Further proof of a paradoxical relationship between high-density lipoprotein levels and adverse cardiovascular outcomes: are there implications for cardiovascular disease prevention?

This editorial refers to ‘U-shaped relationship between apolipoprotein A1 levels and mortality risk in men and women’, by C.C. Faaborg-Andersen et al. https://doi.org/10.1093/eurjpc/zwac263. Cardiovascular disease (CVD), mainly coronary heart disease, persists as the leading cause of disease burden and deaths in the world; in 2019, it was the cause of 6.2 million deaths among individuals between the ages of 30 and 70 years.1 Major cardiovascular risk factors include core health behaviours and health factors such as smoking, physical activity, diet, weight, lipids, blood pressure, and glucose control, which contribute to cardiovascular health.2 High systolic blood pressure and smoking were the first and second leading years of life lost (YLL) risk factors globally.2 The major lipids that are generally accepted as risk factors for CVD include total cholesterol, LDL cholesterol (LDL-C), HDL cholesterol (HDL-C) and triglycerides.3 High LDL-C remains a major threat to public health, due to elevated CVD risk, and its overall burden in terms of mortality and YLLs is on the increase globally.1 HDL is one of the five major groups of lipoproteins. It is an anti-atherogenic lipoprotein and it participates in the reverse cholesterol transport process which involves transportation of excess cholesterol in cells in the periphery to the liver, with ultimate excretion from the body in the faeces.4 It is well documented by several prospective cohort studies conducted across different ethnic and racial populations globally that HDL-C is a consistent and independent predictor of fatal and non-fatal CVD; the relationship is described as strong, graded and inverse.5,6 Compelling evidence of this strong association was first reported by the Framingham Heart Study,7 which led to the concept of HDL being referred to as the ‘good cholesterol’ and that interventions targeted at increasing levels of HDL-C could reduce the risk of atherosclerotic CVD (ASCVD) (the HDL hypothesis).8 However, in recent times, the HDL hypothesis has been challenged based on data from genetic and epidemiologic studies and failed HDL-targeted clinical trials.9–12 The protective role of HDL-C has been seriously questioned and whether there is a causal relationship between HDL-C and ASCVD remains controversial. Nevertheless, the value of HDL-C as a useful predictor of CVD risk is well established and unchallenged; HDL-C remains a valuable component of most cardiovascular risk calculators.5,13 A number of emerging studies, however, suggest that higher levels of HDL-C are also associated with increased risk of adverse cardiovascular outcomes, consistent with a U-shaped relationship.11,14 The European Society of Cardiology/European Atherosclerosis Society have considered the whole evidence and indicated in their dyslipidaemia guidelines that the risk of ASCVD appears to increase when HDL-C levels are above 90 mg/dL (2.3 mmol/L).15 Why there are negative effects of very high levels of HDL-C on ASCVD risk is not clear, but the following explanations have been proposed:16 (i) very high elevations in HDL may represent dysfunctional HDL in some individuals, which may increase rather than reduce the risk of CVD and (ii) genetic mutations that contribute to high HDL levels may also confer adverse vascular risk by unknown mechanisms. The vascular effects of very high levels of HDL-C are still not well understood and remains a subject of debate. Apolipoprotein A1 (ApoA1) is the major structural and functional protein which accounts for ∼70% of total HDL. Just like many plasma apolipoproteins, the main sites for ApoA1 synthesis and secretion are the liver and small intestine.17 The major functions of ApoA1 include interaction with cellular receptors and providing HDL with multiple anti-atherogenic activities.17 As expected, low levels of ApoA1 are also associated with an increased risk of CVD and have been shown to be consistent with a linear dose–response relationship.18–20 There have been reports that ApoA1 may actually be a stronger risk factor than HDL-C.18,19 Whether very high levels of ApoA1 are associated with an increased risk of adverse cardiovascular outcomes is uncertain; it is not known if a similar U-shaped relationship exists. Given the uncertainty in the evidence, Faaborg-Andersen et al.21 in this issue of the journal sought to investigate the nature and magnitude of the relationship between ApoA1 levels and adverse outcomes in greater detail than ever before. Their three linked objectives were to (i) evaluate the sex-specific relationships between very high ApoA1 levels and risk of cardiovascular and all-cause mortality; (ii) investigate if measurements of ApoA1 levels could help categorize individuals with very high HDL-C levels into clinically relevant sub-groups; and (iii) explore the genetic basis of the associations. They employed the UK Biobank cohort comprising 402 786 general population participants (average age of 56 years) without a known history of coronary artery disease at study entry who were followed up for a median duration of 12 years for the specified outcomes. In analysis that adjusted for several traditional cardiovascular risk factors, the authors demonstrated U-shaped relationships between ApoA1 levels and the outcomes of cardiovascular and all-cause mortality.21 Individuals in the highest decile of ApoA1 levels demonstrated higher risk of adverse outcomes compared with those within the lowest risk eighth decile; the risk was more pronounced in men than in women. On accounting for an ApoA1 polygenic risk score that explained 5.5% of the variance in ApoA1 levels across the study cohort, the associations were not attenuated. A reclassification analysis showed that ApoA1 measurements were not of benefit in categorizing individuals with very high HDL-C levels into clinically relevant sub-groups. The authors should be congratulated for the conducting this elegant study, especially for the novelty of the research question (being the first study to show a U-shaped relationship between ApoA1 and adverse cardiovascular outcomes), the use of real-world population-level data with its large sample size, and the comprehensive analysis.21 The limitations were mostly inherent to the study, and these include some of the biases of observational cohort designs such as residual confounding, regression dilution and reverse causation. It is not known if study participants were on any lipid lowering medication as this was not accounted for in their analyses. Statins are known to raise levels of HDL-C and ApoA1, and these elevations are maintained in the long-term.22 Another missing confounder was physical activity, which is also known to influence HDL-C levels. Given the strong relationship between ApoA1 and HDL-C and the high correlation observed (R2 = 0.84),21 it would have been interesting to see if the U-shaped relationship between HDL-C and adverse cardiovascular outcomes was existent in this cohort. Similar to HDL-C, the pathways underlying the paradoxical relationship between ApoA1 and adverse cardiovascular outcomes are not understood. The authors make an effort to propose plausible mechanisms, but these are all currently speculative and yet to be proved. Current accepted knowledge is that high levels of ApoA1 are associated with a decreased risk of adverse cardiovascular outcomes, consistent with graded relationships. The current findings which show that higher ApoA1 levels are not necessarily protective against these adverse outcomes are very timely and relevant. Several clinical questions, however, do arise from these findings. What higher levels of ApoA1 are detrimental to cardiovascular health? Faaborg-Andersen et al.21 do not provide a detail description of their dose response-curves, such as the levels of ApoA1 beyond which the risk begins to increase as well as any threshold values. However, the authors did indicate that men with levels 1.74–2.50 g/L and females with levels 2.01–2.50 g/L had higher risk of adverse cardiovascular outcomes.21 Another pertinent question is that, what are the implications of these findings with regard to lifestyle activities (e.g. physical activity) and medications (e.g. statins) that increase levels of HDL-C and subsequently that of ApoA1? Could these interventions contribute to an increased risk of these adverse cardiovascular outcomes? It is too early to answer these questions as the current evidence is based on a large but single observational cohort study. The U-shaped relationship needs replication in other large cohorts. The influence of very high levels of HDL-C on cardiovascular outcomes persists as a matter of debate and has not been fully clarified. Until future studies have shed some light on these paradoxical relationships and their underlying biological mechanisms, it should be kept in mind that very high levels of HDL-C and ApoA1 may be detrimental to cardiovascular health rather than protective. Deaths due to CVDs are still on the increase globally in some population groups. If the world is to meet the targets for Sustainable Development Goal 3 and achieve at least a 30% reduction in premature mortality due to noncommunicable diseases such as CVD by 2030,23 there should be an urgent focus on implementing cost-effective interventions and health policies. The American Heart Association reports that greater CVD-free survival and total longevity and higher quality of life can be augmented by greater adherence to its Life’s Simple 7 (LS7) goals, which involve improving seven health behaviours and factors (diet, participation in physical activity, exposure to cigarette smoking, body mass index, fasting blood glucose, total cholesterol, and blood pressure levels) through lifestyle changes.24 The LS7 has recently been updated to Life’s Essential 8, with inclusion of a new factor—sleep health—as well as changes to some of the original factors. In the absence of definitive evidence, these remain the cornerstone of CVD prevention. No funding was received for this work. The authors confirm that the data supporting the findings of this study are available within the article.