Food Sources of Plant Sterols in the EPIC Norfolk Population

ORIGINAL ARTICLE Food sources of plant sterols in the EPIC Norfolk population

S Klingberg1, H Andersson1, A Mulligan2, A Bhaniani2, A Welch2, S Bingham3, K-T Khaw2,4, S Andersson1 and L Ellega˚rd1

1Department of Clinical Nutrition, Sahlgrenska Academy at Go¨teborg University, Go¨teborg, Sweden; 2Department of Public Health and Primary Care, University of Cambridge, Strangeways Site, Wort’s Causeway, Cambridge, UK; 3Medical Research Council Dunn Human Nutrition Unit, Cambridge, UK; 4Clinical Gerontology Unit, University of Cambridge School of Medicine, Addenbrooke’s Hospital, Cambridge, UK

Objective: To investigate the intake of plant sterols and identify major dietary sources of plant sterols in the British diet. Subjects: A total of 24 798 men and women recruited during 1993–1997, participating in the European Prospective Investigation into Cancer (EPIC-Norfolk). Interventions: A database of the plant sterol (campesterol, b-sitosterol, stigmasterol, campestanol and b-sitostanol) content in foods, based on gas-liquid chromatography (GLC) analyses, was linked to nutritional intake data from food frequency questionnaires in the EPIC-Norfolk population. Results: The mean (s.d.) intake of total plant sterols was 300 (108) mg/d for men and 293 (100) mg/d for women. Bread and other cereals, vegetables and added fats were the three major food sources of plant sterols representing 18.6 (8.9), 18.4 (8.5) and 17.3 (10.4)% of the total plant sterol intake respectively. Women had a higher plant sterol density than men (36.4 vs 32.8 mg/1000 kJ, Po0.001) and in relation to energy intake higher intakes of plant sterols from vegetables, bread and other cereals, added fats, fruits and mixed dishes (all Po0.001), whilst men had higher intakes of plant sterols from cakes, scones and chocolate, potatoes (all Po0.001) and other foods (Po0.01). Conclusions: The intake of plant sterols in UK, mainly from bread, cereals, fats and vegetables, is much higher than previously reported but comparable to recent European studies. European Journal of Clinical Nutrition (2008) 62, 695–703; doi:10.1038/sj.ejcn.1602765; published online 18 April 2007

Keywords: plant sterols; food sources; population; diet

Introduction The chemical structure of plant sterols resembles that of cholesterol but with an additional methyl- or ethyl group in Plant sterols are bioactive compounds found in all vegetable the side chain. More than 250 different plant sterols have foods at varying concentrations. There is, however, limited been identified (Piironen et al., 2000). Plant sterols are information on the plant sterol intake from common foods known to reduce the absorption of cholesterol and the serum in Britain. level of cholesterol, although the mechanisms are not fully understood. In addition to the well-known effect of decreased intestinal cholesterol absorption due to reduced Correspondence: S Klingberg, Department of Clinical Nutrition, Sahlgrenska incorporation of cholesterol into mixed micelles, plant Academy at Go¨teborg University, Box 459, Gothenburg 40530, Sweden. sterols are suggested to influence cellular cholesterol meta- E-mail: [email protected] bolism within the intestinal enterocytes (Plat and Mensink, Guarantor: S Klingberg. Contributors: K-TK and SB are principal investigators in the EPIC-Norfolk 2005). population study. AW and SB are responsible for the nutritional analyses. HA, It is well documented that commercial products enriched SA and LE developed the plant sterol nutrient database. AM and AB prepared with plant sterols reduce serum cholesterol levels. A the plant sterol data set in the EPIC cohort. SK was principal investigator for metaanalysis of 41 different studies has shown that a plant this study and wrote the paper with contributions from co-authors. Received 3 July 2006; revised 12 March 2007; accepted 14 March 2007; sterol dose of 2 g/day reduces low-density lipoprotein (LDL) published online 18 April 2007 cholesterol levels by 10% (Katan et al., 2003). In contrast to Food sources of plant sterols S Klingberg et al 696 these studies with products supplemented with plant sterols, 0.01 mg/100 g fresh material of the tested product. The much less is known about the effects of dietary plant sterol concentration of the five most frequently occurring plant intake (Ostlund, 2002). Traditionally plant sterols from the sterols (the unsaturated plant sterols campesterol (24a- habitual diet have not been considered to have any methyl-5-cholesten-3b-ol), stigmasterol (5,22-cholestadien- significant effect on the serum cholesterol level but this 24a-ethyl-3b-ol) b-sitosterol (24a-ethylcholest-5-en-3b-ol), view has been questioned (Fraser, 1994; Ostlund, 2002). and the saturated plant stanols campestanol (24a-methyl- Serum cholesterol lowering effects of the plant sterol intake 5a-cholestan-3b-ol), and b-sitostanol (24a-ethyl-5a-cholestan- from the natural diet could probably have been under- 3b-ol)) was analysed. The sum of the five plant sterols estimated (Ostlund, 2002). As recently shown in a large-scale constitutes ‘total plant sterols’. All foods were analysed in epidemiological study using the current database, a high duplicate. All fruits and vegetables were bought in 1996, intake of natural dietary plant sterols is significantly cereals in 1997 and fatty foods in 1997 and 2001 in two inversely related to total and LDL cholesterol levels shops in the Gothenburg area. Specific British food items (Andersson et al., 2004). were bought in Cambridge in 2004 and 2005. Fruits and Knowledge of major food sources of plant sterols could be vegetables were analysed as a mix of two different samples an important tool when trying to further improve the and analysis of cereal and fatty foods were made from a mix natural intake of plant sterols as a component in the of two to seven samples. British food items were analysed in prevention of cardiovascular disease. The aim of the current duplicate from a single sample. Seasonal variation in fruits study was to investigate the natural intake of plant sterols in and vegetables was not taken into account. the Norfolk population of The European Prospective In- vestigation of Cancer (EPIC) and to identify major dietary sources of plant sterols in the British diet. Plant sterol database Analyses of more than 330 food items were collected in a plant sterol database and this database was used to estimate Methods plant sterol intake. This database includes vegetables, fruits, cereals, bread, fats, nuts, confectionery, beverages and The study population specific British food items. Food items from this database The EPIC-Norfolk cohort recruited approximately 25 000 were matched with the food items present in the FFQ. The men and women aged 40–79 years of age between 1993 and plant sterol data on vegetables, fruits, cereals and fatty foods 1997 (Day et al., 1999). EPIC-Norfolk is part of the Europe- have been published (Norme´n et al., 1999, 2002, 2007). Plant wide EPIC study designed to investigate dietary and other sterol content of specific British food items used in this study determinants of cancer. The EPIC-Norfolk cohort also is shown in Appendix 1. included end points other than cancer, for example cardiovascular disease. Informed consent was given by all participants and the study was approved by the Norfolk and Food frequency questionnaire Norwich Hospital Ethics Committee. Food intake and At recruitment participants completed a FFQ consisting of nutrient data from food frequency questionnaires (FFQ) a food list of 130 questionnaire lines corresponding to were available for 24 838 participants. 275 food items (Bingham et al., 2001; Welch et al., 2005). The FFQ was validated against a 16-day weighed records and the biomarkers 24-h urine N and K, plasma carotenoids and Plant sterol analysis plasma vitamin C (Bingham et al., 1994, 1995, 1997). The Analyses of the food items were performed at the Depart- questions about intake of different fruits and vegetables were ment of Clinical Nutrition, Go¨teborg University, Sweden, detailed and types of bread, pasta, breakfast cereals and types using a gas-liquid chromatography (GLC) procedure of fat used for frying, baking and spreading had been modified after Jonker et al. (1985) and validated with gas specified which assisted in the correct assignment of plant chromatography mass spectrometry (Dutta and Norme´n, sterol values. Based either on pure animal origin or on 1998). In short, the method comprised acid hydrolysis (6 M ingredients not containing plant sterols, 63 food items were HCl), alkaline saponification (96% ethanolic KOH), lipid set to zero. For plant sterol containing food items the plant extraction with toluene, and a final step of washing in sterol value was based on direct analysis in 119 food items. de-ionized water until neutral pH was reached. Internal Fifty-six food items were assigned a value from similar standard, containing 5a-cholestane, was added to all samples analysed products or proportions of analysed products. For before saponification to quantify the sterols. Samples were 23 food items the plant sterol content was based on

dehydrated with Na2SO4, filtered and evaporated under calculations of British standard receipts with analysed vacuum at 501C. The residue was dissolved in chloroform ingredients. Fourteen food items were assigned plant sterol and stored at À201C. Silylation of sterols to trimethylsilyl values from the UK food composition database ‘McCance ether derivatives was performed before analysis by gas liquid and Widdowson’s The composition of Foods’ (Food Stan- chromatography (GLC). The limit of detection was set to dards Agency, 2002). As plant sterol enriched products were

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 697 not launched until 1999 in the UK, no such products were higher intakes of plant sterols from cakes, scones and present in the FFQ:s. chocolate, potatoes (all Po0.001) and other foods (Po0.01; Figure 1). According to the age groups 40–52, 53–65 and 66–79 years Calculations the plant sterol density for men was significantly higher in All food items were categorized into nine food groups: fruits; the youngest age group than in the oldest age group (Table 2). vegetables; potatoes; bread and other cereals (breakfast For women the plant sterol density was higher in both cereals, pasta and rice); cakes, scones and chocolate; nuts younger age groups compared to the older age group. In and peanut butter; added fats; mixed dishes (dishes of fish, relation to energy intake, both men and women in the meat and vegetables) and other foods (beverages, coffee youngest age group had significantly higher intakes of plant whitener, confectionery, condiments, crisps, sauce, soya/tofu sterols from added fats, mixed dishes, other foods, potatoes and sugar/jam). and nuts and peanut butter and significantly lower intakes of plant sterols from vegetables, cakes, scones and chocolate and fruits (all Po0.001) compared to the oldest age group Statistical analysis (Figure 2a and b). The younger men had also a lower intake Participants with estimated total plant sterol intakes of plant sterols from bread and other cereals (Po0.05) 4750 mg/day were excluded, as they were considered compared to the oldest age group. extreme outliers, which resulted in exclusion of 17 men and 23 women. In total, 24 798 participants (11 227 men and 13 571 women) were thus included in the analysis. All values Discussion are presented as means and standard deviations (s.d.). Comparisons of two groups with continuous variables were An earlier report on the plant sterol intake from the British performed by t-test for independent samples. Different age diet (Morton et al., 1995) showed considerably lower groups were compared using the one-way analysis of amounts of plant sterols than the present study. The variance (ANOVA). When the overall P value from ANOVA previous study was based on the UK TDS collected in the was o0.05 the post hoc Bonferroni test was performed. years 1987 and 1991, using food consumption data from Statistical calculations were performed using the SPSS for the National Food Survey (NFS). Average daily intake of plant WINDOWS (version 11.5) Software program. sterols was estimated from chemical analysis of 20 food groups corresponding to the average consumption of 119 categories of foods and drinks constituting a representative Results diet. Until 1992 the NFS did not include information of foods bought and consumed outside the home. The Mean (s.d.) intake of total plant sterols was 300 (108) mg/day completeness of the estimation of plant sterol intake can for men and 293 (100) mg/day for women (Table 1). The therefore be questioned and the present findings correspond major plant sterol was b-sitosterol contributing to 66% of to approximately 160% of the previous value. The analyses total plant sterol intake. The other sterols, campesterol and of the earlier study were performed using a thin layer stigmasterol represented 22 and 8% of total plant sterol chromatography (TLC) procedure. Some of the differences in intake, respectively. The plant stanols campestanol and b- the plant sterol intake between the two studies can probably sitostanol were minor contributors and together constituted be explained by different nutritional assessment methods only 4% of the total plant sterol intake. Absolute intakes of and chemical analytical methods. In the present study, plant sterols from different foods and food groups are vegetables made a greater contribution to the total plant presented in Table 1. sterol intake than in the study by Morton et al. (1995) (18 vs Bread and other cereals, vegetables and added fats were 7%) while the intakes from bread and other cereals including the three major food sources of plant sterols representing cakes, scones and chocolate represented 33% of total plant 18.6 (8.9), 18.4 (8.5) and 17.3 (10.4)% of the total plant sterol intake in both studies. sterol intake, respectively. The intake of cakes, scones and More recent studies of plant sterol intake in the Nether- chocolate contributed with 14.4 (10.5)% of the total plant lands (Norme´n et al., 2001) and in Finland (Valsta et al., sterol intake whilst fruits contributed with 12.3 (8.4)%. 2004) show estimated daily intakes of just above 300 mg for The mean (s.d.) energy intake was 9.2 (2.6) MJ for men and men and between 240 and 260 mg for women, with bread, 8.1 (2.3) MJ for women, and women had a significantly cereals, vegetable fats, vegetables and fruits as the main higher plant sterol intake in relation to energy intake (plant sources. The plant sterol intake in Spain has recently been sterol density, mg/1000 kJ) than men (36.4 vs 32.8 mg/ reported to be 276 mg per day with oils, cereals, vegetables 1000 kJ, Po0.001). In relation to energy intake, women and fruits as the major sources (Jime´nez-Escrig et al., 2006). had significantly higher intakes of plant sterols from Accordingly the main food sources in the current data are vegetables, bread and other cereals, added fats, fruits and comparable with main food sources found in these other mixed dishes (all Po0.001), whilst men had significantly contemporary European studies.

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 698 Table 1 Contribution of food groups to the intake of five different plant sterols in men and women (n ¼ 24 798) aged 40–79 years, EPIC Norfolk 1993– 1997 (values are mean (s.d.))

Consumption Unsaturated Saturated Total plant sterols mg/d Edible part g/d Campesterol Sitosterol Stigmasterol Campestanol Sitostanol mg/d mg/d mg/d mg/d mg/d

Fruits Total 246 (184) 2.9 (2.5) 30.9 (23.9) 1.9 (1.5) 0.0 (0.0) 0.4 (0.9) 36.0 (27.7) Citrus 42 (58) 1.3 (1.7) 8.4 (11.7) 0.4 (0.6) 0.0 (0.0) 0.0 (0.0) 10.0 (13.9) Apples, pears 98 (98) 0.4 (0.4) 12.2 (12.1) 0.1 (0.1) 0.0 (0.0) 0.0 (0.0) 12.6 (12.6) Berries 20 (32) 0.3 (0.5) 2.5 (4.0) 0.2 (0.3) 0.0 (0.0) 0.0 (0.0) 2.9 (4.7) Other (bananas, canned fruit, dried fruit, 86 (77) 1.0 (0.9) 7.8 (6.9) 1.3 (1.1) 0.0 (0.0) 0.4 (0.9) 10.4 (9.4) stone fruit, melon, grapes)

Vegetables Total 265 (127) 8.9 (5.0) 35.3 (19.0) 7.3 (3.8) 0.2 (0.1) 0.6 (0.3) 52.2 (27.0) Cabbage 78 (56) 4.7 (3.5) 17.8 (13.1) 1.1 (0.9) 0.0 (0.0) 0.1 (0.0) 23.7 (17.3) Legumes 59 (37) 1.0 (0.7) 9.2 (5.6) 3.9 (2.7) 0.0 (0.0) 0.0 (0.0) 14.1 (8.7) Leafy 2 (7) 0.0 (0.0) 0.1 (0.4) 0.1 (0.2) 0.0 (0.0) 0.0 (0.0) 0.2 (0.6) Other (root veg., fruit veg, onions, 125 (70) 3.1 (2.3) 8.2 (5.1) 2.2 (1.3) 0.1 (0.1) 0.5 (0.3) 14.2 (8.3) mushrooms, sweetcorn, lettuce, celery)

Potatoes Total 117 (66) 1.5 (2.2) 7.1 (5.8) 1.0 (0.8) 0.1 (0.1) 0.7 (0.4) 10.2 (8.9) Without fat 85 (60) 0.2 (0.1) 2.3 (1.6) 0.3 (0.2) 0.0 (0.0) 0.5 (0.4) 3.2 (2.3) With fat 32 (28) 1.3 (2.2) 4.8 (5.6) 0.7 (0.7) 0.1 (0.1) 0.2 (0.2) 7.0 (8.6)

Added fats Total 26 (20) 17.9 (13.5) 31.1 (27.6) 2.1 (1.8) 0.1 (0.1) 0.1 (0.1) 54.2 (42.2) Margarine 5 (9) 3.3 (5.7) 4.5 (7.9) 0.2 (0.4) 0.0 (0.0) 0.0 (0.0) 8.0 (14.0) Spread not polyunsat 5 (11) 3.1 (6.9) 4.4 (9.7) 0.1 (0.2) 0.0 (0.0) 0.0 (0.0) 7.6 (16.8) Spread polyunsat 9 (13) 6.0 (8.8) 16.2 (23.8) 0.9 (1.3) 0.0 (0.0) 0.0 (0.0) 23.1 (33.9) Mayonnaise, dressings 6 (8) 5.5 (7.3) 9.1 (11.8) 0.8 (1.1) 0.1 (0.1) 0.1 (0.1) 15.5 (20.2)

Bread and other cereals Total 157 (88) 11.1 (6.4) 32.0 (17.8) 1.9 (1.3) 4.2 (3.1) 5.2 (3.6) 54.3 (31.6) Wholemeal 33 (47) 3.5 (4.9) 9.6 (13.6) 0.6 (0.8) 1.8 (2.5) 2.1 (3.0) 17.6 (24.9) Brown 13 (26) 1.5 (3.0) 3.7 (7.5) 0.3 (0.6) 0.5 (1.0) 0.7 (1.3) 6.6 (13.3) White 33 (43) 2.4 (3.2) 6.6 (8.7) 0.2 (0.3) 0.6 (0.8) 0.8 (1.0) 10.5 (13.9) Crackers, crisp bread 4 (9) 0.6 (1.4) 1.7 (3.7) 0.2 (0.4) 0.1 (0.4) 0.2 (0.5) 2.8 (6.2) Other cereals (breakfast cereals, rice, 73 (64) 3.1 (2.7) 10.4 (7.9) 0.6 (0.7) 1.2 (1.4) 1.4 (1.6) 16.7 (13.7) pasta)

Cakes, scones, chocolate 97 (82) 11.5 (11.4) 27.2 (24.5) 4.7 (4.8) 0.3 (0.3) 0.7 (0.7) 44.4 (40.1) Nuts, peanut butter 3 (9) 0.5 (1.4) 3.1 (8.4) 0.4 (1.1) 0.0 (0.0) 0.1 (0.2) 4.0 (11.0) Mixed dishes 94 (60) 7.2 (6.1) 12.4 (9.4) 1.5 (1.0) 0.0 (0.0) 0.1 (0.1) 21.3 (16.4) Other (coffemate, marmite, ice cream, 3.4 (4.1) 12.0 (12.0) 3.1 (3.6) 0.2 (0.2) 0.5 (0.8) 19.2 (20.1) sweets, crisps, condiments)

Total plant sterols 64.9 (25.9) 194.1 (67.5) 23.9 (9.4) 4.9 (3.1) 8.3 (4.1) 295.8 (103.5) Men 66.6 (27.1) 195.6 (69.5) 24.5 (9.9) 5.0 (3.2) 8.4 (4.39) 299.8 (107.5) Women 63.5 (24.8) 192.9 (65.8) 23.4 (8.9) 4.9 (3.0) 8.3 (4.0) 292.6 (100.0)

When the present study is compared to the recent study of very similar. The differences in plant sterol intake could thus plant sterol intake in Finland by Valsta et al. (2004), the plant partly be explained by the higher energy intake among the sterol intake of women differed (293 vs 237 mg/day), while British women. the plant sterol intake of males were nearly identical (300 vs The study by Norme´n et al. (2001) of a cohort from the 305 mg/day), although a different dietary assessment meth- Netherlands used the same database as in this study apart od was used (24-h recall). Comparison of the British and the from some foods specific to the different countries. The Finnish women show that the British women reported to dietary assessment methods used in the present study and by have a higher energy intake (8.1 vs 6.8 MJ) and that their Norme´n et al. (2001) were both FFQs. Daily intake of plant plant sterol density seem to be higher (36.4 vs 34.9 mg/ sterol was higher in the British women than in the Dutch 1000 kJ), while the corresponding values for the males are women (293 vs 263 mg/day), while the corresponding intake

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 699

Figure 1 Energy-adjusted intake of plant sterols from different food groups (mg/1000 kJ) in men (n ¼ 11 227) and women (n ¼ 13 571) aged 40–79 years, EPIC Norfolk.

Table 2 Plant sterol density (mg/1000 kJ) by age group in men (n ¼ 11 227) and women (n ¼ 13 571) aged 40–79 years, EPIC Norfolk 1993–1997 (values are mean (s.d.))

40–52 years (mg/1000 kJ) P-level a 53–65 years (mg/1000 kJ) P-level b 66–79 years (mg/1000 kJ) P-level c One-way analysis of variance

Men 33.1 (7.3) NS 32.7 (7.5) NS 32.5 (7.5)** 0.005 Women 36.8 (7.4) NS 36.5 (7.7)*** 35.7 (7.9)*** o0.001

One-way analysis of variance post hoc Bonferroni’s t-test. aGroup 40–52 years vs 53–65 years. bGroup 53–65 years vs 66–79 years. cGroup 40–52 years vs 66–79 years. **Po0.01, ***Po0.001.

for the males were nearly identical (300 vs 307 mg/day). The comparable (34 mg). The British population had a lower energy intake was not reported for the Dutch men and intake of plant sterols from bread and other cereals than the women separately. The reported energy intake of the total Dutch and the Finnish population (54 mg vs over 105– Dutch population was marginally higher (0.05 MJ) than for 110 mg). When cakes, scones and chocolate were included in the British women, thus the Dutch women could be assumed the cereal group the intake in Britain still was lower than in to have a lower energy intake or at least a lower intake as the Netherlands and Finland (99 vs 110–120 mg) but higher estimated by FFQ than the British women. than in Spain (81 mg). Despite a few differences, the results Plant sterol intake from vegetables and legumes was higher are still remarkably similar across four different countries in Britain than in both the Netherlands and Finland (52 and with different dietary assessment methods. vs 30–35 mg), while the intake in Spain was comparable Valsta et al. (2004) found no age group differences in plant (45 mg). Also when compared as percentage intake from sterol density. Small but significant differences in plant sterol vegetables and legumes, the intake in Britain (18%) was density were found between different age groups in the higher than in the Netherlands (11%) and Finland (13%), Norfolk population. Younger men and women had a higher but comparable to Spain (16%). The plant sterol intake from plant sterol density compared to older men and women, fruits was also higher in Britain than in both the Netherlands despite the fact that the younger men and women had lower and Finland (36 vs 25–30 mg), while the intake in Spain was intakes of plant sterols from the important plant sterol

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 700

Figure 2 (a) Energy adjusted intake of plant sterols from different food groups (mg/1000 kJ) in men by age group, EPIC Norfolk 1993–1997. (b) Energy adjusted intake of plant sterols from different food groups (mg/1000 kJ) in women by age group, EPIC Norfolk 1993–1997.

sources fruits, vegetables and cakes, scones and chocolate, in many current smokers compared to UK could perhaps have relation to energy intake. affected the dietary intake. A recent meta-analysis of the The EPIC-Norfolk population was considered representa- dietary effects of smoking has concluded that smokers have a tive of the population of England in terms of anthropometric lower intake of polyunsaturated fat, fibre, vitamin C, vitamin variables, blood pressure and serum lipids (Day et al., 1999) E and b-carotene than non-smokers (Dyer et al., 2003). Foods but the fact that the population had approximately half as rich in polyunsaturated fat and fibre are generally higher in

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 701 plant sterols. Consequently, a lower intake of plant sterols spoonful of corn oil will increase the plant sterol intake from could be expected in the general British population. 23 to 137 mg of plant sterols. The FFQ used was validated against a 16-day weighed As plant sterols from the habitual diet has been shown to records and biomarkers and was considered appropriate to be inversely related to total and LDL cholesterol levels use as one of three nutritional methods in this large EPIC- (Andersson et al., 2004) more attention should be addressed Norfolk cohort (Bingham et al., 1997). The FFQ overesti- to these bioactive compounds in the prevention of CVD. mated vegetable, milk, cheese and coffee intake but only the In the study by Andersson et al. (2004), men consuming high overestimation of vegetables could have affected the plant amounts of plant sterols had a 0.25 mmol/l lower total serum sterol intake to any extent. Food frequency questionnaires cholesterol concentration and a 0.14 mmol/l lower LDL are used to assess individual habitual intake over a defined cholesterol concentration compared with those consuming period of time and are generally associated with different low amounts of plant sterols after adjusting for age, BMI types of bias like systematic intake-related bias, systematic and total energy intake. Corresponding values for women person-specific bias and random within-person errors (Kipis were 0.15 and 0.13 mmol/l lower levels of total and LDL et al., 2002). These types of bias affect the individual intake cholesterol, respectively. Despite these relatively small level although biases tend to skew towards unity. Since all differences in serum cholesterol concentration an effect on analyses are performed on group level the individual biases CVD and mortality might be expected, which will be further probably do not significantly affect trends. studied in this population. The knowledge of trends of plant Quality assurance of analytical work was performed during sterol intake from different food groups, between the sexes the set-up of the methodology and during the course of the and between age groups, concluded in this study could be analyses (Normeń et al., 2002). Duplicate sample analyses valuable when trying to optimize plant sterol intake in were repeated if there was a difference of more than 10%. In different populations. general, this variation was around 3%. Thus, the precision of the analytical method is considered adequate, and the accuracy has been determined by known standards during Conclusion analysis, and also in a multicentre comparison between five European laboratories engaged in plant sterol analysis The database of plant sterol values of analysed food items (Normeń, personal communication). along with the epidemiological nutritional data provided the Very little is known about the seasonal variation of plant basis for assessing the plant sterol intake and major plant sterols in fruits and vegetables. The fact that fruits and sterol food sources in a British population in 1993–1997. vegetables to a great extent are imported to Sweden from Mean (s.d.) intake of total plant sterols was 300 (108) mg/day several countries would have made it extremely hard to take for men and 293 (100) mg/day for women. Bread and other seasonal variation into consideration in the analyses, thus cereals, vegetables and added fats were the three major food this was not done. sources of plant sterols representing 18.6 (8.9), 18.4 (8.5) and The WHO, (2005) recommends an increased intake of 17.3 (10.4)% of the total plant sterol intake respectively. The fruits, vegetables, legumes, whole-grain cereals and nuts and intake or cakes, scones and chocolate contributed to 14.4 unsaturated fats in substitution of saturated fats for preven- (10.5)% of the total plant sterol intake whilst fruits tion of cardiovascular disease (CVD). Compliance with these contributed with 12.3 (8.4)%. recommendations would increase the intake of plant sterols. The results of the current study are comparable with recent Optimization of the plant sterol intake can be made by European studies, despite different nutritional assessment selection of foods particularly high in plant sterols, as the methods and slightly different databases. The serum choles- variation of plant sterol concentrations within food groups terol lowering effects of plant sterols from the diet are is high (Appendix 2). Among fruits and vegetables, high probably underestimated. Therefore more attention needs to concentrations of plant sterols are found in oranges and be addressed to the natural content of plant sterols in diets broccoli while apples and carrots contain much lower plant recommended for prevention of CVD. sterol concentrations. As the plant sterol content in different parts of the grain kernel varies, different milling processes and hence different cereal products contain varying amounts Acknowledgements of plant sterols. Cooking does not affect the plant sterol content of fruits, vegetables and cereals (Normeń et al., 1999, We thank the participants and general practitioners who 2002) while industrial processing of vegetable oils reduces took part in EPIC-Norfolk. EPIC–Norfolk is supported by plant sterol content by about 10% (Normeń, unpublished programme grants from the Cancer Research Campaign and data). Since some vegetable oils are especially high in plant Medical Research Council with additional support from the sterols and as the variation of plant sterol concentration is Stroke Association, British Heart Foundation, Food Standards very high, the choice of vegetable oil used in cooking can Agency, Department of Health and Wellcome Trust. have a great impact on the daily intake of plant sterols. The phytosterol analyses were supported by a grant from Exchanging one tablespoonful of olive oil for one table- the Swedish government under the LUA agreement and the

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 702 Swedish Cancer Foundation. This work was supported by Jimeńez-Escrig A, Santos-Hidalgo AB, Saura-Calixto F (2006). Com- FORMAS, the Swedish Research Council for Environment, mon sources and estimated intake of plant sterols in the Spanish diet. J Agric Food Chem 54, 3462–3471. Agricultural Sciences and Spatial Planning, Grant no 22.2/ Jonker D, van der Hoek GD, Glatz JFC, Homan C, Posthumus MA, 2003-0655 and Swedish Research Council, Grant no 521- Katan MB (1985). Combined determination of free esterified and 2003-3826. glycosilated plant sterols in foods. Nutr Rep Int 32, 943–951. Katan MB, Grundy SM, Jones P, Law M, Miettinen T, Paoletti R (2003). Efficacy and safety of plant stanols and sterols in the management of blood cholesterol levels. Mayo Clin Proc 78, 965–978. References Kipis V, Midthune D, Freedman L, Bingham S, Day NE, Riboli E et al. (2002). Bias in dietary-report instruments and its implications for Andersson SW, Skinner J, Ellega˚rd L, Welch AA, Bingham S, Mulligan nutritional epidemiology. Public Health Nutr 5, 915–923. A et al. (2004). Intake of dietary plant sterols is inversely related to Morton GM, Lee SM, Buss DH, Lawrence P (1995). Intakes and major serum cholesterol concentration in men and women in the EPIC dietary sources of cholesterol and phytosterols in the British diet. Norfolk population: a cross-sectional study. Eur J Clin Nutr 58, J Hum Nutr Diet 8, 429–440. 1378–1385. Normeń L, Brants HA, Voorrips LE, Andersson H, van den Brandt PA, Bingham SA, Cassidy A, Cole TJ, Welch A, Runswick SA, Black AE Goldbohm RA (2001). Plant sterol intakes and colorectal cancer et al. (1995). Validation of weighed records and other methods of risk in the Netherlands Cohort Study on Diet and Cancer. Am J dietary assessment using the 24 h urine nitrogen technique and Clin Nutr 74, 141–148. other biological markers. Br J Nutr 73, 531–550. Normeń L, Bryngelsson S, Johnsson M, Evheden P, Ellega˚rd L, Brants Bingham SA, Gill C, Welch A, Day K, Cassidy A, Khaw KT et al. H et al. (2002). The Phytosterol Content of Some Cereal Foods (1994). Comparison of dietary assessment methods in nutritional Commonly Consumed in Sweden and in the Netherlands. J Food epidemiology: weighed records vs 24 h recalls, food-frequency Comp Anal 15, 693–704. questionnaires and estimated-diet records. Br J Nutr 72, 619–643. Normeń L, Ellega˚rd L, Brants H, Dutta P, Andersson H (2007). A Bingham SA, Gill C, Welch A, Cassidy A, Runswick SA, Oakes S et al. phytosterol database: fatty foods consumed in Sweden and the (1997). Validation of dietary assessment methods in the UK arm of Netherlands. J Food Comp Anal 20, 193–201. EPIC using weighed records, and 24-hour urinary nitrogen and Normeń L, Johnsson M, Andersson H, Van Gameren Y, Dutta P potassium and serum vitamin C and carotenoids as biomarkers. Int (1999). Plant sterols in vegetables and fruits commonly consumed J Epidemiol 26, 137–151. in Sweden. Eur J Nutr 38, 84–89. Bingham SA, Welch AA, McTaggart A, Mulligan AA, Runswick SA, Ostlund Jr RE (2002). Phytosterols in human nutrition. Annu Rev Nutr Luben R et al. (2001). Nutritional methods in the European 22, 533–549. Prospective Investigation of Cancer in Norfolk. Public Health Nutr Piironen V, Toivo V, Lampi A-M (2000). Natural sources of dietary 4, 847–858. plant sterols. J Food Comp Anal 13, 619–624. Day N, Oakes S, Luben R, Khaw K-T, Bingham S, Welch A et al. (1999). Plat J, Mensink RP (2005). Plant stanol and sterol esters in the control EPIC-Norfolk: study design and characteristics of the cohort. Br J of blood cholesterol levels: mechanism and safety aspects. Am J Cancer 80 (Suppl. 1), 95–103. Cardiol 96, 15D–22D. Dutta PC, Normeń L (1998). Capillary column gas-liquid chromato- Valsta LM, Lemstro¨m A, Ovaskainen ML, Lampi AM, Toivo J, graphic separation of D5-unsaturated and saturated phytosterols. Korhonen T et al. (2004). Estimation of plant sterol and cholesterol J Chromatogr A 816, 177–184. intake in Finland: quality of new values and their effect on intake. Dyer AR, Elliott P, Stamler J, Chan Q, Ueshima H, Zhou BF (2003). Br J Nutr 92, 671–678. Dietary intake in male and female smokers, ex-smokers, and never Welch AA, Luben R, Khaw KT, Bingham SA (2005). The CAFE´ smokers: The INTERMAP Study. J Hum Hypertens 17, 641–654. computer program for nutritional analysis of the EPIC-Norfolk Food Standards Agency (2002). McCance andd Widdowson’s The food frequency questionnaire and identification of extreme Composition of Foods, Sixth summary edition. Royal Society of nutrient values. J Hum Nutr Diet 18, 99–116. Chemistry: Cambridge. World Health Organization (2005). Global strategy on diet, physical Fraser GE (1994). Diet and coronary heart disease: beyond dietary activity and health. cardiovascular disease: prevention and fats and low-density-lipoprotein cholesterol. Am J Clin Nutr 59 control. http://www.who.int/dietphysicalactivity/publications/facts/ (Suppl. 1), 1117–1123. cvd/en/index.html,20061221.

European Journal of Clinical Nutrition Food sources of plant sterols S Klingberg et al 703

Appendix 1 The plant sterol contents in six specific British food items and in a selection of foods are given in Table A1.

Table A1 Plant sterol content in six specific British food items (mg/100 g of fresh weight, edible portion)

Unsaturated Saturated Total plant sterols mg/100 g Campesterol Sitosterol Stigmasterol Campestanol Sitostanol mg/100 g mg/100 g mg/100 g mg/100 g mg/100 g

Baked beans 1.5 11.6 8.9 0 0.4 22.4 Tofu 14.2 40.7 12.2 0.8 1.8 69.7 Coffemate 3.8 21.6 2.7 0.9 5.3 34.3 Horlicks powder 7.4 20.4 1.1 1.1 12.5 42.5 Ovaltine powder 1.3 6.5 1.8 0 3.8 13.4 Marmite 0 1.7 0 0 0 1.7

The method used for analyses was the same as earlier published (Norme´n et al., 1999, 2002, 2007).

Appendix 2 Plant sterol content in a selection of foods (mg/100 g of fresh weight, edible portion) is given in Table A2.

Table A2 Plant sterol content in a selection of foods (mg/100 g of fresh weight, edible portion)

Food group Plant sterol content (mg/100 g edible portion)

Fruits Orange 24 Pineapple 17 Apple 13 Kiwi fruit 9.1

Vegetables Broccoli 39 Parsnip 27 Carrot 16 Tomato 4.7

Cereals Wheat bran, coarse 200 Wholemeal bread 86 Rolled oats 39 White bread 29

Vegetable oils Corn oil 909 Rape seed oil 668 Grape seed oil 215 Olive oil 154

Plant sterol content constitutes the sum of campesterol, b-sitosterol, stigmasterol, campestanol and b-sitostanol. The analyses have been published (Norme´n et al., 1999, 2002, 2007).

European Journal of Clinical Nutrition