A number of reviewers have examined studies investigating the relationship between coronary heart disease and stroke prior to 2000. Since then, several key studies have been published. Five studies have examined the relationship between wholegrain consumption, coronary heart disease (CHD) and cardiovascular (CVD) disease and found protection for either or both diseases. The researchers concluded that a relationship between wholegrain intake and CHD is seen with at least a 20% and perhaps a 40% reduction in risk for those who eat wholegrain food habitually vs those who eat them rarely. Notwithstanding the fact that fibre is an important component of wholegrains, many studies have not shown an independent effect of fibre alone on CHD events. Thus in terms of CHD prevention, fibre is best obtained from wholegrain sources. Wholegrain products have strong antioxidant activity and contain phytoestrogens, but there is insufficient evidence to determine whether this is beneficial in CHD prevention. Soluble fibre clearly lowers cholesterol to a small but significant degree and one would expect that this would reduce CHD events. There have been a small number of epidemiological studies showing soy consumption is associated with lower rates of heart disease. Countering the positive evidence for wholegrain and legume intake has been the Nurses Health Study in 2000 that showed women who were overweight or obese consuming a high glycaemic load (GL) diet doubled their relative risk of CHD compared with those consuming a low GL diet. Although the literature relating GL with CHD events is somewhat mixed, the relationship with risk factors such as HDL cholesterol, triglyceride and C reactive protein is relatively clear. Thus, carbohydrate-rich foods should be wholegrain and, if they are not, then the lowest glycaemic index (GI) product should be used. Promotion of carbohydrate foods should be focused on wholegrain cereals because these have proven to be associated with health benefits. There is insufficient evidence about whether the addition of other components of wholegrains such as polyphenolics or minerals (such as magnesium or zinc) would improve the health benefits of refined grain foods and this needs investigation. Whether adding bran to refined carbohydrate foods can improve the situation is also not clear, and it was found that added bran lowered heart disease risk in men by 30%. This persisted after full adjustment (including GL) suggesting, at least in men, that fibre may be more important than GI. Thus there are two messages:
The intake of wholegrain foods clearly protects against heart disease and stroke but the exact mechanism is not clear. Fibre, magnesium, folate and vitamins B6 and vitamin E may be important.
The intake of high GI carbohydrates (from both grain and non-grain sources) in large amounts is associated with an increased risk of heart disease in overweight and obese women even when fibre intake is high but this requires further confirmation in normal-weight women.
Recommendation: Carbohydrate-rich foods should be wholegrain and if they are not, then the lowest GI product available should be consumed. Glycemic index is largely irrelevant for foods that contain small amounts of carbohydrate per serve (such as most vegetables).
Grains (in particular wheat, maize, rice, barley, sorghum, oats, rye and millet) and the foods made from them provide more than 56 per cent of the energy and 50 per cent of the protein consumed by humans worldwide (reported in National Health and Medical Research Council, 2003). Consumption of cereal grains forms the basis of a healthy diet, and all current dietary guidelines have cereal foods as the largest component of the recommended daily intake (National Health and Medical Research Council, 2003).
Notwithstanding these guidelines, most adults do not have enough grains in their diet. In Australia, the 1995 National Nutrition Survey found that even among those adults with the highest intake, on the day of the survey only 34 per cent of men and 21 per cent of women met the recommended seven servings a day (National Health and Medical Research Council, 2003). In the US, the average intake of wholegrains is less than one serving per day, and less than 10 per cent of Americans consume three servings per day (reported in Dietary Guidelines 2005 Advisory Committee, 2004). As a result, Australian dietary guidelines exhort adults to ‘eat plenty of cereals (including breads, rice, pasta and noodles), preferably wholegrain (National Health and Medical Research Council, 2003). Likewise, one of the key recommendations of the United States 2005 Dietary Guidelines is ‘Consume 3 or more ounce-equivalents of wholegrain products per day, with the rest of the recommended grains coming from enriched or wholegrain products. In general, at least half the grains should come from wholegrains’ (Dietary Guidelines 2005 Advisory Committee, 2004).
International attention is being directed towards efforts to increase whole grain consumption, as exemplified by the 2005 Whole Grains and Health Global Summit (Marquart and Jones, 2005) and the recent launch of the European Union HEALTHGRAIN Integrated Project, aimed at ‘exploiting bioactivity of European cereal grains for improved nutrition and health benefits’ (www.healthgrain.org).
Cereal grains are an excellent source of carbohydrate, dietary fibre and protein. They are good sources of B-group vitamins, vitamin E and a number of minerals, notably iron, zinc, magnesium and phosphorus. Additionally, wholegrain foods contain a large number of identified phytochemicals, such as phytoestrogens, antioxidants and phenolics, which, together with vitamins and minerals, may be protective against gastrointestinal cancers and cardiovascular disease (Table 1). These components may act together in a synergistic fashion (Slavin et al., 2001).
Milled wholegrains can be nutritionally superior to intact wholegrains for human consumption because poorly digested compounds are removed during the milling process, and nutrient bioavailability is enhanced (Slavin et al., 2000). Food Standards Australia New Zealand has proposed a definition of the term ‘wholegrain’ as ‘wholegrain means the intact grain or the dehulled, ground, milled, cracked or flaked grain where the constituents – endosperm, germ and bran – are present in such proportions that represent the typical ratio of those fractions occurring in the whole cereal, and includes wholemeal’ (FSANZ (Food Standards Australia New Zealand), 2004). Thus the distinction between intact and milled grain is removed which is consistent with epidemiological investigations that do not make this distinction.
Refined grains, where the bran and germ have been removed, have reduced nutrient content (Table 2) compared to wholegrains because the milling process results in the loss, to varying degrees, of dietary fibre, vitamins, minerals, lignans, phytoestrogens, phenolic compounds and phytic acid (Slavin, 2004). The starchy endosperm remaining after milling is ground to produce refined white flour.
Earlier reviewers of the literature (pre-2000) concluded that consumption of grains exerted a cardioprotective influence (Anderson, 1995; Kushi et al., 1999; Slavin et al., 1999). This report identifies reviews and studies published since 1999. We carried out a search for studies and reviews of studies in PubMed (www.ncbi.nlm.nih.gov/entrez/query.fcgi) using key words relating to coronary heart disease (CHD), cardiovascular disease and stroke, combined with key words describing whole grains, cereal grains, dietary fibre, soy and legumes. A total of 53 reviews or studies for cereal grains and heart disease risk factors in humans, five studies for stroke and 12 studies for legumes and heart disease are included.
Wholegrains and CHD
Wholegrain consumption has been shown to be linked to improvements in body mass index (BMI) (McKeown et al., 2002; Steffen et al., 2003), insulin sensitivity (McKeown et al., 2002; Liese et al., 2003; Steffen et al., 2003) and diabetes (Venn and Mann, 2004), all of which are important risk factors for heart disease. Consumption of wholegrains has been shown to improve insulin sensitivity (Pereira et al., 2002) although Juntunen et al. (2003) could not confirm this.
Several reviews published since 1999 have concluded that nutritional approaches to the prevention of CHD should include the consumption of wholegrains (Anderson et al., 2000; Richardson, 2000; McBurney, 2001; Slavin et al., 2001; Anderson, 2002; Liu, 2002; Truswell, 2002; Anderson, 2003; Hu, 2003; Slavin, 2004). The studies reviewed defined wholegrain as either intact or milled grain with bran, germ and endosperm in the same proportion as the unmilled grain. Wholegrain foods were arbitrarily defined as those foods with >25% by weight wholegrain or bran.
Anderson et al. (2000) carried out a meta-analysis of results from three large prospective studies (Fraser et al., 1992; Jacobs et al., 1998, 1999; Liu et al., 1999) which specifically examined the relationship of risk of CHD to wholegrain (Table 3). They concluded that persons in the highest, compared to those in the lowest, quintile for wholegrain intake had a 28% reduction in their risk for CHD (RR 0.72, 95% CI 0.49–0.94), a result which provides strong support for the hypothesis that generous intakes of wholegrains are cardioprotective. A subsequent update of this data (Anderson, 2003) included a fifth study (Liu et al., 2000a) which scarcely altered the results (relative ratio (RR) 0.72, 95% CI 0.48–0.94).
The fact that the associations found in these studies behaved in a dose–response manner and that the studies had been controlled for other dietary and lifestyle factors, including BMI, led Pereira and Liu (2003) to also conclude that ‘… wholegrain foods, through their fibre, antioxidants, and other components, reduce the risk of CHD in a causal manner’.
Jensen et al. (2004) followed a cohort of 42 850 male health professionals aged 40–75 years who were free from cardiovascular disease, cancer and diabetes at baseline in 1986. Daily wholegrain intakes (grams per day) were derived from a detailed and validated semiquantitative food frequency questionnaire (FFQ). The participants were followed with repeated questionnaires on lifestyle and health every 2 years and with detailed FFQs every 4 years. During 14 years of follow-up, 1818 incident cases of CHD occurred. After CHD risk factors and intakes of bran and germ added to foods were controlled for, the hazard ratio (HR) of CHD between lowest and highest quintiles of wholegrain consumption (median intake 3.5 vs 42.4 g/day) was 0.82 (95% CI 0.70–0.96; P for trend=0.01). The authors acknowledge that a greater intake of wholegrains was related to an overall healthier diet and lifestyle. They performed stratified analyses to address potential confounding by healthy participant characteristics and found no discrepant results with respect to the main finding of protection from heart disease with wholegrains in subgroups including never smokers and non-drinkers.
A further review (Jacobs and Gallaher, 2004) has also reported the results of more recent prospective studies of wholegrain consumption (Table 3). The reviewers concluded that a relationship between wholegrain intake and CHD is seen with at least a 20% and perhaps a 40% reduction in risk for those who eat wholegrain food habitually vs those who eat them rarely.
Jacobs and Gallaher consider that, given the wide variability in study designs, an estimated risk reduction of 20–40% is ‘impressively robust’. Study participants included men and women; participants from the US (Liu et al., 2000a; Bazzano et al., 2003; Liu et al., 2003; Mozaffarian et al., 2003; Steffen et al., 2003) and Norway (Jacobs et al., 2001). The cardio protective benefit is seen in the bulk of studies involving middle-aged participants as well as in the single study of the elderly (Mozaffarian et al., 2003). Findings, which were consistently positive for wholegrains, occurred using a variety of different data collection methodologies. Studies described in Table 3 were conducted in cultural settings in which commercially available wholegrain wheat and oat products predominated.
In conclusion, Jacobs and Gallaher and other reviewers find that there is good evidence that wholegrain foods substantially reduce the risk of CHD. Whether all grains are equal in this respect cannot be concluded from these studies, nor can the effectiveness of different parts of the grain. Results from Jensen et al. (2004) suggest that added bran may confer additional benefits; clearly this proposition needs more verification.
Whole grains and stroke
Liu et al. (2000a) examined the relationship of whole grain intake and risk of ischaemic stroke in a 12-year follow-up of the Nurses Health Study. The baseline population for analysis consisted of 75 521 women between 38 and 63 years, without previous diagnosis of diabetes mellitus, angina, myocardial infarction, stroke or other cardiovascular diseases. Incident cases of stroke were confirmed by reviewing medical records and classified as ischaemic where the stroke emanated from thrombotic or embolic occlusion of a cerebral artery, resulting in infarction. Whole grain intake was reported through a self-administered FFQ. During 861 900 years of follow-up, an inverse association between whole grain intake and ischaemic stroke risk was observed; the relative risk for ischaemic stroke in the highest quintile (median intake 2.7 servings/day) was 0.69 (95% CI, 0.50–0.98, P=0.04 for trend) relative to the lowest quintile (median intake 0.13 servings/day), adjusted for other stroke risk factors. No such inverse relationship was observed for refined grain intake. The authors also examined the possibility that specific constituents (folate, potassium, magnesium, vitamin E and fibre) explained the protective effect of whole grains. Adjustment for those components attenuated the inverse relation of whole grain intake with ischaemic stroke (RR, 0.76; 95% CI, 0.51–1.15, when comparing the highest vs the lowest quintiles, P=0.38 for trend). However, the inverse relation remained (even though not statistically significant) suggesting that other constituents may confer additional protection. The authors acknowledge that their findings may not be generalizable to other populations (98% of their cohort were white).
Steffen et al. (2003) did include more African American men and women in their prospective study (10 and 16 per cent, respectively). They followed 11 940 men and women from four US cities, aged 45–64 years, over 11 years. Usual dietary intakes were assessed using a FFQ at baseline and 6 years later. Overall, they found a beneficial effect of whole grain on risks of total mortality and incident coronary artery disease but not on the risk of ischaemic stroke, after adjustment for potential confounders (RR, 0.75, 95% CI, 0.46–1.22 when comparing the highest vs the lowest quintiles, P=0.15 for trend). It is noteworthy, however, that this inverse result is not far removed from that found by Liu et al. (2000a) as described above, albeit nonsignificant. Steffen and coworkers highlight an important limitation of their FFQ, which potentially misclassified whole-grained consumption as refined-grain consumption, thereby creating the possibility for attenuation of the true HR.
Mozaffarian et al. (2003) followed 3588 men and women aged 65+ years at baseline for 8.6 years to determine the association between fibre consumption from fruit, vegetables and cereal sources and incident cerebrovascular disease (CVD), defined as combined incident stroke, fatal and nonfatal myocardial infarction, and CHD death. In secondary analyses, they found that higher cereal fibre intake was associated with lower risk of total stroke (HR, 0.78, 95% CI, 0.64–0.95) and ischaemic stroke (HR, 0.76, 95% CI, 0.60–0.95) when comparing the 80th percentile with the 20th percentile.
Fung et al. (2004) examined the relationship between stroke risks in women and overall dietary patterns, specifically a ‘prudent’ diet (characterized by higher intakes of fruits, vegetables, whole grains, fish and poultry) and a ‘western’ diet (with higher intakes of red and processed meats, refined grains, full-fat dairy products, desserts and sweets). They followed 71 768 women aged between 38 and 63 years, without a history of CVD or diabetes at baseline, for 14 years. Dietary intake was ascertained by FFQ, which assessed food intake during the previous year, from which dietary pattern scores were calculated. After adjustment for stroke risk factors, they found women in the highest quintile of the Western pattern score had a relative risk of 1.58 (95% CI, 1.15–2.15, P=0.0002 for trend) for total strokes and 1.56 (95% CI, 1.05–2.33; P=0.02 for trend) for ischaemic stroke when compared with the lowest quintiles. For the prudent pattern, after adjustment, the RRs were 0.78 (95% CI, 0.61–1.01; P=0.13 for trend) for total stroke and 0.74 (95% CI, 0.54–1.02, P=0.13 for trend) for ischaemic stroke.
There are few studies to date that specifically examine the relationship between whole grain intakes and risk of stroke. The studies described above have shown mixed results when adjustment has been made for potential confounders. However, the trends are strongly suggestive of a protective effect of whole grain on risk of stroke.
Cereal fibre and CHD
A number of researchers have found that a higher intake of cereal fibre is associated with a lower risk of CHD, although results tend not to be statistically significant after adjusting for multiple confounding factors (Anderson et al., 2000; Truswell, 2002; Bazzano et al., 2003; Mozaffarian et al., 2003). Pereira and associates (2004) pooled the results of nine studies, yielding a total of 310 278 men and women who had a total of 6959 CHD events, of which 1869 were deaths. With cereal fibre fitted as a continuous variable, each 10 g/day increment in cereal fibre had a nonsignificant relative risk of 0.90 for total CHD events and a significant relative risk of 0.75 for CHD deaths (Table 4). The associations appeared to be independent of other dietary factors, sex, age, baseline BMI, history of hypertension, diabetes and hypercholesterolaemia. The authors suggest that failure to reach significance for total CHD events may be because of the limitations of some FFQs to accurately quantify cereal fibre intake or to small numbers of events in women.
Anderson et al. (2000) also performed a pooled analysis that explored the relationship between wholegrains and whole wheat bread, cereal fibre, total dietary fibre, fruits and vegetables and risk for CHD. After adjustment for confounding factors, they found that the strongest inverse association was between wholegrain and whole wheat bread intake and risk of CHD. Cereal fibre by itself had the least influence on CHD risk (RR 0.90; 95% CI, 0.80–1.01).
A similar finding was made by Mozaffarian et al. (2003). After adjustment for potential confounders, the overall reduction in cardiovascular disease risk from consumption of cereal fibre was 14% (HR 0.86, 95% CI, .075–0.99) when the 20th percentile of intake (<1.7 g/day) was compared with the highest quintile (>6.3 g/day). However, when post hoc analyses were conducted to investigate whether the observed lower cardiovascular disease risk was related to fibre from any specific food group, it was found that the lower risk appeared to be predominantly related to fibre intake from dark breads such as whole wheat, rye or pumpernickel (i.e. wholegrain intake) (HR 0.76; 95% CI 0.64–0.90). Other cereal fibres had a nonsignificant influence (high fibre, bran or granola cereals (HR 0.99; 95% CI 0.84–1.17), other cold cereals (HR 0.98; 95% CI 0.94–1.02), cooked cereals (HR 1.01, 95% CI 0.92–1.11)). The authors did not state whether these other cold or cooked cereal fibres were wholegrain or refined grain products.
In contrast, Jensen et al. (2004) investigated the association between daily intake of wholegrains, additional bran and germ and CHD risk in men in a cohort of 42 850 health professionals. As described earlier, they found that men with the highest intake of wholegrain had an 18% lower risk of CHD than men in the lowest quintile. To ascertain whether the effect of added bran and germ conferred even greater benefits, the wholegrain content of all grain foods eaten was determined, and bran or germ contained in those foods measured as an amount that would typically be found in the type of wholegrain. Wheat bran, corn bran, oat bran, rice bran and wheat germ added to foods either during processing or by participants while cooking were considered to be added bran and germ. They found that the HR of CHD in men with the highest intake of added bran (median intake 11.10 g/day) was 0.70 (95% CI 0.60–0.82) compared with men with no intake of added bran (P for trend <0.001). Thus, men with the highest intake of added bran had a 30% lower risk of CHD than men who do not consume added bran. Added germ was not associated with CHD risk (although the intake of added germ was very low in this population).
Bazzano et al. (2003) examined the relationship between dietary fibre intake and risk of CHD and CVD in 9776 adults who participated in the First National Health and Nutrition Examination Survey (NHANES 1) Epidemiologic Follow-up Study (NHEFS) conducted in the US. After adjusting for ‘healthy habits’ (regular exercise, not smoking, low dietary intake of cholesterol and saturated fats), they found that consumption of at least 4.5 g of cereal fibre per 1735 kcal was associated with a nonsignificant 20% lower risk of CHD (RR 0.80, 95% CI, 0.63–1.01; P=0.06 for trend) and a nonsignificant 11% lower risk of death from CHD (RR, 0.89, 95% CI, 0.76–1.02; P=0.10 for trend). A limitation of this study is that dietary fibre was estimated using a single 24-h dietary recall with no follow-up data collected, which may have resulted in misclassification of usual dietary fibre intake (there is no data to show which is the best instrument to assess fibre intake).
Lairon et al. (2003) in France investigated the relationship between dietary fibre intake and cardiovascular risk factors in a subsample of 4080 people (2168 male, 1912 women; 45–60 years at inclusion) drawn from the Supplementation en Vitamines et Mineraux Antioxydants (SU.VI.MAX) study population of 12 735 participants. This study was a randomized double-blind, placebo-controlled primary prevention trial designed to test the efficacy of daily supplementation with antioxidant vitamins and minerals at non-pharmacological doses in reducing cancers and cardiovascular diseases. The study commenced in 1994 and followed subjects for 8 years. Total dietary fibre intake data were ascertained by obtaining 12 different, validated 24-h dietary records from each participant during the 4 years after inclusion in the study. The dietary questionnaire came with instruction manuals for coding foods, including photographs for selecting seven different portion sizes. A number of clinical and biochemical indices were determined when participants joined the study, and preliminary data on the association of these indices with level and type of dietary fibre intake was presented.
For men, a number of significant differences were found between the participants in the highest and lowest quintiles of total dietary fibre intake (>26.3 vs <15.2 g/day). Cereal fibre was most consistently associated with the studied variables, and the highest intake was significantly associated with a lower BMI, systolic blood pressure and plasma glucose (P=0.05). For women, the only significant difference between those participants with the highest vs the lowest quintile of total dietary fibre intake (>21.1 vs. <12.7 g/day) was observed for a lower BMI in those with the highest intake. For women, only vegetable fibre displayed an association (Table 5). To date, detailed clinical data and the incidence of cardiovascular events in relation to fibre intake from this study have not been published.
Ness et al. (2002) followed up a secondary prevention, randomized controlled trial in South Wales, UK, conducted by Burr et al. (1989). Between 1983 and 1987, Burr and co-workers allocated men under 70 years who had survived a myocardial infarction to receive fibre advice or not (other groups received fish advice/or not and fat advice/or not). Participants were encouraged to eat at least six slices of wholemeal bread per day, or an equivalent amount of cereal fibre from a mixture of wholemeal bread, high-fibre breakfast cereals and wheat bran. At 6 months, cereal fibre intake in the fibre advice group was 19 and 9 g/day in those not given fibre advice. At 2 years, fibre intake was 17 and 9 g/day, respectively. After 2 years, the results were unclear, in that total mortality increased (not statistically significant) in those who were advised to eat more wholegrain. There were a number of limitations to this study including compliance, short follow-up and no objective check of fibre intake (Truswell, 2002).
When Ness and co-workers followed up (between 1999 and 2000) the surviving participants from this cohort (n=443 who received fibre advice, n=436 who did not), they found no differences in self-reported weight, current smoking, medication use, aspirin use or dietary supplement use. At follow-up, those allocated to receive fibre advice had a significantly higher fibre intake, although the absolute difference was small (6.2 (s.d. 5.4) vs 5.0 (4.4) g/day, P<0.01). CHD mortality increased in those given fibre advice in the first two follow-up time periods, as reported in Burr et al.'s paper, but reduced in the later time periods such that there was no long-term effect on survival (Table 6).
In following up Burr et al.'s cohort, Ness and co-workers acknowledge that dietary advice stopped after 2 years and recent data were collected when around half of the original participants had died. The researchers cannot discount the possibility that the diets of those who survived are different from those who did not. Follow-up diet was assessed with a limited number of questions focused on fibre intake; questionnaire data were not collected on other aspects of current diet and it is thus possible that important differences in diet were not detected. The authors concluded that the failure of longer term data to confirm the initial but nonsignificant increased risk of CHD deaths in men who were given advice to eat more cereal fibre suggested that increased fibre intake was not harmful. The results do, however, suggest that increasing cereal fibre intake by a small degree does not confer any immediate survival advantage. It must be borne in mind, however, that this was a secondary prevention trial of CHD, in which the influence of environmental factors on the pathological process is likely to be weaker than in a primary prevention trial (Truswell, 2002).
Jacobs et al. (2000) argued that the potential health effects of cereal fibre may depend on its source. They hypothesized that the nutrients eaten with fibre (antioxidants and minerals) play a role in protection against chronic disease and that the results of studies of cereal fibre intake would depend on the mix of fibre from wholegrain compared to refined grain. They examined all-cause, CHD and cancer mortality data from 11 040 participants participating in the prospective Iowa Women's Health Study. They compared two groups of women, matched to consume ∼6 g/day of total cereal fibre. One group (n=3559) consumed on average 77% of their cereal fibre from refined grain sources (<3.6 g/2000 kcal of wholegrain fibre and 3.6 g/2000 kcal of refined grain fibre). The second group (n=7481) consumed on average 71% of their grain fibre from wholegrain sources (3.6–6.0 g/2000 kcal wholegrain fibre and <3.6 g/2000 kcal refined grain fibre). This group was followed from baseline in 1986 to December 1997, during which time 1341 total deaths occurred (274 from CHD; 247 with complete data included in multivariate analysis). After adjusting for healthy lifestyle characteristics, a multivariate regression showed that women who consumed on average 1.9 g refined grain fibre/2000 kcal and 4.7 g wholegrain fibre/2000 kcal had a 17% lower all-cause mortality rate (95% CI 0.73–0.94) and a nonsignificant 11% lower CHD rate (95% CI 0.66–1.20) than women who consumed predominantly refined grain fibre (4.5 g/2000 kcal but only 1.3 g wholegrain fibre/2000 kcal).
Despite these findings reflecting statistically nonsignificant reductions for CHD deaths, the authors state that women who were excluded from the matched design because they consumed more than 6 g/2000 kcal of wholegrain fibre (generally accompanied by small amounts of refined grain fibre) had statistically significantly reduced total, CHD and all-cause mortality rates, compared to women who ate little wholegrain fibre.
Jacobs and co-workers thereby demonstrated that a similar amount of total cereal fibre had dissimilar associations with total mortality, depending on whether the fibre came from foods that contained primarily wholegrain or refined grain. They suggested that consumption of the plant constituents that are botanically linked to fibre may confer health benefits above and beyond effects of the fibre itself.
These results highlight the emerging ‘wholegrain story’ (Anderson et al., 2000) which argues that health benefits stem from more than just the fibre; the wholegrain is nutritionally more important because it delivers a whole package of nutrients and phytoprotective substances that may work synergistically to promote health (Anderson et al., 2000; Richardson, 2000; Jacobs and Steffen, 2003; Jacobs and Gallaher, 2004; Slavin, 2004).
Thus although fibre is an important component of wholegrains, many studies have not shown an independent effect of fibre alone on CHD events and/or deaths. Thus in terms of CHD prevention, fibre is probably best obtained from wholegrain sources.
Soluble fibre and CHD
There is little epidemiological data on the association between soluble fibre and heart disease even though soluble fibre clearly lowers serum cholesterol concentration and also lowers glucose in diabetics. A cholesterol-lowering claim is allowed for oat-derived beta glucan (rolled oats, oat bran, whole oat flour, the soluble fraction of alpha-amylase-hydrolysed oat bran or whole oat flour with a beta-glucan content up to 10 percent) by the United States Food and Drug Administration (USFDA, 1997).
In a study of female health professionals, fibre was associated with protection from heart disease (Liu et al., 2002). The association with insoluble fibre was stronger than for soluble fibre although neither separately was significant. The association with total fibre was insignificant after full adjustment.
In a secondary prevention study in men with angina, Burr et al. (2003) found no effects of advice to increase fruit and vegetables and oats on total or cardiovascular mortality although compliance was not well assessed and, in the small number of people actually assessed, it was low.
A meta analysis of 11 clinical intervention trials (Anderson and Major, 2002) involving legumes (other than soy beans) found overall a 6.2% lowering of LDL cholesterol and 22% lowering of triglycerides. The hypocholesterolaemic effects of legumes appear related, in estimated order of importance, to soluble dietary fibre, vegetable protein, oligosaccharides, isoflavones, phospholipids and fatty acids, and saponins.
Oat beta glucan may be ineffective at lowering serum cholesterol concentration when baked into bread and cookies. Jenkins et al. (2002b) confirmed that 4 servings/day of beta glucan or psyllium delivering 8 g/day of soluble fibre lowered total cholesterol by 2.1% which would have a useful population rather than individual effect on heart disease risk.
Soluble fibre clearly lowers cholesterol to a small and significant degree and one would expect that this would reduce CHD events.
Phenolic compounds, antioxidant activity and disease
The primary phenols in cereals are flavonoids and phenolic acids (Adom and Liu, 2002; Adom et al., 2003), which are found predominantly in the bran (Slavin et al., 2000). Corn had the highest total phenolic content (15.55±0.60 μmol of gallic acid equiv/g of grain) of the grains tested, followed by wheat (7.99±0.39), oats (6.53±0.19) and rice (5.56±0.17). The major portion of phenolics in grains existed in the bound form (85% in corn, 75% in oats and wheat and 62% in rice). Ferulic acid was the major phenolic compound, with free, soluble-conjugated and bound ferulic acids present in the ratio of 0.1:1:100. Corn had the highest total antioxidant activity (181.42±0.86 μmol of vitamin C equiv/g of grain), followed by wheat (76.70±1.38), oats (74.67±1.49) and rice (55.77±1.62).
Bound phytochemicals (mostly phenolics) were the major contributors to the total antioxidant activity: 90% in wheat, 87% in corn, 71% in rice and 58% in oats. Bound phytochemicals survived stomach and intestinal digestion to reach the colon. This may partly explain the mechanism by which grain consumption contributes to the prevention of colon cancer and other digestive diseases. Wheat contains a total flavonoid content of 105–142 μmol gallic acid equivalents/100 g. Barley contains about 4–12 times as many polyphenols per 100 g as red wine, which is similar to soy. However, the amount of polyphenol per serve would be quite similar.
Kris-Etherton et al. (2002) report that several population studies have found an inverse association between flavonoid intake and risk of coronary disease. In the Zutphen Elderly Study, a high intake of flavonoids (around 30 mg/day from tea, onions and apples predominantly) was associated with approximately a 50% reduction in CHD mortality rate compared with individuals who had a low flavonoid intake (<19 mg/day) (Hertog et al., 1993). Rimm et al. (1996b) did not find an association between new diagnosis of nonfatal myocardial infarction in 496 participants and flavonol and flavone intake. However, they did report a significant association (RR=0.63) between flavonoid intake and subsequent coronary mortality in 4814 men with existing CHD.
For those studies that have reported an association, putative mechanisms of action include inhibition of LDL oxidation and inhibition of platelet aggregation and adhesion. One measure of antioxidant activity is expressed as Trolox or water-soluble vitamin E equivalents (TE). Miller et al. (2000) analysed the antioxidant content of cereals, fruit and vegetables and found that breakfast cereals are equal or higher than any of the vegetables and the majority of fruits. A 41 g average serving of ready-to-eat breakfast cereal provided 1120 TE, while an 86 g average serving of vegetables or fruits provided 380 and 1020 TE, respectively. Wholegrain wheat- and oat-based products were on average higher in antioxidants than products from refined rice or corn ingredients. Even so, the rice and corn products had a higher TE than nearly all vegetables and a majority of fruits.
Using a different analytical method, Halvorsen et al. (2002) in Norway analysed the antioxidant content of cereals, fruits and vegetables and found that wholemeal flours of barley, common millet and oats contained the highest levels of antioxidants among the cereals. White flour contained between 23 and 54% of the antioxidant content of the wholemeal variants of the various cereals. Wholemeal wheat flour had a greater mean antioxidant concentration (0.33 mmol/100 g) than apples, bananas, pears, melon and watermelon, but less than grapes (1.45 mmol/100 g), oranges, plums, pineapples, lemons and apricots (among other fruits). Among vegetables, wholemeal wheat flour had a greater antioxidant concentration than garlic, cabbage and squash but less than chilli pepper (2.46 mmol/100 g), red cabbage, spinach, onion and broccoli (among other vegetables).
Wholemeal grain products have strong antioxidant activity. Whether this is beneficial in CHD prevention is not known.
Unsaturated fatty acids and wholegrains
Wholegrain wheat contains about 3% lipids and wholegrain oats around 7.5%, of which there are approximately equal amounts of oleic and linoleic acid and 0.1–0.2% linolenic acid. Oleic acid and linoleic acid are associated with lowering total cholesterol and LDL-cholesterol (McPherson and Spiller, 1995) but the amount in cereals would be unlikely to alter plasma lipids.
Phytoestrogens and wholegrains
Phytoestrogens are divided into three classes: isoflavonones, coumestans and lignans. The isoflavones, genistein and daidzein, are found predominantly in soybeans. The primary source of dietary lignans is flaxseed oil, but it is also found in varying concentrations in soybeans, seaweed, wholegrains (particularly rye), fruits and vegetables (Kris-Etherton et al., 2002). Eating wholemeal rye bread at least doubles serum and urine enterolactone concentrations compared with white wheat bread (Juntunen et al., 2000).
In a Finnish cross-sectional, 24-h dietary recall survey of 2852 men and women aged 25–64 years, lignan intake was positively associated with serum enterolactone concentrations. The mean serum enterolactone concentration in the highest quintile of lignan intake was 50% higher than in the lowest quintile (P<0.001) (Kilkkinen et al., 2003).
Jacobs et al. (2002) conducted a crossover feeding study in which 11 overweight, non-diabetic men and women ate, in random order, wholegrain foods or refined grain foods. Serum enterolactone concentrations were higher when eating the wholegrain diet than with the refined grain diet by 6.2 (within person s.e. 1.7) nmol/l (P=0.0008). Plasma enterolactone concentrations were inversely related to plasma F2 isoprostane – an oxidized lipid believed to be a good marker of oxidative stress and defence (Vanharanta et al., 2002).
In the Kuopio Ischaemic Heart Disease Risk Factor Study, multivariate analyses showed significant associations between elevated serum enterolactone concentration and reduced risk of CHD- and CVD-related mortality, but weaker associations in relation to all-cause mortality. In the 1889 men aged 42–60 years, followed for an average of 12.2 years, 70 CHD-related, 103 CVD-related, and 242 all-cause deaths occurred in participants free of prior CVD. In the Cox proportional hazards regression model adjusting for the most potent confounding factors, the risk of CHD-related (P=0.03 for trend) and CVD-related (P=0.04 for trend) death decreased linearly across quartiles of serum enterolactone concentration (Vanharanta et al., 2003).
Wholegrains, particularly from rye, contain phytoestrogens but whether this is beneficial for CHD prevention is not yet clear.
Wholegrains as part of a prudent diet
An emerging trend in dietary analysis is to consider the whole dietary pattern, rather than the effect of individual nutrients or foods, on health (Hu, 2002). This approach is considered to more closely approximate the ‘real world’, where people eat meals consisting of a variety of foods, rather than isolated nutrients. Thus, complicated interactions among nutrients and non-nutrient substances can be taken into account.
Hu (2003) reported on two cohort studies – the Nurses' Health Study (Fung et al., 2001b) and the Health Professionals' Follow-up Study (Hu et al., 2000). Factor analysis was used to examine associations between major eating patterns and risk of CHD. Diet was identified as a ‘prudent’ pattern (higher intakes of fruit, vegetables, legumes, fish, poultry and wholegrains) or a ‘Western’ pattern (higher intakes of red and processed meats, sweets and desserts, French fries and refined grains).
In the Nurses' Health Study, after coronary risk factors were adjusted for, the prudent diet score was associated with an RR of 0.76 (95% CI, 0.60–0.98; P for trend=0.03) comparing the highest with the lowest quintile. The same quintile comparison yielded an RR of 1.46 (95% CI 1.07–1.99; P for trend=0.02) for the Western pattern.
Similar associations were found in the Health Professionals' Follow-up Study. After adjusting for confounders, the RRs from the lowest to highest quintiles of the prudent pattern score were 1.0, 0.87, 0.79, 0.75 and 0.70 (95% CI, 0.56–0.86; P for trend <0.0001). In contrast, the RRs across the same quintiles of the Western pattern score were 1.0, 1.21, 1.36, 1.40, 1.64 (95% CI 1.24–2.17; P for trend <0.0001).
Fung et al. (2001a) also examined the relation between major eating patterns and biochemical markers of coronary artery disease and obesity. They observed inverse correlations between the prudent pattern score and both fasting insulin (−0.25, P<0.05) and homocysteine (−0.20, P<0.01); a positive correlation was observed with folate (0.28, P<0.0001).
Wholegrains are a source of valuable nutrients (Table 1). It is clear that wholegrains are protective in relation to CHD, whereas refined grains may have an adverse effect, although the data for the latter is not as consistent.
In a study of the 9632 men and women who participated in the First National Health and Nutrition Examination Survey (NHANES 1) Epidemiologic Follow-up Study (NHEFS) and who were free of cardiovascular disease at their baseline examination, Bazzano et al. (2001) found that legume consumption was significantly and inversely associated with risk of CHD (P=0.002 for trend) and CVD (P=0.02 for trend) after adjustment for established CVD risk factors. Over an average of 19 years of follow-up, 1802 incident cases of CHD and 3680 incident cases of CVD were found. Legume consumption four times or more per week, compared with less than once a week, was associated with a 22% lower risk of CHD (relative risk, 0.78; 95% CI, 0.68–0.90) and an 11% lower risk of CVD (relative risk, 0.89; 95% CI 0.80–0.98). Frequency of legume intake was estimated using a 3-month FFQ; participants were asked how often ‘dry beans and peas like pinto beans, red beans, black-eye [sic] peas, peanuts and peanut butter’ were usually consumed in the past 3 months excluding periods of illness or dieting. Information on portion size was not collected.
In an ecological study relating soy consumption to heart disease mortality in 47 prefectures in Japan, a weak inverse correlation was found in women only after adjusting for covariates (Nagata, 2000). The study used 3 day food records from 6000 households.
Sasazuki et al. (2001) examined a total of 632 cases aged 40–79 years, living in Fukuoka City or adjacent areas, with their first episode of acute myocardial infarction (AMI) and 1214 controls matched for age, sex and residence. In women only, tofu consumption was inversely related to the risk of AMI; relative risks for eating tofu <2, 2–3 and 4+ times per week were 1.0, 0.8 and 0.5, respectively, after adjustment for non-dietary factors (P for trend=0.01). Further analysis of the independent relationship of three food groups (fruit, fish and tofu) with AMI did not change the inverse association with tofu observed for women.
In Chinese women in the Shanghai Women's Health Study, there was a clear monotonic dose–response relationship between soy food intake and risk of total CHD (P for trend=0.003) with an adjusted RR of 0.25 (95% CI, 0.10–0.63) observed for women in the highest vs the lowest quartile of total soy protein intake. After a mean of 2.5 years (162 277 person-years) of follow-up, 62 incident cases of CHD (43 nonfatal myocardial infarctions and 19 CHD deaths) were seen. The inverse association was more pronounced for nonfatal myocardial infarction (RR=0.14; 95% CI, 0.04–0.48 for the highest vs the lowest quartile of intake; P for trend=0.001) (Zhang et al., 2003).
Interventions with legumes
Wangen et al. (2001) found that 132 mg of isoflavones per day, from high isoflavone soy protein, given for 3 months lowered LDL cholesterol by 6.5% compared with low isoflavone isolated soy protein. Study subjects were 18 normocholesterolemic and hypercholesterolemic postmenopausal women. A contrary finding was demonstrated by Sanders et al. (2002) who contrasted a soy protein diet containing 56 mg/day of isoflavones against one containing 2 mg/day for 17 days in 22 healthy young normolipidemic subjects in a crossover design. Only HDL cholesterol rose, by 4%, with isoflavones.
Jenkins et al. (2002a) also found no effect on lipids of isoflavone-enriched soy protein compared with isoflavone-poor soy protein (73 vs 10 mg) in 41 subjects in a crossover study, but did find that 52 g of soy protein lowered total cholesterol by 4% and also lowered systolic blood pressure in men compared with a dairy-rich low-fat diet.
Zhan and Ho (2005) performed a meta-analysis of 23 recent trials involving 1381 subjects using soy with intact isoflavones and found an LDL cholesterol-lowering of 5%. The effect was independent of initial cholesterol level. Soy containing 80 mg of isoflavone/day was more effective than lower amounts. Ashton and Ball (2000) replaced 150 g of meat per day with 290 g of tofu in 42 men in a randomized crossover trial with periods of 1 month each. Total cholesterol and triglyceride and the lag period of LDL oxidation were significantly lowered by tofu. Jang et al. (2001) randomized 76 men with heart disease to either refined rice or a wholegrain/legume powder for 16 weeks and showed that the wholegrain/legume diet reduced plasma insulin and glucose by 14% and malondialdehyde and F2 isoprostane in diabetic subjects by 28%. The latter two measures show that legumes contain bioavailable antioxidants.
Hale et al. (2002) tested 80 mg of a soy isoflavone concentrate in 29 healthy menopausal women and demonstrated no effects on endothelial function compared with a placebo, although there is some evidence that soy protein isolate (regardless of isoflavone content) contributes to total plasma antioxidant status, which potentially might protect against heart disease (Swain et al., 2002).
Azadbakht et al. (2003) replaced half of the animal protein (total protein 0.8 g/kg) with soy in the diet of 14 patients with nephropathy for 7 weeks and found that the soy lowered total cholesterol, triglyceride and LDL cholesterol as well as urea nitrogen and urine protein levels.
Overall, the recent data have confirmed that soy protein (at least 25 g/day) can lower LDL cholesterol to a small but significant degree (5–6%). Isoflavones may play a small role in cholesterol-lowering but may also play a separate role in soy's effects in lowering heart disease rates, as in the three studies quoted. The USFDA has allowed a health claim for soy (United States Food and Drug Administration (USFDA), 1999).
Glycaemic index, glycaemic load and vascular disease
Glycaemic index is the glucose excursion above baseline over 2 h of 50 g of available carbohydrate in comparison with either 50 g of glucose or 50 g of available carbohydrate from white bread expressed as a percentage. Glycaemic load (GL) is the product of amount of carbohydrate per serve and glycaemic index (GI). Liu et al. (2000b) examined the association between GL and heart disease in 75 521 women in the Nurses Health Study. After 10 years of follow up, 761 incident cases were found and it was observed that the risk of heart disease was twice as great in the women consuming the highest GL diet. The association was only seen in women with a BMI of greater than 23 kg/m2.
Carbohydrate intake varied from 144 to 226 g per day and GI from 72 to 80 g per day (averaged across all carbohydrate-containing foods). Fibre intake varied from 14 to 18 g/day and cereal fibre from 3 to 5 g/day. The women who consumed a high GL diet ate more total fibre, more cereal fibre and less fat but these components did not blunt the effect of the high GI carbohydrates. Carbohydrate amount was not related to disease risk but GI was related. The biggest contributors to the GL in the study by Liu et al. (2000b) were potato (8%) and cold breakfast cereal (4%).
In relation to stroke, carbohydrate itself was a risk factor, as was GL but the latter was only significant in women with a BMI greater than 25 kg/m2 (Oh et al., 2005). Cereal fibre was inversely associated with risk of both haemorrhagic and thrombotic stroke, with a 29–33% reduction in risk for women in the highest quintile of intake compared to those in the lowest.
Possible mechanisms for the increased risk resulting from a high GL include reduction in HDL cholesterol (Ford and Liu, 2001; Liu et al., 2001; Amano et al., 2004), increased fasting plasma triglycerides (Liu et al., 2001; Amano et al., 2004), higher fasting insulin (Amano et al., 2004), increased CRP (Liu et al., 2002) and glucose intolerance (Schulze et al., 2004), with epidemiological associations being shown between risk factors and either the GL of the diet or GI. CRP is a marker of vascular disease but is also believed to play an active role in the heart disease process (Zwaka et al., 2001; Verma et al., 2002).
In dietary interventions, exchanging carbohydrate for fat is well known to lower HDL cholesterol, increase fasting and postprandial triglycerides and to increase fasting insulin (Mensink and Katan, 1992; Garg et al., 1994; Turley et al., 1998; Mensink et al., 2003), but whether this is adverse in terms of cardiovascular risk is unknown and the only data in support of the latter are the Nurses Health Study referred to above (Liu et al. (2000b).
Glycaemic index has been shown to be unrelated to cardiovascular mortality in the Zutphen study (van Dam et al., 2000) and in a Cochrane meta-analysis (Kelly et al., 2004) of 15 interventions looking at the GI of carbohydrates. The latter concluded that the evidence to prove GI was related to cardiovascular risk factors or improvement in glucose control was weak. This conclusion is critical as many wholegrain cereal products are wholemeal (ie they contain milled rather than intact grains) and have a high GI (Foster-Powell et al., 2002), whereas intact grains have a low GI. In the studies relating wholegrain consumption to protection from vascular disease, the highest quintile of wholegrain intake usually has a high GL (Jensen et al., 2004) so the benefit of wholegrain is seen despite the high GL.
the intake of wholegrain foods clearly protects against heart disease and stroke but the exact mechanism is not clear. Fibre, magnesium, folate and vitamins B6 and vitamin E may be important.
The intake of high GI carbohydrates (from both grain and non-grain sources) in large amounts is associated with an increased risk of heart disease in overweight and obese women even when fibre intake is high, but this requires further confirmation in normal-weight women.
Recommendation: Carbohydrate-rich foods should be wholegrain and if they are not, then the lowest GI product available should be consumed. Glycaemic index is largely irrelevant for foods that contain small amounts of carbohydrate per serve (such as most vegetables).
This study was supported in part by BRI Australia Ltd.
About this article
Current Cardiology Reports (2015)