Last week we discussed some of the basic and relatively "easier" questions regarding serum cholesterol levels. This week we venture out to find answers to the more controversial ones.
The answer can be found in long term population-based studies. These involve evaluation of serum cholesterol levels (or whatever is the parameter of interest) in healthy individuals from the outset, with follow-ups conducted over several years to look at their long-term outcomes.
The most well-known ongoing cohort study is the Framingham Heart Study - initiated in the eponymous town in Massachusetts, USA in 1948. This study has generated over 1500 publications and has provided most of the information we have today on heart diseases and strokes (called cardiovascular diseases- CVD). Dr WB Kannel, the Director of this study also coined the term “risk factors” for high blood pressure, diabetes, elevated cholesterol and obesity, which were associated with an increased risk of these illnesses.
One of their publications in 1987 (whose findings were corroborated by independent researchers in 1993) analysed the impact of serum cholesterol levels on total and CVD mortality with age, over a period of 30 years.
The risk of mortality from CVD increased with rising serum cholesterol levels at ages 40 and 50, became lesser at 60 years, but became non-significant beyond that. An increase of CVD mortality of 9% was reported with a 10 mg/dl rise in cholesterol levels. However no clear association was found between cholesterol levels and risk of mortality due to heart attacks and strokes, at ages 70, 80 or higher.
Surprisingly, an increase in mortality was seen in those who had a reduction in cholesterol levels, especially after 60 years of age. This finding has been reported in several other studies as well. One reason could be the so-called “harvest effect”, meaning that those affected by cholesterol have had events in young age, and those who reach the age of 60 or more are affected more by risk factors other than cholesterol. Another could be that tumours could induce a secondary fall in cholesterol levels in older individuals.
Cholesterol levels correlate with survival majorly in the young. Surprisingly, higher levels of cholesterol are associated with trends towards better survival beyond 60 years, and especially in those over 70-80 years of age.
Why should this happen? Why is high cholesterol associated with higher risk below 50 years and lower risk after the age of 60?
For this, we have to examine the total mortality in these studies.
It was found in the Framingham Study as well as in other studies that CVD mortality tended to rise with rising cholesterol levels till 50-years of age, even 60 in one analysis, but total mortality (all-causes) correlated with cholesterol levels only till 40 years of age. By 50 years, the effect of cholesterol on survival became marginal, and finally negative at 80 and beyond. This was because non-CVD mortality (e.g., due to cancers and infections) was inversely related with cholesterol beyond age 50.
This must sound very confusing. So how important is cholesterol level, really?
Again, let us go back to Framingham. What we find is that aside from people with very low (< 160 mg/dl) or very high (>280 mg/dl) values, cholesterol levels follow a continuum and are not widely different between those with and without CVD.
An ideal risk factor graph should look something like Fig. 1, where there are discrete set of values differentiating the abnormal from the normal.
However, the cholesterol levels in Framingham for individuals with and without CVD are seen in Fig. 2, with a lot of overlap, and clear risk only at the extremes of cholesterol levels.
Any statistician would say that this value is not a decisive test of risk, definitely not the most influential risk factor for the most dreaded diseases which afflict humanity.
Notably, lowering the average serum cholesterol levels of the population by 1 mmol/l (about 38 mg/dl, or nearly 20% of baseline values) may lower the risk of CVD by about 30%, but the benefit is largely for the young who are at high baseline risk.
Further, the protective effect of cholesterol at older ages suggests that there is no exposure-dependent risk with cholesterol, i.e., the same level of cholesterol does not cause more harm the longer it persists, as against say, smoking and lung cancer.
The inferences that we can draw from this entire discussion:
As a final bit of information, it is suitable to add that all markers of cholesterol and other lipoproteins have a lower risk association with CVD compared to three other risk factors, namely:
Risk association of Total cholesterol/HDL-cholesterol at HR 1.32 is lower than the above,and the same as high blood pressure (HR 1.31).
So, what are the key takeaways from these two rather extended statistical discussions?
In summary, all adults should have their cholesterol levels checked for evidence of markedly elevated levels that are often seen in families (Familial Hyperlipidemia). Such individuals should receive counselling and appropriate lipid lowering therapy.
Others, who have mildly elevated levels should be advised regarding healthy diet and exercise; it is worth remembering that most individuals can maintain healthy parameters with a healthy lifestyle alone. Even otherwise, it is advisable to consume a healthy diet, consisting of vegetables, fruits, dairy products like curd and cheese, healthy fats and proteins, while avoiding fried foods, sugars, refined and processed foods. Regular exercise is also a must for optimal health.
Older individuals beyond 60 years of age need to be counselled about the risk-benefit ratio of treating elevated cholesterol levels, in view of the increased risk of non-CVD mortality (cancers and infections) in older individuals.
Know your cholesterol levels, but also check for other risk factors like diabetes, hypertension and obesity before demanding medicines for any minor perturbation of cholesterol levels.