Ingmar Jungner, 1* Santica M. Marcovina, 2 Göran Walldius, 3 Ingar Holme, 4 Werner Kolar, 1 and Eugen Steiner 1. Lipoproteins

Transcription

1 Clinical Chemistry 44: (1998) Lipids and Lipoproteins Apolipoprotein B and A-I values in Swedish males and females, standardized according to the World Health Organization International Federation of Clinical Chemistry First International Reference Materials Ingmar Jungner, 1* Santica M. Marcovina, 2 Göran Walldius, 3 Ingar Holme, 4 Werner Kolar, 1 and Eugen Steiner 1 Serum concentrations of apolipoprotein (apo) B and apo A-I were measured from in a Swedish population sample of males and females, ages <20 to >80 years, using an automated immunoturbidimetric method calibrated against fresh pools of human serum and commercial calibrators. All values were recalculated in 1997 after calibration against the WHO- IFCC First International Reference Materials. The recalculation factor was for apo B, whereas for apo A-I, the correction was: y 0.989x The total CVs for both apo B and A-I were generally <7%. The mean value ( SD) for apo B was g/l in all males and g/l in all females. The mean apo A-I concentration was g/l in males 10% lower than in females ( g/l). The mean value of apo B increased up to 60 years of age in males and up to 70 years of age in females. apo A-I concentrations changed only slightly with age in both males and females. apo A-I concentrations among Swedes are nearly identical to those reported recently by two American studies and those obtained in a Finnish population sample. Mean apo B concentrations differ somewhat between the populations but mirror as expected differences in total 1 CALAB Research and CALAB Medical Laboratories, S:t Göran Hospital, S Stockholm, Sweden. 2 Department of Medicine, Northwest Lipid Research Laboratories, University of Washington, 2121 N 35th St., Seattle, WA King Gustaf V Research Institute, Karolinska Institute, S Stockholm, Sweden, and ASTRA HÄSSLE AB, S Mölndal, Sweden. 4 Institute for Medical Statistics, Ullevål Sykehus, P.O. Box 6, 0407 Oslo 4, Norway. *Author for correspondence. Fax ; ingmar.jungner@ stockholm.mail.telia.com. Received February 2, 1998; revision accepted May 14, cholesterol concentrations. The highest values were noted in Swedish subjects. The Swedish sample population is, to our knowledge, the largest describing the distribution of apo B and A-I in a general population of adult males and females of all ages determined with procedures standardized and traceable to the WHO- IFCC First International Reference Materials. In clinical practice, lipid, lipoprotein, or apolipoprotein (apo) 5 concentrations are measured to detect and diagnose lipid disorders (dyslipidemias), to assess the associated cardiovascular risk, and to decide subsequent treatments (1, 2). Sniderman and co-workers (3, 4) recently proposed that measurements of apo B should be used in routine clinical practice and should replace calculation of LDL-cholesterol by the Friedewald formula (5). These authors (3, 4), as well as ourselves and others (6 8), have argued that now that the measurements of apo B and apo A-I are standardized (9 12), they should be more widely used in clinical settings. Furthermore, apo B appears to give a more accurate predictive estimate of atherogenic risk than total cholesterol (TC) or LDL-cholesterol (13 15). apo B and apo A-I can also be determined with high precision and accuracy, and the measurements are easily automated (16, 17), reducing labor costs. In the past, a wider clinical use of apolipoprotein assays was limited by lack of standardization and, consequently, by lack of common population-based reference values. However, 5 Nonstandard abbreviations: apo, apolipoprotein; TC, total cholesterol; NHANES III, National Health and Nutrition Examination Survey III; AMORIS, Apolipoprotein-related Mortality Risk Study; TG, triglyceride; and CL, confidence limit. 1641

2 1642 Jungner et al.: Apolipoprotein B and A-I in a Swedish population recently two American studies have been published. Contois and co-workers (18, 19) presented reference intervals for apo B and apo A-I, primarily in adults, derived from the population-based Framingham Offspring Study. Bachorik et al. (20) measured the concentration of the two apos in Phase 1 ( ) of the Third National Health and Nutrition Examination Survey (NHANES III). The latter presented the distributions of apo A-I and apo B in persons 4 years of age or older, which were the first measured in a probability sample of the noninstitutionalized US civilian population. Both these American studies, as well as a Finnish study in 1995 by Leino et al. (21), used the new WHO-IFCC International Reference Materials as the basis for accuracy. All three investigations were, however, performed on a smaller number of subjects: 3824, , and 575, respectively. Since 1985, one of our group (I.J.) implemented automated methods for the determination of apo B and apo A-I, applying them in screening programs of large populations to improve information about the composition and degree of dyslipoproteinemias. Measurements of apo B and A-I in Swedes (6) and the basic data of the prospective study AMORIS (Apolipoprotein-related Mortality Risk) on Swedes were reported in 1992 (7). The aim of the present study after increasing more than threefold the investigated population and after recalculating earlier apolipoprotein measurements to the new WHO-IFCC standard was to provide a large basis for apo B and apo A-I reference intervals. Subjects and Methods Subjects Swedish males and females ( ) ranging in age from 20 years to 80 years were studied. The Swedish healthcare system has focused on general health screening performed during the last 20 years as a service to almost all employees of most companies, as well as to adult males and females in the general community (virtually all native Swedes). Calab has served as a major laboratory for the 3000 physicians involved in general healthcare. The present clinical material comprises all blood samples sequentially sent to Calab without any limitations because of age, possible diagnosis, or ongoing treatment with diet and/or drugs. Thus, no sample/subjects were excluded from this population study. Because our database does not contain any information about the use of drugs such as beta blockers and estrogens, the possible effects of such drugs on lipids and apo composition could not be analyzed in this study. The large majority of the subjects were subjectively healthy. The major part of the studied population comprised males and females years of age, as presented in our previous study (7). No samples were obtained from hospitalized patients or from subjects participating in clinical trials. Most subjects lived in the greater Stockholm area ( 20% were from other parts of Sweden). The original AMORIS population investigated during comprised subjects (7). The study has recently been extended to include data collected in The data base now comprises subjects ( males, mean SD age, years; and females, mean age, years). Here we describe findings from a subsample of the whole study population. These subjects had their first complete profile of apo B apo A-I TC triglycerides (TGs) taken simultaneously (n ; mean age, years for males and years for females). The subjects years of age are classified in decades. The classes 20 or 80 years of age have mean ages of and years, respectively; the number of males equaled the number of females. About two-thirds of the subjects (n ) were fasted overnight (n for males and n for females). The others had a light morning meal (n ), or the nutrition status was unknown (n ). methods The methods are the same as described earlier (6). apo B and A-I were determined by an immunoturbidimetric method according to Riepponen et al. (22), using polyclonal antisera from Orion Diagnostica (commercial assay). The company is one of the manufacturers that participated in the WHO-IFCC Standardization Program (9 12). TC was determined with the cholesterol oxidase/ peroxidase (CHOD-PAP) assay and TGs with the glycerol phosphate oxidase/peroxidase (GPO-PAP) assay, using enzymatic methods (reagents from Boehringer Mannheim) for the PRISMA instrument and from Bayer Diagnostics GmbH for the DAX instrument. All four methods were from the outset highly automated; from 1985 to 1992, the Multichannel AutoChemist -PRISMA (New Clinicon) was used and since 1993, the Multichannel DAX-96 (Technicon/Bayer Corp.) has been used. All analyzers were computerized with systems for automatic calibration. An extensive quality-control scheme was used throughout the study in determining serum apos. Table 1 shows the long-term precision data from materials used as controls. The total CV (CV total) was the same as reported earlier (6) for the initial ( ) part of the study: generally 7% for apo B, apo A-I, and TGs, and 3% for TC. For apo B and A-I, zero point and one-level calibration were done using fresh pools of human serum because of the lack of suitable calibrators for daily use. A pool of human serum was prepared each day from 300 clear samples from healthy controls. By using this large number of sera, the same mean values were expected (23). The technique was implemented from the start of the measurements in The pool material was replaced as calibrator as of 1993 by the commercially available Seronorm Lipid (Nycomed Pharma) supplemented by the Apolipoprotein Calibration Set SPQ (Incstar Corp.) to maintain stable quality performance.

3 Clinical Chemistry 44, No. 8, Year a No. of months Controls, n b Table 1. Long-term precision data for apolipoproteins. apo B Concentration, g/l Total CV, % Target value Controls, n b apo A-I Concentration, g/l Total CV, % Target value Instrument Fresh pools of human serum 1987 c Prisma 1988 c Prisma 1989 c Prisma Prisma Prisma Prisma DAX DAX DAX DAX DAX DAX-96 Deep-frozen pools of human serum Prisma Prisma Prisma Prisma DAX DAX-96 Seronorm Lipid Prisma Prisma DAX DAX DAX DAX d d d d d DAX-96 Beckman Triad LINK level I DAX DAX d d d d DAX DAX d d d d DAX-96 Beckman Triad LINK level DAX DAX d d d DAX DAX-96 Beckman Triad LINK level DAX d d d d d DAX DAX d d d d d DAX-96 a Calibrators: , fresh pools of human sera; , Seronorm Lipid; , Apo B and Apo A-1 Calibration/Control Set (SPQ Test System). b Number of analyses. c Data cited from Jungner et al. (6). d New batch.

4 1644 Jungner et al.: Apolipoprotein B and A-I in a Swedish population Table 2. Profile data of groups in the study. A. Selected percentiles for apo B (g/l) distribution. Percentile Age, years Sex n a Mean SD 5th 10th 25th 50th 75th 90th 95th All M F M F M F M F M F M F M F M F M F B. Selected percentiles for apo A-I (g/l) distribution. Percentile Age, years Sex Mean SD 5th 10th 25th 50th 75th 90th 95th All M F M F M F M F M F M F M F M F M F C. Means and SDs b for apo B/A-I ratio, TC (mmol/l), and TGs (mmol/l). Age, years Sex apo B/A-I, mean SD TC, mean SD TGs, mean SD All M F M F M F M F M F M F M F M F M F a Subpopulation (n ) of NEW AMORIS, , divided by sex and age. The number of subjects in each decade is the same throughout the table. b Selected percentiles (log-normal distribution) for different sex and age groups can be requested from corresponding author.

5 Clinical Chemistry 44, No. 8, Daily samples from fresh or frozen pools of human sera were used as controls; since 1993 commercially available materials like Seronorm Lipid and Beckman control sera were utilized (Triad LINK level 1 3). The inaccuracy for TC and TGs was checked against material from the National Institute of Standards and Technology, Gaithersburg, MD, or by analyses performed by lipid reference laboratories certified by the CDC. Traceability to the WHO-IFCC International Reference Materials. Recalculation of collected apolipoprotein data was done in 1997 in collaboration with the Northwest Lipid Research Laboratories, University of Washington, Seattle, WA. This laboratory was the one that organized and coordinated the IFCC standardization program (9 12). They provided Calab with three concentrations of their apo B and apo A-I fresh-frozen quality-control samples prepared in-house, with values assigned against the WHO-IFCC Reference Materials. During 5 days, 36 determinations of apo A-I and apo B were performed on each quality-control sample, and data were sent to Northwest Lipid Research Laboratories for statistical evaluation. On the basis of the results of this analysis, a correction factor of for apo B and the correction y 0.989x for apo A-I was used to ensure that values were traceable to the WHO-IFCC International Reference Materials. Accredited laboratory facilities. All analyses were done at Calab Medical Laboratories, Stockholm, Sweden. For almost every determination of apos and lipids, 20 other laboratory analyses were performed as a basis for clinical evaluation. In 1994, the laboratory received an international accreditation according to European Norm by the Swedish Board for the Technical Accreditation (Borås, Sweden). Calab also received a Statement of Good Laboratory Practice compliance (Läkemedelsverket, Uppsala, Sweden) in The laboratory performed analyses in 1996 and is accredited in clinical chemistry, hematology, immunology, and microbiology. statistical analysis Data are presented as means (and SDs), 95% ranges, medians, or selected, theoretically calculated log-normal percentile limits. For between-group mean differences, 95% confidence limits (95% CL) are used. Results mean values and distributions of apolipoproteins The results of the first complete profile data (apo B, apo A-I, apo B/A-I ratio, TC, and TGs) from subjects are shown in Table 2. Selected percentile values of apo B and apo A-I as well as means and SDs of the other variables for males and females in each decade and in the total population are presented. These values were obtained by pooling results from fasting and nonfasting subjects, as in NHANES (20). The mean apo B concentration (Table 2A) was g/l in males (n ) significantly higher than the mean for the females [ g/l; 95% CL for difference, (0.086, 0.094)]. The mean apo A-I concentration (Table 2B) in males was g/l 10% lower than in females [ g/l; 95% CL for difference, 0.15 (0.148, 0.152)]. apo B and A-I concentrations showed an almost gaussian distribution. For the various age and sex strata, the differences between the mean and the median values ranged from 0.03 to 0.05 g/l (apo B) and from g/l (apo A-I), suggesting that the distributions of apo B and A-I were virtually gaussian. The mean apo B/A-I ratio (Table 2C) was in males and in females [95% CL for difference, 0.16 (0.157, 0.163)], whereas the 95% interval was in males and in females, including all ages. The distributions of the apo B/A-I ratio were also virtually gaussian, the differences between the mean and the median values ranging from 0.02 to 0.05 (data not shown). apo b and apo a-i in fasting and nonfasting subjects A comparison between samples collected from subjects reported to be in a fasting state or in nonfasting state showed for apo A-I no difference, whereas the values for apo B were slightly lower. The maximal deviation was 3.3% for apo A-I and 7.7% for apo B (data not shown). apo b and apo a-i by age Median apo B concentrations in adult males were g/l, and the increase continued from those years of age and tended to decrease after age 60 (Fig. 1A). The median apo B concentrations in adult females were g/l, and the increase continued in those in the age group years, with a more pronounced increase in the age group years, and then decreased after age 70 (Fig. 1A). In both males and females, apo B ultimately reached about the same concentration (1.36 g/l for males and 1.38 g/l for females); however, in males this occurred 10 years earlier than in females. apo B values in younger age cohorts were higher among males than among females, whereas the cohorts at 60 years of age showed a reverse relation. The 95% interval for apo B was g/l for males and g/l for females, including all ages. Median apo A-I concentrations in adult males ranged from 1.30 to 1.36 g/l, with only a slight continuous increase from 20 to 59 years, followed by a plateau at years, and then a decrease at 70 years (Fig. 1B). Median apo A-I concentrations in adult females ranged from 1.45 to 1.52 g/l, and there was a first plateau in the age range years, then an increase up to 59 years, followed by a second plateau in the age range years, and a decrease at 70 years (Fig. 1B). The 95% interval for apo A-I was g/l for males and g/l for females, including all ages. The highest

6 1646 Jungner et al.: Apolipoprotein B and A-I in a Swedish population apo B/A-I ratio for males was seen in males years old, whereas the highest ratio for females was obtained in those years old (Fig. 1C). tc and tg values The mean TC (5.9 mmol/l) was the same for males and females (Table 2C). There was no difference between fasting and nonfasting TC values (data not shown). The median TG concentration, including results from all individuals in fasting or nonfasting states, was 1.35 mmol/l for males (95th percentile, 3.64 mmol/l) and 1.04 mmol/l (95th percentile, 2.56 mmol/l) for females (data not shown); Table 2C shows mean values. The fasting median TG (data not shown) for males was 1.29 mmol/l (95th percentile, 3.44 mmol/l), and 1.00 mmol/l (95th percentile, 2.42 mmol/l) for females; fasting mean TG values are given for comparison (Table 3) and were 1.57 and 1.18 mmol/l for males and females, respectively. The difference for TGs between males and females was significant, at P comparison with other population studies traceable to the who-ifcc This Swedish population sample was compared (Table 3) with characteristics for three other population samples traceable to the WHO-IFCC First International Reference Materials: two American (the Framingham Offspring Study (18, 19) and the NHANES III (20)) and one Finnish (21). These populations are from different ethnic origins and differ in sample size and TC concentrations. Their mean age was about the same. If differences in TC concentrations among the population samples are taken into account, the concentrations and percentile distributions of apo B are very similar and of apo A-I are almost identical. Fig. 1. Differences in apo B and apo A-I concentrations and ratios according to age group. Median concentrations (g/l) for apo B (A) and apo A-I (B) together with apo B/A-I ratio (C) at each 10-year interval of males and females. The classes 20 or 80 years have mean ages of and years, respectively. Discussion The present study is the first large epidemiological study applying automated apo methods over several years to the health screening of males and females covering wide age ranges. However, compared with previously published population studies (18 20), our results have always been obtained on fresh samples, thus avoiding possible effects of long-term storage and of freezing and thawing of the samples. In 1992 (6), we presented the first large population report on apo determinations in males and females. Now we have corrected our original results, so that they are traceable to the WHO-IFCC Reference Materials. Thus, our present results are directly comparable with those obtained in recently presented population studies. Determinations of apos may provide a more specific characterization of the type of lipoprotein abnormality, thus leading to a more accurate prediction of the risk of coronary artery disease than that made using the TC, TG, and HDL values alone, as discussed elsewhere (3, 4). To establish whether apo B and A-I, as well as the ratio apo B/A-I, are better predictors of atherogenic risk than TC and TGs, the AMORIS subjects (7) will be

7 Clinical Chemistry 44, No. 8, Table 3. Characteristics of the Swedish population sample compared with other recently reported population samples traceable to the WHO-IFCC First International Reference Materials. Percentile 10th 25th 50th 75th 90th Mean SD apo B, g/l Males SWE a FIN FOS/USA USA Females SWE FIN FOS/USA USA apo A-I, g/l Males SWE FIN FOS/USA USA Females SWE FIN FOS/USA USA n Age, years b TC, mmol/l b ftgs, mmol/l b Males SWE c FIN (range) d FOS/USA USA e 5.08 e 1.42 e Females SWE c FIN (range) d FOS/USA USA e 5.08 e 1.42 e a SWE, Swedish population sample; FIN, Finnish sample (21); FOS/USA, Framingham Offspring Study (18,19); USA, NHANES III (20); and f, fasting. b Mean SD. c For comparison with Finnish sample. Males, after logarithmic transformation; females, after logarithmic transformation. d After logarithmic transformation. e Not specified or sex-divided. monitored over an extended period of time. Another major question that remains to be answered is how much apo B and apo A-I add to the determination of TC and TGs in prediction of total fatal risk. A second report of the AMORIS study, with extended mortality data in relation to initial TC and TG and especially to initial apo B and apo A-I concentrations, is under preparation. Our present study shows that the measurements of apo B and apo A-I were performed with good precision and accuracy throughout the whole observation period. As controls, pools of both fresh and frozen samples of human serum (Table 1) were used. These gave results similar to those found by other authors (19, 21, 24). Of our originally presented data (6, 7), only minor corrections for both apo B and A-I were needed to harmonize the data to be traceable to the WHO-IFCC International Reference Materials. The errors of the methods are of the same order of magnitude or better than those suggested by Marcovina and Albers (25). They recommended that the between-run values should be 8%, and optimally 5%. Bachorik et al. (20) report that the CVs for apo A-I and B averaged 6% throughout the NHANES III study. Recent CV estimates of the overall biological variation of apo B and apo A-1 were 6 7% (24). In our study, there was virtually no difference (apo A-I) or only a slight (apo B) difference, when we compared values obtained in fasting state with

8 1648 Jungner et al.: Apolipoprotein B and A-I in a Swedish population values obtained in individuals having taken a light meal. Although the data were obtained from different individuals, the total number of subjects was large enough in each group of individuals to support the notion that a light meal does not change apo B or apo A-I concentrations (26 28). Bachorik et al. (20) also found in the NHANES study that there were no significant differences in apo B or apo A-I between fasting or nonfasting subjects; they therefore combined the data from both groups. Thus, apos may be regarded as robust measurements, at least regarding sampling conditions. This has considerable practical impact, both for the patient often traveling far to get his or her blood taken and for the physician in terms of the interpretation of the data. Whereas apo B values for both sexes clearly increased with age, apo A-I values showed only minor age-related variation. Males have higher apo B values than females up to age 60. After menopause, apo B concentrations in females continued to rise. Females had 10% higher apo A-I concentrations than males in all age groups. Similar age-related values for apo B and apo A-I were observed in 1992 (6). The present results for apo A-I are in close agreement with the 10% difference also found in the NHANES III Survey (20) and the 13% found in the Framingham Offspring Study (19) and the Finnish study (21) (Table 3). The results reported in these studies were all obtained by turbidimetric analysis with the exception of the NHANES III where rate immunonephelometry was used most of the time (they changed from a radial immunoassay when automated methods were becoming more widely used). The Framingham study used turbidimetric assay reagents and calibrators from Incstar Corp. The Swedish and Finnish studies both used the same turbidimetric method (22) and the same supplier of reagents (Orion Diagnostics). The percentile distributions for the four studies were rather similar (apo B) and nearly identical (apo A-I). apo B is positively correlated with TC (6, 18, 21); therefore, in accordance with the cholesterol data, mean apo B was highest for the Swedes and lowest for the Americans. Mean apo A-I concentrations were virtually the same for all four population samples, with ranges of g/l for males and g/l for females (Table 3). The often conflicting findings reported earlier for apo analysis have been mainly ascribed to differences in methodological approaches, sample handling and storage, and especially to the previous lack of common reference materials to which the calibration of the methods could be referred. In view of the results from the latest investigations, which all rely on the WHO-IFCC International Reference Materials, it can be concluded that the WHO-IFCC standardization has attained its purpose and that it is now possible to establish that there are only minor differences between different populations. Bachorik et al. (20) compared the results from their survey with the Framingham Offspring Study and reported only small differences between different ethnic groups. Several authors, especially after the introduction of the WHO-IFCC Reference Materials, have discussed and suggested cutpoints for apo B (4, 8, 18, 20) and apo A-I (8, 19, 20) that may be used to assess coronary heart disease risk. Bachorik at al. (20) determined apo B concentrations in adults, ages 20 years, categorized by the National Cholesterol Education Program risk levels for LDL-cholesterol. They conclude that the question of whether apo B might eventually be used instead of LDL-cholesterol as the basis for assessing risk for coronary heart disease remains open. Because apo B varies with age and apo A-I varies with sex, age- and sexstandardized values should be used as markers for cardiovascular risk. However, the introduction of new markers for atherogenic risk that may be used instead of the established and accepted LDL and HDL determinations makes this new approach difficult from a practical standpoint. Too many cutoff values create practical problems related to lack of simplicity. Instead of using sex- and age-standardized values, the alternative approach would be to choose a few cutoff values and evaluate whether such values are clinically significant. We are in the process of analyzing AMORIS data relating baseline apo B, apo A-I, TC, and TG values to final risk/fatal myocardial infarction, using the approach of dichotomizing risk in relation to age and sex, as well as by age-and sexstandardized algorithms. The recommended cutoff value should be proven to discriminate future coronary heart disease risk in a stringent, predictive, and possibly simple way. The final decision of how to best define relevant risk and cutoff values must await outcome studies like AMORIS. This work was supported by grants from Gunnar and Ingmar Jungner Foundation for Laboratory Medicine and from Bure Hälsa & Sjukvård Ltd, Stockholm, to Calab Research. References 1. European Atherosclerosis Society: international task force for prevention of coronary heart disease. Prevention of coronary heart disease: scientific background and new clinical guidelines. Nutr Metab Cardiovasc Dis 1992;2: The Expert Panel. Summary of the second report of the National Cholesterol Education Panel (NCEP) expert panel on detection, evaluation and treatment of high blood cholesterol in adults (Adult Treatment Panel II). JAMA 1993;269: Sniderman AD. To (measure apo)b or not to (measure apo)b: a critique of modern medical decision-making. Clin Chem 1997;43: Sniderman AD, Cianflone K. Measurement of apolipoproteins: time to improve the diagnosis and treatment of atherogenic dyslipoproteinemias [Editorial]. Clin Chem 1996;42: 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: Jungner I, Walldius G. Holme I, Kolar W, Steiner E. Apolipoprotein B, A-I in relation to serum cholesterol and triglycerides in Swedish males and females. Int J Clin Lab Res 1992;21:

9 Clinical Chemistry 44, No. 8, Walldius G, Jungner I, Kolar W, Holme I, Steiner E. High cholesterol and triglyceride values in Swedish males and females: increased risk of fatal myocardial infarction. First report from the AMORIS (Apolipoprotein-related Mortality Risk) Study. Blood Press 1992; 1(Suppl 4): Bhatnagar D, Durrington PN. Measurement and clinical significance of apolipoprotein A-I and B. In: Rifai N, Warnick GR, Dominiczak MH, eds. Handbook of lipoprotein testing. Washington, DC: AACC Press, 1997: Marcovina SM, Albers JJ, Dati F, Ledue TB, Ritchie RF. International Federation of Clinical Chemistry standardization project for measurements of apolipoproteins A-I and B. Clin Chem 1991;37: Albers JJ, Marcovina SM, Kennedy H. International Federation of Clinical Chemistry standardization project for measurements of apolipoproteins A-I and B. II. Evaluation and selection of candidate reference materials. Clin Chem 1992;38: Marcovina SM, Albers JJ, Henderson LO, Hannon WH. International Federation of Clinical Chemistry standardization project for measurements of apolipoproteins A-I and B. III. Comparability of apolipoprotein A-I values by use of international reference material. Clin Chem 1993;39: Marcovina SM, Albers JJ, Kennedy H, Mei JV, Henderson LO, Hannon WH. International Federation of Clinical Chemistry standardization project for measurements of apolipoproteins A-I and B. IV. Comparability of apolipoprotein B values by use of international reference material. Clin Chem 1994;40: Wald NJ, Law M, Watt HC, Wu T, Bailey A, Johnson AM, et al. Apolipoproteins and ischaemic heart disease: implications for screening. Lancet 1994;343: Lamarche B, Moorjani S, Lupien, PJ, Cantin B, Bernard P-M, Dagenais GR, Despres JP. Apolipoprotein A-I, B levels and the risk of ischemic heart disease during a five-year follow-up of men in the Quebec Cardiovascular Study. Circulation, 1996;94: Rosseneu M, Fruchart JC, Bard JM, Nicaud V, Vinaimmont N, Cambien F, et al. Plasma apolipoprotein concentrations in young adults with parental history of premature coronary heart disease and in control subjects. The EARS study. Circulation 1994;89: Labeur C, Shephard J, Rosseneu M, Immunological assays of apolipoproteins in plasma: methods and instrumentation. Clin Chem 1990;36: Stein E A. Clinical significance and measurement of apolipoproteins A-1 and B. In: Rifai N, Warnick GR, eds. Laboratory measurement of lipids, lipoproteins and apolipoproteins. Washington, DC: AACC Press, 1994: Contois JH, McNamara JR, Lammi-Keefe CJ, Wilson PWF, Massov T, Schaefer EJ. Reference intervals for plasma apolipoprotein B determined with a standardized commercial immunoturbidimetric assay: results from the Framingham Offspring Study. Clin Chem 1996;42: Contois JH, McNamara JR, Lammi-Keefe CJ, Wilson PWF, Massov T, Schaefer EJ. Reference intervals for plasma apolipoprotein A-I determined with a standardized commercial immunoturbidimetric assay: results from the Framingham Offspring Study. Clin Chem 1996;42: Bachorik PS, Lovejoy KL, Carroll MD, Johnson CL. Apolipoprotein B, AI distributions in the United States, : results of the National Health and Nutrition Examination Survey III (NHANES III). Clin Chem 1997;43: Leino A, Impivaara O, Kaitsaari M, Järvisalo J. Serum concentrations of apolipoprotein A-I, apolipoprotein B, and lipoprotein (a) in a Finnish sample. Clin Chem 1995;41: Riepponen P, Marniemi J, Rautaoja T. Immunoturbidimetric determination of apolipoprotein A-I and B in serum. Scand J Clin Lab Investig 1987;47: Cembrowski GS, Chandler EP, Westgard JO, Assessment of average of normals quality control procedures and guidelines for implementation. Am J Clin Pathol 1984;81: Marcovina SM, Gaur VP, Albers JJ. Biological variability of cholesterol, triglyceride, low-and high-density lipoprotein cholesterol, lipoprotein (a) and apolipoproteins A-I and apo B. Clin Chem 1994;40: Marcovina SM, Albers JJ. Apolipoprotein assays: standardization and quality control. Scand J Clin Lab Investig 1990;50(Suppl 198): Rifai N, Dufour DR, Cooper GR. Preanalytical variation in lipid, lipoprotein, and apolipoprotein testing. In: Rifai N, Warnick GR, Dominiczak MH, eds. Handbook of lipoprotein testing. Washington, DC: AACC Press, 1997: Seishima M, Noma A, Torizawa H, Muto Y. Changes of serum apolipoprotein levels after oral administration of fat in human subjects. Atherosclerosis 1988;73: Rifai N, Merill JR, Holly RG. Postprandial effect of a high fat meal on plasma lipid, lipoprotein cholesterol and apolipoprotein measurements. Ann Clin Biochem 1990;27: