Traditionally, lipoproteins have been separated on the basis of their hydrated densities; the major density classes of lipoprotein particles include chylomicrons, very low-density lipoproteins (VLDL), intermediate-density lipoproteins, low-density lipoproteins (LDL), and high-density lipoproteins (HDL) [1]. Figure 1 depicts the metabolism of these lipoproteins. Chylomicrons are intestinal lipoproteins that transport dietary lipids to peripheral tissues and the liver. They are triglyceride rich and contain one form of apolipoprotein B (apo B), apo B-48. The triglycerides in chylomicrons are hydrolyzed by the endothelial enzyme lipoprotein lipase, which requires apolipoprotein C-II (apo C-II) as a cofactor [2]. The resulting chylomicron remnants are removed from the circulation by the liver through a process that involves the binding of apolipoprotein E (apo E) on the chylomicron remnants to a putative hepatic remnant receptor (or apo E receptor) [3]. Very low-density lipoproteins are triglyceride-rich lipoproteins secreted by the liver and contain another form of apo B, apo B-100. These triglycerides are also hydrolyzed by lipoprotein lipase, with conversion to the more dense VLDL remnants, or intermediate-density lipoprotein. Some VLDL remnants are removed from the circulation by the liver through an apo E-mediated process, but others are further hydrolyzed by the endothelial enzyme hepatic lipase, ultimately resulting in conversion to LDL. Low-density lipoprotein transports cholesterol ester to various peripheral tissues, but a substantial amount of LDL is eventually removed from the circulation by the liver when apo B-100 is bound to the hepatic LDL receptor [4]. Low-density lipoprotein can undergo oxidative modification, producing a form of oxidized LDL that can cause cholesterol loading in cells [5].

View larger version (21K): [in this window] [in a new window]   Figure 1. Schematic diagram of lipoprotein metabolism. HDL = high-density lipoproteins; IDL = intermediate-density lipoproteins; LDL = low-density lipoproteins; VLDL = very low-density lipoproteins.

High-density lipoproteins are synthesized and secreted by the intestine and the liver and also are generated by hydrolysis of triglyceride-rich lipoproteins [6, 7]. The major apolipoproteins in HDL are apolipoprotein A-I (apo A-I) and apolipoprotein A-II (apo A-II) (Table 1). High-density lipoprotein stimulates the efflux from cells of unesterified cholesterol, which is then converted to the esterified form by lecithin-cholesterol acyltransferase Figure 1, a plasma enzyme activated primarily by apo A-I [8]. As small, dense HDL₃ accumulates cholesteryl ester, it is transformed into larger, less dense HDL₂. High-density lipoprotein cholesteryl ester can be transferred to apo B-containing lipoproteins by the cholesteryl ester transfer protein [9]. This may be one important route of human reverse cholesterol transport [10]. High-density lipoprotein is a substrate for hepatic lipase, which hydrolyzes HDL phospholipids and triglycerides, creating smaller HDL particles [7] (Figure 1).

Plasma Lipoproteins and Coronary Artery Disease

The risk for premature atherosclerotic coronary artery disease is directly correlated with plasma concentrations of LDL cholesterol [11, 12] and inversely correlated with levels of HDL cholesterol (reviewed in reference 13). The independent association of plasma triglycerides with coronary artery disease risk is less certain [14, 15]. Interventions designed to decrease plasma LDL concentrations are effective in the primary prevention of coronary artery disease [16-18]. Lowering plasma LDL is also very effective in secondary prevention of coronary artery disease (reviewed in [19]), decreasing overall mortality rate [20], decreasing cardiovascular events [21-25], and producing an objective regression of atherosclerotic disease [22-26]. Although one primary prevention trial suggested an independent benefit of increasing plasma HDL cholesterol concentrations [18], no clinical trials have been done specifically to evaluate the effect of selectively increasing HDL in primary or secondary prevention of coronary artery disease.

Despite the association of plasma lipid levels with coronary artery disease risk, many patients with premature coronary artery disease do not have very high levels of LDL cholesterol or very depressed HDL cholesterol concentrations. Therefore, investigators continue to search for other clinical markers that will allow better prediction of coronary artery disease risk and can be used to guide therapeutic decisions to prevent or treat coronary artery disease. Quantitation of plasma apolipoproteins was proposed as one such clinical tool. In this review, we assess evidence regarding the clinical utility of apolipoprotein quantitation and review the use of plasma apolipoprotein concentrations in the primary and secondary prevention of coronary artery disease. We focus primarily on the apolipoproteins for which the most data and the most clinical evidence exist that are relevant to coronary artery disease: apo A-I, apo B, and lipoprotein(a) (Lp[a]). For each of these apolipoproteins, we address the question of whether quantitation of the apolipoprotein enhances the ability to predict coronary artery disease risk in healthy persons or recurrent events in patients with established coronary artery disease, and we suggest how knowledge of the plasma apolipoprotein concentration might influence clinical management.

We retrieved 82 articles from the English-language literature for the years 1975 to 1993 using MEDLINE (key words: "apolipoproteins," "quantitation," and "coronary artery disease") and review of article bibliographies. We examined all retrospective and prospective studies of apolipoprotein quantitation that used some measure of coronary artery disease as a criterion for patient selection, including acute myocardial infarction, classic angina pectoris, and angiographic evidence of severe coronary artery disease [22, 23]. Many of the studies were designed to address the predictive value of the test for the development of coronary artery disease; relatively few studies assessed the predictive value of the test for recurrent events in patients with established coronary artery disease. We found 71 retrospective cross-sectional studies, including 7 in children and adolescents, and 11 prospective studies. For each apolipoprotein, we discuss the retrospective studies as a group and specific studies where appropriate; each of the studies in children and each prospective study are discussed individually. More than 90% of studies were done in men, and therefore we cannot generalize results to women. In addition, because assays for apolipoprotein quantitation have not been standardized, we included a section addressing some of the methodologic issues in apolipoprotein quantitation.

Issues regarding Assay Methods and Standardization

The lack of standardization and reference methods for apolipoprotein assays is a limitation to the general application of apolipoprotein quantitation in clinical practice [27-31]. Variation in apolipoprotein measurements among laboratories can be substantial. A collaborative study initiated by the International Federation of Clinical Chemistry evaluated differences in apo A-I and apo B quantitation among 28 laboratories, 25 of which were company laboratories [29]. The overall interlaboratory coefficient of variation was 7% for apo A-I and 19% for apo B. After uniform calibration of assay standards, the coefficients of variation decreased to 5% and 6%, respectively. Among the factors resulting in this variation are preanalytical factors such as differences in sampling and storage conditions [32]. In addition, matrix effects (additives, stabilization processes) on the immunoreactivity of standards also can be a source of bias in apolipoprotein quantitation [33].

Assays for apo A-I and apo B are widely available and are frequently done by commercial laboratories. These laboratories use methods that often result in good intralaboratory reproducibility, with coefficients of variation within laboratories that are generally less than 4% [29]. However, some of these assays may be subject to interference by high plasma triglyceride levels [27]. Therefore, apo A-I and apo B measurements in patients with very high hypertriglyceride levels (>4.5 mmol/L [400 mg/dL]) should be interpreted with caution.

Several commercially available Lp(a) assay kits are used by research laboratories for Lp(a) quantitation. However, an Lp(a) assay has yet to be approved by the Food and Drug Administration for clinical use. As in the case of apo A-I and apo B, there has been no standardization of Lp(a) assays [34, 35]. Lipoprotein(a) is an acute-phase reactant [36] and should not be quantitated within several weeks after an acute illness or surgical procedure.

Apolipoprotein A-I

Background

Apolipoprotein A-I is the major apolipoprotein in HDL and serves various structural and functional roles in HDL metabolism (reviewed in reference 6). It is probably important in protecting against premature atherosclerosis. Genetic defects that cause the inability to synthesize apo A-I cause very low plasma concentrations of HDL cholesterol and premature coronary artery disease in the fourth and fifth decades [37-40]. Conversely, an increased rate of apo A-I production causes high plasma levels of HDL cholesterol and may be associated with protection from premature coronary artery disease based on familial longevity [41]. Furthermore, overexpression of human apo A-I in transgenic mice inhibits the development of atherosclerosis [42].

Cross-Sectional Studies

Many retrospective studies examined apo A-I levels in patients with coronary artery disease compared with matched controls. All studies found that apo A-I levels were substantially lower in patients with coronary artery disease [43-80]; the only studies that did not were small studies lacking adequate statistical power [81-83]. However, only some of these investigators tried to determine the relative predictive value of apo A-I levels compared with HDL cholesterol levels, and the results were conflicting. Although some studies concluded that apo A-I concentration is a better predictor of coronary artery disease risk than HDL cholesterol concentration [46, 47, 49, 50, 52, 56, 57, 61, 62, 64, 68-73, 76, 77], others reported that HDL cholesterol is as good as or better than apo A-I in discriminating persons with coronary artery disease from controls [53, 55, 59, 63, 67, 78-81, 83].

Studies in Children

Another method to assess the predictive value of apolipoproteins is quantitation of levels in children and analysis of the association with premature coronary artery disease in their parents or grandparents (Table 2). In one study, apo A-I levels in children were predictive of premature coronary artery disease in their parents (P = 0.04) and were superior to HDL cholesterol levels as predictors [84]. In another study, apo A-I levels were significantly lower in children of parents with premature coronary artery disease (P < 0.05) but were not more predictive than HDL cholesterol levels [85]. In a third study, apo A-I levels were not different in 11- to 19-year-old children of parents with premature coronary artery disease compared with controls [86]. In a fourth study, apo A-I levels in children were not predictive of premature coronary artery disease in their grandparents [87]. Finally, apo A-I levels were significantly lower in adolescent children of parents who had myocardial infarction before the age of 55 years compared with controls (P < 0.01) but were no more predictive than HDL cholesterol [88].

Prospective Studies

Six prospective studies investigated the value of plasma apo A-I concentrations in predicting coronary artery disease [89-94] (Table 3). All six found a significant inverse correlation between apo A-I levels and coronary artery disease incidence. Only two of these studies included women. In one study, lower apo A-I levels were found only in men with coronary artery disease and not in women [90]. In the other study, which was limited only to women, the correlation of apo A-I levels with coronary artery disease was only significant after adjustment for total cholesterol and body mass [92]. Three of these prospective studies did not determine HDL cholesterol levels [90, 92, 93] and thus could not directly compare the predictive value of apo A-I with that of HDL. Of the prospective studies that directly compared apo A-I with HDL cholesterol, all concluded that apo A-I was no more predictive of coronary artery disease than HDL cholesterol. In the first, a very small study, coronary artery disease developed in 12 of 28 participants [89]. In the second, involving 246 male physicians with a new myocardial infarction and 246 matched controls [91], apo A-I levels had a strong inverse correlation with myocardial infarction but were no more predictive than HDL cholesterol. The most recent was a prospective study of 21 520 men in which 229 men who died of coronary artery disease were matched to 1145 controls [94]. Apolipoprotein A-I levels were significantly lower in affected men than in controls, but in the subset of 62 affected men and 237 controls for whom HDL cholesterol levels were available, apo A-I levels were no more predictive of coronary artery disease than were HDL cholesterol levels. Therefore, quantitation of apo A-I has no role in population-based screening for premature coronary artery disease. However, no prospective studies of the predictive value of apo A-I levels for recurrent cardiovascular events or response to interventional therapy in patients with coronary artery disease have been reported. More prospective studies in defined populations are needed to determine whether quantitation of apo A-I may have a role in the clinical management of patients with or potentially at risk for development of premature coronary artery disease.

Apolipoprotein B

Apolipoprotein B is the major apolipoprotein in chylomicrons, VLDL, intermediate-density lipoprotein, and LDL (see Figure 1). It serves an essential structural role in these lipoprotein particles; the genetic inability to secrete apo B causes the absence of these lipoproteins in plasma [95]. Mutations in the apo B gene can cause low levels of apo B and LDL cholesterol and may be associated with protection from premature coronary artery disease [96]. Apolipoprotein B also acts as a ligand for the LDL receptor, mediating the cellular uptake and degradation of LDL [4]. Only one molecule of apo B exists per lipoprotein particle, and thus the quantity of apo B in fasting plasma is a measure of the number of LDL and VLDL particles. In fact, the plasma levels of "non-HDL cholesterol," which includes both LDL and VLDL, are correlated with plasma apo B levels [97, 98]. However, in contrast to the constant 1:1 molar ratio of apo B per LDL and VLDL particle, the amount of cholesterol in these lipoproteins varies widely. Therefore, plasma apo B levels may be a better assay of the concentration of atherogenic lipoprotein particles than are LDL cholesterol or non-HDL cholesterol levels [99, 100].

Cross-Sectional Studies

Whether apo B levels are a more sensitive predictor of coronary artery disease risk than are total or LDL cholesterol levels has recently been debated [28, 97, 100]. Multiple retrospective, cross-sectional studies comparing apo B levels in patients with coronary artery disease with controls have been reported [47,49-52,54,55,57,59,61-63,67-69,71-74,76,-78,80,81,83,98,101] and have consistently found an association of plasma apo B concentrations with increased risk for premature coronary artery disease. Whether apo B levels are more predictive of coronary artery disease than are total or LDL cholesterol levels has not been solved definitively by these cross-sectional studies alone. However, most of the recent studies that specifically addressed this issue found that apo B levels are better correlated with the presence of coronary artery disease than are total or LDL cholesterol levels [68, 69, 72, 73, 77, 80, 98]. Furthermore, a cross-sectional study in patients who had coronary artery bypass graft surgery determined that apo B concentration was a better discriminator than LDL cholesterol concentration in predicting recurrent atherosclerotic disease in bypass grafts 10 years after surgery [102].

Studies in Children

Apolipoprotein B levels have been measured in children and assessed for their ability to predict premature coronary artery disease in their families (see Table 2). In one study, neither total apo B levels nor LDL cholesterol levels were useful in discriminating between children having a parent with premature coronary artery disease and those without a parental history of coronary artery disease [84]. In another study, apo B levels were substantially higher in children of parents with premature coronary artery disease but were not more predictive than total cholesterol levels [85]. In a third study, apo B levels were not different in 11- to 19-year-old children of parents with premature coronary artery disease compared with controls [86]. In a fourth study, apo B levels in children were predictive of premature coronary artery disease in their grandparents, but total and LDL cholesterol were not measured for comparison [87]. Finally, apo B levels were not substantially different in adolescents of parents with myocardial infarction before the age of 55 years compared with controls [88].

Prospective Studies

There are only five prospective studies of apo B concentration and coronary artery disease incidence [90-94] (see Table 3). The first was a Finnish study in patients with high cholesterol levels and a very high rate of coronary artery disease. The investigators matched 92 persons who died of coronary artery disease to 92 controls; they also matched for total cholesterol [mean, 7.42 mmol/L (287 mg/dL)]. Apolipoprotein B levels were not considerably different in patients with coronary artery disease than in controls; however, these results may be difficult to extrapolate to other populations because of the exceptionally high cholesterol levels and coronary artery disease rates in these persons. A prospective case–control study based on the Physicians' Health Study population found that plasma concentrations of apo B were highly correlated with coronary artery disease risk and were more predictive than total cholesterol levels but provided no additional discriminating power than the total cholesterol/HDL cholesterol ratio [91]. However, the samples were not taken after participant fasting and no LDL cholesterol levels were determined for direct comparison with apo B levels. Another prospective study found that apo B levels were substantially associated with premature coronary artery disease by univariate analysis but were not more predictive than total cholesterol levels by multivariate analysis [93]. A prospective study in women found that although there was a positive correlation between apo B levels and coronary artery disease, the correlation with total cholesterol levels was stronger [92]. A recent large prospective study in 21 520 men found that apo B levels were considerably higher in the 229 men who died of coronary artery disease than in the 1145 matched controls [94]. After multiple logistic regression, the association of apo B was independent of total cholesterol and triglyceride levels and, overall, the apo B level was the most strongly associated with coronary artery disease risk. A 10% decrease in apo B was associated with a 22% increase in coronary artery disease mortality rate. More population-based prospective studies are needed to answer the question definitively of whether apo B levels are more predictive of premature coronary artery disease than are traditional lipid parameters. In addition, prospective studies of apo B levels in targeted populations of patients with premature coronary artery disease or family histories of premature coronary artery disease are required.

Familial Combined Hyperlipidemia

A growing body of data indicates that a subset of persons with premature coronary artery disease but relatively normal LDL cholesterol levels have substantially increased apo B concentrations. Disproportionately elevated plasma levels of apo B are characteristic of a syndrome called familial combined hyperlipidemia. This condition often is associated with other plasma lipid abnormalities, including mild to moderate hypertriglyceridemia and low HDL cholesterol levels [103-108]. The plasma LDL cholesterol concentrations in these persons often are only modestly elevated. A related condition is hyperapobetalipoproteinemia [109, 110], in which levels of plasma apo B are increased but LDL cholesterol levels are normal. Biochemically, these conditions are characterized by increased amounts of small dense LDL in the plasma [106, 108, 111]. In some kindreds, the metabolic basis of these disorders is overproduction of apo B-containing lipoproteins by the liver [112-115]. These syndromes, for which an elevated plasma apo B level is an important diagnostic criterion, are associated with a substantially increased risk for premature coronary artery disease [103-105, 109, 110]. Estimates show that at least 10% of myocardial infarction survivors younger than 60 years have familial combined hyperlipidemia [103]. Quantitation of plasma apo B concentrations may be one method to identify these persons whose lipid parameters often are not especially abnormal. Patients with elevated fasting triglyceride levels or decreased HDL cholesterol levels may have familial combined hyperlipidemia. The plasma apo B concentration may be useful in differentiating these persons from those with other causes of dyslipoproteinemia, such as familial hypertriglyceridemia, which are not associated with increased risk for premature coronary artery disease [15, 106]. A recent study suggested that an elevated apo B level associated with elevated triglyceride levels might be more atherogenic than an increased apo B level alone [116].

Lipoprotein(a)

Background

Lipoprotein(a) is an LDL-like lipoprotein containing a unique apolipoprotein called apo(a) (reviewed in references 117-121). Apolipoprotein(a) is similar to plasminogen in structure [122] and may interfere with plasminogen activation (reviewed in reference 123). Plasma levels of Lp(a) vary across a 1000-fold range, and their distribution is skewed to the left in most populations [117]. However, considerable ethnic and racial differences exist in the distribution of Lp(a) levels. For example, Sudanese blacks [124, 125] and African-Americans [126-128] have higher Lp(a) levels and a more bell-shaped Lp(a) distribution than do whites, whereas Asians have lower median Lp(a) levels [124, 125, 129].

Plasma Lp(a) levels are genetically determined [117, 130]. The apo(a) gene accounts for more than 90% of the variation in plasma Lp(a) concentrations [131]. The apo(a) protein has a variable size polymorphism related to the number of plasminogen-like kringle 4 repeats present in the apo(a) gene [132]. A strong inverse correlation exists between the number of kringle 4 repeats in the apo(a) gene and the plasma Lp(a) concentration [117, 132]. Variation in the number of kringles as assessed by apo(a) genotyping may account for approximately 70% of the variation in plasma Lp(a) concentrations [131]. The physiologic basis for this association is an effect of kringle number on the rate of hepatic apo(a) production [133]. The apo(a) genotype may influence coronary artery disease risk independent of its effect on plasma Lp(a) concentrations [125, 129]. Lipoprotein(a) concentrations also vary substantially within each apo(a) isoform class [117, 132]. Variation in the production rate of Lp(a) is the primary determinant of variation in plasma Lp(a) levels among persons with the same apo(a) phenotype but different Lp(a) levels [134]. These differences in Lp(a) production rates may be related to sequence polymorphisms in the apo(a) gene distinct from the size polymorphism [135]. Of the nongenetic factors affecting Lp(a) levels, the best documented are end-stage renal disease [136] and the nephrotic syndrome [137]. Lipoprotein(a) is also an acute-phase reactant; its levels increase in acute inflammatory states and after myocardial infarction and surgical procedures [37].

Cross-sectional Studies

Lipoprotein(a) was first reported to be associated with premature coronary artery disease in 1975 [138]. Since then, several retrospective studies comparing patients who had coronary artery disease with controls showed that the risk for myocardial infarction or angiographically documented coronary artery disease correlated with plasma Lp(a) concentration [72, 75, 78-80, 129, 139-147]. Many of the earlier studies were small and the highly skewed distribution of Lp(a) may have confounded their statistical interpretation. Nevertheless, recent larger cross-sectional studies supported the independent association of Lp(a) with premature coronary artery disease. One study estimated that the population- attributable risk for coronary artery disease due to elevated Lp(a) levels was approximately 25% in men younger than 60 years [142]. In 180 patients with premature coronary artery disease (150 men, 30 women), 17% had Lp(a) levels greater than the 90th percentile for a control population (39 mg/dL) [145]. Family studies in 102 of these probands indicated that Lp(a) excess was the most frequent familial lipoprotein disorder found in this cohort of persons with premature coronary artery disease [78]. Lipoprotein(a) levels may account for much of the familial predisposition to premature coronary artery disease that cannot be accounted for by other known risk factors, including LDL and HDL cholesterol levels [72].

In two cross-sectional studies of patients with heterozygous familial hypercholesterolemia (who have elevated LDL cholesterol levels due to a defect in the LDL receptor and are at generally increased risk for premature coronary artery disease), Lp(a) levels were considerably higher in patients with coronary artery disease than in those without coronary artery disease [148, 149]. Therefore, Lp(a) levels may predict relative coronary artery disease risk in persons with elevated LDL cholesterol. A recent cross-sectional study found that Lp(a) levels were independently predictive of coronary atherosclerosis in 130 cardiac transplant recipients (P = 0.0006) [150]. Three studies evaluated the predictive value of Lp(a) levels in patients with established coronary artery disease after therapeutic intervention. In a cross-sectional study of 167 patients who had coronary artery bypass surgery, Lp(a) levels were independently correlated with coronary bypass saphenous vein graft stenosis (P = 0.002) [151]. Two studies that evaluated the predictive value of Lp(a) levels in restenosis after percutaneous transluminal coronary angioplasty produced conflicting results [152, 153]. In addition to coronary artery disease, retrospective studies showed that Lp(a) levels are associated with clinical cerebrovascular atherosclerosis [147, 154-156] and carotid artery wall thickness in asymptomatic persons [157].

Studies in Children

Five studies evaluated the relation between Lp(a) levels in children and adolescents and coronary artery disease in their parents or grandparents (see Table 2). A study in 1486 18-year-old men found that Lp[a] levels were significantly higher in those with a parent who had premature coronary artery disease compared with controls (P < 0.05) [158]; parents of 18-year-old men with Lp(a) levels greater than 25 mg/dL had a 2.5-fold higher incidence of coronary artery disease. A second study in 42 children with a parent who had a myocardial infarction before the age of 45 years found that Lp(a) levels were significantly higher than in controls (P < 0.01) [85]. Another study in 8 to 17 year olds determined that white children with parental myocardial infarction had significantly increased levels of Lp(a) (P < 0.01), but no association between Lp(a) and parental coronary artery disease was seen in black children [128]. A fourth study in 322 children aged 11 to 19 years found that 42% of the children with parental coronary artery disease had Lp(a) levels greater than 30 mg/dL compared with only 19% of the controls (P < 0.05) [86]; Lp(a) levels were more strongly predictive of coronary artery disease in parents than were LDL cholesterol, HDL cholesterol, apo A-I, and apo B levels. Finally, in 8-to 12-year-old children, a significant association was found between Lp(a) levels and number of grandparents with coronary artery disease (P < 0.01) [87]. These data strongly support the concept that Lp(a) level is an inherited risk factor for premature coronary artery disease.

Prospective Studies

Six prospective studies of Lp(a) and coronary artery disease incidence have been reported (see Table 3) [9294, 159-161]. One case–control study followed 776 healthy 50-year-old Swedish men (at baseline) for 6 years [159]. The 26 men in whom coronary artery disease developed had higher Lp(a) levels than did 109 matched controls (P = 0.01); an Lp(a) level in the highest quintile was associated with twice the coronary artery disease rate than in other study participants. The effect of Lp(a) level was independent of total cholesterol and triglyceride levels, but HDL cholesterol levels were not measured; furthermore, smoking was not controlled for between participants and controls. Another prospective study in 1332 healthy men from Iceland followed for 8.6 years found that the mean Lp(a) level was higher in the 104 men with coronary artery disease than in controls (P < 0.05) [93]; however, HDL and LDL cholesterol levels were not determined in that study. A third prospective study in 3634 British women found a higher mean Lp(a) level in the 51 participants with coronary artery disease than in controls, but because of the skewed Lp(a) distribution, the P value was only 0.15 [92]. A fourth study was a small nested case–control study of the Helsinki Heart Study participants (138 patients with coronary artery disease and 130 controls): In a stepwise logistic regression analysis, Lp(a) levels were not significantly associated with coronary artery disease incidence (P = 0.365) [160]. However, this study cohort consisted of persons selected for increased non-HDL cholesterol, and the use of gemfibrozil may have modified the association between Lp(a) and coronary artery disease. A fifth prospective study of 14 916 healthy male American physicians compared Lp(a) levels in 296 men with myocardial infarction and 296 matched controls and found no evidence of association between Lp(a) level and the risk for future myocardial infarction (P = 0.73) [161]. The sixth prospective study was in 21 520 British men and found that the 229 men who died of coronary artery disease had substantially higher Lp(a) levels than did the 1145 matched controls [94]. However, this association was much weaker than that of apo B, and a 10% increase in Lp(a) was associated with only a 3% decrease in coronary artery disease risk. The most recently published prospective study was a nested case–control study of the 3806 hypercholesterolemic participants in the Lipid Research Clinics Coronary Primary Prevention Trial [162]. Lipoprotein(a) levels were modestly but significantly higher in a cohort of 233 men who subsequently developed coronary artery disease during the course of the study compared with a group of 390 matched controls (P < 0.02). In addition to the studies in the general population, two prospective studies in diabetic persons found no correlation of Lp(a) concentrations with coronary artery disease incidence [163, 164]. Overall, the epidemiologic data indicate that Lp(a) levels are strongly correlated with premature coronary artery disease incidence in persons with elevated LDL cholesterol levels or a family history of premature coronary artery disease but may not be useful to predict coronary artery disease risk in the general population.

Other Apolipoproteins

Apolipoprotein A-II

Of the remaining apolipoproteins, the most intensively investigated in relation to atherosclerosis have been apo A-II and apo E. Like apo A-I, apo A-II is a major structural protein in HDL. However, in contrast to apo A-I, complete synthetic deficiency of apo A-II does not cause low HDL levels and premature coronary artery disease [165]. Several retrospective studies have not shown a consistent correlation between plasma apo A-II concentrations and coronary artery disease risk [62, 68, 76, 77, 166]. One recent prospective study established that the inverse association of plasma apo A-II levels with myocardial infarction was statistically much weaker (P = 0.07) than that of plasma apo A-I levels (P = 0.0001) [91], and a second large prospective study confirmed this conclusion [94]. In fact, recent studies in transgenic mice suggest that apo A-II negatively offsets the protective effect of apo A-I [167] or may be atherogenic rather than antiatherogenic [168]. The quantitation of plasma apo A-II has little clinical utility.

Apolipoprotein E

Apolipoprotein E is found in association with apo B-containing lipoproteins and HDL and serves as a ligand for binding to the LDL receptor and to the putative lipoprotein remnant (apo E) receptor [169]. Two retrospective case–control studies reported no correlation between apo E levels and coronary artery disease [76, 170]; no prospective studies have been reported. Although we know much about the role of apo E in lipoprotein metabolism, its direct relation to atherosclerosis remains uncertain [171] and there is little clinical rationale for quantifying plasma apo E concentrations. However, another aspect of apo E may be clinically important in the prediction of coronary artery disease risk. Apolipoprotein E is genetically polymorphic, and the three isoforms, E2, E3, and E4, are inherited in an autosomal codominant manner. Apolipoprotein E3 is the most common isoform, whereas apo E2 has an allele frequency of about 7% and apo E4 has an allele frequency of 14% [172]. Homozygosity for apo E2 is associated with type III hyperlipoproteinemia (familial dysbetalipoproteinemia). Patients with this disorder have an increased risk for premature atherosclerosis [173], and confirmation of the apo E2/2 phenotype or genotype is important for the definitive diagnosis of this condition. Heterozygosity for apo E4 is associated with increased plasma LDL cholesterol concentrations [172]. However, recent studies suggest that the apo E4 allele is associated with an increased risk for premature coronary artery disease independent of LDL cholesterol level [174-176]. Therefore, determination of the apo E phenotype or genotype may have independent predictive value regarding coronary artery disease risk.

Clinical Recommendations

Apolipoprotein A-I

Recommendations for the use of apolipoprotein quantitation in clinical practice are summarized in Table 4. In most situations, measurement of plasma apo A-I levels does not affect patient management substantially. The new guidelines of the Adult Treatment Panel II of the National Cholesterol Education Program recommend routine screening of HDL and total cholesterol levels in all adults [177]. However, no guidelines exist for managing low HDL cholesterol levels [178]. Theoretically, quantitation of apo A-I could be useful in persons with low HDL levels; for example, a normal apo A-I concentration in this setting could provide an argument against pharmacologic therapy to increase the HDL cholesterol level. However, prospective studies are needed in the subgroup of persons with low HDL cholesterol to determine their risk for premature coronary artery disease and the utility of apo A-I quantitation in this cohort. Some epidemiologic evidence suggests that plasma concentrations of the HDL subclass containing apo A-I but not apo A-II (called LpA-I) may be more predictive of coronary artery disease risk than are plasma HDL cholesterol or total apo A-I concentrations [70, 79, 91]. In the future, quantitation of apo A-I-containing HDL subclasses may have a clinical role in coronary artery disease risk assessment [179].

Apolipoprotein B and Lipoprotein(a)

Our approach to apo B and Lp(a) quantitation in clinical practice for primary and secondary prevention of coronary artery disease is summarized in Figures 2 and 3. The new guidelines of the Adult Treatment Panel II of the National Cholesterol Education Program [177] recommend that patients with established coronary artery disease Figure 2 who have LDL cholesterol levels greater than 3.4 mmol/L (130 mg/dL) and eat an adequate diet should be considered for pharmacologic therapy to lower the LDL cholesterol level. Quantitation of apo B levels in this group provides little additional information that influences the clinical approach, but measurement of Lp(a) could affect drug selection. For persons with coronary artery disease who are not candidates for pharmacologic therapy based on LDL cholesterol level less than 3.4 mmol/L (130 mg/dL), fasting determination of both apo B and Lp(a) levels should be done. A subset of these patients is likely to have elevation of apoB, Lp(a), or both, and therefore they may be candidates for drug therapy despite having normal LDL cholesterol levels. A similar approach may be appropriate for patients with no evidence but a family history of premature coronary artery disease (Figure 3). When the LDL cholesterol is 3.4 to 4.9 mmol/L (130 to 190 mg/dL) with consumption of an optimal diet and no compelling indications exist to begin treatment with drugs, apo B and Lp(a) levels should be measured as additional potential familial risk factors; elevation of either could influence the decision to initiate drug therapy to decrease LDL cholesterol levels.

View larger version (26K): [in this window] [in a new window]   Figure 2. Algorithm to quantitate lipids and apolipoproteins in patients with coronary artery disease. apo = apolipoprotein; CAD = coronary artery disease; HDL = high-density lipoprotein; LDL = low-density lipoprotein; Lp(a) = lipoprotein(a).

View larger version (31K): [in this window] [in a new window]   Figure 3. Algorithm to quantitate lipids and apolipoproteins in patients with a family history of premature coronary artery disease. apo = apolipoprotein; CAD = coronary artery disease; HDL = high-density lipoprotein; LDL = low-density lipoprotein; Lp(a) = lipoprotein(a).

Apolipoprotein B-containing lipoproteins are heterogeneous in apolipoprotein composition [180, 181], and recent data suggest that the different types of apo B-containing lipoprotein particles may vary in their atherogenic potential [79, 180-184]. Once validated assays are available, quantitation of specific apo B-containing lipoprotein particles may provide more predictive power for coronary artery disease risk than do total plasma apo B levels.

Therapeutic Considerations

Is the use of drugs to decrease elevated plasma apo B or Lp(a) levels beneficial even if LDL cholesterol concentrations are not exceptionally elevated? No prospective data are available to address this question. One secondary prevention trial in patients with established coronary artery disease used a plasma apo B concentration of more than 125 mg/dL, rather than an increased LDL cholesterol level, as the biochemical criterion for entry into the study [23]. Patients who were treated aggressively with lipid-lowering therapy experienced less disease progression and fewer cardiovascular events than did controls treated with conventional therapy. Although many of these patients also had increased LDL cholesterol concentrations, this suggests that intervention based on apo B levels may be clinically beneficial.

Apolipoprotein B levels can be decreased successfully with nicotinic acid (niacin), bile acid sequestrants (cholestyramine, colestipol), fibrates (gemfibrozil, fenofibrate), and hydroxymethylglutaryl coenzyme A reductase inhibitors (lovastatin, pravastatin, simvastatin, fluvastatin). However, the only drug studies that reported reduction in Lp(a) levels have used niacin, either alone [185] or in combination with other drugs [23, 186]. Therefore, patients requiring drug treatment for elevated LDL cholesterol levels who also have elevated Lp(a) levels should be considered for niacin therapy. Preliminary reports show that estrogen replacement therapy in women after menopause may lower plasma Lp(a) concentrations or at least prevent their increase [187, 188]. For patients with coronary artery disease who have persistently elevated LDL cholesterol and Lp(a) levels even with maximal therapy, LDL apheresis is effective in decreasing LDL and Lp(a) levels [189]. More clinical trials of the benefit of apo B and Lp(a) reduction are needed before specific clinical recommendations can be made.

Abbreviations

apo A-I: apolipoprotein A-I

apo B: apolipoprotein B

apo C-II: apolipoprotein C-II

apo E: apolipoprotein E

Lp(a): lipoprotein(a)