Letter to the Editor The fall of the traditional lipid hypothesis Kari Salminen This article deals with the latest developments and recommendations concerning hypercholesterolae- mia and its dietary treatment inFinland. In the light of recent studies, the Finnish target value for blood cholesterol level, less than 5 mmol/1 for the whole population, seems to be too low. Also the replace- ment of animal fats in the diet with margarines manufactured by catalytic hydrogenation, as the latest recommendation suggests, is more than ques- tionable. The most recent findings support the hypothesis that consumption of partially hydrogen- ated vegetable and marine oils may contribute to the occurrence of cardiovascular heart diseases. Therefore the Finnish recommendation of Decem- ber 1992, should be reviewed. The so-called lipid hypothesis has been a pre- dominant issue in the controversy over therole of fats in human nutrition during the past thirty years. According to the traditional hypothesis 1) a high blood cholesterol level increases the risk ofcoronary heart disease (CHD) morbid- ity and mortality, and 2) the saturated fatty acids in the diet raise, and polyunsaturated fatty acids lower, the blood cholesterol level. According to the lipid hypothesis it would thus be logical to limit the intake of saturated fats and increase the intake of polyunsaturated fats from the diet, as numerous international and national recom- mendations have suggested. The development of the situation especially inFinland over the past few years is reviewed below. Finnish recommendation of 1988 In autumn 1988 an ad hoc group of the Finnish Association for Internal Medicine and The Finnish Cardiac Society published the recommendation "Diagnosis and treatment of hypercholesterolaemia and other hyperlipidaemias in adults" (1). The tar- get value for blood cholesterol was set at less than 5 mmol/1 for the entire adult population regardless ofage or sex. At the time only one in five Finns had a cholesterol level below the set limit. Classified according to the recommendation, about half of the population had a "moderately" enhanced risk of CHD with cholesterol levels of 5-6.5 mmol/l, and one third a "considerably" en- hanced risk, as their cholesterol level exceeded 6.5 mmol/1. Consequently, on the basis of the recommenda- tion, four in five Finns were deemed to require some kind of treatment to lower their cholesterol levels. Lowering the cholesterol by dietary means was emphasized in the first place, but in cases where other risk factors were found, medical treat- ment was recommended in addition. By 1990, the sales of cholesterol lowering drugs had grown 30- fold from the year preceding the recommenda- tion (2). The dietary recommendations emphasized the replacement of saturated animal fats with unsatur- ated vegetable oils and fats, thus placing particular pressure on the use ofmilk fat. After publication, the recommendation and espe- cially its strict limits for cholesterol values aroused certain criticism. It was, however, as late as 1992 thatthe Finnish Medical Journal published a critical editorial (3) and some articles on the subject (4, 5). New recommendation in 1992 In December 1992 another ad hoc group assigned by five institutions, published a new recommenda- tion on blood serum cholesterol and its diagnosis and treatment (6). The risk values for cholesterol regarding CHD were now somewhat changed. The recommendation considered the range 5- 6.5 mmol/1 as bearing a "mild", range 6.5-8 "mod- erate" and >8 mmol/1 a "considerably" enhanced risk of CHD. The target value for a single individ- ual as well as for the whole population remained at below 5 mmol/1. In suggesting ways to reach the target value, the new recommendation took a slightly more cautious 447 Agric. Sci. Finl. 2 (1993) https://www.c-info.fi/en/info/?token=zzancOfRdxHLW264.dPYNeSE2o3L9SouTuHPUdA.fqM1ED1ZUml30asUtjy0Md_7sdnVnOD6ua3lMcdu2K9DWQrYl5udC8WA5DVlI2hbwDXjG6UFm9KHWooWNyoaiNPsrnVhddoxzM5k6oCjGwQF_GKX3jAJuB4Yy8sGe8fVBDNx9truK45sblb-gcb-XITybkjNJJV2EUF6XHF8CoxZpthRPcMZIsBsK5gXwhOJXibjKL8m_awTGKNXigZhzhidZg3aw9rOLVJDIEHwfiMQxBh227RPKdBEQQ2U9Rfl4osiLQ attitude towards medication than the previous one. However, the suggested dietary measures and the recommendation for decreasing milk fat intake were, if possible, even stricter. Milk product con- sumption was to favour low fat or fat-free products, and butter was to be changed to margarine at the table and to margarine or vegetable oil in baking and frying. Weakness of the lipid hypothesis Blood cholesterol vs. cardiovascular and total mortality NHLBI conference. A conference on "Low Blood Cholesterol: Mortality Associations" arranged by the National Heart, Lung and Blood Institute, NHLBI (7) included presentations from 19 cohort studies from the United States, Europe, Israel and Japan. The material consisted of 650 000 individu- als including reports of more than 68 000 deaths. Fig. 1 shows therelationship between mortality and serum cholesterol levels of men and women. The curves illustrating the relationship between blood cholesterol and cardiovascular mortality on the one hand and total mortality on the other are not identical. Neither are the cholesterol-mortality curves for males and females. This fact was very clearly stated at the NHLBI conference. Based on the conference report, the editorial of Circulation, "Health Policy on Blood Cholesterol. Time to Change Directions", dealt with the choles- terol dilemma (8) drawing the following three con- clusions. 1) There is an association between low blood cholesterol and noncardiovascular deaths in men and women. The curve illustrating the relationship between blood serum cholesterol level and mortality is U-shaped (Fig. 1). The mortality of the population increases both at low and high cholesterol values, as illustrated by the rising limbs of the U-shape. The bottom part of the U-shape describes an "opti- mum" cholesterol level where mortality is lowest. The left-hand limb indicates mainly non-CHD mor- tality and the right-hand limb mainly CHD mortal- ity. Figure I shows how the mortality of men quite clearly follows the U-shape. The rising left-hand limbreflects increasing non-CHD mortality and the right-hand increasing CHD mortality. The curve illustrating the mortality of women is less clearly U-shaped. The rising left-hand limb in the range of the low cholesterol level illustrates non-CHD mortality which is on the increase. The right-hand limb(or, rather, its absence) will be dealt with in point 2). Referring to the first conclusion the editorial noted that it might be time to review national poli- cies aimed at shifting the distribution of blood cho- lesterol of the entire population towards the left. A cholesterol-lowering diet may not be prudent for those adults whose cholesterol levels place them on the left-hand limb of the total mortality U-shape. 2) There is no association between high blood cholesterol and cardiovascular deaths in women. This message can be clearly seen in Fig. 1. In contrast with the evidence for men, a high choles- terol content in women does not reflect CHD mor- tality. Owing to this there is no clearly rising right- hand limb in the U-shape for women. This conclusion introduces a strong female per- spective to the cholesterol controversy to the effect that it no longer seems wise to screen for and treat Fig. 1. U-shape describing the association between blood cholesterol level and mortality. 448 Letter to the EditorAgric. Sei. Fint. 2(1993) high blood cholesterol in women. An exception maust naturally be made with women who already have coronary disease or who otherwise are at a high risk ofCHD death. 3) Primary prevention trials of cholesterol inter- vention reveal an increase in non-CHI) death rates that is similar in magnitude to the decrease in CHI) death rates. According to the editorial, the above three con- clusions indicate a need to change the current cho- lesterol recommendations. Cholesterol measure- ments and treatments should be limited to the mi- nority of the population to whom the benefits would be clearly greater than the disadvantages. Thus, measures should be directed only at people who already have a coronary heart disease or a sub- sequent risk of CHD. Important Sll Mobile Clinic Survey. The Mobile Clinic Survey carried out by the Finnish Social Insurance Institution (SII) should be mentioned in this connection (9). It reviews comprehensively the relationship between blood values, including cho- lesterol values, and mortality in the Finns. The results of the survey are for the most part similar to those of the NHLBI conference report. The authors of the SII Mobile Clinic Survey write about blood serum cholesterol levels in the Finns as follows: "Optimum values in men appeared to be roughly between 5 and 7 mmol/1, and in women between 6 and 9 mmol/1." One-sided Finnish recommendation. The Finnish recommendation (6) differs markedly from the views introduced by the NHLBI conference and the SII Mobile Clinic Survey. It is based on a study on the cholesterol levels and CHD mortality in Amer- ican middle-aged men. The recommendation thus concentrates primarily on theright-hand limb of the U-shape mortality curve of men (darkened in Fig. 1), but the conclusions are, however, applied to the entire Finnish population. This cannot and should not be done. The editorialofCirculation ends as follows: "We need now to pull back our national policies directed at identifying and treating high blood cholesterol in theprimary prevention setting and put on hold well- meant desires to intervene whilewe await convinc- ing evidence that the net effects will be beneficial." The above shows that the Finnish recommenda- tion looks at our national health from a very narrow angle. It is based only on the cholesterol vs. CHD mortality curve of middle-aged men. Dietary fatty acids and hlood cholesterol The second main argument of the lipid hypothesis, which claims that saturated fatty acids raise and polyunsaturated fatty acids lower blood serum cho- lesterol level, is very vague. The P/S ratio, i.e. the ratio of the dietary polyunsaturated fatty acids to saturated fatty acids, has been a cornerstone of the Finnish dietary recommendations. It is now losing its significance as a result of long-lasting discus- sions (10, 11). While the P/S ratio still occupied a central role in the 1988 recommendation, it was omittedfrom the 1992 recommendation. According to the current conception: - Only lauric, myristic and palmitic (Cl2-Cl6) acids of saturated fatty acids raise the total and LDL cholesterol levels. They evidently also raise the HDL cholesterol level. - Polyunsaturated fatty acids lower the total and LDL cholesterol levels. A high intake decreases the HDL cholesterol level. The earlier opinions emphasizing the beneficial effects of polyunsatur- ated fatty acids have become more cautious also because of the oxidation tendency of those acids. - The lipid hypothesis did not take into account monounsaturated fatty acids, especially oleic acid. According to the current opinion it lowers both the total and LDL cholesterol levels without any real effect on the HDL cholesterol. However, it is the occurrence of trans isomers of fatty acids, especially those of oleic acid, formed in the catalytic hydrogenation of vegetable oils, that now really shakes the lipid hypothesis. 449 Agric. Sci. Pint. 2 (1993)Letter to the Editor Letter to the Editor Partially hydrogenated vegetable oils and trans fatty acids. Katan and his collaborators have shown that the trans isomers ofoleic acid formed in hydro- genation raise the total and LDL cholesterol and lower the HDL cholesterol (12, 13). Some other studies (14, 15, 16, 17) have reached a similar conclusion. A recent article published in Lancet (18) has really caused a sensation. Harvard scientists calcu- lated the intake of trans fatty acids from dietary questionnaires completed by 85 095 women with- out diagnosed CHD, stroke, diabetes, or hypecho- lesterolaemia in 1980. During 8 years offollow-up, there were 431 new cases of CHD. After adjust- ment for age and total energy intake, the intake of trans isomers was found to be directly related to a risk of CHD. Additional control for established CHD risk factors, for example saturated fat and dietary cholesterol, did not change the relative risk substantially. The association was stronger for the 69 181 women whose margarine consumption over the previous 10 years had been stable. Intakes of foods that are major sources of trans isomers (mar- garine, cookies, cake, and white bread) were each significantly associated with higher risks of CHD. The researchers concluded that these findings sup- ported the hypothesis that consumption of partially hydrogenated vegetable oils may contribute to the occurrence of CHD. In a study on ischaemic heart disease and con- sumption of hydrogenated marine oils, Thomas showed that there was an association between the disease and the trans fatty acids formed in catalytic hydrogenation (19). The Finnish recommendation of 1992 (6) underes- timates the significance of trans fatty acids and recommends the use of margarines. If the recom- mendation were followed, the intake of trans fatty acids from the Finnish diet would increase substan- tially. The European recommendation (1992) for its part states; "Hydrogenation of vegetable oils, a common industrial practice, produces trans fatty acids whichbehave similarly to saturated fats (20). Their intake shouldbe minimised." Thus hard mar- garines and products containing hydrogenated oils should be avoided. Table 1 shows the fatty acid compositions of three best-selling Finnish margarines, commercial milk fat-rapeseed oil mixtures and traditional but- ter. According to the current conception of the health effects of fatty acids, the milk fat-rapeseed oil mixtures are superior to hydrogenated margar- ines. Surprisingly the Finnish recommendation does not even mention these products which were, in fact, developed by the country’s own dairy in- dustry. In the light of these most recent findings, the prudence of the Finnish recommendation concern- ing dietary fats can be questioned. Conclusion The Finnish recommendation "Diagnosis and treatment of hypercholesterolaemia and other hy- perlipidaemias in adults"( 6) is mainly based on the traditional lipid hypothesis. The expediency of the recommendation should be reviewed in the light of the latest studies on diet, blood serum cholesterol and mortality. According to this information: 1. The mortality of the population increases both at high and at low cholesterol levels. The U-shape (Fig. 1) illustrating mortality vs. blood cholesterol level is different for men and women. Hence, the optimum blood cholesterol range should be set for men and for women separately. The present target value, below 5 mmol/1 for the entire Finnish popu- lation, is not reasonable. 2. The trans fatty acids formed in the catalytic hydrogenation process of polyunsaturated fatty acids have an unfavourable effect on blood choles- terol and may contribute to the occurrence of CHD. Thus the Finnish recommendation to increase the use of margarines is not well-founded. There are considerably better alternatives on the Finnish yel- low fat market (Table 1). Kari Salminen ValioLtd. Research and Development P.O. Box 390 FIN-00101 Helsinki, Finland 450 Agric. Sei. Fin!. 2(1993) Letter lo the Editor Table 1. Fatty acid compositions (%) of three best-selling margarines, milk fat-rapeseed oil mixtures and butter. Saturated Trans Total Mono- Polyunsaturated Sat. & Trans unsaturated C| 2 -C l6 C| g.„ Oleic acid Linoleic acid Linolenic acid Margarine 1 14 12 26 24 33 3 Margarine 2 25 7 32 31 II 4 Margarine 3 15 11 26 23 35 2 Milk t'at-rapeseed oil mixtures Mixture I" 19 2 21 39 14 Mixture 2 h 26 2 28 33 10 Mixture 3 C 34 3 37 28 3 Butter 43 4 47 20 1 1 “ 40% milk fat, 60% rapeseed oil h 60% milk fat, 40% rapeseed oil • 80% milk fat, 20% rapeseed oil References I. Suomen Sisätautilääkärien Yhdistyksen ja Suomen Kar- diologisen Seuran työryhmän suositus. 1988. 19p. 2. Finnish Statistics on Medicines from 1987-1991. The Finnish Committee on Drug Information and Statistics. Helsinki 1988, 1989, 1990, 1991, 1992. 3. Pasanen, A. 1992. Sepelvaltimotaudin ennaltaehkäisy arvioitava uudelleen. Suom. Lääkäri!. 47: 1120-1121. 4. Salminen, K. 1992. Onko aika määritellä kolesterolin tavoitetasot uudestaan? Suom. Lääkäril. 47: 1192-1193, 5. Nikkilä, M. 1992. Kenen poikkeavia lipidiarvoja pitäisi hoitaa? Suom. Lääkäril. 47: 1952-1956. 6. Frick, M.H., Huttunen, J., Huvinen, M,, Katila, M., KesA-Niemi, A., Kovanen, P., Miettinen, T., Olkinuora, J., Pyörälä, K,. Salmi, J., Savolainen, M., Taskinen. M.-R., Tikkanen, M., Uusitupa, M., Viikari, J. & Winell, K. 1992. Aikuisten hyperkolesterolemian ja muiden hyperlipidemioiden diagnostiikka ja hoito. Duodecim 109: 205-225. 7. Jacobs, D., Blackburn, H., Higgins, M., Reed, D., Iso, H„ McMillan, G., Neaton, J., Nelson, J., Potter, J., Rifkind, 8.. Rossouw, J., Shekelle, R. & Yusuf, S. 1992. Report of the Conference on Low Blood Choles- terol: Mortality Associations. Circulation 86: 1046- 1060. 8. Hulley, 5.8., Walsh, J. & Newman, T.B. 1992. Health Policy on Blood Cholesterol. Time to Change Direc- tions. Editorial. Circulation 86: 1026-1029. 9. Aromaa, A. & Maatela, J. 1981. A search for optimum values. Prognostic evaluation of reference values. In: Gräsbeck, R. & Ahlström, T. (eds.). Reference values in laboratory medicine. John Wiley & Sons. 10. Salminen, K. 1988. P/S-suhde, harhaanjohtava suure. Duodecim 104: 1562-1563. 11. Salminen, K. 1990. Ravinnon tyydyttyneet ja tyydyt- tymättömät rasvat. Duodecim 106: 1684-1685. 12. Mensink, R. & Katan, M. 1990. Effect of dietary trans fatty acids on high density and low density lipoprotein cholesterol levels in healthy subjects. N. Engl. J. Med. 323; 439-445. 13. Zock, P.L. & Katan, M.B. 1992. Hydrogenation alter- natives: effects of trans fatty acids and stearic acid versus linoleic acid on serum lipids and liporoteins in humans. J. Lipid Res. 33: 399-410. 14. Nestel, P., Noakes, M, Belling, 8., McArthur, R., Clifton, P., Janus, E. & Abbey, M. 1992. Plasma lipo- protein lipid and Lp(a) changes with substitution of elaidic acid for oleic acid in the diet. J. Lipid Res. 33: 1029-1036. 15. Troisi, R„ Willett, W.C. & Weiss, S.T. 1992, Trans- fatty acid intake in relation to serum lipid concentrations in adult men. Am. J. Clin. Nutr. 56: 1019-1024. 16. Lichtenstein, A.H., Ausman, L.M., Carrasco, W., Jen ner, J.L., Ordovas, J.M. & Schaefer, E. 1993. Hydro- genation Impairs the Hypolipidemic Effect of Corn Oil in Humans. Arterosclerosis and Thrombosis 13: 154- 161. 17, Wood, R., Kubena, K., O’Brien, 8., Tseng, S. & Mar- tin, G. 1993. Effect of butter, mono- and polyunsatu- rated fatty acid-enriched butter, trans fatty acid margar- ine, and zero trans fatty acid margarine on serum lipids and lipoproteins in healthy men. J, Lipid Res. 34: I -10. 18. Willett, W.C., Stampfer, M.J., Manson, J.E., 451 Agric. Sei. Eini. 2 (1993) Colditz, G.A., Speizer, F.E., Rosner, 8.A., Sampson, L.A. & Hennekens, C.H. 1993. Intake of trans fatty acids and risk of coronary heart disease among women. Lancet 341: 581-585. 19. Thomas, L.H. 1992. Ischaemic heart disease and con- sumption of hydrogenated marine oils in England and Wales, J. Epid. and Comm. Health 46: 78-82. 20. Prevention of Coronary Heart Disease: Scientific Back- ground and New Clinical Guidelines. Recommenda- tions of the European Atherosclerosis Society prepared by the International Task Force for Prevention of Coronary Heart Disease. Nutr. Metab. Cardiovasc. Dis. 1992. 2: 113-156. 3452 Letter to the EditorAgric. Sci. Pint. 2(1993)