ORIGINAL COMMUNICATION

Transcription

1 ORIGINAL COMMUNICATION (2002) 56, ß 2002 Nature Publishing Group All rights reserved /02 $ The macronutrient composition of diet and its relation with the traditional Mediterranean diet. Multi-centre study of the Mediterranean Group for the Study of Diabetes (MGSD) B Karamanos 1 *, A Thanopoulou 1, F Angelico 2, S Assaad-Khalil 3, A Barbato 2, M Del Ben 2, V Dimitrijevic-Sreckovic 4, P Djordjevic 4, C Gallotti 9, N Katsilambros 5, I Migdalis 6, M Mrabet 7, M Petkova 8, D Roussi 1 and MT Tenconi 9 1 Diabetes Center, 2nd Medical Department, Athens University Medical School, Hippokration Hospital, Athens, Greece; 2 Dipartimento di Terapia Medica, University La Sapienza, Rome, Italy; 3 Department of Internal Medicine, Alexandria Faculty of Medicine, Alexandria, Egypt; 4 Diabetes Center, Institute for Endocrinology, Diabetes and Metabolic Diseases, Clinical Center of Serbia, Belgrade, Yugoslavia; 5 Diabetes Center, 1st Department of Propedeutic Medicine, Athens University School of Medicine, Laiko General Hospital, Athens, Greece; 6 Diabetes Center, NIMTS Hospital, Athens, Greece; 7 Service de Medicine Interne A C.H.U. Oran, Algeria; 8 Diabetes Center St Luca, Sofia, Bulgaria; and 9 Dipartimento di Medicina Preventiva, Occupazionale e di Communita. Sezione di Igiene, Universita degli studi di Pavia, Italy Objective: To compare the nutritional habits among six Mediterranean countries and also with the various official recommendations and the Mediterranean diet as originally described. Design: Cross-sectional study. Settings: Three centres in Greece, two in Italy and one in Algeria, Bulgaria, Egypt and Yugoslavia. Subjects: Randomly selected non-diabetic subjects from the general population, of age 35 60, not on diet for at least 3 months before the study. Interventions: A dietary questionnaire validated against the 3-Day Diet Diary was used. Demographic data were collected and anthropometrical measurements done. Results: All results were age adjusted. Energy intake varied in men, from 1825 kcal=day in Italy Rome to 3322 kcal=day in Bulgaria and in women, from 1561 kcal=day in Italy Rome to 2550 kcal=day in Algeria. Protein contribution (%) to the energy intake varied little, ranging from 13.4% in Greece to 18.5% in Italy Rome, while fat ranged from 25.3% in Egypt to 40.2% in Bulgaria and carbohydrates from 41.5% in Bulgaria to 58.6% in Egypt. Fibre intake, g=1000 kcal, ranged from 6.8 in Bulgaria to 13.3 in Egypt and the ratio of plant to animal fat from 1.2 in Bulgaria to 2.8 in Greece. The proportion of subjects following the WHO and the Diabetes and Nutrition Study Group (DNSG) of the EASD recommendations for carbohydrates, fat and protein ranged from 4.2% in Bulgaria to 75.7% in Egypt. Comparison with the Mediterranean diet, as defined in the seven Country Study, showed significant differences especially for fruit, vs 464 g=day of the Mediterranean diet, meat, vs 35 g=day, cheese, vs 13 g=day, bread, vs 380 g=day. Conclusions: (a) Dietary habits of the normal population vary greatly among the Mediterranean countries studied. (b) Egypt is closest to the DNSG recommendations. (c) Significant differences from the originally described Mediterranean diet are documented in most Mediterranean countries, showing a Westernization of the dietary habits. (2002) 56, doi: /sj.ejcn Keywords: dietary habits; Mediterranean diet *Correspondence: B Karamanos, Department of Medicine, Athens University, Diabetes Centre, Hippokration Hospital, Vas. Sofias 114, Athens , Greece. ippokratio@aias.gr Guarantor: B Karamanos. Contributors: BK was principal investigator, secured the funding, collaborated in the protocol design, supervised the statistical analysis and wrote the manuscript. AT did the statistical analysis and wrote the manuscript. DR was responsible for analyzing the data from the dietary questionnaire, collaborated in the protocol design, contributed to data collection. FA, SAK, PD, CG, NK, IM, MM, MP, MTT collaborated in the protocol design and were involved in data collecting and in editing the manuscript. AB, MDB, VDS were involved in data collecting and in editing the manuscript. Received 12 June 2001; revised 5 January 2002; accepted 8 January 2002

2 984 Introduction The relation between diet and chronic diseases, such as cardiovascular disease (CVD), cancer etc is one of the major interests in preventive medicine (Ogawa et al, 1999; American Institute for Cancer Research, 1997). Studies of populations which have abruptly changed their dietary habits (both the amount and quality of food), suggest that switching to more Westernized diets is associated with an increase in the prevalence of diet-related diseases, like CVD (Keys, 1970). In some of these populations, an increase in the prevalence of diabetes has also been documented. However there is still controversy in the literature about the role of dietary habits in the development of glucose intolerance (West & Kalbfleisch, 1971; Pan et al, 1997), although there is evidence that high carbohydrate and high monounsaturated fat diets improve insulin sensitivity and glucose disposal (Perez-Jimenez et al, 2001). The Mediterranean diet is considered the golden standard of healthy nutrition and is associated with decreased morbidity and mortality, especially from cardiovascular causes (Keys, 1980; Keys et al, 1986; Helsing & Trichopoulou, 1989; Kafatos et al, 1991). The rapid changes in dietary habits in the Mediterranean area over recent years and the lack of comparative studies among Mediterranean countries, performed with the same protocol and methodology, urged the Mediterranean Group for the Study of Diabetes (MGSD) to organize the present multi-center study. The main aims were to compare the dietary habits among Mediterranean countries, both in the general and the diabetic population, define what exactly is the contemporary Mediterranean diet, compare the lipid profile among countries and investigate its relationship to dietary habits. The aim of the present paper is to investigate the differences in nutritional intake of the general population among five Mediterranean countries and compare current macronutrient intakes with the traditional Mediterranean diet and the nutritional recommendations of various international bodies. Subjects and methods A total of 4254 subjects were recruited in nine centres in six countries, according to the protocol which has been described in detail elsewhere (Karamanos, 1996). There were three centres in Greece, which were combined for the analysis, two centres in Italy and one centre in Algeria, Bulgaria, Egypt and Yugoslavia. The populations studied were, in Egypt and Greece, both urban and rural, in Algeria, Bulgaria and Yugoslavia mainly urban and in Italy mainly rural. In the Yugoslavian centre only diabetics were studied. In summary at least 150 subjects, randomly selected from a defined geographical region, were recruited in each centre. Moreover at least 150 subjects with diabetes mellitus (followed either in diabetes centres or by private doctors) were recruited from the same area. Inclusion criteria were age and for the non-diabetics a free diet for at least 3 months before the study. All subjects were asked not to change their regular diet before the examination and were studied after an overnight (12 h) fast. A patient report form was filled in, comprising demographic data, medical and family history and smoking habits. Anthropometrical measurements comprised height without shoes, weight with light clothing, waist circumference (midway between the rib cage and the iliac crest), hip circumference (maximal circumference between the iliac crest and the thigh region) and blood pressure in the sitting position, after 5 min rest (mean of two measurements). A 12 ml venous blood sample was taken and the plasma was immediately separated and stored at 7 18 C. All the samples were later shipped, in dry ice, to the Central Laboratory (Diabetes Centre, Athens, Greece), where all measurements were done. Plasma glucose, cholesterol, triglycerides and HDL-cholesterol were measured with standard autoanalyser methodology. A dietary history method was used for the evaluation of the dietary habits (Karamanos, 1993; Klaver et al, 1998). A dietary questionnaire comprising 78 questions was filled in by a specially trained person (dietician, nurse or physician). The questions referred to the daily or weekly consumption of various foodstuffs or groups of foodstuffs. The estimation of the amounts consumed was based on common household measures. The questionnaires were centrally analysed with a specially constructed computer program at the co-ordinating centre (Diabetes Centre, Athens, Greece). Information on food composition was given by each participating centre, especially for foods which were local specialities or had a peculiar composition (bulgur, couscous, etc). In order to validate the above dietary questionnaire, 100 subjects from various participating centres were also evaluated using the 3-Day Diet Diary (two weekdays and a Sunday), which is considered the gold standard (Bingham, 1987; Hartman et al, 1990). The comparison of the two methods, as previously described, is shown in Table 1 and the results (see Results) are good enough to allow the use of the dietary Questionnaire in this epidemiological study (Karamanos, 1996). The results were expressed as daily intake of energy in kcal, while the carbohydrate, protein, fat and alcohol intakes were expressed as percentage contribution to the daily Table 1 Comparison of the questionnaire with the 3-Day Diet Diary in 100 subjects Questionnaire P a 3-Day Diet Diary Energy (kcal=day) 1897 NS 1791 Carbohydrates (g=day) 207 < Proteins (g=day) 83 NS 80 Fat (g=day) 79 NS 81 Fibres (g=day) 21 < Animal fat (g=day) 35 NS 35 Plant fat (g=day) 41 NS 43 Cholesterol (mg=day) 263 NS 263 a Paired t-test.

3 energy intake. Total fibre intake was expressed both as g=day and g=1000 kcal of energy. Finally, the ratios of plant to animal protein and plant to animal fat intake were calculated. In one of the publications of the Seven Country Study (Kromhout et al, 1989), food items were grouped in 16 groups and expressed in g=day. In order to compare our findings with those of the Seven Country Study we further analyzed and grouped our data the same way. It must be noted that in the group of fat we included only what was consumed as pure fat, namely butter, oil added after cooking etc and not the fat content of the foodstuff or that added for cooking. Ethical approval for the study was obtained in each individual country, by the appropriate committee. In the final evaluation 3730 subjects, with complete data for the anthropometrical, biochemical and dietary parameters, were included. Of these there were 1833 non-diabetics, 916 male, 917 female and 1897 diabetics, 913 male, 984 female. The present paper concerns only the nutritional data of the non-diabetic population, from six centres in five of the countries, ie Algeria, Egypt, Italy, Bulgaria and Greece. Statistical analysis Data is presented as means standard error. The statistical analysis was performed with the SPSS statistical package. For continuous variables with normal distribution t-test was used for the comparison of two individual groups, while analysis of variance (ANOVA) was used for the evaluation of the differences among the centres. All tests were performed after age adjustment. Continuous variables with skewed distribution were log-transformed before statistical analysis. Categorical variables were evaluated with the chi-square test. Results Comparison of the dietary intake, in 100 subjects, evaluated both by the Questionnaire method used in this study and the 3 Day Diet Diary, is shown in Table 1. Significant differences were found only for carbohydrate, 207 vs 179 g=day and fibre intake, 21 vs 16 g=day, which were higher with the questionnaire method. All other parameters were not different and in fact very similar with both methods of evaluation. Analysis by centre (data not shown) showed exactly the same trends in each centre. Thus the questionnaire method was judged accurate enough for use in the present study. Age, body mass index (BMI), waist (W) and waist-to-hip ratio (W=H) are shown in Table 2, separately for men and women, in each centre. Age was slightly different among the various centres for both sexes, thus all data were age-adjusted for the rest of the comparisons. All variables in the table differed significantly among centres by ANOVA. The degree of obesity (BMI) varied among countries. Algerians were the leanest, while Greeks the fattest. Fat distribution also varied, but differently from the BMI, at least in women. By the criterion of W=H ratio, in men, Greece had the highest central obesity, in contrast to the findings in women where Egypt had the highest and Bulgaria the lowest. By the criterion of waist circumference, Greece showed the highest central obesity for both sexes, while the least central obesity for men was found in Egypt and, for women, in Pavia (Italy). Defining obesity as BMI 30.0 kg=m 2, the prevalence of obesity also varied among countries. The highest prevalence was observed in Greece both for men and women (30.4 and 35.5%, respectively) and the lowest in Algeria (6.1 and 7.5%) and Egypt (3.7% and 14.7%). In Table 3 the daily energy intake and the percentage contribution of each nutrient to it are shown, by sex and centre. All variables differed significantly among centres by ANOVA, after age-adjustment. The pattern of percentage contribution of each nutrient to the energy intake observed among countries was more or less the same for both sexes. However in most centres there were differences between men and women. Total energy intake was higher in men in all centres except in Egypt and Algeria, while energy intake expressed as calories per kg of body weight was higher in 985 Table 2 Characteristics of the subjects by centre and gender after age-adjustment (mean s.e.m.) Algeria Egypt Italy Pavia Italy Rome Greece Bulgaria P-value a Men n Age < BMI < W < W=H < Obesity % (BMI 30.0) < b Women n Age < BMI < W < W=H < Obesity % (BMI 30.0) < b a Comparison among groups with ANOVA. b Comparison among groups with the w 2.

4 986 Table 3 Daily energy intake and percentage contribution of each individual nutrient to it (mean s.e.m.) Energy (kcal=day) kcal=kg Protein (%) Fat (%) Carbohydrate (%) Alcohol (%) Fibre (g=day) Recommended values a < 30 > 50 > 30 Algeria M F Egypt M * * F Italy Pavia M *** *** *** *** *** F Italy Rome M *** ** * *** *** F Greece M ** *** *** F Bulgaria M *** * *** *** F P b < < < < < < < a By the Diabetes and Nutrition Study Group of the EASD. *P < 0.05, **P < 0.01, ***P < difference between sexes. b Comparison among centres with ANOVA. men only in Pavia. In men Rome had the lowest energy consumption, while Bulgaria had the highest. In women Rome had the lowest and Bulgaria together with Algeria the highest. The same pattern was observed when energy intake was considered in relation to body weight (kcal=kg). Fibre intake was lowest in Rome and highest in Egypt for both sexes. Protein contribution to the daily energy intake was lower in Greece and higher in Algeria and Rome for both sexes. Total fat contributed less than 27% of energy in Egypt, but more than 35% in Greece and Bulgaria, where the highest fat intake was observed for both sexes. Carbohydrates contributed more than 55% of the total energy in Egypt and less than 40% in Bulgaria. Alcohol contributed almost 10% of total energy in Italy and 5% in Greece and Bulgaria in men, while in women the contribution was less than 3%. Alcohol consumption was practically null in Egypt and Algeria. Comparing the percentage contribution of each nutrient to the total energy intake with that recommended by the Diabetes and Nutrition Study Group (DNSG) of the European Association for the Study of Diabetes (EASD), it can be seen that in all centres protein consumption was in the recommended range. However the recommended range for carbohydrates or fibre intake was obtained only in Egypt. Moreover the recommended fat percentage contribution was met again only in Egypt for both sexes and in Italy only in males. In Table 4 protein and fat intake are expressed as ratios of plant=animal origin. Daily fibre intake is also shown as g=1000 kcal of energy. The ratios differed significantly among centres, both for protein and fat. For protein the ratio ranged from the highest, 1.2, in Egypt to the lowest, 0.5, in Algeria and for fat the ratio ranged from 2.8 at the highest in Greece to 1.2 at the lowest in Bulgaria, Pavia and Algeria. No differences were noted between men and women. Fibre intake expressed as g=1000 kcal also differed among centres. It was highest in Egypt, Rome and Greece and lowest in Algeria and Bulgaria. In some centres fibre g=1000 kcal was significantly higher in women. In Table 5 the proportion of subjects in each country who met the recommendations of the DNSG, separately for each nutrient, is depicted. At least 70% of the subjects in each country met the recommendations for protein intake, while for fat intake the percentage ranged from 5 to 77%, for carbohydrates from 7 to 93% and for fibre from 3 to 61%. The proportion of subjects following the DNSG recommendations for all major nutrients (All-A), ie carbohydrates, fat and protein, ranged from 4.2% in Bulgaria to 75.7% in Egypt; however, those who met the recommendations for the above nutrients, including fibre (All-B), were 41 45% in Egypt, while in all other countries it was very low, ranging from 0 to 6.7%, because of the very low fibre consumption. Table 4 The ratio of plant animal intake for fat and protein, and fibre intake expressed as g=1000 kcal (mean s.e.m.) Algeria Egypt Italy Pavia Italy Rome Greece Bulgaria P a Plant animal protein M *** * ** < F < Plant animal fat M < F < Fibre (g=1000 kcal) M *** * *** < F < *P < 0.05, **P < 0.01, ***P < differences between sexes. a Comparison among centres with ANOVA.

5 Table 5 Percentage of subjects who meet the nutritional recommendations of the Diabetes and Nutrition Study Group (DNSG) of the EASD, by centre and gender 987 Percentage of the subjects who meet the recommendations Recommended intake (DNSG) 10 20% protein 30% fat 50% carbohydrate 30 g fibre All-A a fibre All-B b Algeria M F Egypt M F Italy Pavia M F Italy Rome M F Greece M F Bulgaria M F a Recommendations followed for all nutrients except fibre, ie protein, fat and carbohydrate. b Recommendations followed for all nutrients, including fibre. The qualitative composition of the diet observed in each centre is shown in Table 6 in comparison with the recommendations of the DNSG and the WHO and also the composition of the Mediterranean diet, as specified in 1993 (Trichopoulou et al, 1993a,b). It can be seen that Greece and Bulgaria were closer to the Mediterranean diet, but Bulgaria had a very low plant=animal fat ratio. Egypt had a high plant=animal fat ratio, but the contribution of fat to the energy intake was low, while that of carbohydrates was very high. However Egypt is the only country perfectly complying with the DNSG and the WHO recommendations. In the lower part of the table, the countries are grouped according to the population studied, urban or rural. It is clear that the differences among centers for all nutritional parameters persist even when urban and rural populations are compared separately. In Table 7 the various foods consumed are grouped in 16 more or less homogeneous groups and the intake is expressed as g=day. The groups were constructed in the same way as in the seven Countries Study in order to be able to compare the results of the two studies. Large variations were observed among centres. The biggest variations Table 6 Comparison of the percentage contribution of each nutrient to the energy intake with the recommendations of the DNSG, the WHO and the Mediterranean diet as originally described Protein (%) Fat (%) Carbohydrate (%) Fibre (g=day) DNSG a < 30 Plant animal > 50 > 30 WHO b fat ratio Alcohol (%) Recommended values MED=AN c diet Algeria Egypt Italy Pavia Italy Rome Greece Bulgaria Urban Algeria Egypt Greece Bulgaria Rural Egypt Italy Pavia Italy Rome Greece a Diabetes and Nutrition Study Group of the EASD, (1988, 1995, 2000). b WHO (1990). c Trichopoulou et al (1993a), b.

6 988 Table 7 Comparison of the actual diet in the Mediterranean countries in grams per day of various foodstuffs and the Mediterranean diet as originally described Bread Cereals a Potatoes Legumes Vegetables Alcohol Fruit Meat Fish Eggs Cheese Milk Fats Sugar Pastries Rest MD-I b (g=day) MD-GR b (g=day) Algeria M F Egypt M F Italy Pavia M F Italy Rome M F Greece M F Bulgaria M F a Cereals, pasta, rice. b Mediterranean diet according to the seven-country Study for Greece and Italy (Kromhout, 1989). were observed in sugar intake, which was 3 4-fold higher in Algeria compared to Rome, fruit intake, which was 3-fold higher in Greece compared to Egypt, legume intake, which was 4 5-fold higher in Egypt compared to Algeria, and potatoe intake, which was 4-fold higher in Bulgaria compared to Rome. In the Seven Countries Study only males were studied and the results of that study for Greece and Italy are shown on the top two rows of the table. In the present cohort, Greek males consumed more meat, cereals, cheese and sugar and less bread, potatoe, fruit, eggs and fat compared to those studied in the sixties, while in Italy males consumed more legumes, fruit, pastries and less bread, vegetables, alcohol and meat. Discussion This is the only recent study, which describes and compares the dietary habits of a large number of free-living subjects in Mediterranean countries, studied with the same methodology. So far there have only been data deriving from small studies, using differing methodologies and not primarily aimed at the comparison of dietary habits among countries. The methodology used for the dietary assessment has been validated against the 3-Day Diet Diary, which is considered the gold standard (Bingham, 1987; Hartman et al, 1990; EURONUT SENECA investigators, 1991) and it compared satisfactorily. Protein, cholesterol fat and types of fat intakes were identical with both methods. The questionnaire overestimated carbohydrate intake by 10% in all centres. The use of the same questionnaire and especially the same technique of completing it, by specially trained personnel, ensured that the small overestimation of carbohydrates was the same in all cohorts, rendering the results absolutely comparable among countries. Moreover the magnitude of the sample ensured a rather representative cohort from each country. All the comparisons were done after age adjustment, rendering data perfectly comparable. Thus the findings of the present study reflect fairly well the nutritional habits of each country and can be used both for the comparisons among the countries and the evaluation of the contemporary Mediterranean diet. The energy intake varied among the countries, in both sexes. However the differences in energy intake did not coincide with the prevalence of obesity, which also varied among countries. This may reflect differing degrees of exercise, a parameter not measured in this study. The qualitative composition of the diet also varied. The most striking differences were observed in the contribution of fat to the energy intake. Such differences in fat intake among countries in Europe have already been documented in type 1 diabetic patients in the EURODIAB study (Toeller et al, 1996). The high fat intake in Greece is in line with the findings of the EURONUT SENECA Study (EURONUT SENECA, investigators, 1991). The differences in the qualitative composition of the diet did not concern only the nutrients, but also their source of origin. For example Greece and Bulgaria had almost the same contribution of fat to the energy intake, however the plant=animal fat ratio was twice as high in Greece. As a matter of fact Greece showed by far the highest plant=animal fat ratio compared to all other countries, denoting a high consumption of foods of plant origin, a fact corroborated by the high ratio of plant=animal proteins as well. This is in accordance with the findings of previous studies in Greece, where a high consumption of foods of plant origin, ie vegetables, fruit and olive oil has been documented (Willett et al, 1995; Trichopoulou & Lagiou, 1997). Apart from fat intake, fibre and carbohydrate consumption also differed significantly among centres. Thus, at present dietary habits differ in many aspects among Mediterranean countries studied and the differences persist even when urban and rural populations are compared separately. They most probably reflect geographic, cultural and socio-economic variations among countries. Within countries some differences exist between

7 urban and rural populations only in Egypt, while in the more Westernized countries dietary habits tend to be the same in urban and rural areas. We compared the present data with the Nutritional Recommendations of: (a) WHO, referring to the general population (WHO, 1990); (b) the Joint Task Force of the European Society of Cardiology, European Atherosclerosis Society and European Society of Hypertension, referring mainly to the prevention of CHD (Wood et al, 1998; Wood, 1999); and (c) the Recommendations of the Diabetes and Nutrition Study Group (DNSG) of the EASD, which refer mainly to the diabetic population, but also apply to the general population (DNSG of the EASD, 1988,1995,2000). We must note that the recommendations of WHO and the DNSG are quite similar. Considering mean values for each country, protein contribution to the energy intake is in the recommended range in all countries. However carbohydrate contribution and fibre consumption are in the recommended range only in Egypt. Finally fat contribution to the energy intake is in the recommended range only in Egypt and Italy. The percentage of subjects in each country who comply with the Nutritional Recommendations of the DNSG for each individual nutrient varies greatly. Most subjects in all countries follow the recommendations concerning proteins, but there are great variations for fat and fibre, which are in the ranges 7 93, 5 77 and 3 61% respectively. Moreover the percentage of subjects complying with the above recommendations for all items, ie protein, fat, carbohydrate and fibre intake, is extremely low in all centres (0 6.7%), except for Egypt, where it reaches 45.6%. This very low figure is mainly due to the low fibre intake in most countries. However even if only the three major nutrients were considered, the compliance hardly reached 25% of subjects. Thus, if the DNSG and WHO Recommendations actually reflect the healthy diet, this is followed by very few subjects in the Mediterranean countries today. On the other hand, the Joint Task Force has set other goals for nutrients and foods for the general population (fruit and vegetables > 400 g=day, legumes > 30 g=day, fish > 20 g=day and fibre g=day). All countries studied meet these goals, apart from fibre intake. Fat could not be compared since the Joint Task Force sets goals for saturated and polyunsaturated fat, while in the present study the quality of fat was evaluated only by plant-to-animal fat ratio. Thus, more or less, the Mediterranean countries studied were within the range of the less strict goals set by the Joint Task Force. It is very difficult to set the standards for a healthy diet. The various nutritional recommendations are based on the results of epidemiological, clinical and experimental studies. Some of them focus on the effects of individual nutrients on individual health problems, ie lipids and atherosclerosis, proteins and nephropathy etc, while others focus on the relation of nutritional habits in general to morbidity and mortality. Studies of the last type provide the most valid information for the definition of the healthy diet (Wood et al, 1998). The concept of the healthy Mediterranean diet was based on data from the Seven Country Study (Keys et al, 1968; Keys, 1980; Kromhout et al, 1989). In this study approximately men, aged were medically examined in 16 centres, in seven countries. Information on diet was collected in random samples of subjects in each centre. Specially trained dieticians observed and recorded all food items consumed over a week. Moreover duplicate meals (exactly what was consumed over the week of observation) were homogenized and analyzed in a central laboratory. With this procedure the most accurate estimation of the dietary habits was obtained. In some countries follow-up examinations were also done after y. The cohort studied in Crete (Greece) had the lowest prevalence of coronary heart disease and this has been attributed, at least in part, to the dietary and exercise habits of the population. Thus the concept of the healthy, cardioprotective, Mediterranean diet emerged. Further studies in Greece, conducted between 1980 and 1988 in both sexes, in urban and rural areas, comprising in total more than 1000 subjects, confirmed the composition of the Mediterranean diet as shown on Table 6 (Kafatos et al, 1991; Trichopoulou et al, 1993a,b). The health benefits of this diet comprising both lower morbidity and mortality, have been attributed to (a) high intake of monounsaturated fat rather than a low fat intake in general, since fat intake contributed more than 40% of the energy, (b) high complex carbohydrate intake, mainly grains and legumes and (c) high fibre intake, mostly from vegetables and fruit (Trichopoulou et al, 1995a,b; Kouris-Blazos et al, 1999; Lagiou et al, 1999). Some other, less well-defined components of the diet, as polyphenols, alpha linolenic acid etc, may also contribute to the cardioprotective effects of the Mediterranean diet (Trichopoulou et al, 1995a,b). The present study reveals differences between the contemporary dietary patterns in the Mediterranean countries and the traditional Mediterranean diet. This is based on the study of a large number of subjects of both sexes, with the same methodology validated against the 3-Day Diet Diary, and thus the results reflect with acceptable accuracy the nutritional patterns in each country. Fat contribution to the energy intake, compared to the Mediterranean diet, is lower in most countries, presumably because of low consumption of fat of plant origin, since the plant animal fat ratio is significantly lower than the 2 : 1 of the traditional Mediterranean diet. Fibre intake is generally low, while protein and carbohydrate contribution to the energy intake shows only small deviations from the Mediterranean diet. Considering individual foodstuffs the bigger deviations from the traditional Mediterranean diet concern the increase in meat and cheese and the decrease in fruit consumption. Among the countries studied only the contemporary Greek diet seems to be closer to the traditional Mediterranean diet, 989

8 990 with a very high plant to animal fat ratio, high fruit intake but with lower fibre intake. Conclusions Considerable differences in the qualitative composition of the diet among Mediterranean countries have been detected and moreover the nutritional habits in most countries are not in agreement with the widely accepted Nutritional Recommendations, but more importantly, with the traditional Mediterranean diet, which is the only one associated with decreased morbidity and mortality. Longitudinal follow-up studies are needed, with endpoints morbidity and mortality, in order to establish how healthy the present Mediterranean diets are and probably the need to return to the traditional Mediterranean diet. Participating centres Professor B Karamanos (Co-ordinator), Drs A Thanopoulou and D Roussi Diabetes Center, 2nd Medical Department, Athens University Medical School, Hippokration Hospital, Athens, Greece. Professor F Angelico, Drs M Del Ben and A Barbato Dipartimento di Terapia Medica, University La Sapienza, Rome, Italy. Professor S Assaad Khalil Department of Internal Medicine, Alexandria Faculty of Medicine, Drs Alexandria, Egypt. Professor P Djordjevic, Drs V Dimitrijevic-Sreckovic and F Canovic Diabetes Center, Institute for Endocrinology, Diabetes and Metabolic Diseases, Clinical Center of Serbia, Belgrade, Yugoslavia. Professor N Katsilambros Diabetes Center, 1st Department of Propedeutic Medicine, Athens University School of Medicine, Laiko General Hospital, Athens, Greece. Dr I Migdalis Diabetes Center, NIMTS Hospital, Athens, Greece. Drs M Mrabet, S Zarat and A Bentounsi Service de Medicine Interne A CHU Oran, Algeria. Dr M Petkova Diabetes Center St Luca, Sofia, Bulgaria. Professor MT Tenconi and Dr C Gallotti Dipartimento di Medicina Preventiva, Occupazionale e di Communita. Sezione di Igiene. Universita degli studi di Pavia, Italy. Acknowledgements We thank all the subjects who took part in this study. The study is part of the Multicenter Nutritional Study of the Mediterranean Group for the Study of Diabetes (MGSD) and it was financially supported by (a) the MGSD, (b) Novo Nordisk A=S and (c) Farmaserve Hellas. We would like also to thank Ms Elli Antoniadou and Ms Martha Panagiotopoulou for secretarial support, Ms Petroula Maraki for laboratory assistance and Ms Christina Leoussi for statistical help. References American Institute for Cancer Research (1997): Food Nutrition and Prevention of Cancer: a Global Perspective. World Cancer Research Fund. Bingham SA (1987): The dietary assessment of individuals: methods, accuracy, new techniques and recommendations. Nutr. Abstr. Rev. 57: Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes (1988): Nutritional recommendations for individuals with diabetes mellitus. Diabetes Nutr. Metab. 1, Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes (1995): Nutritional recommendations for individuals with diabetes mellitus. Diabet. Nutr. Metab. 8, Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes (2000): Nutritional recommendations for individuals with diabetes mellitus. Eur. J. Clin. Nutr. 54, EURONUT SENECA investigators (1991): Intake of energy and nutrients. Eur. J. Clin. Nutr. 45(Suppl 3), Hartman AM, Brown CC, Palmgren J et al (1990): Variability in nutrient and food intakes among older middle-aged men. Implications for design of epidemiologic and validation studies using food recording. Am. J. Epidemiol. 132, Helsing E & Trichopoulou A (eds) (1989): The Mediterranean diet and food culture: a symposium. Eur. J. Clin. Nutr. 43(Suppl 2), Kafatos A, Kouroumalis I, Vlahonikolis C & Labadarios D (1991): Coronary heart disease risk factor status of the Cretan urban population in 1980s. Am. J. Clin. Nutr. 54, Karamanos B (1993): Diet and diabetes In: Diabetes in Europe, ed. R Williams, L Papoz & J Fuller, p John Libbey: London. Karamanos B (1996): Interim report of the Nutrition Study Group of the Mediterranean Group for the Study of Diabetes (MGSD). Medicographia 18(Suppl 1), Keys A (1970): Coronary heart disease in seven countries. Circulation XLI(Suppl 1), Keys A (1980): Seven Countries: a Multivariate Analysis of Death and Coronary Heart Disease. Cambridge, MA: Harvard University Press. Keys A, Aravanis C & Sdrin A (1968): The diets of middle-aged men in two rural areas of Greece. In: Dietary Studies and Epidemiology of Heart Disease, ed. C Den Hartog, K Buzina, F Fidanze, A Keys & P Roine (eds). Rotterdam: Wyt, Keys A, Menotti A, Karvonen MJ et al (1986): The diet and 15-year death rate in the Seven Countries Study. Am. J. Epidemiol. 124, Klaver W, Burema J, Van Staveren WA & Kuniman JT (1998): Definition of terms In: Manual for Methodology for Food Consumption Studies ed. ME Cameron & WA Van Staveren, P 11. New York: Oxford University Press. Kouris-Blazos A, Gnardellis C, Wahlqvist ML, Trichopoulos D, Lukito W & Trichopoulou A (1999): Are the advantages of the Mediterranean diet transferable to other populations? A cohort study in Melbourne, Australia. Br. J. Nutr. 82, Kromhout D, Keys A, Aravanis C, Buzina R, Fidanza F & Giampaoli S (1989): Food consumption patterns in the 1960s in seven countries. Am. J. Clin. Nutr. 49, Lagiou P, Trichopoulou A, Henderickx HK, Kelleher C, Leonhauser IU, Moreiras O, Nelson M, Schmitt A, Sekula W, Trygg K & Zajkas G (1999). Eur. J. Clin. Nutr. 53, Ogawa K, Tsubono Y, Nishino Y, Watanabe Y, Ohkubo T, Watanabe T, Nakatsuka H, Takahashi N, Kawamura M, Tsuji I & Hisamichi S (1999): Inter- and intra-individual variation of food and nutrient consumption in a rural Japanese population. Eur. J. Clin. Nutr. 52, Pan XR, Li GW, Hu YH, Wang JX, Yang WY, An ZX, Hu ZX, Lin J, Xiao JZ, Cao HB, Liu PA, Jiang YY, Wang Jp, Zheng H, Zhang H, Bennett PH & Howard BV (1997): Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care 20, Perez-Jimenez F, Lopez-Mranda J, Pinillos MD, Gomez P, Paz-Rojas E, Montilla P, Marin C, Velasco Mj, Blanco-Molina A, Jimenez & Ordovas JM (2001): A Mediterranean and a high-carbohydrate diet improve glucose metabolism in healthy young persons. Diabetologia, 44,

9 Toeller M, Klischan A, Heitkamp G, Schumacher W, Milne R, Buyken A, Karamanos B, Gries FA, and the EURODIAB IDDM Complications Study Group (1996): Nutritional intake of 2868 IDDM patients from 30 centres in Europe. Diabetologia. 39, Trichopoulou A, Lagiou P (1997): Healthy traditional Mediterranean diet: an expression of culture, history and lifestyle. Nutr. Rev. 55, Trichopoulou A, Katsouyanni K & Gnardellis Ch (1993a): The traditional Greek diet. Eur. J. Clin. Nutr. 47(Suppl 1), S76 S81. Trichopoulou A, Toupadaki N, Tzonou A, Katsouyanni K, Manousos O, Kada E & Trichopoulos D (1993b): The macronutrient composition of the Greek diet: estimates derived from six case-control studies. Eur. J. Clin. Nutr. 47, Trichopoulou A, Kouris-Blazos A, Vassilakou T, Gnardellis C, Polychronopoulos E, Venizelos M, Lagiou P, Wahlqvist ML & Trichopoulos D (1995a): Diet and survival of elderly Greeks: a link to the past. Am. J. Clin. Nutr. 61(6 Suppl), 1346S 1350S. Trichopoulou A, Kouris-Blazos A, Wahlqvist M, Gnardellis Ch, Lagiou P, Polychronopoulos E, Vassilakou T, Lipworth L & Trichopoulos D (1995b): Diet and overall survival in elderly people. Br. Med. J. 311, West K & Kalbfleisch J (1971): Influence of nutritional factors on prevalence of diabetes. Diabetes 20, WHO (1990): Diet, nutrition and the prevention of chronic disease. Report of a WHO Study Group. WHO Technical Report Series 797. Geneva: WHO. Willett WC, Sacks F, Trichopoulou A, Drescher G, Ferro-Luzzi A, Helsing E & Trichopoulos D (1995): Mediterranean diet pyramid: a cultural model of healthy eating. Am. J. Clin. Nutr. 61(6 Suppl), 1402S 1406S. Wood D (1999): Guidelines a missed opportunity. Atherosclerosis 143(Suppl 1), S7 S12. Wood D, De Backer G, Faergeman O, Graham I, Mancia G, Pyorala K and the Members of the Task Force (1998): Task Force Report. Prevention of coronary heart disease in clinical practice: Recommendations of the Second Joint Task Force of European and other Societies on Coronary Prevention. Atherosclerosis 140,