Does Elevated High-Density Lipoprotein Cholesterol Protect Against Cardiovascular Disease?

Abstract

High-density lipoprotein (HDL) contributes to reverse cholesterol transport, which is 1 of the main explanations for the described inverse association between HDL-cholesterol (HDL-C) and atherosclerotic cardiovascular disease (ASCVD) risk. However, efforts to therapeutically raise HDL-C levels with niacin, fibrates, or cholesteryl ester transfer protein inhibitors have not demonstrated a reduction in ASCVD events when compared with placebo among individuals treated with statins. Furthermore, mendelian randomization studies suggest that HDL-C is unlikely to be a direct biologic variable impacting ASCVD risk. More recently, observations from well-conducted epidemiologic studies have indicated a nonlinear U-shaped relationship between HDL-C and subclinical atherosclerosis, and that very high HDL-C (≥80 mg/dL in men, ≥100 mg/dL in women) is paradoxically associated with higher all-cause and ASCVD-related mortality. These observations suggest that HDL-C is not a universal protective factor for atherosclerosis. Thus, there are several opportunities for reframing the contribution of HDL-C to ASCVD risk and related clinical calculators. Here, we examine our growing understanding of HDL-C and its role in ASCVD risk assessment, treatment, and prevention. We discuss the biological functions of HDL-C and its normative values in relation to demographics and lifestyle markers. We then summarize original studies that observed a protective association between HDL-C and ASCVD risk and more recent evidence indicating an elevated ASCVD risk at very high HDL-C levels. Through this process, we advance the discussion regarding the future role of HDL-C in ASCVD risk assessment and identify knowledge gaps pertaining to the precise role of HDL-C in atherosclerosis and clinical ASCVD.

High-density lipoprotein (HDL) is 1 of 5 major lipoprotein classes and is involved in transporting cholesterol to the liver and steroidogenic organs from lipid-laden macrophages in atherosclerotic arteries (1). This function, coined as reverse cholesterol transport, has been 1 of the main explanations for the classical inverse association between HDL-cholesterol (HDL-C) and atherosclerotic cardiovascular disease (ASCVD) events in population-based studies (2, 3). Several of these observations were initially described within the Framingham Heart Study in the 1970s and showed a strong inverse association between HDL-C and coronary heart disease among persons older than 50 years of age (3-6). However, efforts to therapeutically raise HDL-C levels with niacin (7), fibrates (8), or cholesteryl ester transfer protein (CETP) inhibitors (9) have not demonstrated a reduction in ASCVD events among individuals receiving statin therapy (10). CETP inhibition has been associated with a reduction in HDL-C–mediated reverse cholesterol transport (11, 12), which is associated with an increased ASCVD risk and suggests that HDL-C–mediated reverse cholesterol transport is 1 of the main determining factors in protection against ASCVD. Mendelian randomization studies suggest that HDL-C is unlikely to be a direct biologic variable impacting the risk for ASCVD (13-15). More recent observations indicate a U-shaped relationship between HDL-C and subclinical atherosclerosis (Fig. 1) (17), and that very high HDL-C, ≥80 mg/dL in men and ≥100 mg/dL in women (Table 1), is paradoxically associated with higher all-cause and ASCVD mortality (16, 17). This evidence suggests that HDL-C may not be a universal protective factor for atherosclerosis and that kinetics and reverse cholesterol transport properties should also be evaluated when interpreting HDL-C values and ASCVD risk. Thus, there are opportunities for reframing the contribution of HDL-C to ASCVD risk and for evolving modeling equations in risk assessment.

Figure 1.

U-shaped association of HDL-cholesterol with all-cause and ASCVD mortality. Reproduced with permission from Liu et al. Association Between High-Density Lipoprotein Cholesterol Levels and Adverse Cardiovascular Outcomes in High-risk Populations. JAMA Cardiol. 2022;7(7):672-680 (16).

Table 1.

Definition of high-density lipoprotein cholesterol groups, stratified by sex

	Men (mg/dL)	Women (mg/dL)
Low high-density lipoprotein cholesterol	<40	<50
Normal high-density lipoprotein cholesterol	40-80	50-100
Very high high-density lipoprotein cholesterol	≥80	≥100

This review focuses on our growing understanding of HDL-C and its role in ASCVD risk assessment, treatment, and prevention. We first discuss the biological functions of HDL-C and describe its normative values in relation to demographics and lifestyle markers, including diet, alcohol intake, and exercise. We then summarize original studies that observed a protective association between HDL-C and ASCVD risk. Next, we examine more recent studies that have suggested elevated ASCVD risk at higher HDL-C levels beyond normal range, and how this evidence may eventually contribute to current risk prediction models. Through this process, we will advance the discussion regarding the role of HDL-C in routine ASCVD risk assessment and identify current knowledge gaps pertaining to the precise role of HDL-C in atherosclerosis and clinical ASCVD.

Biological Functions of HDL-C

Composition and Distinguishing Features of HDL

HDL particles in plasma are predominantly spherical and discoidal with many subspecies that vary in lipid and protein composition, size, density, shape, and electrophoretic mobility (1) (Fig. 2). Despite these differences, the overall structure is consistent: a surface monolayer of mainly phospholipids and unesterified cholesterol, in which apolipoproteins are embedded, surrounding a water-insoluble neutral lipid core of mainly cholesteryl esters and some triglycerides (1). The 2 major components of HDL, proteins and lipids, vary in concentration spanning 4 to 5 orders of magnitude (18). On average, HDL carries 50 to 100 molecules of esterified or unesterified cholesterol or phosphatidylcholine and 2 to 3 molecules of ApoA-I, the most abundant apolipoprotein on HDL. However, the lipid composition of HDL is tremendously dynamic (18).

Figure 2.

Schematic of high-density lipoprotein.

There are numerous approaches to distinguish and categorize heterogeneity among HDL particles (Table 2). Distinguishing features of HDL include charge and shape, density that subsequently can be subfractionated into size, and apolipoprotein content. Nevertheless, most published epidemiological studies focusing on HDL use HDL-C levels as the primary marker for predicting ASCVD risk. Future studies may incorporate various HDL components once they are validated (including HDL-C, HDL particle number [HDL-P], HDL subfractions, HDL-C efflux capacity, and HDL kinetics) for a more comprehensive assessment of the relationship between HDL with cardiovascular outcomes.

Table 2.

Characteristics and distinguishing features of high-density lipoproteins

Subclassification	Features
Density	HDL2 (∼40%, larger, more lipid rich, lower density)
	HDL3 (∼60%, smaller, more protein rich, higher density)
Apolipoprotein content	ApoA-I (∼70%)
	ApoA-II (∼15%-20%)
	ApoA-IV (≤10%)
	Apo C (≤10%)
	Apo E (≤10%)
	Apo M (≤10%)
Charge and shape	α-HDL (spherical): preα1, preα2, preα3, α1, α2, α3 and α4
	Pre-β HDL (discoidal): preβ1 and preβ2
Particle number	High (≥32 µmol/L)
	Medium (28.0-31.9 µmol/L)
	Low (<28.0 µmol/L)
Particle size	Small HDL (7.3-8.2 nm)
	Medium HDL (8.2-8.8 nm)
	Large HDL (8.8-10.3 nm)

Abbreviations: Apo, apolipoprotein; HDL, high-density lipoprotein.

Biological Function

Because HDL consists of such a vast heterogeneity in terms of the structural, physiochemical, and composition of its components, it is evident that HDL functionality cannot be inferred from simply measuring plasma HDL-C levels. Several studies have shown that HDL-C functionality may even be more important than HDL-C level itself (19, 20). For example, 1 study performed in more than 2900 participants without ASCVD at baseline found cholesterol efflux capacity but not HDL-C levels predicted ASCVD events in a fully adjusted model (21). Although the major role of HDL is in cholesterol efflux and reverse cholesterol transport (22), there are several other biological properties of HDL that are thought to be important in maintenance of endothelial function, protection against oxidation, inflammation, thrombosis, and cytoprotection in immunomodulatory pathways (Table 3) (23).

Table 3.

Notable biological functions of high-density lipoproteins

Biologic function	Description
Macrophage cholesterol efflux	Accepts free cholesterol from lipid laden macrophage foam cells
	Transports excess cholesterol from peripheral tissues to the liver where it can be further metabolized
Maintenance of endothelial function	Upregulates expression of and stimulates endothelial nitric synthase
	Protects endothelial cells from apoptosis
	Promotes growth of endothelial cells
Reduction of LDL-C oxidation	Protects LDL from oxidative damage from free radicals
	Removes oxidized lipids from LDL
Anti-inflammatory properties	Decreases expression of inflammatory cytokines
	Inhibits the expression of endothelial wall adhesion molecules like VCAM-1 and TNF-α
	Reduces recruitment of inflammatory cells into the arterial wall
	Promotes M2 macrophage phenotype
Antithrombotic properties	Downregulates thrombin generation
	Decreases platelet aggregation
	Activates prostacyclin macrophage foam cells
	Activates antithrombotic proteins (protein C and protein S)
Immunomodulation	Binds bacterial products and neutralizes their toxins
	Modulates activity of macrophages, antigen presenting cells and lymphocytes by modulating cholesterol content in lipid rafts

Abbreviations: αHDL, high-density lipoprotein, LDL, low-density lipoprotein, TNF-α, tumor necrosis factor alpha; VCAM-1, vascular cell adhesion molecule 1.

Demographics and Lifestyle Factors in HDL-C

Ethnicity and Sex in HDL

HDL-C levels vary significantly among ethnic groups and races (24-26). These HDL-C variations appear to be at least partly genetic and socioeconomic in origin. After adjustment for weight, smoking, and alcohol consumption, Black individuals have higher HDL-C levels compared with White individuals (25). However, a more recent analysis performed in the Reasons for Geographic and Racial Differences in Stroke study showed that HDL-C ≥ 60 mg/dL was not protective in either Black or White adults when compared with normal HDL-C values (40-59 mg/dL in men, 50-59 mg/dL in women), whereas low HDL-C (<40 mg/dL in men, <50 mg/dL in women) was associated with a 30% higher risk of ASCVD events in White but not in Black adults (27). Median HDL-C values stratified by ethnicity and sex are presented in Fig. 3. Except for Black and Japanese adults, all other minority ethnicities in the United States are more likely to have a lower prevalence of low HDL-C (28). Interestingly, HDL-C level differences between various ethnic groups can be observed beginning in childhood (29).

Figure 3.

Median high-density lipoprotein-cholesterol, stratified by sex and ethnicity. *Error bars represent first and third quartiles.

With respect to sex, women tend to have higher HDL-C levels than men (25, 30, 31). However, no significant sex differences have been observed among White or Black adults with familial hyperalphalipoproteinemia (5, 32), a lipid disorder characterized by elevated plasma HDL-C beyond the 90th percentile. Even though HDL-C levels decrease after menopause (25), women continue to have significantly higher levels compared with men through the seventh decade of life (33). When adjusted for waist/thigh ratio, HDL-C level differences between men and women become less pronounced (31).

Lifestyle and Behavioral Factors and HDL-C

Several lifestyle parameters including consumption of low carbohydrate/high unsaturated fat diet, weight loss, exercise, smoking cessation, and moderate consumption of alcohol increase HDL-C levels (34, 35). Overall, a high intake of refined carbohydrates is associated with lower HDL-C levels (25, 36), whereas replacement of 10% of carbohydrate energy with saturated fat, monounsaturated fat, and polyunsaturated fat increase HDL-C by 4.7, 3.4, and 2.8 mg/dL (0.12, 0.09, and 0.07 mmol/L), respectively, as seen in a large-scale meta-analyses (37). Furthermore, diets rich in omega-3 fatty acids are associated with higher HDL-C (38). After 8 weeks, participants on a DASH diet (diet high in fruits, vegetables, low-fat dairy products and low in total fat, saturated fat, cholesterol) demonstrated significant reductions in HDL-C as well as in total cholesterol and low-density lipoprotein-cholesterol (LDL-C) (39).

Weight loss and physical activity are also notable contributors to HDL-C levels. A significant inverse association of visceral adiposity with HDL-C (40, 41) and HDL-P (42) has been previously demonstrated. During weight loss, HDL-C is reduced 0.27 mg/dL for every kilogram of weight lost (43). Regarding physical activity, exercise intensity and frequency appear to have different effects on HDL-C levels (44). A meta-analysis showed that moderate (45%-55% maximum heart rate) or high-intensity (65%-75% maximum heart rate) exercise, increased HDL-C by 4.6% on average compared with control groups (44). Among elderly men and women, participants in the exercise group had no significant HDL-C changes compared with the control group (45). In persons with type 2 diabetes, only the group with the highest physical activity experienced statistically significant increase in HDL-C levels, whereas less-intense physical activity did not confer higher HDL-C levels (46).

HDL-C is also influenced by cigarette smoking and alcohol intake. A large meta-analysis showed that smoking cessation was associated with a 3.9 mg/dL (0.10 mmol/L) higher level of HDL-C without any changes in the total cholesterol, LDL-C, or triglycerides levels (47). Furthermore, moderate consumption of alcohol is associated with higher HDL-C. Compared with nonalcoholic beverages, red wine (48, 49), whiskey (50), and beer (51) have all been associated with increased HDL-C levels. The magnitude of HDL-C change correlates with the amount of alcohol intake and does not appear to be affected by different beverage types (52). The quantity of alcohol needed to raise HDL-C by 3.87 mg/dL (0.1 mmol/L) is approximately 30 g/day (52). Although large meta-analyses have shown a linear correlation between alcohol consumption and HDL-C concentrations (52, 53), a recent observational study in a Chinese population showed a U-shaped association, with excessive alcohol consumption eventually leading to lower HDL-C (54).

Combining different lifestyle modifications that each individually associate with higher HDL-C appears to have an additive effect on HDL-C levels. A group of males that lost weight with exercise and a diet including <30% calories from fat, <10% calories from saturated fatty acids, and <300 mg/day cholesterol experienced higher HDL-C compared with a group that lost weight but performed no exercise (55). Similarly, another study showed a 9.9% increase in HDL-C in obese women who experienced significant weight loss via diet and exercise. This change was higher than in patients from other studies that either lost weight or only intensely exercised (56).

Hyperinsulinemia and the metabolic syndrome are also key contributors associated with reduced HDL-C, all of which underlie a sedentary lifestyle. The pathophysiology of hyperinsulinemia and the metabolic syndrome include hypertriglyceridemia, which ultimately leads to a higher proportion of triglyceride-enriched HDL that are more rapidly cleared by the kidneys and thus results in lower concentrations of HDL in the periphery (57-59).

Evidence for HDL-C as a Protective ASCVD Risk Factor

Data from large, prospective observational studies have shown that high HDL-C levels are protective against the incidence of ASCVD. In the Framingham Heart study, lipid parameters were measured in 2815 men and women free of ASCVD and aged 49 to 82 years. At a mean follow-up of 4 years, high HDL-C levels were associated with lower incidence of ASCVD, such that participants with HDL-C above 65 mg/dL had an 8-fold lower crude incidence of ASCVD events compared with those <35 mg/dL (3). Gordon et al subsequently validated these findings in 4 large, prospective observational cohorts and concluded that there was a 2% to 3% reduction in ASCVD risk per 10 mg/dL higher HDL-C level (60). The Framingham risk score was among the first composite ASCVD risk scores to incorporate HDL-C as a negative risk factor in risk equations (61). Subsequently, the American College of Cardiology/American Heart Association endorsed the Pooled Cohort Equations in 2013 for estimating 10-year risk of developing ASCVD based on data from 4 large, diverse, observational cohorts—Atherosclerosis Risk in Communities, Cardiovascular Health, the Coronary Artery Risk Development in Young Adults, and the Framingham Original and Offspring Study cohorts. They included HDL-C levels in their risk prediction model, which continues to be widely incorporated in clinical practice (62).

It is noteworthy that not only do high HDL-C levels confer protection against ASCVD, but low levels have been linked with a higher risk of ASCVD in individuals with and without established ASCVD, independent of LDL-C levels. In a post hoc analysis of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation trial, Acharjee et al evaluated 2193 patients with stable ischemic heart disease on optimal medical therapy and found that participants in the highest quartile of HDL-C levels (48-94 mg/dL) had a 33% lower crude rate of death or myocardial infarction (MI) compared with those in the first quartile of HDL-C (<35 mg/dL), with no interaction with LDL-C levels. Among patients with LDL-C levels <70 mg/dL, those in the highest quintile of HDL-C (48-94 mg/dL) had a 65% relative risk reduction in death or MI compared with the lowest quintile (<35 mg/dL), independent of traditional risk factors (63). In 302 430 participants free of established ASCVD enrolled in the Emerging Risk Factors Collaboration, per SD (15 mg/dL) higher HDL-C was associated with a 22% lower incidence of ASCVD after adjusting for traditional ASCVD risk factors as well as other lipid parameters (2).

Genetics, Therapeutics, and Higher HDL-C

Despite significant epidemiological evidence supporting an increased risk of ASCVD conferred by low HDL-C levels, the exact mechanism behind this association, particularly the potential causal role of HDL-C, remains uncertain. Some authors have argued that not all patients with low HDL-C have an increased risk of ASCVD. In fact, hereditary syndromes characterized by low HDL-C levels, such as Tangier disease, fish eye disease, and lecithin cholesterol acyltransferase deficiency, seldom manifest with premature atherosclerosis (64). Sirtori et al evaluated 21 carriers of the apolipoprotein A-I Milano mutant allele, a genetic mutation that presents with very low HDL-C levels, with age- and sex-matched control subjects. They found no difference in the prevalence of subclinical ASCVD as measured by carotid intima media thickness (65).

In addition to studying individuals with inherited lipid disorders, Mendelian randomization (MR) has been used as an approach to understand the role of HDL-C in ASCVD risk. MR is a method used in epidemiologic research to investigate the potential causal role of a modifiable exposure on an outcome of interest, relying on previously established genetic instrumental variables of the exposure (66). The MR study by Voight et al found that higher HDL-C levels do not have a causal protective effect on MI risk (15). Another MR study by Holmes et al also concluded that the causal role of HDL-C on coronary heart disease is unclear (14). More recently, the large genome-wide association study conducted by Klarin et al analyzed genetic data from more than 300 000 individuals of diverse ethnic backgrounds and identified genetic variants associated with blood lipid levels, including a total of 157 HDL-C loci reaching genome-wide significance (67). Following the genome-wide association study, the authors also found that there is a lack of association between HDL-C loci and coronary artery disease.

Familial hyperalphalipoproteinemia underlies genetically elevated plasma HDL-C and may have several primary causes, including deficiencies in CETP, hepatic lipase, and/or apolipoprotein CIII, as well as mutations in scavenger receptor class B type I or endothelial lipase (68). Familial hyperalphalipoproteinemia is a heterogenous disease and is not generally associated with any hallmark clinical features beyond potential corneal opacity (69). Though earlier studies have suggested a longevity syndrome associated with familial hyperalphalipoproteinemia, more recent evidence has shown a higher proportion of ASCVD in persons with familial hyperalphalipoproteinemia, particularly among those with CETP or hepatic triglyceride lipase deficiencies (70). Such studies involving individuals with familial hypercholesterolemia thus demonstrate that markedly elevated HDL-C is not necessarily associated with an anti-atherosclerosis environment.

Furthermore, clinical trials have demonstrated that raising HDL-C does not reduce ASCVD risk, and that the relationship between HDL-C and ASCVD is complex and not yet fully understood. Both CETP inhibitors and niacin (added to statin therapy) significantly increase HDL-C level; however, neither have been shown to provide a meaningful reduction in ASCVD risk (71-75). A proposed explanation is that CETP inhibitors and niacin preferentially increase the level of large HDL-P and, given the heterogeneity of structure and function of HDL-P, such therapies have weaker effects on the total number of HDL-P (76).

There is also a potential concern that CETP inhibitors may inhibit reverse cholesterol transport by forming a nonproductive complex with CETP and HDL particles (77); evidence regarding CETP inhibition is continuously evolving (78). CETP inhibition by torcetrapib in patients with HDL-C <40 mg/dL increases the amount of apoA-I within alphaI-migrating HDL and decreases the fractional catabolic rate of HDL-C (11); however, torcetrapib also increases serum aldosterone and blood pressure (78). Similar untoward off-target side effects have also been noted for the CETP inhibitor, dalcetrapib (78). Most recently, the CETP inhibitor, obicetrapib has been found to lead to a 51% reduction in LDL-C and 165% increase in HDL-C when compared with placebo among patients taking high-intensity statin therapy in a phase 2 clinical trial (79). There are new therapeutic targets currently being tested related to HDL, among which is ApoA-I infusion for reducing the risk of recurrent events among patients who have suffered an MI (80).

Elevated HDL-C and Atherosclerotic Cardiovascular Disease Risk

Although HDL-C has been traditionally viewed as protective in ASCVD risk, many recent observational studies performed across independent cohorts have noted that very high HDL-C is associated with higher incident ASCVD and mortality (16, 17, 81-94) (Table 4).

Table 4.

Collection of cohort studies assessing the association of very high HDL-C with all-cause and cardiovascular disease mortality

Study (author, journal, and year)	Country	Follow-up duration	Study population	Outcomes	Summary
Studies in cohorts including individuals without cardiovascular disease (primary prevention)
Wilkins et al, Journal of the American Heart Association, 2014 (87)	United States	139 624 person-years for men	24 440	All-cause mortality: 5746 (23.5%)	Very high HDL-C was associated with a higher risk of all-cause mortality only among men
		167 622 person-years for women
Hirata et al, Journal of Atherosclerosis and Thrombosis, 2016 (88)	Japan	Mean 18 y	7019	All-cause mortality: 1598 (22.8%)	Very high HDL-C was not associated with a higher risk of all-cause or cause specific mortality
				Cardiovascular mortality: 450 (6.4%)
Ko et al, Journal of the American College of Cardiology, 2016 (89)	Canada	Mean 4.9 y	631 762	All-cause mortality: 17 952 (2.8%)	Very high HDL-C was associated with a higher risk of all-cause and cause specific mortality, which was more pronounced in men. However, a majority of deaths were not due to cardiovascular disease
				Cardiovascular mortality: 4658 (0.7%)
				Cancer mortality: 6850 (1.1%)
				Other mortality: 6444 (1%)
Madsen et al, European Heart Journal, 2017 (90)	Denmark	Median 6.0 y	116 508	All-cause mortality: 10 678 (9.2%)	Very high HDL-C was associated with a higher risk of all-cause and cardiovascular mortality in both men and women. However, very high HDL-C was not associated with cardiovascular disease events such as MI, IHD, and ischemic stroke
				Cardiovascular mortality: 2777 (2.4%)
				Cancer mortality: 2968 (2.5%)
				Other mortality: 2585 (2.2%)
Hirata et al, Journal of Clinical Lipidology, 2018 (91)	Japan	Mean 12.0 y	43 407	All-cause mortality: 4995 (11.5%)	Very high HDL-C was associated with a higher risk of cardiovascular disease events, CHD events, and stroke mortality, but not cardiovascular and all-cause mortality
				Cardiovascular mortality: 1208 (2.8%)
				CHD mortality: 289 (0.7%)
Mazidi et al, International Journal of Cardiology, 2019 (92)	United States	—	25 541	—	Very high HDL-C was associated with a higher risk of CHD and cancer mortality
Oh et al, Atherosclerosis, 2019 (93)	South Korea	Median 3.5 y	365 457	All-cause mortality: 9350 (2.6%)	A trend toward increased risk in all-cause mortality with very high HDL-C was noted only among men. No such trend was noted between cause-specific mortality and very high HDL-C in either men or women
				Cardiovascular mortality: 1585 (0.4%)
				Cancer mortality: 3750 (1%)
Kobayashi et al, Atherosclerosis, 2019 (94)	Japan	Median 4.8 y	83 100	All-cause mortality: 382 (0.5%)	Very high HDL-C was not associated with a higher risk of all-cause mortality or cardiovascular disease events
				Cardiovascular events: 2023 (2.4%)
Liu et al, American Journal of Cardiology, 2021 (16)	United Kingdom	Median 9 y	415 416	All-cause mortality: 15 320 (4%)	Very high HDL-C was associated with a higher risk of all-cause and cardiovascular mortality only among men
				Cardiovascular mortality: 3881 (0.9%)
Trimarco et al, Hypertension, 2022 (81)	Italy	Mean 2.1 y	11 987	Cardiovascular disease events: 245 (2%)	Very high HDL-C was associated with cardiovascular disease events only among men with hypertension
Studies in cohorts including individuals both with and without established cardiovascular disease (primary and secondary prevention)
Bowe et al, Clinical Journal of the American Society of Nephrology, 2016 (82)	United States	Median 9.1 y	1 764 986	All-cause mortality: 541 682 (30.7%)	Very high HDL-C was associated with a higher risk of all-cause mortality
Hamer et al, Arteriosclerosis, Thrombosis, and Vascular Biology, 2018 (83)	United Kingdom	326 016 person-years	37 059	All-cause mortality: 2250 (6.0%)	Very high HDL-C was associated with a higher risk of all-cause mortality but not cardiovascular mortality
				Cardiovascular mortality: 649 (1.8%)
Li et al, Journal of Clinical Endocrinology and Metabolism, 2019 (84)	United States	Median 5.9 y	7766	All-cause mortality: 1921 (24.7%)	Very high HDL-C was associated with a higher risk of all-cause mortality
				Cardiovascular mortality: 665 (8.6%)
				Noncardiovascular mortality: 1256 (16.2%)
Li et al, Scientific Reports, 2019 (85)	China	Mean 8.8 y	100 070	All-cause mortality: 7362 (7.4%)	Very high HDL-C was associated with a higher risk of all-cause only among individuals under 65 years of age, particularly among men
Studies in cohorts including individuals with established cardiovascular disease (secondary prevention)
Liu et al, Journal of the American Medical Association Cardiology, 2022 (17)	United Kingdom and United States	United Kingdom: median 8.9 y	United Kingdom: 14 478; United States: 5467	United Kingdom	Very high HDL-C was associated with a higher risk of all-cause and cardiovascular mortality, particularly among men in the United Kingdom. A similar association of very high HDL-C and mortality without gender differences was noted in the United States
		United States: median 6.7 y		All-cause mortality: 1795 (12.4%)
				Cardiovascular mortality: 1137 (7.9%)
				United States
				All-cause mortality: 1471 (26.9%)
				Cardiovascular mortality: 757 (13.8%)

Abbreviations: CHD, coronary heart disease; HDL-C, high-density lipoprotein cholesterol; IHD, ischemic heart disease; MI, myocardial infarction.

In particular, a U-shaped association between HDL-C and several ASCVD outcomes has been described. Although no single definition of very high HDL-C has been consistently utilized across studies; we propose that very high HDL-C should generally be defined as ≥80 mg/dL in men and ≥100 mg/dL in women. Approximately 11.2% of women and 2% of men have HDL-C ≥80 mg/dL, whereas 1.3% of women and 0.2% of men have HDL-C ≥100 mg/dL (16). Nevertheless, an argument can be made to define elevated HDL-C according to percentiles because of subtle differences in the patient populations studies and the assays used to measure HDL-C (86). Previous studies to report on the association of elevated HDL-C can generally be categorized according to the patient population (primary vs secondary prevention) as well as study outcome (clinical ASCVD, non-ASCVD, and mortality).

One of the largest studies in the general population without known ASCVD was performed among more than 415 000 participants in the UK Biobank over a median follow-up of 9 years. Investigators found that elevated sex-specific HDL-C values (>80 mg/dL in men, >100 mg/dL in women) were associated with a respective 1.8-fold risk for all-cause and ASCVD mortality in men, but not women (16). The association of very high HDL-C with all-cause and ASCVD mortality persisted after adjusting for traditional risk factors as well as an HDL-C genetic risk score. However, whether the high HDL-C confers sex-specific ASCVD risk remains unknown as separate cohorts have observed elevated mortality risks uniformly in both men and women.

An additional large observational study assessing high HDL-C and outcomes was performed among 100 000 Danish participants in the Copenhagen City Heart Study and Copenhagen General Population Study (90). Beginning at HDL-C values of 97 mg/dL in men and 116 mg/dL in women, there was a 36% and 10% higher risk for all-cause mortality, respectively. Furthermore, hazard ratios (HR) for all-cause mortality approached an approximate 2-fold elevated risk for men (>116 mg/dL) and women (>135 mg/dL) at even higher HDL-C concentrations (90). On assessment of cause-specific mortality, event rates appeared to be driven by ASCVD, as HDL-C >116 mg/dL in men and >135 mg/dL in women conferred at least a 2.5-fold higher risk for ASCVD mortality. More recent follow-up studies have shown that individuals in the highest decile of ApoA-I levels, the major protein component of HDL, experience higher risks for all-cause (HR = 1.21) and ASCVD (HR = 1.15) mortality compared with persons in the lowest decile of ApoA-1 (95). The consistency of these results suggests that upper limits of HDL-C should be used clinically to identify at-risk primary prevention patients.

The association between elevated HDL-C and ASCVD outcomes has also been evaluated among patients with coronary artery disease. Among nearly 15 000 predominantly White patients with coronary artery disease enrolled in the UK Biobank, HDL-C ≥80 mg/dL conferred a 96% and 71% higher risk for all-cause and ASCVD mortality, respectively, when compared with HDL-C values between 40 and 60 mg/dL independent of traditional risk factors as well as common genetic polymorphisms for HDL-C (17).

In addition to hard study endpoints, elevated HDL-C has also been studied with respect to subclinical atherosclerosis outcomes across ethnicity and sex. Among nearly 6000 participants in the Multi-Ethnic Study of Atherosclerosis not on lipid-lowering therapy, continuous HDL-C had a significant U-shaped cross-sectional association with prevalent coronary artery calcium (CAC) >0 as well as CAC >100 (96). Although HDL-C >70 mg/dL and the ratio of HDL-C to HDL-P (a proxy measure for cholesterol-overloaded HDL) were not associated with baseline CAC >0, CAC >100, and/or CAC progression, HDL-C>100 mg/dL nominally conferred a 2-fold higher odds of CAC >100 (odds ratio = 2.05; 95% CI, 0.98-4.29) compared with intermediate values of HDL-C (40-70 mg/dL in men, 50-70 mg/dL in women) (96). Among women, menopausal stage has been found to modify the associations of HDL measure with CAC. Specifically, a higher value of small HDL-P and smaller HDL size have been associated with an approximate 2-fold higher odds of CAC presence in late perimenopause but not in the pre/early perimenopause stage (97). Last, patients with a high vs low ratio of HDL-C to HDL-P have been found to have a 56% higher odds of 5-year progression of carotid atherosclerosis, defined by change in total plaque area (98).

Among individuals with very-high HDL-C in previous studies, there has been no clear phenotypic pattern for adjacent circulating lipids, including total cholesterol, LDL-C, and triglycerides. There has been initial evidence that there is a positive association between HDL-C and lipoprotein(a) [Lp(a)], which may be 1 mechanism by which a higher ASCVD risk is observed among persons with very-high HDL-C (99-103). However, there is also speculation that Lp(a) vertical auto profile assays may misinterpret HDL2-C as Lp(a) among individuals with high blood concentrations of HDL2-C (104). Future studies that specifically assess the pattern of lipids among individuals with very high HDL-C may help to further characterize ASCVD risk in this patient population.

Biological Plausibility for Association of High HDL-C and Elevated ASCVD Risk

Although the exact mechanisms remain unknown, several hypotheses have been suggested for the potential increase in cardiovascular risk with very high HDL-C. The composition and the function of the HDL-P may be more relevant than the overall levels of HDL-C. ApoA-I, the main component of the HDL-C, which is responsible for the reverse cholesterol transportation, is prone to damage by oxidative stress and thereby not efficient in cholesterol efflux (71). HDL-C also has anti-inflammatory properties; however, in an oxidative environment, many proteins in the HDL-P are affected that can convert the HDL-P from an anti-inflammatory to a proinflammatory particle (105). Qualitative measurement of the HDL-C function, such as cholesterol efflux capacity, may be a better marker of cardiovascular events than HDL-C levels as seen in the Dallas Heart Study (106).

Potential genetic variants contributing to extremely high HDL-C levels were thought to potentially have other effects on the human body that could increase the risk, independent of the HDL-C levels. Liu et al demonstrated that the HDL-C genetic risk score was associated with HDL-C in a linear fashion and only accounted for 6% of the variation in HDL-C levels in the UK Biobank study. They also noticed that adjusting for the HDL-C genetic risk score in the UK Biobank study did not alter the outcomes such that the association of HDL-C with all-cause and cardiovascular mortality was thought to predominantly be driven by nongenetic causes (17). Last, alcohol intake has been hypothesized to confound the increased ASCVD risk among those with high HDL-C; however, studies have demonstrated a statistically significant higher mortality and ASCVD event rate after adjusting for alcohol intake (85).

Future Role of HDL-C in Risk Calculators

Recent studies implicating very high HDL-C in mortality, ASCVD events, and subclinical atherosclerosis suggest that the highest risk HDL-C metrics include high absolute concentration, cholesterol particle overload, and small particle size. Contemporary evidence also underlines a need to evolve the contribution of HDL-C to current risk calculators based on traditional risk factors. Currently, HDL-C contributes a negative, linear impact on calculators used to quantify 10-year ASCVD risk, including the Pooled Cohort Equations calculator that helps to guide statin eligibility among primary prevention patients (62).

The 2013 Pooled Cohort Equations calculator downweighs an individual's risk by natural log of 14-fold (White) and natural log of 19-fold (African American) for every milligram per deciliter higher HDL-C (62). However, it is not unreasonable to modify the contribution of HDL-C to resemble its U-shaped association with ASCVD risk and all-cause mortality and/or to consider high HDL-C as a risk enhancing factor within prevention guidelines. For both primary and secondary prevention, we suggest that HDL-C contribute to a higher 10-year ASCVD risk when values are <40 or ≥80 mg/dL in men and <50 or ≥100 mg/dL in women. When HDL-C values are between 40 and 80 mg/dL in men and 50 to 100 mg/dL in women, then HDL-C would not contribute toward increasing or decreasing an individual's 10-year ASCVD risk. Contrastingly, HDL-C may simply be used as a risk enhancing factor at the cutoff values listed previously and not contribute toward newer ASCVD risk calculators altogether.

Conclusion

Very high HDL-C (≥80 mg/dL in men, ≥100 mg/dL in women) should be recognized as a risk factor and/or enhancer for elevated ASCVD risk, especially among patients with established clinical ASCVD who have already experienced an index ASCVD event. Further progressing our approach to the measurement of HDL-C in the clinical setting may help to more precisely measure short-term and long-term ASCVD risk, identifying a subgroup of patients who may benefit from more intensive primary and secondary prevention regimens.

Abbreviations

ASCVD

atherosclerotic cardiovascular disease

CAC

coronary artery calcium

CETP

cholesteryl ester transfer protein

HDL

high-density lipoprotein

HDL-C

high-density lipoprotein cholesterol

HDL-P

high-density lipoprotein particle number

HR

hazard ratio

LDL-C

low-density lipoprotein cholesterol

Lp(a)

lipoprotein(a)

MI

myocardial infarction

MR

Mendelian randomization

Disclosures

The authors have no disclosures or conflicts of interest to report.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.