INTRODUCTION — High-density lipoprotein (HDL) cholesterol is a biomarker inversely associated with an increased risk of atherosclerotic cardiovascular disease (ASCVD) events. However, low levels of HDL cholesterol have not been established as causative of this increase in risk. (See 'Low HDL cholesterol as an ASCVD risk factor' below and 'Effect of increasing HDL cholesterol on clinical outcome' below.)
This topic will address clinical aspects of abnormal HDL (the particle) and HDL cholesterol. A more detailed discussion of the role of HDL in cholesterol metabolism is found elsewhere. (See "Lipoprotein classification, metabolism, and role in atherosclerosis", section on 'High-density lipoprotein'.)
HDL VERSUS HDL CHOLESTEROL — High-density lipoprotein (HDL) is a complex circulating particle with many subspecies that vary in lipid and protein composition [1]. (See "Lipoprotein classification, metabolism, and role in atherosclerosis", section on 'High-density lipoprotein'.)
Cholesterol is a major component of the particle and the amount of cholesterol contained in HDL particles can be directly measured; it is referred to as HDL cholesterol. In clinical practice, HDL cholesterol, rather than HDL, is used to risk stratify patients. (See "Measurement of blood lipids and lipoproteins", section on 'Total and HDL cholesterol'.)
LOW HDL CHOLESTEROL AS AN ASCVD RISK FACTOR — The incidence of atherosclerotic cardiovascular disease (ASCVD) events in multiple population studies has been found to be inversely related to the serum high-density lipoprotein (HDL) cholesterol concentration, with low levels being associated with increased coronary risk (figure 1) [2]. For example, based on data from the Framingham Heart Study, the risk for myocardial infarction increases by about 25 percent for every 5 mg/dL (0.13 mmol/L) decrement in serum HDL cholesterol below median values for males and females [3]. Most of these studies have found that, after adjustment for other known risk factors, a low level of HDL cholesterol is an independent predictor of risk. This does not mean, however, that it is causative of CHD. (See 'Low HDL cholesterol as a cause of ASCVD events' below.)
HDL cholesterol levels are predictive of coronary events in patients with known CHD across a broad range of low-density lipoprotein (LDL) cholesterol levels [4-6]. As an example, a post hoc analysis of 2193 individuals with stable disease who participated in the COURAGE trial of optimal medical therapy found that the rate of all-cause death or myocardial infarction was 33 percent lower in the highest compared with the lowest HDL cholesterol quartile [6]. (See "Chronic coronary syndrome: Indications for revascularization", section on 'Patients with high-risk anatomy'.)
Not all studies have found that HDL cholesterol is predictive of future events in patients with established cardiovascular disease (CVD) treated with statin therapy. For example, in a prospective cohort study of 6111 patients with CVD (SMART study) treated with intensive lipid lowering (n = 2046), HDL cholesterol was not associated with recurrent vascular events irrespective of LDL cholesterol level [7]. However, in those treated with no or usual-dose lipid-lowering medication, low HDL cholesterol levels were related to increased risk. (See "Epidemiology, risk factors, and natural history of lower extremity peripheral artery disease", section on 'Metabolic syndrome'.)
Low levels are common in patients with premature CHD (usually defined as men less than 55 to 60 years of age and women less than 65 years of age) [8-13]. Low HDL-cholesterol levels are more common in patients with a first myocardial infarction than in age-matched controls without CHD (19 versus 4 percent) [9]. The risk associated with low HDL cholesterol is independent of the risk attributed to elevated levels of LDL cholesterol and non-HDL cholesterol in most analyses [2], but the risk diminishes after adjustment for triglycerides, suggesting that triglyceride-rich lipoproteins are more important in contributing to atherosclerotic cardiovascular disease. Since low HDL cholesterol levels are most often associated with increased concentrations of atherogenic lipoproteins (eg, small-dense LDL particles and total numbers of LDL particles), many studies that assess cardiovascular risk in low HDL cholesterol patients have adjusted for LDL particle concentration. Multiple studies have shown that the risk of a low HDL cholesterol diminishes after adjustment for LDL particle concentration [14-16].
Not all disorders associated with low HDL cholesterol are accompanied by a predisposition to premature ASCVD [17]. Examples in which there is not a strong association with atherosclerosis include most patients with lecithin:cholesterol acyl transferase deficiency [18,19], and patients with the apolipoprotein A-1 Milano variant [20,21]. A familial history of premature CHD is often helpful in distinguishing high- from low-risk patients with low HDL cholesterol levels.
As mentioned above, most studies have suggested that low HDL cholesterol is an independent risk factor for CHD in multivariate analyses. However, a few studies challenge this view. In the Cardiovascular Health in Ambulatory Care Research Team (CANHEART) Study, an observational cohort of 631,762 individuals, low HDL cholesterol levels were associated with unhealthy lifestyles, higher triglyceride levels, other cardiac risk factors and medical comorbidities, and lower incomes [22]. These associations between low HDL cholesterol and socio-demographic, lifestyle, comorbidities, and mortality suggest that HDL cholesterol is a marker of overall CVD risk. The authors of the paper concluded that HDL cholesterol was not an independent risk factor when a series of associated conditions that have previously not been included in multivariate analyses were incorporated.
Low HDL cholesterol is also felt to be a precursor of the metabolic syndrome and type 2 diabetes, sometimes by many years [23,24]. (See "Metabolic syndrome (insulin resistance syndrome or syndrome X)".)
Isolated low HDL cholesterol — Low serum HDL cholesterol can occur as sole lipid abnormality or, more often, in association with hypertriglyceridemia or elevated LDL particles or apolipoprotein B levels [8]. Isolated low HDL cholesterol, a condition that occurs when HDL cholesterol is <50 mg/dL in women and <40 mg/dL in men, and triglycerides and LDL cholesterol are <100 mg/dL, predicts an increased risk for cardiovascular events in population studies [25,26]. The following two studies are representative:
●An analysis of the Framingham Offspring Cohort (3590 adults without known CVD) investigated the risk of CVD associated with low or high HDL cholesterol phenotypes (<40 mg/dL in men and <50 mg/dL in women) that were stratified by LDL cholesterol levels <100 mg/dL versus ≥100 mg/dL versus LDL cholesterol levels and stratified by triglycerides [26]. When compared with an isolated low HDL cholesterol (<40 mg/dL in men, <50 mg/dL in women), a high HDL cholesterol was associated with a 20 to 40 percent lower CVD risk. The risks associated with HDL cholesterol varied by the LDL cholesterol and triglyceride level. When compared with isolated low HDL cholesterol, CVD risks were increased when the low HDL cholesterol was accompanied by LDL cholesterol ≥100 mg/dL and triglycerides <100 mg/dL (odds ratio [OR] 1.3 [1, 1.6]), triglycerides ≥100 mg/dL and LDL cholesterol <100 mg/dL (OR 1.3 [1.1, 1.5]) or LDL cholesterol and triglycerides ≥100 mg/dL (OR 1.6 [1.2, 2.2]), after covariate adjustment.
●In an observational study which compared 158 individuals with isolated low HDL cholesterol with 780 individuals with an HDL cholesterol ≥40 mg/dL in men or ≥50 mg/dL in women, and triglycerides and LDL cholesterol <100 mg/dL, the risk of cardiovascular events was significantly increased (adjusted hazard ratio [HR] 1.93, 95% CI 1.11-3.34) [27]. However, the risk of death at 10.2 years was not different.
LOW HDL CHOLESTEROL AS A CAUSE OF ASCVD EVENTS — Despite the substantial body of evidence of an inverse relationship between high-density lipoprotein (HDL) cholesterol and cardiovascular event risk, low levels of HDL cholesterol have not been established as causative of this relationship or with the development of atherosclerosis [28].
The argument for lack of causality (for HDL cholesterol) comes from Mendelian randomization (see "Mendelian randomization") analyses and the difficulty in demonstrating improved outcomes with therapies to raise HDL cholesterol [1,29-32]. The absence of very early coronary heart disease (CHD) in individuals with rare disorders such as Tangier disease, who have very low levels of HDL cholesterol but not the predicted increase cardiovascular disease (CVD), provides some further support for the lack of association. (See 'Inherited causes' below.)
Additional evidence that may explain why low HDL cholesterol has not been shown to be causal comes from the Multi-ethnic Study of Atherosclerosis (MESA) cohort [15]. In this study, the associations between HDL cholesterol and HDL particle number determined by nuclear magnetic resonance spectroscopy with carotid intima-medial thickness (a known predictor of cardiovascular risk) and coronary events (such as myocardial infarction, CHD death, and angina) were evaluated in a group of nearly 6000 individuals without known CHD who were not taking lipid-lowering medication. After adjusting for each other (HDL cholesterol or HDL particle) and for other known predictors of CHD events such as low-density lipoprotein (LDL) particle number, age, sex, hypertension, and smoking, a significant inverse relationship between HDL particle, but not HDL cholesterol, and CHD risk (as measured by carotid thickness and CHD events) persisted. This study adds to the concept that the inverse relationship between HDL cholesterol and cardiovascular event risk may be determined more by some structural or functional component of the HDL particle than by its cholesterol content.
Finally, interventions to raise HDL cholesterol (as the only therapeutic target) have not uniformly demonstrated benefit, thus making it difficult to prove causality.
HDL PARTICLES AND CVD — Since high-density lipoprotein (HDL) is a heterogeneous and complex particle (size, charge, chemical composition, and functionality), it may be that components or properties other than cholesterol concentration are responsible (causative) for the observed inverse relationship between HDL cholesterol and clinical outcomes [33,34]. (See "Lipoprotein classification, metabolism, and role in atherosclerosis", section on 'High-density lipoprotein'.)
Most cardiovascular disease (CVD) events (myocardial infarction, stroke, cardiovascular death, revascularization) evaluated in randomized trials have atherosclerosis as the underlying pathology (see "Pathogenesis of atherosclerosis"). It is possible that one or more of the characteristics of the HDL particle influence the development of atherosclerosis. The genetic determinants of these components or functions may or may not be linked to the HDL cholesterol composition. As an example, two polymorphisms in phospholipid transfer protein are associated with increased concentrations of smaller, cholesterol-depleted HDL particles and a lower cardiovascular event rate [35]. Small, dense, protein-rich HDLs exhibit potent atheroprotective properties that can be attributed to specific clusters of proteins and lipids [36]. In addition, mutations in scavenger receptor class B type I (SR-B1) are associated with high HDL cholesterol levels but increased myocardial infarction risk [37].
A protective effect of HDL on atherosclerosis is suggested by the following observations:
●In humans, plasma HDL cholesterol concentrations above 75 mg/dL (1.9 mmol/L) are associated with prolonged life (called the longevity syndrome) and relative freedom from coronary heart disease (CHD) [3,38]. In a review of 18 kindreds with familial hyperalphalipoproteinemia, for example, individuals lived five and seven years longer than the general United States White population [38].
●Experimental models of atherosclerosis can be prevented in transgenic mice or rabbits expressing high levels of human apolipoprotein A-1 [39,40] by somatic gene transfer of apolipoprotein A-1 [41], by the administration of oral apo A-1 mimetic peptides [42], or by the administration of apolipoprotein A-1 Milano, which is a natural variant of apolipoprotein A-1 [43]. Furthermore, liver-directed gene transfer of human apolipoprotein A-1 results in significant regression of preexisting atherosclerosis after four weeks [44].
HDL particles have a variety of actions that can counteract atherogenesis [45]:
●Macrophage cholesterol efflux (figure 2 and figure 3) [29,45] (see 'Cholesterol efflux capacity and CVD risk' below)
●HDL may promote maintenance of endothelial function [46,47]
●HDL protects against oxidation of low-density lipoprotein (LDL) [48,49]
●HDL protects against inflammation [50]
●Immunomodulation [51]
●HDL may, via a variety of actions, interfere with the thrombotic component of atherosclerosis [52-55]
In contrast to the Mendelian randomization studies that investigated HDL cholesterol as a surrogate for HDL, a genome-wide association study phospholipid transfer protein (PLTP) polymorphisms showed that high concentrations of total and small HDL particles were associated with reduced CVD event risk [35].
HDL subclasses — HDL particles can be subcategorized into subclasses based on physicochemical properties. HDL2 and HDL3 are the major subclasses, with the former being described as larger and more buoyant and the latter as smaller and denser. These particles can be isolated by ultracentrifugation. (See "Lipoprotein classification, metabolism, and role in atherosclerosis", section on 'High-density lipoprotein'.)
Older studies have been inconclusive as to whether CVD risk is linked to HDL2 or HDL3. In the largest observational study (2015), worse cardiovascular outcomes were more strongly associated with HDL3 (lower compared with higher tertiles of HDL cholesterol) rather than HDL2 in nearly 4500 individuals with known CVD [56]. However, a large 2011 study found predictive value of any HDL subclass [36]. The composition and function of HDL subclasses varies with small, dense, protein-rich HDLs exhibiting potent atheroprotective properties that can be attributed to specific clusters of proteins and lipids.
A more robust measure of CVD risk is total HDL particle number. In the JUPITER trial, baseline and on-statin HDL particle number was the strongest HDL-related biomarker that was inversely predictive of incident cardiovascular events. In multivariate models that included adjustment for HDL cholesterol, apolipoprotein A-I, and macrophage cholesterol efflux capacity, the odds ratio [OR] for incident cardiovascular events at baseline was 0.65 (0.51 to 0.83, p<0.0001) and on-statin treatment 0.53 (0.34 to 0.82, p<0.005) [57]. (See "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease".)
Cholesterol efflux capacity and CVD risk — The HDL particle has multiple potentially antiatherogenic properties. Much of its antiatherogenic effect is thought to be mediated by its participation in the removal of cholesterol from macrophages in atherosclerotic plaques during a process termed "macrophage cholesterol efflux" (figure 3) [29,45]. This is the first step in the process of reverse cholesterol transport in which excess cholesterol is removed from the body (figure 4A and figure 4B and figure 5 and figure 2) [29,58]. (See 'Low HDL cholesterol as an ASCVD risk factor' above and "Lipoprotein classification, metabolism, and role in atherosclerosis", section on 'High-density lipoprotein'.)
Cholesterol efflux capacity is an ex vivo measure of HDL function [59]. The potential value of assessing macrophage cholesterol efflux was evaluated in a study of 2924 individuals free from CVD [60]. HDL cholesterol, HDL particle concentration, and cholesterol efflux capacity were measured (as were other lipoprotein variables) at baseline. The primary endpoint was atherosclerotic CVD, defined as a first nonfatal myocardial infarction, nonfatal stroke, or coronary revascularization or death from cardiovascular causes. After a median follow-up period of 9.4 years, there was a strong inverse relationship between cholesterol efflux capacity and the primary endpoint (adjusted hazard ratio [HR] 0.33, 95% CI 0.19-0.55) comparing the highest quartile with the lowest. Baseline HDL cholesterol was not associated with cardiovascular events in this study.
We believe this study helps to explain the limitations of using HDL cholesterol for risk prediction. Functional properties of the HDL particle (eg, macrophage cholesterol efflux capacity) are likely to be better predictors of CVD events. However, the utility of the macrophage cholesterol efflux assay has not been established in patients treated with high-intensity statins [34,57].
Cholesterol efflux capacity has also been evaluated as a CVD risk predictor. In one study it improved risk prediction beyond that of coronary artery calcification, family history, and high-sensitivity C-reactive protein [61]. (See "Overview of possible risk factors for cardiovascular disease", section on 'Coronary artery calcification' and "C-reactive protein in cardiovascular disease", section on 'Association of CRP with cardiovascular risk'.)
CETP AND CVD RISK — Cholesteryl ester transfer protein (CETP) is integrally involved in high-density lipoprotein (HDL) metabolism (figure 4A-B). In brief, it mediates the transfer of cholesteryl esters from HDL particles to the triglyceride-rich lipoproteins low-density lipoprotein (LDL) and very low-density lipoprotein (VLDL). Additional information is found elsewhere. (See 'Inherited causes' below and "Lipoprotein classification, metabolism, and role in atherosclerosis", section on 'Cholesteryl ester transfer protein'.)
Protein truncating variants at the CETP gene were associated with higher HDL cholesterol (22.6 mg/dL), lower LDL cholesterol (-12.2 mg/dL), lower triglycerides (-6.3 mg/dL), and a reduced risk for coronary heart disease (CHD; summary odds ratio [OR] 0.70 95% CI 0.54-0.90) in an analysis from 58,469 participants from 12 case-control studies (18,817 CHD cases, 39,652 CHD-free controls) [62].
As CETP lowers HDL cholesterol (in addition to other functions), therapies to decrease CETP levels have been studied for their ability to lower CHD events. (See 'CETP inhibition' below.)
CAUSES OF LOW PLASMA HDL CHOLESTEROL — Most causes of low high-density lipoprotein (HDL) cholesterol are inherited [63]. Other causes include:
●Drugs such as beta blockers, benzodiazepines, and anabolic steroids [64-66]
●Acute infection [67,68], inflammation [69,70], or gammopathies [71]
Inherited causes — While the following are regarded as heritable causes, the genes that account for the variation in serum HDL cholesterol have not been identified in all cases:
●Familial hypoalphalipoproteinemia – Apolipoprotein A-I plays an important role in HDL function. Low HDL cholesterol levels due to hypoalphalipoproteinemia can occur in one of three ways: impaired synthesis of apolipoprotein A-1 (apolipoprotein A-1 deficiency, apolipoprotein A-1/C-III deficiency, apolipoprotein A-1 structural variants); increased catabolism (familial HDL deficiency and Tangier disease); or enzymatic changes affecting HDL metabolism. The enzymatic changes are either genetically determined or, as with decreased activity of lipoprotein lipase (LPL), secondary to insulin resistance.
Familial hypoalphalipoproteinemia is an autosomal dominant disorder that has been linked to premature coronary heart disease (CHD) and stroke [72,73]. Affected patients have a primary depression in HDL cholesterol to below the 10th percentile compared with age- and sex-matched controls. The defect appears to be due to mutations in the apolipoprotein A-1 gene; it accounts for 6 percent of Japanese subjects with low HDL cholesterol concentrations [74].
●Familial HDL deficiency and Tangier disease – Familial HDL deficiency is an autosomal dominant disorder associated with very low serum HDL concentrations and premature CHD but none of the systemic findings in Tangier disease [75-78].
Tangier disease is an autosomal codominant condition in which homozygotes have extremely low levels of serum HDL cholesterol and heterozygotes have HDL cholesterol concentrations about one-half those in normal individuals [79]. HDL-mediated cholesterol efflux from macrophages and intracellular lipid trafficking are impaired in this disorder, leading to the presence of foam cells throughout the body and hepatosplenomegaly, peripheral neuropathy, and frequently premature coronary disease [77]. (See "Neuropathies associated with hereditary disorders", section on 'Tangier disease'.)
The defect in both familial HDL deficiency and Tangier disease appears to involve mutations in the adenosine triphosphate (ATP)-binding cassette transporter (ABCA1) gene, which, as noted above, encodes for the cholesterol efflux regulatory protein (CERP) [75,79-82]. The above observations suggest that ABCA1 plays a central role in intracellular cholesterol transport, which is impaired in both disorders and is associated with increased HDL catabolism [76,82]. It has been suggested that reduced cholesterol efflux onto nascent HDL particles leads to lipid depleted particles that are then rapidly catabolized [83].
●Familial combined hypolipidemia – This is a rare inherited disorder in which affected members have low plasma levels of HDL cholesterol, low-density lipoprotein (LDL) cholesterol, and triglycerides. Affected individuals do not appear to get premature CHD [84]. Familial combined hypolipidemia is discussed in greater detail elsewhere. (See "Low LDL-cholesterol: Etiologies and approach to evaluation".)
●Elevated cholesteryl ester transfer protein activity – Cholesteryl ester transfer protein (CETP) facilitates transfer of cholesteryl esters from HDL to triglyceride-enriched lipoproteins.
Elevated CETP activity has been documented in normotriglyceridemic men with low HDL cholesterol levels. In one series of 109 such men, 27 had increased CETP activity [85]. The patients with normal CETP activity had 20 percent lower concentrations of lipoprotein lipase and 25 percent higher concentration of hepatic triglyceride lipase, other changes that can lower HDL cholesterol.
The plasma concentration and activity of CETP is associated with polymorphism of the CETP gene, referred to as TaqIB. The presence of a gene variant, termed "B1," is associated with a higher CETP concentration and a lower HDL cholesterol level [86]. In an angiographic study of 807 men with documented coronary atherosclerosis, progression of the coronary disease was greatest in patients with the B1B1 genotype, intermediate in those with the B1B2 genotype, and least for those with B2B2 [86]. Therapy with pravastatin slowed the progression of coronary atherosclerosis in the high-risk B1B1 carriers but not in those with B2B2.
However, other studies have found that higher CETP concentrations and lower HDL concentrations reduce the risk of coronary atherosclerosis, suggesting that the TaqIB polymorphism is probably not the causative mutation responsible for the lower HDL concentrations that result in an increased risk of CHD. Support for this comes from a report that evaluated two other common mutations in CETP, A373P and R451Q, in 8467 healthy subjects and 1636 patients with CHD [87]. Compared with non-carriers, those who were heterozygous or homozygous for the A373P/R451Q polymorphism had lower HDL cholesterol concentrations. Women with this allele had a 36 percent lower risk of CHD (odds ratio 0.64); there was a similar trend in men (odds ratio [OR] 0.86). Similarly, some (but not all) studies have suggested an increased risk of CHD in patients with reduced CETP activity. (See 'CETP and CVD risk' above.)
The potential cardiovascular effect of CETP inhibition is discussed below. (See 'CETP inhibition' below.)
●Lipoprotein lipase deficiency – LPL normally lowers triglyceride levels through hydrolysis of triglyceride-enriched lipoproteins and facilitation of cholesterol transfer from these lipoproteins to HDL. Most mutations of the LPL gene lead in heterozygotes to impaired LPL activity, low HDL levels, and hypertriglyceridemia [88,89]. In one report, this pattern was present in 15 percent of patients with premature coronary disease [8]. Homozygotes have complete LPL deficiency with chylomicronemia and severe hypertriglyceridemia [88]. (See "Hypertriglyceridemia in adults: Management".)
●Elevated hepatic triglyceride lipase activity – Hepatic triglyceride lipase lowers HDL by hydrolyzing triglycerides and phospholipids from HDL particles. Increased activity of this enzyme is frequently found in patients with low HDL cholesterol levels with or without hypertriglyceridemia [90].
●Familial LCAT deficiency – After acquisition of free cholesterol onto HDL, the cholesterol is esterified to cholesterol esters by lecithin:cholesterol acyltransferase (LCAT). Patients with homozygous mutations of the LCAT gene have very low serum HDL cholesterol concentrations and typically present with nephropathy and proteinuria; severe corneal opacities (picture 1A-B), called the fish eye syndrome; and target cells [18,19] (see "Burr cells, acanthocytes, and target cells: Disorders of red blood cell membrane"). Premature coronary disease is usually not part of the phenotype [18] but can occur with some mutations even in heterozygotes [91].
●Insulin resistance – Low HDL cholesterol is a component of the atherogenic lipid phenotype that is characterized by obesity, insulin resistance, type 2 diabetes mellitus, dyslipidemia, and hypertension [92]. The atherogenic dyslipidemia consists of low HDL cholesterol, moderate elevations in triglyceride-enriched remnant particles, and borderline to high LDL cholesterol levels with a predominance of small, dense LDL particles. (See "Inherited disorders of LDL-cholesterol metabolism other than familial hypercholesterolemia".)
EFFECT OF INCREASING HDL CHOLESTEROL ON CLINICAL OUTCOME — In the aggregate, the evidence does not support a beneficial clinical impact with efforts to raise high-density lipoprotein (HDL) cholesterol. A 2009 meta-analysis of 108 randomized trials involving nearly 300,000 patients at risk for cardiovascular events of therapy (drugs or diet) found no association between treatment-induced increases in HDL cholesterol with risk ratios for coronary heart disease (CHD) deaths, CHD events, or total deaths after adjustment for changes in low-density lipoprotein (LDL) cholesterol [93].
Most of the randomized trials designed to evaluate the potential benefit of altering components of the lipid profile focused on the impact of LDL cholesterol. Benefits from increasing concentrations of HDL cholesterol have been suggested by two randomized trials and by multivariate analyses of observations from clinical trials in which the goal was lowering LDL cholesterol.
Two studies have found that raising HDL cholesterol in patients with a relatively low baseline serum concentration may be effective for secondary prevention of CHD:
●VA-HIT trial – The VA-HIT trial included 2531 with CHD who had an LDL cholesterol (≤140 mg/dL or 3.6 mmol/L), an HDL cholesterol (≤40 mg/dL or 1.0 mmol/L), and triglycerides ≤300 mg/dL (3.4 mmol/L); the patients were randomly assigned to treatment with gemfibrozil or placebo [94]. At one year, the following differences were noted in the gemfibrozil group:
•The mean HDL cholesterol level was 6 percent higher (34 versus 32 mg/dL [0.9 versus 0.8 mmol/L]).
•The mean total cholesterol was 4 percent lower (170 versus 177 mg/dL [4.4 versus 4.6 mmol/L]).
•The mean triglyceride concentration was 31 percent lower (115 versus 166 mg/dL [1.3 versus 1.6 mmol/L]).
•The mean LDL cholesterol concentration was the same in both groups.
At five years, the combined primary endpoint of cardiac death and nonfatal myocardial infarction occurred less often in the gemfibrozil-treated group (17.3 versus 21.7 percent for placebo). The beneficial effect did not become apparent until two years after randomization.
Multivariable analysis of the VA-HIT trial found that the reduction in nonfatal myocardial infarction and CHD death was strongly correlated with the serum HDL cholesterol concentration achieved with gemfibrozil therapy and was independent of changes in LDL cholesterol or triglycerides [95]. However, a more important predictor of cardiovascular disease (CVD) events in VA-HIT was high concentrations of total HDL particles (in particular, small HDL particle levels) [96].
●Trials of simvastatin plus niacin – One study suggested that additional benefits may be observed in patients with low HDL cholesterol by combining a statin (which not only lowers LDL cholesterol but may have additional cardiovascular benefits beyond lipid lowering) with a drug that increases HDL cholesterol. This three-year trial included 160 patients with clinical and angiographic evidence of CHD who had an HDL cholesterol less than 35 mg/dL (0.9 mmol/L) and an LDL cholesterol less than 145 mg/dL (3.75 mmol/L) [97]. Patients were randomly assigned to one of four regimens: simvastatin plus niacin, antioxidants, simvastatin plus niacin plus antioxidants, or placebo. The mean serum LDL and HDL cholesterol concentrations were unaltered in the antioxidant and placebo groups but changed substantially (-42 and +26 percent, respectively) in the simvastatin plus niacin group. Compared with placebo, patients receiving simvastatin plus niacin were significantly less likely to sustain a cardiovascular event (death, myocardial infarction, stroke, or revascularization), and they experienced angiographic regression (compared with progression for the placebo group) of the most significant coronary stenosis.
The magnitude of the reduction of clinical events with drug therapy (relative risk [RR] 0.1 to 0.4 compared with placebo) in this study was greater than that typically observed with studies of statins alone, suggesting that treatment designed to raise HDL cholesterol provides additional protection beyond that attributable to simply lowering LDL cholesterol. However, caution should be exercised when comparing results of different study populations. In addition, the study group was relatively small and the confidence intervals wide, suggesting that larger studies comparing combination therapies with statins alone may be warranted. Antioxidants did not provide any additional benefits in this study of secondary prevention (and even appeared to attenuate the positive effects of combination drug therapy), which is consistent with most other clinical trial data. (See "Vitamin intake and disease prevention", section on 'Antioxidants'.)
Role of niacin, fibrates, or estrogen — Among medications that alter the lipid profile, nicotinic acid (niacin), gemfibrozil, and estrogen replacement therapy (in postmenopausal women) produce the most pronounced elevations (15 to 30 percent) in HDL cholesterol (table 1). (See "Low-density lipoprotein cholesterol lowering with drugs other than statins and PCSK9 inhibitors", section on 'Fibrates'.)
While there is some evidence that HDL cholesterol can be raised with additional therapies such as niacin or fibrate, there is little evidence that doing so by adding one of these drugs to statin therapy improves cardiovascular outcomes. While niacin can raise HDL cholesterol levels substantially in patients with no other lipid abnormality, there are no data demonstrating that this effect confers a cardiovascular benefit [98-101]. In the AIM High trial of patients with relatively well controlled LDL cholesterol, no benefit was found with niacin in statin-treated patients despite a significant increase in HDL-C [102] (see "Low-density lipoprotein cholesterol lowering with drugs other than statins and PCSK9 inhibitors", section on 'Nicotinic acid (niacin)'). The combination of gemfibrozil and nicotinic acid is more effective than monotherapy, raising HDL cholesterol by as much as 45 percent [101].
Monotherapy with fibrate therapy (gemfibrozil, fenofibrate) can be used initially in patients with hypertriglyceridemia and low HDL cholesterol levels if LDL cholesterol levels are normal or minimally elevated (100 to 130 mg/dL [2.6 to 3.4 mmol/L]). After the hypertriglyceridemia is reversed, residual elevations of LDL cholesterol should be addressed with a statin or, in statin-intolerant patients, with ezetimibe or a bile acid sequestrant.
Investigational therapies
Infusion of apolipoprotein A-I — There are multiple studies that have evaluated the potential benefit of an infusion of apolipoprotein A-1.
●In patients treated with statins to lower LDL cholesterol, apolipoprotein A-1 levels are strongly associated with a reduced risk of CVD events [103]. The intravenous administration of apolipoprotein A-1 Milano to cholesterol-fed rabbits significantly reduced intimal thickening and macrophage content after balloon injury to the femoral and iliac arteries [43] (see 'HDL particles and CVD' above). Based upon this and related observations, a pilot trial of intravenous therapy with recombinant apolipoprotein A-1 Milano phospholipid complexes (ETC-216) was conducted in 57 patients who were within two weeks of onset of an acute coronary syndrome [104]. Subjects were randomly assigned to five weekly infusions of ETC-216 at 15 or 45 mg/kg or to placebo. Patients treated with ETC-216 (combining both dosing groups) had a significant decrease in the mean percentage of coronary artery volume occupied by atheroma (from 38.96 to 37.91 percent), while there was no significant change for the placebo group (from 34.80 to 34.94 percent).
●In a study of 1258 patients with recent myocardial infarction who were randomly assigned to an infusible, plasma-derived apolipoprotein A-1 (CSL112) for four consecutive weekly infusions or placebo, CSL112 was associated with increases in apolipoprotein A-1 complex and cholesterol efflux capacity [59].
Infusion of reconstituted HDL — The ERASE trial randomly assigned 183 patients with CHD to four weekly infusions of either placebo or one of two doses of reconstituted human HDL [105]. The primary endpoint of the percentage change from baseline in coronary atheroma volume, as measured with intravascular ultrasound two to three weeks after the last infusion, was not statistically different among the three groups. The higher of the two HDL doses was associated with a high incidence of liver function test abnormalities, which led to early study discontinuation in this group.
Theobromine — Cocoa intake has been associated with a decreased risk of CVD in observational studies [106]. Theobromine, as found in cocoa, has been associated with an increase in HDL cholesterol [107]. In one small study, intake of 1000 mg of theobromine for four weeks led to a significant increase in HDL cholesterol (increase of HDL cholesterol by 6.2 mg/dL [0.16 mmol/L]; p<0.0001, compared with baseline) [108].
CETP inhibition — Torcetrapib, anacetrapib, evacetrapib, dalcetrapib, and TA-8995 inhibit cholesteryl ester transfer protein (CETP) and significantly raise HDL cholesterol levels [109-111]. Investigation of torcetrapib was stopped due to the finding of an increased risk of cardiovascular events in the ILLUMINATE trial; dal-OUTCOMES, the major trial of dalcetrapib, was terminated early for futility; and evacetrapib was terminated for futility in the ACCELERATE trial. Anacetrapib is under active investigation [112-114]. (See 'CETP and CVD risk' above.)
●Torcetrapib – In the randomized ILLUMINATE trial of over 15,000 patients treated with atorvastatin, a significant increase in HDL cholesterol (72 percent) and a decline in LDL cholesterol (25 percent) were seen after 12 months of torcetrapib therapy [115]. However, the trial was stopped early because, at a mean follow-up of 550 days, torcetrapib therapy was associated with a significant increase in the risk of cardiovascular events (6.2 versus 5.0 percent with atorvastatin alone, hazard ratio [HR] 1.25, 95% CI 1.09-1.44) and death from any cause (1.2 versus 0.8 percent, HR 1.58, 95% CI 1.14-2.19, respectively). The increase in mortality was due to both cardiovascular and non-cardiovascular causes (primarily cancer and infection).
●Anacetrapib – Anacetrapib was shown in two clinical trials to significantly lower LDL cholesterol by about 40 percent and raise HDL cholesterol by over 130 percent without causing a significant increase in systolic blood pressure, as was seen with torcetrapib [116,117]. In the DEFINE trial, 1623 patients with CHD or at high risk and on statin therapy were randomly assigned to either anacetrapib 100 mg or placebo daily [117]. By 24 weeks, anacetrapib, compared with placebo, was associated with a significantly lower level of LDL cholesterol (45 versus 77 mg/dL [1.2 versus 2.0 mmol/L]) and a significantly higher level of HDL cholesterol (101 versus 46 mg/dL [2.6 versus 1.2 mmol/L]). By 76 weeks, there were no significant changes in blood pressure or electrolyte or aldosterone levels compared with placebo. The REALIZE trial, which evaluated anacetrapib in patients with heterozygous familial hypercholesterolemia, is discussed separately. (See "Treatment of drug-resistant hypercholesterolemia", section on 'Potential future approaches'.)
In the phase 3 REVEAL trial, 30,449 adults with atherosclerotic vascular disease who were receiving intensive atorvastatin therapy (see "Management of low density lipoprotein cholesterol (LDL-C) in the secondary prevention of cardiovascular disease", section on 'Intensity of statin therapy') and who had a mean LDL cholesterol of 61 mg/dL (1.58 mmol/L) and a mean HDL cholesterol level of 40 mg/dL (1.03 mmol/L) were randomly assigned to receive 100 mg of anacetrapib once daily or placebo [118]. During the median follow-up of 4.1 years, the primary endpoint of the first major coronary event (a composite of coronary death, myocardial infarction, or coronary revascularization) occurred less often in the anacetrapib group (10.8 versus 11.8 percent; rate ratio 0.91, 95% CI 0.85-0.97). The magnitude and direction of change in LDL cholesterol and HDL cholesterol were consistent with prior trials and there were no important differences in the risks of serious adverse outcomes. We agree with the authors’ conclusion that most of the improvement in clinical outcome can be explained by the additional LDL cholesterol lowering of 26 mg/dL.
●Dalcetrapib – Dalcetrapib has been investigated in several dose-ranging studies [119-123]. In May of 2012, all clinical trials of the drug were stopped due to an absence of clinically meaningful efficacy at interim analysis.
For example, the dal-OUTCOMES trial randomly assigned 15,871 patients with a recent acute coronary syndrome to receive dalcetrapib or placebo in addition to the best available evidence-based care [124]. The trial was terminated early for futility at a prespecified interim analysis when 71 percent of the projected total number of adverse events had occurred. Although dalcetrapib significantly increased HDL cholesterol compared with placebo, there was no significant difference between the two groups (8.3 versus 8.0 percent, respectively) in the risk of the primary composite endpoint of death from CHD, nonfatal myocardial infarction, ischemic stroke, unstable angina, or cardiac arrest with resuscitation.
A genome-wide association study of 5749 patients in the dal-OUTCOMES study identified a region in of the ADCY9 gene polymorphism rs1967309 (p = 2.4 X 10-8), that was associated with atherosclerotic CVD outcomes in the dalcetrapib arm [125]. Patients homozygous for the minor allele (ie, AA genotype) at rs1967309 had a 39 percent reduction in cardiovascular outcomes when treated with dalcetrapib compared with placebo. By contrast, heterozygotes (AG genotype) had a neutral result and patients homozygous for the major allele at rs1967309 (GG genotype) had a 27 percent increase in cardiovascular events with dalcetrapib compared with placebo.
●Evacetrapib – Evacetrapib has been found to not have demonstrable effects on blood pressure or adrenal synthesis of aldosterone or cortisol in preclinical studies [126]. In a study of nearly 400 patients with elevated LDL cholesterol or low HDL cholesterol, patients were randomly assigned to either placebo; evacetrapib monotherapy at 30, 100, or 500 mg daily; or statin therapy (simvastatin 40 mg daily, atorvastatin 20 mg daily, or rosuvastatin 10 mg daily) with or without evacetrapib 100 mg daily [113]. Evacetrapib monotherapy increased HDL cholesterol by 54 to 129 percent and lowered LDL cholesterol by from 14 to 36 percent in dose-dependent manners. In combination with statin therapy, evacetrapib produced increase in HDL cholesterol from 79 to 89 percent compared with placebo and decreases in LDL cholesterol from 11 to 14 percent compared with statin monotherapy. Although the study was not powered to detect adverse clinical effects, none were observed.
In the ACCELERATE trial, 12,092 patients with atherosclerotic cardiovascular disease were planned to be randomly assigned to receive either evacetrapib or placebo in addition to standard medical therapy. After 1363 of the planned 1670 primary clinical endpoints occurred, the trial was terminated due to lack of efficacy [127].
●TA-8995 – TA-8995, a novel inhibitor of CETP, raised HDL cholesterol levels and lowered LDL cholesterol levels significantly without evidence of serious side effects in the TULIP trial, which is discussed elsewhere [114]. (See "Treatment of drug-resistant hypercholesterolemia", section on 'CETP inhibition'.)
Trials of LDL cholesterol lowering — Possible support for the treatment of low HDL cholesterol is provided by inference from clinical trials of LDL-lowering strategies in which patients were further stratified by change in HDL [128,129]. The importance of HDL cholesterol was best illustrated in an analysis from the Air Force/Texas Coronary Atherosclerosis Prevention Trial (AFCAPS/TexCAPS) [130]. In 6605 adults with low HDL cholesterol levels (<45 mg/dL in men and <50 mg/dL in women), the risk of an initial event was more strongly associated with apoB levels (in LDL) than apolipoprotein A-1 levels (in HDL).
In the Heart Protection Study, 20,536 patients with coronary disease, other occlusive arterial disease, or diabetes and baseline plasma total cholesterol level of at least 135 mg/dL (3.5 mmol/L) were treated with a lipid-lowering diet and then randomly assigned to therapy with placebo or simvastatin (40 mg/day) [131]. The vascular event rate in patients in the placebo arm with baseline HDL cholesterol concentrations below 35 mg/dL (0.9 mmol/L) was 29.9 percent, whereas patients with HDL cholesterol concentrations of 42 mg/dL (1.1 mmol/L) or higher had an event rate of 20.9 percent. The relative reduction in risk for patients treated with simvastatin was similar in patients with low and high baseline HDL cholesterol concentrations, resulting in a higher absolute risk reduction in patients with low baseline HDL cholesterol concentrations (7.3 versus 3.9 percent absolute reduction). Although the authors did not discuss whether this difference was statistically significant, the greater reduction in absolute risk in patients with low HDL cholesterol concentrations is consistent with what was seen in other studies.
A post hoc subset analysis of 2073 patients in the LIPID study with a recent myocardial infarction or unstable angina who had a low LDL cholesterol (140 mg/dL [3.6 mmol/L] or less) and low HDL cholesterol (40 mg/dL [1.0 mmol/L] or less) who also had a triglyceride level of 300 mg/dL [3.4 mmol/L] or less found that absolute and relative risk reductions for cardiovascular events and mortality with pravastatin 40 mg daily were similar to those seen in the overall cohort [132]. (See "Management of low density lipoprotein cholesterol (LDL-C) in the secondary prevention of cardiovascular disease".)
OUR APPROACH TO PATIENTS WITH LOW HDL CHOLESTEROL — As there is no firm evidence of benefit from attempts to target low high-density lipoprotein (HDL) cholesterol, we do not do so. We believe the evidence presented above supports the concept that low-density lipoprotein (LDL) cholesterol lowering should be pursued more vigorously when HDL cholesterol is low. (See 'Effect of increasing HDL cholesterol on clinical outcome' above.)
For all patients at increased cardiovascular risk, we recommend regular exercise, smoking cessation [133], attainment of target body weight [134,135], and a healthy diet. These behaviors have been associated with increases in HDL cholesterol. (See "Dietary fat", section on 'Quality of fat' and "Lipid management with diet or dietary supplements" and "Healthy diet in adults", section on 'Mediterranean diet'.)
Some medications such as beta blockers, benzodiazepines, and androgens are associated with a decrease in HDL cholesterol. We do not stop these drugs if they are important to the treatment of other medical conditions [64-66].
PATIENTS WITH HIGH HDL CHOLESTEROL — In patients with high serum high-density lipoprotein (HDL) cholesterol (>60 mg/dL [1.6 mmol/L]), the cause may be hereditary or acquired in conditions such as alcohol abuse, hypothyroidism, phenytoin treatment, and insulin treatment in type 1 diabetes.
High serum HDL cholesterol (>60 mg/dL [1.6 mmol/L]) is most often associated with a lower risk of coronary heart disease (CHD), implying that an elevated low-density lipoprotein (LDL) cholesterol level may be less important in this setting [136]. However, in the absence of evidence to support adopting a less aggressive LDL cholesterol lowering, we do not change our approach to an elevated LDL cholesterol. Furthermore, several studies have shown that high HDL cholesterol levels can be associated with an increased risk of atherosclerosis and cardiovascular events [137]. In these situations, HDL particles are dysfunctional in their antiatherogenic properties [34]. Large HDL particles have a reduced content of antiinflammatory proteins and lipids that may account for their dysfunctional properties.
Insulin treatment in diabetes tends to increase HDL cholesterol so that it may be normal when it would otherwise have been low, or it may be raised, particularly in type 1 diabetes with well insulin-controlled glycemia [138,139]. Atherosclerotic cardiovascular disease (ASCVD) risk models using an algorithm not involving HDL cholesterol have been developed for patients with type 1 diabetes [140].
In one series of patients with elevated HDL cholesterol levels who had CHD, it was found that the HDL particles were functionally impaired with regard to antioxidant and antiinflammatory activities [137]. In the future, we anticipate that HDL biology will extend beyond static measures of concentration to functional measures of the multiple antiatherothrombotic properties [141].
Several issues need to be addressed in premenopausal women with LDL cholesterol ≥160 mg/dL (4.1 mmol/L) and HDL cholesterol ≥60 mg/dL (1.6 mmol/L), including planned pregnancy, regular use of effective contraception, and family history of early-onset cardiovascular disease or type 2 diabetes mellitus. Even though the cardiovascular event rate is very low, the individual patient may express concerns about long-term risk of hypercholesterolemia. In these settings, we provide nutritional counseling and suggest dietary complementary approaches that include diets high in soluble fiber, soy, phytoestrogens, flaxseed, and garlic. If pharmacologic therapy is desired in those patients adherent to a low-fat and cholesterol diet, a nonsystemic agent (bile acid sequestrant) is advisable if pregnancy is a possibility.
RECOMMENDATIONS OF OTHERS — Our recommendation to not specifically target low high-density lipoprotein (HDL) cholesterol with drug therapy is generally in agreement with recommendations from the 2013 American College of Cardiology/American Heart Association, the 2016 European Society of Cardiology/European Atherosclerosis Society, and National Lipid Association (United States) [142-144].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Lipid disorders in adults".)
SUMMARY AND RECOMMENDATIONS
●Relationship of HDL to CVD – Low levels of high-density lipoprotein (HDL) cholesterol are associated with an increased risk of cardiovascular disease (CVD) events and high levels with a decreased risk. However, a causal relationship has not been established. (See 'Low HDL cholesterol as an ASCVD risk factor' above and 'Low HDL cholesterol as a cause of ASCVD events' above.)
●Causes of low HDL – Most causes of low HDL cholesterol are inherited, although there are other considerations. (See 'Causes of low plasma HDL cholesterol' above.)
●Our approach to patients with low HDL
•Exercise, weight loss (in overweight subjects), smoking cessation, and substitution of monounsaturated for saturated fatty acids raise HDL cholesterol. As multiple health benefits are associated with each of these, we recommend their use in patients with low HDL cholesterol. (See 'Our approach to patients with low HDL cholesterol' above.)
•There is no firm evidence of benefit from drug therapy to target low HDL cholesterol and we do not do so. (See 'Effect of increasing HDL cholesterol on clinical outcome' above.)
●Importance of LDL management – Irrespective of HDL cholesterol level, all adults should have a low-density lipoprotein (LDL) cholesterol target established and a plan to reach that target should be in place. (See 'Introduction' above.)