FRENCH PARADOX
Dr. Michel Craple
ANPA - 20 rue Saint Fiacre 75002 PARTS
Tel. : 42.33.51.04 - Fax 45.08.17.02

Definition:

Epidemiological data prove that there exists in France a low morbidity through cardiovascular coronaries (infarctus) despite the fact the diet is rich in lipids, risk factors in arteriosclerosis: foreign observers have baptised this peculiarity the "French Paradox".

Most ommentators explain this phenomenon by the consumption of alcohol or of wine, which would exert a protective effect. This explanation is at present widely exploited by the producers and taken up by the press greedy for paradox and sensation.

According to certain studies, wine would have a particular protective effect, but in the majority of such studies the nature of the drinks has little importance. Alcohol and/or wine could act by increasing the cholerestol level and reducing the agregration of blood platelets.

Notes:

1. The beneficial effects of alcohol have been praised since the origin of drinks; it is true that the therapeutic effects might have been of value at a time when neither anaesthetics nor tranquillisers existed.

One can consider this protective effect as the final avatar of the myth of alcohol as a medicine. Today in a different scientific context, the defenders of this thesis advise moderate doses of alcohol in order to obtain it. However even with this reservation the phenomenon merits a critical look.

2. Epidemiological studies are difficult in this field where the risk factors are numerous, some linked together (food consumption, drinks, tobacco...), where the social parameters play an important role and where the psychopathological mechanism is complex: atheromatic illness (due to too many calories and to tobacco), thrombosis through a build up of platelets and arterial spasm. Furthermore numerous angles are possible at the time of declaration of consumption and causes of death.

A careful interpretation must be made of even the best studies from the methodological point of view: It is not advisable to limit oneself to geographical studies naively comparing very different populations and attributing the differences in morbidity solely to the consumption of alcohol.

For ethical reasons, the definitive proof will not be easily brought about. A link or correlation is not a causal relationship: statistical analyses will never be able to eliminate the role of unknown and I or unquantifiable factors which can be at the origin of the link.

3. The frequency of numerous diseases (cancers, digestive and neurological disease) is increased by even mimimal consumption of alcohol. This is also the case in accidents and social problems due to heavy drinking. However the relative importance to morbidity by infarctus can explain the minimum of total mortality (J Curve) at doses of 1 to 2 glasses per day.

4. It is possible that the explanation of the French paradox is to be found by looking at the diet as a whole. The role of the distribution of energy intake in the day and the nature of lipids consumed has been spoken of. The predominent factor seems to be the richness of the diet in fruits and fresh vegetables carrying fibres and vitamins which is a caracteristic of the so called mediterranean diet (1). Let us also note that certain components of wine considered as protective are found equally well in the base product as they are contained in the pips or the skin of the fruit.

5. The protective effect on myocardial infarction is limited:

• it reaches its maximum at a consumption of between 5 to log of alcohol per day (according to the studies). It has only been proved in some studies that the effect continued in proportion to the dose consumed beyond 20g per day: this is the case in the Danish study (2) which had recent media publicity. Methodological errors are possible for the numbers of consumers are weak for high doses in the countries where the research was undrtaken.
• it is of interest only amongst those subjects who are at risk of infarctus, ie men above 40 years of age and women after the menopause,
• one recent study (3) seems to have proved that this protective effect only comes into being in a genetically determined sub population possessing a particular genotype of an enzyme for the transport of cholesterol (16% of the general population).

6. To reduce the frequency of coronary attacks, it is preferable to diminish the other risk factors (by the suppression of smoking, a balanced diet, physical exercise), modifications which do not cause harm rather than to encourage a a behaviour which generates other illnesses.

7. If therefore it is not justifiable to ask a heart patient , a moderate consumer, to change his habits, it is irresponsible to recommend an alcohol consumption to 'prevent' heart attack in a general message. (4)

• from the individual point of view, it is impossible to know in advance sensitivity and tolerance to alcohol. This message risks being distorted transforming the scientific reality of the protective action of one glass into an incitement to drink thus increasing all risks. The number of abstainers thus encouraged to drink alcohol will increase their consumption to the point of creating problems which they would have spontaneously avoided.
• from the public health point of view, in a country severely affected by the other consequences of excessive consumption, it is probab~e that the message will be understood principally by those for whom it is inapplicable (young men who with this medical endorsement will be able to drink from 20 to 40 without any advantage but increasing numerous medical and social risks).
  

References

1. Criqui MH, Ringel BL. Does diet or alcohol explain the french paradox?
Lancet 1994 1719-23

2. Grønbach M. et al. Mortality associated with moderate intakes of wine, beer, or spirit.
British Medical Journal, volume 310, 6 May 1995, pp 1165-1169

3. Fumeron F. et al. Alcohol intake modulates the effect of a polymorphism of the cholestetyl ester transfer protein gene on plasma high density lipoprotein and the risk of myocardial infarction.
J Clin Invest 1995; 96:1664-71

4. Skog OJ. Public health consequences of the J-curve hypothesis of alcohol problems.
Addiction 1996; 91 : 325-37