Sept 16 2004
Cholesterol And The French Paradox, The Swiss Paradox, The Russian Paradox, The Lithuanian Paradox…Etc…
I was going to write about something else, but I couldn’t resist bringing the latest results from the MONICA study to your attention (MONItor trends in CArdiovascular disease). This study has been going on for ages now. It was set up by the WHO – not the rock band – to look at heart disease rates and risk factors in twenty six countries across the world.
MONICA is a fantastic study by the way. It creates huge volumes of data which can be relied upon to be accurate and objective. Three cheers for the WHO for having the foresight to put this study in place. Talkin’ bout my generation.
For years the MONICA study can remain silent then, every so often, it bestirs itself and out plops a golden egg. The latest golden egg comes from their review of high cholesterol levels in countries ranging from China to Northern Ireland.
They decided to find out what percentage of the male population in nineteen of the MONICA countries (don’t ask me why they chose nineteen), had a cholesterol level greater than 6.5mmol/l ( 260mg/dl.). This is a fairly good indicator of the average cholesterol level in a country.
The figures ranged from a prevalence of about 2% in China, to 51% in Switzerland. In short, MONICA revealed a twenty five fold difference in the prevalence of high cholesterol between countries, which was the somewhat unimaginative headline attached to the study. ‘Prevalence of high cholesterol varies 25-fold.’ Actually, it varies 25.5 fold.
However, what I found far more interesting than this bald statistic was the fact that there appeared to be absolutely no relationship between the percentage of men with high cholesterol levels, and the rate of heart disease in those countries (with the possible exception of China). A fact that appears to have passed without comment. Least said soonest mended?
For example, the rate of CHD in Russia is enormously high, yet the prevalence of high cholesterol levels in Russia was third lowest. On the other hand the rate of CHD in France is very low, yet the prevalence of high cholesterol levels in France was in the top half of the table.
In order to confirm my suspicions about the lack of association I made a rapid search of my archives in order to establish CHD rates in the nineteen countries chosen by MONICA. Then I matched them to the percentage of men in each country with high cholesterol levels.
I used the latest full year where data is available for CHD rates in each of the nineteen countries (1990) – unfortunately, one of the countries was Yugoslavia, which doesn’t exist any more, then plotted this against hypercholesterolaemia rates, from MONICA, five years later (not perfect but then cholesterol levels haven’t changed that much over five years). This resulted in the graph you can see below.
Each of the points on the graph represents a country. Bottom left is China, with a CHD rate of ‘sixty per hundred thousand per year’, and a raised cholesterol level prevalence of two per cent. The highest point is Lithuania; with a CHD rate of five hundred and eight five and a raised cholesterol level prevalence of just over thirty per cent. Far right is Switzerland, CHD rate one hundred and eighty one, raised cholesterol prevalence fifty one per cent.
I have no doubt that a statistician could apply an elegant mathematical formula to that graph, and end up with a straight line leading from bottom left to top right ‘proving’ that the percentage of the population with hypercholesterolaemia is strongly and consistently related to the level of CHD. I am sure that super-computers around the world are humming away madly as they attempt to find the statistical equation that matches the approved ‘answer.’
For myself, all I can see is a French ‘paradox’ and Swiss ‘paradox’ a Russian ‘paradox’ a Lithuanian ‘paradox….. My interpretation is that there is absolutely no connection between cholesterol levels and CHD rates in these nineteen different countries. Once again MONICA comes up trumps. In fact MONICA has been producing results directly contradictory to the cholesterol hypothesis for years;it’s just that somehow or other the results always get smoothed over.
In this case, however, I would have to take my hat-off to anyone who is capable of smoothing that particular graph. They shall be crowned medical statistician of the decade. (Which is not actually a compliment in my book).
Category: Diabetes/Heart Disease