Key Points
- •
When TG levels are between 200 and 800 mg/dL, TG-rich particles are associated with the presence of small dense LDL, low levels of HDL-C, insulin resistance, and metabolic syndrome, all of which increase risk for atherosclerosis.
- •
TG-rich particles affect the size and atherogenic potential of VLDL remnants and LDL particles, both of which enter the arterial subendothelial space and contribute to atherosclerotic plaque.
- •
In a meta-analysis of 68 prospective studies, TG level was not a predictor of risk for nonfatal myocardial infarction and death after adjustment whereas non–HDL-C was associated with increased risk after adjustment for HDL and log TG.
- •
Non–HDL-C predicted risk of CHD and future cardiovascular events better than LDL-C did, probably because non–HDL-C includes all apo B–containing lipoproteins.
- •
Nonfasting TG levels may predict CHD better than fasting levels but are more difficult to standardize and to measure in the clinical setting; non–HDL-C can be measured in the nonfasting state.
- •
NCEP ATP III recommends optimal TG level <150 mg/dL. NCEP recommends calculation of non–HDL-C when TG >200 mg/dL, with the goal being 30 mg/dL higher than the LDL-C goal.
- •
Lifestyle changes—exercise, diet, and weight loss—are necessary to lower TG levels, especially in patients with metabolic syndrome.
- •
Statins, fibrates, niacin, and omega-3 fatty acids can be used to lower TG levels if lifestyle changes are insufficient to reach the goal.
The contribution of triglyceride (TG) and TG-rich lipoproteins to the development of atherosclerosis, especially as an independent predictor of cardiovascular risk, has been debated for many years. TG levels between 200 and 800 mg/dL may be associated with other lipid abnormalities that predispose to atherosclerosis, including low levels of high-density lipoprotein cholesterol (HDL-C), small, dense low-density lipoprotein (LDL) particles, atherogenic TG-rich remnants, and insulin resistance, all of which increase the risk for coronary heart disease (CHD). Therefore, it is difficult to determine the independent contribution of TG and how aggressively one should treat the hypertriglyceridemia. The percentage of adults in the United States with TG levels above 150 mg/dL (1.7 mmol/L), 200 mg/dL (2.3 mmol/L), 500 mg/dL (5.7 mmol/L), and 1000 mg/dL (11.3 mmol/L) is 33%, 18%, 1.7%, and 0.4%, respectively. Therefore, hypertriglyceridemia affects a significant portion of the population. Reliable assessment of the risk associated with lipid fractions is important for the development of accurate screening and treatment strategies.
This chapter reviews the epidemiologic evidence indicating that hypertriglyceridemia contributes to atherosclerosis. Lipid metabolism is then described to demonstrate the mechanisms by which the level of TG-rich lipoproteins affects the composition and size of remnant, LDL, and HDL particles and how this affects development of atherosclerosis and risk for CHD. Finally, the classification and causes of various types of hypertriglyceridemia are discussed, followed by suggestions for evaluation, treatment, and management.
Epidemiologic Evidence Linking Triglyceride Levels with Risk for Coronary Heart Disease
Several epidemiologic studies have provided important findings on the role of TGs and TG-rich lipoproteins in atherosclerosis. In 1959, Albrink and Man observed that TG levels greater than 175 mg/dL were present in 70% of 100 cases of myocardial infarction (MI) compared with only 7% of 92 healthy controls. In the 8-year Prospective Cardiovascular Münster (PROCAM) study reported in 1996, elevated levels of TG were independently associated with incident CHD events after adjustment for LDL-C and HDL-C levels. For this reason, elevated TG levels are in the European model for calculation of cardiovascular risk. An important contribution of PROCAM was the observation that an increasing TG level is directly associated with CHD incidence up to a level of 800 mg/dL. Levels higher than 800 mg/dL are thought not to be associated with CHD because of lipoprotein particles too large to penetrate the vascular endothelium compared with smaller, atherogenic remnant particles found with mild hypertriglyceridemia. In the Framingham Heart Offspring Study, TG level was not associated with CHD after adjustment for HDL-C and other covariates in both men and women. However, a later analysis showed that the cholesterol in remnant lipoproteins is an independent predictor of CHD risk in Framingham women.
In 1996, the first large meta-analysis of 17 prospective, observational studies of TG and CHD events reported that an increase of 89 mg/dL (1 mmol/L) in TG level was associated with a univariate risk of 32% in men and 76% in women and was an independent risk predictor (adjusted for other covariates including HDL-C) of 14% in men and 37% in women.
The Baltimore Coronary Observational Long-Term Study (COLTS), a retrospective cohort study, observed 350 individuals with CHD for up to 18 years after cardiac catheterization. After adjustment for age, gender, use of beta blockers, and other risk factors, baseline fasting TG level >100 mg/dL was associated with a 50% increased risk of subsequent events compared with those with TG <100 mg/dL ( P 0.008) and was an independent predictor of recurrent cardiovascular events in patients with CHD. These results suggested that an optimal TG level is <100 mg/dL in patients with CHD.
The Copenhagen Male Study was a prospective study of 2906 white men without CHD at baseline. During 8-year-follow-up, there was an increased risk for CHD with increasing TG tertiles: 4.65% for TG of 0.4 to 1.1 mmol/L; 7.7% for TG of 1.1 to 1.6 mmol/L; and 11.5% for TG of 1.6 to 2.2 mmol/L. After adjustment for age, body mass index, hypertension, smoking, alcohol, physical activity, diabetes, socioeconomic status, LDL-C, and HDL-C, the middle and highest level of TGs had a relative risk of 1.5 and 2.2, respectively, compared with the lowest tertile of TGs. Moreover, elevated TG and low HDL-C levels were predictive of CHD events in men with LDL-C both less than and greater than 170 mg/dL.
In 2004, a meta-analysis pooled individual data from 96,224 subjects in 26 prospective studies in New Zealand, Australia, and several Asian countries and included 670 and 667 deaths from CHD and stroke, respectively. After adjustment for age, sex, blood pressure, smoking, ratio of total cholesterol (TC) to HDL-C, and major cardiovascular risk factors, compared with those in the bottom fifth of TG levels, individuals in the top fifth of TG levels had a 70% (95% CI, 47-96) greater risk of CHD death, an 80% (95% CI, 49-119) higher risk of fatal or nonfatal CHD, and a 50% (95% CI, 29-76) increased risk of fatal or nonfatal stroke. The association between levels of TG and CHD death was similar across subgroups defined by ethnicity, age, and sex. These results suggested that serum TGs are an independent predictor of CHD and stroke risk in the Asia-Pacific region even after adjustment for HDL-C.
More recently, a meta-analysis of 29 population-based, Western prospective studies included 10,158 CHD cases from 262,525 participants. After correction of risk estimates for long-term within-individual variation in TG measurements, the risk of CHD adjusted for age, gender, and calendar period was approximately twofold higher in individuals in the top third of log TG levels compared with the bottom third. The attributable risk was similar in men and women and in fasted and nonfasted participants. After adjustment for HDL-C, the odds ratio for CHD was attenuated to 1.72 but remained significant (95% CI, 1.56-1.90) in those in the top third of log TG level compared with the bottom third. An important contribution was that repeated measurements an average of 4 years apart in 1933 participants in the EPIC-Norfolk study and an average of 12 years apart in 379 participants in the Reykjavik study showed that the long-term stability of log TG values is similar to that of blood pressure and total serum cholesterol (within-person correlation coefficients of 0.64 [95% CI, 0.60-0.68] during 4 years and 0.63 [95% CI, 0.57-0.70] during 12 years).
The Metabolic, Lifestyle and Nutrition Assessment in Young Adults (MELANY) study obtained two measurements of fasting TGs 5 years apart, as well as lifestyle variables, in 13,953 healthy men, aged 26 to 45 years, who were then observed for an average of 10.5 years. After adjustment for eating breakfast, smoking, exercise, and changes in body mass index, those in the top quintile of TGs had a fourfold higher risk of angiographically proven CHD compared with those in the lowest quintile. The magnitude of this risk was much greater than the average 1.7-fold increase in risk observed in the 2007 meta-analysis and other large studies, a finding thought secondary to the younger age of the cohort compared with other studies. Another important finding was that change in TG level between the initial and the second measurements was positively associated with change in CHD risk.
In a recent subgroup analysis of the Treating to New Targets (TNT) study and Incremental Decrease in Endpoints through Aggressive Lipid Lowering (IDEAL) study, the utility of TGs to predict new cardiovascular events was examined. IDEAL compared atorvastatin 80 mg with simvastatin 20 to 40 mg and TNT compared atorvastatin 80 mg with atorvastatin 10 mg in patients with CHD or a history of MI. After adjustment for age and gender, the risk of cardiovascular events was 63% higher in patients in the highest quintile of TG (HR, 1.63; 95% CI, 1.46-1.81) compared with the lowest quintile. The ability of TGs to predict risk was attenuated when HDL and apo B/apo A-I were in the model, and it was eliminated with inclusion of diabetes, body mass index, glucose, hypertension, and smoking ( P 0.044 and 0.621, respectively, for the trend across quintiles of TG). Similar results were observed in those in whom LDL-C had been lowered to goal.
The Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction (PROVE IT–TIMI 22) trial randomized patients with acute coronary syndromes to atorvastatin 80 mg or pravastatin 40 mg. An on-treatment TG level <150 mg/dL was independently associated with a significant reduction in risk for the composite primary endpoint of nonfatal MI, death, and recurrent acute coronary syndrome. After adjustment for LDL-C and other covariates in a subanalysis, each 10 mg/dL decline in on-treatment TG level was associated with a 1.6% lower risk of the primary endpoint.
In summary, several studies show TGs to be an independent predictor of CHD after adjustment, whereas in others, risk of TGs is attenuated or eliminated after adjustment, especially for HDL-C levels. In the next section, potential reasons for these findings are examined on the basis of the metabolic interrelationships between levels of TG-rich lipoproteins, small dense LDL, remnant particles, and HDL particles.
Lipoprotein Metabolism
Determination of the contribution of TGs to risk for atherosclerosis is complicated by the fact that the metabolism of TG-rich lipoprotein fractions significantly affects the levels and composition of other lipoprotein fractions that also contribute to cardiovascular risk. Therefore, in evaluating the role of TG and TG-rich lipoproteins in contributing to atherosclerosis, one must consider the complex interrelationships between the various lipoproteins during metabolism.
TGs (triacylglycerols) are composed of a glycerol backbone in which each of the three hydroxyl groups is esterified with a fatty acid. TGs play an important role in lipid metabolism and are the major source of metabolic energy storage. Cholesterol and TGs are almost insoluble in plasma; therefore, they are transported in lipoprotein particles from the liver (endogenous production of very-low-density lipoprotein [VLDL]) and intestine (exogenous production of chylomicrons [CMs] from dietary fat) to various tissues—TGs for energy use in skeletal muscle or storage in adipose tissue and cholesterol for synthesis of steroid hormones, bile acid formation, and cell membrane structural integrity. Lipoprotein particles are spherical and contain a central core of varying amounts of TG and cholesteryl ester (CE) (both nonpolar lipids) covered on the surface by a monolayer of polar lipids (primarily phospholipids), one or more apolipoproteins, and unesterified cholesterol, which is also found in the core as particle size increases. They are divided into five major classes on the basis of density, which is inversely related to size and lipid content: CMs, VLDL, intermediate-density lipoprotein (IDL), LDL, and HDL. The main apolipoproteins include apo B (B100 and B48), apo A (A-I, A-II, A-IV, and A-V), apo C (C-I, C-II, and C-III), and apo E (E2, E3, and E4 isoforms). Apolipoproteins serve as cofactors for enzymes and ligands for receptors and therefore play key roles in the regulation of lipoprotein metabolism.
Apo B exists in two forms in plasma, apo B100 and apo B48, both of which are products of the same structural gene on chromosome 2. Both apo B48 and apo B100 are constitutively synthesized; the availability of TG and CE (the core lipids) regulates their secretion. Containing mainly TG in their core, CMs and VLDL are the major TG carriers in plasma and are the two largest classes of lipoproteins. There are two pathways for the metabolism of TG-rich lipoproteins; the exogenous pathway carries dietary fats by apo B48 in CMs, whereas the endogenous pathway represents hepatic secretion of VLDL, a TG-rich apo B100–containing lipoprotein.
Endogenous Pathway: Assembly of VLDL Apo B-100 Lipoprotein Particles
The full-length apo B100 is a glycoprotein that contains 4536 amino acids. Apo B100 is synthesized by the liver and secreted in the form of VLDL, a TG-rich-lipoprotein that in plasma contains 60% TG by mass and 20% CE by mass. Both fatty acids synthesized de novo from acetyl coenzyme A and fatty acids from lipolysis of stored adipose tissue TG or from core lipids of TG-rich remnant particles returning to the liver stimulate the assembly of VLDL in the liver ( Fig. 15-1 ). Microsomal triglyceride transfer protein (MTP) transfers TGs from the cytosol to the endoplasmic reticulum containing nascent apo B during the assembly of CM and VLDL in enterocytes and hepatocytes, respectively. MTP gene expression is regulated by insulin, possibly through transcriptional activity of the sterol response element–binding protein 1c (SREBP-1c). This may explain why VLDL secretion and TG levels are increased in insulin resistance syndromes. In the plasma (see Fig. 15-1 ), VLDL particles adhere to glycosaminoglycan molecules on endothelial cells of capillaries, primarily in muscle, lung, and adipose tissue, where interaction with lipoprotein lipase (LPL) and glycosylphosphatidylinositol-anchored high-density lipoprotein–binding protein 1 (GPIHBP1) results in hydrolysis of VLDL triglycerides to free fatty acids and glycerol.
The removal of TG from VLDL by LPL exposes the apo E molecules on the lipoprotein surface of VLDL. Apo E functions as a ligand in the receptor-mediated clearance of CM and VLDL remnants in the liver through several receptors: a remnant receptor, the LDL receptor (an apo B/apo E receptor), the LDL receptor–related protein, the VLDL receptor, and apo E receptors. With stable isotope methodology, Welty and colleagues published the first study of simultaneous kinetics of apo B48 and apo B100 in human subjects and showed that about 50% of VLDL is directly removed from plasma (see Fig. 15-1 ) and therefore not converted to IDL or LDL. During the removal of the TG from the remaining 50% of VLDL (referred to as the delipidation cascade), the VLDL particles are hydrolyzed by LPL to smaller particles termed VLDL remnants or IDL (relatively enriched in CE but also containing TG), which then interact with hepatic lipase (HL) and are converted to CE-rich LDL, the major cholesterol-carrying lipoprotein in normal human plasma. Apo B100 is the main structural protein of LDL and contains the LDL receptor–binding domain; therefore, LDL is removed from the circulation by binding mainly to hepatic LDL receptors (see Fig. 15-1 ).
Exogenous Pathway: Assembly of Chylomicrons
Produced in the intestine in response to dietary fat, CMs contain apo B48, the amino-terminal 48% of apo B100, which is synthesized by the intestine and produced by a premature stop codon at the apo B100 codon 2153 by tissue-specific mRNA processing (see Fig. 15-1 ). Genes related to sterol absorption (ABCG5) and nuclear receptors LXR and FXR in bile and sterol metabolism may affect variability in absorption of dietary fats and CM formation. Within the intestinal cell, free fatty acids combine with glycerol to form TGs, and cholesterol is esterified by acyl coenzyme A:cholesterol acyltransferase (ACAT) to form cholesterol esters.
CMs enter lacteals in the intestinal villi and travel through the lymphatics to the thoracic duct and then into the bloodstream. Similar to the metabolism of VLDL, CMs bind to LPL on the surface of endothelial cells, where most of the TGs and some surface glycerophospholipids are catabolized to form CM remnants and free fatty acid by LPL, present in capillary walls primarily of skeletal muscle, adipose tissue, and lung, with apo C-II as a cofactor and apo C-III as an inhibitor (see Fig. 15-1 ). Apo B48 does not contain an LDL receptor–binding domain; therefore, the CM remnants are taken up by the liver by remnant receptors as well as by the LDL receptor that recognizes apo E. CMs also bind to HL on the sinusoidal surface of hepatocytes, where HL further hydrolyzes remnant lipids. In the fed state, both CMs and VLDL transport TGs to peripheral organs, where, through the action of tissue-specific LPL, TG-derived free fatty acids are used as energy in muscle and in other tissues, converted to TG, or stored in adipose tissue. Residual TGs (in the form of free fatty acids) and dietary cholesterol are rerouted to hepatocytes.
To transport TG and cholesterol to peripheral tissues, lipoprotein particles cross the endothelial barrier in blood vessels to reach the extracellular space (see Fig. 15-1 ). The subendothelial retention of apo B100–containing lipoproteins by a charge-mediated interaction with proteoglycans in the extracellular matrix is thought to be the initiating event in atherogenesis. Smaller, electronegative LDL particles penetrate the endothelial barrier 1.7-fold better than large LDL particles do; CM and VLDL remnants also can cross. All of these particles interact with positively charged intimal proteoglycans. Oxidation, by reactive oxygen species, of fatty acids of surface phospholipids of the apo B–containing lipoprotein particles results in modification of lysine residues of apo B. Scavenger receptors on macrophages recognize modified apo B, and unregulated uptake of the modified lipoprotein particle causes macrophage accumulation of lipids, a process leading to a foamy cytoplasm and the term foam cells (see Fig. 15-1 ). The most important scavenger receptor is CD36 (also called scavenger receptor B). As foam cells increase in number, the fatty streak develops. VLDL particles from patients with hypertriglyceridemia are enriched in apo E, which can lead to a conformational change in the VLDL particle that facilitates binding to the macrophage scavenger receptor, resulting in unregulated uptake similar to that seen with oxidized LDL. CM remnants are also small enough to enter the subendothelial space, where they are taken up by macrophages that promote atherogenesis. Similar to LDL, IDL is taken up by macrophages and can also cause foam cell formation; endothelium-dependent vasomotor function in human coronary arteries is impaired by both IDL and LDL. Several angiographic trials of cholesterol-lowering therapy have shown that serum IDL concentrations are predictive of an increased incidence of CHD and an increased incidence of coronary events in those with CHD, independently of other factors. VLDL and IDL have been identified in human atherosclerotic plaques, and these particles are associated with progression of mild to moderate coronary lesions. In the Monitored Atherosclerosis Regression Study (MARS) angiographic trial, IDL, but not VLDL or LDL, was associated with progression of carotid artery intima-media thickness. Moreover, the total TG level and markers for TG metabolism predicted risk of progression of low-grade but not of high-grade coronary artery lesions.
Effect of Triglyceride-Rich Lipoproteins on Level of HDL-C in Humans
Levels of TG-rich lipoproteins also affect HDL level and particle size. HDL levels are inversely related to risk of CHD. In contrast to LDL and VLDL, HDL has antiatherogenic properties that include reverse cholesterol transport, antioxidation (protecting apo B lipoproteins from oxidation), antithrombotic and anti-inflammatory properties, and maintenance of endothelial function. Apo A-I is secreted from the liver in a lipid-poor form. The major formation of HDL particles occurs when lipid-poor apo A-I interacts with the ABCA1 receptor on the surface of peripheral cells. This results in transfer of free cholesterol and phospholipid to the apo A-I, forming a pre-β particle. When the free cholesterol is esterified under the action of lecithin-cholesterol acyltransferase (LCAT), this particle becomes a mature HDL particle. The majority of the proteins of HDL, apo A-I, A-II, and A-IV, are secreted as components of VLDL, which are then transferred to the apo A-I particle in the plasma. Plasma HDL is mainly assembled extracellularly during transfer of surface components of TG-rich lipoproteins, including phospholipids and cholesterol.
HDL particles play a significant role in delivery of cholesterol from peripheral cells to the liver after esterification within the particle to CE through plasma LCAT, a process known as reverse cholesterol transport. There are two pathways by which this can occur. In the first, the scavenger receptor class B type 1 (SR-B1) mediates hepatic uptake of CE from HDL particles without uptake of apo A-I or the whole HDL particle. In the second pathway, cholesteryl ester transfer protein (CETP) catalyzes the transfer of CE from HDL to apo B–containing lipoproteins (VLDL and LDL) in exchange for TG from the apo B–containing lipoproteins ( Fig. 15-2 ). This exchange results in apo B–containing lipoproteins, which are enriched with CEs and depleted of TGs, and HDL particles, which are depleted of CEs and enriched with TGs. The TG-rich and CE-poor HDL particles are catabolized faster than the large, CE-rich HDL particles are (apo A-I fractional catabolic rate [FCR] is increased as noted in Fig. 15-2 ), resulting in lower levels of HDL-C in the setting of high TG levels. The apo B–containing lipoproteins, now enriched in CE, are taken up by the liver receptors, as previously described. This exchange through CETP action is thought to be responsible for the inverse relationship between levels of TG and HDL-C. The cardioprotective effect of HDL has been largely attributed to its role in reverse cholesterol transport. HL then hydrolyzes the TGs within the TG-rich LDL to release free fatty acids, a process that remodels the LDL particles into smaller and denser LDL particles that can enter the arterial intima more easily than larger LDL particles, thus making them more atherogenic (see Fig. 15-2 ). Small, dense LDL particles also bind less avidly to the LDL receptor, thus prolonging their half-life in the circulation, making these particles more susceptible to oxidative modification and to subsequent uptake by the macrophage scavenger receptors.
Using stable isotopes in the fed (nonfasting) state in humans, Welty and colleagues showed that apo A-I FCR is inversely correlated with the FCR of apo B48 (see Fig. 15-2 for details) but not with VLDL apo B100 FCR or production rate. Thus, when CM apo B48 clearance is delayed (represented by decreased apo B48 FCR in Fig. 15-2 ), TG-rich apo B particles accumulate and the TG is transferred to HDL apo A-I particles in exchange for CE. These results suggest that in the fed state, levels of TG-rich apo B48 of intestinal origin are more important determinants of levels of HDL-C than the amount of TG-rich lipoproteins of hepatic origin, which contain apo B100.
Mutations in Enzymes of VLDL and Chylomicron Metabolism Affecting Risk for Atherosclerosis
Alterations in the enzymes involved in lipid metabolism may affect risk for atherosclerosis. As noted earlier, LPL hydrolyzes TGs contained in the core of CMs and VLDL. In partial LPL deficiency, TG-enriched lipoproteins have a prolonged circulation time that allows more interaction with the endothelium. An Asn291Ser substitution in LPL causes impaired function of this enzyme and is associated with an increase in plasma TG. Female but not male carriers of this mutation have a twofold increase in the risk of CHD and nonfatal cerebrovascular disease.
Apo C-II is an activator of LPL that hydrolyzes the core TGs, thereby releasing free fatty acids and making the CMs and VLDL progressively smaller and forming remnants. Apo C-II thus increases the catabolism of both CM and VLDL, thereby lowering TG levels. Apo C-III is an inhibitor of LPL, and in contrast to apo C-II, which lowers TG levels, apo C-III can raise TG levels by stimulating VLDL synthesis, inhibiting LPL, and inhibiting the binding of remnants to the LDL receptor mediated by apo E. Thus, high levels of apo C-III are associated with TG-enriched VLDL particles that are ultimately converted to TG-rich remnants. These remnants are then lipolyzed to small, dense LDL particles by HL. Thus, high levels of apo C-III are associated with TG-rich VLDL particles that circulate longer and are therefore ultimately converted to TG-remnants, which are then lipolyzed by HL to small, dense LDL particles. Most patients with elevated levels of VLDL have excess amounts of both TGs and apo C-III within their lipid particles. Persons lacking apo C-III have efficient lipolysis of TGs and therefore low levels of TG.
Produced in the liver, apo A-V activates proteoglycan-bound LPL and thus accelerates TG hydrolysis from VLDL and CMs independently of other apoproteins. A sequence element between residues 185 and 228 functions in binding of apo A-V to heparin sulfate proteoglycans, members of the LDL receptor family and GPIHBP1. Plasma levels of apo A-V are extremely low, and this factor, plus the association of apo A-V with cytosolic lipid droplets, suggests that apo A-V may modulate TG metabolism within the cell. The gene for apo A-V is located at the apo A1/C3/A4/A5 gene cluster on chromosome 11q23. Several single-nucleotide polymorphisms are associated with significantly higher plasma TG levels in patients (i.e., −1131T>C, S19W, G185C). The structural mutations Q139X, Q148X, and IVS3 + 3G>C predispose to familial hypertriglyceridemia and late-onset chylomicronemia. Thus, apo A-V is an important regulator of plasma TG levels in humans.
Polymorphisms in HL may also affect risk for atherosclerosis. There are four common sequence polymorphisms in the HL gene promoter; the most frequent is a C to T substitution. The presence of a C allele is associated with higher HL activity; smaller, denser, and more atherogenic LDL particles; and lower levels of HDL-C.
Classification and Causes of Hypertriglyceridemia
Hypertriglyceridemia can occur as an isolated hypertriglyceridemia or in combination with hypercholesterolemia (familial combined hyperlipoproteinemia and familial dysbetalipoproteinemia) ( Table 15-1 ). Both forms can be further subdivided into primary and secondary causes. LPL, HL, and their apolipoproteins regulate levels of TG; therefore, abnormalities in any of these can affect TG levels.
Isolated Hypertriglyceridemia
Familial Hypertriglyceridemia—Severe
Familial hypertriglyceridemia is characterized by elevated TG levels with normal cholesterol and can be divided into severe and mild forms (see Table 15-1 ). In the severe form, TGs exceed 1000 mg/dL because of increases in both CMs and VLDL particles (type V in original Fredrickson-Levy classification, also called primary mixed hypertriglyceridemia). Most patients with mixed hypertriglyceridemia have familial hypertriglyceridemia due to partial deficiency of LPL or apo C-II (the ligand for LPL on CMs and VLDL) deficiency exacerbated by one or more of the secondary disorders noted in Table 15-1 .
The other major primary cause of TGs >1000 mg/dL is exogenous hyperlipemia or familial chylomicronemia due to CMs (hyperlipoproteinemia type I in original Fredrickson-Levy classification). The most common primary cause of type I is complete absence of either LPL activity or apo C-II. When LPL is absent (prevalence is 1 in a million), TG is generally >2000 mg/dL. Patients with TGs >2000 mg/dL usually have both a genetic form of hypertriglyceridemia and a secondary cause. Similar clinical manifestations of both types I and V include hepatosplenomegaly and occasional eruptive xanthomas (see Table 15-1 ). Features that distinguish type I from type V include (1) presentation of type I in childhood and of type V in adulthood; (2) absence of LPL or apo C-II activity or homozygous gene mutations in type I; (3) presence of a secondary factor in type V (alcohol, obesity, type 2 diabetes mellitus, hypothyroidism, or poor diet); (4) higher prevalence of type V than of type I; and (5) increased CMs alone in type I compared with elevations in both CMs and VLDL in type V.
The primary risk associated with TG levels >1000 mg/dL is pancreatitis. Minimal atherosclerotic risk is reported for patients with hyperchylomicronemia (type I) or the severe form of familial hypertriglyceridemia (type V), probably because the lipoprotein particles are too large to enter the arterial wall.
Patients with marked hypertriglyceridemia (>1000 mg/dL [11.3 mmol/L]) may develop the chylomicronemia syndrome. This can include recent memory loss, abdominal pain or pancreatitis, dyspnea, eruptive xanthoma, flushing after alcohol ingestion, and lipemia retinalis.
Familial Hypertriglyceridemia—Mild
The mild form of familial hypertriglyceridemia (type IV hyperlipoproteinemia phenotype) is an autosomal dominant disorder characterized by mild to moderate elevations in TG from 200 to 500 mg/dL, often in association with insulin resistance, obesity, hyperglycemia, hypertension, hyperuricemia, and low HDL-C. Mutations in the LPL gene decrease enzyme activity and therefore delay the degradation of CMs and VLDL that carry endogenous TGs. Gly188Glu, Asp9Asn, and Asn291Ser are N-terminal mutations that reduce the activity of LPL, resulting in an increase in serum TGs by 20% to 80% and also lower levels of HDL-C. More marked elevations require some other factor, such as one of the drugs or acquired disorders (e.g., estrogen replacement therapy in postmenopausal women).
Familial hypertriglyceridemia is common in patients with premature CHD. The prevalence of familial hypertriglyceridemia with low HDL-C levels was 15% in patients undergoing coronary arteriography before the age of 55 years. Among first-degree relatives of affected patients, baseline serum TG levels predicted cardiovascular mortality, independent of serum total cholesterol.
Hypertriglyceridemia and Hypercholesterolemia
Hypertriglyceridemia can occur in two phenotypes in combination with hypercholesterolemia. The first is familial combined hyperlipoproteinemia (FCHL), and the second is familial dysbetalipoproteinemia.
Familial Combined Hyperlipoproteinemia
In FCHL (type IIb), overproduction of hepatically derived VLDL apo B100–containing lipoproteins results in plasma TG levels of 200 to 500 mg/dL and plasma cholesterol levels of 200 to 400 mg/dL and small, dense LDL. FCHL has an autosomal dominant mode of inheritance with variable penetrance and a population prevalence of 1% to 2%. It is the most common familial lipid disorder in post-MI patients and accounts for one third to one half of familial causes of CHD. The molecular basis includes mutations in LPL and apo C-III and upstream stimulatory factor 1 (USF1), which encodes an upstream stimulatory factor. FCHL is also linked with insulin resistance due to both increased free fatty acid flux from the periphery and insulin-stimulated lipogenesis, which increases the production of VLDL. Families with premature CHD often have either familial dyslipidemia (high TGs and low HDL-C [type IV]) or FCHL (high TGs, high LDL-C, and low HDL-C [type IIb]).
Patients with FCHL can present with combined hypercholesterolemia and hypertriglyceridemia or either abnormality alone. Thus, subjects with FCHL who overproduce VLDL particles and also synthesize TG at an increased rate will secrete an increased number of large, TG-rich VLDL particles. If they are unable to efficiently catabolize these VLDL particles because of an LPL mutation or low LPL activity, they will have a high TG level but a normal or reduced number of LDL particles and thus a normal LDL-C level. On the other hand, with efficient catabolism of the increased numbers of large, TG-rich VLDL particles, the number of LDL particles is increased, resulting in both increased TG and LDL-C levels. Finally, subjects who synthesize a normal quantity of TG but an increased number of VLDL particles (with normal TG load) have increased numbers of LDL particles and elevated plasma LDL-C levels but a normal TG level. LPL may be responsible for part of this phenotypic variability as hypertriglyceridemia is more prominent in patients with LPL deficiency or an LPL gene mutation.
Familial Dysbetalipoproteinemia
As noted before, the apo E ligand is necessary for receptor-mediated catabolism of CM and VLDL remnants. There are three isoforms of the apo E allele: apo E2, apo E3, and apo E4. Apo E3/3 is the most common apo E genotype. Apo E2 differs from apo E3 by substitution of cysteine for the normal arginine at residue 158 in the receptor-binding domain. Consequently, apo E2 does not bind as well as apo E3 to the apo B/E (LDL receptor). Subjects with familial dysbetalipoproteinemia have the apo E2/E2 genotype, which is an autosomal recessive disorder characterized by an accumulation of VLDL and CM remnants in the plasma due to inefficient uptake through apo E2, which binds poorly to hepatic LDL-related receptors. Consequently, they have an increase in cholesterol-enriched VLDL (β-VLDL, also termed IDL) and CM remnants. The prevalence of the apo E2 isoform is 1 in 100; however, only approximately 1 in 10,000 carriers exhibits the dyslipidemia, which is thought to be triggered by a secondary cause, such as marked hyperglycemia (type 2 diabetes), hyperuricemia (gout), hypothyroidism, or obesity. Classic physical findings include tuberoeruptive xanthomas and xanthomas of the palmar creases; the risk for CHD is increased.
Secondary Causes
The major secondary causes (see Table 15-1 ) of hypertriglyceridemia include poorly controlled type 2 diabetes, obesity, excessive alcohol intake, renal disease, pregnancy, medications ( Table 15-2 ), excessive ingestion of saturated fats and simple sugars, nonalcoholic hepatosteatosis, and physical inactivity. Alcohol intake can cause elevated TG by inhibition of LPL or increased VLDL TG production. For every gram of alcohol consumed per day, TG concentration can increase on average by 0.19 mg/dL, which is about 5.7 mg/dL for 30 g of alcohol. These secondary causes must always be considered during the evaluation and treatment of hypertriglyceridemia.
| Drug | Triglycerides | LDL-C | HDL-C |
|---|---|---|---|
| Alcohol | Increased | No effect | Increased |
| Estrogens, estradiol | Increased | Decreased | Increased |
| Androgens, testosterone | Increased | Increased | Decreased |
| Progestins | Decrease | Increase | Decrease |
| Glucocorticoids | Increased | No effect | Increased |
| Cyclosporines | Increased | Increased | Increased |
| Tacrolimus | Increased | Increased | Increased |
| Thiazide diuretics | Increased | Increased | Decreased |
| Beta blockers | Increased | No effect | Decreased |
| Sertraline | Possible increase | Increased | No effect |
| Protease inhibitors | Increased | No effect | No effect |
| Valproate and related drugs | Increased | No effect | Decreased |
| Isotretinoin | Increased | No effect | Decreased |
| Clozapine, olanzapine † | Increased | No effect | Decreased |
* Alcohol, estrogens, estradiol, glucocorticoids, thiazide diuretics, beta blockers, sertraline, protease inhibitors, valproate and related drugs, and isotretinoin can cause severe hypertriglyceridemia and the chylomicronemia syndrome in patients with a familial form of hypertriglyceridemia.
† Second-generation antipsychotics: clozapine and olanzapine have most effect; risperidone and quetiapine have intermediate effects; and aripiprazole and ziprasidone have least effect.
Metabolic Syndrome
The association of hypertriglyceridemia with obesity and diabetes or glucose intolerance is termed the metabolic syndrome, which is defined clinically according to the National Cholesterol Education Program by at least three of the following: central obesity (waist circumference >35 inches in women and >40 inches in men), fasting blood glucose concentration ≥100 mg/dL, TGs ≥150 mg/dL, low HDL-C (<40 mg/dL in men and <50 mg/dL in women), and systolic or diastolic blood pressure ≥130/≥85 mm Hg. Atherogenic dyslipidemia in metabolic syndrome and people with type 2 diabetes (termed diabetic dyslipidemia) is characterized by elevated TGs and small, dense, cholesterol-depleted LDL and HDL particles. Insulin resistance increases mobilization of free fatty acids from adipose tissue to liver, where increased production of VLDL occurs; thus, hypertriglyceridemia in type 2 diabetes and metabolic syndrome is usually secondary to increased VLDL concentrations in plasma, with or without chylomicronemia. Downregulation of LPL expression in insulin resistance leads to decreased catabolism of TG-rich VLDL. The higher levels of TG promote CETP-mediated transfer of CE from HDL, thus producing TG-rich small, dense HDL that are catabolized more rapidly, leading to low levels of HDL-C. Small, dense HDL also have reduced antioxidant and anti-inflammatory properties. The metabolic syndrome and increased TG-rich lipoproteins are also associated with a proinflammatory and prothrombotic state due to the presence of atherogenic lipoproteins, clotting factors, and increased plasma viscosity. In clinical practice, elevated serum TGs are most often observed in persons with the metabolic syndrome, although secondary or genetic factors can raise TG levels. Metabolic syndrome has a prevalence of 24% in U.S. adults and 43% in adults older than 60 years ; therefore, it is a major health problem.
Laboratory Evaluation of Hypertriglyceridemia
In this section, the literature on several clinical laboratory approaches to assess the risk of TGs and TG-rich lipoproteins is reviewed.
Fasting Versus Nonfasting Triglyceride Levels
The Friedewald equation is often used to estimate LDL-C levels: LDL-C TC − (TG/5 + HDL-C); it requires measurement of fasting TG levels. TG/5 is an estimate of the cholesterol in VLDL and IDL particles (VLDL-C + IDL-C). This equation has been used for more than 40 years to calculate LDL-C; therefore, the majority of research studies have examined the association of TG with CHD on the basis of fasting TG levels. As shown earlier in the section on lipoprotein metabolism, levels of TG-rich- and apo B48–containing CMs of intestinal origin are probably more important determinants of levels of HDL-C than the amount of TG-rich lipoproteins of hepatic origin, which contain apo B100. In the fasting state, very little apo B48 is being produced; therefore, it would make sense that nonfasting levels of TG might be a better predictor not only of the concentration of TG-rich lipoproteins circulating in plasma most of the time but also of the HDL-C level.
In 2007, Bansal and colleagues and Nordestgaard and coworkers reported that in two long-term prospective cohort studies, TG levels obtained 2 to 4 hours postprandially predicted risk of CHD better than TG levels measured after a 12- to 14-hour fast and better than LDL-C calculated by the Friedewald equation. In both studies, elevated postprandial TG levels increased risk of CHD for both sexes; however, women had greater risk of CHD associated with hypertriglyceridemia than men did, confirming prior studies that showed higher risk in women than in men. Postprandial lipoproteins are TG rich, and if their catabolism is delayed (insulin resistance, LPL mutations), the products of their metabolism, CM remnants and small dense LDL, can remain in the plasma for 12 hours or more, with exposure of the endothelium to TG-rich, atherogenic remnant particles, a finding accounting for the greater CHD risk with postprandial increases in TG levels.
Although postprandial TG levels may predict risk better than fasting TG levels, nonfasting TG measurements may be difficult to incorporate into clinical practice. First, a standard fat-feeding protocol would need to be developed and prepared fresh for each test. Second, a fat load may cause nausea or vomiting. Third, requirement for a 2- to 4-hour postprandial peak may not be practical in outpatient care. In the long run, fasting TG measurements may be more reliable because of controlled conditions. The strong correlation between postprandial and fasting TG levels may obviate the need for repeated postprandial measurements.
Lipoprotein Particle Subclasses
Lipoprotein particle composition has been associated with differences in the relative atherogenicity of lipoproteins. Small, dense LDL particles are more susceptible to accelerated oxidation and are incorporated more readily by vascular wall macrophages than other lipoprotein particles are. Several prospective cohort studies have reported that the number of small, dense LDL particles is a greater predictor of CHD risk than are measured levels of serum LDL-C. Thus, unlike the linear relationship of risk with LDL-C, the risk of CHD associated with elevated TG levels may be a function of the associated lipoprotein disorder more than a direct numerical correlation with TGs. In the Veterans Affairs High-Density Lipoprotein Intervention Trial (VA-HIT), a change in TG concentration did not predict the magnitude of reduction in risk of CHD events; rather, it was the change in LDL and HDL particles that predicted change in risk.
Patients with CHD often have increases in small, dense LDL; however, LDL particle size is not an independent predictor of CHD, and in fact, the major factor regulating LDL particle size is plasma TG level. Therefore, the measurement of particle size has not been proven to provide better information than standard lipid and lipoprotein measurements. Patients with TG levels >150 mg/dL generally have increased levels of small, dense LDL particles. As noted earlier, after conversion of large VLDL particles to TG-rich remnants, HL hydrolyzes the remnants to small, dense LDL particles. Consequently, levels of TG and VLDL are strongly and positively correlated with levels of small, dense LDL particles. In some prospective, nested case-control studies, subjects with small, dense LDL particles have an increased risk for CHD; however, other studies have concluded that increased levels of small LDL particles are not an independent predictor of CHD but rather are products of elevated levels of TG-rich lipoproteins that are associated with an atherogenic milieu—elevated TG, reduced HDL-C, and potentially other biomarkers of the metabolic syndrome. Because both large and small LDL subtypes have been shown to predict CHD, apo B concentrations, which estimate particle numbers, rather than LDL particle size, may be the better predictor of CHD. Moreover, apo C-III enrichment of apo B–containing lipoproteins has been linked to the atherogenicity of apo B–containing lipoproteins and therefore may also be a better predictor of CHD than particle size.
Apo B
Much data supports apo B as a predictor of CHD. Fredrickson and associates recognized more than 40 years ago that atherosclerosis is more closely related to the total number of apo B–containing particles rather than to LDL-C or TG concentrations alone. One apo B molecule is present on the surface of VLDL, IDL, LDL, and lipoprotein(a), a molecule of apo B100 covalently bound to apoprotein (a) ; therefore, the apo B level may provide a more direct measure of circulating atherogenic lipoproteins. As noted earlier, not all forms of hypertriglyceridemia are atherogenic ; however, the relative atherogenicity of apo B is well established by the fact that modification of lysine residues on apo B is necessary for uptake by scavenger receptors on macrophages. However, the routine measurement of apo B is not always practical because of cost and technical limitations that preclude measurement in routine laboratory assays in hospital chemistry laboratories.
Non–HDL-C
For all of these reasons, as suggested by the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III), non–HDL-C is a practical and useful surrogate measure of atherogenic particle concentration and more predictive of CHD risk than LDL-C level alone (especially when TG levels are elevated) because this measure is a sum of all atherogenic lipoproteins. Non–HDL-C is TC − HDL-C, which is the sum of VLDL-C, IDL-C, and LDL-C. Therefore, non–HDL-C includes the cholesterol in all of the atherogenic apo B–containing lipoproteins: TG-enriched lipoproteins, CMs, CM remnants, VLDL and VLDL remnants, IDL, LDL, and lipoprotein(a). Non–HDL-C is accurate and reliable in a nonfasting state; therefore, measurement of non-HDL is practical and easy. When lipid levels are normal, non–HDL-C is highly correlated with apo B levels. Because high LDL-C and TG levels confer greater risk for CHD than high LDL-C alone, NCEP ATP III guidelines recommended non–HDL-C as a secondary target of therapy when the serum TG level is ≥200 mg/dL after the LDL-C target is achieved ( Table 15-3 ).
| Risk Category | LDL-C Goal | Non–HDL-C Goal † |
|---|---|---|
| Very high risk ‡ | <70 mg/dL (optional) | <100 mg/dL |
| High risk: CHD § or CHD risk equivalents ∥ | <100 mg/dL | <130 mg/dL |
| Moderately high risk: ≥2 risk factors ¶ (10-year risk 10%-20%) | <130 mg/dL | <160 mg/dL |
| Moderate risk: >2 risk factors (10-year risk <10%) | <130 mg/dL | <160 mg/dL |
| Lower risk: 0 or 1 risk factor | <160 mg/dL | <190 mg/dL |
* NCEP ATP III guidelines for non–HDL-C state that in addition to the primary goal of LDL-C reduction, non–HDL-C is a secondary target of therapy in patients with TG levels of 200 to 499 mg/dL. Because a normal VLDL-C level is <30 mg/dL, the therapeutic goal for non–HDL-C is 30 mg/dL higher than the goal for LDL-C. Factors that place patients at very high risk favor a decision to reduce LDL-C levels to <70 mg/dL. The optional goal of <70 mg/dL does not apply to subjects who are not very high risk.
† When TGs are 200 to 499 mg/dL.
‡ Presence of established cardiovascular disease plus (1) multiple major risk factors (especially diabetes mellitus), (2) severe and poorly controlled risk factors (especially continued cigarette smoking), (3) multiple risk factors for the metabolic syndrome (especially high TGs [>200 mg/dL] plus non–HDL-C >130 mg/dL with low HDL-C [<40 mg/dL]), and (4) acute coronary syndromes.
§ CHD includes history of myocardial infarction, unstable angina, stable angina, coronary artery procedures, or evidence of clinically significant myocardial ischemia.
∥ CHD risk equivalents include clinical manifestations of noncoronary forms of atherosclerotic disease, diabetes, and ≥2 risk factors, with 10-year risk for hard CHD >20%.
¶ Risk factors include cigarette smoking, hypertension (blood pressure <140/90 mm Hg or taking antihypertensive medication), low HDL-C (<40 mg/dL), family history of premature CHD, and age (men <45 years, women <55 years).
Predictive Value of Non–High-Density Lipoprotein Cholesterol for Coronary Heart Disease
Several observational and intervention studies have reported that elevated levels of non–HDL-C are predictive of cardiovascular disease and cardiovascular disease mortality, similar to the predictive value of apo B and as good as or better than that of LDL-C. In the Bypass Angioplasty Revascularization Investigation (BARI), 1514 patients with multivessel CHD were observed for 5 years. Non–HDL-C, but not HDL-C or LDL-C, was a significant univariate and multivariate predictor of nonfatal MI (RR, 1.049; 95% CI, 1.006-1.093; P < 0.05) and angina pectoris (RR,1.049; 95% CI, 1.004-1.096; P < 0.05) at 5 years. Non–HDL-C did not predict mortality in this study. In long-term follow-up from the Lipid Research Clinics Program Follow-up Study, increases of 30 mg/dL in non–HDL-C and LDL-C levels corresponded to increases in CVD mortality of 19% and 15%, respectively, in men and 11% and 8%, respectively, in women. The risk for CHD death was lowest in men and women with LDL-C levels <100 mg/dL in the Atherosclerosis Risk in Communities (ARIC) study; elevated TG levels were associated with substantially greater relative risks in women (4.7) than in men (2.1) after adjustment for LDL-C, HDL-C, and lipoprotein(a). In prospective follow-up of a cohort of 15,632 healthy women older than 45 years at baseline in the Women’s Health Study, non–HDL-C was as good as or better than apolipoprotein fractions for prediction of risk of a first cardiovascular event. In 6-year follow-up of 18,225 men in the Health Professionals Follow-up Study free of CHD at baseline, the relative risk of CHD in the highest quintile of non–HDL-C compared with the lowest quintile was 2.76 after adjustment (95% CI, 1.66-4.58), which was better than LDL-C, 1.81 (95% CI, 1.12-2.93), but not quite as good as apo B, 3.01 (95% CI, 1.81-5.00). After mutual adjustment of non–HDL-C and LDL-C, only non–HDL-C was predictive of CHD. In prospective follow-up of 1562 men and 1760 women older than 30 years and free of CHD at baseline in the Framingham Heart Study, the predictive value of non–HDL-C was better than LDL-C and comparable to apo B for prediction of risk of CHD. After multivariate adjustment in a subsequent analysis of the combined original Framingham Heart Study cohort and the offspring (2693 men, 3101 women), VLDL-C was an independent predictor of risk and non–HDL-C level was a stronger predictor of CHD risk than LDL-C alone at TG levels both greater than and less than 200 mg/dL. These results suggest that VLDL-C contributes to the development of CHD in addition to LDL-C and thus support the fact that both VLDL-C and LDL-C are essential in predicting CHD risk. In fact, non–HDL-C was better than LDL-C.
The Emerging Risk Factors Collaboration analyzed records of 302,430 people without initial vascular disease from 68 long-term prospective studies (Europe and North America) for a total of 12,785 cases of CHD (8857 nonfatal MIs and 3928 deaths due to CHD) during 2.79 million person-years of follow-up (median, 6.1 years to first outcome). The hazard ratio for the primary outcome (nonfatal MI and CHD death) for TG was 1.37 (95% CI, 1.31-1.42) after adjustment for nonlipid risk factors. However, after further adjustment for HDL-C and non–HDL-C, the hazard ratio for TG was reduced to 0.99 (95% CI, 0.94-1.05) ( Fig. 15-3 ). The hazard ratio for CHD with non–HDL-C was 1.56 (95% CI, 1.47-1.66) after adjustment for nonlipid risk factors. After adjustment for HDL-C and log TG, the hazard ratio for non–HDL-C remained significant at 1.50 (95% CI, 1.39-1.61) ( Fig. 15-4 ). There was no difference between those who fasted and those who were nonfasting and no difference by gender for either TG or non–HDL-C. In a subset with available measurements, the hazard ratio for directly measured LDL-C was 1.38 (95% CI, 1.09-1.73), which was similar to the hazard ratio of 1.42 (95% CI, 1.06-1.91) for non–HDL-C in the same subset. Analysis by HDL-C showed that after adjustment for nonlipid risk factors, non–HDL-C, and log TG, the hazard ratio for CHD with HDL-C was 0.78 (95% CI, 0.74-0.82). When analyzed by quintiles of HDL-C levels, the hazard ratio was 0.35 (95% CI, 0.30-0.42) for a 15 mg/dL higher HDL-C and 80 mg/dL lower non–HDL-C; this was not changed by addition of TG level. The hazard ratios for non–HDL-C and apo B were very similar in magnitude and shape, as were those for HDL-C and apo A-I ( Fig. 15-5 ), findings suggesting that apolipoprotein measurements are no better in predicting risk than cholesterol levels. In summary, this largest meta-analysis suggests that TG level provides no additional information about risk when non–HDL-C is calculated from TC and HDL-C levels in either the fasting or nonfasting state.
