Reference Materials for the Standardization of the Apolipoproteins A-I and B, and Lipoprotein(a).

Measurement of the apolipoproteins A-I and B, and lipoprotein(a) enable identification of individuals at increased risk of cardiovascular disease. However, the lack of standardized methods to measure these risk markers has resulted for many years in the non-comparability of values and often a conflicting interpretation of clinical studies. Due to the collaborative efforts of the International Federation of Clinical Chemistry and Laboratory Medicine, research organizations, clinical chemistry laboratories and diagnostic companies, secondary reference materials for the apolipoproteins A-I and B, and lipoprotein(a) have been prepared and tested for their ability to harmonize test values. SP1-01 and SP3-07 WHO-IFCC reagents are now available to manufacturers for use in the value transfer of apolipoprotein A-I and B values to master calibrators, and PRM is proposed to be the secondary reference material for Lp(a). By the worldwide use of such reference materials a better traceability and standardization of measurement is being achieved in the clinical laboratories.

Introduction

Apolipoprotein A-I (apo A-I), the major protein in high-density lipoproteins (HDLs), apolipoprotein B (apo B), the major protein in low-density lipoproteins (LDLs), and lipoprotein(a) [Lp(a)], can serve as important predictors of cardiovascular disease risk. Until recently, the lack of internationally accepted standardization impeded the broad application of apolipoproteins in laboratory medicine. The International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) through its Committee on Apolipoproteins and Working Group on Lp(a), and together with research institutions and several diagnostic companies have succeeded in their effort to achieve a consensus on a practical standardization procedure. This included the preparation of suitable secondary reference preparations needed for calibrating all commercially available immunoassays for measurement of apo A-I and B, and Lp(a).

Lipoproteins and apolipoproteins: Classification and properties

Disorders of lipid metabolism are important causes of atherosclerosis (1). Atherogenic risk is associated with changes in the composition of blood lipids (cholesterol, triglycerides, and phospholipids), and in their transport forms, i.e., the lipoproteins (2). Different atherogenicity is attributed to the different lipoprotein classes with LDLs considered as atherogenic, and HDLs as protective against atherosclerotic vascular changes caused by accumulation of cholesterol in the arteries (3, 4). Lp(a) is considered to act as an atherogenic particle especially when other risk factors are present (5, 6). The functional properties of lipoproteins and their atherogenicity depend mainly on the type and composition of their protein components, i.e., the apolipoproteins (1, 3). The major apoproteins of HDLs are apo A-I and apo A-II, of LDLs is apo B, and of Lp(a) are apolipoprotein (a) [apo(a)] and apo B.

Diagnostic relevance of apolipoprotein measurements

The quantitative measurement of apo A-I and apo B is an alternative to conventional lipid and lipoprotein analysis. Apolipoprotein measurements permit an assessment of cardiovascular risk and response to lipid-lowering therapy, and also enable identification of patients with certain inheritable abnormalities of lipoprotein metabolism not detectable using other methods (7). Prospective studies have confirmed the usefulness of plasma apo B concentration in determining risk (8, 9, 10, 11). The ratio of apo B/apo A-I allows those subjects with and without coronary artery disease (CAD) to be distinguished (12, 13) and seems to be useful for assessment of the effectiveness of pharmacological treatment with lipid-lowering drugs (14, 15). Elevated levels of apo B are also found in normolipidemic patients with early CAD even when levels of total and LDL-cholesterol are normal. Apo B measurement may be more relevant to CAD risk assessment because the amount of apo B per LDL particle is relatively constant, whereas that of cholesterol is variable (16, 17). In hypertriglyceridemic patients with impaired glucose tolerance and type II diabetes and where total cholesterol is similar to the non-diabetic population, apo B can differentiate small, dense LDL enriched with cholesterol ester from less atherogenic particles (16).

Numerous case-control and prospective clinical studies have linked elevated levels of plasma Lp(a) to an increased risk of atherosclerotic-related diseases (5, 18, 19, 20, 21). In a meta-analysis of twelve prospective studies and after conversion to the same Lp(a) concentration units for 1617 cases of ischemic heart disease (IHD) and 10,035 controls, it was concluded that Lp(a) is an independent prospective risk factor for IHD, with the effect similar in men and women the populations being 95% Caucasian (22). Lp(a) increases the risk of IHD when LDL-cholesterol is also elevated (6, 18), but is not associated with progression of CAD in subjects without prominent primary or secondary hyperlipidemia (23).

Harmonization of methods for the measurement of apo A-I, apo B and Lp(a)

In routine clinical laboratories the immunochemical measurement of apo A-I, apo B and Lp(a) is a state-of-the-art procedure using commercially available immunoassays based mainly on fully automated immunonephelometric (INA) or immunoturbidimetric (ITA) methods (1, 3, 5, 24). The precision of the tests performed on automated analyzers is high. Nephelometer systems yield a mean interassay variability of ~4% for apo A-I and apo B measurements (1) and a similar precision was obtained for Lp(a) (25, 26). Until recently, the large-scale introduction of apolipoprotein measurement was hampered by the lack of reference materials and methods which resulted in the use of different calibration methods and a lack of comparability of results (27, 28). Other major sources of variability of apolipoprotein results are non-linearity of methods, matrix effects causing non-parallelism between calibrator and samples, and inadequate relative apolipoprotein concentrations in the available reference materials (25, 27). It was soon realized that the preparation, selection, and calibration of appropriate serum-based reference materials were the key to obtaining international standardization and harmonization of immunoassays for measurements of apo A-I and B (29, 30) and Lp(a) (25, 26, 31). Thus, use of common reference materials in a physical state equivalent to that of patient samples is necessary to reduce the variation in values among laboratories and between different measurement systems. Primary standards for apo A-I and B, and Lp(a), i.e., pure proteins, are not suitable for manufacturers’ assay systems because they show instability and/or matrix effects. Pure apo B is particularly unsuited as a primary standard due to self-association, sensitivity to oxidation, and irreversible matrix aggregation. Therefore the use of secondary matrixed (serum) reference materials (SSRMs) is the only possible approach to harmonizing calibration. A hierarchy of reference materials and reference methods constitute the reference measurement system (Fig.1) and form the basis for the establishment of standardized reference values and decision cutoff limits in different ethnic populations.

IFCC standardization programs

In 1989 the IFCC Committee on Apolipoproteins (Chairperson: Professor Santica Marcovina, Seattle, WA, USA), the Centers for Disease Control and Prevention (CDC), Atlanta, GA, USA, and representatives of 25 diagnostic companies producing test systems for apo A-I and B reached a consensus on the implementation of a collaborative study in 3 phases. Similarly, in 1995 the IFCC Working Group for Lp(a) Assay Standardization (Chairperson: Professor Armin Steinmetz, Andernach, Germany), working in collaboration with manufacturers of diagnostic assays for Lp(a), was formed with a view to selecting a suitable secondary reference material for Lp(a) that would result in closer comparability of Lp(a) values between assay systems.

Preparation of serum reference materials for apo A-I, apo B and Lp(a)

Special effort is needed to prepare suitable SSRMs for apo A-I, apo B, and Lp(a) that can be used to calibrate different test systems uniformly (26, 32, 33). Reference preparations must have the same immunochemical behavior as patient specimens in all systems. The SSRM for Apo A-I was prepared in lyophilized form to ensure long-term stability. In contrast, as the lyophilization process alters apo B structure resulting in changes in immunoreactivity and matrix effects between methods, a stabilized liquid preparation shock-frozen immediately after production and kept at -70 ºC to improve long-term stability, was used as reference material for apo B. The SSRM for Lp(a) was prepared in lyophilized form from a serum pool and has three predominantly expressed apo(a) polymorphs containing 16, 17 and 18 K4 domains, and three minor polymorphs of 14, 20 and 32 K4 domains.

Apolipoprotein A-I and B Standardization In Phase 1 the harmonization of test results showed that uniform calibration of 26 test systems for apo A-I and 28 for apo B had been accomplished using the same reference materials (32). Apo B levels before the uniform calibration showed a large among-assay variation (>19%) due to different assay standardization. This was significantly reduced to a mean of 6% after uniform calibration. Of the 26 reference materials for apo A-I and apo B proposed by the manufacturers, 15 for apo A-I and 11 for apo B, two lyophilized preparations for apo A-1 and two liquid-stabilized preparations for apo B were selected for further evaluation in Phase 2 of the study (33). Using these materials as the common calibrator, results again showed that by uniform calibration of the 37 test systems harmonization of results could be accomplished. Among-systems variation in apo A-l levels improved from 9% to 5.4% using the common calibrator and from ~20% to 7% for apo B. The linearity and parallelism study showed that the immunoreactivity exhibited by the candidate SSRMs, SP1 and SP3, was closest to that of the fresh and frozen serum pools and there were no significant matrix effects in the tested assays.

In Phase 3, SP1-01 (the batch prepared first) and SP3-07 (a new batch) were examined to determine their ability to transfer assigned values to each calibrator (34, 35). An apo A-I value of 1.50 g/L was assigned to SP1-01 and an apo B value of 1.22 g/L was assigned to SP3-07 using primary reference materials and consensus methods (Table 1). Following a common protocol, the values for apo A-I and apo B were transferred from the two reference materials to the respective calibrators of the test systems of the participating diagnostic companies. Then 50 fresh-frozen serum samples covering a wide range of apo A-I and apo B concentrations were analyzed to test for comparability between various immunoassays. The results obtained by the diagnostic companies were excellent with an interlaboratory variation for results using of 2-6% for apo A-I and 3-7% for apo B. Therefore these SSRMs were proposed as WHO international reference materials.

Table 1.

The reference measurement systems for apo A-I, apo B and Lp(a).

Analyte	1° Reference Method	1° Reference Material	2° Referenc e Material	Consensus Method for Value Assignment
Apo A-I	HPLC-MS (candidatea) [CDCb]	BCR – CRM 393, Purified apo A-I	WHO RR SP1-01 (1.50 ± 0.08 g/L)	RIA (CDC)
Apo B	Not available	d=1.03-1.05 kg/L, UC purified LDL	WHO RR SP3-07 (1.22 ± 0.02 g/L)	INA (NWLRLc)
Lp(a)	Not available	Two Lp(a) purified preparationsd	PRM 2B	a-40 ELISAe (NWLRL)
			(107.1 ± 8.6 nmol/L)	a1-1 ELISA (NWLRL)

a Barr JR, Maggio VL, Patterson DG Jr., Cooper GR, Henderson LO, Turner WE, et al. Isotope dilution-mass spectrometric quantification of specific proteins: model application with apolipoprotein A-I. Clin Chem 1996;42:1676-82.

b Centers for Disease Control and Prevention

c Northwest Lipid Research Laboratories

d Lp(a) isolated by sequential ultracentrifugation and molecular sieve chromatography or by lysine-Sepharose and CsCl density gradient ultracentrifugation

e Marcovina SM, Albers JJ, Gabel B, Koschinsky ML, Gaur VP. Effect of the number of apolipoprotein (a) crinkle 4 domains on immunochemical measurements of lipoprotein(a). Cline Chem 1995;41:246-55.

The materials SP1-01 and SP3-07 were submitted to the WHO Expert Panel on Standardization as proposed reference materials for apo A-I and B, respectively. Based on the results of the reports submitted, the materials were accepted as WHO-IFCC International Reference Reagents (Table 1). The materials are stored at CDC and are available to manufacturers and research institutes for the evaluation and standardization of new assays.

Lipoprotein(a) Standardization

In Phase 1 forty test systems were evaluated for analytical performance by testing serum samples for precision, linearity and parallelism characteristics. Eight manufactured Lp(a) calibrator materials were tested for commutability and method harmonization after reference to an arbitrary calibration (25). In Phase 2, testing was repeated in 27 Lp(a) test systems using four newly prepared materials (26). The Lp(a) material with maximum harmonization achieved a variation of <8% for 18 INA and ITA systems. On the basis of acceptable analytical performance, maximal harmonization effect and documented stability, the lyophilized material PRM 2B was selected as the common calibrator for Lp(a) and assigned a value of 107 nmol/L using a consensus, reference measurement system (Table 1).

In Phase 3, 16 manufacturers and six research laboratories evaluated PRM for its ability to transfer an accuracy-based value to the immunoassay calibrators and for the extent of concordance in results between methods (31). After uniformity of calibration was demonstrated in the 22 evaluated systems, Lp(a) was measured on 30 fresh frozen sera covering a wide range of Lp(a) values and apo(a) sizes. The among-laboratory CV of Lp(a) measurement of the 30 samples ranged from 6% to 31%. In general, these CVs were higher than those obtained for the PRM (2.8%) and higher than those observed in three quality control samples (14%, 12%, and 9% respectively), reflecting the broad range of apo(a) isoforms in the samples and the sensitivity of most of the analytical systems to apo(a) size heterogeneity.

Accuracy of Lp(a) measurement for each system was determined by comparison with target values assigned using the consensus, reference measurement system (31). A Lp(a) latex-enhanced immunoturbidimetric method was the most accurate of the tested systems with an average absolute bias of only 4.4 nmol/L compared with 12.4 to 23.8 nmol/L for other methods. Only two of the tested Lp(a) systems showed minimal bias to apo(a) isoform heterogeneity. From the Phase 3 results it appears that many of the current Lp(a) methods are isoform-biased and hence inaccurate, and that their use has possibly led to incorrect conclusions in clinical studies.

The proposed secondary reference material for Lp(a), PRM, has the characteristics of a suitable reference material with a negligible matrix effect in the tested Lp(a) assays. Through the use of PRM superimposable Lp(a) results were produced for a manufacturer’s assay and the reference method indicating that PRM is able to standardize Lp(a) measurement. The incomplete harmonization of values is not a problem of the calibrator but of the apo(a) size-sensitive nature of current assays. No reference material, either a primary or secondary standard will be able to eliminate substantial differences in Lp(a) values measured by analytical systems that are affected by apo(a) size heterogeneity. Based on the results of this study together with the stability data for PRM, the IFCC WG Lp(a) will seek international recognition of PRM as the secondary reference material for Lp(a).

Conclusions The initiative started by the IFCC Committee on Apolipoproteins has contributed to a practical solution to the harmonization of test results of immunoassays for apo A-I and apo B. Therefore through the efforts of the IFCC a worldwide consensus on the calibration of test systems for apo A-I and B has been achieved. The availability of internationally accepted reference materials will eliminate a major factor of variability among test systems. Through the use of IFCC standardized methods traceable to SP1-01 and SP3-07 the value distributions of apo A-I and apo B in various populations have been described (Table 2; 36, 37, 38, 39, 40).

Table 2.

Reference ranges of apo A-I and B for males and females in various populations.

Population	Apo A-I (10th percentile) in g/L		Apo B (75th percentile) in g/L
	Men	Women	Men	Women
US (Framingh am)	1.07	1.22	1.18	1.11
Finland	1.09	1.25	1.41	1.29
Sweden	1.1	1.22	1.53	1.42
Northern Italy	1.16	1.27	1.29	1.31