- Email: [email protected]
WINE, GARLIC, AND CHD IN SEVEN COUNTRIES
145
WINE, GARLIC, AND CHD IN SEVEN COUNTRIES SIR,-Wine lovers everywhere will have been delighted with the report of St Leger, Cochrane, and Moore.’ Bacchus himself would applaud: "The principle finding is a strong and negative association between ischaemic heart death and alcohol consumption. This is shown to be wholly attributable to wine "
consumption." McMichael objected that French death certificates on heart disease are notoriously suspect ;2 indeed they are, but the negative association between wine and coronary mortality persists
Cord
plasma insulin values (pmol/1)
in neonates.
Mothers had received no intravenous fluids in labour (left) or more than 4 h of infusion of either <10 g/h of glucose (centre) or 10-20 g/h of glucose (right).
(200-400 ml of 5% dextrose) per hour intravenously, for over 4 h before delivery. The infusion was given as a vehicle for oxytocin. and/or for hydration. Their infants had very high plasma insulin concentrations at birth, the highest being fifteen times the level seen in the "no infusion" group. We need to know whether iatrogenic hyperinsulinaemia could do harm by causing a rebound fall in blood glucose after birth. Under experimental conditions maternal glucose infusions raise the fetal blood glucose and plasma insulin in parallel,’ so, theoretically, the fetal hyperinsulinamua induced might be "appropriate" to the concomitant hyperglycoemia. However, we found that in clinical practice when mothers were given "high glucose" infusions for prolonged periods individual cord insulin and glucose levels correlated poorly (r=0-39; not significant), so some of their infants will start life with a high plasma insulin mismatching a low glucose concentration and could be particularly at risk of hypoglycaemia. Znamenacec et al. found that a maternal glucose infusion for 30 min before delivery resulted in a tendency for blood glucose to fall further than it did in controls. Insulin values were not reported, but fetal insulin secretion was unlikely to2have been maximal because the glucose infusion time was short. Certain groups of susceptible infants, (e.g., those who are preterm or small for dates) could have symptomatic hypoglycaemia as a result of the additional effect of an excessive maternal dextrose load. Until prospective studies are done some doubt must hang over the safety of prolonged intravenous infusions of more than 10 g of glucose per hour to mothers in labour. A. L.
was
in
receipt
of a Wellcome research
University Department of Pædiatrics, John Radcliffe Hospital, Oxford OX3 9DU; and Hammersmith Hospital, London W12
fellowship. A. LUCAS T. E. ADRIAN A. AYNSLEY-GREEN S. R. BLOOM
7. Znamenacek K, Pribylova H, Sabata V. The effect of prenatal glucose and insulin infusion on carbohydrate metabolism in the new born. Acta Pœdiat
Scand 1966; 55:452-56.
when France is omitted from the analysis.3 Still, there are strong reasons for resisting the temptation to prescribe, for patients or for ourselves, an abundance of wine in the hope of warding off heart-attacks. Comparisons of death rates by cause as given in international vital statistics always raise questions about the comparability of diagnoses. However, the recent vital statistics of developed countries do not grossly misrepresent the true picture for heart-disease, except, perhaps, in France. In any case we can disregard vital statistics and point to the findings of the Seven Countries Study, covering over 12 000 men aged 40-59 who were followed up for ten years.4,5 The incidence of coronary heart disease and death from that cause showed major differences among the sixteen cohorts, differences not explained by relative body-weight, physical activity, or smoking habits. The Americans and the Finns had the worst experience, the Greeks and the Japanese the best. For Europe there was a marked gradient from north to south-from Finland to the Netherlands to Italy and Yugoslavia and on to Greece where the men on the island of Crete rivalled the Japanese in their relative freedom from the disease. The cohorts of the Seven Countries Study were never offered as being necessarily representative of the countries or regions where they lived, but it turned out that their relative incidence rates were strikingly in accord with the relative coronary death rates in the national statistics. More to the point is that we knew a great deal about those men and their mode of life before there was any evidence of disease. Our Greeks and Italians, like their countrymen in general, are wine drinkers ; the men of Zutphen, like other Dutchmen, drink some wine but are much more likely to drink beer, with gin on occasion ; our Finns and Americans seldom drank wine. Our heaviest wine drinkers seem to have been the men on the Dalmatian coast of Yugoslavia and those men proved to be only slightly more prone to heart-attacks than were our Greeks. So the experience in the Seven Countries Study could be cited as being in accord with the negative association between wine and CHD found by St Leger et al. in international statistics. Another consistency could be pointed out. The remarkable decline in the coronary death rate in the United States in recent years has coincided with a spectacular increase in the consumption of wine by Americans. The findings in the Seven Countries Study might also be cited as supporting Slater6and Tyrrell’ who suggest that thought should be given to garlic as a possible preventive against heart-attacks. For our men in Greece, Italy, and Dalmatia use far more garlic than do their counterparts in the
1. St
2. 3. 4.
5.
Leger AS, Cochrane AL, Moore F. Factors associated with cardiac mortality in developed countries with particular reference to the consumption of wine. Lancet 1979; i: 1017-20. McMichael J. Ischæmic heart disease and wine. Lancet 1979; i: 1186. St Leger AS, Cochrane AS, Moore F. Ischæmic heart disease and wine. Lancet 1979; i:1294. Keys A, ed. Coronary heart disease in seven countries. Am Heart Assoc Monogr 1970, no 29 (also Circulation suppl, April, 1970). Keys A, Aravanis C, Blackburn H, van Buchem FSP, Buzina R, Djordjević BS, Dontas AS, Fidanza F, Karvonen MJ, Kimura N, Menotti A, Puddu V, Punsar S, Taylor HL. Seven countries: Death and coronary heart disease in ten years. Cambridge, Massachusetts, Harvard University Press (in press).
6. Slater NGP. On garlic. Lancet 1979; i: 1294. 7. Tyrrell H. On garlic. Lancet 1979; i: 1294.
146 Netherlands and the U.S. while the "very strong smell" and "acrid, pungent taste" (to quote the Oxford English Dictionary) of Allium sativum almost never contribute to the cookery of rural Finland. All around the Mediterranean garlic is used in abundance-in excess Britons might say-and in all that region CHD is far less common than it is in Britain and northern Europe where garlic is used timidly if at all. But there is no lack of other parallel differences between the regions-wine v. milk or beer, olive oil v. butter, small v. large portions of meat, strong v. weak coffee or tea, masses of leafy vegetables v.... you name it. The development of atherosclerosis and the clinical disease may be promoted or hindered by a host of things, known and unknown, and simple associations, such as those with the use of wine or garlic, warrant no conclusions about cause and effect. At best it might be stated that the observed association is or is not consistent with a certain hypothesis, a hypothesis derived from a reasonable concept of pathogenesis and a clear understanding of relevant facts of physiology and biochemistry. But in regard to wine and garlic the hypothesis is missing. Most of the differences in the incidence of ischaemic heartdisease among the cohorts of the Seven Countries Study are accounted for by two variables-the concentration of cholesterol in the serum and the arterial blood-pressure. Now if it could be shown that the habit of drinking wine and/or eating garlic lowers the serum cholesterol or the blood pressure, or both, we could urge a campaign to use more wine and garlic. I live much of the time in southern Italy and need no such persuasion happily to embrace the local customs of the kitchen and the dinner table. Laboratory of Physiological Hygiene,
ANCEL KEYS
U.S.A.
WHY BLAME CHOLESTEROL? in your Dec. 1 issue
again illusattribute unifactorial causality to multifactorial disease. By failing to cite smoking and blood-pressure among the "physical, chemical, organismal, or perhaps immunological" factors involved in the reaction of the vessel wall to injury and subsequent development of cholesterol-containing, raised plaques, Sir John McMichael ignores established fact. He also reveals something of the contrived nature of his contentions by commenting on high mortality in New Zealand Maoris with low cholesterol while omitting any reference to blood-pressure, body-weight, and smoking. The Report of the Joint Working Party of the Royal College of Physicians and the British Cardiological Society devoted about a fifth of its content to lipids. It emphasised the variation in CHD risk according to the total burden of risk factors and the need to consider the overall degree of CHD risk rather than whether any particular factor had reached a "critical" level, when dealing with an individual. The same report was unequivocal in stating that the cost of any of the preventive measures recommended, including that of reducing saturated fat and partly substituting polyunsaturated fat, did not approach the cost of inaction. In our own work in the West of Scotland,2,3 the independent and significant nature of the relative and attributable risks in SiR,— I he correspondence
trates
the
fallacy
and
futility of attempting to
of the Joint Working Party of the Royal college of Physicians and British Cardiac Society on Prevention of Coronary Heart Disease, J Roy Coll Phys 1976; 10(3): 20. 2. Hawthorne VM, Greaves DA, Beevers DG. Blood pressure in a Scottish town. Br Med J 1974; iii: 600. 3. Hawthorne VM, Greaves DA, Beevers DG. Community studies of hypertension in Glasgow: Epidemiology and control of hypertension. In: Oglesby P, ed. Proceedings of second International symposium on the epidemiology of hypertension. Chicago Heart Association, 1975: 537-51. 1.
Report
Whatever the basis for the 8% decline and 12% increase in CHD mortality between 1969 and 1973 in U.S. White males age 45-54 years compared with England and Wales, North Ireland, and Scotland, respectively,5,6 cholesterol is one of the many factors contributing in varying degree to this disturbing difference. We hope it will encourage others in U.K. to know that a noteworthy inclusion in the fifteen programmes proposed to prevent disease and promote health into the 1990s in U.S., is one with the specific aim to reduce population cholesterol levels to 180 mg/dl within this decade.7 of Epidemiology, School of Public Health,
Department
University of Michigan, Ann Arbor, Michigan 48109,
VICTOR M. HAWTHORNE
U.S.A.
EXTRA CHROMOSOME 12 IN CHRONIC LYMPHOCYTIC LEUKÆMIA
SIR,-Polyclonal B cell mitogens can induce both DNA synthesis and immunoglobulin production in lymphocytes from patients with chronic lymphocytic leukxmia (CLL).1 We have reported that stimulation with one such mitogen, the EpsteinBarr virus (EBV), revealed a chromosomal aberration in a CLL patient.2 We have now investigated eight patients with this disease, seven of whom were untreated. were stimulated by EBV in all Peripheral blood patients. In seven of them another polyclonal B cell mitogen, lipopolysaccharide from Escherichia coli (LPS), was also used for stimulation. Metaphases were analysed after five days of incubation. The Q-banding technique was used for chromosome identification. Four of the patients had an extra chromosome 12 in more than a third of the metaphases investigated, after stimulation with EBV. Three of them had the same abnormality after stimulation with LPS. Two of these patients had additional aberrations. One had a deletion of chromosome 2 and the other a deletion of chromosome 14 (table). The conclusion is that polyclonal B cell mitogens can stimulate CLL lymphocytes and reveal previously undetectable
lymphocytes
School of Public Health,
University of Minnesota, Minneapolis, Minnesota,
identified populations of Renfrew and Paisley for CHD mortality by the 80-95th and the top 5th percentile of baseline cholesterol values, has already been described in an illustrative sample of 5616 males observed over 5 years.4 No contrary evidence has yet emerged. census
4. Hawthorne VM. Diet and coronary heart disease. Br Med J 1977; ii: 186-87. 5. WHO statistics annual 1969: vol I, vital statistics and causes of death. Geneva: W.H.O., 1972. 6. W.H.O. statistics annual 1973: vol I, vital statistics and causes of death. Geneva: W.H.O., 1976. 7. Preventing disease/promoting health—objectives for the nation: draft working papers of Atlanta conference. Department of Health Education and
Welfare, August, 1979. 1. Robèrt K-H, Bird AG, Möller E. Mitogen-induced differentiation of human CLL lymphocytes to antibody secreting cells. Scand J Immunol (in press). 2. Gahrton G, Zech L, Robèrt K-H, Bird AG. Mitogenic stimulation of leukemia cells by Epstein-Barr virus. N Engl J Med 1979, 301:438.
EXTRA CHROMOSOME
12
IN FOUR PATIENTS WITH CLL
WINE, GARLIC, AND CHD IN SEVEN COUNTRIES SIR,-Wine lovers everywhere will have been delighted with the report of St Leger, Cochrane, and Moore.’ Bacchus himself would applaud: "The principle finding is a strong and negative association between ischaemic heart death and alcohol consumption. This is shown to be wholly attributable to wine "
consumption." McMichael objected that French death certificates on heart disease are notoriously suspect ;2 indeed they are, but the negative association between wine and coronary mortality persists
Cord
plasma insulin values (pmol/1)
in neonates.
Mothers had received no intravenous fluids in labour (left) or more than 4 h of infusion of either <10 g/h of glucose (centre) or 10-20 g/h of glucose (right).
(200-400 ml of 5% dextrose) per hour intravenously, for over 4 h before delivery. The infusion was given as a vehicle for oxytocin. and/or for hydration. Their infants had very high plasma insulin concentrations at birth, the highest being fifteen times the level seen in the "no infusion" group. We need to know whether iatrogenic hyperinsulinaemia could do harm by causing a rebound fall in blood glucose after birth. Under experimental conditions maternal glucose infusions raise the fetal blood glucose and plasma insulin in parallel,’ so, theoretically, the fetal hyperinsulinamua induced might be "appropriate" to the concomitant hyperglycoemia. However, we found that in clinical practice when mothers were given "high glucose" infusions for prolonged periods individual cord insulin and glucose levels correlated poorly (r=0-39; not significant), so some of their infants will start life with a high plasma insulin mismatching a low glucose concentration and could be particularly at risk of hypoglycaemia. Znamenacec et al. found that a maternal glucose infusion for 30 min before delivery resulted in a tendency for blood glucose to fall further than it did in controls. Insulin values were not reported, but fetal insulin secretion was unlikely to2have been maximal because the glucose infusion time was short. Certain groups of susceptible infants, (e.g., those who are preterm or small for dates) could have symptomatic hypoglycaemia as a result of the additional effect of an excessive maternal dextrose load. Until prospective studies are done some doubt must hang over the safety of prolonged intravenous infusions of more than 10 g of glucose per hour to mothers in labour. A. L.
was
in
receipt
of a Wellcome research
University Department of Pædiatrics, John Radcliffe Hospital, Oxford OX3 9DU; and Hammersmith Hospital, London W12
fellowship. A. LUCAS T. E. ADRIAN A. AYNSLEY-GREEN S. R. BLOOM
7. Znamenacek K, Pribylova H, Sabata V. The effect of prenatal glucose and insulin infusion on carbohydrate metabolism in the new born. Acta Pœdiat
Scand 1966; 55:452-56.
when France is omitted from the analysis.3 Still, there are strong reasons for resisting the temptation to prescribe, for patients or for ourselves, an abundance of wine in the hope of warding off heart-attacks. Comparisons of death rates by cause as given in international vital statistics always raise questions about the comparability of diagnoses. However, the recent vital statistics of developed countries do not grossly misrepresent the true picture for heart-disease, except, perhaps, in France. In any case we can disregard vital statistics and point to the findings of the Seven Countries Study, covering over 12 000 men aged 40-59 who were followed up for ten years.4,5 The incidence of coronary heart disease and death from that cause showed major differences among the sixteen cohorts, differences not explained by relative body-weight, physical activity, or smoking habits. The Americans and the Finns had the worst experience, the Greeks and the Japanese the best. For Europe there was a marked gradient from north to south-from Finland to the Netherlands to Italy and Yugoslavia and on to Greece where the men on the island of Crete rivalled the Japanese in their relative freedom from the disease. The cohorts of the Seven Countries Study were never offered as being necessarily representative of the countries or regions where they lived, but it turned out that their relative incidence rates were strikingly in accord with the relative coronary death rates in the national statistics. More to the point is that we knew a great deal about those men and their mode of life before there was any evidence of disease. Our Greeks and Italians, like their countrymen in general, are wine drinkers ; the men of Zutphen, like other Dutchmen, drink some wine but are much more likely to drink beer, with gin on occasion ; our Finns and Americans seldom drank wine. Our heaviest wine drinkers seem to have been the men on the Dalmatian coast of Yugoslavia and those men proved to be only slightly more prone to heart-attacks than were our Greeks. So the experience in the Seven Countries Study could be cited as being in accord with the negative association between wine and CHD found by St Leger et al. in international statistics. Another consistency could be pointed out. The remarkable decline in the coronary death rate in the United States in recent years has coincided with a spectacular increase in the consumption of wine by Americans. The findings in the Seven Countries Study might also be cited as supporting Slater6and Tyrrell’ who suggest that thought should be given to garlic as a possible preventive against heart-attacks. For our men in Greece, Italy, and Dalmatia use far more garlic than do their counterparts in the
1. St
2. 3. 4.
5.
Leger AS, Cochrane AL, Moore F. Factors associated with cardiac mortality in developed countries with particular reference to the consumption of wine. Lancet 1979; i: 1017-20. McMichael J. Ischæmic heart disease and wine. Lancet 1979; i: 1186. St Leger AS, Cochrane AS, Moore F. Ischæmic heart disease and wine. Lancet 1979; i:1294. Keys A, ed. Coronary heart disease in seven countries. Am Heart Assoc Monogr 1970, no 29 (also Circulation suppl, April, 1970). Keys A, Aravanis C, Blackburn H, van Buchem FSP, Buzina R, Djordjević BS, Dontas AS, Fidanza F, Karvonen MJ, Kimura N, Menotti A, Puddu V, Punsar S, Taylor HL. Seven countries: Death and coronary heart disease in ten years. Cambridge, Massachusetts, Harvard University Press (in press).
6. Slater NGP. On garlic. Lancet 1979; i: 1294. 7. Tyrrell H. On garlic. Lancet 1979; i: 1294.
146 Netherlands and the U.S. while the "very strong smell" and "acrid, pungent taste" (to quote the Oxford English Dictionary) of Allium sativum almost never contribute to the cookery of rural Finland. All around the Mediterranean garlic is used in abundance-in excess Britons might say-and in all that region CHD is far less common than it is in Britain and northern Europe where garlic is used timidly if at all. But there is no lack of other parallel differences between the regions-wine v. milk or beer, olive oil v. butter, small v. large portions of meat, strong v. weak coffee or tea, masses of leafy vegetables v.... you name it. The development of atherosclerosis and the clinical disease may be promoted or hindered by a host of things, known and unknown, and simple associations, such as those with the use of wine or garlic, warrant no conclusions about cause and effect. At best it might be stated that the observed association is or is not consistent with a certain hypothesis, a hypothesis derived from a reasonable concept of pathogenesis and a clear understanding of relevant facts of physiology and biochemistry. But in regard to wine and garlic the hypothesis is missing. Most of the differences in the incidence of ischaemic heartdisease among the cohorts of the Seven Countries Study are accounted for by two variables-the concentration of cholesterol in the serum and the arterial blood-pressure. Now if it could be shown that the habit of drinking wine and/or eating garlic lowers the serum cholesterol or the blood pressure, or both, we could urge a campaign to use more wine and garlic. I live much of the time in southern Italy and need no such persuasion happily to embrace the local customs of the kitchen and the dinner table. Laboratory of Physiological Hygiene,
ANCEL KEYS
U.S.A.
WHY BLAME CHOLESTEROL? in your Dec. 1 issue
again illusattribute unifactorial causality to multifactorial disease. By failing to cite smoking and blood-pressure among the "physical, chemical, organismal, or perhaps immunological" factors involved in the reaction of the vessel wall to injury and subsequent development of cholesterol-containing, raised plaques, Sir John McMichael ignores established fact. He also reveals something of the contrived nature of his contentions by commenting on high mortality in New Zealand Maoris with low cholesterol while omitting any reference to blood-pressure, body-weight, and smoking. The Report of the Joint Working Party of the Royal College of Physicians and the British Cardiological Society devoted about a fifth of its content to lipids. It emphasised the variation in CHD risk according to the total burden of risk factors and the need to consider the overall degree of CHD risk rather than whether any particular factor had reached a "critical" level, when dealing with an individual. The same report was unequivocal in stating that the cost of any of the preventive measures recommended, including that of reducing saturated fat and partly substituting polyunsaturated fat, did not approach the cost of inaction. In our own work in the West of Scotland,2,3 the independent and significant nature of the relative and attributable risks in SiR,— I he correspondence
trates
the
fallacy
and
futility of attempting to
of the Joint Working Party of the Royal college of Physicians and British Cardiac Society on Prevention of Coronary Heart Disease, J Roy Coll Phys 1976; 10(3): 20. 2. Hawthorne VM, Greaves DA, Beevers DG. Blood pressure in a Scottish town. Br Med J 1974; iii: 600. 3. Hawthorne VM, Greaves DA, Beevers DG. Community studies of hypertension in Glasgow: Epidemiology and control of hypertension. In: Oglesby P, ed. Proceedings of second International symposium on the epidemiology of hypertension. Chicago Heart Association, 1975: 537-51. 1.
Report
Whatever the basis for the 8% decline and 12% increase in CHD mortality between 1969 and 1973 in U.S. White males age 45-54 years compared with England and Wales, North Ireland, and Scotland, respectively,5,6 cholesterol is one of the many factors contributing in varying degree to this disturbing difference. We hope it will encourage others in U.K. to know that a noteworthy inclusion in the fifteen programmes proposed to prevent disease and promote health into the 1990s in U.S., is one with the specific aim to reduce population cholesterol levels to 180 mg/dl within this decade.7 of Epidemiology, School of Public Health,
Department
University of Michigan, Ann Arbor, Michigan 48109,
VICTOR M. HAWTHORNE
U.S.A.
EXTRA CHROMOSOME 12 IN CHRONIC LYMPHOCYTIC LEUKÆMIA
SIR,-Polyclonal B cell mitogens can induce both DNA synthesis and immunoglobulin production in lymphocytes from patients with chronic lymphocytic leukxmia (CLL).1 We have reported that stimulation with one such mitogen, the EpsteinBarr virus (EBV), revealed a chromosomal aberration in a CLL patient.2 We have now investigated eight patients with this disease, seven of whom were untreated. were stimulated by EBV in all Peripheral blood patients. In seven of them another polyclonal B cell mitogen, lipopolysaccharide from Escherichia coli (LPS), was also used for stimulation. Metaphases were analysed after five days of incubation. The Q-banding technique was used for chromosome identification. Four of the patients had an extra chromosome 12 in more than a third of the metaphases investigated, after stimulation with EBV. Three of them had the same abnormality after stimulation with LPS. Two of these patients had additional aberrations. One had a deletion of chromosome 2 and the other a deletion of chromosome 14 (table). The conclusion is that polyclonal B cell mitogens can stimulate CLL lymphocytes and reveal previously undetectable
lymphocytes
School of Public Health,
University of Minnesota, Minneapolis, Minnesota,
identified populations of Renfrew and Paisley for CHD mortality by the 80-95th and the top 5th percentile of baseline cholesterol values, has already been described in an illustrative sample of 5616 males observed over 5 years.4 No contrary evidence has yet emerged. census
4. Hawthorne VM. Diet and coronary heart disease. Br Med J 1977; ii: 186-87. 5. WHO statistics annual 1969: vol I, vital statistics and causes of death. Geneva: W.H.O., 1972. 6. W.H.O. statistics annual 1973: vol I, vital statistics and causes of death. Geneva: W.H.O., 1976. 7. Preventing disease/promoting health—objectives for the nation: draft working papers of Atlanta conference. Department of Health Education and
Welfare, August, 1979. 1. Robèrt K-H, Bird AG, Möller E. Mitogen-induced differentiation of human CLL lymphocytes to antibody secreting cells. Scand J Immunol (in press). 2. Gahrton G, Zech L, Robèrt K-H, Bird AG. Mitogenic stimulation of leukemia cells by Epstein-Barr virus. N Engl J Med 1979, 301:438.
EXTRA CHROMOSOME
12
IN FOUR PATIENTS WITH CLL