You are on page 1of 14

European Journal of Clinical Nutrition (2002) 56, 114 2002 Nature Publishing Group All rights reserved 09543007/02

2 $25.00
www.nature.com/ejcn

REVIEW Cereal grains and coronary heart disease


AS Truswell1*
1

Biochemistry Department, University of Sydney, Sydney, New South Wales, Australia

Cereal grains and their products provide around 30% of total energy intake in British adults, (much more than any of the other major food groups). Coronary heart disease (CHD) is the largest single cause of death in Britain and many other Western countries. This review examines the question whether there is a relation between cereal consumption and CHD. Several of the nutrients in cereals have known potential for reducing risk factors for CHD: the linoleic acid, bre, vitamin E, selenium and folate. Cereals also contain phytoestrogens of the lignan family and several phenolic acids with antioxidant properties. Processing generally reduces the content of these nutrients and bioprotective substances. Although cereals at the farm gate are very low in salt, processed cereal foods, eg bread and some breakfast cereals, are high-salt foods and thus could contribute to raising blood pressure. Human experiments have clearly shown that oat bre tends to lower plasma total and LDL cholesterol but wheat bre does not. Rice bran and barley may also lower cholesterol but most people do not eat enough barley to have an effect. Cereal foods with low glycaemic index such as pasta and oats are benecial for people with diabetes and might lower plasma lipids. Between 1996 and 2001 an accumulation of ve very large cohort studies in the USA, Finland and Norway have all reported that subjects consuming relatively large amounts of whole grain cereals have signicantly lower rates of CHD. This conrms an earlier report from a small British cohort. The protective effect does not seem to be due to cholesterol-lowering. While cohort studies have shown this consistent protective effect of whole grain cereals, there has been (only one) randomised controlled secondary prevention trial of advice to eat more cereal bre. In this there was no reduction of the rate of reinfarction. The trial had some weaknesses, eg there were eight different diets, compliance was not checked objectively, and duration was for only 2 y. It appears valid to make health claims (as now permitted by the US FDA) that whole grain cereal foods and oat meal or bran may reduce the risk of CHD. European Journal of Clinical Nutrition (2002) 56, 1 14. DOI: 10.1038=sj=ejcn=1601283 Keywords: cereal grains; dietary bre; coronary heart disease; plasma cholesterol; prospective cohort studies

Introduction
Cereals are grasses, monocotyledonous plants. Cereal grains occupy the largest box (USA) or most of the equal largest box (Australia) in pictorial food guides for nutrition education. Eat plenty of cereal foods, preferably wholegrain and without added fat, salt or sugar can be found in several sets of national dietary guidelines. Coronary heart disease (CHD; ICD 410 414) is still the disease that causes more deaths than any other in Europe. The question of the relation between the cereal food group and CHD is therefore impor-

tant for public health nutrition. Is there evidence that all or some of the cereals group affect the risk of CHD? If there is, then it follows that if people increase their intake of this (these) cereal food(s) it would be a contribution to further reducing the national burden of CHD. Cereal foods can be subdivided in several ways: (i) into cereals and cereal products OR cereal-based products or dishes; (ii) into botanical species wheat, rice, maize, barley, oats, rye, etc; (iii) into whole grain cereals (and products) or rened cereals; (iv) into the many different ways of eating cereals and products for wheat these include bread, breakfast cereals, pasta, biscuits, cakes, mufns, pancakes, pizza, as thickener, etc; rice is mostly eaten boiled, as are oats

*Correspondence: AS Truswell, Biochemistry Department, University of Sydney, Sydney, NSW 2006, Australia. Guarantor: AS Truswell. Contributors: AS Truswell. Received 13 February 2001; revised 30 May 2001; accepted 6 June 2001

Cereal grains and CHD AS Truswell

2
(which are also eaten raw in muesli). Maize is mostly eaten as popcorn or cornakes; sweet corn is also eaten as a vegetable. Rye makes a less common bread, and most of the barley is used to make beer or whisky. When cereal is mixed with other foods its components are diluted and their effects on health become very difcult to discern. This review considers cereals and CHD in three sections: (A) What food composition analysis suggests; (B) Experiments to see the effect of cereals or concentrated cereal components on risk factors for CHD; and (C) Epidemiological studies, relating intake of cereals to outcome of coronary heart disease. Several of these nutrients in cereals have the known potential, if in adequate amounts, for reducing risk factors for CHD. The predominantly polyunsaturated oil (50% linoleic acid) and possibly some of the bre could lower plasma LDL-cholesterol, the vitamin E and selenium are antioxidants, and folic acid might lower plasma homocysteine. Processing can make large changes in the composition of cereal products. Rening reduces many of the nutrients

(A) Components of cereals nutrients, bre and phytochemicals


Nutrients As an important part of a balanced diet (ie one that provides all the food groups in the nutrition education pyramid or plate and all the recommended dietary intakes) the cereal group provides important amounts of most nutrients (Table 1) but not all. Man cannot live by bread alone. Against human requirements cereals are naturally lacking in calcium, vitamin A, vitamin C and vitamin B12 (and also in sodium before processing; Table 1). An example of the contribution to the diet of major nutrients in cereals is illustrated by Figure 1 from Hansen (1973).

Figure 1 Nutrient density (as dened here) is the ratio of the content of a nutrient (expressed as % of recommended nutrient intake (or RDA)) to energy content of the food (expressed as % of a standard energy intake). Thus if one were to obtain all ones dietary energy from bread and cereals, requirements for protein and iron would be met but the diet would not supply enough calcium; vitamins A and C would be very inadequate while the vitamin B1 (thiamin) would be twice the requirement. (Ref Hansen, 1973).

Table 1 Nutrients which cereals provide in important amounts. (Figures in brackets are % total intake from cereals in Dietary and Nutritional Survey of British Adults (Gregory et al, 1990; Ministry of Agriculture, Fisheries & Food, 1994))
Available carbohydrate (46%), mostly starch* Protein (23%) (with lysine the rst limiting amino acid) Oil (rich in o-6 linoleic acid (22%), and some phytosterols) Fibre (47%)-insoluble and soluble (and resistant starch) Potassium (14%) Iron (42%) (low availability) Phosphorus (25%), magnesium (30%) Zinc (22%), manganese Selenium (depending on soil content) Thiamin (and more in fortied bread and breakfast cereals) Riboavin (including fortied breakfast cereals) (22%) Niacin (low bioavailability) (28% as niacin equivalents) Vitamin B-6 (? Low bioavailability) (19%) Pantothenate (17%) and biotin (23%) Folate (21%) (before extensive fortication) Vitamin E (21%).

A predominantly cereal diet is low or lacking in: Calcium Vitamin A (retinol) & b-carotene (except yellow maize) Vitamin C Vitamin B-12 Sodium chloride**- before processing
*A variable minority of this is resistant to human amylases. **Cereals as harvested are very low in sodium and chloride, eg 3 mg Na=100 g wholemeal our. But after processing many cereal products are high sodium foods, eg 520 mg=100 g in white bread (Mugford et al, 1996).

European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

3
Table 2 Nutrients (per 100 g) compared in wholemeal and white ours, from McCance & Widdowsons The Composition of Foods, 5th edition (Holland et al, 1991)
Wholemeal our
Protein (g) Fat (g) Carbohydrate (g) Fibre (g) Sodium (mg) Potassium (mg) Magnesium (mg) Phosphorus (mg) Iron (mg) Zinc (mg) Manganese (mg) Selenium (mg) Thiamin (mg) Riboavin (mg) Niacin (mg) Vitamin B-6 (mg) Folate (mg) Vitamin E (mg) *Before processing. 12.7 2.2 63.9 8.6 3. * 340. 120. 320. 3.9 2.9 3.1 53. 0.46 0.09 5.7 0.50 57. 1.4

would not benet the circulatory system unless they could be absorbed in adequate amounts.

White our
9.4 1.3 77.7 3.6 3. * 150. 20. 110. 1.5 0.6 0.6 4. 0.10 0.03 1.7 0.15 22. 0.3

Retained % after rening


74% 59% 122% 42% 100% 44% 17% 34% 38% 21% 19% 8% 22% 33% 30% 30% 39% 21%

(B) Experiments on effects of cereal foods or fractions on risk factors for CHD
It is very unlikely that a randomised controlled human trial could be achieved in which CHD outcomes are recorded after people have taken either a high cereal diet or a low cereal diet in a randomised controlled trial for years, with all other dietary and other variables kept similar. It is, however, relatively simple to give a particular food, or fraction of it, to subjects under controlled conditions for a few weeks and measure if there is an effect on, say, plasma cholesterol, one of the major risk factors for CHD. If the particular food consistently and signicantly lowers plasma LDL-cholesterol, it is reasonable to expect that the food may help to reduce the risk of CHD. It is on the basis of such short-term, safe and comfortable human experiments that the US FDA (but no other country at present) allows a health claim that oatmeal or oat bran (as part of a healthy diet) may reduce the risk of CHD.

which are concentrated in the germ, aleurone layer, scutellum or outer coat. Table 2 compares the content of important nutrients in wholemeal wheatour and white our. All the nutrients which could be protective against CHD are reduced, linoleic acid and bre and folate by about half and selenium and vitamin E by much more. The other way in which processing affects composition is by addition of micronutrients when breakfast cereals, bread, etc are fortied with some vitamins or minerals. The added nutrients are usually calcium or iron, and several B vitamins (thiamin, riboavin, niacin, folic acid).

Phytochemicals Cereals contain phytoestrogens of the lignan family (Figure 2). Plant lignans are changed by intestinal bacteria to mammalian lignans. They include: secoisolariciresinol diglycoside?enterdiol; matairesinol?enterolactone. In people who consume relatively large amounts of whole grain cereals these phytoestrogens in adults may have a protective effect against hormone-related cancers (the structure of enterodiol is similar to that of tamoxifen). It is possible that, via a hormonal effect on HDL, there could be some inuence on risk of CHD (Thompson, 1993), but there is no clear evidence for this at present. Other likely candidates for some protective potential against CHD are phenolic acids, found in the outer layers of cereal grains, which are antioxidants in vitro (Thompson, 1993). The most abundant are ferulic acid, vanillic acid, pcoumaric acid, protocatechuic acid and caffeic acid (Figure 3). While vitamin E, another antioxidant, is fairly well absorbed in humans, these phenolic acids in cereal brans

Wheat bre and plasma cholesterol{* For researchers setting out to test the dietary bre hypothesis in the early 1970s this was the obvious type of bre to experiment with. Hugh Trowell (1975) had reasoned that wheat bre would lower plasma cholesterol. It is available in concentrated, fairly reproducible yet natural form as bran or whole grain wheat breads, which can be readily compared in controlled experiments with isocaloric low-bre white bread. There are at least 34 published reports of controlled human experiments in which plasma cholesterol was measured when subjects ate extra wheat bre (Table 3). In 27 of these experiments plasma total cholesterol did not go down it even rose signicantly in two trials (van Dokkum, 1978; Stasse-Wolthuis, 1979). Most of these 27 experiments were well designed, with control test control or crossover design or parallel control groups, and some were of long duration. In some studies all food was provided for the subjects in a metabolic unit (Jenkins et al, 1975; Kay & Truswell, 1977; Raymond et al, 1977; Stasse-Wolthuis et al, 1980). In the best of these reports with negative results, frequent measurements were made of plasma cholesterol concentrations. Five of the researchers in Table 3 who could not demonstrate an effect of wheat bre on plasma cholesterol did nd a cholesterol-lowering effect with pectin, using similar methods. By contrast, in the seven papers reporting lower plasma cholesterol on wheat bre the subjects total less than one-quarter of those in the 27 negative reports. Several used only one-way designs (control test) and only one
*All references for this section and Table 3 are in Truswell and Beynen (1992).

European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

Figure 2

LIGNANS, phytoestrogens found in cereals. Plant lignans on left and forms produced by intestinal bacteria on right.

plasma cholesterol per feeding period. The subjects were outpatients, and in some trials, they ate ad libitum. In some trials the plasma cholesterols were measured in the routine hospital clinical biochemistry laboratory. In one interesting paper in Table 3, Munoz et al (1979) reported reduced plasma cholesterol when subjects ate bran from hard red spring wheat but not with ordinary soft white wheat bran. Except for the possibility this raises that there might be some reproducible super bran it is clear that wheat bre does not have a cholesterol-lowering effect. In recent years wheat bran has been used as the placebo control in trials on the effect of guar gum and of oat bran on plasma cholesterol. It was still possible that wheat bre might raise plasma HDL-cholesterol, even if it does not lower total cholesterol. This had been reported by OMoore et al (1978) and LindEuropean Journal of Clinical Nutrition

garde and Larsson (1984). In the earlier trials with wheat bre only total cholesterol was measured, not lipoprotein fractions. However in most papers that do report HDLcholesterol it has not been changed signicantly by wheat bre (McDougall et al, 1978; Dixon, 1978; Stasse-Wolthuis et al, 1979; Flanagan et al, 1980; Gariot et al, 1986; Kestin et al, 1990) and wheat bre has no consistent effect on fasting plasma triglycerides (Truswell & Kay, 1976).

Oat bre and plasma cholesterol In nice contrast to wheat bre, most researchers who have tried the effect of eating increased oat bre, either as rolled oats or as oat bran have found reductions of plasma total and LDL-cholesterol.

Cereal grains and CHD AS Truswell

Figure 3 Phenolic acids found in cereal brans.

Table 3 (References in Truswell & Beynen (1992)) Wheat bre and plasma total cholesterol
Reference
Eastwood (1969) Eastwood et al (1973) Heaton and Pomare (1974) Connell et al (1975) Truswell and Kay (1975); Kay and Truswell (1977a) Jenkins et al (1975a, b) Durrington et al (1975) Bremner et al (1975) Persson et al (1975, 1976) Brodribb and Humphreys (1976) Heaton et al (1976) Weinreich et al (1977) Mathur et al (1977) Rhodes et al (1977=8) Avgerinos et al (1977) Raymond et al (1977) McDougall et al (1978) Letchford et al (1978) Farrell et al (1978) OMoore et al (1978) Tarpila et al (1978) Dixon (1978) Henry et al (1978) Angelico et al (1978) Van Dokkum (1978) Watts et al (1978) Munoz et al (1979) Munoz et al (1979) Van Berge-Henegouwen et al (1979) Stasse-Wolthuis et al (1979, 1980) Flanagan et al (1980) Liebman et al (1983) Lindgarde and Larsson (1984) Gariot et al (1986) Kestin et al (1990)

Effects
0 0 0 0 0 0 0 0 77% 0 0 0 ; 0 720% 0 0 ; ; 0 0 0 0 0 : Sig 0 0 ; 710% : 0 0 0 0 0

Notes
28 monks; 14 g bran; 2 weeks 8 men; 16 g bran; 3 weeks 6 normal, 8 cholelithiasis; 38 g bran; 5 weeks 23 students; 28 57 g bran; 11 weeks 6 normals; 23 g bran 167 g whole wheat bread; 3 weeks 6 men; 30 g bran, 77 g bran products, WW bread; 3 weeks 12 men; 30 g bread; 6 weeks 5 men hypertriglyceridaemia; 50 g bran; 12 weeks 14 elderly; 20 g bran; 12 weeks 40 with diverticulosis; 24 g bran; 52 weeks 19 students; 180 g WW bread; 19 weeks 25 nurses; 24 g bran; 5 weeks 20 normal, 10 HL; 40 g bran; 4 weeks 6 normal; 90 g All Bran; 3 weeks (abstract only) 12 subjects; 72 g bran; 4 weeks metabolic ward 9 normal, 9 cholelithiasis; 50 g bran; 4 weeks 17 subjects, no dietary control (abstract only) Same group; 14 males; short experiment 12 irritable bowel; bran; 4 weeks 22 diverticulosis; 25 g bran; 26 52 weeks 16 normal; bran WW bread; 13 weeks Bran-enriched bread 8 normal, metabolic ward; bread bran cf white bread 11 normal; 30 g bran; 8 weeks 6 normal; soft white bran; 4 weeks 9 normal; hard red spring wheat bran; 4 weeks 7 males; 35 g bran; 4 weeks Signicantly : 16 subjects, all food provided; 38 g bran; 5 weeks 16 irritable bowel; bran; 12 weeks 20 overweight men; 0.5 g bran=kg; 6 weeks 12 men HC; 11 g conc wheat bre; 8 weeks 4 normal; 20 g bran; 7 weeks 24 men, mild HC; 35 g bran; 4 weeks

WW, whole wheat; HL, hyperlipidaemia; HC, hypercholesterolaemia.

European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

Table 4 Human experiments on effect of oat (bre) on plasma cholesterol


Authors
DeGroot et al Luyken et al Kretsch et al Judd & Truswell Kirby et al Anderson et al Anderson et al Storch et al Roth Leitzmann Van Horn et al Turnbull Leeds Gold Davidson Van Horn et al Hamilton et al Reynolds et al Birkeland et al Swain et al Denmark-Wahnefried et al Hegsted et al Kestin et al Leadbetter et al Van Horn et al Anderson et al Anderson et al Bremer et al Davidson et al

Year
1963 1965 1979 1981 1981 1984a 1984b 1984 1985 1986 1987 1988 1988 1989 1989 1990 1990 1990 1990 1990 1991 1991 1990 1991 1990 1991

n 21 95 6 10 8 HC 10 4 HC 12 18 208 17 HC 19 238 (half controls) 12 HC 43 HC 54 HC 20 71 HC 11 24 HC 40 HC 42 OM 38 C 12 HC 20 HC 12 HC 148

Oats
OM 140 g=d OM 50 g=d OB? OM 125 g=d OB 100 g=d OB 100 g=d OB 100 g=d OB 53 g=d OM 19 g=d OB 39 g=d OM 35 g=d OM 150 g=d OB 17 g=d OM 56 g=d OB 30 g=d OB 24 g=d OB OB 87 g=d OB 50 g=d OB 100 g=d Cf rice bran OB 95 g=d OB 30,60 or 90 g=d or more OM 56 g=d OB 25 g=d OB 110 g=d OB 45 g=d OM 28,56,84 g=d OB 28,56,84 g=d OB 56 g=d OB 88 g=d OB suppt OB 60 g=d 55 g OB OB 84 g=d OB 50 g=d OB concentrate 11 g b-glucan=d OB 123 g=d OB 135 g=d OB cereal (4.5 g bre) Oat gum concentrate in drink

Design

Plasma lipids
TC 710% TC 76% TC NS TC 78% TC 713% (not HDL or TG) TC 719% TC 723% TC 712% TC 2% TC 73% on OB or OM (OM sig) TC 75% (not HDL or TG) TG 75% TC 73% probably not sig TC 77% (not HDL or TG) TC 74% (not HDL or TG) TC 0 change TC 0 change LDL 73% NS TC 712% (not HDL) TC 77% similar fall on rice bran TC 75.6% (not HDL) NC TC 73% sig (not HDL) TC 75.4 sig (not HDL) TC 712.8 (or 78.4!) NS TC 72.7% on 56 g OM TC 79.5% on 56 g OB TC 72.3 cf start TC 74.5 cf WB TC 79% TC 711% OB, 75% WB hence 76% net TC 72.8 (NS) LDL 75.8% sig NC TC probably lower 254?239,254, 242 mg=dl NS TC 73.3% NS TC 74% sig LDL 75.5% sig TC 79% (not HDL or TG) TC 72.3% sig LDL 74.6% sig TC & LDL both 79%

Latin square Parallel

Keenan et al Spencer et al Kashtan et al Lepre Crane Mackay Ball Saudia et al Stewart et al Torronen et al Whyte et al Zhang et al Poulter et al Braaten et al

1991 1991 1992 1992 1992 1992 1992 1992 1992 1992 1993 1994

145 11 84 30 HC 39 mild HC 20 24 H 23 mild HC 9 ileostomy 59 19 HC

Cross over

OB oat bran; OM oat meal; HC hypercholesterolaemia; WB wheat bran TC total cholesterol; NS not signicant; NC no change; good design. The reference for the authors in this table can be found in (Truswell & Beynen, 1992; Truswell, 1995; Brown et al, 1999).

Table 4 shows that the percentage reductions of total cholesterol have ranged from 718% to 0 in subjects receiving rolled oats (oatmeal) or oat bran. In 28 of the 38 studies there was a signicant reduction of total cholesterol, with a similar percentage reduction of LDL-cholesterol. HDL-cholesterol and triglycerides did not change. In 10 studies, however, there was no signicant change in total or LDLcholesterol. Most of the trials showing no effect were conducted with free-living out patients. In the most publicised trial with no effect (Swain et al, 1990) the subjects had quite low plasma cholesterols (4.40 mmol=l), their fat and energy intakes were higher in the oat bran period and, although total cholesterols were only 3% lower on oat bran than on
European Journal of Clinical Nutrition

the comparison low-bre diet, the ratio of LDL=HDL was actually 9% lower (statistical signicance not reported). It is also remarkable that all four of the trials in New Zealand (Auckland, Christchurch and Dunedin) found no effect. This raises the possibility that the oat bran used in that country was atypical (? low in b-glucan). Very few of all the published trials reported the bre composition of the oat bran. On the other hand the trials with the most optimistic results (Anderson et al, 1984a and 1984b; Storch et al, 1984) had very small subject numbers and a suboptimal one-way design (control-oat bran). Another question is whether the predominantly polyunsaturated oil in oats (7 10 g=100) contributes to the cholesterol-lowering effect. In a number

Cereal grains and CHD AS Truswell

7
of the trials wheat bran or rened wheat or corn cereal were used in contrasting periods but they contain less oil. Judd & Truswell (1981) gave a balanced amount of polyunsaturated oil in the control period. Some of the later trials have used more concentrated oat gum, presumably low in oil. A further difculty in interpreting these results is that some authors (eg Keenan et al, 1991; Kashtan et al, 1992) report the difference between oat bre and start (day 0) as the effect, whereas a different result is seen if plasma cholesterol on oat bre is compared with that in the synchronous (control) wheat bre period. In the majority of reports (Table 4) there appears to be a tendency to a dose-response effect, though even in the bestdesigned trial (Davidson et al, 1991) this was rather inconsistent. Two estimates have been published of the size of the average cholesterol-lowering effect. Truswell (1995) noted that the unweighted mean plasma cholesterol reduction in 30 trials was 5.5%, with daily intakes of around 50 g oat bran. Brown et al (1999) calculated from 25 trials they accepted for meta-analysis that plasma cholesterol falls by 70.04 mmol=l per gram of soluble oats bre consumed. The trouble with this estimate is that the soluble bre content was not actually measured in most of the trials. But if we take the soluble bre in rolled oats at 4.5 g=100 g and in oat bran 7 g=100 g, then 50 g oatmeal should lower cholesterol by 0.09 mmol=l and 50 g of oat bran by 0.14 mmol=l. The latter is 3% of a starting plasma cholesterol of 5.0 mmol=l. The active part of oats total bre is the b-glucan. This is soluble bre, not brous, and it is the main component of oat gum. Concentrated oat gum lowers plasma cholesterol in humans (Braaten et al, 1994) and animals (Welch et al, 1988). Treatment of oat bran with a b-glucanase increased its free sugars, decreased its viscosity and abolished the cholesterollowering action in rats (Tietyen et al, 1990). Hot extrusion of oat bran or our has been reported to increase the solubility of its b-glucan (Hamilton et al, 1989). This may explain the 7 and 4% plasma cholesterol falls in well-controlled experiments with oat bran breakfast cereals taken two or three times a day (Hamilton et al, 1989; Reynolds et al, 1989). The most plausible mechanism of action of oats b-glucan is by its viscosity interfering with reabsorption of bile acids and hence producing a negative sterol balance (cholestyramine effect). In oat bre periods increased bile acids have been reported in faeces (Judd & Truswell; 1981; Kirby et al, 1981) and in ileostomy efuent (Zhang et al, 1993). ment the lipids in rice bran were balanced with oil of similar fatty acid pattern in the control (wheat bran) period. Rice bran does not contain b-glucan; its soluble bre content (hemicellulose) is about half that of oat bran. The oil, and particularly g-oryzanol in its unsaponiable fraction (Yoshino et al, 1989) may be a major reason for cholesterol reduction by large intakes of rice bran. The hemicelluloses of rice bran have been shown to bind bile acids (Normand et al, 1987). The properties of its bran, are unlikely to apply to rice as it is usually eaten, ie polished and white, unlike oats, which is eaten wholegrain. It is not only the vitamins in rice grain that are lost when it is rened and polished. Barley contains soluble bre as b-glucan. In a well controlled 11 week crossover trial with 21 mildly hypercholesterolaemic men in Adelaide (McIntosh et al, 1991), a diet high in barley bre gave 6% lower plasma total and LDL-cholesterol than a control diet with the same 38 g=day of bre from wheat. The subjects in this trial ate 170 g=day of barley foods, including barley bran and barley akes. There is at least one other report of a cholesterol-lowering effect of barley in humans (Newman et al, 1989). Barley, in western countries is nearly all used for brewing. As human food it is only a minor cereal. Its consumption is unlikely to make a substantial contribution to plasma cholesterol-lowering.

Cereal foods with low glycaemic index Whole grain cereal products: whole grain wheat breads not wholemeal (Jenkins et al, 1988) oatmeal, pasta, high amylose maize and rye bread have low glycaemic indices (Foster-Powell & Brand Miller, 1995). When they are eaten the rise of blood glucose is lower than after the same amount of carbohydrate from most breads, most breakfast cereals and most boiled rice. Diets with low glycaemic index foods predominating have fairly clear health advantages for people with diabetes (Brand Miller, 1994). As well as improved glycaemic control, plasma cholesterol declined in some of the controlled dietary trials (Brand Miller, 1994). Reductions of plasma cholesterol have also been reported in non-diabetic subjects on low, compared with high glycaemic index diets (Brand Miller, 1994).

Other cereal bres and plasma cholesterol Rice bran at intakes of 100 g=day reduced plasma total cholesterol by 7% (no signicant effect on other plasma lipids) (Hegsted et al, 1993). Sixty g=day of rice bran reduced LDLcholesterol slightly, increased HDL-cholesterol slightly, increased the total cholesterol=HDL-cholesterol ratio signicantly and lowered triglycerides (Kestin et al, 1990). Fifteen and 30 g=day rice bran reduced plasma triglycerides but not other lipids (Sanders & Reddy, 1991). In this latter experi-

Cereal foods with disproportionately high salt Australian and other countries dietary guidelines recommend people to choose low salt foods and use salt sparingly (eg, National Health & Medical Research Council, 1992). This is mainly as a contribution to reducing the prevalence of high blood pressure. The salt intake that many humans consume produces hypertension in chimpanzees. The chimpanzee experiment (Denton et al, 1995) is the nearest we can get to a controlled human experiment. The Nutrition Research Foundation at the University of Sydney held a symposium in December 1993 to consider whether and to what extent the dietary guideline to choose low salt foods is
European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

8
compatible with the guideline to eat plenty of breads and cereals . . . The paradox is that cereals are very low in sodium when they are harvested but after processing and as eaten some of them are among our highest sodium foods. An objective approach to grading the sodium content of cereal foods would be to use RG Hansens index of nutritional quality method (Hansen et al, 1979) and compare sodium content numbers against 2300 mg=day (100 mmol) as the upper end of desirable intake for adults (National Health & Medical Research Council, 1992). Two slices of bread weigh (approx) 60 g. This much bread contains (Mugford et al, 1996) 636 kJ 6.4% of standard 10 MJ energy intake day. Its sodium content, 312 mg is 13.6% of standard 2300 mg. Bread is therefore a donor of sodium to the diet. 13.6=6.4 sodium (% standard) is 2.12energy (% standard). Again, All Bran even in its new formulation contains 190 mg sodium in a 50 g standard serving (Salt Skip News, 1994): 190=2300 8%. The energy in this serving is 554 kJ which is 5.5% of 10 MJ. The sodium is still higher against the daily reference than the energy content: 8=5.5 1.45x. The sodium content at which breads sodium would equal its per cent energy contribution is 243 mg=100 g. Similarly the sodium content at which the sodium would equal the energy contribution of a breakfast cereal (with 1500 kJ per 100 g) is 345 mg sodium. There are a few breads and a number of breakfast cereals with sodium content lower than this, ie not disproportionately high relative to energy content, some as low as 1 mg=100 g. Sodium levels are tending to be brought down in breakfast cereals. followed for 10 to 20 y (Morris et al, 1977). CHD incidence was lower in these who ate more food energy (could be related to physical activity), and independently CHD was strikingly lower in those who were originally eating more cereal bre. When these results were rst presented at the Royal Society of Medicine, the audience was puzzled. The reliability of the investigators could hardly be doubted but bre of wheat (the main British cereal) does not reduce plasma cholesterol (Truswell et al, 1978). Perhaps people who eat plenty of cereal bre are more health conscious in a variety of ways. Morris did note that they were less likely to smoke (Morris et al, 1977). Burr and Sweetnam (1982) identied a cohort of 11 000 who shopped at health foods shops or subscribed to health food magazines and followed them up for 5 y. There was no evidence that their bre intake made any difference to the heart disease mortality. Some prospective studies that report a negative (protective) association of total dietary bre with CHD incidence have not been corrected for energy intake (Kromhout et al, 1982; Fehily et al, 1993). If people eat more megajoules per day (being better reporters or having larger energy expenditure) they are likely to eat more bre (and other food components). Other studies report negative association for total dietary bres (vegetable fruit cereal) (Khaw & Barrett-Connor, 1987) or for grain intake only in women (Knekt et al, 1994). Between 1996 and 1999, however, an accumulation of very large prospective studies (Table 5) have been reported that all conrm and extend the nding of Morris et al (1977). Rimm et al (1996) reported from the Health Professionals Follow-up Study. The cohort consisted of male dentists, veterinarians, pharmacists, optometrists, osteopaths and podiatrists. After exclusions there were 43 757 men. Their diets at the start were assessed with a 131 item food frequency questionnaire. Fibre intake was adjusted for energy intake by regression analysis. During 6 y of follow-up there were 511 non fatal cases of myocardial infarction and 229 coronary deaths (total 734 events). Relative risk of CHD for the top quintile (fth) of total dietary bre was 0.64 compared with 1.00 in the lowest bre quintile. On further analysis insoluble bre was found to be protective but

(C) Epidemiological studies relating cereal grains to CHD


Cohort studies Professor Jerry Morris of London, the epidemiologist who rst clearly demonstrated that exercise protects against CHD (Morris et al, 1953), published with Jean Marr (an expert on food intake methodology (Marr, 1971)) the results of a cohort of 337 men working for London Transport or banks,

Table 5 Recent Prospective Epidemiological Studies


Adjusted relative risk in highest quintile
0.71 0.77 0.60 0.82 0.68 sig sig sig sig sig

Authors
Rimm et al, 1996 Pietinen et al, 1996 Jacobs et al, 1998 Jacobs et al, 1999 Wolk et al, 1999 Liu et al, 1999 Jacobs et al 2001

Subjects
43 757 male health professionals USA 21 930 male smokers, Finland ATBC Study 34 492 females Iowa Womens Health Study 38 740 females 68 782 females US Nurses 75 521 females 33 848 men & women. Three countries prospective in Norway

CHD events
734 fatal and non-fatal 581 coronary deaths 438 coronary deaths 682 coronary deaths 591 fatal and non-fatal 761 fatal and non-fatal 553 coronary deaths

Dietary component
6 y follow-up cereal bre 6 y follow-up cereal bre 9 y follow-up whole grains

0.75 sig 0.76 sig

cereal bre 10 y follow-up whole grain 11 to 17 y follow-up whole grains

European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

9
(surprisingly) not soluble bre. A separate multivariate analysis was conducted for bre from cereals, vegetables and fruit. The negative association for cereal bre was the strongest relative risk for highest intake quintile 0.71 (95% condence interval 0.55 to 0.91). This relative risk did not appear to be much changed by adjustments for several micronutrients. Rimm et al made further calculations, eg they checked their results with bre calculated both by Southgates method and by Englysts. They concluded that they could not completely exclude the possibility that some unknown factor, associated with high bre intake is responsible, but it would need to have a strong inuence. The lack of any substantial confounding by known predictors of cardiovascular disease and by other dietary factors, together with demonstrated benets in metabolic studies, suggests that higher intake of dietary bre, particularly from cereal and grain sources can reduce substantially the risk of coronary heart disease (Rimm et al, 1996). Pietinen et al (1996) in Finland gave a 276 food item dietary questionnaire to most of the participants at the start of the Alpha-Tocopherol, Beta Carotene (ATBC) Cancer Prevention Study. They were men, all smokers 50 to 60 y of age at entry. The ATBC study was a randomised prevention trial: a-tocopherol or b-carotene or both or neither were given and subjects followed up for 6 y. The dietary questionnaire was taken home to be lled in, brought back two weeks later and checked by a nurse. The 21 930 subjects with good dietary data experienced 1399 CHD events (581 of which were deaths). In the analysis of association of dietary components with CHD events, adjustment had to be made for vitamin treatment groups (there were fewer coronary deaths in the vitamin E group and more in the b-carotene group). Intake of dietary bre ranged two-fold between the highest quintile (mean 34.8 g=day) and the lowest (16.1 g=day). Higher bre intake was associated with lower intakes of saturated fat, cholesterol and alcohol and higher intakes of rye products, dietary b-carotene, vitamins C and E. It was also associated with more physical activity but there were no differences in age, body mass index (BMI), smoking or serum cholesterol or blood pressure across quintiles of dietary bre intake. In age and treatment group-adjusted analysis, water soluble bre, cellulose and vegetable and fruit bre were signicantly inversely associated with risk of CHD events (fatal and non-fatal). For total dietary bre (energy adjusted) quintiles the relative risks were 1.00, 0.91, 0.88. 0.86. 0.84 (P for trend 0.03). Coronary deaths were more strongly negatively associated with total dietary bre. They were also signicantly inversely associated with soluble and insoluble bre and with cereal bre (RR 0.77). Simultaneous adjustment for each of the main food group sources showed that only the inverse association between cereal bre and CHD mortality remained signicant. Food groups inversely associated with CHD deaths were rye products, potatoes, vegetables and fruits and berries. An unusual feature of this study, apart from its huge size and statistical power is the relatively large range of bre intakes. A major contributor of cereal bre in Finland is rye. Those in the highest quintiles of total dietary bre consumed 161 g=day of rye. The overall mean daily intake of bre in the ATBC study was 18.9 insoluble bre and 5.4 g soluble bre. Adjustment for serum cholesterol did not change the results, so the cholesterol-lowering effect of soluble bre cannot be the explanation. Pietinen et al could only speculate that dietary bre may inuence risk of coronary disease via postprandial lipid response or glucose and insulin responses or haemostatic factors. DR Jacobs et al have published two papers about the Iowa Womens Health Study, relating the risk of death to whole grain intake. Whole grain cereals are clearly much the same as high bre cereals. The main difference is one of emphasis. Jacobs et al are thinking about other components in whole cereal grains as possibly benecial, not just the bre. In their rst paper (Jacobs et al, 1998) postmenopausal women were followed for 9 y. Food intake at the start was assessed with a 127 item food frequency questionnaire. The questionnaire was sent by mail to a random sample of all women 55 69 y of age who had an Iowa drivers licence. After exclusions there were 34 492 in the cohort. Disease outcomes were ascertained from state and national health registers. 3320 women died, including 438 from CHD. The authors found a striking inverse association of whole grain intake with risk of death from CHD. Age and energy-adjusted relative risks from lowest to highest quintiles of whole grain intake were 1.0, 0.84, 0.58, 0.45 and 0.60- (P for trend 0.0002). For total rened grain intake, on the other hand, there was a small positive, non signicant association; relative risks (multivariable adjusted) were 1.0, 0.99, 1.14, 1.04 and 1.12 (P 0.57) from lowest to highest intake quintile. The responsible types of whole grain cereal were mainly dark bread and whole-grain breakfast cereals. The inverse association of whole-grain intake with risk of CHD death was somewhat weakened after adjustment for various typical constituents such as dietary bre, phytic acid and vitamin E. The authors concede that in this study dietary assessment was made once, by a mailed, self-administered questionnaire and non-fatal CHD events were not collected. In their second paper Jacobs et al (1999) show in tables the risks of 10 major causes of death, by quintiles of whole-grain and rened grain intake, adjusted for 23 other variables. Whereas Jacobs et al (1998) excluded women with CHD at baseline, in this 1999 article they were included. With the multiple adjustments the relative risks for CHD deaths by increasing quintiles of whole-grain intake were 1.0, 1.03, 0.86, 0.70, 0.82, (P 0.03). The foods classied as whole grain cereals were (% of total) dark bread (61%), whole grain breakfast cereal (18%), popcorn (13%), cooked oatmeal (7%), wheat germ (1.5%), brown rice (1.3%), bran (0.6%) and other grains: bulgar, couscous (0.3%). Two reports by the Harvard epidemiology and nutrition group appeared in 1999 from the US Nurses cohort study.
European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

10
One is about dietary bre, the other about whole-grain consumption in relation to CHD. In the paper by Wolk et al (1999) 68 782 women were followed up for 10 y from 1984. Dietary data were collected by mailed food frequency questionnaires in 1984, 1986 and 1990. The authors documented 591 major CHD events (429 non-fatal myocardial infarction and 162 deaths). They found a signicant inverse association between total dietary bre intake and risk of CHD. This was due to a signicant inverse relation with cereal bre, but not with vegetable or fruit bre. Women in the highest quintile of cereal bre intake had a 34% lower risk of CHD events, compared with those in the lowest quintile. Relative risks in ascending cereal bre intake quintiles were 1.0, 1.06, 0.71, 0.76, 0.68 (P for trend 0.002). The inverse association was not explained by higher intakes of vitamin E, folate, vitamin B-6, magnesium, vegetables or fruit. Wolk et al (1999) speculate that cereal bre might act via increased insulin sensitivity or perhaps favourable effects on plasminogen activator type 1. In the other paper from the same group, Liu et al (1999) evidently conducted a separate analysis of data from the same large cohort of US nurses. Four of the authors of this paper are co-authors of Wolk et al (1999). In this second paper whole grain consumption (rather than cereal bre) was assessed against CHD events. Seventy-ve thousand, ve hundred and twenty-one women subjects completed food frequency questionnaires in 1984, 1986 and 1990 and were followed up for 10 y. Whole-grain food included dark breads, whole-grain breakfast cereals, popcorn, cooked oatmeal, wheat germ, brown rice, bran and other less common grains (the same classication as used by Jacobs et al (1999) Rened grain included sweet rolls, cake, desserts, white bread, pasta, mufns, biscuits, rened grain breakfast cereals, white rice, pancakes, wafes and pizza. (Breakfast cereals were classied as whole grain if they contained over 25% whole grain or bran by weight). Seven hundred and sixty one cases of CHD were documented, 208 of them fatal. After adjusting for age and smoking the relative risks of ascending quintiles of whole grain intake were 1.0, 0.86, 0.82, 0.72 and 0.67 (P for trend 0.001). Women with higher intakes of whole grain foods not only smoked less, they were more likely to exercise, to take hormone replacement or use supplements of multivitamins or vitamin E. After additional adjustment for these behaviours, as well as BMI, alcohol intake, aspirin use and type of fat intake, the relative risks of CHD events by quintile of wholegrain intake were 1.0, 0.92, 0.93, 0.83 and 0.75 (P for trend 0.001). The lower risk associated with higher whole-grain intake was partly but not fully explained by its contribution to intakes of dietary bre, folate, vitamin B-6 and vitamin E. There was no signicant association (negative or positive) between rened grain intake and CHD events. DR Jacobs of the University of Minnesota School of Public Health has also been one of the leaders of a large prospective cohort of the general population (men and women 35 56 y) in three Norwegian counties (Finnmark, Sogn og Fjordane
European Journal of Clinical Nutrition

and Oppland). They were followed from between 1977 and 1983 to 1994 (for 17 to 11 y). The response rate was 79%. Diet was assessed with a 66 item food frequency questionnaire. The authors combined self-report of number and type of bread slices (white, light whole grain, dense whole grain) to form a whole grain bread score. This could range from 0.05 (1 slice per day made with 5% whole grain our) to 5.4 (9 slices per day, made with 60% whole grain our). These Norwegian whole grain bread eaters were less likely to be smokers, were more physically active, had lower serum cholesterol and blood pressure and ate less saturated fat than white bread eaters. CHD deaths (expressed as hazard rate ratios went down stepwise with increasing whole grain quintiles (1, 0.92, 0.85, 0.75 and 0.63). When adjusted for smoking physical activity, use of cod liver oil and multivitamins, blood pressure, serum cholesterol and BMI, the hazard rate ratios were still inversely and stepwise related to whole grain intake quintiles (1, 0.99, 0.94, 0.88. 0.76; P for linear trend 0.04). Cancer deaths were also inversely related to whole grain intake quintiles but adjusted hazard rate ratios did not show a signicant trend. The authors did not estimate total vegetable and fruit intake in their food frequency questionnaire; they focussed strongly on intakes of bread, meat, sh, milk, coffee, oranges, potatoes, cakes and fat.

A Single randomised controlled trial The only randomised controlled trial in which people were asked to eat more cereal bre was carried out in South Wales, UK and reported in 1989 (Burr et al, 1989). Two-thousand and thirty-three men who had recovered from a myocardial infarction were given dietary fat advice or not, and=or fatty sh advice or not, and=or cereal bre advice or not. There were thus eight possible combinations of dietary advice and this was a secondary prevention trial for 2 y the diet and reinfarction trial (DART). According to a dietary questionnaire at 6 and 24 months the cereal bre advice group were eating 19 and 17 g cereal bre per day respectively, compared to 9 g cereal bre=day in the no bre advice group. After 2 y there were signicantly fewer CHD recurrences in the sh advice group but not in the bre advice group or the fat advice group. Compliance with fat advice was checked objectively with plasma cholesterols, compliance and with sh advice was checked with plasma eicosapentaenoic acid (EPA). Twenty-two percent of the sh group took Maxepa sh oil capsules which helped ensure increased EPA intake. There was, however, no objective check of bre intake. The authors admit that compliance must have been variable. Having eight different dietary groups looks as if it could lead to confusion. Follow-up was short and this was a secondary prevention trial of CHD (in which inuence of environmental factors on the pathological process is likely to be weaker than in a primary prevention trial). By contrast, people with a history of CHD were excluded from the Nurses

Cereal grains and CHD AS Truswell

11
and the Health Professionals cohorts described above (Rimm et al, 1996; Wolk et al, 1999; Lui et al, 1999). Although the DART trial had its weaknesses, the results of its bre arm means that there has to be a question mark at the end of the results of the cohort studies (Jacobs et al, 1999). An ecological between-countries study Recently, household budget data from 10 European countries were related to national mortality statistics for CHD, breast cancer and colorectal cancer (Lagiou et al, 1999) (Table 6). Simple correlation between total cereals and CHD mortality gave a coefcient of 0.13, ie a small positive association. The countries represented were Belgium, Germany, Greece, Hungary, Ireland, Luxembourg, Norway, Poland, Spain and UK and the data is for the start of the 1990s. This epidemiological approach is of course less reliable than cohort studies with multivariate analysis. It is affected by multiple confounding factors but it gives a view of the real life situation. This paper also showed that in the country with the lowest rate of CHD, cereal consumption was second highest. Beyond Europe CHD rates are low in countries with high cereal, consumption, in Japan, China and Africa.
Table 6 CHD mortality and cereal consumption (early 1990s) in 10 European countries from Lagiou et al, (1999)
CHD mortality*
Spain Greece Belgium Poland Luxembourg Germany Norway UK Hungary Ireland 73 95 103 119 124 150 187 223 235 255

3.

Cereals consumption, g**


168 268 177 283 174 167 176 167 264 239

*Gender and age adjusted, per 100 000 person-y. **Availability per head per day, based on household budget survey data.

Conclusions
1. There is no clear association, negative or positive between total cereal consumption and CHD. The pattern of consumption of the different cereal species and products varies between communities. The majority of over 40 human trials of the effect of oatmeal or oat bran on plasma lipids have found a modest reduction of total and LDL cholesterol. The US Food & Drug Administration permits a health claim for rolled oats, whole oat our or oat bran eligible sources of soluble bre as possibly reducing the risk of CHD (Food & Drug Administration, 1999). However, four of the ve trials of oat bre conducted in Australasia did not show a signicant effect. It is also likely that whole barley and rice bran share the cholesterol-lowering effect of oats if sufciently large amounts

2.

4.

are eaten. Barley and brown rice are not, however, eaten in large amounts in Europe. Wheat bre does not lower plasma total or LDL-cholesterol. In the late 1990s an impressive set of ve separate prospective (cohort) epidemiological studies were published from USA, Finland and Norway, based on over 98 000 men and 109 000 women. The ve cohorts all showed a signicant inverse association of cereal bre or whole grains with CHD. Only a minor part of this apparent protective effect can be explained by the cholesterol lowering effect of soluble bre. Part of the effect may be due to folate and=or vitamin E and=or effects on glucose=insulin responses and=or haemostatic factors. From the available evidence none of these can be singled out as the major mechanism for protection from CHD. It is more likely that cereal grains contain several protective factors. Because of these recent epidemiological ndings, Professor Willett (senior investigator of the Health Professionals Study and the US Nurses Study) has called for revision of the dietary pyramid used for nutrition education in the USA, and Australasia (Willett, 1998). Willett thinks the pyramids foundation is in need of repair: greater emphasis should be placed on the health benets of whole grains, not complex carbohydrates (Willett, 1998). Complex carbohydrates is a term that the FAO=WHO Consultation (Joint FAO= WHO Expert Consultation, 1998) advises against using. The US FDA now also permits a health claim (Food & Drug Administration, 1999) for whole grain foods (as dened) Diets rich in whole grain foods and other plant foods and low in total fat, saturated fat and cholesterol may reduce the risk of heart disease and some cancers. For purposes of this health claim whole grain foods contain 51% or more whole grain ingredients by weight per reference amount with dietary bre 2.3 g=50 g or 1.7 g per 35 g and the food low in fat. Note in passing that the FDA claim adds reduced risk of some cancers to the the heart disease claim for whole grain foods. In Australia meanwhile the NH & MRC in its 1999 clinical practice guidelines recommends poorly soluble cereal bres (eg wheat bran) as one of the dietary measures for primary prevention of colorectal cancer that is better supported by the scientic evidence (National Health & Medical Research Council, 1999). There is another possible way in which cereal foods, rened or whole grain, have the potential to contribute to reduction of atherothrombotic vascular diseases. This is as carriers of folic acid fortication. A high plasma (total) homocysteine level appears to be an independent risk factor for CHD, stroke and venous thrombosis (ie atherothrombotic vascular diseases in general, including but not conned to CHD; Hankey & Eikelboom, 1999). Folic acid at intakes in the nutritional range is the most effective of the three B vitamins in lowering high plasma homocysteine (Homocysteine
European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

12
Lowering Trialists Collaborative, 1998). In the USA folic acid fortication of cereal foods has been mandatory since 1996 and in one representative group of people (Framingham, MA, participants offspring) the prevalence of low plasma folate decreased from 22 to 2% and the prevalence of high plasma homocysteine dropped from 19 to 10% (Jacques et al, 1999). Voluntary fortication of cereals and some other staple foods is being encouraged by the Australian Commonwealth Department of Health and the ANZ Food Authority. Although the reason for folic acid fortication is ofcially stated to be reduction of neural tube defects in babies, it could in fact have a bigger public health impact on cardiovascular diseases. could be made that whole grain cereal foods and oat meal or bran may reduce the risk of CHD. The scientic evidence would support such claims and they are permitted in the USA. In Australia health claims on individual foods are not at present permitted by ANZFA. This is the big picture. Two minor considerations are that foods with lower glycaemic index and with lower salt (NaCl) content may be benecial. One priority for nutrition research on human cereal consumption is to increase understanding of the mechanisms for the protectiveness of whole grains (Slavin et al, 1999). Another priority is to organise, perhaps internationally, a second CHD prevention trial with added whole grain cereals having more rigorous design, and larger numbers and longer follow-up than the DART trial. The authors of that trial called for more randomised controlled trials of prevention of CHD (Elwood et al, 1992).

Historical perspective
The nutritional evidence and considerations in this review about the relative merits of whole grain and rened cereals are the latest chapter in a very long story of opinions about brown vs white breads (McCance & Widdowson, 1956). Until recently white bread and rened breakfast cereals were regarded as nutritionally suboptimal because they had lost major nutrients in the rening process. Addition of thiamin and some other nutrients to white bread and of several vitamins and minerals to many breakfast cereals was thought to at least partly answer these concerns. In 1974 Cleave (1974) argued that rened our may be a cause of several chronic degenerative diseases. Cereals provide the bulk of our starch consumption. In the 1970s starch was regarded as poor quality food. Though starch may still have a lowly image among consumers, its nutritional benets have become better known to nutritional scientists (Stephen et al, 1995), starting with the 1977 Dietary Goals for the United States (Select Committee on Nutrition, 1977). The low bre content of white bread and rened ready-to-eat cereals remained a disadvantage but except for Morriss paper (Morris et al, 1977) and oat bran this was not thought relevant to heart disease, and the discovery of resistant starch reduced the biological difference between brown and white breads. The present nutritional position of cereal foods, mostly rened, often with added nutrients and mixed ingredients and providing some resistant starch is that they are unobjectionable staple components of our usual western diet. For example, cereal foods and products provided 30% of the energy intake in the Dietary and Nutritional Survey of Australia (McLennan & Podger, 1997). Within the total cereals food group, whole grain products, only a minority of all the cereal products, are now emerging as nutritionally benecial, not only for large bowel health but also for prevention of CHD.

References
Anderson JW, Story L, Sieling B, Chen WJL, Petro MS & Story J (1984a): Hypocholesterolemic effects of oat bran or bean intake for hypercholesterolemic men. Am. J. Clin. Nutr. 40, 1146 1155. Anderson JW, Story L, Sieling B & Chen WJL (1984b): Hypocholesterolemic effects of high-bre diets rich in water-soluble plant bres. J. Can. Diet. Assoc. 45, 14 148. Anderson JW, Spencer DB, Hamilton CC, Smith SF, Tietyen J, Bryant CA & Oeltgen P (1990): Oat-bran cereal lowers serum total and LDL cholesterol in hypercholesterolemic men. Am. J. Clin. Nutr. 52, 495 499. Anderson JW, Gilinsky NH, Deakins DA, Smith SF, ONeal DS, Dillon DW & Oeltgen P (1991): Lipid responses of hypercholesterolemic men to oat-bran and wheat-bran intake. Am. J. Clin. Nutr. 54, 678 683. Birkeland KI, Gullestad L, Falch D & Torsvik H (1990): Oat bran and serum cholesterol. New Engl. J. Med. 322, 1748 1749. Braaten JT, Wood PJ, Scott FW, Wolynetz MS, Lowe MK, BradleyWhite P & Collins MW (1994): Oat b-glucan reduces blood cholesterol in hypercholesterolemic subjects. Eur. J. Clin. Nutr. 48, 465 474. Brand Miller J (1994): Importance of glycemic index in diabetes. Am. J. Clin. Nutr. 59, 747S 752S. Bremer JM, Scott RS & Lintott CJ (1990): Oat bread versus bread for lipid lowering. Proc. Nutrition Soc. N.Z. 15, 171 172. Brown L, Rosner B, Willett WW & Sacks FM (1999): Cholesterollowering effects of dietary bre: a meta-analysis. Am. J. Clin. Nutr. 69, 30 42. Burr ML & Sweetnam PM (1982): Vegetarianism, dietary ber and mortality. Am. J. Clin. Nutr. 36, 873 877. Burr ML, Fehily AM, Gilbert JF, Rogers S, Holliday RM, Sweetnam PM, Elwood PC & Deadman NM (1989): Effects of changes in fat, sh and bre intakes on death and myocardial reinfarction: diet and reinfarction trial (DART). Lancet ii, 757 761. Cleave TL (1974): The Saccharine Disease. Conditions Caused by the Taking of Rened Carbohydrates, Such as Sugar and White Flour. John Wright: Bristol. Davidson MH, Dugan LD, Burns JH, Bova J, Story K & Drennan KB (1991): The hypocholesterolemic effects of b-glucan in oatmeal and oat bran. J.A.M.A. 265, 1833 1839. De Groot AP, Luyken R & Pikaar NA (1963): Cholesterol-lowering effect of rolled oats. Lancet, ii, 303 304. Demark-Wahnefried W, Bowering J & Cohen PS (1990): Reduced serum cholesterol with dietary change using fat-modied and oat bran supplemented diet. J. Am. Diet. Ass. 90, 223 229.

Summary
While most cereal foods provide a range of nutrients, without special effect on risks of CHD, it appears that claims
European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell


Denton D, Weisinger R, Mundy NI, Wickings EJ, Dixson A, Moisson P et al (1995): The effect of increased salt intake on blood pressure of chimpanzees. Nature Med. 1, 1009 1016. Elwood PC, Burr ML & Sweetnam PM (1992): Fish, bre and heart disease. In Coronary Heart Disease Epidemiology. From Aetiology to Public Health, M Marmot & P Eliott (eds). pp 203 216. Oxford: Oxford Medical Publications. Fehily AM, Yarnell JWG, Sweetnam PM & Elwood PC (1993): Diet and incident ischaemic heart disease: The Caerphilly study. Br. J. Nutr. 69, 303 314. Flanagan M, Little C, Milliken J et al (1980): The effects of diet on high density lipoprotein. J. Human Nutr. 34, 34 45. Food and Drug Administration (1999): Washington DC: Health Claims Chart-Authorised Claims. Foster-Powell K & Brand Miller J (1995): International tables of glycemic index. Am. J. Clin. Nutr. 62, 871S 893S. Gariot P, Digby JP, Genton P, Lambert D, Bau RMH & Debry G (1986): Long-term effect of brain ingestion on lipid metabolism in healthy men. Ann. Nutr. Metab. 30, 369 373. Gold KV & Davidson DM (1988): Oat bran as a cholesterol-reducing dietary adjunct in a young, healthy population. West. J. Med. 148, 299 302. Gregory J, Foster K, Tyler H & Wiseman M (1990): The Dietary and Nutritional Survey of British Adults. London: HMSO. Hamilton CC, Tietyen J, Spencer BA & Anderson JW (1989): Serum lipid responses of hypercholesterolemic men to a ready to eat oat bran cereal. J. Am. Diet. Ass. 89(Suppl), A19. Hankey GJ & Eikelboom JW (1999): Review. Homocysteine and vascular diseases. Lancet 354, 407 413. Hansen RG (1973): An index of food quality. Nutrition Reviews 31, 1 7. Hansen RG, Wyse BW & Sorenson AW (1979): Nutritional Quality Index of Foods. Avi Publishing, Connecticut: Westport. Hegsted M, Windhauser MM, Lester SB & Morris K (1990): Stabilized rice bran and oat bran lower cholesterol in humans. F.A.S.E.B. J. 4, A 368. Hegsted M, Windhauser MM, Morris K & Lester SB (1993): Stabilized rice bran and oat bran lower cholesterol in humans. Nutrition Research 13, 387 298. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA & Southgate DAT (1991): McCance & Widdowsons The Composition of Foods, 5th Edition. Cambridge: Royal Society of Chemistry. Homocysteine Lowering Trialists Collaboration (1998): Lowering blood homocysteine with folic acid based supplements: metaanalysis of randomised trials. Br. Med. J. 316, 894 898. Jacques PF, Selhub J, Boston AG, Wilson PWF & Rosenberg IH (1999): The effect of folic acid fortication on plasma folate and total homocysteine concentrations. New Engl. J. Med. 340, 1449 1454. Jacobs DR Jr, Meyer KA, Kushi LH & Folsom AR (1998): Whole-grain intake may reduce the risk of ischemic heart disease death in postmenopausal women: the Iowa Womens Health Study. Am. J. Clin. Nutr. 68, 248 257. Jacobs DR, Jr, Meyer KA, Kushi LH & Folsom AR (1999): Is wholegrain intake associated with reduced total and cause-specic death rates in older women? The Iowa Womens Health Study. Am. J. Public Health 89, 322 329. Jacobs DR, Meyer HE & Solvoll K (2001): Reduced mortality among whole grain bread eaters in men and women in the Norwegian County Study. Eur. J. Clin. Nutr. 55, 137 143. Jenkins DJA, Hill MS & Cummings JH (1975): Effect of wheat ber on blood lipids, fecal steroid excretion and serum iron. Am. J. Clin. Nutr. 28: 1408 1411. Jenkins DJA, Wesson V, Wolever TMS, Jenkins AL, Kalmusky J, Giudici S, Csima A, Josse RG & Wong GS (1988): Wholemeal versus wholegrain breads: proportion of whole or cracked grain and the glycaemic response. Br. Med. J. 297, 958 960. Joint FAO=WHO Expert Consultation, Rome 14 18 April, 1997 (1998): Carbohydrates in Human Nutrition. FAO Food & Nutrition Paper 66, Rome: FAO. Judd PA & Truswell AS (1981) The effect of rolled oats on blood lipids and fecal steroid excretion in man. Am. J. Clin. Nutr. 34, 2061 2067. Kay RM & Truswell AS (1977): The effect of wheat ber on plasma lipids and faecal steroid excretion in man. Br. J. Nutr. 37: 227 235. Kashtan H, Stern HS, Jenkins DJA, Jenkins AL, Hay K, Marcon N, Minkin S & Bruce WR (1992): Wheat-bran and oat-bran supplements effects on blood lipids and lipoproteins. Am. J. Clin. Nutr. 55, 976 980. Keenan JM, Wenz JB, Myers S, Ripsin C & Huang Z (1991): Randomized, controlled, crossover trial of oat bran in hypercholesterolemic subjects. J. Family Practice 33, 600 608. Kestin M, Moss R, Clifton PM & Nestel PJ (1990): Comparative effects of three cereal brans on plasma lipids, blood pressure and glucose metabolism in mildly hypercholesterolemic men. Am. J. Clin. Nutr. 52, 661 666. Khaw KT & Barrett-Connor E (1987): Dietary ber and reduced heart disease mortality rates in men and women: a 12 y prospective study. Amer. J. Epidemiol. 126, 1092 1103. Kirby RW, Anderson JW, Sieling B, Rees ED, Chen WJL, Miller RE & Kay RM (1981): Oat-bran intake selectively lowers serum lowdensity lipoprotein cholesterol concentrations of hypercholesterolemic men. Am. J. Clin. Nutr. 34, 824 829. Knekt P, Reunaen A, Jarvinen R, Seppanen R, Heliovaara M & Aroura A (1994): Antioxidant vitamin intake and coronary mortality in a longitudinal population study. Am. J. Epidemiol. 139, 1180 1189. Kretsch MJ, Crawford L & Calloway DH (1979): Some aspects of bile acid and urobilinogen excretion and faecal elimination in men given a rural Guatemalan diet and egg formula with and without oat bran. Am. J. Clin. Nutr. 32, 1492 1496. Kromhout D, Bosschieter EB & de Lezenne Coulander C (1982): Dietary bre and 10 y mortality from coronary heart disease, cancer and all causes. Lancet ii, 518 522. Lagiou P, Trichopoulou A, Hendrickx HK, Kelleher C, Leonhauser IU, Moreiras P, Nelson M, Schmitt A, Sekula W, Trygg K & Zajkas G (1999): Household budget survey nutritional data in relation to mortality from coronary heart disease, colorectal cancer and female breast cancer in European countries. Eur. J. Clin. Nutr. 53, 328 332. Leadbetter J, Ball MJ & Mann JI (1991): Effect of increasing quantities of oat bran in hypercholesterolemic people. Am. J. Clin. Nutr. 54, 841 845. Lepre F & Crane S (1992): Effect of oatbran on mild hyperlipidaemia. Med. J. Aust. 157, 305 308. Lindegarde F & Larsson L (1984): Effect of a concentrated bran bre preparation on HDL-cholesterol in hypercholesterolaemic men. Human Nutr. Clin. Nutr. 38C, 39 45. Lui S, Stampfer MJ, Hu FB, Giovanucci E, Rimm E, Manson JE, Hennekens CH & Willett WC (1999): Whole grain consumption and risk of coronary heart disease: results from the Nurses Health study. Am. J. Clin. Nutr. 70, 412 419. Luyken R, de Wijn JF, Pikaar NA & van der Meer R (1965): De invloed van havermout op het serum cholesterolgehalte van het bloed. Voeding 26, 229 244. Mackay S & Ball MJ (1992): Do brans and oatbran add to the effectiveness of a low fat diet? Eur. J. Clin. Nutr. 9, 641 648. McCance RA & Widdowson EM (1956): Breads White and Brown. Their Place in Thought and Social History. London: Pitman Medical. McDougall RM, Yakymyshyn L, Walker K & Thurston OG (1978): Effect of wheat bran on serum lipoproteins and biliary lipids. Can. J. Surg. 21, 433 435. McIntosh GM, Whyte J, McArthur R & Nestel PJ (1991): Barley and wheat foods: inuence on plasma cholesterol concentrations in hypercholesterolemic men. Am. J. Clin. Nutr. 53, 1205 1209. McLennan W & Podger A (1997): National Nutrition Survey Selected Highlights Australia 1995. Canberra: Australian Bureau of Statistics, Department of Health. Marr JW (1971): Individual dietary surveys: purposes and methods. World Review of Nutrition & Dietetics 13, 106 164. Ministry of Agriculture, Fisheries & Food (1994): The Dietary and Nutritional Survey of British Adults Further Analysis. London: HMSO.

13

European Journal of Clinical Nutrition

Cereal grains and CHD AS Truswell

14

Morris JN, Heady JA, Rafe PAB & Parks JW (1953): Coronary heart disease and physical activity of work. Lancet ii, 1053 1057 and 1111 1120. Morris JN, Marr JW & Clayton DG (1977): Diet and heart: A postscript. Br. Med. J. 2, 1307 1314. Mugford DC, Grifth PJ & Walker AR (1996): Nutrient levels in white, mixed grain and wholemeal bread. An Australia-wide survey of breads from different bakeries and different States. Food Australia 48, 264 269. Munoz JM, Sandstead HH & Jacobs RA (1979): Effect of some cereal brans and TVP on plasma lipids. Am. J. Clin. Nutr. 32, 580 592. National Health & Medical Research Council (1992): Dietary Guidelines for Australians. Canberra: Australian Government Publishing Service. National Health & Medical Research Council (1999): The prevention, early detection and management of colorectal cancer. Clinical Practice Guidelines. Canberra: NH & MRC. Newman RK, Lewis SE, Newman CW, Boik RJ & Ramage RT (1989): Hypercholesterolemic effect of barley foods on healthy men. Nutr. Rep. Int. 39, 749 759. Normand FL, Ory RL & Modd RR (1987): Binding of bile acids and trace minerals by soluble hemicelluloses of rice. Food Technol. 41, 86 99. OMoore RR, Flanagan M, McGill AR, Wright EA, Little C & Weir DG (1978): Diet and heart disease. Br. Med. J. i: 1213. Pietinen P, Rimm EB, Korhonen P, Hartman AM, Willett WC, Albanes D & Virtamo J (1996): Intake of dietary ber and risk of coronary heart disease in a cohort of Finnish men: the Alpha-Tocopherol, Beta Carotene Cancer Prevention Study. Circulation 94, 2720 2727. Poulter N, Chang CL, Cugg A, Poulter C, Sever P & Thom S (1993): Lipid proles after the daily consumption of an oat-based cereal: a controlled crossover trial. Am. J. Clin. Nutr. 58, 66 69. Raymond TL, Connor WE, Lin DS, Warner S, Fry MM & Connor SL (1977): The interaction of dietary bers and cholesterol upon the plasma lipids and lipoproteins, sterol balance and bowel function in human subjects. J. Clin. Invest. 60, 1429 1437. Reynolds HR, Lindeke E & Hunninghake DB (1989): Effect of oat bran on serum lipids. J. Am. Diet. Ass. 89(Suppl), A112. Rimm EB, Ascherio A, Giovanucci E, Spiegelman D, Stampfer MJ & Willett WC (1996): Vegetable, fruit and cereal ber intake and risk of coronary heart disease among men. J.A.M.A. 275, 447 541. Roth G & Leitzmann C (1985): Langezeiteinuss ballaststoffereicher Fruhstuckscerealien auf die Blutlipide beim Menschen. Akt. Ernahr. 10, 106 109. Salt Skip News 1994, August. Kelloggs still reducing salt in breakfast cereals. Hobart: Menzies Centre for Population Health Research. Sanders TAB & Reddy S (1991): The inuence of rice bran on plasma lipids and lipoproteins in human volunteers. Eur. J. Clin. Nutr. 46, 161 172. Saudia TL, Bareld BR & Barger J (1992): Effect of oat bran consumption on total cholesterol levels in healthy adults. Mil. Med. 157, 567 568. Select Committee on Nutrition and Human Needs, United States Senate (1977): Dietary Goals for the United States. Washington DC: US Government Printing Ofce. Slavin JL, Martini MC, Jacobs DR & Marquart L (1999): Plausible mechanisms for the protectiveness of whole grains. Am. J. Clin. Nutr. 70, 459S 463S. Spencer H, Norris C, Derler J & Osis D (1991): Effect of oat bran mufns on calcium absorption and calcium, phosphorus, magnesium and zinc balance in men. J. Nutr. 121, 1976 1983. Stasse-Wolthuis M, Katan MN, Hermus RJJ, Hautvast JGAJ (1979): Increase of serum cholesterol in men fed a bran diet. Atherosclerosis 45, 87 91. Stasse-Wolthuis M, Albers HFF, van Jeveren JGC et al (1980): Inuence of dietary ber from vegetables and fruits, bran or citrus pectin on serum lipids, fecal lipids and colonic function. Am. J. Clin. Nutr. 33, 1745 1756.

Stephen AM, Sieber GM, Gerster YA & Morgan DR (1995): Intake of carbohydrate and its components-international comparisons, trends over time, and effects of changing to low-fat diets. Am. J. Clin. Nutr. 62, 815S 867S. Stewart FM, Neutze J & Newsome-White R (1992): The addition of oatbran to a low fat diet has no effect on lipid values in hypercholesterolaemic subjects. N.Z. Med. J. 105, 398 400. Storch K, Anderson JW & Young VR (1984): Oat-bran mufns lower serum cholesterol of healthy young people. Clin. Res. 34, 740A. Swain JF, Rouse IL, Curley CB & Sacks FM (1990): Comparison of the effects of oat bran and low-bre wheat on serum lipoprotein levels and blood pressure. New Engl. J. Med. 322, 147 152. Thompson LU (1993): Potential health benets and problems associated with antinutrients in foods. Food Research International 26, 131 149. Tietyen JL, Nevins DJ & Schneeman BO (1990): Characterization of the hypercholesterolemic potential of oat bran. F.A.S.E.B. J. 4, A527. Torronen R, Kansanen L, Uusitupa M, Hanninen O, Myllmaki O, Harkonen H & Malkki Y (1992): Effect of an oat bran concentrate on serum lipids on free-living men with mild to moderate hypercholesterolaemia. Eur. J. Clin. Nutr. 9, 621 627. Trowell H (1975): Coronary heart disease and dietary ber. Am. J. Clin. Nutr. 28, 798 800. Truswell AS & Beynen AC (1992): Dietary bre and plasma lipids: potential for prevention and treatment of hyperlipidaemias. In Dietary Fibre a Component of Food. Nutrition functions in Health and Disease, TF Schweizer & CA Edwards (eds). pp 295 332. ILSI Human Nutrition Reviews. London: Springer Verlag. Truswell AS & Kay RM (1976): Bran and blood lipids. Lancet i, 367. Truswell AS (1995): Dietary bre and blood lipids. Current Opinion in Lipidology 6, 14 19. Truswell AS, Morris, Sinclair, Low-Beer, McLean-Baird, Heaton, Eastwood & Trowell. In Discussion after paper by JN Morris, JW Marr & DG Clayton (1978): In KW Heaton (ed). Dietary Fibre, 3rd Kellogg Nutrition symposium, pp 54 56. London: Newman Publishing. Turnbull WH & Leeds AR (1987): Reduction of total and LDLcholesterol in plasma by rolled oats. J. Clin. Nutr. Gastroenterol. 2, 177 181. Van dokkum W (1978): Zemelen in brood: verteerbaarheid en invloed op het defaecatiepatroon, de mineralen-balans en de serumlipidconcentraties bij de mens. Voedingsmiddelentechnologie 11, 18 21. Van Horn L, Lui K, Parker D, Emidy L, Liao Y, Pan WH, Giumetti D, Hewitt J & Stampler J (1986): Serum response to oat product intake with a fat-modied diet. J. Am. Diet. Ass. 86, 759 764. Van Horn L, Emidy L, Lui K, Liao Y, Ballew C, King J & Stanley J (1988): Serum lipid responses to fat-modied oatmeal-enhanced diet. Prev. Med. 17, 377 386. Van Horn L, Moag-Stahlberg A, Lui K, Ballew C, Ruth K, Hughes R & Stamler J (1991): Effects on serum lipids of adding instant oats to usual American diets. Am. J. Pub. Hlth. 81, 183 188. Welch RW, Peterson DM & Schranka N (1988): Hypercholesterolemic and gastrointestinal effects of oat bran fractions in chicks. Nutr. Rep. Int. 38, 551 561. Whyte JL, McArthur R, Topping D & Nestel P (1992): Oat bran lowers plasma cholesterol levels in mildly hypercholesterolemic men. J. Am. Diet. Ass. 92, 446 449. Willett WC (1998): The dietary pyramid: does the foundation need repair? Am. J. Clin. Nutr. 68, 218 219. Wolk A, Manson JE, Stampfer MJ, Colditz GA, Hu FB, Speizer FE, Hennekens CH & Willett WC (1999): Long-term intake of dietary ber and decreased risk of coronary heart disease among women. J.A.M.A. 281, 1998 2004. Yoshino G, Kazumi T, Amaino M, Tateiwa M, Yamasaki T, Takashima S, Iwai M, Hatanaka H & Baba S (1989): Effects of gamma-oryzanol on hyperlipidemic subjects. Curr. Ther. Res. Clin. Experience 45, 543 552. Zhang J-X, Hallman SG, Andersson H, Bosaeus I, Aman P, Tidehag P, Stenling R, Lundin E & Dahlgren S (1992): Effect of oat bran on plasma cholesterol and bile acid excretion in nine subjects with ileostomies. Am. J. Clin. Nutr. 56, 99 105.

European Journal of Clinical Nutrition