Evaluation of Lipoprotein(a) Electrophoretic and Immunoassay Methods in Discriminating Risk of Calcific Aortic Valve Disease and Incident Coronary Heart Disease: The Multi-Ethnic Study of Atherosclerosis.
Despite the renewed clinical interest in Lp(a), its measurement is not well-standardized. Indeed, there remain analytical difficulties in accurately quantifying Lp(a) because of its unique structural properties, as reviewed by Marcovina and Albers (12). Clinical laboratories across the US offer a number of methodologies that quantify different aspects of the Lp(a) particle population, including electrophoresis- or ultracentrifugation-determined cholesterol content [Lp(a)-C] (13,14), electrophoresis-determined particle number [Lp(a)-P] (15), and immunoassay-determined Lp(a) mass concentration [Lp(a)-M] or particle concentration. To date, limited research has compared these metrics in the context ofcardiovascular outcomes, with inconsistent results reported (16-20). No study has tested whether there may be methodology-based differences in discriminating risk of CAVD. Further complicating Lp(a)-related risk assessment, race/ethnicity is a recognized determinant of Lp(a) concentrations, and black individuals are known to have 2- to 3-fold greater Lp(a) concentrations compared with whites and Hispanics (21-23). Yet, whether Lp(a) assays differentially discriminate CHD or CAVD risk across different races has not been tested.
Although Lp(a)-M is the most widely used assay, a few clinical reference laboratories offer Lp(a)-C or Lp(a)-P assessment. Thus, a study comparing the performance of these assays would be of use to clinicians faced with a choice of assays all purported to have advantages in risk stratification of CHD. Given these gaps in our knowledge, the current study aimed to compare Lp(a)-M, Lp(a)-C, and Lp(a)-P to determine whether they are differentially associated with risk of CAVD or incidence of CHD over a median follow-up period of 12 years.
Materials and Methods
The design of the Multi-Ethnic Study of Atherosclerosis (MESA) has been previously described (24), and information about the MESA protocol is available online (25). Briefly, 6814 men and women between the ages of 45 and 84 years, without clinical evidence of cardiovascular disease, were recruited from 6 communities in the US between 2000 and 2002. Institutional Review Board approval was obtained at all MESA sites, and all participants gave informed consent.
The current study excluded (a) participants who were taking lipid-lowering medication at baseline and those with otherwise limited specimen available (n = 1090), and (b) a random subcohort of 1000 MESA participants whose specimens were relatively depleted because of selection for more extensive phenotyping. The remaining subcohort was composed of 4679 individuals of the following race/ethnicities: 36.5% white (n = 1709), 28.8% blacks (n = 1347), 22.7% Hispanic (n = 1064), and 11.9% Chinese-Americans (n = 559). Age, race/ethnicity, sex, baseline measurements including the use of hypertension medication, systolic blood pressure, diabetes (treated or untreated diabetes mellitus as determined by 2003 American Diabetes Association fasting criteria), smoking status, and education status were recorded.
Laboratory measurements were conducted on specimens collected at MESA examination 1. Lp(a)-M was measured with a latex-enhanced turbidimetric immunoassay (Denka Seiken, Tokyo, Japan). This assay used 5 independent calibrators to control for apolipoprotein(a) [apo(a)] size heterogeneity, and the total imprecision was found to be <5%. At a concentration of 4 mg/dL, the imprecision was 6%. Lp(a)-C concentrations were obtained by electrophoresis of serum to separate lipoprotein classes followed by enzymatic staining of cholesterol in the separated Lp(a) particle fraction (Health Diagnostics Laboratory Inc., Richmond, VA) (13). Densitometric scans were performed to identify each lipoprotein subclass band and to quantify the Lp(a)-related cholesterol content based on total cholesterol concentration. The total imprecision was <8%, and the assay had a lower limit of quantification (LOQ) of 3 mg/dL. Below the LOQ at concentrations of 2.0 to 3.0 mg/dL Lp(a)-C, the imprecision was 18.5%. Lp(a)-P was measured using the same electrophoresis approach accompanied by immunostaining of apolipoprotein(B) [apo(B)] in the apo(B)-containing lipoprotein bands, which were composed of LDL, intermediate density lipoprotein, very low-density lipoprotein, and Lp(a) (26). Total apo(B) was quantified by immunoassay (Roche Diagnostics, Indianapolis, IN), and the percentage of apo(B) in the Lp(a) band and the molecular weight of apo(B) were used to calculate Lp(a)-P. The total imprecision was <10%, and the LOQ was 50 nmol/L. Below the LOQ at concentrations of 40 to 50 nmol/L Lp(a)-P, the imprecision was 13% to 25%.
Plasma fasting triglyceride, total cholesterol, and high-density lipoprotein cholesterol (HDL-C) concentrations were measured as described previously (27). LDL-C was calculated based on the Friedewald formula in participants with triglycerides <400 mg/ dL. The calculated LDL-C included the cholesterol contained in Lp(a) particles. To account for this overlap in the LDL-C covariate with the Lp(a) exposure variable, we subtracted Lp(a)-C from LDL-C to get the non-Lp(a) LDL-C.
CALCIFIC AORTIC VALVE DISEASE
CAVD was determined by the presence of aortic valve calcification that was obtained at study baseline through cardiac computed tomography imaging as previously described (28). Each participant was scanned twice; scans were interpreted at the core laboratory at the Los Angeles Biomedical Research Institute of the Harbor-University of California Los Angeles Medical Center by experienced readers, blinded to clinical information. Any observed calcified focus that extended to the aortic root was deemed aortic valve calcium as described previously (29).
CORONARY HEART DISEASE
After the baseline examination, study participants were followed up every 9 to 12 months by telephone to obtain information on interim hospital admissions, cardiovascular outpatient diagnoses, and deaths. Two study physicians blinded to other study data independently reviewed the medical records and adjudicated CHD events through December 2013. Incident CHD was defined as the first occurrence of any of the following: myocardial infarction (n = 150), resuscitated cardiac arrest (n = 24), CHD death (n = 70), or definite angina (n = 132). Definite angina was defined as symptoms of typical chest pain and physician diagnosis of angina followed by coronary artery bypass grafting and percutaneous coronary intervention, evidence of ischemia by stress tests or resting electrocardiogram, or [greater than or equal to]70% obstruction on coronary angiography. In addition, there were 61 cases of "probable angina," but only 18 were included as CHD cases in the present analysis. Probable angina cases that were included showed symptoms of typical chest or atypical symptoms and physician diagnosis of angina followed by coronary artery bypass grafting. Four cases of probable angina cases followed by percutaneous coronary intervention were excluded, as obstruction did not reach 70%. An additional 14 individuals who did not experience angina and underwent percutaneous coronary intervention without evidence of obstruction [greater than or equal to]70% were also excluded. Some individuals suffered multiple CHD events.
Statistical analyses were conducted using Stata (version 12.1, Stata Corp). Baseline characteristics are presented as medians (interquartile range) for continuous variables and frequencies (%) for categorical variables. The difference between continuous variables among racial/ethnic groups was tested using the Tukey-Kramer test, and z-test was used to compare categorical variables. Missing data were excluded when calculating frequencies. Additionally, correlations among Lp(a)-M, Lp(a)-P, and Lp(a)-C methods were evaluated using Deming regression, which accounted for assay imprecision. Martingale residuals showed departures from linearity for all 3 measures of Lp(a). The Lp(a)-C and Lp(a)-P assays had relatively high LOQ at 3.0 mg/dL and 50 nmol/L, respectively. To avoid analyzing individuals with Lp(a) concentrations below the LOQ, the upper 15th percentile was selected as a cutoff in race-stratified analyses to identify risk of CHD and CAVD associated with this selected designation of "high" Lp(a). For each Lp(a) measure, a generalized linear model with log link function was used to estimate the relative risk of Lp(a) for the presence of CAVD (>0 vs = 0) with a 95% CI. Statistical adjustments were made for age, sex, hypertension (systolic blood pressure [greater than or equal to]140 mmHg or on hypertension medication), smoking, education status, diabetes, non-Lp(a)-LDL-C, and HDL-C. Race/ethnicity was included as a covariate in unstratified analyses. Cox regression was used to test for associations between each Lp(a) measurement and incident CHD, adjusting for the same covariates in the cross-sectional CAVD analysis.
Baseline characteristics of 4679 MESA participants are shown in Table 1, stratified by race/ethnicity. White MESA participants had a nominally higher incidence of CHD (138 cases, 8.1%) and prevalence of CAVD (248 cases, 14.5%). Chinese-American MESA participants had a lower incidence of CHD (27 cases, 4.8%) and lower prevalence of CAVD (37 cases, 6.6%). Black MESA participants showed significantly higher median concentrations of Lp(a)-M than other races/ethnicities (P < 0.001). Because of the relatively high LOQ values for Lp(a)-C (3.0 mg/dL) and Lp(a)-P (50 nmol/L), differences among races were not evaluated.
Deming regression analyses were conducted to determine interassay correlations among Lp(a) measures. The correlation coefficient between Lp(a)-M and Lp(a)-C was estimated at 0.89; the coefficient between Lp(a)-C and Lp(a)-P was 0.92. The correlation coefficient between Lp(a)-M and Lp(a)-P was 0.91. To control for the relatively high LOQvalues for Lp(a)-C and Lp(a)-P, additional regression analyses were conducted that excluded individuals below these limits. In subjects with Lp(a)-C results above the cohort 25th percentile (3.7 mg/dL), the Deming regression correlation coefficient between Lp(a)-M and Lp(a)-C was estimated at 0.83. In those with Lp(a)-P values above the cohort 25th percentile (70 nmol/L), the coefficient between Lp(a)-P and Lp(a)-M was 0.83.
Lp(a)-related risks of CAVD and incident CHD are shown in Table 2. Both upper 25th and 15th percentile cutoffs were examined using a dichotomized statistical approach (cutoff selection is outlined in the Methods and Discussion sections). Relative regression analysis determined cross-sectional associations between Lp(a) measurements and the prevalence of CAVD. Significant associations between all Lp(a) analytes and CAVD were observed using either upper 25th or 15th percentile cutoff values (P < 0.0001) following adjustments for age, sex, race/ethnicity, hypertension, smoking, education status, diabetes, non-Lp(a)-LDL-C, and HDL-C.
Cox regression analysis determined associations between Lp(a) measurements and incident CHD over a 12-year follow-up period with adjustment for the same covariates as above. A greater risk of incident CHD was observed for study participants in the upper 25th percentiles across all Lp(a) analytes: Lp(a)-M (P = 0.002), Lp(a)-C (P = 0.003), and Lp(a)-P (P = 0.004), as well as for those in the upper 15th percentiles: Lp(a)-M (P = 0.02), Lp(a)-C (P < 0.0001), and Lp(a)-P (P = 0.02).
RACE-STRATIFIED ANALYSES: RACE/SPECIFIC CUTOFFS
The study sample was next stratified by race/ethnicity and analyzed by race-specific cutoff values (Tables 3 and 4). Notably, race/specific upper 25th percentiles were found to be an unsuitable cutoff because they would have included white, Chinese-American, and Hispanic participants with Lp(a)-C and Lp(a)-P concentrations that were below assay LOQs. Therefore, a race-specific upper 15th percentile cutoff was used for each race/ethnicity and for each Lp(a) analyte, e.g., the upper 15th percentile of Lp(a)-M corresponded to 83.9 mg/dL in black individuals and 49.3 mg/dL in whites (all cutoff values are indicated in Tables 3 and 4). Because race stratification resulted in 12 associations per outcome, a Bonferroni correction was applied whereby associations were deemed significant at P = 0.004.
Lp(a)-related risk of CAVD prevalence using the above approach is presented in Table 3. In white participants, significant 71% to 74% greater risks of CAVD were found in the upper 15th percentiles for all Lp(a) analytes (P < 0.0001). No significant associations were observed with CAVD in black, Chinese-American or Hispanic participants. Race-stratified Lp(a)-related CHD risk is presented in Table 4. No significant associations with CHD were observed in any race/ethnicity.
RACE-STRATIFIED ANALYSES: SUBCOHORT-BASED CUTOFFS
A subcohort-specific cutoff of the upper 25th percentile was used across all races/ethnicities for each Lp(a) analyte. Lp(a)-related risk of CAVD prevalence using this approach is presented in Table 1 of the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol63/issue12. In terms of significance and effect sizes, the strongest relations were found in white participants across all Lp(a) analytes (P < 0.0001) with 67% to 74% greater risks of CAVD in those in the upper 25th percentiles compared with individuals below this cutoff. No significant associations were observed in other racial/ethnic groups for any Lp(a) target with CAVD following correction for multiple comparisons. Lp(a)-related CHD risk stratified by race/ ethnicity is presented in Table 2 of the online Data Supplement. No significant associations were observed in any racial/ethnic groups for any Lp(a) target with CAVD following correction for multiple comparisons.
In this subcohort of 4679 MESA participants, we found that, regardless of the targeted component (cholesterol content, particle number, or particle mass), Lp(a) was significantly associated with prevalent CAVD and incident CHD following adjustment for other risk factors, including race/ethnicity. Race-stratified analyses were limited by statistical power except for the white group, in which Lp(a) was shown to associate with CAVD regardless of targeted Lp(a) component.
Although the current study represents the first to examine relations of Lp(a)-M, Lp(a)-C, and Lp(a)-P with CAVD and incident CHD, smaller studies examining coronary artery disease as the outcome have been conducted. A prospective study of466 patients over a 4-year period (19) showed that Lp(a)-C, but not Lp(a)-M, was associated with angiographically determined coronary artery disease (P < 0.001) and cardiovascular events (P = 0.01). And yet the Lp(a)-M assay used in the above study relied on an apo(a) isoform-sensitive antibody, which may have contributed to the null finding. A subsequent study in Framingham Offspring participants (16) reported the opposite finding in that men in the upper tertile of Lp(a)-M were at greater risk of incident CHD over a 12.3-year follow-up period. No association was found for Lp(a)-C and CHD, and investigators concluded that Lp(a)-C may not account for Lp(a)-related risk or may be an inaccurate proxy of Lp(a). Results in MESA participants contradict this latter finding. This discrepancy between MESA and Framingham Offspring findings regarding Lp(a)-C may be because of differences in study samples, the Lp(a)-C analytical technique, or statistical approach. Ultimately, our findings support the conclusion that the 3 Lp(a) assays highly correlate and associate with CAVD and incident CHD risk when examined in the entire cohort. However, the high LOQs of Lp(a)-C and Lp(a)-M assays posed limitations on our statistical approaches used in this study and likely limit their utility in other settings.
As expected, stratifying analyses by race/ethnicity attenuated most associations such that almost all results were rendered nonsignificant after accounting for multiple comparisons testing, with 1 exception. For CAVD, the strongest relations of Lp(a) and CAVD that were observed in the entire study sample were found to be primarily driven by whites in whom associations among all Lp(a) analytes and CAVD remained significant (P < 0.0001), and effect sizes nominally increased. No differences among Lp(a) analytes were apparent in whites. By comparison, nonsignificant findings were observed for other races/ethnicities. Although hazard ratio effect sizes were largely similar across race groups, there was substantial overlap in confidence intervals among these associations. As such, results for race-stratified analyses were largely inconclusive. Here again, the high LOQs of 2 of the assays, in addition to the reduced number of individuals in each group, limited statistical power in these analyses.
CONSIDERATIONS FOR LP(A) ASSAY SELECTION
All 3 Lp(a) methods are assays offered in clinical laboratories, but only the mass assay is automated and widely accepted. Although results for the entire sample were highly correlated between the 3 assays, each assay has limitations that should be considered in assay selection. The Lp(a)-M method used in the current study is an immunoassay that measures concentrations of Lp(a) expressed in milligrams per deciliter rather than nanomoles per liter--the latter being recommended by the IFCC Lp(a) standardization program (12). In addition, the antibody used in this assay is directed against a size-heterogeneous component of Lp(a), the apo(a) protein. Although calibrators were included in the Lp(a)-M assay, overestimations or underestimations of Lp(a)-M concentrations in individuals with extreme apo(a) sizes remain possible (13). In terms of its advantages, Lp(a)-M is the most extensively used in research and clinical settings and is comparatively less labor-intensive, making it less costly compared with Lp(a) assessment through electrophoresis.
In contrast to the above Lp(a)-M immunoassay, the Lp(a)-C and Lp(a)-P methods are based on electrophoretic separation of the lipoprotein subclasses followed by gel staining for quantifying either cholesterol or apo(B)-100, respectively. Both are insensitive to apo(a) size variability, and yet, both Lp(a)-C and Lp(a)-P assays are limited in terms of their analytical sensitivity and have relatively high LOQs at 3.0 mg/dL and 50 nmol/L, respectively. Depending on the race/ethnic group, approximately 50% to 85% of MESA participants were below these concentrations. As a consequence of the relatively high LOQs, Lp(a)-C and Lp(a)-P assays showed imprecision at concentrations that would be common in the general population. Specifically, CVs for the Lp(a)-C assay were calculated to be 18.5% at concentrations between 2 and 3 mg/dL. Similarly, CVs for the Lp(a)-P assay ranged from 13% to 25% at concentrations between 40 and 50 nmol/L. Taken together, the LOQs and assay imprecision may limit clinical application in treatment monitoring of patients undergoing Lp(a)-lowering therapy and also may limit their utility in long-term risk stratification cohort studies.
A further consideration in evaluating these results is the unresolved issue of appropriate Lp(a) cutoff values that best discriminate disease risk. One could observe from the current results that cutoff selection, particularly across race/ ethnicity, nominally influences effect sizes, confidence intervals, and P values among the different Lp(a) analytes. We speculate that this is less related to Lp(a)-associated risk, but that more accurate method-specific and race/ethnicity-specific cutoffs must be established. We also made this observation in our previous studies of Lp(a) (7,21). To establish more accurate cutoff values in terms of risk discrimination, a multicohort approach is required to first optimize the Lp(a) cutoff value with respect to clinical sensitivity and specificity and then test it in a replicate cohort(s). Given the current and previous results (7, 21-23), exploring Lp(a) race-specific cutoffs using such an approach may be warranted.
STRENGTHS AND LIMITATIONS
There are a number of strengths and limitations of the current study. We believe that this is the largest study to examine Lp(a)-M, Lp(a)-P, and Lp(a)-C and their relations with CAVD and incident CHD in a multiethnic cohort. In terms of limitations apart from those related to analytical techniques, our definition of CAVD was restricted to the presence of aortic valve calcification at baseline. It also must be acknowledged that the number of CHD events was limited in certain groups, e.g., 27 CHD cases were observed in Chinese-American individuals. Likewise, CAVD was least prevalent in Chinese-American participants; therefore, the results should be interpreted with caution. Finally, multiple adjustments were made within the statistical model, but the possibility of residual confounding cannot be excluded.
In summary, we found that 3 analytical assays that measure different components of the Lp(a) particle were associated with risks ofincident CHD and prevalent CAVD independent of typical risk factors. On stratifying by race/ ethnicity, most associations were rendered nonsignificant, and we cannot conclude that assay performances were different among races/ethnicities. The analytical limitations of the Lp(a)-C and Lp(a)-P methods, specifically their relatively high LOQs and imprecision, will constrain their utility in research environments and clinical settings. As such, despite its inaccuracy because of the different isoform sizes of apo(a), the automated Lp(a)-M immunoassay remains the most attractive option for routine clinical laboratory assessment of Lp(a)-related CHD and CAVD risk in the general population. However, to further define Lp(a) cutoff values that identify disease risk across different populations, additional large multiethnic cohort studies will be required that rely on an Lp(a) methodology that is well-validated, apo(a)-isoform insensitive, and generates values traceable to a common reference standard.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contribution to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: J.R. Kizer, Gilead Sciences, Inc., Pfizer, Inc. Honoraria: None declared.
Research Funding: Contracts HHSN268201500003I, N01-HC95159, N01-HC-95160, N01-HC- 95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01- HC-95167, N01-HC-95168 and N01-HC-95169 from the National Heart, Lung, and Blood Institute and by grants UL1-TR-000040 and UL1-TR-001079 from NCRR.
Expert Testimony: J.R. Kizer, Cooney, Sculley, and Dowling.
Patents: None declared.
Role of Sponsor: The funding organizations played a direct role in the design of study, review and interpretation of data, preparation of manuscript, and final approval of manuscript.
Acknowledgment: The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.
(1.) Berg K. A new serum type system in man-the Lp system. Acta Pathol Microbiol Scand 1963;59:369-82.
(2.) Erqou S, Kaptoge S, Perry PL, DiAngelantonio E, Thompson A, White IR, et al. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. JAMA 2009;302:412-23.
(3.) Kamstrup PR, Tybjaerg-Hansen A, Steffensen R, Nordestgaard BG. Genetically elevated lipoprotein(a) and increased risk of myocardial infarction. JAMA 2009; 301:2331-9.
(4.) Nordestgaard BG, Chapman MJ, Ray K, Boren J, Andreotti F, Watts GF, et al. Lipoprotein(a) as a cardiovascular risk factor: current status. Eur Heart J 2010;31: 2844-53.
(5.) Arsenault BJ, Boekholdt SM, Dube MP, Rheaume E, Wareham NJ, Khaw KT, et al. Lipoprotein(a) levels, genotype, and incident aortic valve stenosis: a prospective Mendelian randomization study and replication in a case-control cohort. Circ Cardiovasc Genet 2014;7: 304-10.
(6.) Thanassoulis G, Campbell CY, Owens DS, Smith JG, Smith AV, Peloso GM, et al. Genetic associations with valvular calcification and aorticstenosis. N Engl J Med 2013;368:503-12.
(7.) Cao J, Steffen BT, Budoff M, Post WS, Thanassoulis G, Kestenbaum B, et al. Lipoprotein(a) levels are associated with subclinical calcific aortic valve disease in white and black individuals: the Multi-Ethnic Study of Atherosclerosis. Arterioscler Thromb Vasc Biol 2016;36:1003-9.
(8.) Raal FJ, Giugliano RP, Sabatine MS, Koren MJ, Langslet G, Bays H, et al. Reduction in lipoprotein(a) with PCSK9 monoclonal antibody evolocumab(AMG145): a pooled analysis of more than 1,300 patients in 4 phase II trials. J Am Coll Cardiol 2014;63:1278-88.
(9.) Santos RD, Raal FJ, Catapano AL, Witztum JL, Steinhagen-Thiessen E, Tsimikas S. Mipomersen, an antisense oligonucleotide to apolipoprotein B-100, reduces lipoprotein(a) in various populations with hypercholesterolemia: results of 4 phase III trials. Arterioscler Thromb Vasc Biol 2015;35:689-99.
(10.) Tsimikas S, Viney NJ, Hughes SG, Singleton W, Graham MJ, Baker BF, et al. Antisense therapy targeting apolipoprotein(a): a randomised, double-blind, placebo-controlled phase 1 study. Lancet 2015;386: 1742-83.
(11.) Viney NJ ,van Capelleveen JC, Geary RS, Xia S, Tami JA, Yu RZ, et al. Antisense oligonucleotides targeting apolipoprotein(a) in people with raised lipoprotein(a): two randomised, double-blind, placebo-controlled, dose-ranging trials. Lancet 2016;388:2239 -53.
(12.) Marcovina SM, Albers JJ. Lipoprotein (a) measurements for clinical application. J Lipid Res 2016;57: 526-37.
(13.) Baudhuin LM, Hartman SJ, O'Brien JF, Meissner I, Galen RS, Ward JN, et al. Electrophoretic measurement of lipoprotein(a) cholesterol in plasma with and without ultracentrifugation: comparison with an immunoturbidimetric lipoprotein(a) method. Clin Biochem 2004; 37:481-8.
(14.) Kulkarni KR, Garber DW, Marcovina SM, Segrest JP. Quantification of cholesterol in all lipoprotein classes by the VAP-II method. J Lipid Res 1994;35:159-68.
(15.) Kulkarni KR. Cholesterol profile measurement by vertical auto profile method. Clin Lab Med 2006;26:787-802.
(16.) Lamon-Fava S, Marcovina SM, Albers JJ, Kennedy H, Deluca C, White CC, et al. Lipoprotein(a) levels, apo(a) isoform size, and coronary heart disease risk in the Framingham Offspring Study. J Lipid Res 2011;52: 1181-7.
(17.) Konerman M, Kulkarni K, Toth PP, Jones SR. Lipoprotein(a) particle concentration and lipoprotein(a) cholesterol assays yield discordant classification of patients into four physiologically discrete groups. J Clin Lipidol 2012;6:368-73.
(18.) Baudhuin L, Bryant S, Spears G, Hartman S, Somers V, Berger P, et al. Lipoprotein(a) cholesterol but not lipoprotein(a) mass is significantly correlated with angiographic coronary artery disease and major adverse events. Circulation 2015;132:A11043.
(19.) McConnell JP, Baudhuin LM, Berger PB, Kullo IJ, O'Brien JF, Bryant SC, et al. Lipoprotein(a) cholesterol, but not Lp(a) mass, is an independent predictor of angiographic coronary artery disease and subsequent cardiovascular events in patients referred for coronary angiography [Abstract]. Circulation 2007;116(Suppl 16): 3608.
(20.) Yeang C, Clopton PC, Tsimikas S. Lipoprotein(a)cholesterol levels estimated by vertical auto profile correlate poorly with Lp(a) mass in hyperlipidemic subjects: implications for clinical practice interpretation of Lp(a)-mediated risk. J Clin Lipidol 2016;10: 1389-96.
(21.) Guan W, Cao J, Steffen BT, Post WS, Stein JH, Tattersall MC, et al. Race is a key variable in assigning lipoprotein(a) cutoff values for coronary heart disease risk assessment: the Multi-Ethnic Study of Atherosclerosis. Arterioscler Thromb Vasc Biol 2015;35:996-1001.
(22.) Virani SS, Brautbar A, Davis BC, Nambi V, Hoogeveen RC, Sharrett AR, et al. Associations between lipoprotein(a) levels and cardiovascular outcomes in black and white subjects: the Atherosclerosis Risk in Communities (ARIC) Study. Circulation 2012;125:241-9.
(23.) Hoogeveen R, Sun W, Virani SS, Ballantyne CM. Circulating lipoprotein(a)[Lp(a)] levels and risk of cardiovascular disease: the Atherosclerosis Risk in Communities (ARIC)study. Atherosclerosis 2016;e125-6.
(24.) Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epide miol 2002;156:871-81.
(25.) MESA. Available from: www.mesa-nhlbi.org (Accessed August 2017).
(26.) Guadagno PA, Summers Bellin EG, Harris WS, Dayspring TD, Hoefner DM, Thiselton DL, et al. Validation of a lipoprotein(a) particle concentration assay by quantitative lipoprotein immunofixation electrophoresis. Clin ChimActa 2015;439:219-24.
(27.) Tsai MY, Johnson C, Kao WH, Sharrett AR, Arends VL, Kronmal R, et al. Cholesteryl ester transfer protein genetic polymorphisms, HDL cholesterol, and subclinical cardiovascular disease in the Multi-Ethnic Study of Atherosclerosis. Atherosclerosis 2008;200:359-67.
(28.) Katz R, Wong ND, Kronmal R, Takasu J, Shavelle DM, Probstfield JL, et al. Features of the metabolic syndrome and diabetes mellitus as predictors of aortic valve calcification in the Multi-Ethnic Study of Atherosclerosis. Circulation 2006;113:2113-9.
(29.) Budoff MJ, Takasu J, Katz R, Mao S, Shavelle DM, O'Brien KD, et al. Reproducibility of CT measurements of aortic valve calcification, mitral annulus calcification, and aortic wall calcification in the multiethnic study of atherosclerosis. Acad Radiol 2006; 13:166-72.
Jing Cao, [1, 2]([dagger]) Brian T. Steffen,  ([dagger]) Weihua Guan,  Matthew Budoff,  Erin D. Michos,  Jorge R. Kizer,  Wendy S. Post,  and Michael Y. Tsai  *
 Department of Pathology and Immunology, Baylor College of Medicine, Houston, TX;  Department of Pathology, Texas Children's Hospital, Houston, TX;  Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, MN;  Division of Biostatistics, School of Public Health, University of Minnesota, Minneapolis, MN;  Department of Medicine, University of California, Los Angeles, CA;  Division of Cardiology, Johns Hopkins University School of Medicine, Balti more, MD;  Department of Medicine and Department of Epidemiology and Population Health, Albert Einstein College of Medicine, Bronx, NY.
* Address correspondence to this author at: University of Minnesota, Twin Cities, 420 Delaware St. SE, Mayo Mail Code 609, Minneapolis, MN 55455. Fax 612-625-1121; e-mail firstname.lastname@example.org.
([dagger]) Cao and B.T. Steffen contributed equally to this work.
Received December 23, 2016; accepted August 17, 2017.
Previously published online at DOI: 10.1373/clinchem.2016.270751
 Nonstandard abbreviations: Lp(a), lipoprotein(a); LDL, low-density lipoprotein; CHD, coronary heart disease; CAVD, calcific aortic valve disease; LDL-C, low-density lipoprotein cholesterol; Lp(a)-C, Lp(a) cholesterol content; Lp(a)-P, Lp(a) particle concentration; Lp(a)-M, Lp(a) mass; MESA, Multi-Ethnic Study of Atherosclerosis; apo(a), apolipoprotein(a); LOQ, lower limit of quantification; apo(B), apolipoprotein(B); HDL-C, high-density lipoprotein cholesterol.
Table 1. Characteristics of MESA participants across 4 race/ethnicity groups. (a) Blacks Whites N 1347 1709 Age (years) 61 (52-70) 62 (54-71) Sex (male) 621 (46.1%) 813 (47.6%) Smoking 726 (53.9%) 929 (54.5%) Diabetes 196 (14.6%) 86 (5.0%) Hypertension 428 (31.8%) 325 (19.0%) On hypertension 613 (45.5%) 493 (28.9%) medications Non-Lp(a) LDL-C 113 (92-133) 115 (97-136) (mg/dL) HDL-C (mg/dL) 50 (41-61) 50 (41-62) Lp(a)-M (mg/dL) 35.1 (20.4-61.6) (d) 13.0 (5.8-29.6) Lp(a)-C(mg/dL) (e) 2.8 (1.3-7) 1.2 (0.6-2.8) Lp(a)-P (nmol/L) (f) 54 (22-119) 19 (9-48) CHD (12-year 87 (6.5%) 138 (8.1%) follow-up) Presence of aortic 157(11.7%) 248 (14.5%) valve calcification Chinese-Americans Hispanics N 559 1064 Age (years) 62 (53-71) 61 (52-69) Sex (male) 217 (38.8%) 517 (48.6%) Smoking 137 (24.5%) 504 (47.4%) Diabetes 55 (9.8%) 171 (16.1%) Hypertension 126 (22.5%) 257 (24.2%) On hypertension 138 (24.7%) 305 (28.7%) medications Non-Lp(a) LDL-C 114 (96-132) 116 (97-137) (mg/dL) HDL-C (mg/dL) 48 (40-58) (b) 45 (38-54) (c) Lp(a)-M (mg/dL) 12.9 (7.7-23.4) 13.1 (6.3-28.8) Lp(a)-C(mg/dL) (e) 1.0 (0.5-1.9) 1.0 (0.5-2.4) Lp(a)-P (nmol/L) (f) 18 (10-36) 19 (9-49) CHD (12-year 27 (4.8%) 69 (6.5%) follow-up) Presence of aortic 37 (6.6%) 140 (13.2%) valve calcification (a) Data are shown as medians (interquartile range) for continuous variables and as counts (%) for categorical variables. Definition: smoking (former and current), diabetes (treated and untreated), hypertension (systolic blood pressure [greater than or equal to]140 mm Hg). (b,c,d) P <0.001 compared with other race/ethnicity groups. (e) The LOQ for the Lp(a)-C assay was 3.0 mg/dL; median values must therefore be interpreted with caution. (f) The LOQ for the Lp(a)-P assay was 50 nmol/L; median values must therefore be interpreted with caution. Table 2. Lp(a)-related risks of prevalent CAVD and development of CHD in a subcohort of MESA participants (n = 4593). (a) Lp(a)-M Lp(a)-C (b) (mg/dL) (mg/dL) Calcific aortic valve disease (d) Upper 25th percentile 39.9 3.7 Estimated RR 1.49 1.48 95% CI (1.27-1.74) (1.26-1.73) P value <0.0001# <0.0001# Upper 15th cutoff value 59.7 6.9 Estimated RR 1.54 1.49 95% CI (1.29-1.83) (1.23-1.82) P value <0.0001# <0.0001# Incident coronary heart disease (e) Upper 25th percentile 39.9 3.7 Estimated HR 1.49 1.46 95% CI (1.16-1.91) (1.14-1.87) P value 0.002# 0.003# Upper 15th cutoff value 59.7 6.9 Estimated HR 1.40 1.76 95% CI (1.05-1.88) (1.33-2.34) P value 0.02# <0.0001# Lp(a)-P (c) (nmol/L) Calcific aortic valve disease (d) Upper 25th percentile 70 Estimated RR 1.49 95% CI (1.27-1.75) P value <0.0001# Upper 15th cutoff value 119 Estimated RR 1.52 95% CI (1.27-1.81) P value <0.0001# Incident coronary heart disease (e) Upper 25th percentile 70 Estimated HR 1.45 95% CI (1.13-1.86) P value 0.004# Upper 15th cutoff value 119 Estimated HR 1.41 95% CI (1.06-1.87) P value 0.02# The P values are in bold to highlight the significance. (a) Adjustments were made for age, sex, race/ethnicity, hypertension, smoking, education status, diabetes, LDL-C, and HDL-C. (b) The LOQ for the Lp(a)-Cassay was 3.0 mg/dL. (c) The LOQ for the Lp(a)-P assay was 50 nmol/L. (d) Relative risk ratios (RRs) and 95% CI are presented for individuals in the upper 25th and 15th percentiles for each Lp(a) marker. (e) Hazard ratios(HRs)and 95% CI are presented for individuals in upper 25th and 15th percentiles for each Lp(a) marker. Note: The P values are in bold to highlight the significance are indicated with #. Table 3. Race-stratified Lp(a)-related risk of prevalent CAVD. (a) Lp(a)-M Lp(a)-C Lp(a)-P (mg/dL) (mg/dL) (nmol/L) Black Upper 15th% cutoff value 83.9 9.9 167 Estimated RR 1.50 1.50 1.36 95% CI (1.02-2.21) (1.02-2.21) (0.95-1.96) P value 0.04 0.04 0.10 White Upper 15th% cutoff value 49.3 5.7 102 Estimated RR 1.71 1.74 1.74 95% CI (1.37-2.14) (1.37-2.21) (1.37-2.20) P value <0.0001# <0.0001# <0.0001# Chinese-American Upper 15th% cutoff value 36.2 3.0 62 Estimated RR 1.847 1.949 1.725 95% CI (0.45-7.53) (0.72-5.29) (0.40-7.54) P value 0.39 0.19 0.47 Hispanic Upper 15th% cutoff value 46.0 4.6 93 Estimated RR 1.26 1.50 1.50 95% CI (0.84-1.88) (1.07-2.11) (1.06-2.12) P value 0.26 0.02 0.02 The P values are in bold to highlight the significance. (a) Relative risk ratios (RR) and 95% CI are presented for individuals in respective upper 15th percentiles for each Lp(a) marker. Adjustments were made for age, sex, hypertension, smoking, education status, diabetes, LDL-C, and HDL-C. Upper 15th percentile cutoff values are shown for each analyte and racial/ethnic group. Note: The P values are in bold to highlight the significance are indicated with #. Table 4. Race-stratified Lp(a)-related risk of developing CHD over a 12-year follow-up period. (a) Lp(a)-M Lp(a)-C Lp(a)-P (mg/dL) (mg/dL) (nmol/L) Black Upper 15th% cutoff value 83.9 9.9 167 Estimated HR 2.01 1.92 1.59 95% CI (1.22-3.30) (1.16-3.18) (0.95-2.66) P value 0.006 0.01 0.08 White Upper 15th% cutoff value 49.3 5.7 102 Estimated HR 1.45 1.55 1.53 95% CI (0.95-2.24) (1.01-2.38) (1.01-2.32) P value 0.09 0.05 0.05 Chinese-American Upper 15th% cutoff value 36.2 3.0 62 Estimated HR 2.54 2.43 2.02 95% CI (1.05-6.16) (1.02-5.78) (0.82-4.96) P value 0.04 0.04 0.13 Hispanic Upper 15th% cutoff value 46.0 4.6 93 Estimated HR 1.69 1.14 1.04 95% CI (0.92-3.08) (0.59-2.18) (0.54-2.00) P value 0.09 0.70 0.91 (a) Hazard ratios (HRs)and 95% Clare presented for individuals in upper 15th percentiles for each Lp(a) marker. Adjustments were made for age, sex, hypertension, smoking, education status, diabetes, LDL-C, and HDL-C. Upper 15th percentile cutoff values are shown for each analyte within each racial/ethnic group.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Lipids, Lipoproteins, and Cardiovascular Risk Factors|
|Author:||Cao, Jing; Steffen, Brian T.; Guan, Weihua; Budoff, Matthew; Michos, Erin D.; Kizer, Jorge R.; Post,|
|Date:||Nov 1, 2017|
|Previous Article:||Simultaneously Monitoring Immune Response and Microbial Infections during Pregnancy through Plasma cfRNA Sequencing.|
|Next Article:||High Lipoprotein(a) and Low Risk of Major Bleeding in Brain and Airways in the General Population: a Mendelian Randomization Study.|