How to measure lipoprotein(a) and in whom

Lipoprotein(a): a structurally complex and challenging analyte

Lipoprotein(a) (Lp[a]) is a structurally complex lipoprotein that provides us with some significant analytical challenges. Lp(a) is structurally similar to low-density lipoprotein (LDL) in that it contains one apolipoprotein B (ApoB) molecule in each lipoprotein particle and has a similar lipid composition, being rich in cholesterol. Each Lp(a) particle also contains a single molecule of apolipoprotein(a) (Apo[a]). Apo(a) is bound to ApoB via covalent disulphide bridging and displays a large degree of variation in peptide length (figure 1).¹ The Apo(a) peptide sequence shows some similarities to plasminogen, but the protease domain in Apo(a) is inactive. Peptide sequences for both molecules consist of a series of kringle domains, which are looped polypeptides stabilised by internal disulphide bridges. Plasminogen contains five kringle domain types (I to V), while Apo(a) only contains types IV and V. There are 10 subtypes of kringle IV domains in Apo(a) with different amino acid sequences. Only one copy of kringle IV type 1 and types 3–10 are present in an Apo(a) molecule, but the number of kringle IV type 2 (KIV-2) domains varies, leading to the existence of isoforms of different molecular weight. The number of KIV-2 repeats varies from three to over 40,² so that the molecular weight of Apo(a) ranges from <250 to >650 kDa between individuals. The Apo(a) isoform present in an individual is genetically determined, which means that Apo(a) molecular weight can also vary within an individual due to the presence of two different alleles.

The molecular weight of the isoform(s) present in an individual is correlated with the number of Lp(a) particles present. Hepatic production and secretion of low molecular weight isoforms of Apo(a) is more rapid than production of higher molecular weight forms. This means that concentrations of Lp(a) by mass can be discordant relative to concentrations of Lp(a) determined by number of particles (i.e. molar units). Individuals with low molecular weight isoforms have particles of lower mass, but a greater number of particles, than those with high molecular weight isoforms, who have particles of higher mass, but a smaller number of particles. This picture becomes even more complicated when we consider that many individuals have two different Apo(a) isoforms. The complex structural variations that exist in Apo(a) help to explain the extreme variations in plasma Lp(a) concentrations, which vary by up to 1,000-fold between individuals.

Requirements of an Lp(a) assay and the difficulties in meeting them

The structural heterogeneity of Lp(a) caused by variation in the molecular weight of Apo(a), makes accurate measurement of Lp(a) in serum particularly challenging. Routine assays for measurement of Lp(a) are based on immunoassay. Immunoassays depend on an antibody–antigen reaction, where antibodies in the assay bind to target epitopes in the molecule to be measured (the analyte). Formation of an antibody–analyte complex leads to the generation of a signal, the magnitude of which is in proportion to the number of antibody–analyte complexes formed and, hence, the concentration of the analyte. The concentration of analyte is determined by comparing the strength of the signal generated in the sample to a calibration standard containing a known concentration of the target. The requirements for an accurate immunoassay include:

1. The antibodies used must be specific for the analyte and not cross-react with structurally similar molecules.
2. Assays should be calibrated according to a recognised reference material to allow comparability between assays.
3. The analyte present in the calibrators should be representative of the analyte in the patient sample in terms of molecular structure.

All three of these requirements pose a potential difficulty in the case of Lp(a). The majority of routine assays use polyclonal antibodies, so it is important to ensure that these antibodies do not cross-react with plasminogen. Polyclonal antibodies raised against intact Lp(a) may contain antibodies that react with plasminogen and ApoB, which would have to be removed in order to develop antibodies that perform well in a Lp(a) assay.³

The heterogeneous nature of Lp(a) has made establishing a well-characterised reference material assigned with appropriate values challenging. The first assays developed for the measurement of Lp(a) used calibrators with values assigned in mass units (mg/dL), which represented the total mass of Lp(a), including the LDL-like portion and Apo(a).³ The problem with this approach is that the signal generated by a well-designed assay is proportional to Apo(a) particle number, rather than mass of Lp(a). Mass and particle number are not interchangeable in the case of Lp(a) due to the large variations in molecular weight between individuals. This variation in the mass of Lp(a) is caused by, not only marked differences in the molecular weight of Apo(a) isoforms, but also differences in the lipid content (cholesterol, triglycerides, phospholipids) between Lp(a) particles. Assigning mass units to Lp(a) calibrators and assays is, therefore, fundamentally unsound, as the assays measure the protein Apo(a) and not the lipid content of the Lp(a) particle. A lack of standardised calibration practices has led to poor agreement between different Lp(a) assays,⁴ and this has confounded the interpretation of important epidemiological studies.⁵

A well-characterised reference material is now available (WHO/IFCC SRM-2B) that has an assigned value in molar units rather than mass.⁶ This material was isolated from the plasma of an individual with high Lp(a) concentrations and the molar concentration of Apo(a) was determined using amino acid composition analysis. The reference material underwent evaluation by members of an International Federation of Clinical Chemistry (IFCC) working group and was ultimately accepted as the first World Health Organisation (WHO)/IFCC International Reference Material for Lp(a). Use of this reference material to assign molar values to immunoassay calibrators affords more consistent calibration for Lp(a) assays. However, use of a single reference material for calibration of Lp(a) assays does not automatically achieve an acceptable level of between-assay agreement.Marcovina et al. evaluated the impact of using the WHO/IFCC reference material to calibrate 22 different assays for measuring Lp(a).² Uniformity of calibration was demonstrated through analysis of the reference material with each assay. Agreement between the 22 assays was very close with a between-assay relative standard deviation (RSD) of 2.8%. However, although uniformity of calibration was achieved, agreement between assays for patient samples remained poor. Between-assay RSDs for 30 patient serum samples varied from 6% to 31%. This study revealed that despite common calibration procedures, agreement was poor for patient samples due to many of the assays being sensitive to differences in the Apo(a) isoform sizes found in these samples.

Accurate measurement of Lp(a) can only be achieved if assays are designed to be isoform-insensitive. In other words, one particle of high molecular weight Lp(a) should produce the same signal in the assay system as one particle of low molecular weight Lp(a). This condition is not met by the majority of routine assays for Lp(a) as they use polyclonal antibodies that bind to the KIV-2 repeat region. In the case of higher molecular weight isoforms, more antibody–analyte complexes can be formed with the larger number of KIV-2 repeats, leading to an increased signal relative to low molecular weight forms where fewer antibody–analyte complexes can be formed. This leads to an overestimate of Lp(a) particle number in patients with high molecular weight isoforms of Apo(a) and underestimate of Lp(a) particle number in those with low molecular weight forms (figure 2).³ Patients with high molecular weight isoforms tend to have lower Lp(a) concentrations, on average, and patients with low molecular weight isoforms, higher Lp(a) concentrations. Since Lp(a)-associated cardiovascular risk is correlated with Lp(a) particle numbers, patients with higher Lp(a)-associated cardiovascular risk (low weight isoforms) suffer from underestimation of Lp(a). Patients with lower Lp(a)-associated cardiovascular risk (high weight isoforms) suffer from overestimation of Lp(a). This phenomenon has likely led to the underestimation of the degree of cardiovascular risk conferred by Lp(a) in earlier epidemiological studies.⁵ More recent studies, utilising more accurate, isoform-independent assays, have revealed a stronger relationship between Lp(a) and cardiovascular risk.^5,7

Assays that are truly insensitive to Lp(a) isoform have been developed in reference and research laboratories. The Northwest Lipid Research Laboratory (University of Washington, Seattle) have developed an ELISA (enzyme-linked immunosorbent assay) method using monoclonal antibodies in a sandwich format. In this assay, a capture antibody is directed against the KIV-2 repeat region of Apo(a) and a detection antibody is directed against the kringle IV type 9 region of Apo(a) (of which there is a single copy in each Apo(a) molecule – figure 1). This means that only one signalling antibody–analyte complex is formed for each particle of Lp(a) present in the sample, regardless of isoform size.³ An alternative approach is to use tandem mass spectrometry, which through the measurement of specific fragments of the Apo(a) molecule allows determination of both molar concentration of Lp(a) and the size of the Apo(a) isoforms present.⁸ However, while these assays are useful as reference methods, neither the ELISA nor tandem mass spectrometry assays are commercially available, or suitable for high-volume analysis in clinical laboratories (although a commercial ELISA using monoclonal antibodies against ApoB and the KIV-2 repeat region has recently become available⁹). None of the currently available commercial assays in routine use can be considered to be truly isoform-independent. The Denka Seiken assay is the least isoform-dependent commercial assay.⁷ This is achieved through the use of five independent calibrators across the working concentration range, with values traceable to the WHO/IFCC reference material. Each of these calibrators contains a distribution of isoforms likely to be encountered in patient sera with similar Lp(a) concentrations (lower concentration calibrators contain relatively high molecular weight Lp(a) isoforms and higher concentration calibrators contain lower molecular weight Lp(a) isoforms).⁴ A number of other manufacturers have used the Denka Seiken reagents to calibrate their assays and achieve acceptable agreement with isoform-independent reference assays. The Northwest Lipid Research Laboratory (University of Washington) provide a certification process that evaluates Lp(a) assay performance and traceability to WHO/IFCC SRM-2B by comparing the Lp(a) results obtained to those from the ELISA reference method.

Despite the availability of a recognised International Reference Material, and a process for determining the performance and traceability of routine methods relative to a reference method, standardisation between commercially available Lp(a) assays can still be poor. A recent study compared Lp(a) concentrations measured using six different commercially available assays and found substantial variation in the results between assays.⁴ When compared with the Denka Seiken method all of the other five assays showed a concentration-dependent bias. While some assays exhibited an acceptable degree of agreement at lower and moderate Lp(a) concentrations, they all displayed a significant bias at very high Lp(a) concentrations. There was also variation in the units of reporting used in the assays, with some reporting in molar units and others in mass units.

A number of recommendations that have been suggested in order to improve harmonisation of Lp(a) assays include:^3,10

• All assays should use calibrators with values traceable to the WHO/IFCC reference material in molar units.
• Assays that are not inherently isoform-independent should use a Denka Seiken calibration process, using a number of independent calibrators of varying concentration containing appropriate Apo(a) isoforms.
• Ensure the accuracy of commercial assays has been certified as WHO/IFCC traceable with acceptable performance in relation to variations in the Lp(a) isoforms present.
• Molar units should not be converted to mass units (assays with calibrator concentrations assigned with mass units cannot be considered standardised).

Addressing the problem of poor standardisation and reliability of the available assays for Lp(a) has become more important as measurement of Lp(a) in order to detect increased cardiovascular risk associated with raised Lp(a) is recommended in guidance from a number of bodies.^11–13 The availability of accurate Lp(a) assays will be vital, if and when, medical therapies specifically targeting Lp(a) reduction become available, both to identify patients likely to benefit from such therapy and to monitor the effectiveness in terms of Lp(a) reduction.

In whom should we measure Lp(a)?

In recent years it has been clearly proven that Lp(a) concentrations are a significant risk factor for cardiovascular disease. Historically, the impact of Lp(a) on cardiovascular risk has been controversial and underestimated in epidemiological studies due, at least in part, to the difficulties in the accurate measurement of Lp(a) described above.⁵ Serum Lp(a) concentrations are predominantly genetically determined and remain relatively stable throughout life. Currently, there are no medical therapies that are specifically targeted at lowering Lp(a) concentrations. PCSK9 (proprotein convertase subtilisin/kexin type 9) inhibitors can reduce Lp(a) concentrations by 20–30%,¹⁴ and there is some evidence that subjects with high Lp(a) treated with these agents have a reduced cardiovascular risk. However, it is not clear if this reduction in risk is due to lowering Lp(a) or LDL-cholesterol concentrations. Lipoprotein apheresis has been recommended as a means of reducing Lp(a) concentrations in those with high Lp(a) and progressive coronary disease.¹¹ Apheresis therapy is a cumbersome and expensive form of treatment, however, and only available in specialist centres. Although it is difficult to modify Lp(a) concentrations, measurement of Lp(a) is recommended in certain populations in a number of guidelines,^11–13 as modifiable CVD risk factors (LDL-cholesterol, blood pressure, smoking, etc.) can be aggressively targeted with the aim of mitigating the additional risk associated with raised Lp(a).

A HEART UK consensus statement published in 2019¹¹ recommends measuring Lp(a) in specific population groups at high risk of CVD. These groups are individuals with:

• A personal or family history of premature atherosclerotic cardiovascular disease (<60 years of age).
• First degree relatives of patients with raised serum Lp(a) levels (>200 nmol/L).
• Familial hypercholesterolaemia or other inherited dyslipidaemias.
• Calcific aortic valve stenosis.
• A borderline increase (but <15%) 10-year risk of a cardiovascular event.

Similar recommendations have been made in a US National Lipid Association statement,¹² which also recommends measurement of Lp(a) in similarly defined groups as well as patients suffering premature ischaemic stroke (<55 years). Recommendations for measurement of Lp(a) in a European Atherosclerosis Society consensus statement are also consistent with the HEART UK recommendations.¹³ A recent joint European Society of Cardiology/European Atherosclerosis Society guideline on the management of dyslipidaemias goes as far as recommending that Lp(a) is measured at least once in all adults.¹⁵ The justification provided for this is that individuals with very high Lp(a) have a similar lifetime CVD risk to individuals with heterozygous familial hypercholesterolaemia.

Defining the ‘normal’ range for serum Lp(a) concentrations, and at what level individuals should be considered at higher CVD risk, is difficult. This is, in part, due to the lack of standardisation between assays and use of different units. There is also evidence of significant differences in Lp(a) distribution between different ethnic groups.¹⁶ Various cut-offs for defining individuals as ‘at-risk’ have been suggested in the literature, but the likely optimal approach is to base decision limits on population centiles. The US National Lipid Association statement¹³ suggests that Lp(a) concentrations above 100 nmol/L (or 50 mg/dL) should be considered as a risk factor for CVD. This is based on the 80^th percentile of a largely Caucasian population. The 80^th percentile for an African American population is significantly higher (and lower for Japanese Americans), but it is not clear if different decision limits should be used for different ethnicities. The HEART UK statement recommends grading the increased risk of CVD conferred by Lp(a) concentrations as follows: 32–90 nmol/L minor; 90–200 nmol/L moderate; 200–400 nmol/L high; >400 nmol/L very high.¹² The authors note that these values are based on percentile cut-offs determined by the ongoing Copenhagen General Population Study. Measurements in this study were carried out using the Roche assay. In view of the poor standardisation of Lp(a) assays, these values should be transferred between centres using different assays with caution.

Conclusion

Lp(a) is a complex and heterogeneous lipoprotein, which is challenging to measure accurately due to the large degree of variation in the molecular weight of Apo(a) isoforms between individuals. Accurate measurement of Lp(a) is vital in determining which patients are at increased risk of CVD as a result of raised Lp(a) concentrations, but the currently available commercial assays suffer from inaccuracies due to poor standardisation and sensitivity to isoform size. Various measures, such as ensuring traceability to the WHO/IFCC reference material (in molar units) and using a calibration scheme that minimises the impact of isoform-sensitivity, should be implemented to improve the comparability between Lp(a) assays.

UK, US and European consensus statements all recommend the measurement of Lp(a) in certain high-risk populations to identify individuals with raised Lp(a), so as to allow them to be targeted with aggressive management of other CVD risk factors. Accurate measurement of Lp(a) is likely to become even more important in the future as clinical trials of medical therapies targeted specifically at lowering Lp(a) are underway.

Key messages

• The unique structure of lipoprotein(a) (Lp[a]) and variable molecular weight of apoliprotein(a) (Apo[a]) isoforms make accurate measurement of Lp(a) a challenge
• Current commercial assays for Lp(a) are not well standardised and are affected by variations in the molecular weight of the Lp(a) isoforms present
• All assays for Lp(a) should be traceable to the WHO/IFCC international standard and calibrated using independent calibrators to minimise isoform sensitivity
• A HEART UK consensus statement recommends measurement of Lp(a) in certain high-risk populations
• Patients with raised Lp(a) may benefit from aggressive management of modifiable cardiovascular risk factors

Conflicts of interest

None declared.

Funding

None.

Christopher S Boot
Consultant Clinical Scientist

Blood Sciences, Royal Victoria Infirmary, Newcastle upon Tyne Hospitals NHS Foundation Trust, Queen Victoria Road, Newcastle upon Tyne, NE1 4LP

([email protected])

Articles in this supplement

Lipoprotein(a): a historical perspective
Lipoprotein(a): mechanisms of pathogenicity
Lipoprotein(a): marker and target in calcific aortic valve disease
Current management of the patient with high lipoprotein(a)

References

1. van der Hoek YY, Wittekoek ME, Beisiegel U et al. The apolipoprotein(a) kringle-IV repeats which differ from the major repeat kringle are present in variably-sized isoforms. Hum Mol Genet 1993;2:361–6. https://doi.org/10.1093/hmg/2.4.361

2. Marcovina SM, Albers JJ, Scanu AM et al. Use of a reference material proposed by the International Federation of Clinical Chemistry and Laboratory Medicine to evaluate analytical methods for the determination of plasma lipoprotein(a). Clin Chem 2000;46:1956–67. https://doi.org/10.1093/clinchem/46.12.1956

3. Marcovina SM, Albers JJ. Lipoprotein (a) measurements for clinical application. J Lipid Res 2016;57:526–37. https://doi.org/10.1194/jlr.R061648

4. Scharnagl H, Stojakovic T, Dieplinger B et al. Comparison of lipoprotein (a) serum concentrations measured by six commercially available immunoassays. Atherosclerosis 2019;289:206–13. https://doi.org/10.1016/j.atherosclerosis.2019.08.015

5. Rifai N, Ma J, Sacks FM et al. Apolipoprotein(a) size and lipoprotein concentration and future risk of angina pectoris with evidence of severe coronary atherosclerosis in men, The Physicians Health Study. Clin Chem 2004;50:1364–7. https://doi.org/10.1373/clinchem.2003.030031

6. Dati F, Tate JR, Marcovina SM, Steinmetz A. First WHO/IFCC international reference reagent for lipoprotein(a) for immunoassay – Lp(a) SRM 2B. Clin Chem Lab Med 2004;42:670–6. https://doi.org/10.1515/CCLM.2004.114

7. Erqou S, Kaptoge S, Perry PL et al. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. JAMA 2009;302:412–23. https://doi.org/10.1001/jama.2009.1063

8. Lassman ME, McLaughlin TM, Zhou HH et al. Simultaneous quantitation and size characterization of apolipoprotein(a) by ultra-performance liquid chromatography/mass spectrometry. Rapid Commun Mass Spectrom 2014;28:1101–06. https://doi.org/10.1002/rcm.6883

9. Contois JH, Nguyen RA, Albert AL. Lipoprotein(a) particle number assay without error from apolipoprotein(a) size isoforms. Clin Chim Acta 2020;505:119–24. https://doi.org/10.1016/j.cca.2020.02.030

10. Cegla J, France M, Marcovina SM, Neely RDG. Lp(a): when and how to measure it. Ann Clin Biochem 2021;58:16–21. https://doi.org/10.1177/0004563220968473

11. Cegla J, Neely RDG, France M et al. HEART UK consensus statement on lipoprotein (a): a call to action. Atherosclerosis 2019;291:62–70. https://doi.org/10.1016/j.atherosclerosis.2019.10.011

12. Wilson DP, Jacobson TA, Jones PH et al. Use of lipoprotein(a) in clinical practice: a biomarker whose time has come. A scientific statement from the National Lipid Association. J Clin Lipidol 2019;13:374–92. https://doi.org/10.1016/j.jacl.2019.04.010

13. Nordestgaard BG, Chapman MJ, Ray K et al. Lipoprotein(a) as a cardiovascular risk factor: current status. Eur Heart J 2010;31:2844–53. https://doi.org/10.1093/eurheartj/ehq386

14. Raal FJ, Giugliano RP, Sabatine MS et al. Reduction in lipoprotein(a) with PCSK9 monoclonal antibody evolocumab (AMG 145). J Am Coll Cardiol 2014;63:1278–88. https://doi.org/10.1016/j.jacc.2014.01.006

15. Mach F, Baigent C, Catapano AL et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J 2020;41:111–88. https://doi.org/10.1093/eurheartj/ehz455

16. Kraft HG, Lingenhel A, Pang RWC et al. Frequency distributions of apolipoprotein(a) kringle IV repeat alleles and their effects on lipoprotein(a) levels in Caucasian, Asian, and African populations: the distribution of null alleles is non-random. Eur J Hum Genet 1996;4:74–87. https://doi.org/10.1159/000472175