Introduction
The general consensus in medicine on LDL has been that “lower is better”. However, recent data is challenging this notion. There is a growing body of data that shows the lower your cholesterol the HIGHER your chance of all-cause mortality, especially cancer. This is the paradox currently facing the scientific and medical community. Cholesterol and lipids are an extremely complex part of our physiology and it seems this general approach to lowering LDL in medicine has likely missed the mark on the more important causes of cardiovascular disease and mortality. So, in this post I’m going to look at some of the recent data on LDL and disease/mortality.
Before I jump in, it should be noted that many (at least those staying up to date, notwithstanding many GPs who are still only looking at LDL…) in the medical community have moved away from focusing only on total cholesterol and LDL towards more specific markers like ApoB and LDL-P (particle counts). Towards the end I will discuss how even these markers are likely still missing the mark, however, I will not get into those markers in this post and will save that for a follow up article.
And of course, none of this is medical advice and I am not a medical doctor or cardiologist. This is just purely for educational purposes and you should always speak to your doctor in a shared decision making framework about what is best for you. Always seek informed consent from your doctor. If they are not advising you fully of the pros and cons of any medications or interventions, you should find another doctor.
This article assumes basic physiology and lipidology knowledge and is written in an academic manner – so it does get somewhat detailed. If you want to discuss your lipid markers or overall health in more detail you can contact me for a consultation.
What the Evidence Shows
Cardiovascular disease (CVD), along with cancer, is the top killer of most people, especially once you make it past 60 years old. So, if you want to optimise your longevity and live a long and healthy life, this is a very important part of the equation. Early research suggested that LDL was causative to cardiovascular disease and follow up studies using cholesterol lowering medications have shown a reduction in the rate of cardiovascular events, particularly in those who already have CVD. However, much of this research was focused only on ischemic heart disease (IHD – which can lead to heart attacks) and did not account for many confounding variables or look into the association with all-cause mortality. As we learn more about lipid metabolism and the pathogenesis of atherosclerotic CVD (ASCVD), there is a growing body of research challenging this oversimplified model. Many people still believe that LDL is causative although there are many other precipitating factors that make LDL necessary but not sufficient to cause CVD. I won’t get into this debate here, what I will focus on is the all-cause mortality research and what this suggests is the optimal total and LDL cholesterol levels with the lowest risk of dying from all-causes (which is what we should focus on if we want to live a long life).
Total Cholesterol
Let us first look at the role of TOTAL cholesterol (TC) by reviewing a study by Sang-Wook Yi et al. 2019 [1], where they stated the above importance well - “Although disease-specific morbidity and mortality, such as IHD mortality, have their analytical merits, all-cause mortality is arguably the most important endpoint for patients or the general population when assessing risk factors and the effectiveness of a treatment or a public health intervention for life-threatening diseases”. They go on to discuss how the current cholesterol guidelines have mostly been determined based on IHD and not all-cause mortality. They studied an extremely large cohort of over 12.8 million participants from Korea over 10 years and assessed total cholesterol levels and all-cause mortality. What they found was a reverse J curve across all age groups whereby the LOWEST cholesterol was associated with the HIGHEST risk of all-cause mortality.
Specifically, “In the spline analysis (Figure 1), the TC ranges associated with the lowest mortality were approximately 200–240mg/dL (5.172- 6.206 mmol/L), except for men at 18–34 years (approximately 180–220mg/dL) and for women at 18–34 years (approximately 160–200mg/dL) and at 35–44 years (approximately 180–220mg/dL).” This would put it well above the current guidelines to optimal cholesterol ranges. Interestingly, the people in the higher centiles of TC were also older and had higher levels of blood glucose, blood pressure and BMI – all risk factor of CVD and all-cause mortality. They go on to state “Our study suggested that the optimal ranges for overall survival are higher than that those for IHD.” This study aligns with many others showing the same trend – high cholesterol is protective against all-cause mortality.
Figure 1
Ok, but IHD is important as well since we also don’t want to die of IHD... This is where there is perhaps a lot more nuance about TC and LDL and it’s causative role in IHD. Let us look at another paper published in the Lancet back in 2007 [2] which was a meta-analysis of 61 prospective observational studies looking at over 890,000 healthy adults which also showed similar results. Although there was an increased rate of IHD with higher TC (figure 2), this was mostly attenuated when age and blood pressure were accounted for. Moreso, when they accounted for HDL levels, “the ratio of total/HDL cholesterol was substantially more informative about IHD mortality than either, and was more than twice as informative as total cholesterol” (figure 2). They also found LOWER rates of stroke with HIGHER levels of TC and go on to state “The absence of any independently positive association between total cholesterol and stroke mortality in middle age (after allowing for systolic blood pressure) or in those with systolic blood pressure below 145 mm Hg, and the negative association of cholesterol with stroke mortality at older ages or at higher blood pressures, are unexplained”. So clearly this is again telling us that there is more to the equation than just looking at TC or LDL. The protective benefit of HDL-C is often not discussed enough and there are plenty of papers showing a protective effect of higher HDL-C when LDL-C is also elevated (which we won’t cover here).
Figure 2
Ok, but this is just TC which most medical practitioners nowadays are hopefully not looking too deep at. What about LDL as this is supposedly the cause of ASCVD that we want to lower?
LDL Cholesterol
Like TC, research has also shown that LDL is inversely associated with all-cause mortality, meaning that higher levels are associated with less risk.
Let’s look at another recent paper from 2021 by Ya Lui et al. titled Association between low density lipoprotein cholesterol and all-cause mortality: results from the NHANES 1999–2014 [3]. The authors of this paper also say, “Because the main goal of disease prevention is to prolong life, all-cause mortality is the most important and easy to determine result”.
As the title suggests, they used the database from NHANES with a cohort of >19,000 participants that met the criteria with a mean follow up period of 7.83 years. What they found was that in the age adjusted model the lowest LDL-C (<70 mg/dL or 1.8102 mmol/L) had the highest risk of all-cause mortality (HR 1.708). After adjusting for multiple confounders this was still present (HR 1.6 and 1.373 for Model 2 and 3 respectively) (Figure 3). They too found a U-shaped association with LDL and all-cause mortality, whereby the lowest and highest LDL have the highest risks, although the higher levels are less significant (Figure 4). This was still present after adjusting for multiple confounding variables like other diseases.
Figure 3
Figure 4
Interestingly, this study also showed the LOWEST LDL-C had the HIGHEST risk of cardiovascular mortality (HR 1.412, P 0.035) and the fourth highest had the lowest incidence (130-159 mg/dL or 3.362-4.112 mmol/L) (HR 0.807, P 0.248) and the highest LDL (>160 (4.138, P 0.660) with a HR of only 1.088 (virtually no increased risk) – which goes against the notion of “lower is better”.
Focusing back on all-cause mortality, let us look at a paper from Ditlev et al. 2020 titled Association between low density lipoprotein and all cause and cause specific mortality in Denmark: prospective cohort study [4]. As the title suggests, this was a cohort from Denmark of over 100,000 participants followed for ~9 years. They looked at various centiles of LDL levels and associated CVD, cancer and all-cause mortality and also found a U-shaped curve with all-cause mortality where the lowest (<1.8) and highest (>4.8) had the highest mortality (HR 1.25 and 1.15 respectively – note higher risk for lower LDL). They reported the lowest risk was at 3.6 mmol/ for the overall population and in individuals not receiving lipid lowering medications. However, when we look at the data and hazard ratios, we can see that even 4-4.8 had nearly the same HR in the adjusted models (HR 1.02 and 1.04) (Figure 5). This is WELL above the LDL guidelines. In fact, they even show “in the fractional polynomials, the concentration of LDL-C associated with the lowest risk of all cause mortality was 4.1 mmol/L”!
Figure 5
This is even more significant when we consider that those in the top 4 centiles had the highest blood pressure and those in the top 2 had the highest pack years of smoking cigarettes (Figure 6). Of course these are accounted for with the statistical analysis, but even so, these are not perfect tools and its not unusual for the effects of these to still be present. And yet, they had the lowest risks at the 3rd and 2nd highest LDL levels…
Figure 6
We can see the same sort of results in a study by Sung et al. 2019 [5] where they state, “A higher risk of all-cause mortality was observed in lowest LDL-C group compared with the LDL-C 100–129 mg/dL as the reference group, this remained after accounting for confounding variables”. When we look at the HRs in this study we see a similar pattern; higher LDL (130-159 mg/dL) is protective against all-cause mortality (Figure 7) even after accounting for various confounders.
Figure 7
And this is again in the presence of significant confounding disease states in those with higher LDL such as diabetes, elevated triglycerides, smoking and hypertension – all major risk factors for mortality (Figure 8)! And note again, when HDL is accounted for, we see the lowest risks (model 3).
Figure 8
As this study excluded those with statins (lipid lowering drugs) we can confidently say that the risk was not associated with the use of these medications. However, the argument is often made that other factors are at play when considering the increased mortality rate and lower LDL. Often reverse causality is blamed, however, many of the studies listed here accounted for this by excluding those who died within 1-3 years of the start of the study. The other arguments are that there is some other pathophysiology occurring where low LDL is associated with said pathological state. But yet again, we see in many studies that account for other diseases that LDL is still independently associated. Research suggests that is because LDL plays a pivotal role in various aspects of health such as the immune system where it acts as part of the immune system to inhibit/clear pathogens. In such a case, low LDL would predispose you to pathogens that can lead to other disease states or frailty.
Frailty
Frailty itself has often been a scapegoat for people to blame as a confounder for low LDL – which means that it’s not LDL that is the cause of the increased all-cause mortality but associated frailty itself (of which low LDL tends to be present because of the frailty or other diseases). For that reason, Schatz et al. [6] conducted a study to specifically account for frailty to determine if this was a significant confounder with LDL leading to mortality. They state “that the usual statistical adjustments for traditional coronary heart disease risk factors (ie, excluding older persons from cholesterol screening) do not account for possible changes associated with frailty, and are therefore inappropriate.” They assessed frailty by multiple measures: ability to undertake 17 activities of daily living, forced expiratory volume, grip strength, low haemoglobin, weight loss >10%. What they showed was that those in the highest quartiles of LDL (>5.15) had the lowest risks (Figure 9), even after accounting for frailty (Figure 10, cox 2) and age and other risk factors (Figure 10, cox 1 and 3) with the lowest risk in quartile 3 then 4 (4·87–5·43 and 5·44–9·88 mmol/L, respectively) – although only cox 1 was statistically significant (Figure 10). Furthermore, they also showed the only significant risk with mortality was associated with the lowest cholesterol (2.04-4.5 mmol/L, HR 1.64). So, at worst we can say that high cholesterol is not associated with mortality and at best we can see it’s associated with a protective effect regardless of frailty.
Figure 9
Figure 10
And we can also remove cancer as a cause of frailty and low LDL because we have very clear evidence (some shown within this article) that low LDL is still associated with all-cause mortality even when adjusting for or excluding those with cancer.
Selective Mortality
One thing to note with this U-shaped association with all-cause mortality, especially when we break down groups by age, is that there seems to be a stronger affect the older you get – typically over 60 years old. So it could be that there is some sort of selective mortality affecting the results whereby those who have some genetic protection against CVD/mortality will survive past 60 and those who do not have these same genetics will have a higher mortality at a younger age. This may be the case, although it seems unlikely given that many of the papers have used populations with a broad age range well under the age of 60 with long follow ups. What is likely going on is that there are other pathological states predisposing these people to earlier death such as more endothelial dysfunction or a higher tendency towards clotting/plaques not associated with LDL per se.
Closing
So looking at this data, should we still be supporting this notion of “lower is better” when it comes to total cholesterol and LDL? My position is no, but there's some nuance.
If we’re concerned about all-cause mortality, then it would appear that a HIGHER total cholesterol (5.2-6.2 mmol/L or 200-240 mg/dL) and LDL (3-4 mmol/L or 116-155 mg/dL) is beneficial, keeping in mind there does still appear to be a U-shaped curve whereby excessively high TC and LDL may be a higher risk – albeit a much lower risk than the lowest ranges. And if we account for confounding variables, its likely that TC and LDL are not the main factors.
But what about IHD? Although there is data showing association and direct improvement with treatment of TC and LDL on IHD, this is likely oversimplified. And this takes us to the crux of this issue – we are likely looking at the wrong things. For so long we have focused on a reductionist view with the lipid hypothesis and ASCVD and CVD that we have likely missed the more important factors and complexity of lipid metabolism and pathogenesis. What are the proximate causes of ASCVD and many other CVD diseases? Metabolic dysfunction and oxidative damage – particularly damage to the endothelial lining and glycocalyx. The LDL just so happens to be present because it is a repair mechanism against this damage - meaning it is not the cause per se (hence necessary but not sufficient).
What is far more valuable to look at are markers of metabolic health. And research that has looked into markers like LPIR (Lipoprotein Insulin Resistance) Index and Triglyceride ratios have shown far higher associations with CVD and all-cause mortality. And new research is showing that in people with good metabolic health but high LDL, there is potentially no increased risk of plaque formation (see research on Lean Mass Hyper-responders by Dave Feldman et al.) - although this is early data which needs to be repeated.
In my opinion, an elevated TC and LDL (and perhaps even ApoB) in the context of a healthy metabolism (good LPIR, insulin, HbA1c, LDL-P, trigs, HDL, homocysteine), is likely not a major concern for ASCVD and is likely to have health and all-cause mortality benefits across the lifetime, but particularly as you age beyond 50 years old.
However, if you’re metabolically unhealthy and generally don’t live a healthy lifestyle and you have poor lipid/metabolic markers along with high blood pressure, then in all likelihood, an elevated TC and LDL will eventually lead to plaque/ASCVD/CVD – which is still the number 1 killer. In which case, people may benefit from lipid lowering medications to a moderate level (not as low as possible).
Of course there is a lot more nuance to this story, and there are other markers associated with lipids like LDL-P (particle count), ApoB and Lp(a) which have shown stronger associations with disease which we’ll save for another article. And we now have the ability to image the coronary arteries around the heart at fantastic resolution with a Coronary CT Angiogram (CCTA) to determine our total plaque burden which would lead to a heart attack or stroke. I’ll cover these in a follow up article.
Thoughts/Comments
I’m just a simple clinician trying to find truth and understand this very complex area of physiology and recognise I may be wrong or miss things. So let me know your thoughts by leaving a comment and let’s start a conversation on this very important and exciting area of research and medicine!
References
Yi, S.-W., Yi, J.-J. & Ohrr, H. total cholesterol and all-cause mortality by sex and age: a prospective cohort study among 12.8 million adults. doi:10.1038/s41598-018-38461-y.
The Lancet, 370(9602), 1829–1839 | 10.1016/s0140-6736(07)61778-4. https://sci-hub.se/10.1016/s0140-6736(07)61778-4.
Liu, Y. et al. Association between low density lipoprotein cholesterol and all-cause mortality: results from the NHANES. Scientific Reports | 11, 22111 (123AD).
Ditlev, C., Johannesen, L., Langsted, A., Mortensen, M. B. & Nordestgaard, B. G. Association between low density lipoprotein and all cause and cause specific mortality in Denmark: prospective cohort study. doi:10.1136/bmj.m4266.
Sung, K.-C. et al. Clinical Medicine Low Levels of Low-Density Lipoprotein Cholesterol and Mortality Outcomes in Non-Statin Users. doi:10.3390/jcm8101571.
Schatz, I. J. et al. Cholesterol and all-cause mortality in elderly people from the Honolulu Heart Program: A cohort study. Lancet 358, 351–355 (2001).
Comments