Coronary disease THE FRENCH PARADOX: LESSONS FOR OTHER COUNTRIES Correspondene to: Professor Jean Ferrières, Department of Epidemiology, INSERM U558, University Shool of Mediine, 37, allées Jules Guesde, 31073 Toulouse edex, Frane; ferriere@it.fr T Jean Ferrières Heart 2004; 90:107 111 Life is the art of drawing suffiient onlusions from insuffiient premises Samuel Butler he Frenh paradox is the observation of low oronary heart disease (CHD) death rates despite high intake of dietary holesterol and saturated fat. 1 2 The Frenh paradox onept was formulated by Frenh epidemiologists 3 in the 1980s. Frane is atually a ountry with low CHD inidene and mortality (table 1). The mean energy supplied by fat was 38% in Belfast and 36% in Toulouse in 1985 86. 4 More reently, in 1995 97, the perentage of energy from fat was 39% in Toulouse aording to a representative population survey. THE FRENCH PARADOX AND CAUSES OF DEATH The first soure of error ould ome from an underestimated CHD mortality. Aording to this hypothesis, Frenh physiians may not delare all the CHD deaths as CHD. If standardised data for example those provided by the MONICA (monitoring of trends and determinants in ardiovasular disease) projet are used, the results onerning CHD attak and mortality rates show that Frane is at a low risk for CHD (table 1). Under ertifiation of CHD deaths in Frane is a possible bias, but after orretion, it remains a low bias. Thus, validated data on CHD mortality and inidene show that Frane is haraterised by CHD risk, orroborating the first part of the Frenh paradox definition. THE FRENCH PARADOX AND SATURATED FAT In orrelation studies, measures that represent harateristis of an entire population (onsumption of animal fat, daily milk, and alohol) are used to desribe disease (CHD mortality). Limitations of orrelational studies are the inability to link exposure with disease in partiular individuals, the lak of ability to ontrol the effets of potential onfounding fators, and the use of average exposure levels rather than atual individual values. Numerous orrelation studies have been arried out in various ountries onerning the relation between onsumption of fat and CHD mortality. 1 2 In one of the most interesting ones, Artaud-Wild and olleagues examined the relation of CHD mortality to the intake of foodstuffs and nutrients in 40 ountries. 2 After having defined a holesterol saturated fat index (CSI), they studied the relation between the CSI and CHD mortality (per 100 000 men aged 55 64 years) for all 40 ountries (fig 1). Frane had a CSI of 24 per 1000 kal and a CHD mortality rate of 198, whereas Finland had a CSI of 26 per 1000 kal and a CHD mortality rate of 1031. In the seven ountries study, 12 763 men from 16 ohorts in seven ountries were examined for CHD risk fators in 1958. Information on diet was olleted by use of seven day food reords. 5 The average onsumption of animal food groups, with the exeption of fish, was positively assoiated with 25 year CHD mortality rates. Furthermore, the average population intake of saturated fat was strongly related to 10 and 25 year CHD mortality rates. It is interesting to ompare the World Health Organization dietary reommendations for saturated fat (, 10% of total energy) with atual intakes in Frane. In representative ross setional surveys of the Frenh population performed in 1986 87 and 1995 97, the saturated fat intake was 15% of the total energy intake in the first survey and 16% in the latter survey. This high onsumption of saturated fatty aids is suh that Frenh subjets are exposed to a high risk of CHD. Why a high onsumption of saturated fatty aids does not lead to a high CHD risk in Frane (and maybe elsewhere) is a entral question behind the Frenh paradox onept. The Frenh paradox is a way of presenting provoative results from epidemiologial studies and does not take into aount ausality between risk fators and CHD mortality. 107 Heart: first published as 10.1136/heart.90.1.107 on 15 Deember 2003. Downloaded from http://heart.bmj.om/ on 11 May 2018 by guest. Proteted by opyright. www.heartjnl.om
108 Table 1 Age standardised oronary heart disease (CHD) mortality and event rate in seleted European regions (men, aged 35 64 years) Population Offiial CHD mortality rate per 100000* THE FRENCH PARADOX AND WINE DRINKING Even if ausality is not part of the Frenh paradox, some authors have put forward several hypotheses in order to explain it. Debates have foused on alohol onsumption and, more speifially, on red wine. A superfiial investigation of the lifestyles of south European ountries, haraterised by low CHD inidene, onfirms wine as one of their ommon identifying elements. In a seminal paper, 1 Renaud and De Lorgeril ompared age standardised annual mortality from CHD and related risk fators in MONICA populations, inluding the Frenh entres. The mean serum total holesterol onentrations were similar in Frane, in the USA, and the UK. After having performed a regression analysis between death rate from CHD and onsumption of dairy fat and wine, these authors onluded that the Frenh paradox may be aused by a high onsumption of wine. Their artile has started a long series of researh studies analysing the relation between wine and CHD. Final results do depend on the ultural bakground of the ountries where they have been arried out. The most striking illustration is the lassi ontrast between ultures with moderate and regular alohol onsumption and those where binge drinking is ommon. It would take too long to report CHD mortality per 100000À Glasgow (UK) 332 365 777 Belfast (UK) 280 279 695 Lille (northern Frane) 89 172 298 Strasbourg (north eastern Frane) 80 141 292 Toulouse (south western Frane) 53 91 233 Barelona (north eastern Spain) 63 76 210 Coronary events per 100 000` *Based on death ertifiate enumeration. ÀFatal events inluded definite, possible, and unlassifiable (mainly sudden deaths with no available diagnosti information) oronary deaths. `Coronary events inluded non-fatal events (definite myoardial infartion) and fatal events. Figure 1 Plot of death rate from oronary heart disease (1977) orrelated with daily dietary intake (from 1976 to 1978) of holesterol and saturated fat as expressed by the holesterol fat index (CSI) per 1000 kal. Reprodued from Artaud-Wild et al, 2 with permission. on all the studies dealing with the relations between alohol and CHD. 6 The only lear message is that moderate alohol drinking (two or three times a day) has a protetive effet against CHD. Alohol intake raises high density lipoprotein (HDL) holesterol onentrations and approximately 50% of the risk redution attributable to alohol onsumption is explained by hanges in HDL holesterol. However, the differential effets of wine, beer, and spirits have been examined. European researh arried out in Frane and Denmark has shown that wine onsumption has been assoiated with a derease of 24 31% in all ause mortality; little to moderate wine drinking leads to lower mortality from ardiovasular disease than an equivalent onsumption of beer or spirits. 7 Whih mehanisms are responsible for the potential different effets of the various types of alohol? We have to turn towards urrent theories on atherogenesis. The dominant theory is the oxidative modifiation hypothesis. Oxidised lipids, espeially phospholipids, generated during low density lipoprotein (LDL) oxidation or within oxidatively stressed ells, are the triggers for many of the events seen in developing lesions. 8 Flavonoids are omponents of a wide variety of edible plants, fruit, and vegetables and of beverages suh as tea, offee, beer, and wine. In vitro, inhibition of LDL oxidation by flavonoids derived from red wine has been demonstrated. In men, onsumption of red wine with meals redues the suseptibility of human plasma and LDL to lipid peroxidation. Red wine ontains a variety of polyphenols derived from grape skins. In free living subjets, 9 onentration of atehin in plasma was threefold higher in a diet with fruit and vegetables but without wine, than a diet without fruit, vegetables, and wine, and fourfold higher than in the latter in a diet with red wine but without vegetables and fruit. It seems that red wine and the Mediterranean diet may onfer an additional protetive effet. It is possible that, after adjustment for total alohol intake, wine onsumption might be assoiated with a more favourable lipid profile than beer onsumption: higher HDL holesterol, lower triglyerides and fibrinogen onentrations. Indeed, alohol onsumption and assoiated lifestyles may have an effet on CHD above and beyond their impat on lipids. The impat on haemostasis has also been debated. 1 6 In the Caerphilly prospetive heart disease study, platelet aggregation indued by adenosine diphosphate was inhibited in subjets who drank alohol. Furthermore, in the Western group of the Lyon diet heart study, there was an inverse relation between ethanol (wine represented 88% of total ethanol) and ex vivo platelet aggregation; this orrelation was absent in the Mediterranean group. Platelet aggregation Heart: first published as 10.1136/heart.90.1.107 on 15 Deember 2003. Downloaded from http://heart.bmj.om/ on 11 May 2018 by guest. Proteted by opyright. www.heartjnl.om
was low and the same in both groups, suggesting that platelet aggregation was low in the Mediterranean group beause of the diet (rih in a linoleni aid) and in the Western group (diet relatively rih in saturated fat) beause of wine. The effet on blood pressure is worthy of attention, deserving an analysis. In Frane, alohol onsumption is rather homogeneous throughout the week whereas in Northern Ireland, Fridays and Saturdays aount for 66% of total alohol onsumption. 10 After adjustment for all the lassial risk fators, blood pressure levels are higher in drinkers in Northern Ireland on Monday and derease until Thursday whereas blood pressure levels are onstant throughout the week in Frenh drinkers. 10 One an onlude that the binge drinking pattern observed among drinkers in Northern Ireland leads to disadvantageous onsequenes regarding blood pressure levels, onsistent with a higher inidene of CHD on Mondays and Tuesdays and a lower inidene on Fridays in the Belfast MONICA register. The ethanol ontent of alohol beverages seems to be a determining protetive fator against CHD. However, the effets of regular alohol onsumption are omplex. So it an be onluded that it is diffiult to ompare the effets of wine onsumption in populations with low intake or with different wine onsumption patterns with populations of wine drinkers showing very different lifestyles (see below). THE FRENCH PARADOX AND THE VARIABILITY OF CHD The MONICA projet has provided an overview of CHD event rates in 37 populations from 1985 to 1993 and these data show a large variability in CHD event rates depending on persons, plae, and time. Some ountries present low or high CHD inidene and dissimilar CHD mortality trends (fig 2). Analysis of fig 2 shows that Frane has low CHD inidene, but not very dissimilar to that observed in other ountries suh as Belgium (Ghent) or Spain (Barelona). Moreover, fig 2 shows that ountries exhibit different CHD mortality trends. Some regions with rather low CHD attak rates show a signifiantly delining trend in CHD mortality (Ghent, Toulouse, Strasbourg) or a more limited derease (Augsburg, Lille) or even an inrease (Barelona). Some regions with high CHD attak rates present dissimilar trends in CHD mortality, a signifiant derease (Belfast), a more limited derease (Glasgow), or even a signifiant inrease (Kaunas, Tarnobrzeg). In a reent study in Oxfordshire, 11 the CHD attak rates in men were 273 per 100 000, muh lower than rates reported by MONICA entres in Glasgow and Belfast and similar to rates reported by MONICA entres in Frane and northern Italy. Frane has atually low CHD inidene and, in reality, it is very lose to Germany, Italy or Spain. New paradoxes may emerge in CHD mortality and in CHD ase fatality trends or in epidemiologial situations within eah ountry. In any ase, whatever the terms Frenh paradox, North-South paradox, or Mediterranean paradox the most diffiult issue is not the desription but the explanation. Reently, the Framingham risk funtion has been tested on other populations. In these studies, it has overestimated absolute oronary risk in diverse populations: Japanese, Hispani males living in the USA; Italian males in Italy; Danish and Frenh males and females. 12 The general onlusion was that the Framingham models should not be used to predit absolute CHD risk in populations with different inidenes. A more appropriate CHD risk funtion ould be produed loally. However, a single risk funtion derived in one plae at a partiular time may not be appliable elsewhere and must be adjusted for geographial and temporal fators. Furthermore, the ontribution of lassial risk fators to the trends in CHD over 10 years aross the WHO MONICA projet populations has been studied: variability of CHD event rates have been explained by trends of major risk fators in only 15% in women and 40% in men. These results suggest that other CHD risk fators should be measured or disovered, partiularly in ountries with a low CHD inidene. THE FRENCH PARADOX AND INTERVENTION STUDIES In the Cohrane Collaborative meta-analysis of 27 randomised intervention trials lasting more than six months, no signifiant effet was shown with redued or modified dietary fat on overall mortality, ardiovasular mortality or ardiovasular events after sensitivity analysis. 13 In another review, 14 Hu and Willet aknowledged that simply lowering the perentage of energy from total fat in the diet is unlikely to improve lipid profile or redue CHD inidene. On the ontrary, studies using a whole diet approah showed signifiant and onlusive results. The Lyon diet heart study 15 randomised 605 Frenh patients with CHD to either a diet rih in fruit, vegetables, nuts and a linoleni aid-rih margarine or their usual diet (a prudent Western diet). After two and four years, the intervention diet group had signifiantly redued ardiovasular end points, although Figure 2 Relation between trends in oronary heart disease mortality rates and oronary heart disease event rates in seleted MONICA populations 1985 to 1993 (men aged 35 64 years). 109 Heart: first published as 10.1136/heart.90.1.107 on 15 Deember 2003. Downloaded from http://heart.bmj.om/ on 11 May 2018 by guest. Proteted by opyright. www.heartjnl.om
110 CHD risk fators were omparable between groups before and after enrolment. The Indo-Mediterranean diet heart study 16 randomised 1000 patients with CHD or surrogate risk fators to either a diet rih in whole grains, fruit, vegetables, walnuts, and almonds, or their loal diet. The mean intake of a linoleni aid was signifiantly greater in the intervention group. After two years, total ardia end points were signifiantly fewer in the intervention group than the ontrols. The spetaular results obtained in the Lyon and the Indo-Mediterranean diet heart studies, and in other studies, onern mainly seondary prevention. Nevertheless, the Frenh paradox refers to primary prevention and to lifestyles firmly rooted in ountries with low CHD inidene. But what do we know about the relation between CHD risk fators, lifestyles, and CHD risk? THE FRENCH PARADOX AND THE DIET HEART HYPOTHESIS The lassial risk fators are those of CHD found in Frane and any other ountry in western Europe. However, signifiant differenes in CHD risks are observed for similar CHD risk fator levels. In the seven ountries study, at a holesterol value of 5.2 mmol/l, the CHD mortality rates were five times higher in northern Europe than in Mediterranean southern Europe. Within Europe, similar relative risks for CHD in relation to holesterol are to be observed, but with notably different absolute risks. 17 In the same study, the authors observed that the relative inrease in 25 year mortality from CHD for a given inrease in blood pressure was similar among the populations. However, at systoli and diastoli blood pressure of about 140 and 85 mm Hg, respetively, 25 year rates of mortality from CHD varied by a fator of more than three among the populations, with high rates in the USA and northern Europe and low rates in Japan and Mediterranean ountries. In the Gerona provine (Spain), a high prevalene of risk fators ontrasted with a low inidene of CHD. 18 The prevalene of total holesterol, blood pressure, smoking, and obesity in the Gerona provine were lower than those of North Karelia (Finland) but higher than those of Minnesota (USA). The authors stated that the Frenh paradox ould be extended to the paradoxial oexistene of high prevalene of risk fators and low CHD inidene in southern Europe. 18 Other risk fators may partly aount for the differenes of CHD risk between northern and southern European ountries. In the MONICA projet, there was a signifiant assoiation between CHD mortality and plasma homoysteine, Toulouse (Frane) having one of the lowest values of the latter. For some researhers, Frane and other regions have a high onsumption of fruit and vegetables, rih soures of folate, and this may explain the Frenh paradox. Fibrinogen is a strong risk marker for CHD. In a prospetive ohort (PRIME) study involving 10 600 men living in four regions (Lille, Strasbourg, and Toulouse in Frane, Belfast in Northern Ireland), the assoiation of plasma fibrinogen with the inidene of CHD was signifiant even after adjustment for other CHD risk fators. These fators explained 25% of the exess risk of CHD in Belfast ompared with Frane, while fibrinogen alone aounted for 30%. 19 Physial ativity patterns ould also explain, at least partly, the gradient in CHD observed between these two ountries. In the same PRIME study, subjets with higher, regular, leisure time physial ativity energy expenditure had a lower inidene of CHD over a five year follow up. Although total physial ativity levels were similar, Frenh men performed more leisure time physial ativity, whih may ontribute to their lower inidene of CHD in omparison with Northern Ireland. The diet heart hypothesis is a very important but rather omplex onept. The above debate about the relation between wine, other alohols, and CHD inidene is at the entre of the diet heart hypothesis. Indeed, all the faets of wine drinkers behaviour annot be summed up in wine onsumption. In Frane, total alohol, wine, and beer onsumption was negatively related to soioeonomi status and eduational level; on the other hand in Northern Ireland, total alohol, beer, and spirits onsumption was negatively related whereas wine onsumption was positively related to soioeonomi status and eduational level. Data indiate that even within Frane, the total amount and the onsumption of the different types of aloholi beverages vary onsiderably aording to the geographial region. Thus, onsidering the relation between CHD and alohol onsumption per se, not taking into aount the general nutritional pattern of the population studied, might lead to spurious results, sine differenes in alohol onsumption might also reflet differenes in nutritional intake. For example, in Toulouse there is a higher onsumption of wine than in Belfast, but also a higher onsumption of vegetables and fruit. 4 Similar results were obtained in a ross setional study in the USA. The US researhers suggested that people who onsume wine have health habits that ould influene other health outomes, and their behaviour might result in a lower inidene of aner and all ause mortality. Central to the diet heart hypothesis are the attitudes to food and its role in life. In an interesting study, 20 adults and ollege students from Flemish Belgium, Frane, USA, and Japan were surveyed and questioned about beliefs and attitudes onerning the diet health link. Generally, the US group assoiated food most with health and least with pleasure, while the Frenh group were the most food pleasure orientated and the least food health orientated. The authors onluded that diet health and food onerns do differ aross ultures and may exert a negative influene on health. 20 So far, in this artile, we have barely mentioned the roles played by soial epidemiology and psyhology in CHD. A reently published study exploring the question of whether psyhosoial variables ould partly explain the Frenh paradox, within the Northern Irish-Frenh PRIME study, provided little support for the psyhosoial risk hypothesis. However, other interesting studies suggest that ardiovasular reativity an predit the development of some prelinial states and perhaps new linial events. THE FRENCH PARADOX AND PUBLIC HEALTH IMPLICATIONS Some fats ited in this artile are inonsistent with the known risk fators for CHD and are thus paradoxial. These paradoxes are neither linked to a speifi ountry nor to speifi mehanisms operating in atheroslerosis. This review has not disussed the role of geneti fators, despite the fat that gene environmental interations are probably fundamental in atheroslerosis. However, urrently available data and, more speifially, the hanging trends in CHD inidene and mortality rates support the hypothesis of a fundamental role played by environmental and behavioural fators. The Frenh paradox onept disusses lassial risk fators and probably numerous protetive fators whih are speifi to determined regions and largely unidentified. The Frenh Heart: first published as 10.1136/heart.90.1.107 on 15 Deember 2003. Downloaded from http://heart.bmj.om/ on 11 May 2018 by guest. Proteted by opyright. www.heartjnl.om
The Frenh paradox: key points A high intake of dietary holesterol and saturated fat but low CHD death rates define the Frenh paradox Variability of CHD rates is the rule, and low CHD rates are observed in southern or Mediterranean European ountries Classial risk fators do not embrae the totality of CHD risk, partiularly in Frane and in other southern European ountries Complex behaviour onerning wine drinking and attitudes to food ould lower CHD inidene The Frenh paradox suggest that the promotion of primary prevention, based on an optimal diet rih in fruit and vegetables, regular physial exerise, and life without smoking, is worthwhile The Frenh paradox is an inentive for more researh in ountries with low CHD inidene and probably more protetive CHD risk fators paradox does not teah other ountries any lessons sine eah ountry has probably its own appropriate preventive solutions. Like Butler, we must end by drawing suffiient onlusions from insuffiient premises. In primary prevention, adherene to lifestyle guidelines is assoiated with a very low risk of CHD. The Frenh paradox onept should foster researh on protetive CHD risk fators while onveying, in primary prevention, messages promoting healthy behaviour suh as regular physial exerise, optimal diet, and life without smoking. Time is still needed to define new fators in the field of gene environment interations or soial epidemiology, or to provide an even better definition of optimal diet and regular physial exerise. REFERENCES 1 Renaud S, de Lorgeril M. Wine, alohol, platelets, and the Frenh paradox for oronary heart disease. Lanet 1992;339:1523 6. This referene gives the definition of the Frenh paradox and suggests the inhibition of platelet reativity by wine may be one explanation for protetion from CHD in Frane. 2 Artaud-Wild SM, Connor SL, Sexton G, et al. Differenes in oronary mortality an be explained by differenes in holesterol and saturated fat intakes in 40 ountries but not in Frane and Finland. A paradox. Cirulation 1993;88:2771 9. Frane and Finland have similar intakes of holesterol and saturated fat, but onsumption of vegetables and vegetable oil ontaining monounsaturated and polyunsaturated fatty aids is greater in Frane than in Finland. 3 Rihard JL, Cambien F, Duimetière P. Epidemiologi harateristis of oronary disease in Frane. Nouv Presse Med 1981;10:1111 4. 4 Evans A, Ruidavets JB, MCrum E, et al. Autres pays, autres oeurs? Dietary patterns, risk fators and ishaemi heart disease in Belfast and Toulouse. QJM 1995;88:469 77. The lassial CHD risk fators do not explain the large differenes in CHD between Belfast (UK) and Toulouse (Frane), and total fat onsumption do not differ between the entres. 5 Kromhout D, Keys A, Aravanis C, et al. Food onsumption patterns in the 1960s in seven ountries. Am J Clin Nutr 1989;49:889 94. 6 Chadwik DJ, Goode JA, eds. Alohol and ardiovasular diseases: Novartis Foundation Symposium 216. New York: John Wiley and Sons, 1998:272. 7 Gronbaek M, Beker U, Johansen D, et al. Type of alohol onsumed and mortality from all auses, oronary heart disease, and aner. Ann Intern Med 2000;133:411 9. A study of the relation between different types of alohol and mortality. Wine may have a benefiial effet on all ause mortality attributable to a redution in death from both CHD and aner. 8 Steinberg D. Atherogenesis in perspetive: hyperholesterolemia and inflammation as partners in rime. Nat Med 2002;8:1211 7. 9 Ruidavets JB, Teissedre PL, Ferrières J, et al. Catehin in the Mediterranean diet: vegetable, fruit or wine? Atheroslerosis 2000;153:107 17. 10 Marques-Vidal P, Arveiler D, Evans A, et al. Different alohol drinking and blood pressure relationships in Frane and Northern Ireland. The PRIME study. Hypertension 2001;38:1361 6. In Frane, alohol onsumption is rather homogeneous throughout the week. Binge drinking patterns observed among Northern Ireland drinkers leads to high blood pressure, onsistent with a higher inidene of myoardial infartion on Mondays and Tuesdays. 11 Volmink JA, Newton JN, Hiks NR, et al. Coronary event and ase fatality rates in an English population: results of the Oxford myoardial infartion inidene study. The Oxford myoardial infartion inidene study group. Heart 1998;80:40 4. 12 Menotti A, Lanti M, Puddu PE, et al. Coronary heart disease inidene in northern and southern European populations: a reanalysis of the seven ountries study for a European oronary risk hart. Heart 2000;84:238 44. 13 Hooper L, Summerbell CD, Higgins JP, et al. Redued or modified dietary fat for preventing ardiovasular disease. Cohrane Database Systemati Revue 2001;(3):CD002137. 14 Hu FB, Willett WC. Optimal diets for prevention of oronary heart disease. JAMA 2002;288:2569 78. 15 de Lorgeril M, Renaud S, Mamelle N, et al. Mediterranean alpha-linoleni aid-rih diet in seondary prevention of oronary heart disease. Lanet 1994;343:1454 9. The first demonstration of the effiay of a Mediterranean diet on various outomes in seondary prevention. 16 Singh RB, Dubnov G, Niaz MA, et al. Effet of an Indo-Mediterranean diet on progression of oronary artery disease in high risk patients (Indo- Mediterranean diet heart study): a randomised single-blind trial. Lanet 2002;360:1455 61. A final onfirmation that a Mediterranean diet rih in a linoleni aid indues a better prognosis in patients with CHD. 17 Kromhout D. On the waves of the seven ountries study: a publi health perspetive on holesterol. Eur Heart J 1999;20:796 802. Within Europe, similar relative risks for CHD in relation to holesterol are observed, but with notably different absolute risks: high risks in northern Europe and low risks in the Mediterranean southern Europe. 18 Masia R, Pena A, Marrugat J, et al. High prevalene of ardiovasular risk fators in Gerona, Spain, a provine with low myoardial infartion inidene. REGICOR investigators. J Epidemiol Community Health 1998;52:707 15. The prevalene of CHD risk fators in Gerona (Spain) is higher than that of Minnesota (USA), where the CHD inidene is three times higher. 19 Sarabin PY, Arveiler D, Amouyel P, et al. Plasma fibrinogen explains muh of the differene in risk of oronary heart disease between Frane and Northern Ireland. The PRIME study. Atheroslerosis 2003;166:103 9. 20 Rozin P, Fishler C, Imada S, et al. Attitudes to food and the role of food in life in the U.S.A., Japan, Flemish Belgium and Frane: possible impliations for the diet-health debate. Appetite 1999;33:163 80. A study of the attitudes to food in different ountries. The Amerians, who do the most to alter their diet, are the least likely to lassify themselves as healthy eaters. Additional referenes appear on the Heart website http://www.heartjnl.om/supplemental 111 Heart: first published as 10.1136/heart.90.1.107 on 15 Deember 2003. Downloaded from http://heart.bmj.om/ on 11 May 2018 by guest. Proteted by opyright. www.heartjnl.om