Lipid measurement is done with the main goal of identifying and treating individuals with or at risk for atherosclerotic cardiovascular disease (ASCVD).(1) Cholesterol circulates in plasma within lipoprotein particles, of which apolipoproteins are essential structural and functional components. There are two main isoforms of apoB found on lipoproteins: apoB-48, which is present in lipoproteins of intestinal origin including chylomicrons and chylomicron remnants, and apoB100, which is found in lipoproteins of hepatic origin including very-low-density lipoprotein (VLDL), intermediate-density lipoproteins (IDL), low-density lipoprotein (LDL), and lipoprotein (lp(a)).(2) While chylomicrons are too large to enter the arterial wall, the other apoB containing lipoproteins, including chylomicron remnants, can be trapped in the arterial wall and deposit cholesterol, therefore driving the atherogenesis process.(3) High-density lipoprotein (HDL) particles do not contain apoB. Because each atherogenic lipoprotein particle only contains one molecule of apoB, the plasma apoB measurement provides a measurement of the total number of atherogenic particles in plasma (Figure 1).(4)LDL-C particles derived from the VLDL lipolysis, contain different types of lipids and proteins -including apoB- in a micelle-like arrangement (Figure 1). Different types of assays have been developed to quantify LDL in plasma, the most widely used being LDL-C, which quantifies the amount of cholesterol carried collectively by the LDL lipoproteins.(5)All atherogenic lipoproteins, including VLDL and LDL, contain a single apoB molecule (Figure 1). Accordingly, measurement of apoB provides a direct assessment of the number of atherogenic lipoprotein particles. The use of apoB clinically, however, remains low. Rather, mass-based estimates of LDL-C (the total mass of cholesterol carried collectively by all LDL particles) have been, and remain, the primary measure used to guide treatment.(6–11) LDL-C can be measured directly with ultracentrifugation, but is usually calculated indirectly based on total cholesterol, HDL-C, and triglycerides (TG) using one of several formulas.There are multiple reasons for apoB to be used clinically. Although LDL-C, and apoB are highly correlated, a large body of prior research indicates that apoB is a better predictor of ASCVD risk than LDL-C for several reasons.(3, 12, 13) First, the cholesterol mass per LDL particle is variable – as shown in figure 2-, which can lead to cholesterol-depleted or cholesterol-enriched lipoproteins which vary in their atherogenicity. Persons with large number of cholesterol-depleted lipoproteins (and a high apoB) may have an apparently low LDL-C despite being at higher risk of ASCVD.(14) Similarly, patients may have high LDL-C levels but low apoB due to the presence of cholesterol-enriched lipoproteins. In these patients, ASCVD risk is over-estimated by LDL-C.(2) Second, LDL-C measurements fail to capture the risk conferred by VLDL and IDL particles, which are also atherogenic.(15) As illustrated in Figure 2 with two hypothetical cases, patients with the same level of LDL-C can have different risk profiles based on the number of atherogenic particles in plasma, therefore the importance of measuring apoB.
Another reason for using apoB clinically is that it does not suffer the same challenges related to measurement as LDL-C. LDL-C is usually estimated from total cholesterol, HDL cholesterol, and TG, based on a formula. The most commonly used formula, the Friedewald equation, has known limitations, particularly in patients with hypertriglyceridemia. Newer formulas including the Martin/Hopkins equation and the Sampson method have been proposed, but there remains uncertainty about the ideal method. Further, all methods have limited performance at very low levels of LDL-C, which is now increasingly common with the introduction of novel lipid lowering therapies,(16–18) and in the presence of severe hypertriglyceridemia (in whom fasting LDL-C measurements are recommended).(19) In contrast, the measurement of apoB is standardized, accurate, inexpensive, and can be applied on automated chemistry platforms widely available in the clinical laboratories.(20) The International Federation of Clinical Chemistry and the World Health Organization endorsed the standardization of apoB measurement in 1994, introducing a reference material that minimized variability across laboratories.(21) Nevertheless, to further enhance the standardization, there are ongoing efforts to develop a new Reference Measurement System employing direct apoB measurement through mass spectrometry.(22)
ApoB is also unaffected by fasting state. Although the apoB immunoassay measures both apoB100 (which is found on VLDL, IDL, and LDL particles) and apoB48 (which is found on chylomicrons), even in a postprandial state the number of chylomicrons relative to apoB100 containing particles is low.(10) On average, there are 9 LDL apoB100 particles for every VLDL apoB100 particle. Likewise, there are 9 VLDL apoB100 particles for each apoB48 particle. This ratio underlines why total apoB, even in a postprandial state, essentially equates to total apoB100 and is largely unaffected by fasting.(3)
In epidemiologic studies, apoB outperforms LDL-C to predict ASCVD. The first evidence that apoB serves as a more precise indicator of risk emerged in 1980, with a study showing that apoB, particularly LDL apoB, more effectively distinguished between patients with and without coronary atherosclerosis.(23) Since then, the evidence supporting superiority of apoB has grown significantly. A meta-analysis by Sniderman et al of 12 epidemiological studies containing estimates of relative risk of ischemic cardiovascular events of different lipid markers, found that apoB was the most potent marker of cardiovascular relative risk ratio compared to non-HDL.(24) Additionally, changes in apoB may better capture the potential benefit of lipid lowering therapy. A meta-analysis of 29 randomized clinical trials involving 332,912 patients on lipid-lowering therapy (statins, ezetimibe, PSCK9 inhibitors, cholesterylester transfer protein inhibitors, fibrates, niacin, and n-3 fatty acids) demonstrated that absolute reduction in apoB was associated with decreased all-cause and cardiovascular mortality (relative risk for every 10mg/dL decrease in apoB of 0.95: 0.92–0.99 and 0.93:0.88–0.98, respectively).(25) Moreover, Thanassoulis et al., using a frequentist and a Bayesian approach with data from 7 placebo-controlled statin trials, found that mean relative risk reduction of cardiovascular events per standard deviation change in lipid marker from statin therapy was more closely related to reductions in apoB than to reductions in either non-HDL-C or LDL-C.(26) In the ODYSSEY-Outcomes trial, achieved levels of LDL-C or non–HDL-C were not predictive of major adverse cardiovascular events (MACE) after taking achieved apoB into account.(18) Data from other trials have also supported the superiority of apoB over LDL-C as predictor of cardiovascular events. For example, an analysis that included individuals from the UK biobank (primary prevention cohort) and FOURIER and IMPROVE-IT trials (secondary prevention cohort) demonstrated that apoB was the only lipid parameter independently associated with incident myocardial infarction.(13) The risk captured by apoB was independent from the type of lipid (cholesterol or TG) or lipoprotein (either LDL or TG-rich as VLDL and IDL) across both cohorts.(13) Furthermore, data from the Copenhagen General Population Study, showed that in statin-treated patients, apoB was a superior marker for all-cause mortality risk than both non-HDL-C and LDL-C.(27)
Using non-HDL-C instead of LDL-C can help improve risk prediction by capturing the burden of non-LDL, apoB containing lipoproteins such as VLDL and IDL particles. However, there are still a significant proportion of individuals (ranging between 8–23%) in whom apoB and non-HDL-C remain discordant, and in whom apoB remains a better predictor of ASCVD risk.(28) Welsh et al., analyzing data from the UK biobank for patients without cardiovascular disease and not taking statins, found that 18% of participants had discordant apoB and LDL-C values (defined as ≥10% absolute difference in baseline percentile of direct LDL and apoB). In those patients, only apoB was associated with increased ASCVD risk (adjusted hazard ratio per standard deviation, 1.23: 1.12–1.35, P<0.001), but not directly measured or calculated LDL-C (by either Friedewald or Martin-Hopkins formula), nor non-HD-CL.(29)
Measurement of apoB is also needed for accurate differential diagnoses and targeted treatment of hypertriglyceridemia, i.e. to differentiate hypertriglyceridemia due to excess VLDL particles vs hyperchylomicronemia.(30) Furthermore, for the identification of Familial Dysbetalipoproteinemia, also known as Type III Hyperlipoproteinemia, apoB measurement is necessary. This is a highly atherogenic and treatable dyslipoproteinemia characterized by elevated levels of both cholesterol and triglycerides due to a significant increase in cholesterol-rich chylomicron and VLDL remnant lipoprotein particles.(30) Utilizing apoB alongside a traditional lipid panel allows for the reliable diagnosis of Type III Hyperlipoproteinemia and its differentiation from mixed hyperlipidemia.(31)
How have guidelines incorporated apoB?
Although major clinical guidelines for management of blood cholesterol focus on LDL-C, measurement of apoB is increasingly recommended in certain populations, in particular those with elevated TG.(6, 8) Specific apoB measurement recommendations and statements about apoB measurement by the principal lipid guidelines are shown in Table 1.
Canadian and European guidelines have recommended apoB measurement in certain populations for many years. Since 2012, the Canadian Cardiovascular Society (CCS) guideline for the diagnosis and treatment of dyslipidemia has recognized the utility of measuring apoB in people with high TG (defined as TG>1.5mmol/L – 131mg/dl- in the CCS guideline).(33) In the last (2021) guideline, the CCS strongly recommended using either apoB or non-HDL-C instead of LDL-C as the preferred lipid parameter for screening in any patient with TG > 1.5mmol/L (133mg/dL).(6) The guideline further recognized that there could be discordance between apoB and non-HDLC in some patients, and suggested that the decision on which to use should be guided by availability and clinical expertise. Meanwhile, since 2016, the European Society of Cardiology and the European Atherosclerosis Society (ESC/EAS) recommended considering apoB, when available, in patients with high TG, and expanded this recommendation in the 2019 guideline.(8, 34) In this most recent document, ESC/EAS recommend apoB measurement as an alternative to LDL-C (and may even be preferred over non-HDL-C) as the primary marker for screening, diagnosis, and management in people with high TG levels, diabetes (DM), obesity, or very low LDL-C (Class I Level C). The 2019 ESC/EAS guideline stated that in the aforementioned individuals, LDL-C may underestimate ASCVD risk by underestimating both the total concentration of cholesterol carried by LDL and the total concentration of apoB-containing lipoproteins Compared with Canadian and European guidelines, American guidelines have been more limited in recommending measurement of apoB.(7) One of the oldest recommendations for utilization of apoB in the United States comes from the American Association for Clinical Chemistry (AACC), whose 2009 Lipoproteins and Vascular Diseases Division Working Group on Best Practices statement supported the use of apoB for assessing and monitoring patients at risk for cardiovascular disease. In that statement, the authors state that “changing perceptions and practice will not be easy, considering that physicians and patients are accustomed to LDL-C. Significant education efforts will be required”.(10) Indeed, over the subsequent decade, LDL-C remained the main recommended measure of lipid related risk by US Guidelines. The 2018 American Heart Association and the American College of Cardiology, in collaboration with multiple leading medical organizations (AHA/ACC Multisociety Guideline) state that those with TG ≥200 mg/dL have a “relative indication” for apoB measurement.(7) The guideline also states that an apoB >130mg/dL should be considered a “risk enhancing factor” for considering lipid-lowering therapy but do not recommend universal testing. The 2020 American Association of Clinical Endocrinologists and the American College of Endocrinology (AACE/ACE) guideline stated that apoB may provide a more accurate assessment of atherogenicity as apoB can be elevated in individuals with normal LDL-C, particularly in those with insulin resistance who could often have small-dense LDL.(9) It also recognized that apoB is useful in assessing the success of lipid-lowering therapy, as it may remain above the goal after achieving the LDL-C target. However, the recommendation to measure apoB– or LDL particles, lp(a), and high sensitivity C reactive protein-was only made “based on individual patient clinical circumstances”. In 2021 a scientific statement from the National Lipid Association (NLA) endorsed that the measurement of apoB is reasonable, not only as part of the initial lipid evaluation, but also in patients on lipid therapy.(32) The document recognized that in high-risk patients with LDL-C <70mg/dL on lipid-lowering therapy, apoB measurement may effectively identify a patient with an increased ASCVD risk due to residual atherogenic lipoprotein burden who could benefit from additional therapy. While various guidelines have recognized value in measuring apoB, there is considerably less agreement in apoB thresholds for risk/treatment.
Guidelines consistently support that apoB is a more accurate predictor of cardiovascular risk, particularly in the presence of cholesterol-depleted apoB particles and is recommended as the marker of choice for individuals with high TG levels. However, it should be noted that there is still no unanimous agreement on the specific TG levels that define hypertriglyceridemia. Evidence also indicates that cholesterol-depleted apoB lipoproteins can occur outside of cases with high TG levels. In further exploring the link between TG, apoB, and LDL-C, De Marco et al., analyzed LDL-C/apoB ratios in a cohort of 6,272 participants from National Health and Nutrition Examination Survey (NHANES).(35) They found that cholesterol-depleted apoB particles, as indicated by a low LDL-C/apoB ratio, can be present across all TG levels, reporting that 21.4% of individuals with TG levels <100 mg/dL exhibited a low LDL-C/apoB ratio.(35) These findings reinforce the argument for the routine inclusion of apoB in clinical assessments for all patients, not just those with elevated TG levels.
Guideline-based LDL-C targets vary significantly, partly based on guideline publication year and temporal updates from iterative trial data, and partly due to differences between guidelines. However, even when the same LDL-C target is used, (i.e. <70 mg/dL) guidelines vary in the corresponding apoB level. This reflects the lack of a universally agreed upon conversion between apoB and LDL-C and different possible ways to establish an apoB target. Table 2 synthesizes the target/trigger levels present in different guidelines.
The 2021 CCS guideline established apoB-based recommendations by generating percentile equivalents of apoB and non-HDL-C for each of their recommended LDL-C target using data from NHANES and UK Biobank.(6) They recommend starting statin treatment with an LDL-C ≥5.0mmol/L (193.35mg/dL) or an apoB≥145mg/dL in low-risk patients, and an LDL-C ≥3.5mmol/L (135.34mg/dL) or an apoB of ≥105mg/dL in intermediate or low-risk patients with other cardiovascular risk factors. The CCS also recommended using an LDL-C or apoB level of ≥70mg/dL as a threshold to intensify lipid-lowering treatment in patients with ASCVD. On the other hand, the 2019 ESC/EAS guideline establishes secondary apoB goals of <65, <80, and <100 mg/dL for very-high-, high-, and moderate-risk patients, respectively, which corresponded to LDL-C targets of <55 mg/dL, <70 mg/dL and <100 mg/dL.(8)
In the USA, the AHA/ACC Multi-Society Cholesterol Guideline states that an apoB level ≥130 mg/dL corresponds to an LDL-C ≥160 mg/dl and constitutes a risk-enhancing factor (particularly when accompanied by persistently elevated TG) for which statin initiation or intensification can be considered.(7) Meanwhile, the AACE/ACE suggests an apoB goal of <70, <80, and <90 mg/dl for individuals at extremely high-risk (persistent ASCVD), very high-risk (established ASCVD or DM plus ≥1 additional risk factor), and high and moderate-risk (risk of ASCVD, including those with DM), respectively.(9) In the 2015 NLA Recommendations for Patient-Centered Management of Dyslipidemia, the apoB recommendations are < 80 mg/dL for high-risk patients and < 90 mg/dL in primary prevention, both based on NHANES III data.(11) In 2009, an AACC working group established apoB cut points equivalent to LDL- cut points based on Framingham Offspring Study.(10) In this document, apoB levels of 80 and 100 mg/dL are considered equivalent to LDL-C levels of 100 and 130mg/dL.
Population percentiles of apoB relative to LDL-C
ApoB equivalents for LDL-C have been determined through various approaches using population percentiles. One method involves determining corresponding apoB values in individuals with a particular LDL-C level, while the other involves aligning the LDL-C percentile corresponding to a particular LDL-C value and determining the same percentile value for apoB in that population. One challenge to this approach globally is that the population distribution of apoB in a population varies by region. Figure 3 shows the levels of apoB corresponding to the 10th, 50th, and 90th population percentiles based on studies from the United States, Canada, South Korea, India, Mexico, Finland, and Sweden.(43–49) Not only do apoB levels vary in different populations, so do the levels of LDL-C and non-HDL-C.(50, 51) The CCS was the only guideline that established the apoB triggers based on LDL-C population percentiles equivalents using data from NHANES and the UK Biobank.(6) In contrast, the other guidelines, the equivalent apoB level suggested relative to LDL-C, can be significantly higher in terms of population percentile.(52) For example, the ESC/EAS, AACE/ACE, and NLA guidelines all use an apoB of 80 mg/dL to correspond to an LDL-C of 70 mg/dL. However, based on NHANES and the Very Large Database of Lipids, a LDL-C of 70 mg/dL corresponded to a population-percentile–equivalent apoB value of 60 mg/dL (≈7th–9th percentile), while an apoB of 80 mg/dL corresponded to a population-percentile–equivalent LDL-C value of 100 mg/dL (≈31st–36th percentile).(52)
Discordance between LDL and apoB
As mentioned above, there can be significant variability in the amount of cholesterol carried by every lipoprotein, hence; apoB, LDL-C, and non-HDL-C are not equivalent cardiovascular risk markers.(53) For most patients, apoB and LDL-C are highly correlated, such that using LDL-C accurately captures the majority of the patient’s atherogenic risk. Multiple “discordance analyses” have attempted to identify patients with discordant apoB and LDL-C levels and analyze their risk of cardiovascular events. As there is no definitive ratio between LDL-C and apoB that defines a cholesterol-depleted vs cholesterol-enriched lipoprotein phenotype, most of the discordance studies have used median values of LDL-C and apoB to establish four different groups (concordantly high and low, and discordantly high and low apoB in relation to LDL-C or non-HDL-C) to compare cardiovascular outcomes between them. Other studies have used a variety of other methods. The manner of defining discordance, as well as the results of these discordance analyses are summarized in Table 3.(27, 28, 54–61)
Because of differences in the populations studied as well as the definition of discordance, the cutoffs used to identify adults with “discordantly high” and “discordantly low” apoB varied. Most of the studies utilized median values to define discordance, therefore there was not a specific difference between apoB and LDL-C to consider someone “discordantly high”, and the difference between apoB and LDL-C levels in these groups ranged from −11.5 to 33.52. The percentage of people with discordantly high apoB also ranged from 5.3–24.9%. Overall, men, people with metabolic syndrome, DM, hypertension, smokers, higher BMI, higher levels of TG, and lower levels of HDL were most likely to have discordantly high apoB relative to LDL-C, highlighting the need to test apoB in these populations as LDL-C can be insufficient to fully capture ASCVD risk.
ApoB levels achieved in randomized clinical trials.
While LDL-C has been the main target for clinical trials of lipid lowering therapy, many trials have also measured apoB both at baseline and on-treatment. Data from trials that reported apoB at baseline and achieved apoB levels are presented in Table 4 and Figure 4. While the degree of LDL-C lowering varied across and within therapeutic classes, in statin, ezetimibe, PCSK9i, inclisiran and bempedoic acid trials the proportional reduction in LDL-C was greater than the proportional reduction in apoB. The degree of apoB reduction was highly correlated with the degree of LDL-C reduction, and the correlation between the % reduction in apoB and the % reduction in LDL-C across all other trials is high (R2=0.85). Overall, across all trials, the % reduction of apoB was slightly lower than LDL-C. For statin, PCSK-9i, inclisiran, and bempedoic acid trials, the mean reduction in apoB was ~6–8% lower than the mean reduction in LDL-C, meanwhile, for the ezetimibe trials, the mean reduction in LDL-C was 16% higher than the apoB reduction. This difference could be explained by the mechanism of action of ezetimibe – inhibiting cholesterol absorption from the small intestine – or due to differences in the populations enrolled in trials.
With respect to what the actual target apoB level should be, trial data can provide further evidence. Figure 5 shows the achieved LDL-C and apoB levels across statin, ezetimibe, PCSK9i, inclisiran and bempedoic acid trials from Table 3. Across all trials, the on-treatment LDL-C and apoB levels attained are highly correlated (R2=0.85). For most trials of statins, ezetimibe, and inclisiran on-treatment LDL-C levels were slightly lower than on-treatment apoB levels, while for PCSK9i and bempedoic acid, on-treatment LDL-C levels were slightly higher than on-treatment apoB levels. However, across all trials, the on-treatment LDL-C levels and on-treatment apoB levels were roughly equivalent.
Across clinical trials, the levels of apoB achieved in the treatment arms has varied based on the population being studied and the type of treatment used, though trial data have consistently shown that with respect to apoB and LDL-C, lower is better. For instance, in statin trials, apoB levels ranged from 119 mg/dL in the REVERSAL trial (Pravastatin 40 mg) to 66 mg/dL in the JUPITER trial (Rosuvastatin 20 mg).(63, 71) When ezetimibe was combined with simvastatin in the IMPROVE-IT trial, the apoB level reached 70 mg/dL.(38) Contemporary outcomes trials of PSCK-9 inhibitor monoclonal antibodies achieved the lowest on-treatment apoB levels, with the ODYSSEY and FOURIER trials averaging 39 mg/dL and 45 mg/dL, respectively.(75–77) The latter two trials would suggest a target apoB level in secondary prevention of at least <55 mg/dL, consistent with current European guideline recommendations.(7)
Summary of Results and apo B Targets
The totality of data suggests that apoB may be a superior marker not only of risk of ASCVD but also of the benefit of lipid lowering therapy compared with LDL-C. The use of apoB has also practical benefits worth mentioning, including that patients don’t need to fast, and is more accurate than LDL-C in patients with high TG or very low LDL-C levels. As guidelines are increasingly recommending more aggressive lipid reduction, obtaining precise lipid measurements is essential. This has led many to call for the preferential use of apoB over LDL-C. However, to date, consistent recommendations for how clinicians should use apoB results remain lacking, and different guidelines have provided variable apoB targets for given LDL-C levels.
Previously, some guidelines have compared population or study-specific percentiles of LDL-C and apoB to identify equivalent levels. While this approach has a benefit of potentially identifying similar numbers of individuals for treatment, variability across populations in the distribution of LDL-C and apoB could lead to continued variability in global guidelines for optimal apoB targets. Discordance analyses have confirmed that persons with metabolic syndrome and hypertriglyceridemia are at highest risk of discordantly high apoB and reinforce current guideline recommendations to measure apoB in these populations. Unfortunately, variability across analyses in how discordance was defined limits their utility in guiding clinical management. However, all discordance analyses demonstrated that regardless of LDL-C, those with elevated apoB are at increased risk of ASCVD, reinforcing the importance of measurement of apoB.
Given this, trial data appear to be most useful in developing clinical recommendations for how to implement apoB to guide treatment recommendations. In terms of proportional reduction, all the lipid lowering therapy included appear to reduce apoB slightly less than LDL-C. Thus, if targeting a 50% or more LDL-C reduction (as recommended by many guidelines) with statins, ezetimibe, PSCK-9i, inclisiran, or bempedoic acid, a comparable apoB reduction would be 40–45%.
To guide treatment targets, the trial data provide a very straightforward path. As the on-treatment LDL-C levels and on-treatment apoB levels were nearly identical, the same number can be used for apoB as LDL-C. Thus, if a target LDL-C level is <70 mg/dL, the target apoB should also be <70 mg/dL (0.70 g/L). In high-risk patients with a target of LDL-C <55mg/dL, the corresponding goal for apoB should be <55mg/dL (0.55 g/L). This simple approach can be easily communicated to patients and treating physicians and could eliminate confusion between guidelines.
Conclusions
The use of apoB offers a standardized, accurate, and cost-effective measurement of the total number of atherogenic lipoprotein particles in plasma, providing a more accurate assessment of ASCVD risk and effectiveness of lipid lowering therapy. Unlike LDL-C, apoB is particularly advantageous in patients with discordant lipid profiles, such as those with high TG, and insulin resistance, who are more likely to have cholesterol depleted apoB particles. ApoB is also more accurate when the levels of LDL-C are lower, a situation very common now days in clinical practice when guidelines suggest more aggressive treatment targets. Despite this, apoB faces barriers to widespread adoption in clinical practice, primarily due to the absence of consistent guidance on its interpretation and application. To address the existing challenges and based on evidence in RCT, this review suggests the use of the same number for LDL-C as for apoB when establishing treatment targets. The simplification of the interpretation of apoB, by aligning apoB targets with LDL-C levels, could enhance its integration into routine practice. The use of apoB in clinical decision-making holds promise for more accurate ASCVD risk prediction and tailored lipid-lowering strategies.
AK
Dr. Amit Khera
Cardiology · Preventive CardiologyThe University of Texas Southwestern Medical Centeramitk
