Dear JULES,
Cholesterol & mortality – world graphs
Executive summary
* This post shares my presentation, which was part of the Low Carb All Stars conference on the weekend of June 19th – 20th, and a key update from within it.
* A Monday note from November 2010 examined World Health Organisation (WHO) data for the average cholesterol levels and death rates for 192 countries. The deaths examined were from cardiovascular disease (CVD) and all-cause mortality. The associations between average cholesterol levels and deaths were collated for males and females separately. Four charts were thus able to be drawn. All associations were inverse – the higher the cholesterol levels, the lower the death rates and vice versa.
* Over many years, I kept looking for the data to be able to update these charts and some became available in 2020. The data for CVD were no longer available but data for non-communicable disease could be used instead. All-cause mortality data were available. Four charts could again be produced from the data – males and females, non-communicable disease and all-cause mortality – plotted against average country cholesterol levels. This time data were available for 181 countries. All associations were inverse – the higher the cholesterol levels, the lower the death rates and vice versa.
* These updated data were included in an academic publication in December 2020. If you need to be able to cite an inverse association between cholesterol levels and deaths, such a reference is now available.
Introduction
On June 19th and 20th, 2021, Dr Paul Mason hosted a Low Carb All Stars conference (online) (Ref 1). I wasn’t able to join in the Q&A sessions (I’ll be in Cornwall when you read this), but they were probably lively if last year was anything to go by.
You can see my presentation here.
I decided to present a selection of Monday notes in the field of cholesterol, statins, saturated fat and the diet heart hypothesis generally, as this continues to be a subject area of much interest in the low carb world. The first post covered was written in November 2010. The post was called “Cholesterol & heart disease – there is a relationship, but it’s not what you think” (Ref 2). In the post about the 500th Monday note, I shared that this post had been the second most popular on the site over the years (Ref 3).
It only took a couple of hours to find the data for the original post but finding the data to update it proved much more difficult. The data for the original and the update have come from the World Health Organization web site.
The original data
All the data for the 2010 post came from 2002, as that was the most recent year for which both cholesterol and mortality data were available.
The first set of charts plotted death rates from all causes and from cardiovascular disease (CVD) and country average cholesterol levels for men and women. Four charts were drawn:
i) CVD death rates and cholesterol for men;
ii) CVD death rates and cholesterol for women;
iii) All-cause death rates and cholesterol for men; and
iv) All-cause death rates and cholesterol for women.
All four charts showed an inverse relationship – the higher the average cholesterol level, the lower the death rates and the lower the average cholesterol level, the higher the death rates (read that carefully). That’s the exact opposite of the relationship that we are told exists. The relationships got stronger from (i) through to (iv) such that the correlation coefficient (r) was 0.66 for all deaths and cholesterol in men and 0.74 for all deaths and cholesterol in women. The r2 (the strength of the relationship) is shown in the charts below.
Updating the data
I was unable to get data for CVD death rates for countries, but data were available for non-communicable disease (NCD). The three sources for the most recent work are:
1) Mean (average) cholesterol levels for males and females by country up to 2009. I used the most recent year, 2009 (Ref 4).
2) Total non-communicable disease (NCD) mortality data by country – for males and females – year 2010 was the closest to 2009, so I used that (Ref 5).
3) Total all-cause mortality by country – for males and females – I used 2010 again (Ref 6).
The World Health Organisation defines non-communicable disease, also known as chronic diseases, as those of long duration, which are “the result of a combination of genetic, physiological, environmental and behaviours factors.” There are approximately 41 million NCD deaths annually (Ref 7). CVD accounts for most NCD deaths (17.9 million), followed by cancer (9 million), respiratory diseases (3.9m), and diabetes (1.6m). Those four conditions account for approximately 80% of all NCD deaths.
As we can see, total deaths from non-communicable disease are different to cardiovascular deaths, for which I was able to get data in 2010. On the downside, we can’t directly compare cholesterol levels and heart deaths, and this is the most often claimed direct relationship. On the upside, it is possibly more interesting to review the relationship between cholesterol levels and all NCD deaths, as claims are made for health benefits of lowering cholesterol beyond heart disease – cancer for example. Conversely, diabetes is reported as a possible harm resulting from taking cholesterol lowering medication (Ref 8).
The same four relationships have been examined for 2009/2010 and the graphs are as follows:
Academic publication
Dr Uffe Ravnskov has been trying to get these charts published since he first saw them in 2010, as he wanted to reference them in academic papers. I’m delighted to say that we finally achieved this in December 2020 (Ref 9). A paper called “The LDL Paradox: Higher LDL-Cholesterol is Associated with Greater Longevity” was published in the Annals of Epidemiology & Public Health. The two charts for mean cholesterol and all-cause mortality for males and females were published in this paper. If you would also like to reference the fact that there is an inverse association between average cholesterol levels and all-cause mortality for 181 countries – the reference to cite is that December paper.
Interpretation
When I first looked at this in 2010, using data from 2002 for all-cause deaths, cardiovascular deaths and average cholesterol levels, there were data for 192 countries. All four charts – for all deaths and CVD deaths and for men and women – showed inverse relationships. The higher the cholesterol level, the lower the death rate and vice versa.
The 2020 update was still only able to use data from 2009-2010 – recent data do not seem to be available for this research. The 2020 update also had to use non-communicable disease (NCD) and not CVD, which has pros and cons. Data were available for 181 countries for the 2020 update. Again, all four charts – for all deaths and NCD deaths and for men and women – showed inverse relationships. The higher the cholesterol level, the lower the death rate and vice versa.
It is interesting that these relationships have held true for data from 2002 to 2010 (from my analyses in 2010 and 2020 respectively).
It is interesting that all four charts have been inverse both times.
It is interesting that the inverse relationships are stronger in women than men.
It is interesting that this is the exact opposite of the relationship we are told exists. We are told that the lower the cholesterol, the lower the death rate. This is the basis upon which cholesterol lowering activity takes place. But it doesn’t hold true.
These charts show association, not causation and we must take care not to make the mistake of claiming causation. We cannot claim from these charts that lower cholesterol causes higher deaths or higher cholesterol causes lower deaths, but we can say that the association does not support the relationships that are consistently claimed.
As Dr Malcolm Kendrick said of association and causation, “association does not prove causation, but lack of association disproves causation.” That’s the territory these charts have put us in.
Until the next time
All the best – Zoë
References
Ref 1: https://www.tickettailor.com/events/lowcarballstars/505717
Ref 2: https://www.zoeharcombe.com/2010/11/cholesterol-heart-disease-there-is-a-relationship-but-its-not-what-you-think/
Ref 3: https://www.zoeharcombe.com/2020/10/the-500th-monday-note/
Ref 4: http://apps.who.int/gho/data/node.main.A891
Ref 5: https://apps.who.int/gho/data/node.main.A860
Ref 6: http://apps.who.int/gho/data/view.main.1340
Ref 7: https://www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases
Ref 8: https://www.thennt.com/nnt/statins-for-heart-disease-prevention-without-prior-heart-disease-2/
Ref 9: Ravnskov et al. The LDL Paradox: Higher LDL-Cholesterol is Associated with Greater Longevity. Annals of Epidemiology & Public Health 2020. https://meddocsonline.org/annals-of-epidemiology-and-public-health/the-LDL-paradox-higher-LDL-cholesterol-is-associated-with-greater-longevity.pdf
