reasonable that the health benefits of the dietary changes reviewed would extend into older adulthood.Reference Johnson¹⁶⁹ There are, however, a number of specific considerations when focusing on older people. Changes in body composition, hormonal status and gut function can have an effect on appetite, energy intake and the absorption of and requirements for particular nutrients (e.g. vitamin B12), whilst co-morbidities, concurrent medication use, cognitive decline, vision and hearing loss, loss of taste, change in dentition, pain, isolation, bereavement, decreased mobility and reduced income, which can all occur with ageing, will affect the ability to buy, prepare and consume food, as well as affecting appetite.Reference Gariballa, Sinclair, Geissler and Powers¹⁷⁰ Such ageing-related changes must be considered when developing dietary recommendations specifically for older people.

Conflicts of interest

The authors declare they have no conflicts of interest.