CHAPTER 29: HYPERLIPIDEMIA

The major lipids of plasma, cholesterol and triglyceride, are transported in molecular complexes called lipoproteins. The basic structure of all lipoproteins is similar. They contain a core of neutral lipid (triglyceride and cholesterol ester) that is surrounded by a polar coat containing nonesterified cholesterol, phospholipid, and proteins (called apolipoproteins) (Fig. 29–1). The major categories of lipoproteins consist of low-density lipoprotein (LDL), very-low-density lipoprotein (VLDL), high-density lipoprotein (HDL), and chylomicrons. These lipoproteins vary in size and density (Fig. 29–2). Because lipoproteins can be separated by electrophoresis, they also have been named according to their migration relative to serum proteins. LDL is called beta lipoprotein, VLDL is pre-beta lipoprotein, and HDL is alpha lipoprotein.

The primary function of lipoproteins is to transport triglycerides and cholesterol.^1,2,3,4,5 The major forms of triglyceride-rich lipoproteins (TGRLs) are chylomicrons (Fig. 29–3)^6,7 and VLDLs (Fig. 29–4). Chylomicrons are synthesized in the liver from dietary fat. They are secreted into the circulation for transport of lipids to tissues. In the circulation, chylomicron triglyceride undergoes hydrolysis by an enzyme, lipoprotein lipase (LPL), located on the surface of capillary endothelial cells; fatty acids released during lipolysis are taken up by adipose tissue and muscle. In addition, the liver synthesizes triglyceride from fatty acids and incorporates it into VLDL particles. VLDL triglyceride likewise undergoes lipolysis, releasing fatty acids for adipose tissue and muscle. Following triglyceride lipolysis, an intermediate lipoprotein, called a VLDL remnant, is formed. A portion of VLDL remnants is removed by the low-density lipoprotein receptor (LDLR), but most VLDL is converted through intermediate lipoproteins to LDL. The latter is removed from the circulation by LDLRs that are located on the surface of liver cells.^8,9

Circulating lipoproteins contain multiple proteins, called apolipoproteins.^{10,11,12,13,14,15} The major apolipoprotein of LDL and VLDL is called apo B-100.¹⁶ LDL contains only apo B-100 whereas VLDL also contains apo Cs and apo Es^13,17,18; these apolipoproteins are smaller than apo B-100. LDLRs recognize apo B-100 and, to a lesser extent, apo Es. Apo C-II is required for activation of LPL, whereas apo C-III interferes with LPL activity^13,19 (Fig. 29–5).

Following ingestion of dietary fat, the intestine also forms triglyceride-rich lipoproteins (TGRLPs) called chylomicrons. The triglyceride of chylomicrons follows a metabolic path similar to that of VLDL-triglyceride. Chylomicrons interact with LPL and release fatty acids, which go mainly to adipose tissue for reesterification into triglyceride. The major apolipoprotein of chylomicrons is apo B-48, which is a truncated version of apo B-100.¹⁶ Apo B-48 is not recognized by LDLRs. Removal of most chylomicron-triglyceride results in a chylomicron remnant, which is then removed by the liver. The removal pathway is not fully understood.

An elevated plasma LDL is designated a major risk factor for atherosclerotic cardiovascular disease (ASCVD). Several lines of evidence support a strong association: animal studies, genetic forms of elevated LDL, epidemiological associations, and randomized controlled trials (RCTs) of LDL-lowering therapies.²⁰

The biological mechanisms whereby elevated LDL produces atherosclerosis have been extensively studied.^{4,21,22,23,24,25,26} Circulating LDL particles filter into the arterial intima, where they are retained by interstitial proteoglycans. Through processes not fully understood, LDL undergoes structural modifications that render it susceptible to macrophage attack. Macrophages ingest modified LDL particles and are thereby transformed into lipid-rich foam cells.

Steps in progression of atherosclerosis have been studied in detail.^{27,28,29,30,31,32,33} The major stages of atherogenesis are summarized in Fig. 29–6.³³ Accumulation of large numbers of foam cells gives rise to fatty streaks. Some foam cells die and release their cholesterol esters into the interstitium, and over time the core of extracellular lipid expands. Later, smooth muscle cells from the medium began to produce fibrous connective tissue. This tissue forms a covering of the fatty streak; here the lesion is called a fibrous plaque (fibroatheroma). Continuous filtration of LDL into the arterial wall leads to several steps to plaque progression; after many years, atheromas degenerate into complicated lesions, become unstable, and are prone to rupture.

When rupture occurs, plaque material discharges into the lumen of the artery, producing thrombosis and cardiovascular events.^34,35,36,37 At any stage of this process, LDL-lowering therapy reduces plaque progression and decreases the likelihood of plaque rupture and acute ASCVD events.³⁸

A unique form of LDL is called lipoprotein(a) [Lp(a)].^39,40,41,42 This is an LDL particle in which apo B is covalently linked to another protein named apolipoprotein(a) [apo(a)]. Metabolism of Lp(a) is illustrated in Fig. 29–7.⁴³ The structure of apo(a) is genetically variable. This variability determines the size of the Lp(a) particle. Apo(a) is structurally related to plasminogen. It contains a variable number of “kringle IV” repeats, each of which consists of 114 amino acids. The greater the number of repeats, the lower is the plasma concentration of Lp(a). Considerable evidence indicates that Lp(a) carries atherogenic potential equal to or even greater than LDL.

There is growing evidence that VLDL, like LDL, is atherogenic.^{7,44,45,46,47,48,49,50} For this reason, many investigators favor combining LDL and VLDL into a single fraction of atherogenic lipoproteins. This can be done in either of two ways. First, the cholesterol (C) content of both LDL-C and VLDL-C can be combined; this fraction is named non-HDL cholesterol (non–HDL-C).^51,52,53 Because LDL and VLDL contain apo B-100 (Fig. 29–8),⁵⁴ some investigators favor measurement of total apo B-100 over non–HDL-C.^55,56,57 But there is a high correlation between non–HDL-C and apo B,⁵⁸ so either can be used as a surrogate for atherogenic lipoproteins. Measurement of non–HDL-C has the advantage that it is more readily available in clinical practice and is less expensive than apo B. Nonetheless, apo B may be more informative for treatment decisions such as titration of statins or lowering atherogenic lipoprotein concentrations with combination therapy.

Chylomicrons seemingly are not atherogenic. Native chylomicrons likely are too large to filter into the arterial wall. There is a minimal amount of triglyceride in atherogenic lesions. On the other hand, chylomicron remnants may promote atherosclerosis.^{59,60,61,62,63} Normally, chylomicron remnants are rapidly removed from the circulation; consequently plasma levels are very low. In some circumstances, however, concentrations of chylomicron remnants can be relatively high.⁶⁴ When this occurs, chylomicron remnants may contribute to atherosclerosis.⁶⁵

HDLs do not contain apo B. Instead, HDL particles are held together by apo A-I and apo A-II.⁶⁶ HDLs are considered antiatherogenic (Fig. 29–9),⁶⁷ but under some circumstances HDL apo A-I may be proatherogenic (Fig. 29–10).⁶⁸ It has been conventionally believed that the reason is an absence of apo B; however, HDL have multiple potentially atheroprotective properties. High levels of HDL-C nonetheless are associated with reduced risk of ASCVD.^69,70 Whether an elevation of HDL is actually protective against ASCVD or instead is a marker for reduced risk is uncertain.⁷¹ Although the epidemiologic evidence for an inverse association is strong, to date no convincing RCTs have demonstrated that raising HDL-C levels prevents ASCVD events. Nonetheless, the search for efficacious HDL-raising drugs continues to be active.⁷² If HDL is truly antiatherogenic, therapies that enhance these atheroprotective pathways are being explored. Various possibilities have been proposed and continue to be investigated.^68,73

The two main categories of hyperlipidemia are hypercholesterolemia and hypertriglyceridemia. Neither has a precise definition. In the past, they were defined according to their gaussian distribution. This approach no longer holds, because of their graded associations with atherosclerotic disease. The National Cholesterol Education Program (NCEP),²⁰ identified multiple levels of LDL-C based on their association with ASCVD; a modification of this classification, including values for non–HDL-C, is shown in Table 29–1.

Hypertriglyceridemia likewise is subdivided into normal, borderline high, high, and very high (Table 29–2).²⁰

Borderline-high levels are commonly associated with obesity or diabetes. High levels are often found in persons with a genetic predisposition to hypertriglyceridemia. Most people with a high triglyceride concentration have elevations of VLDL-triglyceride; less commonly they have a condition called dysbetalipoproteinemia, which signifies an excess of cholesterol-enriched VLDL remnants.¹⁰ When fasting triglycerides occur in the very high range, chylomicrons typically are present.¹⁰ Excessively high levels of chylomicrons signify an increased risk of acute pancreatitis.

The term combined hyperlipidemia indicates an increased concentration of both cholesterol and triglycerides. It can be present under several circumstances (eg, combined elevations of LDL and VLDL, dysbetalipoproteinemia, or very high triglycerides).

Genetic Hypercholesterolemias

Several genetic defects have been identified to underlie hypercholesterolemia. In most cases, these are associated with high levels of LDL-C (Table 29–3).⁷⁴ They can be briefly summarized.

LDL receptor deficiency, or LDLR deficiency, is an autosomal dominant genetic disorder accompanied by very high levels of LDL-C and premature atherosclerotic disease.^75,76,77 This was the first genetic cause of hypercholesterolemia identified and typically is called familial hypercholesterolemia (FH). Most people carry mutations in the LDLR gene, which is responsible for encoding the LDLR protein. This protein is located mainly on the surface of liver cells, which removes LDL from the circulation. It does this by binding to apo B-100 molecules of LDL particles. LDLR deficiency can occur in two forms: heterozygous or homozygous. In the heterozygous form, one allele of the LDLR gene is defective or absent. A large number of different mutations have been found in the LDLR gene, many of which produce very high LDL-C levels. When these mutations are present in heterozygotes, only half the normal number of LDLRs is expressed and LDL-C levels are approximately doubled. In the homozygous form of LDLR deficiency, both alleles are absent or defective, and LDL-C levels are approximately increased by a factor of four. Classic estimates were that heterozygous FH occurs in about one in 500 persons; homozygous FH occurs only in one in a million persons. In some populations, the prevalence of both forms can be higher. Patients with heterozygous FH commonly develop premature ASCVD, most often between the ages of 30 to 60 years, and estimates are that it accounts for at least 2% of premature myocardial infarctions (MIs). Homozygous FH leads to very premature disease, often in the teens or earlier. Another clinical feature of FH is the presence of cholesterol deposition in tendons (xanthomas) of the hands, elbows, knees, and feet, especially the Achilles tendon. Deposition of cholesterol in the skin about the eyes is called xanthelasma; in individuals with hypercholesterolemia, this finding suggests but is not pathognomonic of FH. The same is true for corneal arcus.

Familial defective apo B-100 (FDB) is a genetic defect in apo B-100 that causes moderate to severe hypercholesterolemia.^78,79,80 It results from a mutation in a single amino acid (Arg-Gln at position 3500 of the apo B molecule); this mutation reduces the ability of LDL to bind to the LDLR. It is relatively common (1/500 to 1/1000) in central Europe but much less common in other populations. FDB is accompanied by increased risk of ASCVD, but apparently less so than with LDLR deficiency.

Autosomal recessive hypercholesterolemia (ARH) is a recessive disorder characterized by very high levels of LDL-C, xanthomatosis, and premature ASCVD.^81,82 It results from mutations in the ARH protein’s phosphotyrosine binding domain, needed for internalization of LDL into the liver. ARH protein functions as an adapter that couples the LDL receptor to the endoplasmic reticulum. A defect in the protein impairs the normal function of the LDL receptor and results in severe hypercholesterolemia.

Sitosterolemia is a rare autosomal recessive disorder that manifests hypercholesterolemia, tuberous and tendon xanthomas, and in some patients, premature ASCVD.^83,84 Metabolic abnormalities include hyperabsorption of cholesterol and sitosterol and decreased biliary secretion of these sterols. Severe hypercholesterolemia commonly occurs in affected children and is called pseudo–homozygous familial FH. As children grow older, serum cholesterol levels typically return to normal. Plasma sitosterol levels are several-fold elevated. Sitosterolemia results from mutations in two ATP-binding cassette transporter (ABC transporter) proteins that are encoded by ABCG5 and ABCG8 genes.^84,85 In the liver, these transporters pump cholesterol and sitosterol into bile; with mutations, both sterols are retained in the liver and whole body. Retention of sitosterols leads to tissue accumulation, resulting in xanthomas and accelerated atherosclerosis.

Autosomal dominant hypercholesterolemia caused by gain-of-function mutations in the proprotein convertase subtilisin/kinexin 9 (PCSK9) gene is another form of hypercholesterolemia, which has recently been discovered. This is characterized by missense variants in PCSK9. PCSK9 is a circulating protein that promotes degradation of the LDLR (Fig. 29–11).⁷⁴ A gain-of-function mutation in PCSK9 causes autosomal dominant hypercholesterolemia.^86,87 The discovery of this mutation was critical for the development of a new class of drugs for treatment of elevated LDL-C—namely PCSK9 inhibitors.

Genetic Forms of Hypertriglyceridemia

Genetic abnormalities affecting all types of TGRLP have been identified. These include VLDL, VLDL remnants, chylomicrons, and their combinations. These various hypertriglyceridemias sometimes clinically manifest only when associated with conditions that lead to VLDL overproduction (ie, obesity and diabetes). Each category will be reviewed.

Familial hypertriglyceridemia is defined by the presence of elevated triglyceride levels in a single family.⁸⁸ Available evidence suggests that impaired catabolism of VLDL underlies this condition.^89,90 The most common abnormality found consists of heterozygous variants of LPL.^91,92 Other lipolytic defects also may exist. When such abnormalities occur in obese patients, production of VLDL-TG is enhanced, which worsens hypertriglyceridemia.⁹⁰ Whether familial hypertriglyceridemia is associated with ASCVD has been uncertain. But the most definitive long-term follow-up of families with familial hypertriglyceridemia strongly suggests that risk of ASCVD mortality is in fact increased.⁹³

Familial combined hyperlipidemia is said to be present when different lipoprotein phenotypes (eg, hypercholesterolemia, hypertriglyceridemia, or combined hyperlipidemia) occur in different members of the same family.⁸⁸ To date, no single genetic defect has been identified to account for this condition. Most likely, it is a polygenic disorder.⁹⁴ Familial combined hyperlipidemia is more frequently accompanied by premature ASCVD than is familial hypertriglyceridemia.⁹⁵ Obesity and/or type 2 diabetes are common in familial combined hyperlipidemia; thus, a component of the abnormal lipoprotein patterns likely results from overproduction of apo B–containing lipoproteins by the liver.^96,97 Another cause of apparent overproduction of apo B in this disorder could be concomitant LDLR deficiency.⁹⁸

A disorder closely related to familial combined hyperlipidemia is called hyperapobetalipoproteinemia.^99,100 It is characterized by elevations in apo B, small LDL particles, and frequent ASCVD. Like familial combined hyperlipidemia, where apo B levels are typically increased, the lipid phenotype of hyper–apo B is variable. In some affected families, triglycerides and/or cholesterol are elevated. It is generally believed that the production of apo B is increased,¹⁰¹ although the actual mechanism has not been determined.

Familial dysbetalipoproteinemia denotes accumulation of cholesterol-enriched remnants of VLDL and chylomicrons.^102,103 Patients with this disorder frequently show xanthoma striatum palmare (orange or yellow discoloration of palmar creases) and tuberoeruptive xanthomas over the elbows and knees. They carry increased risk of ASCVD, notably coronary heart disease (CHD) and peripheral arterial disease. Dysbetalipoproteinemia typically results from less-common isoforms of apo E. There are three apo E isoforms—E2, E3, and E4. Frequencies of apo E alleles in the general population are 7% for E2, 79% for E3, and 14% for E4 (http://www.alzgene.org/meta.asp?geneID=83). Persons who are homozygous for E2 (E2:E2) are those most likely to show dysbetalipoproteinemia. The E2 isoform appears to be a poor ligand for the LDLR, leading to accumulation of remnant lipoproteins. There is also impaired conversion of VLDL remnants to LDL, which results in excess cholesterol-enriched remnants. The presence of the E2:E2 pattern does not necessarily lead to hyperlipidemia. Most patients who manifest hyperlipidemia have a concomitant condition that causes overproduction of lipoproteins, such as obesity and/or type 2 diabetes.¹⁰⁴ The presence of approximately equal elevations in plasma cholesterol and triglycerides suggests a diagnosis of dysbetalipoproteinemia. This diagnosis can be confirmed by detection of the E2:E2 genotype through genetic analysis.

Familial chylomicronemia is characterized by very high triglyceride levels consequent to accumulation of chylomicrons in plasma. It results from a defect in lipolysis of chylomicron triglycerides. The most common cause is a genetic deficiency in LPL.¹⁰⁵ Other causes of defective lipolysis are deficiencies in apo C-II^106,107 and mutations in GPIHBP1.¹⁰⁸ Apo C-II is required for activation of LPL. GPIHBP1 is a glycosylphosphatidylinositol-anchored protein of capillary endothelial cells; it binds LPL in the subendothelial spaces and shuttles it to the capillary lumen, where it can hydrolyze the triglyceride of chylomicrons. In patients with any of these genetic disorders, high-fat diets produce severe chylomicronemia and predispose to acute pancreatitis. Mechanisms underlying development of pancreatitis are unknown. Other clinical features of the chylomicronemia syndrome include eruptive xanthomas, lipemia retinalis, hepatosplenomegaly, and cognitive impairment.¹⁰⁹

Familial type V hyperlipoproteinemia is characterized by very high triglycerides in both VLDL and chylomicrons and present in multiple family members.^{90,110,111,112} This lipoprotein pattern may represent a severe form of familial hypertriglyceridemia.⁹⁰ Many affected patients are obese or have type 2 diabetes. In them, overproduction of VLDL-TG and defective clearance of TGRLP are reported.^90,113 This dual defect in metabolism of TGRLP likely accounts for elevations in both VLDL and chylomicrons. The defect in catabolism of TGRLP may be caused by a variety of catabolic abnormalities.¹¹⁴

Familial hypoalphalipoproteinemia is characterized by several genetic causes of low HDL-C. These include apo A1 deficiency, apo A1 structural mutations, Tangier disease (homozygous deficiency of adenosine triphosphate (ATP)-binding cassette transporter A1 (ABCA1), and lecithin-cholesterol acyl transferase (LCAT).^115,116,117 Apo A1 deficiency can result from mutations in the apo A1 gene.^116,118 Affected patients show very low HDL-C levels and often have corneal opacifications, planar xanthomas and, sometimes, premature atherosclerosis. One notable structural variant of apo A1, among others, is called apo A-I Milano¹¹⁹; most affected individuals are heterozygous for this mutation. They nonetheless usually have very low HDL-C levels. Even so, they do not manifest premature ASCVD. Tangier disease is a genetic disorder having near absence of HDL-C, reduced LDL-C, and mild elevations of triglyceride.^115,116,120 Affected patients are characterized by enlarged, yellow/orange tonsils, hepatosplenomegaly, and peripheral neuropathy. Some patients have premature ASCVD. The molecular defect in Tangier disease is a mutation in the ABC1 gene. ABC1 on the surface of cells appears to initiate the first step of reverse cholesterol transport; its deficiency reduces the availability of cholesterol for formation of HDL. Premature atherosclerosis has been observed in some affected patients. Genetic LCAT deficiency is accompanied by extremely reduced HDL-C concentrations.^121,122,123 LCAT is required for conversion of nonesterified cholesterol to cholesterol esters. In the absence of cholesterol esters, normal HDL particles cannot form. Persons with LCAT deficiency have corneal opacification, anemia, and progressive renal disease.

Abetalipoproteinemia is a rare disease characterized by an absence of lipoproteins containing apo B-48 and apo B-100.^124,125 It is an autosomal recessive disorder caused by a mutation in microsomal triglyceride transfer protein (MTP).^126,127 The absence of MTP prevents the incorporation of triglyceride into apo B-containing lipoproteins, thereby preventing their secretion in plasma. There is a fat malabsorption caused by an inability to form chylomicrons. A parallel defect in the liver prevents the formation of VLDL, which eliminates both VLDL and LDL from the circulation. These defects lead to an absence of apo B-48 and apo B-100 in the circulation. Patients with abetalipoproteinemia are subject to serious side effects, notably, steatorrhea and a failure to thrive; fat-soluble vitamin deficiency, including vitamin E, neurological disease (degeneration of the spinocerebellar and dorsal column tracts), retinitis pigmentosa; hepatic steatosis; and cirrhosis.^128,129

Familial hypobetalipoproteinemia is characterized by very low (but not absent) LDL-C levels. A precise lower cut point to define a very low LDL-C is not well defined. Many cases exhibit truncating mutations in apo B-100.^{130,131,132,133,134} However, in the majority of those with LDL-C in the low range, truncations of apo B seemingly are not present; beyond apo B truncations, other causes of very low LDL-C have not been identified.¹³⁵ A potential side effect of hypobetalipoproteinemia is development of a fatty liver; this is secondary to failure to adequately mobilize triglycerides into VLDL.¹³⁶

Genetic Variants Affecting Lipoprotein Pathways and ASCVD

Lipid and lipoprotein levels are influenced by multiple conditions that may influence atherosclerosis risk by mechanisms that are partly unrelated to lipoprotein metabolism. Some examples include diabetes, cigarette smoking, hormonal replacement therapy, and many others. More than 100 variants are associated with changes in various lipid and lipoprotein parameters. The mechanisms of many other lipid loci are not fully understood. Recent efforts have evaluated polymorphisms in specific lipoprotein pathways that are associated with ASCVD. Genetic risk scores of these variants support the concept that hyperlipidemia is often polygenic or polygenic with environmental influence. Mendelian randomization has been used to infer causality of a genome-wide association trait involved in lipoprotein metabolism with ASCVD.¹³⁷ This approach has been used in the identification of established and emerging pathways of interest, to stratify responders and nonresponders to lipid-modifying therapy, to explore the genetic basis for adverse reactions and drug elimination pathways, and to provide support for the investigation of pharmacological inhibitors of a selected pathway. The phenotypic consequences of DNA sequence variation in genes targeted by pharmacological agents is summarized in Table 29–3.^138,139

LDL Receptor (LDLR)

Rare LDLR mutations have been shown to contribute to increased risk of MI in individual families,^140,141 whereas more common variants at nearly 50 loci have been associated with MI risk in the general population.^{142,143,144,145,146} About 2% of individuals with early-onset MI (≤ 50 years in males and ≤ 60 years in females) carry an LDLR mutation.¹⁴⁶ More severe LDLR mutations were associated with higher MI risk in the National Heart, Lung, and Blood Institute’s exome sequencing project. Carriers of rare nonsynonymous mutations had a 4.2-fold increased risk of MI versus non-MI controls, and carriers of null alleles had a higher 13-fold risk of MI. These associations with MI presumably are mediated through elevations of LDL-C.

Although mutations in LDLR have been emphasized as loss-of-function and associated with increased CHD risk, LDLR sequence noncoding variants have been shown to be protective. In an Icelandic population, a splice variant in LDLR (rs72658867-A, c.2140+5G>A) and an intronic variant (rs17248748-T) are associated with lower non–HDL-C levels and lower CAD risk.¹⁴⁷ The splice variant alters the LDLR reading frame and increases mRNA expression of LDLR that was unrelated to increased expression in wild-type LDLR.

HMG-CoA Reductase (HMGR)

Hydroxymethyl-CoA (3-hydroxy-3-methylglutaryl-coenzyme A) reductase (HMGR) is the rate-limiting step in cholesterol biosynthesis, Inhibition of HMGR increases LDLR synthesis. The magnitude of HMG-CoA reduction with statins results in a correlative reduction in cardiovascular events.¹⁴⁸ In an analysis of randomized trials of statin therapy, a genetic risk score based on 27 genetic variants was used to stratify the risk of both incident and recurrent CHD events.¹⁴⁹ The benefit of statin therapy on incident and recurrent coronary events increased across the low (13%), intermediate (29%), and high (48%) genetic risk categories. The high genetic risk participants had larger absolute risk reduction, resulting in a threefold decrease in the number needed to treat to prevent one CHD event in primary prevention trials.

The mendelian randomization principle has been used to investigate the mechanism underlying the risk of type 2 diabetes observed in randomized controlled trials of statins.^150,151 Single nucleotide polymorphisms (SNPs) in the HMGCR rs1723848-G allele and rs12916 were associated with weight gain and incident type 2 diabetes.¹⁵² These data suggest a genetic basis for weight gain and type 2 diabetes in 2% to 6% of patients treated with statin therapy.

Proprotein Convertase Subtilisin Kinexin 9 (PCSK9)

The genetics of PCSK9 have identified missense and nonsense loss-of-function variants that are associated with lower LDL-C levels and lower rates of MI (see Table 29–3).^139,153 PCSK9 variants were associated with a 35 mg/dL lower LDL-C concentration and a 49% lower CHD risk in African Americans—and a 13 mg/dL lower LDL-C and 18% lower CHD risk in whites. These associations support the concept that lifelong lower levels of LDL-C are accompanied by lower risk of CHD. Thus, every 1 mmol/L lower LDL-C level is associated with a 54% lower CHD risk. This risk is twofold greater than anticipated from meta-analysis of statin-based RCTs¹⁵⁴ and supports the concept that the duration of low LDL-C levels influences long-term risk.

Niemann-Pick C1-Like1 (NPC1L1)

Niemann-Pick C1-like1 (NPC1L1) is expressed in the small intestine, where it functions as a transporter of dietary cholesterol from the intestinal lumen to the enterocytes.^155,156 The Myocardial Infarction Genetics Consortium Investigators identified 15 naturally occurring inactivating mutations that were associated with an average LDL-C level that was 12 mg/dL lower than in noncarriers.¹⁵⁷ Carrier status for an inactivating mutation was associated with a 53% relative reduction in the risk of CHD (carrier frequency 0.04% vs 0.09% in controls). The effects of inhibiting NPC1L1, HMGCR, or both on CHD events were examined in a mendelian randomization study.¹⁵⁸ NPC1L1 polymorphisms were associated with a 2.4 mg/dL lower LDL-C and 4.8% lower risk of CHD, whereas the group with HMRCR polymorphisms had a 2.9 mg/dL lower LDL-C and 5.3% lower risk of CHD. Individuals with NPC1L1 and HMGCR polymorphisms had an additive 5.8 mg/dL lower LDL-C and larger 10.8% lower risk of CHD. This 2 × 2 genetic analysis provides support for NPC1LI as a valid target for pharmacological intervention in combination with statin therapy, which was recently supported by the IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial).¹⁵⁹

Cholesteryl Ester Transfer Protein (CETP)

Cholesteryl ester transfer protein (CETP) variants located near or in the CETP gene at 16q13 were associated with higher levels of HDL-C and lower risk of incident MI.^160,161 SNP rs708272 in the CETP gene was associated with a per-allele increase in HDL-C levels of 3.1 mg/dL and a corresponding 24% lower risk of future MI (age-adjusted hazard ratio [HR] 0.76; 95% confidence interval [CI] 0.62-0.94). Concordant effects on HDL-C and incident MI were also observed at the CETP locus for rs4329913 and rs7202364. In a meta-analysis, the CETP polymorphism was associated with lower LDL-C levels of 1.2 mg/dL. Based on these genetic studies, it would be anticipated that CETP is a valid therapeutic target; however, pharmacological inhibition with multiple agents has not reduced CHD risk.^162,163 Several explanations for the failed trials include off-target toxicity with torcetrapib and lack of LDL-C lowering with dalcetrapib. In contrast, anacetrapib lowered LDL cholesterol by about 28% in ACCELERATE (A Study of Evacetrapib in High-Risk Vascular Disease), but this trial was terminated after 2 years based on the futility analysis (https://investor.lilly.com/releasedetail.cfm?ReleaseID).

A pharmacogenomics approach has been proposed for the identification of individuals who derive benefit versus harm from CETP inhibition. In the dal-Outcomes trial, polymorphisms in the ADCY9 gene on chromosome 16 were associated with reductions in cardiovascular events.¹⁶⁴ Among patients with genotype AA at rs1967309, there was a 39% reduction in the composite cardiovascular end point with dalcetrapib compared with placebo. In contrast, patients with genotype GG had a 27% increase in events with dalcetrapib versus placebo. Surprisingly, the GG genotype was associated with slightly lower LDL-C levels. These data have been replicated in dal-PLAQUE with carotid intima-medial thickness. This pharmacogenomic approach to CETP inhibition has been suggested as strategy to investigate whether dalcetrapib may prove an effective target for the prevention of atherosclerotic cardiovascular events among individuals with the certain variants in ADCY9. The mechanism for the participation of ADCY9 in dalcetrapib-associated cardiovascular risk remains uncertain.

Triglyceride-Rich Lipoprotein (TGRL)

The genetic architecture for triglycerides in the population is comprised of rare large-effect variants, common small-effect variants, and environmental influences.^7,165 Variants associated with common, complex traits such as plasma triglyceride range from common (> 1:20), to low (1:200 to 1:20), to rare (< 1:200). Approximately 50% of interindividual plasma triglyceride variability is estimated to come from DNA sequence variants (Table 29–4).

Elevated triglyceride levels have been associated with increased risk of cardiovascular events; however, the causality of TGRL has been challenging because of the associations with other major cardiovascular risk markers. As will be discussed in the following sections, mendelian randomization studies provide evidence that mutations in apo V5, LPL, and apo C-III are associated with increased risk of MI.

Apolipoprotein A5 (APOA5)

Apo A5 is involved in multiple stages of TGRL metabolism, including VLDL production, TGRL remnant clearance, and regulation of LPL activity.¹⁶⁶ Genome-wide association studies have shown that polymorphisms in or near the apo A5 locus are associated with nonfasting TG levels. APOA5 genetic variants associate with incident CHD that is in direct relation to the increase in elevated nonfasting TG and remnant cholesterol.^167,168,169 Carriers of rare nonsynonymous mutations in APOA5 had higher plasma triglyceride levels and were at 2.2-fold increased risk of MI.¹⁴⁶

Apolipoprotein C3 (APOC3)

Two mendelian randomization studies have established a central role for apo C3 in TGRL and CHD risk.^170,171 APOC3 variants are strongly associated with triglyceride levels. Approximately 1 in 150 individuals carried any of four protein-altering or splice-site variants of APOC3.¹⁷⁰ Heterozygous carriers of any of these four APOC3 mutations had 39% lower plasma triglyceride levels, 46% lower circulating plasma APOC3, and 40% lower risk of CHD. In two general population studies from Denmark that included data from 75,725 participants, participants with nonfasting triglyceride levels of less than 90 mg/dL had a 57% lower incidence of cardiovascular disease than those with levels of 350 mg/dL or more.¹⁷¹ Loss-of-function mutations in one allele for APOC3 was associated with lower triglyceride levels and lower ischemic vascular disease risk that were similar to the first study (44% nonfasting triglycerides and 41% for ischemic vascular disease).

Adipokine Angiopoietin-Like 3 and 4 (ANGPTL3)

Angiopoietin-like protein 3 (ANGPTL3) renders LPL and endothelial lipase more susceptible to proteolytic inactivation by proprotein convertases. In Angptl3-knockout mice, circulating triglycerides and HDL-C are elevated.^172,173,174

Individuals with distinct nonsense mutations in ANGPTL3 exhibit low plasma levels of LDL-C, triglycerides, and HDL-C or a phenotype referred to as familial combined hypolipidemia (FHBL2).¹⁷⁵ ANGPTL3 polymorphisms are most strongly associated with triglyceride level (polymorphism rs2131925, change of 4.9 mg/dL per allele) but also LDL-C (rs2131925, change of 1.6 mg/dL). In a study of 115 FHBL2 individuals carrying 13 different mutations in the ANGPTL3 gene (14 homozygotes, 8 compound heterozygotes, and 93 heterozygotes) and 402 controls, carriers of two mutant alleles had nondetectable plasma levels of ANGPTL3 protein, whereas heterozygotes exhibited a reduction ranging from 34% to 88% depending on the genotype.¹⁷⁶

ANGPTL4 is a secreted protein induced in adipose tissue that inactivates LPL in the subendothelial space.¹⁷⁷ Loss-of-function variants in ANGPTL4 (E40K, T266M) are associated with increased LPL activity and low triglycerides.¹⁷⁸ In the Atherosclerosis Risk in Communities Study (ARIC), carriers of one or two copies of the 40K variant had age- and sex- adjusted lower levels of triglycerides (–19 mg/dL), LDL-C (–5 mg/dL) and higher HDL-C (+4 mg/dL). Carriers also had a 37% adjusted reduced risk of CHD.

Diabetes

Patients with type 2 diabetes commonly manifest atherogenic dyslipidemia (diabetic dyslipidemia).¹⁷⁹ When this occurs, it is essentially part of the metabolic syndrome. Dyslipidemia is usually the result of accumulation of ectopic fat in the liver.¹⁸⁰ The causes of hepatic ectopic fat are attributed to overnutrition, obesity, and insulin resistance.^180,181 Two major responses to ectopic fat are overproduction of VLDL-triglyceride^182,183 and increased expression of hepatic lipase.¹⁸⁴ VLDL particles in type 2 diabetes are particularly enlarged and enriched with triglycerides.¹⁸⁵ Elevated hepatic lipase contributes to the low HDL-C characteristic of diabetic dyslipidemia.^186,187 LDL particles in diabetic dyslipidemia are typically small and dense.¹⁷⁹

Another factor contributing to diabetic dyslipidemia may be a decrease in expression or activity of LPL.^188,189 This possibility accords with early studies showing that patients with type 2 diabetes have a decreased clearance of TGRL.^190,191

Occasionally, persons with type 2 diabetes will present with severe hypertriglyceridemia and chylomicronemia. Most of these have poor glycemic control and reduced circulating insulin. Two factors likely contribute. First, they have an elevation of nonesterified fatty acids (NEFAs) caused by lack of insulin to suppress NEFA release from adipose tissue. This elevation drives an overproduction of VLDL-triglyceride. Lack of insulin further reduces LPL, which reduced clearance of TGRL. It should be noted, however, that poor glycemic control per se usually does not cause chylomicronemia. One study strongly suggests that chylomicronemia can be explained by a combination of insulin deficiency and a familial form of hypertriglyceridemia.¹⁹¹ Insulin deficiency in type 1 diabetes also can result in chylomicronemia when it coexists with a genetic hypertriglyceridemia.

Nephrotic Syndrome

The characteristic lipid abnormality in patients with the nephrotic syndrome is hypercholesterolemia. This abnormality has been attributed to low levels of serum albumin (or low oncotic pressure). The response to hypoalbuminemia was early proposed to be overproduction of apo B–containing lipoproteins by the liver.^192,193,194 On the other hand, elevated LDL-C levels could be secondary to reduced clearance of LDL from the circulation. Isotope kinetic studies have noted both patterns of response in patients with nephrotic syndrome.^195,196,197 The overproduction of LDL can readily be explained by response to reduced oncotic pressure when plasma albumin levels are low. On the other hand, the observed reduction in clearance of LDL suggests downregulation of LDLRs. Another cause of increased LDL-C with nephrotic syndrome appears to be an overloading of LDL particles with cholesterol.¹⁹⁵ This could result from prolonged circulation of LDL particles in plasma. In addition, however, this abnormality may be secondary to increased plasma levels of cholesteryl ester transfer protein, which has been reported in nephrotic syndrome.¹⁹⁸

An elevation of LDL-C is not the only lipid abnormality seen in many patients with nephrotic syndrome. In fact, there can be a concomitant increase in plasma triglycerides. The latter response appears to be secondary to decreased clearance of TGRL.^193,195 A decreased clearance of these lipoproteins may be the result of a loss of apo C-II in the urine.¹⁹⁹ Apo C-II is required for activation of LPL; therefore, its deficiency in VLDL particles could contribute to lipolysis of VLDL triglyceride.²⁰⁰

A persistent question has been whether the hyperlipidemia associated with nephrotic syndrome can promote development of ASCVD. Long-term follow-up of large numbers of nephrotic patients is not available; nonetheless, limited evidence supports the atherogenic potential of nephrotic syndrome.^201,202

Chronic Kidney Disease

Chronic kidney disease (CKD) is a risk factor for ASCVD.²⁰³ Mechanisms are believed to be multifactorial, including various components of the metabolic syndrome (ie, dyslipidemia, prothrombotic state, proinflammatory state). Several patterns of dyslipidemia are reported. The most common is hypertriglyceridemia.²⁰⁴ Evidence supports a defective clearance of triglyceride from the circulation, most likely secondary to defective triglyceride lipolysis.^205,206 A high ratio of apo C-III:apo C-II has been implicated in reduced lipoprotein lipase activity.^205,206,207 Reduced levels of HDL also are common.²⁰⁸ Although elevated LDL-C is typically not present in CKD, the demonstration that LDL-lowering therapy reduces risk for ASCVD implicates LDL as one factor.²⁰⁹ Because of the likelihood of many factors contributing to ASCVD in CKD, the quantitative contribution of dyslipidemia to cardiovascular disease is uncertain.

Hypothyroidism

Hypothyroidism has long been known to raise serum LDL-C levels.²¹⁰ The mechanism for this response is not well understood, but most likely there is a decreased expression of LDLRs.^211,212,213 Clinical experience mandates that all patients with hypercholesterolemia or other dyslipidemia be tested for hypothyroidism. Many hypothyroid patients are asymptomatic and may go undetected if their thyroid-stimulating hormone values are not measured.²¹⁴ Not only can hypothyroidism cause hypercholesterolemia, it can unmask other forms of latent dyslipidemia.^215,216,217 Another reason to test for hypothyroidism is because it predisposes to severe myopathy when statins are used for treatment of hypercholesterolemia.²¹⁸ A long-debated issue is whether hypothyroidism (or subclinical hypothyroidism) predisposes to ASCVD. The data are conflicting, but several reports suggest a greater risk.^{219,220,221,222,223,224}

Human Immunodeficiency Virus

Dyslipidemia is higher among human immunodeficiency virus (HIV)-infected patients than among uninfected persons, occurring in over one-third of patients.²²⁵ Dyslipidemia is a major risk factor for CHD and cerebrovascular events (Data Collection on Adverse Events of Anti-HIV Drugs [DAD] Study Group). Changes over time in risk factors for cardiovascular disease and use of lipid-lowering drugs in HIV-infected individuals impact on MI.^{226,227,228,229,230}

Statins have a number of drug-drug interactions that affect choice of specific statin and dosing in HIV-infected patients.²³¹ A number of these drugs are commonly used in HIV-infected patients, particularly HIV and hepatitis C, and treatment with protease inhibitors and cobicistat.²³² In addition, comorbid conditions that may affect the use and dosing of statins, such as chronic liver and kidney diseases, are more common among persons with HIV.^233,234,235 Thus, HIV-infected patients may be at higher risk for inappropriate statin prescription (contraindicated agent or higher than recommended dose) and toxicity. Despite this, there are scant data regarding patterns of statin use and dosing among HIV-infected patients.

Medications

Androgens

In hypogonadal men, testosterone replacement has little effect on serum lipoproteins.²³⁶ However, there is evidence that testosterone increases the activity of hepatic lipase, which can reduce HDL-C.^237,238,239 This effect could explain the reduction of HDL-C that occurs during puberty in body.^240,241,242 Anabolic steroids more potently raise hepatic lipase activity and cause marked reduction in HDL-C.²⁴³

An important question relates to the role of testosterone in causing the differences in body fat distribution between men and women.²⁴⁴ Men characteristically exhibit accumulation of adipose tissue mainly in the upper body subcutaneous and visceral regions. In women, fat is located predominantly in the lower body. This difference in body fat distribution may contribute to the higher cardiovascular risk in men. In women with polycystic ovarian disease, testosterone levels are elevated, which may account for their male-pattern of body fat distribution.²⁴⁵ On the other hand, low testosterone levels have been found to associate with upper body obesity,²⁴⁶ so the role of testosterone in determining body fat distribution is uncertain.

Estrogens

It is well known that women are at lower risk for ASCVD. Many investigators have speculated that estrogens protect against cardiovascular disease. A considerable body of research indicates that estrogens lower LDL-C and raise HDL-C.²⁴⁷ LDL-C levels commonly rise after the menopause, most likely due to decrease in LDLR activity.^248,249 Occasionally, estrogen replacement can cause severe hypertriglyceridemia.^250,251,252 This may be more likely in women with an underlying hypertriglyceridemia. The mechanisms for this response are not well understood, but most likely represent a defect in triglyceride lipolysis.

Retinoids

It is well known that use of oral retinoids for treatment of skin diseases, such as acne, can induce hypertriglyceridemia.²⁵³ In some cases, severe hypertriglyceridemia can occur and can be accompanied by acute pancreatitis.²⁵⁴

ASCVD rates vary in different countries.²⁵⁵ Difference in diet seemingly contributes importantly to this variability.^{256,257,258,259,260,261,262,263} Besides the composition of the diet, obesity, physical activity levels, and cigarette smoking are risk factors.^264,265 Diet composition, obesity, and physical active influence LDL-C levels and other lipoproteins.

Dietary Lipids

Dietary fats have been extensively studied for their effects on lipoproteins.²⁶⁶ Dietary cholesterol as well as saturated and trans fatty acids are LDL-raising nutrients.²⁰ In populations with high intake of these saturated fatty acids, LDL-C levels are raised by 10% to 15%.^267,268 In contrast, unsaturated fatty acids (monounsaturated and polyunsaturated) do not raise LDL-C levels.^269,270,271 Therefore, these fats can be used in the place of saturated fatty acids.²⁷² High-carbohydrate diets, in contrast, cause mild-to-moderate increases in VLDL and reduce HDL-C levels.^273,274 Because of these effects, monounsaturated fatty acids may be preferable to carbohydrate as a replacement of saturated fatty acids.²⁷³ A diet high in monounsaturated fatty acids is characteristic of the traditional Mediterranean diet.^275,276

Populations with high intakes of saturated fats and cholesterol are at increased risk for ASCVD, compared to those with lower intakes.^277,278,279 Several smaller RCTs indicate that unsaturated fats reduce CHD events when substituted for saturated fats.^280,281,282

Cardioprotective Foods and Food Patterns

Much epidemiologic evidence indicates that certain foods other than dietary fats offer some protection against ASCVD. Among these foods are fruits and vegetables, fish, nuts, seeds, perhaps modest alcohol consumption, and low sodium/high potassium intakes.^{265,283,284,285,286,287,288,289} Also, some natural foods, such as tree nuts and peanuts, legumes, whole grains rich in soluble fiber such as oats and barley, and cocoa products such as chocolate, have been reported to reduce plasma cholesterol.²⁹⁰ Furthermore, high intake of soluble fiber will reduce cholesterol levels.^291,292 Finally, plant sterols/stanols and intake of about 2 g/d will lower serum cholesterol by about 10%.^{293,294,295,296,297} None of these factors has undergone rigorous a RCT, but epidemiologic data are highly suggestive of benefit. In the case of n-3 fatty acids, however, an RCT from Japan found that eicosapentaenoic acid incrementally lowered risk of major coronary events when given in combination with a statin.²⁹⁸ Another impressive RCT examined influence of Mediterranean-type diet on CHD risk.²⁹⁹ This diet was rich in virgin olive oil or mixed nuts; it was reported to reduce risk for ASCVD.²⁹⁹

Obesity, Overnutrition, and Metabolic Syndrome

Obesity in the general population is associated with increased risk of ASCVD and other disorders. Among the latter are hypertensive cardiovascular disease, type 2 diabetes, fatty liver disease, gallstone disease, obstructive sleep apnea, and certain forms of cancer (colon, breast, and endometrial). Excess body weight can be defined as overweight (body mass index [BMI] 25-29 kg/m²) and obesity (BMI ≥ 30 kg/m²). According to the Centers for Disease Control and Prevention (CDC), more than one-third (34.9% or 78.6 million) of US adults are obese. In addition, more than one-third of American adults (35.7%) and 17% (12.5 million) of American children and teens are clinically obese. The CDC notes that non-Hispanic blacks have the highest age-adjusted rates of obesity (47.8%) followed by Hispanics (42.5%), non-Hispanic whites (32.6%), and non-Hispanic Asians (10.8%). Moreover, obesity is higher among middle-aged adults (age 40-59 years; 39.5%) than among younger adults (age 20-39 years; 30.3%) or older adults (≥ age 60 years; 35.4%).

Obesity denotes an excess of adipose tissue, which is a storage tissue for triglycerides. It is a strong indicator of chronic ingestion of greater intakes of nutrient energy (calories) than needed to maintain normal body weight. Ingestion of excess calories is at the core of metabolic abnormalities associated with obesity.¹⁸¹ The primary function of adipose tissue is to safely store excess triglycerides and to protect vital tissues from nutrient overload. In some obese individuals, adipose tissue efficiently serves this function and prevents the metabolic consequences of caloric overload. But in others, adipose tissue becomes overloaded with triglyceride and cannot mount an adequate defense against overnutrition.

Fat (triglycerides) can be stored in either upper body or lower body adipose tissue pools.¹⁸¹ Lower body fat occurs predominantly in subcutaneous adipose tissue in the gluteofemoral region. This appears to be a preferable location for fat storage, because the metabolic consequences of gluteofemoral obesity tend to be minimal.³⁰⁰ It is a common storage site for triglyceride in women. In contrast, men usually store excess fat in the upper body (abdominal obesity). Here fat storage can occur in either subcutaneous or visceral pools. Upper body obesity is accompanied by greater nutrient overload in other issues. The latter is called ectopic fat.¹⁸⁰

Dangers for cardiovascular disease and overnutrition fall under the category of metabolic syndrome. The latter includes the following features: atherogenic dyslipidemia (elevated triglycerides and total apo B; reduced HDL-C; and small, dense LDL particles), dysglycemia (prediabetes or diabetes), elevated blood pressure, a proinflammatory state, and a prothrombotic state.³⁰¹ Most patients, but not all, with the metabolic syndrome exhibit abdominal obesity. Nonobese individuals with metabolic syndrome most likely have genetic susceptibility to metabolic syndrome, and minimal overnutrition elicits the condition. Even in obese persons with the syndrome, susceptibility factors other than triglyceride-storage capacity most certainly contribute development of metabolic risk factors.³⁰²

An essential defect in the metabolic syndrome appears to be the accumulation of ectopic fat in skeletal muscle, liver, and possibly the heart and pancreatic beta cells. In muscle, ectopic fat contributes to insulin resistance and type 2 diabetes.¹⁸⁰ In the liver, it predisposes to nonalcoholic fatty liver as well as to atherogenic dyslipidemia.^180,300,303 In beta cells, ectopic fat may impair insulin secretion.³⁰⁴ And in the heart, it may lead to systolic or diastolic dysfunction.^305,306

A clinical diagnosis of the metabolic syndrome can be made in accord with a clinical definition (Table 29–5).³⁰⁷

The risk factors of this definition include increased blood pressure, dyslipidemia (increased triglycerides and lowered HDL-C), increased fasting glucose, and abdominal obesity (increased waist circumference). This diagnosis harmonized previous recommendations from the International Diabetes Federation and the American Heart Association/National Heart, Lung, and Blood Institute. Any three of five abnormal findings constitute a diagnosis of the metabolic syndrome. A single set of cut points are used for all components except waist circumference, for which different national or regional thresholds for waist increased circumference were recommended.

Several meta-analyses or studies in large cohorts have been carried out to determine risk of CHD or stroke in patients with metabolic syndrome.^{308,309,310,311,312,313} All of these studies reveal that patients with the syndrome are at higher risk for ASCVD events or mortality than are those without the syndrome. Epidemiological studies show that obesity itself is an underlying risk factor for ASCVD,^314,315 and this risk appears to be mediated largely through the metabolic syndrome. In patients with this syndrome, risk for ASCVD events is increased 1.5- to 2-fold, and for total ASCVD mortality, the risk is increased in the range of 1.5-fold.

Physical Inactivity

Epidemiological data indicate that physical activity offers some protection against ASCVD.³¹⁶ Vigorous physical activity can lower triglycerides, raise HDL-C, overall improve the metabolic syndrome, and may protect against ASCVD in other ways.^317,318,319

Cigarette Smoking

Cigarette smoking is a major cause of ASCVD worldwide. All prevention strategies should emphasize smoking prevention or cessation as a component of lifestyle intervention.²⁵⁵ It lowers HDL-C and creates a dysfunctional HDL particle.³²⁰

Statins

Statins inhibit cholesterol synthesis in the liver, which increases LDLRs, and increased LDLR activity promotes removal of LDL and VLDL remnants from the circulation. Depending on the dose and agent chosen, statins can reduce LDL-C levels by 25% to 55%. These powerful cholesterol-lowering drugs reduce risk of ASCVD events in both primary and secondary prevention.^148,321,322 RCTs, carried out up to 5 years, demonstrate risk reductions of 25% to 45%; long-term treatment likely reduces risk even more.³²³ When drugs are indicated for risk reduction, statins are first-line treatment.

Bile Acid Resins

Bile acid resins include cholestyramine, colestipol, and colesevelam. These agents bind bile acids in the intestine. They reduce the return of bile acids to the liver and reduce feedback inhibition on the conversion of cholesterol into bile acids.³²⁴ This change reduces hepatic cholesterol concentrations and stimulates the synthesis of LDLRs. The net result is a reduction of LDL-C levels of 15% to 25%, depending on the dose of drug given. A major RCT revealed that treatment of hypercholesterolemia with cholestyramine reduces the risk of ASCVD.³²⁵ The use of bile acid resins is limited by the fact that they are less efficacious than statins and are often accompanied by constipation and other gastrointestinal side effects. These agents have been shown to enhance LDL lowering when given with statins; however, no large-scale clinical trials have ever tested the efficacy of this combination.

Ezetimibe

Ezetimibe inhibits intestinal absorption of cholesterol and lowers LDL-C by 15% to 25%.³²⁶ It is largely safe and very well tolerated. There are no large monotherapy prevention trials with ezetimibe. However, in one large secondary prevention study, addition of ezetimibe to maximal dose statins incrementally reduced risk of ASCVD events.¹⁵⁹ In another study, the combination of ezetimibe and simvastatin was significantly lowered cardiovascular events in persons with CKD.³²⁷

Monoclonal Antibody Inhibitors of PCSK9

Monoclonal antibody inhibitors of PCSK9 are a new class of LDL-lowering drugs. They bind PCSK9 from the circulation and thereby prevention the action of PCSK9 to promote degradation of LDL receptors.^328,329 As a result, more LDLRs are expressed by the liver, and serum LDL-C levels are reduced.^330,331 US Food and Drug Administration (FDA)–approved PCSK9 inhibitors are evolocumab^330,331 and alirocumab.³³² Another agent (bococizumab) is under investigation. These drugs must be given systemically once or twice a month. They cause marked incremental reductions of LDL-C even when given with statins. To date, they appear to be safe and have not been reported to cause myopathy. Interim reports from RCTs indicate that they enhance risk reduction in high-risk patients treated with statins.^333,334 Meta-analysis of several smaller trials further suggest ASCVD benefit.³³⁵

Mipomersen

Mipomersen is an RNA antisense agent that targets the messenger RNA for synthesis of apo B. It reduces the hepatic secretion of apo B–containing lipoproteins and thus lowers VLDL and LDL.³³⁶ Mipomersen has been approved by the FDA for treatment of homozygous FH. The agent is given weekly by injection. Because it inhibits the secretion of lipoproteins, development of fatty liver is a risk; this appears to be its major side effect. Outcome RCTs have not been carried out.

Lomitapide

Lomitapide is an inhibitor of MTP.^337,338 This protein delivers triglyceride into VLDL particles and reduces their secretion into the circulation. Both VLDL and LDL levels are reduced. High doses markedly lower serum lipoprotein, but this causes retention of triglyceride in the liver (fatty liver). Recent investigations show that low doses of lomitapide produce a substantial reduction of LDL levels with lesser risk of fatty liver.³³⁹ The FDA has approved lomitapide for treatment of homozygous FH.

Fibrates

Fibrates (eg, gemfibrozil and fenofibrate) are primarily triglyceride-lowering agents that also lower VLDL-C. They also mildly reduce LDL-C and raise HDL-C. They are useful in patients with severe hypertriglyceridemia to reduce risk of acute pancreatitis. They also seemingly lower risk of ASCVD, but considerably less than do statins. In a meta-analysis of all fibrates trials, average risk of CHD morbidity was lowered by approximately 10%.³⁴⁰ On the other hand, when meta-analysis was carried out in subgroups of patients with hypertriglyceridemia, CHD was reduced by of approximately 25%.³⁴¹ In addition, in large RCTs, significant reduction in risk was found during monotherapy with gemfibrozil^342,343; they are thus an alternative for patients who are statin intolerant. Although subgroup analysis favors use of fibrates in patients with hypertriglyceridemia, a major RCT is required to provide convincing evidence that the combination is beneficial.

Niacin

Niacin is similar to fibrates in that it reduces triglycerides and increases HDL-C. It also can be more efficacious in lowering LDL-C; reductions in the range of 10% to 15% can be obtained. One monotherapy, secondary prevention RCT showed that niacin reduces CHD events and total mortality.^344,345 Niacin further appears to reduce subclinical atherosclerosis when combined with a statin.^346,347 On the other hand, in large secondary prevention RCTs, combining with maximal statin therapy did not further lower ASCVD events.^348,349 Niacin can cause a variety of side effects—flushing, skin rashes, rises in liver transaminases, gastrointestinal upset, and elevations of plasma glucose; many people cannot tolerate niacin for long periods.³⁵⁰ In a recent clinical trial, niacin plus simvastatin increased risk in the Chinese population.³⁵¹ On the other hand, patients who can tolerate niacin but not statins may find the former useful when combined with ezetimibe.³⁵²

CETP Inhibitors

CETP inhibitors inhibit CETP,³⁵³ which transfers cholesterol ester from HDL-C to VLDL or LDL. As a result, HDL-C levels are increased, whereas VLDL-C and LDL-C are decreased. By raising HDL-C, the so-called good cholesterol, this exchange theoretically could be beneficial. However, to date three RCTs with CETP inhibitors have not shown an ASCVD risk reduction. One inhibitor, torcetrapib, gave an increase in total mortality, and the trial was discontinued.¹⁶² Trials with two other CETP inhibitors, dalcetrapib and evacetrapib, were discontinued before completion because of futility.³⁵⁴ A fourth drug, anacetrapib, continues to be evaluated in clinical trials.³⁵⁵

LDL Apheresis

LDL apheresis is an extracorporeal method for the removal of circulating apo B–containing lipoproteins, including LDL, Lp(a), and VLDL. There are multiple apheresis methods, including dextran sulphate cellulose adsorption, heparin-induced extracorporeal LDL cholesterol precipitation, immunoadsorption, and double filtration plasma pheresis of lipoproteins.³⁵⁶ The procedure is performed weekly or biweekly, depending on the rate at which LDL-C levels return to baseline after therapy. LDL apheresis lowers LDL-C acutely by 50% to 76%.^357,358 However, the time-averaged LDL-C reduction is about 30% after 6 months and 38% after 18 months in the absence of statin therapy.³⁵⁹ Larger LDL-C reductions may be observed in patients with higher baseline lipid levels, who appear to have a relatively greater response to this therapy.³⁶⁰ The most frequent adverse effects include hypotension, anemia, nausea, flushing, and headache. Venous access is necessary, and many patients require construction of an arteriovenous fistula or a central venous port that increases the risk of long-term complications.³⁵⁶

Clinical outcome studies with LDL apheresis are limited, and no study has demonstrated significantly improved survival with LDL apheresis. In an observational study of 130 patients with heterozygous FH who were treated with either LDL apheresis plus statin therapy (n = 43) or statin therapy (n = 87), there was a significantly lower rate of total coronary events (nonfatal MI, percutaneous transluminal coronary angioplasty, coronary artery bypass grafting, and death from CHD) in the LDL apheresis group (10% vs 36%, respectively) at 6 years.³⁶¹

Statin intolerance is the inability to tolerate statins at the recommended dosage to reduce cardiovascular events caused by adverse symptoms, signs or laboratory studies that result in interruption, discontinuation, or dose reduction of the statin therapy.³⁶² The major cause of statin intolerance includes adverse muscle complaints, with far less common reports of cognitive impairment and peripheral neuropathy. A crucial component of assessment of statin intolerance is dechallenge and rechallenge of statin therapy at a reduced dosage or selection of an alternate statin with different pharmacodynamics and pharmacokinetic properties.^363,364 Consensus statements differ with respect to the number of statins that require rechallenge, but this process is essential because 60% to 70% of individuals tolerate reintroduction of a statin on blinded rechallenge.³⁶⁵

Statin-associated Adverse Muscle Complaints

Although the main concern with statins is the risk of rhabdomyolysis, this is a rare occurrence that affects 0.1% of patients.^363,364 Terminology relating to statin-associated adverse muscle events is variable and has changed over time. According to the 2014 National Lipid Association Statin Muscle Safety Task Force:³⁶³

• Myalgia: A symptom of muscle-discomfort, including muscle aches, soreness, stiffness, tenderness, or cramps with or soon after exercise, with a normal creatine kinase (CK) level. Myalgia symptoms can be described as flu-like symptoms.

• Myopathy: Muscle weakness (not due to pain), with or without an elevation in CK level

• Myositis: Muscle inflammation

• Myonecrosis: Elevation in muscle enzymes compared with either baseline CK levels (while not on statin therapy) or the upper limit of normal that has been adjusted for age, race, and sex:

  • Mild: 3- to 10-fold elevation in CK above baseline

  • Moderate: 10- to 50-fold elevation in CK above baseline

  • Severe: 50-fold or greater elevation in CK above baseline or an absolute level of 10,000 U/L or more

• Clinical rhabdomyolysis: Myonecrosis with myoglobinuria or acute renal failure (an increase in serum creatinine of least 0.5 mg/dL [44 μM/L])

Much of the information regarding adverse events with statins is derived from large clinical trials of statin therapy and from observational studies of statins in clinical use.³⁶⁶ However, experience in clinical practice suggests that muscle side effects are more common, resulting in discontinuation of statin therapy.³⁶⁷ Differences in selection criteria in randomized trials from clinical practice may limit the ability to generalize their results to the adverse event profiles seen in a broader population of patients.³⁶⁸

Risk Factors for Statin-Associated Adverse Muscle Events

Statin Characteristics

The ability to cause muscle injury varies among statins. Myositis has been reported in less than 0.5% percent, but the frequency increases at higher doses. With simvastatin, the incidence in clinical trials, in which patients were carefully monitored and potential interacting drugs were prohibited, was 0.02% at 20 mg/day, 0.07% at 40 mg/day, and 0.3% at 80 mg/day.³⁶⁹ In one clinical trial, simvastatin 80 mg/day was accompanied by a 10-fold increase in rhabdomyolysis when compared to simvastatin 20 mg/day.³⁷⁰ Other clinical trials have reported more common adverse muscle events than simvastatin 80 mg/d.^351,371 Thus, simvastatin 80 mg/day is not recommended.

The risk of adverse muscle events appears to be lowest with fluvastatin.³⁷² Pravastatin also has a low risk of myonecrosis, as demonstrated with pravastatin (40 mg/day) in an analysis of more than 112,000 patient-years of experience in three large controlled trials.³⁷³ In addition to this difference in direct toxicity, pravastatin, fluvastatin, and pitavastatin are also less likely to be involved with drug interactions since they are not extensively metabolized by CYP3A4 (Table 29–6).

Preexisting Neuromuscular Disorders

Underlying neuromuscular disorders may become clinically manifest after initiation of statin therapy.³⁶³ In some cases, statins can cause new neuromuscular disorders. However, others involve increased risk of toxicity in patients with known underlying neuromuscular disorders or the development of clinically manifest disease in patients who had likely had preclinical disease before statins were initiated.

Hypothyroidism and Other Disorders

Enhanced susceptibility to statin-associated myopathy occurs in patients with hypothyroidism, acute or chronic renal failure, and obstructive liver disease. Hypothyroidism may predispose to the development of statin-associated myopathy that “unmasks” hypothyroid myopathy. In addition, hypothyroid patients who are pharmacologically euthyroid have higher rates of statin-associated adverse muscle complaints.³⁷⁴

Patient Characteristics

Genetic factors may increase the risk of statin myopathy. A genome-wide association study found that common variants of the SLCO1B1 gene, which encodes the organic anion transporting polypeptide 1B1 (OATP1B1) that mediates hepatic uptake of most statins, substantially increased the risk of adverse muscle complaints and myonecrosis (mild to severe) in patients treated with simvastatin.³⁷¹ A subsequent randomized trial of statins found an association between a specific SLCO1B1 variant (SLCO1B1*5) and an increased risk of myalgias and myonecrosis.³⁷⁵

Certain ethnic populations have an increased risk of statin muscle adverse events. Simvastatin prescribing information recommends caution when doses above 20 mg daily are used in Chinese patients who are also receiving niacin and states that such patients should not be treated with simvastatin 80 mg daily in combination with niacin.³⁵¹ In a large randomized trial in which all patients were treated with simvastatin 40 mg daily, participants from China had a higher rate of adverse muscle symptoms, including myonecrosis, than patients from Europe (1.3 vs 0.4 events per 1000 patients per year).

Other patient factors associated with an increased risk of statin-associated adverse muscle events include advanced age (> 75 years), frailty, female sex, small body frame, acute or decompensated liver disease, and severe renal disease.^363,376

Concurrent Drug Therapy

The increase in susceptibility to myopathy is substantially greater in patients receiving concurrent therapy with a number of drugs, particularly those that inhibit cytochrome P450 3A4 (CYP3A4) (see Table 29–6), gemfibrozil and peroxisome proliferator-activated receptor (PPAR)-alpha agonists.²³¹ Concurrent use of a drug or drug class that is independently considered a risk factor for myopathy (ie, glucocorticoids, cyclosporine, daptomycin, zidovudine) may further increase risk. The uptake, metabolism, and clearance of each statin is different, and therefore each statin is subject to different types of drug interactions (see Table 29–6).

Exercise

Unaccustomed vigorous exercise in those on statins may increase the risk of muscle injury. However, the association has not been consistently observed across trials.^377,378 Clinical judgment is necessary in interpreting elevated CK levels in patients on statins. CK elevations can be related to hypothyroidism or muscle injury during sports (eg, running, hockey, impact or collision sports), and patients who engage in high-impact sports should be advised to have a CK measured before engaging in particularly high-intensity exercise that day. Although there have been concerns about the effects of statin therapy on exercise tolerance even in asymptomatic individuals, no statistically significant differences were reported in muscle strength, muscle endurance, or aerobic performance.³⁷⁹ In contrast to these results, a smaller trial conducted in less fit patients found that treatment with simvastatin diminished exercise-associated improvements in aerobic fitness and muscle mitochondrial content.³⁸⁰

Time Course of Muscle Events

The onset of muscle symptoms is usually within weeks to months after the initiation of statin therapy but may occur at any time during treatment. Myalgias and myopathy resolve and serum CK concentrations return to normal over days to weeks after discontinuation of the drug.

Diagnosis of Statin-Associated Adverse Muscle Events

The diagnosis of symptomatic and more severe statin-associated muscle events with laboratory abnormalities (ie, elevated serum CK) is based on a temporal association for both onset with initiation of statin therapy and resolution with statin withdrawal. Definitions of the categories of statin-associated adverse muscle events are provided in Table 29–7.

However, some patients report muscle symptoms from statin therapy without an elevation in serum CK. Distinguishing among these is typically based on a careful history and trials of cessation and reintroduction of statin therapy. Even with this, in some patients it will be difficult to be certain whether muscle symptoms result from statin therapy. A scoring system has been proposed for statin-associated symptoms,³⁶³ but this tool has not yet been validated (Table 29–8).

Management of Statin-Associated Adverse Muscle Events

Patients with significant muscle toxicity from a statin should discontinue therapy. Based on clinical experience, in the absence of clinical symptoms, a CK level more than 10 times the upper limit of normal that is believed to be caused by a statin is an indication for discontinuing the medication. Patients should drink large quantities of fluids to facilitate renal excretion of myoglobin. If a patient requires a statin and experiences adverse muscle events other than rhabdomyolysis, once symptoms have resolved and the CK has returned to baseline off statin therapy, we suggest the following approach.

If switching statin therapy is unsuccessful, initiate alternate-day (or less frequent) dosing with careful monitoring. Patients with a history of statin-induced rhabdomyolysis should generally not be treated with another statin because of the risk of recurrent myonecrosis and acute kidney injury.³⁸¹

Cognitive Impairment

Statin therapy has been associated with case reports of cognitive impairment. However, prospective analysis of cognitive impairment in controlled clinical trials has been limited. A randomized trial of pravastatin versus placebo did not show an association with cognitive impairment in 5777 white adults 70 years of age or older with existing vascular disease or at high risk for vascular disease because of smoking, hypertension, or diabetes.³⁸² This study was restricted to white participants and limited by the inclusion of a select population of older, high-risk, clinical trial participants. Additionally, incident cognitive impairment was not reported.

Type 2 Diabetes

Statin therapy is associated with a small increased risk of type 2 diabetes mellitus, which is more pronounced among nondiabetes patients treated with high-intensity compared with low- to moderate-intensity statin therapy.¹⁵¹ The mendelian randomization principle was used to investigate the mechanism underlying the glucose-raising effect of statins. Single SNPs in the HMGCR gene, rs17238484 and rs12915, were used to explore the association of incident and prevalent type 2 diabetes. The rs172348484G allele was associated with a 1.62% higher plasma insulin concentration (95% CI 0.53-2.72; P = .04), 0.23% high plasma glucose concentration and 0·30 kg (95% CI: 0·18-0·43) higher body weight. These polymorphisms explained 2% of 6% of the risk of type 2 diabetes.^{150,151,152,383} It is important to note that the decreased risk of ASCVD with statins usually trumps the small increased risk of type 2 diabetes.

LDL-C and Non–HDL-C as Major Targets of Therapy

Most guidelines identify LDL as the primary atherogenic lipoprotein—and therefore the dominant target of cholesterol-lowering therapy.^20,384 But growing evidence indicates that VLDL is similarly atherogenic.^20,385 Many investigators thus contend that combining LDL-C and VLDL-C into non–HDL-C is preferable to LDL-C alone.³⁸⁶ Non–HDL-C can also be called atherogenic cholesterol. The major apolipoprotein of LDL and VLDL is apo B; for this reason, the measurement of total apo B is an alternative to non–HDL-C.⁵⁶ The measurement of total apo B can be converted to the total lipoprotein particle number for apo B–containing lipoproteins; this conversion is possible because there is one apo B molecule per lipoprotein particle. According to several reports, total apo B (or lipoprotein particle number) is a better predictor of ASCVD than is LDL-C.^{320,387,388,389,390,391,392,393,394,395} A few reports claim that apo B is superior to non–HDL-C in prediction of ASCVD,^55,396,397 but other investigators report that non–HDL-C is superior to apo B.^53,398,399 Regardless, the difference in predictive power between the two is small.

Cholesterol-Lowering Therapy in Secondary Prevention

Efficacy of Therapy

Many RCTs document that statin therapy reduces risk of recurrent cardiovascular events in patients with established ASCVD.^{20,322,400,401,402} Risk reduction occurs regardless of baseline LDL-C (or non–HDL-C) levels. Several RCTs report that risk reduction with statin therapy extends to LDL-C ranges of 70 to 80 mg/dL.^{403,404,405,406,407,408,409} Statin treatment reduces strokes as well as CHD, as shown in several RCTs and, particularly, in one trial designed to test effects on stroke.⁴¹⁰ Based on RCT evidence, it is reasonable to maximize statin therapy in patients with established CHD.

A critical question is whether further cholesterol lowering beyond maximal statin therapy will give additional risk reduction. Recent American College of Cardiology (ACC)/American Heart Association (AHA) cholesterol guidelines⁴¹¹ concluded that RCT evidence is insufficient to justify adding other nonstatin drugs to statin therapy. Nonetheless, it may be worthwhile to review their potential. Examples include bile acid resins, ezetimibe, nicotinic acid, fibrates (eg, fenofibrate), and n-3 fatty acids.

Until recently, the only agent systematically tested as an add-on drug was nicotinic acid. As mentioned, adding nicotinic acid to statin therapy failed to achieve additional risk reduction for ASCVD.^348,349 This was true even though previous studies showed that adding nicotinic acid to statins favorably reduced carotid atherosclerosis in imaging studies.³⁴⁷ Even so, in the light of large outcome trials, nicotinic acid cannot be considered a first-line, add-on drug in secondary prevention.

Fenofibate, a fibrate, resembles nicotinic acid in that it lowers VLDL and LDL and raises HDL-C.⁴¹² Among fibrates, fenofibrate is preferred over gemfibrozil when used with statins because it is accompanied by a lower risk of severe myopathy.⁴¹³ Nonetheless, in one recent study in patients with diabetes, the combination of statin plus fenofibrate failed to reduce ASCVD risk more than statin alone.⁴¹⁴ Still, in meta-analysis of results in hypertriglyceridemic patients from this study, combination therapy showed a strong trend toward greater benefit compared to statin alone.⁴¹⁵ This result accords with subgroup meta-analysis of hypertriglyceridemic in all fibrate trials; together, they showed greater risk reduction were used compared to placebo.³⁴¹ A similar favorable trend from fibrate therapy was not found in subgroups without hypertriglyceridemia. In sum, meta-analysis supports the concept that adding fenofibrate to statin therapy is an option for high-risk patients with hypertriglyceridemia.

Many providers add n-3 fatty acids to statin therapy on the belief that additional risk reduction will be achieved. In two secondary prevention trials, adding n-3 fatty acids to statin treatment, however, did not produce additional benefit.^416,417 But in contrast, in the JELIS [Japan Eicosapentaenoic acid (EPA) Lipid Intervention Study] trial,²⁹⁸ adding eicosapentaenoic acid to statins in patients with ASCVD caused a reduction in cardiovascular events. Thus, the benefit of adding n-3 fatty acids to statin therapy remains open to question. Some investigators prefer n-3 fatty acids in combination with statins to treat hypertriglyceridemia because they do not raise the risk of severe myopathy.

Another potential add-on drug is ezetimibe. This agent was combined with a statin in the IMPROVE-IT trial.¹⁵⁹ The combination of maximal simvastatin therapy plus ezetimibe significantly reduced recurrent ASCVD events in patients who had recent acute coronary syndromes compared with simvastatin alone. The addition of ezetimibe reduced LDL-C levels to well below 70 mg/dL, which was the approximate level attained by simvastatin alone. This study makes several important points. First, it undercuts any notion that statins are the only cholesterol-lowering drug that can reduce ASCVD events. Second, it supports the concept of “the lower, the better” for atherogenic cholesterol in patients with established ASCVD. Third, it reinforces the cholesterol hypothesis by showing that a drug that uniquely lowers cholesterol levels will decrease risk.

Hot on the heels of the IMPROVE-IT trial was approval of PCSK9 inhibitors by US and European regulatory agencies. Monoclonal antibodies against PCSK9 free up LDLRs and dramatically reduce serum LDL-C levels.⁴¹⁸ Recent reports^333,334 further document a striking reduction in LDL-C when PCSK9 inhibitors are combined with statins in high-risk patients. They further showed preliminary data in which this combination enhanced reduction of CHD and stroke.

Cholesterol Goals in Secondary Prevention

There are two ways to approach the question of cholesterol goals in secondary prevention. First, the meta-analysis of the Cholesterol Treatment Trialists’ (CTT) Collaboration³²² showed that for every mmol/L lowering of LDL-C there was an approximate 22% reduction in relative risk of ASCVD events (see Fig. 29–11). On average, for every 1% reduction in LDL-C, risk is reduced by 1%. This means that a 50% reduction in LDL-C should reduce risk by 50%. For simplicity, this meta-analysis supported the concept of “the more, the better” for LDL-C reduction. The 2014 ACC/AHA guidelines implicitly accepted this conclusion when it recommended high-intensity statins in patients with ASCVD. For example, treatment with either atorvastatin 80 mg/day or rosuvastatin 20 mg/day should achieve a “more-the-better” goal of 50% reduction of LDL-C. Based on the IMPROVE-IT trial, an even greater reduction of LDL-C (eg, ≥ 60%), attained by adding ezetimibe to a high-intensity statin, could be recommended by the same rationale.

An alternate approach to a cholesterol goal can be based on a meta-analysis of statin trials that examines maximal risk reduction that can be achieved by cholesterol lowering.⁴¹⁹ The results of this analysis showed “the lower, the better” for cholesterol lowering. That is, the lowest levels of LDL-C (or non–HDL-C) were accompanied by the lowest risk of future ASCVD events. This relationship held to an LDL-C of less than 50 mg/dL. Similar results were obtained for non–HDL-C. This “lower-is-better” relationship is supported by the IMPROVE-IT trial. At the least, an LDL-C goal for secondary prevention can be reasonably set at less than 70 mg/dL, but based on IMPROVE-IT, a case can made for a still lower goal. In the future, RCTs with PCSK9 inhibitors may call for a very low goal for atherogenic cholesterol levels.

Cholesterol-Lowering Therapy for Primary Prevention

Optimal Cholesterol Levels in Primary Prevention

There is no simple answer to the question of what is the optimal LDL-C (or non–HDL-C) for primary prevention. Several epidemiological studies indicate that CHD risk is progressively lower as total cholesterol reduces to 150 mg/dL.^420,421 This level corresponds to an LDL-C of about 100 mg/dL.²⁰ Genetic epidemiology further demonstrates that lifetime levels of LDL-C of approximately 100 mg/dL are accompanied by a low lifetime risk of CHD.^153,422,423 Moreover, RCTs with cholesterol-lowering drugs show that reducing LDL-C concentrations to near 100 mg/dL or below produce significant reduction in ASCVD rates.^20,424 Based on these congruent lines of evidence, the Adult Treatment Panel (ATP) III²⁰ defined an LDL-C level of less than 100 mg/dL as being optimal, whereas 100 to 129 mg/dL was called near optimal. On the other hand, this so-called optimal level may not be the most efficacious level for high-risk patients, who may benefit from still further cholesterol lowering.

Selection of Patients for Drug Therapy in Primary Prevention

Three methods have been used for selection of patients for initiation of cholesterol-lowering drugs. These are (1) 10-year risk for ASCVD based on multiple-risk-factor algorithms (global risk assessment), (2) lifetime risk based on single major risk factors, and (3) 10-year risk based on imaging for subclinical atherosclerosis. Each approach can be briefly reviewed for its strengths and weaknesses.

Ten-Year Risk Estimates Based on Multiple-Risk-Factor Algorithms

These algorithms have been developed from various test cohorts. The most widely used in the United States is that derived from a population in Massachusetts, the Framingham Heart Study.⁴²⁵ The Framingham cohort included several thousand subjects studied over many years. Recently, the National Heart, Lung and Blood Institute commissioned the creation of a new algorithm based on five cohorts, including Framingham. This algorithm is currently sponsored by the AHA and the ACC. It was published in 2013 and will be called the 2013 ACC/AHA risk algorithm.⁴²⁶ It can be accessed online (http://tools.acc.org/ASCVD-Risk-Estimator/).

Similarly, QRISK is an ASCVD risk assessment tool developed in the United Kingdom that derives from a large prospective, open-cohort study.⁴²⁷ The derivation cohort for this algorithm included 1.28 million individuals (aged 35-74 years and free of diabetes and ASCVD). The validation study contained 610,000 subjects from 160 practices. A comparison study indicated that the Framingham study overpredicts risk compared to QRISK by about 35%.⁴²⁷ This algorithm is available online (http://qrisk.org/). A major strength of QRISK lies in its much larger size compared to US algorithms.

In essence, multiple–risk factor algorithms use population averages to predict risk in individual patients. When applied in guidelines, they assume that individuals resemble population means. A major limitation of this approach is that risk in individuals can vary substantially from the population mean. This is particularly so in older persons who exhibit wide variability in atherosclerotic burden. Moreover, population risk tends to decline over time, as suggested by QRISK versus Framingham.⁴²⁷ Consequently, population-risk algorithms derived in the past tend to overestimate current risk. This is illustrated by recent reports showing that application of the ACC/AHA algorithm overestimates population risk when tested in more recently studied cohorts.^{428,429,430,431,432,433} A risk algorithm developed in one country does not necessarily hold in another; for example, various studies have shown that Framingham algorithms overestimate CHD risk in several countries.⁴³⁴

Lifetime Risk Assessment Based on Algorithms and on Individual High-Risk Conditions

If lifetime risk estimates were to be reliable, they would have utility for deciding on treatment strategies. Online estimates for lifetime risk are provided by both 2013 ACC/AHA (http://tools.acc.org/ASCVD-Risk-Estimator/) and QRISK (http://www.qrisk.org/lifetime/). These are not as reliable as 10-year risk estimates because long-term risk is not as well substantiated. An alternate and simpler approach is to identify higher risk conditions and to use these as a guide to lifetime risk reduction (Table 29–9).

Any of these conditions carry a relatively high lifetime risk, as can be shown by risk algorithms. The presence of higher risk conditions makes 10-year risk assessment unnecessary. Most such patients deserve serious consideration of statin treatment. Use of this strategy will shift drug administration to an earlier age than will current 10-year risk algorithms. Practical considerations make RCTs testing lifetime therapies unrealistic. But RCTs have been carried out in persons with each high-risk condition, and they show efficacy for shorter term risk.^435,436 RCTs demonstrating efficacy of cholesterol-lowering therapy have been documented in patients with diabetes,⁴³⁷ metabolic syndrome,^438,439,440 CKD,³²⁷ cigarette smoking,⁴⁴¹ hypertension,⁴⁴² and hypercholesterolemia.⁴⁴³ Therefore, lifetime risk reduction is almost certain when used in patients with higher risk conditions.

Ten-Year Risk Assessment Based on Atherosclerosis Imaging

A major drawback of global risk algorithms is that they place excessive weight on age as a risk factor. In essence, age is used as a surrogate for progressive atherosclerosis. The impact of age on estimated risk is so powerful that most men over 60 years and most women over 70 years are recommended for statin therapy.⁴¹¹ Atherosclerotic burden increases with age in the population as a whole; yet many older individuals have little or no atherosclerosis. They will not benefit from taking statins. One way to improve identification of individuals without atherosclerotic burden is to measure subclinical atherosclerosis. The most readily available and intensively studied modality for atherosclerosis imaging is coronary artery calcium (CAC).^444,445 Measurement of CAC in older individuals who are at relatively low risk by global risk assessment will uncover a substantial portion who are devoid of coronary atherosclerosis.^446,447 Recent studies show that actual 10-year risk for clinical CHD when CAC is zero is very low (see Table 29–9).⁴⁴⁸ In the absence of CAC, statin treatment is not needed.⁴⁴⁹

Atherosclerosis imaging may be particularly useful under the following circumstances: absence of high-risk conditions (Table 29–10), estimated 10-year risk for ASCVD in the range of 5% to 19% by global risk assessment,⁴²⁶ presence of emerging risk factors in otherwise lower risk individuals (see Table 29–10), and statin intolerance in persons without ASCVD. As shown in Table 29–10, individuals with a zero CAC Agatston units are virtually free of CHD after 10 years.⁴⁴⁸ These individuals will not need statin therapy. If CAC scores are in the range of 1 to 99 Agatston units, risk is borderline, and statins are optional. The decision should be made on a variety of factors previously considered. At higher CAC scores (> 100 Agatston units), use of statins is reasonable.

Risk Classification for Primary Prevention

The following are classifies risk categories that can be applied to primary prevention.

• High risk. ATP III guidelines categorized patients according to absolute risk. These guidelines pointed out that patients with established ASCVD have a 10-year risk of recurrent events exceeding 20%. Patients who exceeded this risk but had no history of ASCVD were said to have CHD risk equivalents. This nomenclature is no longer used, but identifying those at high risk, whose risk is as high as ASCVD patients and who deserve more intensive cholesterol-lowering therapy, may still be useful.

• Moderately high risk. There is no agreement at present on what constitutes moderately high risk. At this level, risk is high enough to justify cholesterol-lowering drugs, but it is less than that of patients with ASCVD. The ACC/AHA guidelines set a 10-year risk for statin consideration of 7.5% or more. The National Institute of Health and Care Excellence (NICE) guidelines recommend a threshold of ≥ 10%. In neither of the recent guidelines was an upper boundary set, and moderately high risk and high risk were not distinguished. In this document, moderately high risk will be defined as a 10-year risk of ASCVD of 7.5% to 19% and thus will be distinguished from high risk. This distinction may be useful in selection of intensity of statin therapy.

• Higher lifetime risk. At the next lower level, some patients can be said to be at relatively low 10-year risk but a higher lifetime risk (ie, 35%-50%). One way to identify these patients is through a multiple risk factor algorithm—available from the ACC/AHA (http://tools.acc.org/ASCVD-Risk-Estimator/) or QRISK (http://www.qrisk.org/lifetime/). Another way to identify a high lifetime risk in current lower risk individuals is through the presence of higher risk conditions (see Table 29–10). Among these are persistently high LDL-C levels (≥ 160 mg/dL), and especially those with very high LDL-C (≥ 190 mg/dL).

Cholesterol Goals in Primary Prevention

There are two reasons to establish goals for atherogenic cholesterol in primary prevention. The first is to prevent ASCVD events in the short-term (eg, 10 years) and the second is the need to prevent events over a lifetime. Table 29–11 poses a strategy to achieve these aims.

For high-risk individuals, whose 10-year risk is equivalent to that of a patient with established ASCVD (ie, ≥ 20%), cholesterol lowering should be as intense as in ASCVD patients. Accordingly, the goal can be either a 50% reduction in LDL-C levels, or alternatively, a target non–HDL-C of less than 100 mg/dL (LDL-C < 70 mg/dL). For most patients, a high-intensity statin can be justified. If the goals of therapy are not achieved, a nonstatin add-on agent can be considered. If only a moderate-intensity statin can be tolerated, consideration can be given to adding ezetimibe or bile acid resin.

For those at moderately high risk (ie, 10-year risk in the range of 7.5%-19%), the use of a high-intensity statin, as recommended by ACC/AHA guidelines, or a moderate-intensity statin, as proposed by NICE, is an option. At this risk level, atherogenic cholesterol should be reduced by at least 40%, or alternatively, to an LDL-C goal of less than 100 mg/dL (or a non–HDL-C goal of < 130 mg/dL).

For lower risk individuals whose lifetime risk is high (eg, 35%-50%), the LDL-C likewise should be reduced to less than 100 mg/dL. The same goal should be set for patients with was very high LDL-C. In both cases, the aim is to slow progression of atherosclerosis so as to reduce lifetime risk.

Physician-Patient Discussion

The decision to initiate a lifetime of statin therapy should not be taken lightly. For therapy to be successful, the patient becomes a key partner in the treatment regimen. Before starting a statin, several issues should be discussed or considered as options. These can be reviewed briefly.

Adherence to Therapy

Discontinuation of prescribed statin is a common issue in clinical practice. It has been shown that failure to sustain statin therapy results in higher rates of ASCVD after MI. Poor adherence to statins can be attributed to many factors, among which are health system factors, provider behavior, and patient behavior.⁴⁵⁰ The first step in overcoming failure to adhere is to review strategies for maintaining compliance in patient discussion.

Consideration of Drug Side Effects

Most patients tolerate statins and other cholesterol-lowering drugs without significant side effects. As already mentioned, some patients complain of a variety of side effects of statins: statin-associated adverse muscle events, hyperglycemia, cognitive dysfunction, and neuropathy. Statins also can increase plasma glucose, which may cross the threshold for categorical diabetes. This may be alarming, but is not a serious side effect. Unfortunately, many patients have a preconceived anticipation of drug side effects and often blame various symptoms on their drug. If complaints arise, physicians should thoroughly discuss symptoms and make an effort to separate real from imagined statin intolerance. By encouragement, it is often possible to successfully restart and maintain statin therapy. For those who cannot tolerate statins, a variety of strategies, as discussed before, can be used to achieve a satisfactory cholesterol lowering.

Alternatives to Statin or Add-on Drugs

Statins are first-line cholesterol-lowering therapy, but if they are not tolerated, other ways to achieve cholesterol goals can be considered. Efforts should be doubled to start and maintain lifestyle modification, which can lower cholesterol levels by 10% to 15%.²⁰ Both bile acid resins and ezetimibe can lower LDL-C by 15% to 25%. Ezetimibe plus a moderate-intensity statin will reduce cholesterol levels as much as high-intensity statins. Fibrates and niacin are alternative add-on drugs for patients with hypertriglyceridemia. Finally, in patients with severe hypercholesterolemia, PCSK9 inhibitors may be an option for patients who are statin intolerant, especially in those with ASCVD or very high LDL-C levels.

Cost-Effectiveness

With introduction of generic statins, drug costs have declined greatly. For some patients, however, overall cost of therapy is a significant issue. Expenses mount with regular monitoring for response, clinical management of side effects, and time and expense of travel to provider locations. In patients who are at borderline risk, these factors deserve consideration as to whether to institute statin therapy.

Number Needed to Treat

Some physicians find the number need to treat (NNT) to be useful in discussions with patients. NNT is the number of people needed to treat to prevent one ASCVD event over a given period of time. Periods of time may be 5 years (the duration of most RCTs), 10 years (the usual absolute risk projection), or a lifetime. Table 29–11 illustrates the NNT for statin therapy at different 10-year risk estimates for moderate-intensity and high-intensity statins.

The NNT for a lifetime of therapy will be much lower, depending the age at which statin therapy is started. A low lifetime NNT constitutes the strongest rationale for a lifetime of statin therapy in primary prevention.

Emerging Risk Factors

Most patients who have persistent major risk factors deserve consideration for statins (see Table 29–11); they have a relatively high lifetime risk. In addition, some investigators use emerging risk factors in risk assessment (Table 29–12).⁴⁵¹

The most commonly used factor is C-reactive protein.⁴²⁴ Others include low HDL-C, hypertriglyceridemia, family history of premature ASCVD, elevated fibrinogen, prediabetes, and elevations of Lp(a). Although measurements of any or all of these emerging risk factors are not routinely recommended by guideline committees, they are an option for patients whose indications for statins are borderline.

Elderly Patients

Current cholesterol guidelines place high emphasis on use of statins in older persons. This is because risk assessment algorithms give heavy weight to age as a risk factor. Unfortunately, these algorithms become less reliable with advancing age.⁴⁵² This places a greater responsibility on clinicians who care for older persons. One solution to this dilemma is to measure subclinical atherosclerosis. For example, older individuals who are free of CAC have a very low risk for future ASCVD events. On the other hand, if CAC levels are high, the clinician is justified in advocating statin therapy. In addition, if an older person has long-standing, major risk factors, statin therapy is warranted. There is little doubt that high-risk, older persons will benefit from statins treatment; this is been adequately demonstrated in RCTs.^453,454

Statins in Women

The benefit of statin therapy for women having ASCVD is well documented.¹⁵⁴ Their utility for primary prevention has been questioned^{455,456,457,458} (http://tinyurl.com/7ywndqe). At comparable ages, women generally have lower absolute risk than men. Therefore, close attention should be given to absolute risk estimates when contemplating statin usage and women. Some of the factors that favor statin therapy include diabetes, heavy cigarette smoking, multiple risk factors (eg, metabolic syndrome), hypercholesterolemia, and a strong family history of premature ASCVD.