http://informahealthcare.com/lab ISSN: 1040-8363 (print), 1549-781X (electronic) Crit Rev Clin Lab Sci, 2013; 50(6): 163–171 ! 2013 Informa Healthcare USA, Inc. DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol: Diagnosis and cardiovascular risk management Tjerk de Nijs1, Allan Sniderman2, and Jacqueline de Graaf1 1
Department of General Internal Medicine, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands, and 2Mike Rosenbloom Laboratory for Cardiovascular Research, McGill University, Royal Victoria Hospital, Montreal, Canada Abstract
Keywords
The most recent guidelines released by the EAS/ESC and the Canadian Cardiovascular Society (CCS) retain low-density lipoprotein cholesterol (LDL-C) as the primary measure of the atherogenic risk of the apolipoprotein B (apoB) lipoproteins and the primary target of LDL-C lowering therapy. Both organizations endorse non-high-density lipoprotein cholesterol (nonHDL-C) and apoB as ‘‘alternate/secondary’’ targets, but neither group offers evidence supporting the continued preference of LDL-C as the primary target over non-HDL-C and apoB. Further, both suggest that non-HDL-C and apoB more or less measure the same thing and, therefore, are essentially interchangeable. But what is the evidence that LDL-C should remain the primary target, and are apoB and non-HDL-C mirror images of one another? Furthermore, are estimation of risk and establishment of treatment targets the only relevant issues, or is diagnosis also an essential objective? These are the questions this article will address. Our principal objectives are: (1) to clarify the differences between LDL-C, non-HDL-C, and apoB and to distinguish what they measure; (2) to summarize the evidence relating to LDLC, non-HDL-C, and apoB as predictors of cardiovascular risk and as targets for treatment; and (3) to demonstrate that diagnosis of atherogenic dyslipoproteinemias should be a fundamental clinical priority.
ApoB, atherosclerosis, apolipoproteins, CVD, lipoproteins, non-HDL-C History Received 22 July 2013 Revised 7 September 2013 Accepted 19 September 2013 Published online 29 November 2013
Abbreviations: apoB: apolipoprotein B; APP: application; CCS: Canadian Cardiovascular Society; EAS: European Atherosclerosis Society; ESC: European Society of Cardiology; FCHL: familial combined hyperlipidemia; FDB: familial dysbetalipoproteinemia; FHTG: familial hypertriglyceridemia; HDL: high density lipoprotein; HR: hazard ratio; IDL: intermediate density lipoprotein; LDL-C: low-density lipoprotein cholesterol; LDL: low-density lipoprotein; NonHDL-C: non-high density lipoprotein cholesterol; TG: triglyceride; VLDL: very low density lipoprotein
Introduction Cardiovascular disease (CVD) is now a worldwide affliction and dyslipidemia is one of the most important and most modifiable causes. The major organizations responsible for creating CVD guidelines state that low-density lipoprotein cholesterol (LDL-C) is the best measure of the proatherogenic risks of low-density lipoproteins. However, recent evidence indicates that both non-high-density lipoprotein cholesterol
Referee: Dr. John Burnett, Clinical Professor in the School of Medicine & Pharmacology, University of Western Australia; Consultant Medical Biochemist in the Department of Clinical Biochemistry at PathWest Laboratory Medicine WA, Royal Perth Hospital (RPH). Address for correspondence: Prof. Dr Jacqueline de Graaf, Department of General Internal Medicine 463, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail: [email protected]
(non-HDL-C) and apolipoprotein B (apoB) are more accurate indices of cardiovascular risk than LDL-C. While the importance of LDL-C is well understood and the evidence supporting the clinical use of LDL-C is extensive, little is known about non-HDL-C and apoB. Furthermore, in today’s clinical practice, where statins are easily accessible, diagnosis of lipid disorders is often neglected. Part of this may be explained by the fact that the Frederickson classification of dyslipoproteinemia is rather complex and often requires additional, expensive, and complex measurements of lipoproteins lipids. With the introduction of apoB as a biomarker, diagnosis of dyslipoproteinemia in clinical practice is now simple. Accordingly, this article will focus on the strengths and limitations of apoB and non-HDL-C compared to LDL-C in cardiovascular risk evaluation and will evaluate how apoB facilitates diagnosis of dyslipoproteinemia to achieve complete cardiovascular risk management.
20 13
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
REVIEW ARTICLE
164
T. de Nijs et al.
What do LDL-C, non-HDL-C, and apoB measure?
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
LDL-C LDL-C is a measure of the mass of cholesterol within lowdensity lipoprotein (LDL) particles. LDL particles were originally defined as those lipoproteins with a density between 1.019 and 1.063 g/ml; however, more recently, the definition has been expanded to encompass all such particles between 1.006 and 1.063 g/ml. That is, the present conventional definition includes both classical intermediate density lipoprotein particles (IDL) and classical LDL particles. An LDL-C value calculated using the Friedewald equation includes all the cholesterol within apoB particles1 between 1.006 and 1.063 g/ml, as would an LDL-C value measured directly by any commercially available method. Relying on LDL-C assumes that cardiovascular risk due to LDL is equal to the mass of cholesterol within all LDL particles. Since the mass of cholesterol within individual LDL particles can vary substantially, relying on LDL-C assumes that the amount of cholesterol that is deposited within the arterial wall is independent from the number of particles. That is, this cholesterol model predicts that individuals with a small number of cholesterol-enriched particles will deposit more cholesterol within their arteries than individuals with a lower LDL-C but a greater number of cholesterol-depleted particles. However, there is no evidence to suggest that this is the case and compelling evidence indicates that it is not. The cholesterol within apoB particles enters the arterial wall by way of those apoB particles. The primary determinant of the number of apoB particles that enter and are trapped within the arterial wall and, therefore, the mass of cholesterol deposited within the arterial wall is the number of apoB particles within the lumen of the artery. Moreover, cholesterol is not the only injurious substance within the LDL particle. ApoB itself can stimulate an inflammatory response as well as angiogenesis2,3. ApoB There are two types of apoB in plasma: (1) apoB-100, which is synthesized by the liver and provides structural integrity to VLDL, LDL, and Lp(a) particles, LDL particles to which an apo(a) has been attached, and (2) apoB-48, which is synthesized by the intestine and encircles and provides structural integrity to chylomicron and chylomicron remnant particles4. After a meal, plasma triglycerides rise and fall as apoB-48 chylomicron particles secreted by the intestine enter the circulation and are rapidly cleared by the liver. Nevertheless, even at peak postprandial levels, there are so few apo-B48 particles compared to the number of apo-B100 particles that, for practical purposes, apoB-100 accounts for virtually all measured apoB. Thus, apoB equals5 apoB-100. VLDL transport triglyceride from the liver into adipose tissue and skeletal muscle. As the triglyceride within them is progressively hydrolyzed, the VLDL particles (d51.006 g/ ml) become progressively smaller and cholesterol-enriched. They are transformed sequentially into IDL particles (d 1.006–1.019 g/ml) containing roughly equal quantities of cholesterol and triglycerides, and then into LDL particles (d 1.019–1.063 g/ml), which are even smaller and more enriched
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
with cholesterol. IDL and LDL particles are unquestionably atherogenic. In most instances, even in hypertriglyceridaemia (except for familial dysbetalipoproteinemia: see next paragraph) or severe hypertriglyceridemia, LDL particles make up more than 90% of the total plasma apoB6–8. VLDL particles are much larger than LDL particles and, therefore, enter the arterial wall much less easily than LDL particles. Moreover, in almost all individuals, there are nine times as many LDL particles as VLDL particles. In other words, LDL particles are much more atherogenic than VLDL particles because there are many more of them and they are much smaller, explaining why LDL and not VLDL drives the atherogenic risk due to apoB lipoproteins. In summary, apoB equals the sum of the VLDL, IDL, LDL and Lp(a) particles in plasma, i.e. all potentially atherogenic lipoprotein particles (Figure 1). The utility of using the number of apoB particles as a marker of cardiovascular risk is based on the fact that, although each apoB particle contains one molecule of apoB, the mass of cholesterol within the apoB particles can vary substantially. When this occurs, LDL-C does not equal LDL particle number. Cholesterol-enriched LDL particles are larger and more buoyant than normal LDL particles, whereas cholesterol-depleted LDL particles are smaller and denser than normal LDL particles. When LDL particles are normal in composition, LDL-C and apoB are concordant and LDL-C is an accurate measure of the number of LDL particles. However, when cholesterol-enriched or cholesterol-depleted LDL particles are dominant, LDL-C and apoB are discordant and LDL-C does not provide an accurate measure9 (Figure 2). Fasting samples are not required to measure apoB because there are so few apo-B48 particles. The measurement of apoB is standardized and the American Association of Clinical Chemistry has determined that apoB can be measured more accurately in routine clinical laboratories than either LDL-C or non-HDL-C10,11. The perception of difficulties in the measurement of apoB relates to research studies, which used in-house, non-standardized methods or analyzed samples that have been preserved for extended periods of time. Non-HDL-C Non-HDL-C is the arithmetic sum of the cholesterol in VLDL and LDL particles and therefore, by definition, it equals all the atherogenic cholesterol in lipoproteins. Put this way, apoB and non-HDL-C appear to be mirror images of each other, with apoB measuring all the atherogenic particles and non-HDL-C measuring all the atherogenic cholesterol. This may explain why many argue that nonHDL-C is an acceptable functional surrogate for apoB. Given that no extra charge is required to calculate nonHDL-C, these individuals may conclude that non-HDL-C is superior to apoB. However, similar to LDL-C and apoB, when VLDL and LDL particles are either cholesterolenriched or cholesterol-depleted – that is, when apoB and non-HDL-C are discordant – non-HDL-C is not an accurate measure of the number of total apoB particles. Therefore, non-HDL-C does not necessarily equal total apoB particle number, as will be discussed further below.
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol
165
Figure 1. Every apoB lipoprotein particle contains one molecule of apoB – so ApoB equals the sum of the VLDL, IDL, and LDL particles in plasma, i.e. all the potential atherogenic lipoprotein particles. Triglyceride-rich VLDL particles and LDL particles differ in composition: VLDL particles can contain variable amounts of triglyceride (TG) or cholesterolesters (CE). LDL particles can be cholesterol-depleted or cholesterol-enriched. Remnant particles when present are in IDL density range. The (blue) circle reflects one molecule of apoB per apoB lipoprotein particle.
Figure 2. Is an LDL cholesterol level of 3.0 mmol/l high risk or not? It depends on the number of atherogenic LDL particles. In the subject on the left side, numerous small, dense LDL particles result in an LDL-C of 3 mmol/L. As each LDL particle has one apoB, apoB will be high (41.2 g/L – 90th percentile of the population). In the subject on the right side, normal, relatively large, buoyant LDL.
LDL-C, non-HDL-C, and apoB in cardiovascular risk prediction The conventional approach to determining which marker of risk is superior to another is to compare the predictive power of each method in all subjects in a study and then integrate the resulting values in a meta-analysis. This was the approach of
Sniderman et al., who demonstrated a hierarchy of accuracy in which non-HDL-C was found to be superior to LDL-C and apoB was superior to non-HDL-C. Importantly, this study showed that these differences were clinically as well as statistically significant12. In contrast, the Emerging Risk Factor Collaboration (ERFC) determined that non-HDL-C and apoB
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
166
T. de Nijs et al.
are equivalent markers of risk13,14. While this result garnered significant attention, the ERFC also showed that LDL-C and total cholesterol were equal to non-HDL-C and apoB. That is, the ERFC found that TC was equal to any other measure of the lipoprotein-related risk of vascular disease, which contradicts such a large number of studies that, in our view, it is not credible. Despite the refined design and mathematical analyses that entered into this study, it may be that, since such a large number of studies were evaluated, many featured inadequate design, methods, and execution15. The ERFC is a participantlevel analysis based on actual individual data, which means that a new population is constructed from variable population subsections of all studies that chose to participate. The strength of participant-level analysis is that it is based on individual rather than summary data, as in a meta-analysis. However, in contrast to meta-analyses, there is no commitment to include all of the relevant studies. Nevertheless, neither meta-analysis nor participant-level analysis may be the most appropriate method to resolve the dispute. The reason that some studies show apoB to be superior to non-HDL-C, whereas others do not, is that the differences in predictive power can be apparent only in discordant subgroups – the subgroups in which apoB particles are either cholesterol-enriched or cholesterol-depleted. In groups for which LDL-C and apoB are concordant, both markers will predict risk equally and, since the concordant groups are much larger than the discordant ones, any potential difference in the two markers will be diluted by the much larger number of subjects in the concordant group. By restricting the comparison of the predictive powers of these measures to the discordant groups, it is possible to clearly determine whether it is the mass of cholesterol or the number of apoB particles that matters the most9. Discordance analyses have shown striking superiority for LDL particle number over LDL-C16,17 and of apoB over nonHDL-C18. Accordingly, discordance analysis appears to be the most appropriate way to demonstrate the limitations of the conventional approach to risk evaluation.
Pathophysiology/mechanism to explain the role of non-HDL-C and apoB in CV prediction The commonsense and, therefore, commonly-offered explanation for the superiority of non-HDL-C over LDL-C is that non-HDL-C includes the cholesterol in VLDL as well as LDL and thus represents the total atherogenic cholesterol burden. But the evidence that VLDL cholesterol is atherogenic is weak to non-existent. Despite the fact that fibrates reduce triglycerides and VLDL-C, there is no evidence suggesting that these changes reduce cardiovascular risk. Indeed, if VLDL-C ‘explained’ why non-HDL-C was superior to LDL-C, the hazard ratio (HR) of VLDL-C would have to be substantially higher than to the HR of LDL-C19. That is, VLDL-C would have to be more dangerous than LDL-C, and there is no evidence that this is the case. On the contrary, there is a good reason to believe that the superior predictive power of non-HDL-C over LDL-C is based on the fact that non-HDL-C is a more accurate marker of LDL particle number than LDL-C. As illustrated in Figure 3, cholesterol-depleted LDL particles are generated by
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
the transfer of cholesterol ester from LDL to VLDL. Accordingly, the decrease in LDL-C that results is offset by a reciprocal increase in VLDL-C. This explains why the correlation between non-HDL-C and either apoB or LDL particle number is much stronger than the correlation observed between LDL-C and either apoB or LDL particle number19. Non-HDL-C, therefore, is an indirect way of measuring LDL particle number and related to apoB, not cholesterol. NonHDL-C does not validate the importance of VLDL particles in atherogenesis. On the contrary, non-HDL-C reinforces the fundamental importance of LDL particles in atherogenesis. The mistaken interpretation that VLDL-C is highly atherogenic leads to a misplaced faith in the atherogenic risk of triglycerides and a misdirection of therapeutic effort towards triglycerides and away from LDL particles. Indeed, we know that it is not triglycerides and VLDL-C that must be lowered, but LDL particle number.
Lowering LDL: what parameter is best to measure? apoB, non-HDL-C or LDL-C? To date, all guidelines have chosen and promoted LDL-C as the primary target for LDL-lowering therapy20–22. Yet the great majority of statin trials were not designed to test target levels of LDL-C or any other marker. On the contrary, as Hayward and Krumholz have pointed out23, most of these trials were really tests of regimens – either an active arm versus a placebo or two different statin regimens – as opposed to rigorous, protocol-driven comparisons of those who reached a particular level versus those who did not. Because these studies were not tests of targets, LDL-C has no special merit over other markers in terms of assessing the adequacy of LDL-lowering therapies. Both apoB and non-HDL-C are equally valid descriptors of the changes produced by the regimens that were tested. On the other hand, there can be no serious doubt that lowering of LDL is the major, if not the exclusive, mechanism of benefit from statin therapy. Indeed, the Cholesterol Treatment Trialists (CTT) metaanalysis24 demonstrated a continuous relationship between risk and benefit over the range of levels of LDL-C evaluated in the trials they examined. They stated that this relationship was linear, although the possibility of a curvilinear relation was not excluded and, on the basis of all we know from observational studies, this seems more likely. The difference between the two models is significant. In a linear model, relative and absolute benefit remain constant for any degree of LDL-lowering at any level of LDL, whereas in a curvilinear model, relative benefit is constant but absolute benefit becomes exponentially lower as starting levels of LDL become much lower as well. As it turns out, a flexible curvilinear model appears to fit the clinical trial results significantly better than the conventional linear model25. What is indisputable is that all recorded observational studies and, indeed, all clinical trials are consistent with the conclusion that the absolute risk due to LDL depends on the absolute level of LDL. The higher the level of LDL-C, the higher the absolute risk; conversely, the lower the level of LDL-C, the lower the absolute risk. This is crucial background for the next issue: the relative utility of the three markers in monitoring the adequacy of LDL-lowering therapy. As we have
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol
167
Figure 3. Why non-HDL-C is a more accurate marker of LDL particle number than LDL-C. This figure demonstrates how small cholesterol-depleted LDL particles are created when CE and TG are exchanged between VLDL and LDL particles by cholesterol ester transfer protein (CETP). On the left are VLDL1 and LDL1, the particles before the transfer and on the right are VLDL2 and LDL2, the particles after the transfer. The amount of CE in VLDL2 is4than the amount of CE in VLDL1 by the amount of CE that was transferred from LDL1 to VLDL1 to create VLDL2. Similarly, the amount of CE in LDL2 is less than the amount of CE in LDL1 by the amount of CE that was transferred from LDL1 to VLDL1 to create VLDL2. LDL2 is now cholesterol-depleted and would no longer be equal to LDL particle number. However, non-HDL-C2 (¼VLDL2-C þ LDL2-C) is the same as non-HDL-C1 (VLDL1-C þ LDL1-C). Thus the relation between non-HDL-C and the number of LDL particles (one) is the same. And that is how non-HDL-C compensates for the error in LDL-C in estimating the number of small cholesterol-depleted particles.
reviewed, the observational studies assessed support a hierarchy of predictive power of the three markers, with nonHDL-C superior to LDL-C and apoB superior to non-HDL-C. By contrast, in their meta-analysis of 8 of the major statin trials, Boekholdt et al. demonstrated that the HR of non-HDL-C was statistically superior to apoB and LDL-C, although the differences were trivial and of no clinical significance26. The Heart Protection Study27 found no difference at all amongst the three markers in their utility to predict subsequent events. The conclusion would seem to be that it does not matter what you measure to estimate the risk of future events after treatment. That conclusion is wrong.
Interpretation of the HR of LDL-C, non-HDL-C, and apoB after statin treatment The HR for future events after treatment for all three markers are remarkably similar. More noteworthy is that they are substantially less than those found in prospective observational studies. Almost certainly the explanation is that, on average, the levels of LDL-C in individuals receiving therapy as part of a trial are substantially lower than those found in individuals taking part in observational studies, and this means that the hazard posed by LDL is also substantially lower. The assumption that residual risk with statin therapy continues to be driven by LDL is not supported by these results. When LDL is low, LDL is not a major risk and lowering LDL further will not substantially diminish residual risk28.
But this is also not the whole truth. The HR is an average and, while it tends to be much lower in patients on statin therapy, this does not exclude the possibility that a subgroup of these patients – presumably those whose LDL remains the most elevated – might benefit from more intensive LDLlowering therapy. Accordingly, if we group the results of the 6 statin trials that achieved the lowest absolute levels of LDL-C, the average value for LDL-C is 1.8 mmol/L (71 mg/dl), a level equivalent to the 13th percentile of the American population. The average value for non-HDL-C is 2.6 mmol/L (100 mg/dl), a level equivalent to the 14th percentile of the American population. Thus, the levels of non-HDL-C and LDL-C, expressed in terms of the population, are virtually identical. By contrast, the average level of apoB is 88 mg/dl, a level equivalent to the 34th percentile of the population, which means half the patients treated had a value greater than this29. The HR expresses the risk of an event per a standard increment in the level of a marker. Since one standard deviation represents the same extent of change for any marker, the HR has become an accepted index to compare the relative predictive powers of different markers. However, the actual absolute hazard posed by any marker represents the risk posed at any particular level expressed in terms of the actual absolute number of standard deviations it represents. Thus, the higher the percentile of a marker in the population, the greater the absolute hazard posed by that marker. Accordingly, even though the HRs of a subsequent event while on statin treatment for LDL-C, non-HDL-C, and apoB
168
T. de Nijs et al.
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
are all equal, the risk predicted by a level of apoB at the 34th percentile is substantially greater than the risk predicted by an LDL-C or a non-HDL-C level falling within the 13th percentile. Therefore, apoB identifies those individuals who still face substantial risk from LDL and who might well benefit from further LDL-lowering therapy.
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
Summary of the roles of apoB and non-HDL-C as risk predictors and treatment targets The current European, American, and Canadian guidelines regarding these markers present no evidence for why they retain LDL-C as the primary measure of the atherogenic risk due to apoB lipoproteins and the primary measure of the adequacy of LDL-lowering. They acknowledge that both nonHDL-C and apoB are superior, but propose them as ‘‘alternate’’ targets. Non-HDL-C and apoB are closely related measures but they do not perform equally. As discordance analysis confirms, apoB is a more accurate measure of risk than non-HDL-C. Moreover, as we have shown above, apoB more accurately identifies individuals with residual elevations of atherogenic particle number who might benefit from further LDL-lowering therapy. If apoB is not available, nonHDL-C should be calculated and included on laboratory reports. European and Canadian guidelines have proposed an apoB target level of 80 mg/dl. However, an apoB of 65 mg/dl would be the population equivalent of the LDL-C and nonHDL-C targets, and that is the target we suggest in very highrisk subjects (Table 1).
Apolipoprotein B, diagnosis of lipid disorders, and the apoB App While hailed at the time, the classification of lipid disorders created by Frederickson, Levy and Lees30 soon fell out of fashion. Currently, apoB lipid disorders are more simply described as hypercholesterolemia, hypertriglyceridemia, or combined hyperlipidemia. However, this approach is inadequate because it assumes that all forms of hypertriglyceridemia are equally atherogenic, and this is clearly not the case. Moreover, this approach assumes that LDL-C accurately reflects LDL particle number, and this is also false. The original Fredrickson approach was lipid-based, whereas cholesterol and triglyceride circulate within apoB lipoprotein particles. Accordingly, we have developed a simple diagnostic algorithm which changes a lipid classification to a lipoprotein particle-based classification. Our algorithm is based on total cholesterol, triglycerides, and apoB and, except for Lp(a), allows for the identification of elevated apoB lipoprotein particles31. When a clinician knows which lipoproteins are elevated, the differential diagnosis becomes
Table 1. Guideline and population equivalent targets of LDL lowering. LDL-C
non-HDL-C
apoB
PE apoB
Concentration 51.8 mmol/L 52.6 mmol/L 580 mg/dl 565 mg/dl Population 13th 14th 34th 13th Percentile PE – Population equivalent apoB concentration when targeting similar population percentiles for LDL-C¼non-HDL C¼apoB.
straightforward and, therefore, therapeutic approaches are easily determined (Figure 4 and apoB APP).
From hypertriglyceridemia to VLDL, VLDL remnants and chylomicrons particles Hypertriglyceridemia can be broken down into hypertriglyceridemia due to elevations in chylomicron particles versus hypertriglyceridemia due to elevations in VLDL particles, versus hypertriglyceridemia due to a rise in the number of both chylomicron and VLDL particles versus hypertriglyceridemia due to a high number of chylomicron and VLDL remnants. Hypertriglyceridemia due to excess VLDL may either be associated with markedly increased numbers of LDL particles and, therefore, markedly increased cardiovascular risk, or with a normal number of LDL particles and no detectable increase in cardiovascular risk. Because cholesterol-depleted LDL particles are so common with hypertriglyceridemia, LDL-C will not distinguish the two and non-HDL-C may be high in both. However, apoB will identify which patients have elevated total atherogenic particle number and elevated LDL particle number. Importantly, hypertriglyceridemia may be a familial condition. Two syndromes are common and need to be distinguished: familial combined hyperlipidemia (FCHL) is characterized by high apoB (41200 mg/l), whereas familial hypertriglyceridemia (FHTG) is characterized by normal apoB (51200 mg/l). Cardiovascular event rates are much lower in patients with FHTG than in those with FCHL32,33. Another major advantage of this algorithm is that the diagnosis of familial dysbetalipoproteinemia (FDB) (type III hyperlipoproteinemia in Frederickson’s classification of the hyperlipoproteinemias), which is due to the presence of remnant chylomicron and VLDL particles and which is highly atherogenic, can be easily made by any physician with access to a routine clinical laboratory. FDB is characterized by a decreased clearance of chylomicron VLDL remnants due to variants in the APOE phenotype, usually the APOE2/E2 variants. In most cases, an additional genetic, hormonal, or environmental factor is needed for hyperlipoproteinemia to manifest34. Conventionally, a ratio of VLDL lipoprotein cholesterol to plasma triglyceride concentration (CVLDL/ TG ¼ r) greater than 0.30 was said to be diagnostic35, but this can only be measured using ultracentrifugation – an expensive, time-consuming research technology. Using cholesterol, triglyceride, and apoB, type III, hyperlipoproteinemia can be easily and reliably diagnosed7. Other disorders characterized by hypertriglyceridemia are LPL-deficiency and primary apoCII deficiency. In these disorders, chylomicrons cannot be cleared effectively and the associated clinical scenarios are characterized by the combination of severely elevated triglyceride levels and very low ApoB levels (575 mg/dl)36.
Familial hypercholesterolemia ApoB levels are extremely elevated in patients with the familial hypercholesterolemia phenotype, which can be monogenic or polygenic in origin. Mutations in the LDL receptor are by far the most common monogenic cause, but
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol
169
Figure 4. Diagnostic algorithm for the atherogenic ApoB dyslipoproteinemias. This diagnostic algorithm can be downloaded for free in the APP store (apoB APP)31.
Figure 5. Summary of LDL-C, non-HDL-C and apoB and their performances as risk predictors, treatment targets, and diagnostic markers of lipid disorders. 2012 Update of the Canadian Cardiovascular Society guidelines for the diagnosis and treatment of dyslipidemia for the prevention of cardiovascular disease in adults22 (reprinted with permission).
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
170
T. de Nijs et al.
mutations in apoB or proprotein convertase subtilisin/kexin 9 (PCSK9), a protein that regulates LDL receptor number at the cell surface, can also occur. Other unusual variants are possible such as autosomal recessive hypercholesterolemia (ARH). Importantly, it has recently been shown that about 30% of the cases may be polygenic in origin and, in a substantial number of others with this phenotype, no genetic basis can be demonstrated37. Notwithstanding the genotype, clinical risk and management are based on the phenotype – markedly elevated levels of cholesterol-rich apoB particles. These are only a few examples to illustrate the type of diagnoses that can be made using apoB in combination with total cholesterol and triglycerides. The algorithm – the apoB app – can be downloaded for free from the Apple app store. Using the apoB app, the practicing clinician can easily diagnose all the apoB atherogenic dyslipoproteinemias and quickly read about the primary and secondary causes of these disorders and their therapeutic approaches.
Summary of the role of ApoB in diagnosis and cardiovascular risk management In summary, incorporating apoB with conventional lipids allows more precise estimation of cardiovascular risk, more accurate determination of the adequacy of LDL-lowering therapy, and, for the first time, rapid diagnosis of the atherogenic apoB dyslipoproteinemias. In Figure 5, the performance of apoB and non-HDL-C versus LDL-C as risk predictors, treatment targets, and diagnostic markers of lipid disorders is summarized.
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
References 1. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18: 499–502. 2. Ketelhuth DFJ, Rios FJO, Wang Y, et al. Identification of a danger-associated peptide from apolipoprotein B100 (ApoBDS-1) that triggers innate proatherogenic responses. Circulation 2011; 124:2433–43. 3. Avraham-Davidi I, Ely Y, Pham VN, et al. ApoB-containing lipoproteins regulate angiogenesis by modulating expression of VEGF receptor 1. Nat Med 2012;18:967–73. 4. Young SG. Recent progress in understanding apolipoprotein B. Circulation 1990;82:1574–94. 5. Stalenhoef AF, Malloy MJ, Kane JP, Havel RJ. Metabolism of apolipoproteins B-48 and B-100 of triglyceride-rich lipoproteins in normal and lipoprotein lipase-deficient humans. Proc Natl Acad Sci 1984;81:1839–43. 6. Sniderman A, Vu H, Cianflone K. Effect of moderate hypertriglyceridemia on the relation of plasma total and LDL apo B levels. Atherosclerosis 1991;89:109–16. 7. Sniderman A, Tremblay A, Bergeron J, et al. Diagnosis of type III hyperlipoproteinemia from plasma total cholesterol, triglyceride, and apolipoprotein B. J Clin Lipidol 2007;1:256–63. 8. Sniderman AD, Tremblay A, De Graaf J, Couture P. Phenotypes of hypertriglyceridemia caused by excess very-low-density lipoprotein. J Clin Lipidol 2012;6:427–33. 9. Sniderman AD, St-Pierre AC, Cantin B, et al. Concordance/ discordance between plasma apolipoprotein B levels and the cholesterol indexes of atherosclerotic risk. Am J Cardiol 2003;91: 1173–7.
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
10. Contois JH, McConnell JP, Sethi AA, et al; AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Apolipoprotein B and cardiovascular disease risk: position statement from the AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Clin Chem 2009;55:407–1. 11. AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Association of apolipoprotein B and nuclear magnetic resonance spectroscopy-derived LDL particle number with outcomes in 25 clinical studies: assessment by the AACC Lipoprotein and Vascular Diseases Division Working Group on Best Practices. Clin Chem 2013;59:752–70. 12. Sniderman AD, Williams K, Contois JH, et al. A meta-analysis of low-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, and apolipoprotein b as markers of cardiovascular risk. Circ Cardiovasc Qual Outcomes 2011;4:337–45. 13. The Emerging Risk Factors Collaboration. Major lipids, apolipoproteins, and risk of vascular disease. JAMA. 2009;302: 1993–2000. 14. The Emerging Risk Factors Collaboration. Lipid-related markers and cardiovascular disease prediction. JAMA 2012;20;307: 2499–506. 15. Sniderman AD Furberg CD. Cardiovascular disease risk prediction factors. JAMA 2012;308:1969. 16. Cromwell WC, Otvos JD, Keyes MJ, et al. LDL particle number and risk of future cardiovascular disease in the Framingham Offspring Study – implications for LDL management. J Clin Lipidol 2007;1:583–92. 17. Otvos JD, Mora S, Shalaurova I, et al. Clinical implications of discordance between low-density lipoprotein cholesterol and particle number. J Clin Lipidol 2011;5:105–13. 18. Sniderman AD, Islam S, Yusuf S, McQueen MJ. Discordance analysis of apolipoprotein B and non-high density lipoprotein cholesterol as markers of cardiovascular risk in the INTERHEART study. Atherosclerosis 2012;225:444–9. 19. Sniderman A, McQueen M, Contois J, et al. Why is nonhighdensity lipoprotein cholesterol a better marker of the risk of vascular disease than low-density lipoprotein cholesterol?. J Clin Lipidol 2010;4:152–5. 20. Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the national cholesterol education program (ncep) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel iii). JAMA 2001;285: 2486–97. 21. Catapano AL, Reiner Zˇ, De Backer G, et al. ESC/EAS Guidelines for the management of dyslipidaemias: the task force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Atherosclerosis 2011;217:1–44 (and in Eur Heart J. 2011 Jul;32(14):1769-818). 22. Anderson TJ, Gre´goire J, Hegele RA, et al. 2012 Update of the Canadian Cardiovascular Society Guidelines for the Diagnosis and Treatment of Dyslipidemia for the Prevention of Cardiovascular Disease in the Adult. Can J Cardiol 2013;29: 151–67. 23. Hayward RA, Krumholz HM. Three reasons to abandon lowdensity lipoprotein targets: an open letter to the Adult Treatment Panel IV of the National Institutes of Health. Circ Cardiovasc Qual Outcomes 2012;5:2–5. 24. Baigent C, Blackwell L, Emberson J, et al.; Cholesterol Treatment Trialists’ (CTT) Collaboration. Cholesterol treatment trialists collaboration efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170 000 participants in 26 randomised trials. Lancet 2010; 376:1670–81. 25. Takagi H, Umemoto T. Limit to benefits of large reductions in low-density lipoprotein cholesterol levels: use of fractional polynomials to assess the effect of low-density lipoprotein cholesterol level reduction in metaregression of large statin randomized trials. JAMA Intern Med 2013;173:1028–9. 26. Boekholdt S, Arsenault BJ, Mora S, et al. Association of LDL cholesterol, non–HDL cholesterol, and apolipoprotein b levels with risk of cardiovascular events among patients treated with statins: a meta-analysis. JAMA 2012;307:1302–9. 27. Parish S, Offer A, Clarke R, et al; Heart Protection Study Collaborative Group. Lipids and lipoproteins and risk of different
DOI: 10.3109/10408363.2013.847897
28.
29. 30.
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
31. 32.
vascular events in the MRC/BHF Heart Protection Study. Circulation 2012;125:2469–78. Sniderman A, Thanassoulis G, Couture P, et al. Is lower and lower better and better? A re-evaluation of the evidence from the Cholesterol Treatment Trialists’ Collaboration meta-analysis for low-density lipoprotein lowering. J Clin Lipidol 2012;6:303–9. Sniderman AD, De Graaf J, Couture P. Low-density lipoproteinlowering strategies: target versus maximalist versus population percentile. Curr Opin Cardiol 2012;27:405–11. Fredrickson DS, Levy RI, Lees RS. Fat transport in lipoproteins – an integrated approach to mechanisms and disorders. N Engl J Med 1967;276:34–44. Sniderman A, Couture P, de Graaf J. Diagnosis and treatment of apolipoprotein B dyslipoproteinemias. Nat Rev Endocrinol 2010; 6:335–46. Brunzell JD, Albers JJ, Chait A, et al. Plasma lipoproteins in familial combined hyperlipidemia and monogenic familial hypertriglyceridemia. J Lipid Res 1983;24:147–55.
ApoB versus non-HDL-cholesterol
171
33. Veerkamp MJ, de Graaf J, Hendriks JC, et al. Nomogram to diagnose familial combined hyperlipidemia on the basis of results of a 5-year follow-up study. Circulation 2004;109: 2980–5. 34. Mahley RW, Huang Y, Rall SC. Pathogenesis of type III hyperlipoproteinemia (dysbetalipoproteinemia): questions, quandaries, and paradoxes. J Lipid Res 1999;40:1933–49. 35. Fredrickson DS, Morganroth J, Levy RI. Type III hyperlipoproteinemia: an analysis of two contemporary definitions. Ann Intern Med 1975;82:150–7. 36. Zambon A, Brunzell JD, Torres A, et al. Prevention of raised low-density lipoprotein cholesterol in a patient with familial hypercholesterolaemia and lipoprotein lipase deficiency. Lancet 1993;341:1119–21. 37. Talmud PJ, Shah S, Whittall R, et al. Use of low-density lipoprotein cholesterol gene score to distinguish patients with polygenic and monogenic familial hypercholesterolaemia: a casecontrol study. Lancet 2013;381:1293–301.
ApoB versus non-HDL-cholesterol: Diagnosis and cardiovascular risk management Tjerk de Nijs1, Allan Sniderman2, and Jacqueline de Graaf1 1
Department of General Internal Medicine, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands, and 2Mike Rosenbloom Laboratory for Cardiovascular Research, McGill University, Royal Victoria Hospital, Montreal, Canada Abstract
Keywords
The most recent guidelines released by the EAS/ESC and the Canadian Cardiovascular Society (CCS) retain low-density lipoprotein cholesterol (LDL-C) as the primary measure of the atherogenic risk of the apolipoprotein B (apoB) lipoproteins and the primary target of LDL-C lowering therapy. Both organizations endorse non-high-density lipoprotein cholesterol (nonHDL-C) and apoB as ‘‘alternate/secondary’’ targets, but neither group offers evidence supporting the continued preference of LDL-C as the primary target over non-HDL-C and apoB. Further, both suggest that non-HDL-C and apoB more or less measure the same thing and, therefore, are essentially interchangeable. But what is the evidence that LDL-C should remain the primary target, and are apoB and non-HDL-C mirror images of one another? Furthermore, are estimation of risk and establishment of treatment targets the only relevant issues, or is diagnosis also an essential objective? These are the questions this article will address. Our principal objectives are: (1) to clarify the differences between LDL-C, non-HDL-C, and apoB and to distinguish what they measure; (2) to summarize the evidence relating to LDLC, non-HDL-C, and apoB as predictors of cardiovascular risk and as targets for treatment; and (3) to demonstrate that diagnosis of atherogenic dyslipoproteinemias should be a fundamental clinical priority.
ApoB, atherosclerosis, apolipoproteins, CVD, lipoproteins, non-HDL-C History Received 22 July 2013 Revised 7 September 2013 Accepted 19 September 2013 Published online 29 November 2013
Abbreviations: apoB: apolipoprotein B; APP: application; CCS: Canadian Cardiovascular Society; EAS: European Atherosclerosis Society; ESC: European Society of Cardiology; FCHL: familial combined hyperlipidemia; FDB: familial dysbetalipoproteinemia; FHTG: familial hypertriglyceridemia; HDL: high density lipoprotein; HR: hazard ratio; IDL: intermediate density lipoprotein; LDL-C: low-density lipoprotein cholesterol; LDL: low-density lipoprotein; NonHDL-C: non-high density lipoprotein cholesterol; TG: triglyceride; VLDL: very low density lipoprotein
Introduction Cardiovascular disease (CVD) is now a worldwide affliction and dyslipidemia is one of the most important and most modifiable causes. The major organizations responsible for creating CVD guidelines state that low-density lipoprotein cholesterol (LDL-C) is the best measure of the proatherogenic risks of low-density lipoproteins. However, recent evidence indicates that both non-high-density lipoprotein cholesterol
Referee: Dr. John Burnett, Clinical Professor in the School of Medicine & Pharmacology, University of Western Australia; Consultant Medical Biochemist in the Department of Clinical Biochemistry at PathWest Laboratory Medicine WA, Royal Perth Hospital (RPH). Address for correspondence: Prof. Dr Jacqueline de Graaf, Department of General Internal Medicine 463, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail: [email protected]
(non-HDL-C) and apolipoprotein B (apoB) are more accurate indices of cardiovascular risk than LDL-C. While the importance of LDL-C is well understood and the evidence supporting the clinical use of LDL-C is extensive, little is known about non-HDL-C and apoB. Furthermore, in today’s clinical practice, where statins are easily accessible, diagnosis of lipid disorders is often neglected. Part of this may be explained by the fact that the Frederickson classification of dyslipoproteinemia is rather complex and often requires additional, expensive, and complex measurements of lipoproteins lipids. With the introduction of apoB as a biomarker, diagnosis of dyslipoproteinemia in clinical practice is now simple. Accordingly, this article will focus on the strengths and limitations of apoB and non-HDL-C compared to LDL-C in cardiovascular risk evaluation and will evaluate how apoB facilitates diagnosis of dyslipoproteinemia to achieve complete cardiovascular risk management.
20 13
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
REVIEW ARTICLE
164
T. de Nijs et al.
What do LDL-C, non-HDL-C, and apoB measure?
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
LDL-C LDL-C is a measure of the mass of cholesterol within lowdensity lipoprotein (LDL) particles. LDL particles were originally defined as those lipoproteins with a density between 1.019 and 1.063 g/ml; however, more recently, the definition has been expanded to encompass all such particles between 1.006 and 1.063 g/ml. That is, the present conventional definition includes both classical intermediate density lipoprotein particles (IDL) and classical LDL particles. An LDL-C value calculated using the Friedewald equation includes all the cholesterol within apoB particles1 between 1.006 and 1.063 g/ml, as would an LDL-C value measured directly by any commercially available method. Relying on LDL-C assumes that cardiovascular risk due to LDL is equal to the mass of cholesterol within all LDL particles. Since the mass of cholesterol within individual LDL particles can vary substantially, relying on LDL-C assumes that the amount of cholesterol that is deposited within the arterial wall is independent from the number of particles. That is, this cholesterol model predicts that individuals with a small number of cholesterol-enriched particles will deposit more cholesterol within their arteries than individuals with a lower LDL-C but a greater number of cholesterol-depleted particles. However, there is no evidence to suggest that this is the case and compelling evidence indicates that it is not. The cholesterol within apoB particles enters the arterial wall by way of those apoB particles. The primary determinant of the number of apoB particles that enter and are trapped within the arterial wall and, therefore, the mass of cholesterol deposited within the arterial wall is the number of apoB particles within the lumen of the artery. Moreover, cholesterol is not the only injurious substance within the LDL particle. ApoB itself can stimulate an inflammatory response as well as angiogenesis2,3. ApoB There are two types of apoB in plasma: (1) apoB-100, which is synthesized by the liver and provides structural integrity to VLDL, LDL, and Lp(a) particles, LDL particles to which an apo(a) has been attached, and (2) apoB-48, which is synthesized by the intestine and encircles and provides structural integrity to chylomicron and chylomicron remnant particles4. After a meal, plasma triglycerides rise and fall as apoB-48 chylomicron particles secreted by the intestine enter the circulation and are rapidly cleared by the liver. Nevertheless, even at peak postprandial levels, there are so few apo-B48 particles compared to the number of apo-B100 particles that, for practical purposes, apoB-100 accounts for virtually all measured apoB. Thus, apoB equals5 apoB-100. VLDL transport triglyceride from the liver into adipose tissue and skeletal muscle. As the triglyceride within them is progressively hydrolyzed, the VLDL particles (d51.006 g/ ml) become progressively smaller and cholesterol-enriched. They are transformed sequentially into IDL particles (d 1.006–1.019 g/ml) containing roughly equal quantities of cholesterol and triglycerides, and then into LDL particles (d 1.019–1.063 g/ml), which are even smaller and more enriched
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
with cholesterol. IDL and LDL particles are unquestionably atherogenic. In most instances, even in hypertriglyceridaemia (except for familial dysbetalipoproteinemia: see next paragraph) or severe hypertriglyceridemia, LDL particles make up more than 90% of the total plasma apoB6–8. VLDL particles are much larger than LDL particles and, therefore, enter the arterial wall much less easily than LDL particles. Moreover, in almost all individuals, there are nine times as many LDL particles as VLDL particles. In other words, LDL particles are much more atherogenic than VLDL particles because there are many more of them and they are much smaller, explaining why LDL and not VLDL drives the atherogenic risk due to apoB lipoproteins. In summary, apoB equals the sum of the VLDL, IDL, LDL and Lp(a) particles in plasma, i.e. all potentially atherogenic lipoprotein particles (Figure 1). The utility of using the number of apoB particles as a marker of cardiovascular risk is based on the fact that, although each apoB particle contains one molecule of apoB, the mass of cholesterol within the apoB particles can vary substantially. When this occurs, LDL-C does not equal LDL particle number. Cholesterol-enriched LDL particles are larger and more buoyant than normal LDL particles, whereas cholesterol-depleted LDL particles are smaller and denser than normal LDL particles. When LDL particles are normal in composition, LDL-C and apoB are concordant and LDL-C is an accurate measure of the number of LDL particles. However, when cholesterol-enriched or cholesterol-depleted LDL particles are dominant, LDL-C and apoB are discordant and LDL-C does not provide an accurate measure9 (Figure 2). Fasting samples are not required to measure apoB because there are so few apo-B48 particles. The measurement of apoB is standardized and the American Association of Clinical Chemistry has determined that apoB can be measured more accurately in routine clinical laboratories than either LDL-C or non-HDL-C10,11. The perception of difficulties in the measurement of apoB relates to research studies, which used in-house, non-standardized methods or analyzed samples that have been preserved for extended periods of time. Non-HDL-C Non-HDL-C is the arithmetic sum of the cholesterol in VLDL and LDL particles and therefore, by definition, it equals all the atherogenic cholesterol in lipoproteins. Put this way, apoB and non-HDL-C appear to be mirror images of each other, with apoB measuring all the atherogenic particles and non-HDL-C measuring all the atherogenic cholesterol. This may explain why many argue that nonHDL-C is an acceptable functional surrogate for apoB. Given that no extra charge is required to calculate nonHDL-C, these individuals may conclude that non-HDL-C is superior to apoB. However, similar to LDL-C and apoB, when VLDL and LDL particles are either cholesterolenriched or cholesterol-depleted – that is, when apoB and non-HDL-C are discordant – non-HDL-C is not an accurate measure of the number of total apoB particles. Therefore, non-HDL-C does not necessarily equal total apoB particle number, as will be discussed further below.
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol
165
Figure 1. Every apoB lipoprotein particle contains one molecule of apoB – so ApoB equals the sum of the VLDL, IDL, and LDL particles in plasma, i.e. all the potential atherogenic lipoprotein particles. Triglyceride-rich VLDL particles and LDL particles differ in composition: VLDL particles can contain variable amounts of triglyceride (TG) or cholesterolesters (CE). LDL particles can be cholesterol-depleted or cholesterol-enriched. Remnant particles when present are in IDL density range. The (blue) circle reflects one molecule of apoB per apoB lipoprotein particle.
Figure 2. Is an LDL cholesterol level of 3.0 mmol/l high risk or not? It depends on the number of atherogenic LDL particles. In the subject on the left side, numerous small, dense LDL particles result in an LDL-C of 3 mmol/L. As each LDL particle has one apoB, apoB will be high (41.2 g/L – 90th percentile of the population). In the subject on the right side, normal, relatively large, buoyant LDL.
LDL-C, non-HDL-C, and apoB in cardiovascular risk prediction The conventional approach to determining which marker of risk is superior to another is to compare the predictive power of each method in all subjects in a study and then integrate the resulting values in a meta-analysis. This was the approach of
Sniderman et al., who demonstrated a hierarchy of accuracy in which non-HDL-C was found to be superior to LDL-C and apoB was superior to non-HDL-C. Importantly, this study showed that these differences were clinically as well as statistically significant12. In contrast, the Emerging Risk Factor Collaboration (ERFC) determined that non-HDL-C and apoB
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
166
T. de Nijs et al.
are equivalent markers of risk13,14. While this result garnered significant attention, the ERFC also showed that LDL-C and total cholesterol were equal to non-HDL-C and apoB. That is, the ERFC found that TC was equal to any other measure of the lipoprotein-related risk of vascular disease, which contradicts such a large number of studies that, in our view, it is not credible. Despite the refined design and mathematical analyses that entered into this study, it may be that, since such a large number of studies were evaluated, many featured inadequate design, methods, and execution15. The ERFC is a participantlevel analysis based on actual individual data, which means that a new population is constructed from variable population subsections of all studies that chose to participate. The strength of participant-level analysis is that it is based on individual rather than summary data, as in a meta-analysis. However, in contrast to meta-analyses, there is no commitment to include all of the relevant studies. Nevertheless, neither meta-analysis nor participant-level analysis may be the most appropriate method to resolve the dispute. The reason that some studies show apoB to be superior to non-HDL-C, whereas others do not, is that the differences in predictive power can be apparent only in discordant subgroups – the subgroups in which apoB particles are either cholesterol-enriched or cholesterol-depleted. In groups for which LDL-C and apoB are concordant, both markers will predict risk equally and, since the concordant groups are much larger than the discordant ones, any potential difference in the two markers will be diluted by the much larger number of subjects in the concordant group. By restricting the comparison of the predictive powers of these measures to the discordant groups, it is possible to clearly determine whether it is the mass of cholesterol or the number of apoB particles that matters the most9. Discordance analyses have shown striking superiority for LDL particle number over LDL-C16,17 and of apoB over nonHDL-C18. Accordingly, discordance analysis appears to be the most appropriate way to demonstrate the limitations of the conventional approach to risk evaluation.
Pathophysiology/mechanism to explain the role of non-HDL-C and apoB in CV prediction The commonsense and, therefore, commonly-offered explanation for the superiority of non-HDL-C over LDL-C is that non-HDL-C includes the cholesterol in VLDL as well as LDL and thus represents the total atherogenic cholesterol burden. But the evidence that VLDL cholesterol is atherogenic is weak to non-existent. Despite the fact that fibrates reduce triglycerides and VLDL-C, there is no evidence suggesting that these changes reduce cardiovascular risk. Indeed, if VLDL-C ‘explained’ why non-HDL-C was superior to LDL-C, the hazard ratio (HR) of VLDL-C would have to be substantially higher than to the HR of LDL-C19. That is, VLDL-C would have to be more dangerous than LDL-C, and there is no evidence that this is the case. On the contrary, there is a good reason to believe that the superior predictive power of non-HDL-C over LDL-C is based on the fact that non-HDL-C is a more accurate marker of LDL particle number than LDL-C. As illustrated in Figure 3, cholesterol-depleted LDL particles are generated by
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
the transfer of cholesterol ester from LDL to VLDL. Accordingly, the decrease in LDL-C that results is offset by a reciprocal increase in VLDL-C. This explains why the correlation between non-HDL-C and either apoB or LDL particle number is much stronger than the correlation observed between LDL-C and either apoB or LDL particle number19. Non-HDL-C, therefore, is an indirect way of measuring LDL particle number and related to apoB, not cholesterol. NonHDL-C does not validate the importance of VLDL particles in atherogenesis. On the contrary, non-HDL-C reinforces the fundamental importance of LDL particles in atherogenesis. The mistaken interpretation that VLDL-C is highly atherogenic leads to a misplaced faith in the atherogenic risk of triglycerides and a misdirection of therapeutic effort towards triglycerides and away from LDL particles. Indeed, we know that it is not triglycerides and VLDL-C that must be lowered, but LDL particle number.
Lowering LDL: what parameter is best to measure? apoB, non-HDL-C or LDL-C? To date, all guidelines have chosen and promoted LDL-C as the primary target for LDL-lowering therapy20–22. Yet the great majority of statin trials were not designed to test target levels of LDL-C or any other marker. On the contrary, as Hayward and Krumholz have pointed out23, most of these trials were really tests of regimens – either an active arm versus a placebo or two different statin regimens – as opposed to rigorous, protocol-driven comparisons of those who reached a particular level versus those who did not. Because these studies were not tests of targets, LDL-C has no special merit over other markers in terms of assessing the adequacy of LDL-lowering therapies. Both apoB and non-HDL-C are equally valid descriptors of the changes produced by the regimens that were tested. On the other hand, there can be no serious doubt that lowering of LDL is the major, if not the exclusive, mechanism of benefit from statin therapy. Indeed, the Cholesterol Treatment Trialists (CTT) metaanalysis24 demonstrated a continuous relationship between risk and benefit over the range of levels of LDL-C evaluated in the trials they examined. They stated that this relationship was linear, although the possibility of a curvilinear relation was not excluded and, on the basis of all we know from observational studies, this seems more likely. The difference between the two models is significant. In a linear model, relative and absolute benefit remain constant for any degree of LDL-lowering at any level of LDL, whereas in a curvilinear model, relative benefit is constant but absolute benefit becomes exponentially lower as starting levels of LDL become much lower as well. As it turns out, a flexible curvilinear model appears to fit the clinical trial results significantly better than the conventional linear model25. What is indisputable is that all recorded observational studies and, indeed, all clinical trials are consistent with the conclusion that the absolute risk due to LDL depends on the absolute level of LDL. The higher the level of LDL-C, the higher the absolute risk; conversely, the lower the level of LDL-C, the lower the absolute risk. This is crucial background for the next issue: the relative utility of the three markers in monitoring the adequacy of LDL-lowering therapy. As we have
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol
167
Figure 3. Why non-HDL-C is a more accurate marker of LDL particle number than LDL-C. This figure demonstrates how small cholesterol-depleted LDL particles are created when CE and TG are exchanged between VLDL and LDL particles by cholesterol ester transfer protein (CETP). On the left are VLDL1 and LDL1, the particles before the transfer and on the right are VLDL2 and LDL2, the particles after the transfer. The amount of CE in VLDL2 is4than the amount of CE in VLDL1 by the amount of CE that was transferred from LDL1 to VLDL1 to create VLDL2. Similarly, the amount of CE in LDL2 is less than the amount of CE in LDL1 by the amount of CE that was transferred from LDL1 to VLDL1 to create VLDL2. LDL2 is now cholesterol-depleted and would no longer be equal to LDL particle number. However, non-HDL-C2 (¼VLDL2-C þ LDL2-C) is the same as non-HDL-C1 (VLDL1-C þ LDL1-C). Thus the relation between non-HDL-C and the number of LDL particles (one) is the same. And that is how non-HDL-C compensates for the error in LDL-C in estimating the number of small cholesterol-depleted particles.
reviewed, the observational studies assessed support a hierarchy of predictive power of the three markers, with nonHDL-C superior to LDL-C and apoB superior to non-HDL-C. By contrast, in their meta-analysis of 8 of the major statin trials, Boekholdt et al. demonstrated that the HR of non-HDL-C was statistically superior to apoB and LDL-C, although the differences were trivial and of no clinical significance26. The Heart Protection Study27 found no difference at all amongst the three markers in their utility to predict subsequent events. The conclusion would seem to be that it does not matter what you measure to estimate the risk of future events after treatment. That conclusion is wrong.
Interpretation of the HR of LDL-C, non-HDL-C, and apoB after statin treatment The HR for future events after treatment for all three markers are remarkably similar. More noteworthy is that they are substantially less than those found in prospective observational studies. Almost certainly the explanation is that, on average, the levels of LDL-C in individuals receiving therapy as part of a trial are substantially lower than those found in individuals taking part in observational studies, and this means that the hazard posed by LDL is also substantially lower. The assumption that residual risk with statin therapy continues to be driven by LDL is not supported by these results. When LDL is low, LDL is not a major risk and lowering LDL further will not substantially diminish residual risk28.
But this is also not the whole truth. The HR is an average and, while it tends to be much lower in patients on statin therapy, this does not exclude the possibility that a subgroup of these patients – presumably those whose LDL remains the most elevated – might benefit from more intensive LDLlowering therapy. Accordingly, if we group the results of the 6 statin trials that achieved the lowest absolute levels of LDL-C, the average value for LDL-C is 1.8 mmol/L (71 mg/dl), a level equivalent to the 13th percentile of the American population. The average value for non-HDL-C is 2.6 mmol/L (100 mg/dl), a level equivalent to the 14th percentile of the American population. Thus, the levels of non-HDL-C and LDL-C, expressed in terms of the population, are virtually identical. By contrast, the average level of apoB is 88 mg/dl, a level equivalent to the 34th percentile of the population, which means half the patients treated had a value greater than this29. The HR expresses the risk of an event per a standard increment in the level of a marker. Since one standard deviation represents the same extent of change for any marker, the HR has become an accepted index to compare the relative predictive powers of different markers. However, the actual absolute hazard posed by any marker represents the risk posed at any particular level expressed in terms of the actual absolute number of standard deviations it represents. Thus, the higher the percentile of a marker in the population, the greater the absolute hazard posed by that marker. Accordingly, even though the HRs of a subsequent event while on statin treatment for LDL-C, non-HDL-C, and apoB
168
T. de Nijs et al.
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
are all equal, the risk predicted by a level of apoB at the 34th percentile is substantially greater than the risk predicted by an LDL-C or a non-HDL-C level falling within the 13th percentile. Therefore, apoB identifies those individuals who still face substantial risk from LDL and who might well benefit from further LDL-lowering therapy.
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
Summary of the roles of apoB and non-HDL-C as risk predictors and treatment targets The current European, American, and Canadian guidelines regarding these markers present no evidence for why they retain LDL-C as the primary measure of the atherogenic risk due to apoB lipoproteins and the primary measure of the adequacy of LDL-lowering. They acknowledge that both nonHDL-C and apoB are superior, but propose them as ‘‘alternate’’ targets. Non-HDL-C and apoB are closely related measures but they do not perform equally. As discordance analysis confirms, apoB is a more accurate measure of risk than non-HDL-C. Moreover, as we have shown above, apoB more accurately identifies individuals with residual elevations of atherogenic particle number who might benefit from further LDL-lowering therapy. If apoB is not available, nonHDL-C should be calculated and included on laboratory reports. European and Canadian guidelines have proposed an apoB target level of 80 mg/dl. However, an apoB of 65 mg/dl would be the population equivalent of the LDL-C and nonHDL-C targets, and that is the target we suggest in very highrisk subjects (Table 1).
Apolipoprotein B, diagnosis of lipid disorders, and the apoB App While hailed at the time, the classification of lipid disorders created by Frederickson, Levy and Lees30 soon fell out of fashion. Currently, apoB lipid disorders are more simply described as hypercholesterolemia, hypertriglyceridemia, or combined hyperlipidemia. However, this approach is inadequate because it assumes that all forms of hypertriglyceridemia are equally atherogenic, and this is clearly not the case. Moreover, this approach assumes that LDL-C accurately reflects LDL particle number, and this is also false. The original Fredrickson approach was lipid-based, whereas cholesterol and triglyceride circulate within apoB lipoprotein particles. Accordingly, we have developed a simple diagnostic algorithm which changes a lipid classification to a lipoprotein particle-based classification. Our algorithm is based on total cholesterol, triglycerides, and apoB and, except for Lp(a), allows for the identification of elevated apoB lipoprotein particles31. When a clinician knows which lipoproteins are elevated, the differential diagnosis becomes
Table 1. Guideline and population equivalent targets of LDL lowering. LDL-C
non-HDL-C
apoB
PE apoB
Concentration 51.8 mmol/L 52.6 mmol/L 580 mg/dl 565 mg/dl Population 13th 14th 34th 13th Percentile PE – Population equivalent apoB concentration when targeting similar population percentiles for LDL-C¼non-HDL C¼apoB.
straightforward and, therefore, therapeutic approaches are easily determined (Figure 4 and apoB APP).
From hypertriglyceridemia to VLDL, VLDL remnants and chylomicrons particles Hypertriglyceridemia can be broken down into hypertriglyceridemia due to elevations in chylomicron particles versus hypertriglyceridemia due to elevations in VLDL particles, versus hypertriglyceridemia due to a rise in the number of both chylomicron and VLDL particles versus hypertriglyceridemia due to a high number of chylomicron and VLDL remnants. Hypertriglyceridemia due to excess VLDL may either be associated with markedly increased numbers of LDL particles and, therefore, markedly increased cardiovascular risk, or with a normal number of LDL particles and no detectable increase in cardiovascular risk. Because cholesterol-depleted LDL particles are so common with hypertriglyceridemia, LDL-C will not distinguish the two and non-HDL-C may be high in both. However, apoB will identify which patients have elevated total atherogenic particle number and elevated LDL particle number. Importantly, hypertriglyceridemia may be a familial condition. Two syndromes are common and need to be distinguished: familial combined hyperlipidemia (FCHL) is characterized by high apoB (41200 mg/l), whereas familial hypertriglyceridemia (FHTG) is characterized by normal apoB (51200 mg/l). Cardiovascular event rates are much lower in patients with FHTG than in those with FCHL32,33. Another major advantage of this algorithm is that the diagnosis of familial dysbetalipoproteinemia (FDB) (type III hyperlipoproteinemia in Frederickson’s classification of the hyperlipoproteinemias), which is due to the presence of remnant chylomicron and VLDL particles and which is highly atherogenic, can be easily made by any physician with access to a routine clinical laboratory. FDB is characterized by a decreased clearance of chylomicron VLDL remnants due to variants in the APOE phenotype, usually the APOE2/E2 variants. In most cases, an additional genetic, hormonal, or environmental factor is needed for hyperlipoproteinemia to manifest34. Conventionally, a ratio of VLDL lipoprotein cholesterol to plasma triglyceride concentration (CVLDL/ TG ¼ r) greater than 0.30 was said to be diagnostic35, but this can only be measured using ultracentrifugation – an expensive, time-consuming research technology. Using cholesterol, triglyceride, and apoB, type III, hyperlipoproteinemia can be easily and reliably diagnosed7. Other disorders characterized by hypertriglyceridemia are LPL-deficiency and primary apoCII deficiency. In these disorders, chylomicrons cannot be cleared effectively and the associated clinical scenarios are characterized by the combination of severely elevated triglyceride levels and very low ApoB levels (575 mg/dl)36.
Familial hypercholesterolemia ApoB levels are extremely elevated in patients with the familial hypercholesterolemia phenotype, which can be monogenic or polygenic in origin. Mutations in the LDL receptor are by far the most common monogenic cause, but
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
DOI: 10.3109/10408363.2013.847897
ApoB versus non-HDL-cholesterol
169
Figure 4. Diagnostic algorithm for the atherogenic ApoB dyslipoproteinemias. This diagnostic algorithm can be downloaded for free in the APP store (apoB APP)31.
Figure 5. Summary of LDL-C, non-HDL-C and apoB and their performances as risk predictors, treatment targets, and diagnostic markers of lipid disorders. 2012 Update of the Canadian Cardiovascular Society guidelines for the diagnosis and treatment of dyslipidemia for the prevention of cardiovascular disease in adults22 (reprinted with permission).
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
170
T. de Nijs et al.
mutations in apoB or proprotein convertase subtilisin/kexin 9 (PCSK9), a protein that regulates LDL receptor number at the cell surface, can also occur. Other unusual variants are possible such as autosomal recessive hypercholesterolemia (ARH). Importantly, it has recently been shown that about 30% of the cases may be polygenic in origin and, in a substantial number of others with this phenotype, no genetic basis can be demonstrated37. Notwithstanding the genotype, clinical risk and management are based on the phenotype – markedly elevated levels of cholesterol-rich apoB particles. These are only a few examples to illustrate the type of diagnoses that can be made using apoB in combination with total cholesterol and triglycerides. The algorithm – the apoB app – can be downloaded for free from the Apple app store. Using the apoB app, the practicing clinician can easily diagnose all the apoB atherogenic dyslipoproteinemias and quickly read about the primary and secondary causes of these disorders and their therapeutic approaches.
Summary of the role of ApoB in diagnosis and cardiovascular risk management In summary, incorporating apoB with conventional lipids allows more precise estimation of cardiovascular risk, more accurate determination of the adequacy of LDL-lowering therapy, and, for the first time, rapid diagnosis of the atherogenic apoB dyslipoproteinemias. In Figure 5, the performance of apoB and non-HDL-C versus LDL-C as risk predictors, treatment targets, and diagnostic markers of lipid disorders is summarized.
Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
References 1. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18: 499–502. 2. Ketelhuth DFJ, Rios FJO, Wang Y, et al. Identification of a danger-associated peptide from apolipoprotein B100 (ApoBDS-1) that triggers innate proatherogenic responses. Circulation 2011; 124:2433–43. 3. Avraham-Davidi I, Ely Y, Pham VN, et al. ApoB-containing lipoproteins regulate angiogenesis by modulating expression of VEGF receptor 1. Nat Med 2012;18:967–73. 4. Young SG. Recent progress in understanding apolipoprotein B. Circulation 1990;82:1574–94. 5. Stalenhoef AF, Malloy MJ, Kane JP, Havel RJ. Metabolism of apolipoproteins B-48 and B-100 of triglyceride-rich lipoproteins in normal and lipoprotein lipase-deficient humans. Proc Natl Acad Sci 1984;81:1839–43. 6. Sniderman A, Vu H, Cianflone K. Effect of moderate hypertriglyceridemia on the relation of plasma total and LDL apo B levels. Atherosclerosis 1991;89:109–16. 7. Sniderman A, Tremblay A, Bergeron J, et al. Diagnosis of type III hyperlipoproteinemia from plasma total cholesterol, triglyceride, and apolipoprotein B. J Clin Lipidol 2007;1:256–63. 8. Sniderman AD, Tremblay A, De Graaf J, Couture P. Phenotypes of hypertriglyceridemia caused by excess very-low-density lipoprotein. J Clin Lipidol 2012;6:427–33. 9. Sniderman AD, St-Pierre AC, Cantin B, et al. Concordance/ discordance between plasma apolipoprotein B levels and the cholesterol indexes of atherosclerotic risk. Am J Cardiol 2003;91: 1173–7.
Crit Rev Clin Lab Sci, 2013; 50(6): 163–171
10. Contois JH, McConnell JP, Sethi AA, et al; AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Apolipoprotein B and cardiovascular disease risk: position statement from the AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Clin Chem 2009;55:407–1. 11. AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Association of apolipoprotein B and nuclear magnetic resonance spectroscopy-derived LDL particle number with outcomes in 25 clinical studies: assessment by the AACC Lipoprotein and Vascular Diseases Division Working Group on Best Practices. Clin Chem 2013;59:752–70. 12. Sniderman AD, Williams K, Contois JH, et al. A meta-analysis of low-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, and apolipoprotein b as markers of cardiovascular risk. Circ Cardiovasc Qual Outcomes 2011;4:337–45. 13. The Emerging Risk Factors Collaboration. Major lipids, apolipoproteins, and risk of vascular disease. JAMA. 2009;302: 1993–2000. 14. The Emerging Risk Factors Collaboration. Lipid-related markers and cardiovascular disease prediction. JAMA 2012;20;307: 2499–506. 15. Sniderman AD Furberg CD. Cardiovascular disease risk prediction factors. JAMA 2012;308:1969. 16. Cromwell WC, Otvos JD, Keyes MJ, et al. LDL particle number and risk of future cardiovascular disease in the Framingham Offspring Study – implications for LDL management. J Clin Lipidol 2007;1:583–92. 17. Otvos JD, Mora S, Shalaurova I, et al. Clinical implications of discordance between low-density lipoprotein cholesterol and particle number. J Clin Lipidol 2011;5:105–13. 18. Sniderman AD, Islam S, Yusuf S, McQueen MJ. Discordance analysis of apolipoprotein B and non-high density lipoprotein cholesterol as markers of cardiovascular risk in the INTERHEART study. Atherosclerosis 2012;225:444–9. 19. Sniderman A, McQueen M, Contois J, et al. Why is nonhighdensity lipoprotein cholesterol a better marker of the risk of vascular disease than low-density lipoprotein cholesterol?. J Clin Lipidol 2010;4:152–5. 20. Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the national cholesterol education program (ncep) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel iii). JAMA 2001;285: 2486–97. 21. Catapano AL, Reiner Zˇ, De Backer G, et al. ESC/EAS Guidelines for the management of dyslipidaemias: the task force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Atherosclerosis 2011;217:1–44 (and in Eur Heart J. 2011 Jul;32(14):1769-818). 22. Anderson TJ, Gre´goire J, Hegele RA, et al. 2012 Update of the Canadian Cardiovascular Society Guidelines for the Diagnosis and Treatment of Dyslipidemia for the Prevention of Cardiovascular Disease in the Adult. Can J Cardiol 2013;29: 151–67. 23. Hayward RA, Krumholz HM. Three reasons to abandon lowdensity lipoprotein targets: an open letter to the Adult Treatment Panel IV of the National Institutes of Health. Circ Cardiovasc Qual Outcomes 2012;5:2–5. 24. Baigent C, Blackwell L, Emberson J, et al.; Cholesterol Treatment Trialists’ (CTT) Collaboration. Cholesterol treatment trialists collaboration efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170 000 participants in 26 randomised trials. Lancet 2010; 376:1670–81. 25. Takagi H, Umemoto T. Limit to benefits of large reductions in low-density lipoprotein cholesterol levels: use of fractional polynomials to assess the effect of low-density lipoprotein cholesterol level reduction in metaregression of large statin randomized trials. JAMA Intern Med 2013;173:1028–9. 26. Boekholdt S, Arsenault BJ, Mora S, et al. Association of LDL cholesterol, non–HDL cholesterol, and apolipoprotein b levels with risk of cardiovascular events among patients treated with statins: a meta-analysis. JAMA 2012;307:1302–9. 27. Parish S, Offer A, Clarke R, et al; Heart Protection Study Collaborative Group. Lipids and lipoproteins and risk of different
DOI: 10.3109/10408363.2013.847897
28.
29. 30.
Critical Reviews in Clinical Laboratory Sciences Downloaded from informahealthcare.com by University of California San Francisco on 08/18/14 For personal use only.
31. 32.
vascular events in the MRC/BHF Heart Protection Study. Circulation 2012;125:2469–78. Sniderman A, Thanassoulis G, Couture P, et al. Is lower and lower better and better? A re-evaluation of the evidence from the Cholesterol Treatment Trialists’ Collaboration meta-analysis for low-density lipoprotein lowering. J Clin Lipidol 2012;6:303–9. Sniderman AD, De Graaf J, Couture P. Low-density lipoproteinlowering strategies: target versus maximalist versus population percentile. Curr Opin Cardiol 2012;27:405–11. Fredrickson DS, Levy RI, Lees RS. Fat transport in lipoproteins – an integrated approach to mechanisms and disorders. N Engl J Med 1967;276:34–44. Sniderman A, Couture P, de Graaf J. Diagnosis and treatment of apolipoprotein B dyslipoproteinemias. Nat Rev Endocrinol 2010; 6:335–46. Brunzell JD, Albers JJ, Chait A, et al. Plasma lipoproteins in familial combined hyperlipidemia and monogenic familial hypertriglyceridemia. J Lipid Res 1983;24:147–55.
ApoB versus non-HDL-cholesterol
171
33. Veerkamp MJ, de Graaf J, Hendriks JC, et al. Nomogram to diagnose familial combined hyperlipidemia on the basis of results of a 5-year follow-up study. Circulation 2004;109: 2980–5. 34. Mahley RW, Huang Y, Rall SC. Pathogenesis of type III hyperlipoproteinemia (dysbetalipoproteinemia): questions, quandaries, and paradoxes. J Lipid Res 1999;40:1933–49. 35. Fredrickson DS, Morganroth J, Levy RI. Type III hyperlipoproteinemia: an analysis of two contemporary definitions. Ann Intern Med 1975;82:150–7. 36. Zambon A, Brunzell JD, Torres A, et al. Prevention of raised low-density lipoprotein cholesterol in a patient with familial hypercholesterolaemia and lipoprotein lipase deficiency. Lancet 1993;341:1119–21. 37. Talmud PJ, Shah S, Whittall R, et al. Use of low-density lipoprotein cholesterol gene score to distinguish patients with polygenic and monogenic familial hypercholesterolaemia: a casecontrol study. Lancet 2013;381:1293–301.
