Multi-center determination of galectin-3 assay performance characteristics:

Results Galectin-3 results were equivalent when measured in serum or EDTA plasma. Imprecision studies demonstrated that the total CV was < 10% at a low concentration of 6 ng/mL, 7% near the mid-level concentration of 21 ng/mL, and 15% at the high level of 70 ng/mL. The limit of blank was 0.86 ng/mL, the limit of detection was 1.13 ng/mL, and the limit of quantitation was 0.96 ng/mL. The linear measurement range was 0.96–130 ng/mL and there was no high-dose hook at levels up to 500 ng/mL. No cross-reactivity with nine compounds structurally related to galectin-3 and no interference from 22 common medications, icterus or lipemia was found. Hemolysis and presence of human anti-mouse antibodies or rheumatoid factor were found to be potential sources of interference. Samples can be stored for up to 15 days at either 22–28 °C or 2–8 °C before analysis; measurements are stable after storage at − 20 °C or − 70 °C for at least 6 months and through six freeze–thaw cycles. The 90th, 95th and 97.5th percentile of the normal reference interval was 17.6, 20.3 and 22.1 ng/mL, respectively. Galectin-3 in the acute decompensated heart failure patients ranged from 4.0 to 75 ng/mL; using a cutpoint of 22.1 ng/mL in an unadjusted analysis, galectin-3 values were associated with an outcome of death ( p = 0.035). Galectin-3 is associated with severity of heart failure as indicated by NYHA Class ( p -value for trend, 0.017). Conclusion This ELISA for galectin-3 measurement demonstrated acceptable analytical characteristics and was associated with mortality in a preliminary unadjusted analysis. Keywords Galectin-3 Assay Analytical performance Heart failure Introduction Galectin-3 is a carbohydrate-binding lectin best known for its regulatory functions. First thought to largely play a physiological role, galectin-3 research conducted over the past few years has revealed a role in the development of pathological conditions such as tumorigenesis, autoimmune and inflammatory diseases [1–4] . Expression of galectin-3 has been associated with the epithelium and inflammatory cells including macrophages, neutrophils and mast cells. Galectin-3 has been implicated in a variety of biological processes important in heart failure (HF), including myofibroblast proliferation, fibrogenesis, tissue repair, cardiac remodeling and inflammation [5–10] . These observations have lead to current research directions that focus on the modulation of galectin-3 as a potential diagnostic and therapeutic target. Galectin-3 has received much attention as a mediator of HF development and progression; it has been recognized for some time that galectin-3 is involved in the process of apoptosis [11] . Myocardial biopsies in failure-prone hypertrophied rat hearts and hypertrophied hearts of patients with aortic stenosis showed the up-regulation of galectin-3 [10] . Similarly, elevated circulating levels of galectin-3 have been detected in patients with acute decompensated [12] and chronic [13,14] HF. Perhaps most relevant from a clinical standpoint is the fact that galectin-3 is involved in the development of fibrosis and adverse cardiac remodeling, two key pathophysiological mechanisms responsible for the progression of heart failure. The availability of an accurate and reproducible assay with well-defined performance characteristics, including pre-analytical variables, are fundamental for laying the foundation for establishment of a novel biomarker [15] and for evidence-based laboratory medicine [16] . The purpose of this study was to define key performance characteristics of a new assay for galectin-3 for use in the clinical assessment of patients with HF. Methods Galectin-3 assay Galectin-3 levels were determined using an optimized enzyme-linked immunosorbent assay (ELISA) (Galectin-3 assay; BG Medicine, Waltham, MA, USA). This is a two-site immunoassay in which the capture antibody, a rat IgG2a monoclonal antibody raised against mouse galectin-3 protein, is immobilized on 96-well plates. The overall homology between mouse and human galectin-3 protein is 85% [17] . In the N-terminal portion of the protein where the epitope for the assay is located, there is 100% homology between human and murine galectin-3. The assay requires 25 μl of patient sample or control which is pre-diluted with 225 μl of assay diluent for duplicate measurement. For standardization of the assay, galectin-3 was prepared from commercially available recombinant human galectin-3 protein expressed in E. coli ; galectin-3 concentrations were assigned by spectrophotometric absorption at 280 nm using an extinction coefficient of 1.29 cm −1 as determined using the method described by ref. [17] . Serial dilutions of lyophilized standard material provided in the kit were prepared with each run to produce an eight-point standard curve ranging from 0 ng/mL to 10 ng/mL. One hundred microliters of the standards, diluted sample and controls was then applied to the 96-well plate, in duplicate. After 1-hour of incubation at 20–25 °C, the plate was rinsed with buffer and the wells dried. The detection antibody (tracer), a mouse monoclonal antibody that targets human galectin-3 protein and is conjugated with horseradish peroxidase (HRP) was then added to each well and incubated for 1 h at 20–25 °C. After washing, rinsing and drying the plate, tetramethyl benzidine substrate was added, followed by incubation for 20 min in the dark. An acidic solution was then added to each well to stop the reaction. Reading of the plates was performed at 450 nm with data reduction using a third order polynomial model. High and low controls for the assay are provided with each kit; the controls were prepared by spiking a known amount of recombinant human galectin-3 into SeraCon II CD, a delipidated, defibrinated, charcoal-stripped human serum substitute (SeraCare Life Sciences, Milford, MA). The overall assay performance time is approximately 3.5 h. Testing sites Measurements were performed at University of Maryland School of Medicine, Baltimore, MD, USA (UMB), Lahey Clinical Medical Center, Burlington, MA, USA (LCMC) and University of California at San Francisco, San Francisco, CA USA (UCSF). Appropriate institutional review board approval was obtained for testing at each site. Imprecision Studies at each site were designed and conducted according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) EP5-A2. EDTA plasma purchased from Research Blood Components, LLC (Brighton, MA) was divided into aliquots and spiked with human recombinant galectin-3 to prepare controls which spanned the range of galecin-3 concentrations. These EDTA plasma samples were analyzed in quadruplicate (two sets of duplicates) for two runs per day, over 20 days. In accordance with CLSI EP5-A2, a fully nested Model II ANOVA with adjustment for estimating within-run, run-to-run, day-to-day and total imprecision. Limit of blank (LoB), limit of detection (LoD), limit of quantitation (LoQ) LoB, LoD and LoQ were determined according to CLSI guideline EP-17A. Assay diluent, a phosphate buffered saline solution, pH 7.4, containing 1% bovine serum albumin, was used as the matrix for establishing the LoB. SeraCon II CD was spiked with recombinant human galectin-3 for determining the LoD and LoQ. Calculation of LoB was accomplished by ranking the signal of 48 galectin-3 replicate measurements from highest to lowest. The 46th ranked measurement was the 95.5th percentile and represents the LoB. LoD was determining from calculating the SD of 16 replicates of a low galectin-3 sample at 1.09 ng/mL with the equation: LoD = LoB + (1.645⁎SD). LoQ was specified as the smallest galectin-3 concentration corresponding to a total coefficient of variation (CV) of no more than 20% and was determined by performing 16 replicate measurements for each of four low concentration galectin-3 serum pools. The lowest concentration having a CV ≤ 20% was extrapolated from the second order polynomial curve of the galectin-3 and is the specified LoQ for the assay. Linearity Linearity was assessed using high/low dilution, according to the recommendations of CLSI guideline EP06-A. Six high concentration samples were prepared by spiking with recombinant galectin-3 to produce concentrations of approximately 125 ng/mL. Each of these six high samples was then serially diluted with the corresponding native (low) samples at constant intervals (9 parts high + 1 part low, 8 parts high + 2 parts low, etc.). The measured values of the high and low concentration samples were used to determine the expected value for each of the serial dilutions. Linearity was evaluated by linear regression comparing the expected value ( x -axis) versus the measured value ( y -axis) of each sample. High-dose hook Serum and EDTA plasma samples that contained 500 ng/mL in galectin-3 were analyzed neat and diluted 1:2, 1:4, 1:8 and 1:16 in diluent, serum and EDTA plasma. The results were compensated for the dilution factors used and % recovery was calculated. Interfering substances Potential interference from endogenous and exogenous substances was determined according to guidance in CLSI EP7-A2. The paired-difference approach was utilized where a “base sample” of EDTA plasma (from Research Blood Components, LLC, Brighton, MA) was divided to create a “test sample” spiked with the potential interferent and a “control sample” spiked with the diluent or solvent only. Recovery was calculated according to: % Recovery = [(Test Pool Galectin-3 Dose / Control Pool Dose) × 100%]. Endogenous substances tested were unconjugated bilirubin, conjugated bilirubin, triglycerides, total cholesterol, creatinine and purified hemoglobin. Recovery of packed cell lysate measurements was also examined in two different experiments. Possible interference from human anti-mouse antibodies (HAMA) and rheumatoid factor (RF) on the galectin-3 measurements was also investigated. Medications (exogenous substances), including acetaminophen, acetylsalicylic acid, amlodipine, ampicillin, ascorbic acid, atenolol, caffeine, carvedilol, captopril, chloramphenicol, diclofenac, digoxin, diltiazem, disopyramide, dopamine, enalaprilat, furosemide, hydrochlorothiazide, ibuprofen, indomethacin, lidocaine, lisinopril, losartan, lovastatin, methyldopa, metoprolol, naproxen, nifedipine, quinidine, ramipril, spironolactone, theophylline, verapamil and warfarin were tested as potential interferences. Cross-reactivity Human recombinant members of the galectin protein family and collagens, including galectin 1, galectin-2, galectin-4, galectin-7, galectin-8, galectin-9, galectin-12, collagen I and collagen III were tested for potential interference with the assay by spiking a “base” galectin-3 EDTA plasma sample with each potential interferent to produce a “test” sample. Potential cross-reactivity was calculated according to: Cross-reactivity = (Measured concentration in test sample − measured concentration in base sample) / Measured dose of test sample ⁎ 100%). Measurement matrix Galectin-3 measurements were performed in matched serum and EDTA plasma samples from 50 individuals that spanned the concentration range of the assay. Linear regression and Bland-Altman analysis were performed to assess agreement between the matrices. Storage stability Eighteen EDTA plasma and serum pools were prepared from at least two samples each from using leftover de-identified specimens from suspected HF patients; galectin-3 was measured for each pool within 4 h of collection to determine the base ( t 0 ) value. Aliquots of each of the 18 serum and plasma pools were immediately stored at 22–28 °C, 2–8 °C or frozen at − 20 °C and − 70 °C. Measurement occurred at various times after t 0 , designated t n , according to the storage conditions. Measurements were carried out to 6 months. Sample stability under a given storage condition was established by comparing the t 0 value of each EDTA plasma or serum pool with values at subsequent time points, t n . The absolute difference of each of the 18 samples was averaged and expressed as the percent change in galectin-3 level between t 0 and t n . If the average galectin-3 level at an individual t n was within ± 25% of the t 0 value, then the sample was considered to be stable up to that t n under the storage condition. Freeze–thaw stability in serum and EDTA plasma was evaluated after storage at − 20 °C or − 70 °C overnight. Samples were then thawed at room temperature for 1 h to complete one freeze–thaw cycle and then refrozen at − 20 °C or − 70 °C overnight. Samples underwent three complete freeze–thaw cycles in this fashion with measurement at day 3. An additional three complete freeze–thaw cycles were then completed with measurement on day 8 (total six freeze cycles). Normal distribution The normal distribution for galectin-3 was determined for 1092 subjects (men: n = 520; women: n = 572) enrolled in the BioImage Study (NCT00738725), a large, ongoing, prospective, observational study that has recruited approximately 6800 men age 55–80 years and women age 60–80 years without known cardiac disease. Details on this cohort have been published and presented previously [18,19] . Clinical samples Galectin-3 was measured in EDTA plasma samples collected at hospital presentation with approval of the local Institutional Review Board from patients enrolled in a prospective, five-center, observational study that were ≥ 18 years of age and treated in an emergency department with a primary complaint of dyspnea. Patients included here ( n = 129) were diagnosed with acute decompensated HF (ADHF) as determined by an expert panel of physicians trained in cardiovascular medicine; the distribution of NYHA classification in the cohort was: Class I–II 26%, Class III 52% and Class IV 23%. The cohort was 73% male, age 61 ± 13 years; the race distribution was: white 39%, black 57%, other 4%; the body mass index (kg/m 2 ) was 31 ± 9 and eGFR was 56 ± 26 mL/min/1.73 m 2 . Further details on this cohort have been published previously [20] . A preliminary unadjusted analysis of the association between galectin-3 values above and below the 97.5th percentile of the normal subject distribution and mortality was conducted. The association of NYHA classification and galectin-3 values was plotted and the trend examined for significance in an unadjusted analysis. Results Imprecision Table 1 shows the precision data determined at the three sites for three galectin-3 levels at low, mid and elevated concentrations. Overall, within-run precision was less than 8%, run-to-run was less than 10% in all but two cases and day-to-day imprecision was less than 4%. Total CV at the low galectin-3 concentration of about 6 ng/mL was < 10%, total CV at the mid-level concentration of 21 ng/mL was approximately 7%, and total CV at the high level of approximately 70 ng/mL was about 15%. LoB, LoD, LoQ The 48 blank replicates performed for LoB determination yielded a signal corresponding to galectin-3 concentrations ranging from 0.19 to 1.23 ng/mL. The 46th highest ranked replicate was 0.86 ng/mL and represents the LoB. The mean value of the 16 replicates for LoD determination was 1.09 ng/mL and the SD was 0.163 ng/mL. Thus the LoD is 0.86 ng/mL + [1.645 (0.163)] = 1.13 ng/mL. The galectin-3 concentrations of the four pools utilized for LoQ determination were 0.40, 1.09, 1.32, and 1.45 ng/mL. Respective CVs for these pools were 59.8%, 16.9%, 10.4% and 10.5%. The galectin-3 concentration having ≤ 20% total CV was 0.97 ng/mL and represents the LoQ. Linearity Fig. 1 shows expected and measured galectin-3 values. Both the first order and linear regression and the second order polynomial curves shown in Fig. 1 indicate acceptable linearity for the assays. These data indicate that the galectin-3 assay is linear in from the LoQ at 0.96 to130 ng/mL. High-dose hook Samples up 500 ng/mL in galectin-3 were evaluated for the high-dose hook effect. The diluent, serum and EDTA plasma matrices at 500 ng/mL and 250 ng/mL concentrations had high optical density levels that precluded ELISA measurement. For the diluent, serum and EDTA plasma matrices at the target concentration of 125 ng/mL, measurements were 103, 119 and 94.3 ng/mL, respectively; at the 62.5 ng/mL target the respective measured concentrations were 50.6, 49.0 and 51.5 ng/mL; at the 31.3 ng/mL target, recoveries were 24.4 32.0 and 31.1 ng/mL, respective. No evidence of a high dose hook effect was observed with this assay for galectin-3 concentrations up to 500 ng/mL. Interfering substances Table 2 shows the recovery of endogenous materials, which ranged from 92.7% to 111.9%, mean 101%, and presented no substantial interference. Table 3 displays interference data for packed cell lysates; these studies showed recoveries between 144% and about 999%, mean 472%. This interference may represent intracellular galectin-3 released from white cells during lysis. Specimens for galectin-3 measurement should be examined for signs of hemolysis; hemolyzed samples should be not be used in this assay. Table 4 shows that recoveries for 34 medications tested were 91.7% to 108.1%, mean 100%; none of these substances presented a substantial interference. Tables 5a and 5b display results for three samples tested for HAMA and three for RF interference, respectively. Both HAMA and RF have the potential to show high recovery and users must interpret galectin-3 values with caution in patients who have had prior treatment with antibodies or exposure to mouse antigens (i.e., HAMA) or patients who have positive RF (e.g., rheumatoid arthritis). Cross-reactivity Testing of nine galectin proteins and collagen I and III resulted in mean cross-reactivity ranging from −0.32% to 0.30% with the galectin-3 assay. Thus this galectin-3 assay does not cross-react substantially with any of the structurally similar compounds tested. Measurement matrix Linear regression analysis of the 50 matched serum and EDTA plasma samples ranging from 4.0 to 100 ng/mL yielded the following equation: EDTA-plasma = 0.997⁎Serum + 0.5 ng/mL. Bland-Altman analysis showed a slope mean bias of 0.998, 95% CI 0.97–1.02, which is not significantly different from 1.0. This is equivalent to a mean bias of −0.2% (100⁎0.988–1.000/1.000) with a 95% confidence interval of − 2.7% to + 2.4%. These data demonstrate that either serum or the EDTA plasma samples are appropriate for measurement with this galectin-3 assay. Stability on storage Table 6 displays the stability data for galectin-3. Measurements in either serum or EDTA plasma showed less than ± 25% average difference from the index measurements when stored for 15 days at 22–28 °C and for 15 days at 2–8 °C. Data in Table 6 indicate that galectin-3 can be measured in samples stored at either − 20 °C or − 70 °C for at least 6 months and is stable through at least six freeze–thaw cycles when stored at either temperature for this time. Normal distribution Fig. 2 shows the distribution of galectin-3 values for reference interval (normal) samples included in this study. The 90th percentile (17.6 ng/mL), 95th percentile (20.3 ng/mL), and 97.5th (22.1 ng/mL) percentile of the normal cohort are indicated. Clinical samples The distribution of galectin-3 concentrations in samples from individuals presenting with dyspnea and diagnosed with ADHF is displayed in Fig. 3 . Using the 97.5th percentile of 22.1 ng/mL from the normal cohort as a cutpoint, there was a significant association between galectin-3 and mortality (log-rank p = 0.035). Fig. 4 shows the association between increasing severity of HF, as reflected by NYHA classification, and galectin-3 values. The trend for the association between increasing NYHA and galectin-3 in this unadjusted analysis was significant ( p -value for trend = 0.017). Discussion Examination of key performance characteristics for this galectin-3 assay demonstrates that it is appropriate for clinical measurement using either serum or EDTA plasma specimens. Heparin and citrated plasma matrices were not tested, but there is no fundamental reason that these specimen types could not be used after performing appropriate validation studies. This galectin-3 assay has a wide linear range from the LoQ of 0.96 ng/mL up to130 ng/mL. No high-dose hook effect was evident in samples that had up to 500 ng/mL of galectin-3. Therefore this analytical measurement range and assessment of the hook effect appear to be appropriate. The LoB, LoD and LoQ parameters determined for this assay yielded values below the galectin-3 distribution of both the normal distribution ( Fig. 2 ) and ADHF cohorts ( Fig. 3 ). Along with evidence from the linearity study, these low-end and high-end characteristics translate into the ability of reporting virtually all clinical samples as a continuous “ng/mL” concentration value. This assay is robust based on stability and specificity data for the galectin-3 molecule from cross reactivity and interference studies. Measurements performed after storage for over 2 weeks at either room temperature or refrigerated were not different from fresh samples; also, results of the assay are stable through at least six freeze–thaw cycles. Numerous drugs that would be commonly encountered in HF patients were examined, and no interference was identified. In addition, no interference was observed for substances with similar molecular structures to galectin-3, including galectin-1, galectin-2, galectin-4, galectin-7, galectin-8, galectin-9, galectin-12, collagen I, or collagen III. Imprecision is a central performance variable for clinical assays and was rigorously characterized at all three sites. It is noteworthy that the day-to-day imprecision component was 0 (zero) for a number of the levels tested (see Table 1 ). This result indicates that virtually all of the imprecision variability was due to the run-to-run data component, as analyzed by the ANOVA model used in CLIS EP5-A2. Also, the total CV determined by one of the three sites was significantly lower (better) than the other two. However, it is recommended that users have a conservative expectation in accordance with the performance demonstrated by the other two sites. The demonstrated performance is consistent with imprecision expectations for ELISA type assays. Use of galectin-3 measurements for assessing the risk for adverse outcomes among ADHF patients was suggested in studies using an earlier research use only (RUO) galectin-3 assay [12–14] . However, the RUO test is no longer available so method comparisons with the assay presented here were not possible. Therefore conclusions made previously with the RUO test cannot necessarily be extrapolated to this assay. Employing the 97.5th percentile of the normal distribution (22.1 ng/mL) as the cutpoint with the assay presented here, galectin-3 measurements in a cohort of ADHF patients showed an association with an outcome of death in an unadjusted analysis. Potential confounders were not adjusted for in the analysis shown in Fig. 3 ; although these data indicate promise, appropriate clinical studies are needed to fully elucidate the appropriate cutpoint for galectin-3, as well as its potential impact on management strategies. There was also a significant relationship between severity of heart failure, i.e., NYHA class, and galectin-3 measurements with this assay in an unadjusted analysis. Users should be cognizant of limitations for this galectin-3 assay. Firstly, any clinical information reported here must be considered preliminary. Before clinical use of galectin-3 for prognosis is appropriate the clinical characteristics must be established in properly powered studies that are able to adjust for potentially important variables. Secondly, serum and EDTA plasma were used in this study. Any other specimen matrix must be validated before use. Thirdly, readers must be aware of interferences. Although bilirubin and purified hemoglobin did not interfere with the assay, hemolyzed blood demonstrated a significant interference which may be due to galectin-3 released from white blood cells during lysis. Specimens for galectin-3 measurement should be examined for hemolysis as this may represent the presence of a potential positive interference, and any samples showing hemolysis should be recollected or used for measurement with caution. Analysis of samples with known HAMA also showed a substantial positive interference, as did samples positive for RF. Users must be cautious for patients who have had prior treatment with antibodies or exposure to mouse antigens and who have positive RF values. Determining if novel biomarkers contribute to information that is already available to clinicians can only be properly assessed after the pre-analytical and analytical studies presented here are completed. Characterization of the pre-analytical and analytical performance showed that this galectin-3 assay is reproducible, reliable, robust, and ready for use in fully evaluating clinical risk stratification and patient management applications. Acknowledgment This study was supported by BG Medicine, Inc., Waltham, MA. As stated in the affiliations, Shunguang Wang, Aram Adourian, Carol Adiletto, and Peter Gardiner are employees of BG Medicine. Robert Christenson, Alan Wu and Christopher DeFilippi have acted as consultants to, and received compensation from, BG Medicine. None of the other individuals involved with the manuscript are consultants to, or have received compensation from, BG Medicine. References [1] R. de Boer A.A. Voors P. Muntendam Galectin-3: a novel mediator of heart failure development and progression Eur. J. Heart Fail. 11 2009 811 817 [2] R.Y. Yang G.A. Rabinovich F.T. Liu Galectins: structure, function and therapeutic potential Expert Rev. Mol. Med. 13 2008 e17 e39 [3] G.A. Rabinovich L.G. Baum N. Tinari Galectins and their ligands: amplifiers, silencers or tuners of the inflammatory response? Trends Immunol. 23 2002 313 320 [4] J. Dumic S. Dabelic M. Flögel Galectin-3: an open-ended story Biochim. Biophys. Acta 1760 2006 616 635 [5] N.C. Henderson A.C. Mackinnon S.L. Farnworth F. Poirier F.P. Russo J.P. Iredale C. Haslett K.J. Simpson T. Sethi Galectin-3 regulates myofibroblast activation and hepatic fibrosis Proc. Natl. Acad. Sci. U. S. A. 103 2006 5060 5065 [6] I. Kuwabara F.T. Liu Galectin-3 promotes adhesion of human neutrophils to laminin J. Immunol. 156 1996 3939 3944 [7] Y.H. Liu M. D'Ambrosio T.D. Liao H. Peng N.E. Rhaleb U. Sharma S. André H.J. Gabius O.A. Carretero N -acetyl-seryl-aspartyl-lysyl-proline prevents cardiac remodeling and dysfunction induced by galectin-3, a mammalian adhesion/growth-regulatory lectin Am. J. Physiol. Heart Circ. Physiol. 296 2009 H404 H412 [8] M. Papaspyridonos E. McNeill J.P. de Bono A. Smith K.G. Burnand K.M. Channon D.R. Greaves Galectin-3 is an amplifier of inflammation in atherosclerotic plaque progression through macrophage activation and monocyte chemoattraction Arterioscler. Thromb. Vasc. Biol. 28 2008 433 440 [9] H. Sano D.K. Hsu L. Yu J.R. Apgar I. Kuwabara T. Yamanaka M. Hirashima F.T. Liu Human galectin-3 is a novel chemoattractant for monocytes and macrophages J. Immunol. 165 2000 2156 2164 [10] U. Sharma S. Pokharel T.J. van Brakel J.H. van Berlo J.P.M. Cleutjens B. Schroen S. André H.J.G.M. Crijns H.J. Gabius J. Maessen Y.M. Pinto Galectin-3 marks activated macrophages in failure-prone hypertrophied hearts and contributes to cardiac dysfunction Circulation 110 2004 3121 3128 [11] R.Y. Yang D.K. Hsu F.T. Liu Expression of galectin-3 modulates T-cell growth and apoptosis Proc Natl Acad Sci U S A 93 1996 6737 6742 [12] R.R. van Kimmenade J.L. Januzzi P.T. Ellinor Utility of amino-terminal pro-brain natriuretic peptide, galectin-3, and apelin for the evaluation of patients with acute heart failure J. Am. Coll. Cardiol. 48 2006 1217 1224 [13] H. Milting P. Ellinghaus M. Seewald Plasma biomarkers of myocardial fibrosis and remodeling in terminal heart failure patients supported by mechanical circulatory support devices J. Heart Lung.Transplant. 27 2008 589 596 [14] D. Lok P. van der Meer Bruggink-André de la Porte, et al. Galectin-3, a novel marker of macrophage activity, predicts outcome in patients with stable chronic heart failure J. Am. Coll. Cardiol. 98A 2007 49 [Abstract] [15] D.A. Morrow J.A. de Lemos Benchmarks for the assessment of novel cardiovascular biomarkers Circulation 115 2007 949 952 [16] R.H. Christenson Evidence-based laboratory medicine – a guide for critical evaluation of in vitro laboratory testing Ann. Clin. Biochem. 44 2007 111 130 [17] B.J. Cherayil S. Chaitovitz C. Wong S. Pillai Molecular cloning of a human macrophage lectin specific for galactose Proc. Natl. Adac. Sci. U. S. A. 87 1990 7324 7328 [18] P. Muntendam A. Adourian R. Christenson Reference interval for plasma galectin-3 in healthy subjects age 55 years and older Clin. Chem. 55 2009 A73 [19] P. Muntendam P. Gardiner The BioImage Study: novel approaches to risk assessment in the primary prevention of atherosclerotic cardiovascular disease Atherosclerosis 10 2009 P799 [20] K.B. Shah W.J. Kop R.H. Christenson D.B. Diercks D. Kuo S. Henderson K. Hanson M.R. Mehra C.R. deFilippi Lack of diagnostic and prognostic utility of circulating plasma myeloperoxidase concentrations in patients presenting with dyspnea Clin. Chem. 55 2009 59 67