Clinical Chemistry

Ishibashi,, Midori

doi: 10.1373/clinchem.2005.061325pmid: 16391325

The landscape of prostate cancer has changed since the appearance of the first prostate-specific antigen (PSA) assay. PSA testing has gained rapid recognition since M. Kuriyama et al. (1) of Roswell Park first reported clinical studies using an enzyme immunoassay with anti-PSA rabbit antibody. Mikolajczyk and coworkers (2)(3) reported the presence of several free PSA isoforms and the potential application of free PSA isoforms as serum markers. Today, more than 30 types of total PSA assay reagent sets and ∼10 types of free PSA and PSA–α1-antichymotrypsin (ACT) assays, based on various principles, are available. Key elements in PSA measurement are interchangeability of assays and stability of serum samples before tests (4)(5). Efforts to standardize PSA assays were initiated in 1992 at the First Stanford Conference, organized by T. Stamey. At the Second Stanford Conference on International Standardization of Prostate-Specific Antigen (1994), Stamey et al. (6) proposed a primary calibrator consisting of 90% purified PSA-ACT and 10% free PSA (90:10) on a molar basis. Subsequently, the Clinical and Laboratory Standards Institute (CLSI; formerly NCCLS) issued a document (7) recommending a set of 3 distinct materials containing 100% free PSA, 100% PSA-ACT, and 90% PSA-ACT:10% free PSA. This document led to further activity in the harmonization of PSA assays. In 1999, Robert M. Nakamura, a member of the International Consultation Committee on Prostate Cancer, made recommendations, with his colleagues (8), on standardization and quality assessment of PSA immunoassays. The recommendations included the following: (a) free PSA should be the primary standard for free PSA and total PSA immunoassays; (b) assays should exhibit an equimolar response to free PSA and PSA-ACT complexes; (c) assay antibodies to free PSA should not cross-react with human kallikrein 2 (hK2); (d) reference values for free PSA/total PSA need to be established for each assay combination; and (e) clinical laboratories should specify the name of the assay on their reports. Kuriyama et al. (9) studied 9 commercially available assays for total PSA and described equations to calculate the expected Tandem-R PSA (Beckman Coulter) results from the PSA concentrations obtained with the other methods studied. That report has been widely used in clinical studies. At the time of the report by Kuriyama et al., however, the concept of equimolarity of assays was not well established, and the risk of converting results among assays, without knowing the reactivities of the antibodies in the different reagent sets, was not assessed. The heterogeneity of PSA forms in serum was later described in detail (10), and the effect of that heterogeneity on results of immunoassays became fully appreciated. In 1997, the Committee of Investigation on PSA Assay (sponsored by the Japanese Urological Association and the Japanese Society of Laboratory Medicine) carried out a survey of the variability among assays. Patient serum, purified 100% free PSA, and purified 100% PSA-ACT were prepared for survey samples. We studied 28 total PSA assays from 22 manufacturers and analyzed the reactivity ratios of free PSA to PSA-ACT in those assays by dilution tests using purified 100% free PSA and purified 100% PSA-ACT. The study revealed that the differences among measurement schemes were related to differences in the reactivities of antibodies for free PSA and PSA-ACT caused by the variety of PSA molecular forms. In 2000, we performed another survey focusing on the clinically important total PSA range up to 20 μg/L (11). In that survey, 26 assays from 18 companies were investigated. During the 3 years between our surveys, the reactivities of most assays for total PSA had been improved. In the initial survey, only 34% of the assays showed an equimolar response to free PSA and PSA-ACT. The minimum and maximum measured values were 4.1 and 32.3 μg/L for the serum sample containing >40% free PSA (11). Large disparities persist among assays, however. We attribute this variability, in part, to matrix differences between serum samples and the WHO buffer-based reference material used for assignment of values to the assay calibrators. Anti-PSA antibodies show different reactivities and affinities for the various forms of free PSA and PSA-ACT in buffer- or serum-based samples. The goal of PSA standardization efforts is to improve harmonization of results for patient serum samples, but the WHO primary standard materials use buffer with added bovine serum albumin, and assay calibrators use various matrices. Ideally, the total PSA assay should have an equimolar response for free PSA and PSA-ACT. We examined the matrix effect among the buffer- and serum-based samples by measuring purified free PSA and PSA-ACT. The differences in the measured concentrations for buffer-based vs serum-based samples reported by the AxSYM (Abbott Diagnostics) and ADVIA Centaur (Bayer) were <4%, in contrast to those reported by the Elecsys (Roche Diagnostics), Access (Beckman Coulter), and Immulite (Diagnostic Products Corp.), which were 9%, 7%, and 18%, respectively. In all cases, PSA-ACT showed greater reactivity than did free PSA. The standardization of PSA assays is limited because only primary reference materials (WHO reference materials) have been prepared and because PSA does not have a total reference measurement system including reference materials and a reference measurement procedure. To further promote standardization, a reference measurement system should be established to transmit the accuracy of the buffer-based primary standard material to patient serum, which is the final goal. In a comprehensive study in this issue of Clinical Chemistry, Stephan et al. (12) show that interchangeability among methods is a problem for clinical samples. In that study, they took great care when storing samples and performing the assay, and the reliability of the measured values was high. The percentage free PSA (%fPSA) is related to both total PSA and free PSA and needs to be evaluated carefully. As Stephan et al. indicate, for patients in whom the %fPSA determines whether a prostate biopsy is performed, the result might depend on the assay used. Thus it seems important that total and free PSA be measured by compatible methods. Some important requirements for studies that compare several assays include that the comparison method chosen gives equimolar responses for free PSA and PSA-ACT and that it has minimal susceptibility to matrix effects. We have confirmed that the Access (Hybritech) total PSA assay, which was used as the comparison method in the study by Stephan et al. (12), does not have an adequate equimolar response to free PSA and PSA-ACT in spite of using the same antibodies as the Tandem-R PSA assay. One difficulty in the study of Stephan et al. (12) may be the calculation of free PSA from the PSA-ACT and total PSA values obtained by the ADVIA Centaur assay. Although most molecules of complexed PSA are combined with ACT, some complexed PSA that is measurable by immunoassay is combined with other protease inhibitors, such as α1-proteinase inhibitor, inter-α-trypsin inhibitor, and others (13). Thus, the concentrations of total PSA, free PSA, and PSA-ACT measured by assays of Eiken Chemical Company do not necessarily satisfy the equation fPSA + cPSA = tPSA in our study (where fPSA is free PSA, cPSA is complexed PSA, and tPSA is total PSA). In conclusion, at present, samples should be preserved until measurement is carefully performed, and the assay name and reference interval should be listed on reports. 1 Kuriyama M, Wang MC, Papsidero LD, Killian CS, Shimano T, Valenzuela L, et al. Quantitation of prostate-specific antigen in serum by a sensitive enzyme immunoassay. Cancer Res 1980 ; 40 : 4658 -4662. 2 Mikolajczyk SD, Marks L, Partin AW, Rittenhouse HG. Free prostate-specific antigen in serum is becoming more complex. Urology 2002 ; 59 : 797 -802. 3 Mikolajczyk SD, Catalona WJ, Evans CL, Linton HJ, Millar LS, Marker KM, et al. Proenzyme forms of prostate-specific antigen in serum improve the detection of prostate cancer. Clin Chem 2004 ; 50 : 1017 -1025. 4 Jung K, von Klinggräff P, Brux B, Sinha P, Schnorr D, Loening SA. Preanalytical determinations of total and free prostate-specific antigen and their ratio: blood collection and storage condition. Clin Chem 1998 ; 44 : 685 -688. 5 Woodrum D, York L. Two-year stability of free and total PSA in frozen serum samples. Urology 1998 ; 52 : 247 -251. 6 Stamey TA, Chen Z, Prestigiacomo AF. Reference material for PSA: the IFCC Standardization Study. Clin Biochem 1998 ; 31 : 475 -481. 7 . Clinical and Laboratory Standards Institute. Primary reference preparations used to standardize calibration of immunochemical assays for serum prostate specific antigen (PSA); approved guideline. CLSI document I/LA 10-A 1997 CLSI Wayne, PA. . 8 Nakamura RM, Abrahamsson PA, Chopin D, et al. Progress in standardization and quality assessment of free PSA (prostate specific antigen), total PSA, and complexes PSA immunoassays. Murphy G Khoury S Partin A Denis L eds. Prostate cancer: proceedings of the 2nd International Consultation on Prostate Cancer - June 27–29, 1999 1999 : 205 -217 WHO Paris. Geneva, Switzerland. . 9 Kuriyama M, Akimoto S, Akaza H, Arai Y, Usami M, Imai K, et al. Comparison of various assay system for prostate-specific antigen standardization. Jpn J Clin Oncol 1992 ; 22 : 393 -399. 10 Zhou AM, Tewari PC, Bluestein BI, Caldwell GW, Larsen FL. Multiple forms of prostate-specific antigen in serum: differences in immunorecognition by monoclonal and polyclonal assays. Clin Chem 1993 ; 39 : 2483 -2491. 11 Kano S, Ishibashi M, Itoh Y. The present status of standardization in serum total PSA measurement: a result of “Survey 2000” conducted by PSA ad hoc committee of Japanese Urological Association. Rinsho Byori 2001 ; 49 : 967 -973. 12 Stephan C, Klaas M, Müller C, Schnorr D, Loening SA, Jung K. Interchangeability of measurements of total and free prostate-specific antigen in serum with 5 frequently used assay combinations: an update. Clin Chem 2006 ; 52 : 59 -64. 13 Stenman UH, Leinonen J, Alfthan H, Rannikko S, Tuhkanen K, Alfthan O. A complex between prostate-specific antigen and α1-antichymotrypsin is the major form of prostate-specific antigen in serum of patients with prostatic cancer: assay of the complex improves clinical sensitivity for cancer. Cancer Res 1991 ; 51 : 222 -226. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Burgers, Jako, S

doi: 10.1373/clinchem.2005.059345pmid: 16391326

Evidence-based medicine (EBM) challenges clinicians and laboratory professionals to make rational decisions in healthcare. Clinical practice guidelines embody the principles of EBM in making options and choices explicit, considering the scientific evidence and resources available (1). The evidence is often limited or controversial, however, and resources differ among countries and regions. Therefore, the translation of research evidence into recommendations for clinical practice is not straightforward. Judgment beyond the evidence is necessary, taking into consideration the balance between benefits and harms and risks, patients’ views and preferences, and potential organizational and financial barriers (2). Addressing these issues requires careful discussions within a working group that includes representatives from all relevant disciplines. In contrast to the development of systematic reviews, guidelines cannot be produced in an academic “ivory tower” by a few experts. When the evidence is not strong or not fully applicable to the patient population targeted in the guideline, the working group may depart from the evidence. In laboratory medicine, high-level diagnostic evidence is particularly scarce, and the link between diagnostic tests and better patient outcomes is often unknown (3). It thus comes as no surprise that the composition of the working group and group dynamics can influence the outcome of discussions and lead to variation among the recommendations of different groups (4)(5). High-quality guidelines are based on evidence as well as a broad consensus of opinions, which facilitates the acceptance and effective use of the guideline in the target group (6). To ensure high quality, guidelines should be developed within a structured and coordinated program according to the principles of evidence-based guideline development (7). Explicit reporting of the methods and procedures followed will improve the guideline quality (8). On the other hand, lack of information about the methods and procedures does not automatically mean that the recommendations in the guideline are not valid. The AGREE (Appraisal of Guidelines Research and Evaluation) Instrument, first published in 2001 (9), is a tool for assessing the quality of clinical guidelines according to 23 criteria, grouped into 6 domains (scope and purpose, stakeholder involvement, methodology, clarity and presentation, applicability, and editorial independence). The instrument is available in more than 10 languages and is currently used in many countries around the world. The AGREE criteria mainly concern the methods used for developing the guideline and the quality of the reporting. The clinical content of the recommendations and the quality of the supporting evidence are not addressed, however (10), a deficit common in appraisal tools for clinical guidelines (11). One might assume that a guideline with a high quality score would contain valid recommendations that are not in conflict with the best available evidence. If this were true, quality assessment of the evidence behind the recommendations would not be needed, and much appraiser time would be saved. In this issue of Clinical Chemistry, Watine et al. (12) report the results of their test of the correlation between guideline quality and clinical validity of recommendations for 11 guidelines on a specific topic (non-small cell lung cancer) in laboratory medicine. Guideline quality was measured with the AGREE Instrument summarized in a global score. The validity of the recommendations was assessed by comparison with the results of a systematic review performed by the authors themselves. This study is the first to test the relationship between the AGREE quality scores and the clinical content of guidelines, a process that could be considered as a next step in validating the AGREE Instrument. The results revealed that there was no relationship: “good” as well as “not so good” guidelines contained “good” recommendations, corresponding with the evidence of the systematic review, and even “good” guidelines included “not so good” recommendations. These findings confirm the variability of the translation of evidence into practice recommendations by different guideline groups. Such variability, however, is not necessarily undesirable. Limited availability of resources could explain the fact that some guidelines do not recommend the whole series of laboratory tests determined to be useful according to the evidence from the systematic review. On the other hand, a lack of budget restrictions might account in part for the recommended use of tumor markers, whereas this was not recommended in the systematic review. Ideally, these considerations should have been made explicit in the guideline. Guideline documents offer the opportunity to include balanced thoughts behind the recommendations, to identify gaps in knowledge, and to include different options when the evidence is unclear or lacking. Such information is a better alternative to nonspecific recommendations that could potentially harm patients (13). One important limitation of the study by Watine et al. (12) is that they used their own systematic review as a gold standard. This would be appropriate if sufficient evidence were available, including data from high-quality trials or prospective cohorts in the case of specific diagnostic research questions. Most of their conclusions, however, were not based on the highest level of evidence but on validation by experts. Thus, their review itself could be considered a guideline based on evidence and consensus, inevitably including some subjective judgments. Moreover, guidelines are particularly needed in areas of uncertainty. If the evidence is clear, guidelines would not have an added value compared with a review or textbook. Some variation in guidelines is acceptable because of context-specific decisions. One should therefore be cautious in promoting international guidelines (14). Nevertheless, international standards similar to those for randomized clinical trials and diagnostic studies could be developed for guideline quality and reporting(15)(16). The Guidelines International Network, founded in 2002, provides good opportunities to exchange methods and new approaches to guideline development(17). Further validation of the AGREE Instrument would be the next step to achieve more international consensus about guideline quality and methodology. 1 Grimshaw J, Russel I. Achieving health gain through clinical guidelines. I. Developing scientifically valid guidelines. Qual Health Care 1193 ; 2 : 243 -248. 2 Scottish Intercollegiate Guidelines Network. SIGN 50: a guideline developers’ handbook (February 2001; last updated May 2004). http://www.sign.ac.uk/guidelines/fulltext/50/index.html (accessed August 2005).. 3 Oosterhuis WP, Bruns DE, Watine J, Sandberg S, Horvath AR. Evidence-based guidelines in laboratory medicine: principles and methods. Clin Chem 2004 ; 50 : 806 -818. 4 Pagliari C, Grimshaw J. Impact of group structure and multidisciplinary evidence-based guideline development: observed study. J Eval Clin Pract 2002 ; 8 : 145 -153. 5 Raine R, Sanderson C, Hutchings A, Carter S, Larkin K, Black N. An experimental study of determinants of group judgments in clinical guideline development. Lancet 2004 ; 364 : 429 -437. 6 Grol R, Dalhuijsen J, Thomas S, Veld C, Rutten G, Mokkink H. Attributes of clinical guidelines that influence use of guidelines in general practice: observational study. BMJ 1998 ; 317 : 858 -861. 7 Burgers JS, Cluzeau FA, Hanna SE, Hunt C, Grol R. Characteristics of high quality guidelines: evaluation of 86 clinical guidelines developed in ten European countries and Canada. Int J Technol Assess Health Care 2003 ; 19 : 148 -157. 8 Fervers B, Burgers JS, Haugh MC, Brouwers M, Browman G, Cluzeau FA, et al. Predictors of high quality clinical practice guidelines: examples in oncology. Int J Qual Health Care 2005 ; 17 : 123 -132. 9 . The AGREE Collaboration. Development and validation of an international appraisal instrument for assessing the quality of clinical practice guidelines: the AGREE project. Qual Saf Health Care 2003 ; 12 : 18 -23. 10 The AGREE Collaboration. AGREE Instrument training manual, 2003. http://www.agreecollaboration.org (accessed August 2005).. 11 Vlayen J, Aertgeerts B, Hannes K, Sermeus W, Ramaekers D. A systematic review of appraisal tools for clinical practice guidelines: multiple similarities and one common deficit. Int J Qual Health Care 2005 ; 17 : 235 -242. 12 Watine J, Friedberg B, Charet JC, Nagy E, Onody R, Oosterhuis W, et al. Conflict between guideline methodologic quality and recommendation validity: a potential problem for practitioners. Clin Chem 2006 ; 52 : 65 -72. 13 Shekelle PG, Kravitz RL, Beart J, Marger M, Wang M, Lee M. Are nonspecific practice guidelines potentially harmful? A randomized comparison of the effect of nonspecific versus specific guidelines on physician decision making. Health Serv Res 2000 ; 34 : 1429 -1448. 14 Eisenberg JM. Globalize the evidence, localize the decision: evidence-based medicine and international diversity. Health Aff (Millwood) 2002 ; 21 : 166 -168. 15 Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet 2001 ; 357 : 1191 -1194. 16 Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Clin Chem 2003 ; 49 : 1 -6. 17 Ollenschläger G, Marshall C, Qureshi S, Rosenbrand K, Slutsky J, Burgers J, et al. Improving the quality of health care: using international collaboration to inform guideline programmes by founding the Guidelines International Network (G-I-N). Qual Saf Health Care 2004 ; 13 : 455 -460. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Myers, Gary, L;Miller, W, Greg;Coresh,, Josef;Fleming,, James;Greenberg,, Neil;Greene,, Tom;Hostetter,, Thomas;Levey, Andrew, S;Panteghini,, Mauro;Welch,, Michael;Eckfeldt, John, H

doi: 10.1373/clinchem.2005.0525144

Azzazy, Hassan, ME;Pelsers, Maurice, MAL;Christenson, Robert, H

doi: 10.1373/clinchem.2005.056143pmid: 16269514

Abstract Background: A biomarker that reliably detects myocardial ischemia in the absence of necrosis would be useful for initial identification of unstable angina patients and for differentiating patients with chest pain of an etiology other than coronary ischemia, and could provide clinical utility complementary to that of cardiac troponins, the established markers of necrosis. Unbound free fatty acids (FFAu) and their intracellular binding protein, heart-type fatty acid–binding protein (H-FABP), have been suggested to have clinical utility as indicators of cardiac ischemia and necrosis, respectively. Methods: We examined results of clinical assessments of FFAu and H-FABP as biomarkers of cardiac ischemia and necrosis. Data published on FFAu and H-FABP over the past 30 years were used as the basis for this review. Results: Although little clinical work has been done on FFAu since the initial reports, recent studies documented an association between increased serum FFAs and ventricular dysrhythmias and death in patients with acute myocardial infarction (AMI). Recent data suggest that serum FFAu concentrations increase well before markers of cardiac necrosis and are sensitive indicators of ischemia in AMI. H-FABP is abundant in cardiac muscle and is presumed to be involved in myocardial lipid homeostasis. Similar to myoglobin, plasma H-FABP increases within 3 h after AMI and returns to reference values within 12–24 h. Conclusions: FFAu may have a potential role in identifying patients with cardiac ischemia. H-FABP is useful for detecting cardiac injury in acute coronary syndromes and predicting recurrent cardiac events in acute coronary syndromes and in congestive heart failure patients. Assays are available for both markers that could facilitate further clinical investigations to assess their possible roles as markers of cardiac ischemia and/or necrosis. Myocardial ischemia, a major cause of myocardial injury and necrosis, is initiated whenever the coronary arterial flow cannot supply sufficient oxygen to the myocardium. Within seconds of myocardial ischemia, several changes occur within myocytes, such as termination of aerobic metabolism, onset of anaerobic glycolysis, potassium ion leakage, and cessation of contraction to reduce the energy demands, causing wall motion abnormalities that can be detected by echocardiography. Within minutes, other changes follow, including leakage of metabolites, a decrease in pH, and increased intracellular calcium concentrations and osmotic load. Early ultrastructural changes include swollen mitochondria, edema, and cytoplasmic blebbing. The general cause of irreversible changes, within hours of ischemia, is progressive and prolonged ATP depletion. The hallmarks of this stage, which represents the “point of no return”, are sarcolemmal disruption and leakage of cardiac macromolecules such as cardiac troponins I and T (cTnI1 and cTnT) and creatine kinase-MB (CK-MB). The pathophysiologic changes and metabolic progression of ischemia to necrosis have been described in detail in a recent text (1). The release of cardiac biomarkers is influenced by a variety of factors: Cytosolic enzymes: An increase in intracellular calcium activates a variety of enzymes, including phospholipases and the protease calpase. Calpase contributes to the early degradation, dissociation, and release of myofibrillar proteins (such as cTnI and cTnT) after myocardial damage. pH-dependent dissociation of structural proteins could also affect release of such markers. On the other hand, lysosomes are stable within the first 3–4 h after onset of ischemia and do not affect the breakdown of subcellular structures. Subcellular location: Soluble cytosolic molecules, such as fatty acid–binding proteins (FABPs), are released more rapidly than structurally bound molecules. Molecular mass: Within a particular intracellular localization, smaller molecules such as myoglobin and FABPs may enter the vascular system to a large extent directly via the microvascular endothelium. Plasma clearance: Smaller molecules such as FABPs (also myoglobin) pass through the glomerular membranes and are reabsorbed and metabolized in tubular epithelial cells (2). Falsely increased plasma concentrations (caused by impaired clearance attributable to renal failure) or falsely decreased concentrations (in patients with hypermetabolic states) for both markers may therefore be observed. Concentration gradients: Concentration gradients between cardiomyocytes and interstitial spaces as well as local blood and lymphatic flow may also affect the appearance of markers in the general circulation. Whether the release of biomarkers from the injured myocardium indicates irreversible damage and cardiac necrosis remains an issue of debate. The classic hypothesis suggests that release of biomarkers from the cardiomyocyte is possible only from irreversibly injured myocytes and is based on the hypothesis that plasma membranes are physiologically impermeable to macromolecules (3). An alternative hypothesis proposes that release is a metabolically controlled property of cell membranes and that small extracellular increases in cardiac biomarkers may be caused by reversible disturbance of cell metabolism (4). Recent evidence suggests that under moderate ischemic stress, myocardial tissue can release small amounts of macromolecules from the cytosol by mechanical mechanisms other than persistent membrane perforation (4). The prevention of membrane leaks is an energy-dependent process in which myocardial plasma membranes become permeable to intracellular macromolecules under conditions of energy shortage. However, the appearance of mitochondrial enzymes and prolonged increases in cardiac proteins in serum are generally accepted as indicators of myocardial necrosis. Testing the specificity of novel biomarkers of ischemia is challenged by the absence of a “gold standard” for myocardial ischemia. Comparison with troponin concentrations will not be valid because the ischemia marker would be expected to increase in unstable angina patients, who should not have any detectable increases in cardiac troponins. Achieving clinical acceptance of the proposed biomarker as the new gold standard will require extensive laboratory and translational research. In this review, we discuss the physiology and pathophysiology of unbound free fatty acids (FFAu) and heart-type FABP (H-FABP) and their proposed clinical applications as new biomarkers of cardiac ischemia and necrosis, respectively. Unbound FFAs physiology and pathophysiology FFAs play several essential roles in physiologic homeostasis. Under aerobic conditions, nonesterified long-chain FFAs represent the primary metabolic sources for the myocardium, accounting for almost two-thirds of the ATP generated, with glucose metabolism generating the remaining one-third of myocardial oxygen demand (5). Plasma long-chain fatty acids are either esterified to glycerol or nonesterified (or FFAs), most of which are bound to albumin. The mechanism for uptake of FFAs into myocytes remains unclear but involves passive diffusion and/or active carrier-mediated transport (6). In the cytoplasm, long-chain FFAs are bound to FABP, which presumably facilitates their transport to the outer mitochondrial membrane where they become esterified/activated by long-chain acyl-CoA synthetase (Fig. 11 ). Once activated, acyl-CoA esters are directed mainly to β-oxidation, but some may be stored as triglycerides or converted into membrane phospholipids. During hypoxia and ischemia, nonesterified fatty acids/FFAs have damaging effects on heart tissue and have been associated with an increased incidence of ventricular dysrhythmias and death in patients with acute myocardial infarction (AMI) (7)(8). Proposed mechanisms for the damaging effects of FFAs during ischemia include accumulation of toxic intermediates of fatty acid metabolism, suppression of glucose utilization, and uncoupling of oxidative metabolism from electron transfer (5). Inhibitors of FFA metabolism have been shown experimentally to reduce the infarct size and decrease the postischemic cardiac dysfunction in animal models (9). Shown in Table 11 is a comparison of FFAu and H-FABP concentrations in circulation and in myocytes under physiologic and ischemic conditions. FFAu assays Although most of the FFAs in serum are bound to albumin, a small amount is unbound; this is frequently referred to as the “free” fraction. Serum FFAu concentrations are determined from the ratio of total serum FFAs to total serum albumin (10). A method for estimating serum FFA concentrations is based on the breakdown of pyrophosphate, which is formed by thioesterification of FFAs with ATP and CoA in the presence of acyl-CoA synthetase, to inorganic phosphate, which is measured by reaction with molybdate (11). A recently developed method for measurement of serum FFAu uses a fluorescent probe of FFAu, termed acrylodated intestinal fatty acid–binding protein (ADIFAB), which is prepared by derivatizing a recombinant intestinal FABP with the fluorescent molecule acrylodan (12). Binding of a single FFAu molecule to ADIFAB, which does not interact with other serum molecules, displaces the fluorescent tag, producing a spectral shift from 432 nm to 505 nm that can be measured with a fluorometer. Human serum contains a mixture of 6 FFAs: palmitate (25%), stearate (10%), oleate (38%), linoleate (22%), arachidonate (3%), and linolenate (2%) (13). Richieri and Kleinfeld (13) reported ADIFAB dissociation constants, determined at 37 °C and at concentrations below the critical micelle concentrations, of 0.31, 0.08, 0.28, 0.97, 1.63, and 2.50 μmol/L for palmitate, stearate, oleate, linoleate, arachidonate, and linolenate, respectively. Variations in dissociation constants are highly correlated with the solubility of the specific fatty acid in water, suggesting that all of these fatty acids bind to intestine FABP with a similar conformation. A second-generation version of the assay uses a handheld reader and 15 μL of plasma and provides turnaround times <1 min (14). The assay shows improved low-end sensitivity and is not affected by hemoglobin. The CV for duplicate measurements is 7%. The FFAu upper reference limit (URL), determined at the 97.5th percentile of value distribution, is 2.7 nmol/L [mean (SD), 1.5 (0.6) nmol/L; range, 0.6–4.5 nmol/L] (14). Reports have suggested that heparin may cause FFA increases in vivo and perhaps in vitro as well. This is because heparin is known to stimulate the activity of lipoprotein lipase, which releases FFAs from triglycerides associated with blood lipoproteins. Blood collected into heparin-containing flasks or tubes or from patients receiving therapeutic heparin may therefore not be suitable for the FFAu test. There has been controversy regarding this issue, however. Thus, assays considered for clinical use must be evaluated in appropriate studies addressing the potential of heparin interference. clinical applications The mean (SD) serum FFAu concentration, measured with the ADIFAB assay, in 283 samples from healthy donors was 7.5 (2.5) nmol/L (13), and the distribution of FFAu values was independent of donor age and sex. Mean FFAu values increased significantly (by 1.5 nmol/L; P <0.001) after overnight fasting. The clinical uses of FFAu concentrations are summarized in Table 22 . Using the fluorescent probe ADIFAB, Kleinfeld et al. (15) measured serum FFAu concentrations in 22 patients 5 min before and 30 min after percutaneous transluminal coronary angioplasty (PCTA). Post-PCTA concentrations were higher than baseline values in all patients, with the mean FFAu concentration increasing 5-fold compared with the mean value [7.5 (2.5) nmol/L] observed in healthy patients. Ischemic ST changes monitored by electrocardiography (ECG) were observed in only 50% of patients. In addition, FFAu concentrations were significantly higher in the ECG-positive group than the ECG-negative group. An increase in FFAu concentrations was suggested as an early marker of ischemia caused by PCTA (15). In a different study, 9 MI patients had increased FFAu concentrations, whereas only 2 of the 9 had increased cTnI (14). In addition, 93% and 30% of other chest pain patients had increased FFAu and cTnI concentrations, respectively. FFAu concentrations were increased in every instance that cTnI was increased. In addition, there was a positive correlation between peak FFAu and cTnI concentration. At presentation, all of the MI patients had increased FFAu, whereas only 22% had increased cTnI. Some of these patients had additional diagnoses (such as cocaine abuse, sepsis, and cardiac contusion) that can cause myocardial ischemia and injury. The authors therefore proposed that FFAu concentrations may increase in the presence of acute myocardial injury independent of plaque rupture (14). Circulating FFA concentrations have also been suggested as putative for ventricular arrhythmias and sudden death after MI (16). In the Paris Prospective Study I, plasma FFA concentrations were measured in a cohort of 5250 middle-aged men free of known ischemic cardiac disease (17). After 22 years of follow-up, increased FFA concentrations were found to be an independent risk factor for sudden death (relative risk, 1.70; 95% confidence interval, 1.21–2.13) but not for fatal MI. Pirro et al. (18) examined the relationship between circulating FFA concentrations and risk of ischemic heart disease (IHD) in 2130 men with insulin resistance syndrome who were without IHD at enrollment. During a 5-year follow-up, 114 of these individuals developed IHD. After adjustment for nonlipid risk factors, increased circulating FFA concentrations conferred a 2-fold increase in the risk of IHD (odds ratio, 2.1; P = 0.05) compared with lower plasma FFA concentrations. However, after adjustment for triglyceride concentrations, HDL-cholesterol, small, dense LDL, apolipoprotein B, and fasting plasma insulin, the relationship between plasma FFA concentrations and IHD did not achieve statistical significance. In another study, FFAu concentrations were measured in serum samples from 458 AMI patients (75 females and 383 males), enrolled in the TIMI II trial (19), who were treated with tissue plasminogen activator (tPA) (20). FFAu concentrations were measured with the ADIFAB2 assay in blood samples drawn on admission and 50 min, 5 h, and 8 h after initiation of tPA treatment. Relative to the control population, results of this study indicated an ∼4-fold increase in serum FFAu concentrations at enrollment, a further 2-fold increase after tPA administration, and then a gradual decrease within 5 h of tPA administration. At cutoff a 5 nmol/L, the predicted sensitivity was 98% based on the results for admission and 50-min samples. The specificity, based on comparison with healthy individuals and patients with noncardiovascular disease, was 93%. Although interesting, these data must be considered exploratory for indicating the potential clinical performance of FFAu. This is because the TIMI II trial included only well-characterized AMI patients, who were compared with patients with known noncardiovascular disease and with healthy individuals, which will not be the way the test will be used in practice. FFAs also correlated well with mortality at 30 days in the TIMI II cohort (20). summary Current data, although limited, suggest that monitoring of FFAu concentrations in patients presenting with ischemic symptoms may provide an early indication of cardiac ischemia. Cohort trials that enroll a broad spectrum of suspected myocardial ischemia patients need to be performed to fully evaluate the true potential of this biomarker. ROC curves need to be plotted for relevant populations. H-FABP physiology FABPs bind long-chain fatty acids reversibly and noncovalently. FABPs are relatively small (15 kDa) intracellular proteins that are abundantly produced in tissues having active fatty acid metabolism, including the heart, liver, and intestine (21). FABPs each contain 126–137 amino acids, and their tertiary structure resembles a clam shell in which the ligand is bound between the 2 halves of the clam by interaction with specific amino acid residues within the binding pocket, the so-called β-barrel (22). Currently, 9 distinct FABP types have been identified, with each type showing a characteristic pattern of tissue distribution and a stable intracellular half-life of 2–3 days (21). H-FABP was first shown to be released from injured myocardium in 1988, after which several studies investigated its application as a biochemical marker of myocardial injury. The H-FABP isoform is produced not only in cardiomyocytes but also, to a lesser extent, in skeletal muscle (23), distal tubular cells of the kidney (24), specific parts of the brain (25), lactating mammary glands, and placenta (23). Human H-FABP contains 132 amino acid residues and is an acidic protein (pI 5) (26). The primary biological function of FABPs is to facilitate intracellular translocation of long-chain fatty acids (see Fig. 11 ), which is usually hampered by the very low solubility of these compounds in aqueous solutions (21). H-FABP can therefore be regarded as the cytoplasmic counterpart of plasma albumin. H-FABP knock-out mice have a markedly lower (∼50%) fatty acid uptake rate and oxidation (24). Other functions of H-FABP include participation in signal transduction pathways, such as regulation of gene expression by mediating fatty acid signal translocation to peroxisome proliferator–activated receptors (27), and putative protection of cardiac myocytes against the detergent-like effects of locally high concentrations of long-chain fatty acids, particularly during ischemia (21)(28). The cellular production of FABPs is regulated primarily at the transcriptional level. In experimental animals, FABP was increased by endurance training (29) and diabetes (30). immunologic assays for h-fabp The characteristics of several assays for human H-FABP are shown in Table 33 . H-FABP is a stable protein; both plasma samples and recombinant protein solutions can be subjected to at least 8 freeze–thaw cycles without loss of immunoreactivity (31). Samples can be stored for at least 2 years at −80 °C (32). Recombinant H-FABP is immunochemically equivalent to the tissue-derived protein and generally serves as a calibration material in immunoassays (26)(31). enzyme-linked immunoassays Tanaka et al. (2) developed a competitive enzyme immunoassay for H-FABP in plasma and urine samples. However, the assay required a long assay time and was not suitable for clinical application. Wodzig et al. (31) developed a one-step ELISA with a total performance time of 45 min; this assay has sensitivity and specificity comparable to the two-step ELISA developed by Okhura et al. (33). Both assays are commercially available and are used in clinical research. Automated immunoassays. Several automated assays have been developed, including an enzyme immunoassay, an automated sandwich immunoassay (34), and a fully automated microparticle-enhanced immunoassay (COBAS® MIRA Plus analyzer; Hoffmann-La Roche). These assays use carboxylated latex particles coated with 3 monoclonal anti-human H-FABP antibodies (35) and are not commercially available at present. Very recently, a new concept of precipitation ellipsometry has been reported (36), with a rapid assay time of 10 min, but this assay is still in prototype form. Lateral-flow assays. Qualitative H-FABP lateral-flow assays have also been developed, and 2 whole-blood tests are commercially available (37)(38)(39). These qualitative tests have a 15-min analysis time, and the cutoff for normal vs high H-FABP concentrations is 6 μg/L. Drawbacks of these lateral-flow assays include substantial interobserver differences in interpretation of color development and the inability to differentiate between moderate and high H-FABP concentrations. Immunosensors. Siegmann-Thoss et al. (40) developed a sandwich immunoassay that uses glucose oxidase–labeled detection antibodies. In the current format, this system requires sample predilution and is susceptible to plasma matrix effects. Real-time optical immunosensors have been developed (41), but these require large sample volumes, have high detection limits, and are susceptible to interference from plasma lipids. As part of the EUROCARDI project, Schreiber et al. (42) and Key et al. (43) developed the first amperometric immunosensor for plasma H-FABP measurement. This rapid 20-min semiautomated analyzer gave results comparable to those obtained with the ELISA developed by Wodzig et al. (31), but did not exhibit sufficient sensitivity in the low-normal concentration range of 5–15 μg/L. In 2002, O’Reagan et al. (44) described an H-FABP immunosensor that used whole blood and had an assay time of 50 min. Recently, a prototype of an online immunodisplacement sensor has been developed for continuous monitoring of H-FABP (45). In this immunosensor, blood samples are obtained via a microdialysis probe or via continuous ultrafiltration of venous blood. clinical interpretation of plasma h-fabp concentrations Under nonpathologic conditions, H-FABP is not present in plasma or interstitial fluid, and cytoplasmic concentrations of this protein are 2 × 105-fold higher than its vascular concentrations (46). The plasma H-FABP concentration measured in apparently healthy individuals (<5 μg/L) is suggested to result from continuous release from damaged skeletal muscle cells. The biological variation attributable to age, sex, and circadian rhythm significantly influences H-FABP reference values (47). Probably because of their larger muscle mass, men have higher plasma H-FABP concentrations than women. Because H-FABP is eliminated from the circulation predominantly by renal clearance (2) and renal function decreases with age, plasma H-FABP concentrations increase during aging. In addition, H-FABP release from skeletal muscle may increase with age or exercise, as has been described for myoglobin (48). A URL of 6 μg/L has been proposed independently by several groups (47)(49). Selected studies indicating key clinical uses of H-FABP are summarized in Table 44 . clinical applications of h-fabp Early marker of AMI. H-FABP was initially reported to be rapidly released from injured myocardium (50). Because of the recent redefinition of MI (51), biochemical markers have become even more important for assessment of suspected cardiac ischemia patients with non–ST-segment elevation. Because the plasma release characteristics of H-FABP after myocardial injury closely resemble those of myoglobin (52), the application of H-FABP as a sensitive early marker for myocardial injury has been investigated by several groups (53). In general, H-FABP was found to perform better than or similar to myoglobin (53). The areas under the ROC curves for these comparisons, which used the admission blood samples from all patients, were significantly larger for H-FABP than for myoglobin, indicating better performance of H-FABP within 6 h after onset of symptoms. Furthermore, subgroup analysis of patients presenting <6 h after onset of symptoms showed better performance for H-FABP compared with myoglobin (54). The observed higher sensitivity of H-FABP may be related to the higher cardiac tissue content of H-FABP compared with myoglobin. In addition, the reference values for H-FABP in plasma are far lower than those for myoglobin. Therefore, after myocardial injury, H-FABP increases to above the URL more rapidly than does myoglobin or troponin (54)(55)(56). This rapid increase to above the URL can also be used to further improve the diagnostic value of the marker (i.e., rule-out power) by use of sequential plasma H-FABP measurements. When they evaluated plasma H-FABP values at admission and 1–2 h after admission, Haastrup et al. (57) reported an increased probability of detecting an AMI. Okamoto et al. (54) measured concentrations of H-FABP, myoglobin, and CK-MB in 140 AMI patients, 49 non-AMI chest pain patients, and 75 healthy volunteers. The area under the ROC curve for H-FABP was significantly higher (0.921) than those of myoglobin (0.843) and CK-MB (0.654). In another study, H-FABP, cTnI, and creatine phosphokinase concentrations were measured in 218 patients with chest pain and suspected MI; 94 of these patients were eventually diagnosed with MI (55). H-FABP showed 100% sensitivity and negative predictive value at 1 h after admission (55). The areas under the ROC curves for H-FABP, creatine phosphokinase, and cTnI calculated at admission and 1 h after admission were 0.871 and 0.995, 0.711 and 0.856, and 0.677 and 0.845, respectively (55). Measurement of H-FABP in serum or plasma was suggested to allow the earliest exclusion of non-AMI patients. Seino et al. (56) compared the diagnostic efficacy of a newly developed whole blood rapid test for H-FABP with that of a rapid cTnT test in 129 consecutive patients with suspected cardiac ischemia, 31 of whom had a diagnosis of AMI. The respective temporal sensitivities of H-FABP and cTnT tests were 100% vs 50% at 3 h and 100% vs 100% at >12 h after onset of symptoms. The respective specificities were 63% vs 96.3% at 3 h and 75% vs 87.5% at >12 h. The negative predictive values were 100% vs 86.7% at 3 h and 100% vs 100% at >12 h. The rapid H-FABP assay was suggested to effectively exclude non-AMI patients within 3 h of onset (56). Differentiation of cardiac and skeletal muscle injury. H-FABP is produced mainly in the heart, but to a lesser extent, it is also produced in skeletal muscle (58). When patients suffered skeletal muscle injury as a result of cardioversion, multiorgan failure, postoperative states, or vigorous exercise such as running (59) or rowing (60), H-FABP was released into the blood. The myoglobin/H-FABP ratio has been used to differentiate between heart muscle (ratio = 2–10) and skeletal muscles (ratio = 20–70), depending on the type of muscle (58). In patients with AMI, the plasma myoglobin/H-FABP ratio was ∼5 during the entire period of increased plasma concentrations, whereas for patients with aortic surgery (causing no-flow ischemia of the lower extremities), the plasma myoglobin/H-FABP ratio was 45 (58). During defibrillation after AMI, the plasma myoglobin/H-FABP ratio increased from 8 to 60 in the 24 h after AMI as a result of injury of the intercostal pectoralis muscles (58). In cases in which a second increase in plasma concentrations of marker proteins occurs, this ratio can be used to differentiate a recurrent infarction (ratio remains at 2–10) from additional skeletal muscle injury (ratio increases to 20–70). Infarct size, reperfusion, and coronary bypass grafting. To evaluate the effect of thrombolytic therapy, the size of a myocardial infarct can be estimated by measuring the cumulative release of H-FABP. In patients treated with standard thrombolytic therapy after AMI, plasma concentrations of H-FABP and myoglobin peaked at ∼4 h after first symptoms, whereas creatine kinase (creatine phosphokinase or CK-MB) and lactate dehydrogenase peaked at ∼12 and 20 h, respectively (52). Because H-FABP and myoglobin rapidly return to their respective URLs (within 24 h after AMI) as a result of renal clearance (23)(52), both proteins can be used to assess a recurrent infarction within 10 h after first AMI (58), which might be missed by CK-MB, cTnT, and cTnI because plasma concentrations of these markers return much more slowly to reference values (61)(62). If no thrombolytic therapy is administered, the H-FABP concentration in plasma peaks at 8 h and returns to within reference values after only 36 h, comparable to myoglobin (62). These differences in release kinetics do not impact the measurement of cardiac proteins in plasma, however (32)(58)(62). Because H-FABP is cleared by the kidneys, renal insufficiency could potentially impact its clinical utility; however, data from de Groot et al. (32) indicate that individually estimated clearance rates can be applied successfully for infarct size estimation. The rapid release of H-FABP can also be used for the detection of successful coronary reperfusion in patients with AMI (63)(64)(65). Both plasma H-FABP and myoglobin were found to increase sharply after successful reperfusion, but in patients with failed reperfusion, both proteins increased more slowly. The relatively low sensitivity and specificity of ∼70% could be improved to 80% by normalization to infarct size. The characteristics of rapid release and ability to differentiate between skeletal or cardiac muscle injury can be useful for early detection of postoperative myocardial tissue loss in patients undergoing coronary bypass surgery (66). Clinical assessment of congestive heart failure (CHF). Preliminary studies in patients with CHF indicate that increased plasma concentrations of H-FABP and cTnT are associated with progressive deterioration of ventricular function and a worse prognosis (67). H-FABP concentrations were related not only to CHF severity (New York Heart Association classes 3 and 4) and serum cTnT concentrations (67), but also to the occurrence of recurrent cardiac events (68)(69). Knowledge regarding the significance of H-FABP as a marker of myocardial injury in CHF is continuing to evolve and needs further study. Prognostic value. In the early hours of acute coronary syndrome (ACS), selection of patients who are at high risk for cardiac events is an important factor for determining the appropriate treatment strategy. The use of plasma H-FABP concentrations for early prediction of adverse clinical outcomes in patients with suspected ACS has only recently been the subject of investigation but shows promising results; increased plasma H-FABP concentrations significantly correlated with increased cardiac event rates and cardiac mortality (70)(71). Pelsers et al.(68) showed that when plasma H-FABP was <6 μg/L, the negative predictive value for a recurrent event in CHF patients within 90 days was 81%, whereas cTnT <0.02 μg/L had a negative predictive value of 57%. This difference is most likely explained by insufficient sensitivity of the cTnT assay. Although for cTnT a cutoff value of 0.1 μg/L for indication of myocardial injury is commonly used, cutoff values of 0.05 and 0.02 μg/L are now being evaluated for more sensitive immunoassays. h-fabp as a potential marker of cardiac ischemia Although H-FABP is generally regarded as a marker of necrosis, one recent study has indicated its additional potential utility as a marker of ischemia (72). H-FABP concentrations measured in pericardial fluid samples collected immediately after median sternotomy were significantly increased in 17 patients with unstable angina who had anginal symptoms and/or ST changes compared with 17 other patients who did not have these symptoms [mean (SD) values were 16.3 (2.0) vs 9.6 (1.0) μg/L; P = 0.0046]. H-FABP secretion into the interstitial space may be mediated by increased permeability of the myocardial cell membrane associated with severe ischemia. The main advantage of H-FABP is its ability to exclude non-AMI patients very early after onset of symptoms. The fact that H-FABP may be present in the circulation in the absence of AMI makes it difficult to distinguish between patients with an AMI or unstable angina and warrants more investigation to definitively establish the diagnostic cutoff for H-FABP. In addition, only a few reports (68)(70)(71) have shown the prognostic value of H-FABP measurements in ACS patients. Further investigation of the prognostic value of H-FABP measurements is needed. In combination with cardiac troponins, H-FABP may be useful to cover the complete diagnostic window of patients presenting with ACS in the emergency department, along with the electrocardiographic and clinical symptoms. Widespread availability on automated analyzers is necessary for routine applicability of H-FABP. Conclusions Preliminary data suggest that FFAu concentrations have potential in identifying patients with cardiac ischemia. More work is needed, however, to clinically validate this marker and to meet quality specifications. H-FABP is a useful biomarker for detection of cardiac injury in ACS within 6 h of symptoms onset. Limitations include a lack of complete cardiac specificity, a relatively small diagnostic window of 24–30 h after the acute event, and the probability of falsely increased values in patients with renal insufficiency. Although a relatively small number of clinical studies have been performed (12 studies involving a total of 2130 patients), all of these studies showed better or similar performance of H-FABP compared with myoglobin for the early diagnosis of AMI. H-FABP also has prognostic value to predict recurrent cardiac events in patients with ACS or CHF. The use of H-FABP in ruling out MI in patients with ACS is promising but needs further study. Figure 1. Open in new tabDownload slide Schematic overview of the molecular mechanism of cellular uptake and use of long-chain fatty acids (FA). After dissociation from plasma albumin, fatty acids are translocated through the lipid bilayer (gray) via passive diffusion, membrane-associated proteins, or a combination of both (right side of schematic). The membrane-associated fatty acid transporters fatty acid-binding protein (FABPpm), fatty acid translocase (CD36), and fatty acid transport protein (FATP) are involved. Intracellular fatty acids are bound to cytoplasmic H-FABP (FABPc) and, after activation to fatty acyl-CoA, to acyl-CoA–binding protein (ACBP). ACSy, acyl-CoA synthetase. Figure 1. Open in new tabDownload slide Schematic overview of the molecular mechanism of cellular uptake and use of long-chain fatty acids (FA). After dissociation from plasma albumin, fatty acids are translocated through the lipid bilayer (gray) via passive diffusion, membrane-associated proteins, or a combination of both (right side of schematic). The membrane-associated fatty acid transporters fatty acid-binding protein (FABPpm), fatty acid translocase (CD36), and fatty acid transport protein (FATP) are involved. Intracellular fatty acids are bound to cytoplasmic H-FABP (FABPc) and, after activation to fatty acyl-CoA, to acyl-CoA–binding protein (ACBP). ACSy, acyl-CoA synthetase. Table 1. Serum and intracellular concentrations of H-FABP and FFAu under physiologic and pathophysiologic conditions. . Physiologic conditions . Ischemia . Necrosis . Plasma concentrations  FABP 0–6 μg/L 6–20 μg/L (72) 6–2000 μg/L (47)  FFAu 7.5 (2.5) nmol/L (13) High High Intracellular concentrations  FABP 500 μg/g (22) Low Very low  FFAu Normal High Low Main source of myocytic ATP Anaerobic metabolism (β-oxidation of fatty acids) Aerobic metabolism (accumulation of lactate) . Physiologic conditions . Ischemia . Necrosis . Plasma concentrations  FABP 0–6 μg/L 6–20 μg/L (72) 6–2000 μg/L (47)  FFAu 7.5 (2.5) nmol/L (13) High High Intracellular concentrations  FABP 500 μg/g (22) Low Very low  FFAu Normal High Low Main source of myocytic ATP Anaerobic metabolism (β-oxidation of fatty acids) Aerobic metabolism (accumulation of lactate) Table 1. Serum and intracellular concentrations of H-FABP and FFAu under physiologic and pathophysiologic conditions. . Physiologic conditions . Ischemia . Necrosis . Plasma concentrations  FABP 0–6 μg/L 6–20 μg/L (72) 6–2000 μg/L (47)  FFAu 7.5 (2.5) nmol/L (13) High High Intracellular concentrations  FABP 500 μg/g (22) Low Very low  FFAu Normal High Low Main source of myocytic ATP Anaerobic metabolism (β-oxidation of fatty acids) Aerobic metabolism (accumulation of lactate) . Physiologic conditions . Ischemia . Necrosis . Plasma concentrations  FABP 0–6 μg/L 6–20 μg/L (72) 6–2000 μg/L (47)  FFAu 7.5 (2.5) nmol/L (13) High High Intracellular concentrations  FABP 500 μg/g (22) Low Very low  FFAu Normal High Low Main source of myocytic ATP Anaerobic metabolism (β-oxidation of fatty acids) Aerobic metabolism (accumulation of lactate) Table 2. Clinical use of FFAu measurements. Clinical use . Study description . Findings . Conclusions/Comments . Reference . Marker of ischemia FFAu measured in 22 patients before and 30 min after coronary angioplasty FFAu increased 14-fold over baseline concentrations; highest concentrations were observed in patients with ischemic ST-segment changes Abnormal FFAu concentrations may be a sensitive marker of cardiac ischemia (15) Marker of ischemia 458 patients enrolled in TIMI II trial; blood was collected at presentation and 50 min, 5 h, and 8 h after tPA Sensitivity of FFAu was 91% at admission and 98% at 50 min after tPA (cutoff, 5 nmol/L); specificity was 93% for noncardiovascular patients and healthy persons; higher FFAu concentrations correlated with mortality (4-fold higher rate of death) Increased FFAu may be a sensitive marker of cardiac ischemia (20) Prognostic value 5250 men, followed for 22 years; FFAu concentrations were measured FFAu were found to be an independent risk factor for sudden death (OR1 = 1.70; 95% CI, 1.21–2.13) FFAu may have an arrhythmogenic role and contribute to a higher frequency of premature ventricular complexes and may therefore contribute to death (17) Clinical use . Study description . Findings . Conclusions/Comments . Reference . Marker of ischemia FFAu measured in 22 patients before and 30 min after coronary angioplasty FFAu increased 14-fold over baseline concentrations; highest concentrations were observed in patients with ischemic ST-segment changes Abnormal FFAu concentrations may be a sensitive marker of cardiac ischemia (15) Marker of ischemia 458 patients enrolled in TIMI II trial; blood was collected at presentation and 50 min, 5 h, and 8 h after tPA Sensitivity of FFAu was 91% at admission and 98% at 50 min after tPA (cutoff, 5 nmol/L); specificity was 93% for noncardiovascular patients and healthy persons; higher FFAu concentrations correlated with mortality (4-fold higher rate of death) Increased FFAu may be a sensitive marker of cardiac ischemia (20) Prognostic value 5250 men, followed for 22 years; FFAu concentrations were measured FFAu were found to be an independent risk factor for sudden death (OR1 = 1.70; 95% CI, 1.21–2.13) FFAu may have an arrhythmogenic role and contribute to a higher frequency of premature ventricular complexes and may therefore contribute to death (17) 1 OR, odds ratio; CI, confidence interval. Table 2. Clinical use of FFAu measurements. Clinical use . Study description . Findings . Conclusions/Comments . Reference . Marker of ischemia FFAu measured in 22 patients before and 30 min after coronary angioplasty FFAu increased 14-fold over baseline concentrations; highest concentrations were observed in patients with ischemic ST-segment changes Abnormal FFAu concentrations may be a sensitive marker of cardiac ischemia (15) Marker of ischemia 458 patients enrolled in TIMI II trial; blood was collected at presentation and 50 min, 5 h, and 8 h after tPA Sensitivity of FFAu was 91% at admission and 98% at 50 min after tPA (cutoff, 5 nmol/L); specificity was 93% for noncardiovascular patients and healthy persons; higher FFAu concentrations correlated with mortality (4-fold higher rate of death) Increased FFAu may be a sensitive marker of cardiac ischemia (20) Prognostic value 5250 men, followed for 22 years; FFAu concentrations were measured FFAu were found to be an independent risk factor for sudden death (OR1 = 1.70; 95% CI, 1.21–2.13) FFAu may have an arrhythmogenic role and contribute to a higher frequency of premature ventricular complexes and may therefore contribute to death (17) Clinical use . Study description . Findings . Conclusions/Comments . Reference . Marker of ischemia FFAu measured in 22 patients before and 30 min after coronary angioplasty FFAu increased 14-fold over baseline concentrations; highest concentrations were observed in patients with ischemic ST-segment changes Abnormal FFAu concentrations may be a sensitive marker of cardiac ischemia (15) Marker of ischemia 458 patients enrolled in TIMI II trial; blood was collected at presentation and 50 min, 5 h, and 8 h after tPA Sensitivity of FFAu was 91% at admission and 98% at 50 min after tPA (cutoff, 5 nmol/L); specificity was 93% for noncardiovascular patients and healthy persons; higher FFAu concentrations correlated with mortality (4-fold higher rate of death) Increased FFAu may be a sensitive marker of cardiac ischemia (20) Prognostic value 5250 men, followed for 22 years; FFAu concentrations were measured FFAu were found to be an independent risk factor for sudden death (OR1 = 1.70; 95% CI, 1.21–2.13) FFAu may have an arrhythmogenic role and contribute to a higher frequency of premature ventricular complexes and may therefore contribute to death (17) 1 OR, odds ratio; CI, confidence interval. Table 3. Characteristics of immunoassays for human H-FABP. Assay . Assay time, min . Sample . Calibration range, μg/L . Detection limit, μg/L . URL, μg/L . Year (Reference) . ELISA 45 Serum; plasma 0–60 0.3 6 1997 (31) IFMA1 50 Serum; plasma 0–300 0.1 6 1997 (73) EIA Serum; plasma 0–100 1 7 1997 (35) Immunosensor 20 Plasma 0–350 5 10 1997 (42) Latex 10 Serum; plasma 0–150 1.1 14 1998 (74) Lateral flow 15 Whole blood 6.2 6.2 2001 (39) Immunosensor 50 Whole blood 0–250 4 2002 (44) Lateral flow 15 Whole blood 0–125 2.8 7 2003 (37) Assay . Assay time, min . Sample . Calibration range, μg/L . Detection limit, μg/L . URL, μg/L . Year (Reference) . ELISA 45 Serum; plasma 0–60 0.3 6 1997 (31) IFMA1 50 Serum; plasma 0–300 0.1 6 1997 (73) EIA Serum; plasma 0–100 1 7 1997 (35) Immunosensor 20 Plasma 0–350 5 10 1997 (42) Latex 10 Serum; plasma 0–150 1.1 14 1998 (74) Lateral flow 15 Whole blood 6.2 6.2 2001 (39) Immunosensor 50 Whole blood 0–250 4 2002 (44) Lateral flow 15 Whole blood 0–125 2.8 7 2003 (37) 1 IFMA, immunofluorometric assay; EIA, enzyme immunoassay. Table 3. Characteristics of immunoassays for human H-FABP. Assay . Assay time, min . Sample . Calibration range, μg/L . Detection limit, μg/L . URL, μg/L . Year (Reference) . ELISA 45 Serum; plasma 0–60 0.3 6 1997 (31) IFMA1 50 Serum; plasma 0–300 0.1 6 1997 (73) EIA Serum; plasma 0–100 1 7 1997 (35) Immunosensor 20 Plasma 0–350 5 10 1997 (42) Latex 10 Serum; plasma 0–150 1.1 14 1998 (74) Lateral flow 15 Whole blood 6.2 6.2 2001 (39) Immunosensor 50 Whole blood 0–250 4 2002 (44) Lateral flow 15 Whole blood 0–125 2.8 7 2003 (37) Assay . Assay time, min . Sample . Calibration range, μg/L . Detection limit, μg/L . URL, μg/L . Year (Reference) . ELISA 45 Serum; plasma 0–60 0.3 6 1997 (31) IFMA1 50 Serum; plasma 0–300 0.1 6 1997 (73) EIA Serum; plasma 0–100 1 7 1997 (35) Immunosensor 20 Plasma 0–350 5 10 1997 (42) Latex 10 Serum; plasma 0–150 1.1 14 1998 (74) Lateral flow 15 Whole blood 6.2 6.2 2001 (39) Immunosensor 50 Whole blood 0–250 4 2002 (44) Lateral flow 15 Whole blood 0–125 2.8 7 2003 (37) 1 IFMA, immunofluorometric assay; EIA, enzyme immunoassay. Table 4. Key clinical uses of H-FABP. Clinical use . Study description . Findings . Conclusions/Comments . Reference . Early marker of MI H-FABP was measured in samples from 22 AMI patients In 18 of 22 AMI patients, H-FABP concentrations were at or above threshold in samples taken 3.5 h after first onset of symptoms Within 0.5–3.5 h after symptom onset, H-FABP had >80% sensitivity for AMI diagnosis [In a different study, sensitivity of CK-MB, CK mass, or CK activity and troponins within 0–6 h of symptom onset was reported to be <65% (75)] (76) Urinary marker of AMI Serum and urinary concentrations of H-FABP were determined in serial samples obtained from 11 AMI patients H-FABP was significantly increased in serum and urine samples obtained 5–10 h after symptoms developed and decreased sharply afterward H-FABP is a urinary marker of myocardial injury; H-FABP is eliminated from the circulation by the kidneys, but the exact mechanism is unknown; the only other urinary cardiac marker tested was myoglobin (2) Marker of myocardial injury after cardiac surgery H-FABP, CK-MB, and TnT concentrations were measured in blood samples serially collected from 10 patients undergoing CABG1 The time to peak after aortic declamping was shorter for H-FABP [1.4 (0.5) h] than for CK-MB [2.5 (0.5) h] and TnT [6.6 (1.3) h] H-FABP may be an early marker of myocardial injury in patients undergoing cardiac surgery (66) Detection of coronary reperfusion H-FABP and myoglobin concentrations were measured in serum samples from 45 patients with AMI treated with intracoronary thrombolysis or direct PCTA The predictive accuracies for H-FABP ratios >1.8 for detection of reperfusion within 15, 30, and 60 min of initiation of treatment were 93%, 98%, and 100%, respectively H-FABP and myoglobin ratios had similar predictive accuracies for early detection of successful coronary reperfusion (63) Infarct sizing H-FABP and myoglobin concentrations, CK-MB activity, and HBDH were assayed serially in plasma samples obtained from 20 AMI patients In 15 AMI patients with normal renal function, agreement was good between infarct size estimated from H-FABP or myoglobin curves and that estimated with CK-MB or HBDH Serial plasma H-FABP or myoglobin concentrations may be used for infarct sizing within the first 24 h after symptom onset only in AMI patients with normal renal function (77) Clinical use . Study description . Findings . Conclusions/Comments . Reference . Early marker of MI H-FABP was measured in samples from 22 AMI patients In 18 of 22 AMI patients, H-FABP concentrations were at or above threshold in samples taken 3.5 h after first onset of symptoms Within 0.5–3.5 h after symptom onset, H-FABP had >80% sensitivity for AMI diagnosis [In a different study, sensitivity of CK-MB, CK mass, or CK activity and troponins within 0–6 h of symptom onset was reported to be <65% (75)] (76) Urinary marker of AMI Serum and urinary concentrations of H-FABP were determined in serial samples obtained from 11 AMI patients H-FABP was significantly increased in serum and urine samples obtained 5–10 h after symptoms developed and decreased sharply afterward H-FABP is a urinary marker of myocardial injury; H-FABP is eliminated from the circulation by the kidneys, but the exact mechanism is unknown; the only other urinary cardiac marker tested was myoglobin (2) Marker of myocardial injury after cardiac surgery H-FABP, CK-MB, and TnT concentrations were measured in blood samples serially collected from 10 patients undergoing CABG1 The time to peak after aortic declamping was shorter for H-FABP [1.4 (0.5) h] than for CK-MB [2.5 (0.5) h] and TnT [6.6 (1.3) h] H-FABP may be an early marker of myocardial injury in patients undergoing cardiac surgery (66) Detection of coronary reperfusion H-FABP and myoglobin concentrations were measured in serum samples from 45 patients with AMI treated with intracoronary thrombolysis or direct PCTA The predictive accuracies for H-FABP ratios >1.8 for detection of reperfusion within 15, 30, and 60 min of initiation of treatment were 93%, 98%, and 100%, respectively H-FABP and myoglobin ratios had similar predictive accuracies for early detection of successful coronary reperfusion (63) Infarct sizing H-FABP and myoglobin concentrations, CK-MB activity, and HBDH were assayed serially in plasma samples obtained from 20 AMI patients In 15 AMI patients with normal renal function, agreement was good between infarct size estimated from H-FABP or myoglobin curves and that estimated with CK-MB or HBDH Serial plasma H-FABP or myoglobin concentrations may be used for infarct sizing within the first 24 h after symptom onset only in AMI patients with normal renal function (77) 1 CABG, coronary artery bypass grafting; HBDH, hydroxybutyrate dehydrogenase. Table 4. Key clinical uses of H-FABP. Clinical use . Study description . Findings . Conclusions/Comments . Reference . Early marker of MI H-FABP was measured in samples from 22 AMI patients In 18 of 22 AMI patients, H-FABP concentrations were at or above threshold in samples taken 3.5 h after first onset of symptoms Within 0.5–3.5 h after symptom onset, H-FABP had >80% sensitivity for AMI diagnosis [In a different study, sensitivity of CK-MB, CK mass, or CK activity and troponins within 0–6 h of symptom onset was reported to be <65% (75)] (76) Urinary marker of AMI Serum and urinary concentrations of H-FABP were determined in serial samples obtained from 11 AMI patients H-FABP was significantly increased in serum and urine samples obtained 5–10 h after symptoms developed and decreased sharply afterward H-FABP is a urinary marker of myocardial injury; H-FABP is eliminated from the circulation by the kidneys, but the exact mechanism is unknown; the only other urinary cardiac marker tested was myoglobin (2) Marker of myocardial injury after cardiac surgery H-FABP, CK-MB, and TnT concentrations were measured in blood samples serially collected from 10 patients undergoing CABG1 The time to peak after aortic declamping was shorter for H-FABP [1.4 (0.5) h] than for CK-MB [2.5 (0.5) h] and TnT [6.6 (1.3) h] H-FABP may be an early marker of myocardial injury in patients undergoing cardiac surgery (66) Detection of coronary reperfusion H-FABP and myoglobin concentrations were measured in serum samples from 45 patients with AMI treated with intracoronary thrombolysis or direct PCTA The predictive accuracies for H-FABP ratios >1.8 for detection of reperfusion within 15, 30, and 60 min of initiation of treatment were 93%, 98%, and 100%, respectively H-FABP and myoglobin ratios had similar predictive accuracies for early detection of successful coronary reperfusion (63) Infarct sizing H-FABP and myoglobin concentrations, CK-MB activity, and HBDH were assayed serially in plasma samples obtained from 20 AMI patients In 15 AMI patients with normal renal function, agreement was good between infarct size estimated from H-FABP or myoglobin curves and that estimated with CK-MB or HBDH Serial plasma H-FABP or myoglobin concentrations may be used for infarct sizing within the first 24 h after symptom onset only in AMI patients with normal renal function (77) Clinical use . Study description . Findings . Conclusions/Comments . Reference . Early marker of MI H-FABP was measured in samples from 22 AMI patients In 18 of 22 AMI patients, H-FABP concentrations were at or above threshold in samples taken 3.5 h after first onset of symptoms Within 0.5–3.5 h after symptom onset, H-FABP had >80% sensitivity for AMI diagnosis [In a different study, sensitivity of CK-MB, CK mass, or CK activity and troponins within 0–6 h of symptom onset was reported to be <65% (75)] (76) Urinary marker of AMI Serum and urinary concentrations of H-FABP were determined in serial samples obtained from 11 AMI patients H-FABP was significantly increased in serum and urine samples obtained 5–10 h after symptoms developed and decreased sharply afterward H-FABP is a urinary marker of myocardial injury; H-FABP is eliminated from the circulation by the kidneys, but the exact mechanism is unknown; the only other urinary cardiac marker tested was myoglobin (2) Marker of myocardial injury after cardiac surgery H-FABP, CK-MB, and TnT concentrations were measured in blood samples serially collected from 10 patients undergoing CABG1 The time to peak after aortic declamping was shorter for H-FABP [1.4 (0.5) h] than for CK-MB [2.5 (0.5) h] and TnT [6.6 (1.3) h] H-FABP may be an early marker of myocardial injury in patients undergoing cardiac surgery (66) Detection of coronary reperfusion H-FABP and myoglobin concentrations were measured in serum samples from 45 patients with AMI treated with intracoronary thrombolysis or direct PCTA The predictive accuracies for H-FABP ratios >1.8 for detection of reperfusion within 15, 30, and 60 min of initiation of treatment were 93%, 98%, and 100%, respectively H-FABP and myoglobin ratios had similar predictive accuracies for early detection of successful coronary reperfusion (63) Infarct sizing H-FABP and myoglobin concentrations, CK-MB activity, and HBDH were assayed serially in plasma samples obtained from 20 AMI patients In 15 AMI patients with normal renal function, agreement was good between infarct size estimated from H-FABP or myoglobin curves and that estimated with CK-MB or HBDH Serial plasma H-FABP or myoglobin concentrations may be used for infarct sizing within the first 24 h after symptom onset only in AMI patients with normal renal function (77) 1 CABG, coronary artery bypass grafting; HBDH, hydroxybutyrate dehydrogenase. 1 Nonstandard abbreviations: cTnI and cTnT, cardiac troponin I and T, respectively; CK-MB, creatine kinase-MB; H-FABP, heart-type fatty acid–binding protein; FFAu, unbound free fatty acids; AMI, acute myocardial infarction; ADIFAB, acrylodated intestinal fatty acid–binding protein; URL, upper reference limit; PCTA, percutaneous transluminal coronary angioplasty; ECG, electrocardiography; IHD, ischemic heart disease; tPA, tissue plasminogen activator; CHF, congestive heart failure; and ACS, acute coronary syndrome(s). Portions of the FABP diagnostic studies performed in the laboratory of Jan F.C. Glatz (Netherlands Heart Foundation Professor of Cardiac Metabolism) were supported by the Ministry of Economic Affairs (BTS Grant 97.188) and the Dutch Technology Foundation (Grant GGN4860). 1 Wu AHB eds. Cardiac markers, 2nd ed 2003 : 467pp Humana Press Totowa, NJ. . 2 Tanaka T, Hirota Y, Sohmiya K, Nishimura S, Kawamura K. Serum and urinary human heart fatty acid-binding protein in acute myocardial infarction. Clin Biochem 1991 ; 24 : 195 -201. 3 Mair J. Tissue release of cardiac markers: from physiology to clinical applications. Clin Chem Lab Med 1999 ; 37 : 1077 -1084. 4 Piper HM, Schwartz P, Spahr R, Hutter JF, Spieckermann PG. Early enzyme release from myocardial cells is not due to irreversible cell damage. J Mol Cell Cardiol 1984 ; 16 : 385 -388. 5 Liedtke AJ. Alterations of carbohydrate and lipid metabolism in the acutely ischemic heart. Prog Cardiovasc Dis 1981 ; 23 : 321 -336. 6 van der Vusse GJ, van Bilsen M, Glatz JF, Hasselbaink DM, Luiken JJ. Critical steps in cellular fatty acid uptake and utilization. Mol Cell Biochem 2002 ; 239 : 9 -15. 7 Oliver MF, Kurien VA, Greenwood TW. Relation between serum-free-fatty acids and arrhythmias and death after acute myocardial infarction. Lancet 1968 ; 1 : 710 -714. 8 Gupta DK, Jewitt DE, Young R, Hartog M, Opie LH. Increased plasma-free-fatty-acid concentrations and their significance in patients with acute myocardial infarction. Lancet 1969 ; 2 : 1209 -1213. 9 Hendrickson SC, St Louis JD, Lowe JE, Abdel-aleem S. Free fatty acid metabolism during myocardial ischemia and reperfusion [Review]. Mol Cell Biochem 1997 ; 166 : 85 -94. 10 Spector AA. Fatty acid binding to plasma albumin. J Lipid Res 1975 ; 16 : 165 -179. 11 Kiziltunc A, Akcay F. An enzymatic method for the determination of free fatty acids in serum/plasma. Clin Chem Lab Med 1998 ; 36 : 83 -86. 12 Richieri GV, Ogata RT, Kleinfeld AM. A fluorescently labeled intestinal fatty acid binding protein. Interactions with fatty acids and its use in monitoring free fatty acids. J Biol Chem 1992 ; 267 : 23495 -23501. 13 Richieri GV, Kleinfeld AM. Unbound free fatty acid levels in human serum. J Lipid Res 1995 ; 36 : 229 -240. 14 Apple FS, Wu AH, Mair J, Ravkilde J, Panteghini M, Tate J, et al. Future biomarkers for detection of ischemia and risk stratification in acute coronary syndrome. Clin Chem 2005 ; 51 : 810 -824. 15 Kleinfeld AM, Prothro D, Brown DL, Davis RC, Richieri GV, DeMaria A. Increases in serum unbound free fatty acid levels following coronary angioplasty. Am J Cardiol 1996 ; 78 : 1350 -1354. 16 Tansey MJ, Opie LH. Relation between plasma free fatty acids and arrhythmias within the first twelve hours of acute myocardial infarction. Lancet 1983 ; 2 : 419 -422. 17 Jouven X, Charles MA, Desnos M, Ducimetiere P. Circulating nonesterified fatty acid level as a predictive risk factor for sudden death in the population. Circulation 2001 ; 104 : 756 -761. 18 Pirro M, Mauriege P, Tchernof A, Cantin B, Dagenais GR, Despres JP, et al. Plasma free fatty acid levels and the risk of ischemic heart disease in men: prospective results from the Quebec Cardiovascular Study. Atherosclerosis 2002 ; 160 : 377 -384. 19 Comparison of invasive and conservative strategies after treatment with intravenous tissue plasminogen activator in acute myocardial infarction. Results of the thrombolysis in myocardial infarction (TIMI) phase II trial. The TIMI Study Group. N Engl J Med 1989 ; 320 : 618 -627. 20 Kleinfeld AM, Kleinfeld KJ, Adams JE. Serum levels of unbound free fatty acids reveal high sensitivity for early detection of acute myocardial infarction in patients samples from the TIMI II trial. J Am Coll Cardiol 2002 ; 39 : 312A . 21 Glatz JF, van der Vusse GJ. Cellular fatty acid-binding proteins: their function and physiological significance. Prog Lipid Res 1996 ; 35 : 243 -282. 22 Young AC, Scapin G, Kromminga A, Patel SB, Veerkamp JH, Sacchettini JC. Structural studies on human muscle fatty acid-binding protein at 1.4 Å resolution: binding interaction with three C18 fatty acids. Structure 1994 ; 2 : 523 -534. 23 Zschiesche W, Kleine AH, Spitzer E, Veerkamp JH, Glatz JF. Histochemical localization of heart-type fatty acid-binding protein in human and murine tissues. Histochem Cell Biol 1995 ; 103 : 147 -156. 24 Maatman RG, van de Westerlo EM, van Kuppevelt TH, Veerkamp JH. Molecular identification of the liver- and the heart-type fatty acid-binding proteins in human and rat kidney. Use of the reverse transcriptase polymerase chain reaction. Biochem J 1992 ; 288 : 285 -290. 25 Pelsers MM, Hanhoff T, Van der Voort D, Arts B, Peters M, Ponds R, et al. Brain- and heart-type fatty acid-binding proteins in the brain: tissue distribution and clinical utility. Clin Chem 2004 ; 50 : 1568 -1575. 26 Schreiber A, Specht B, Pelsers MM, Glatz JF, Börchers T, Spener F. Recombinant human heart-type fatty acid-binding protein as standard in immunochemical assays. Clin Chem Med Lab 1998 ; 36 : 283 -288. 27 Wolfrum C, Borrmann CM, Borchers T, Spener F. Fatty acids and hypolipidemic drugs regulate peroxisome proliferator-activated receptors α- and γ-mediated gene expression via liver fatty acid binding protein: a signaling path to the nucleus. Proc Natl Acad Sci U S A 2001 ; 98 : 2323 -2328. 28 Glatz JF, Storch J. Unraveling the significance of cellular fatty acid-binding proteins. Curr Opin Lipidol 2001 ; 12 : 267 -274. 29 van Breda E, Keizer HA, Vork MM, Surtel DA, de Jong YF, van der Vusse GJ, et al. Modulation of fatty-acid-binding protein content of rat heart and skeletal muscle by endurance training and testosterone treatment. Pflugers Arch 1992 ; 421 : 274 -279. 30 Glatz JF, van Breda E, Keizer HA, de Jong YF, Lakey JR, Rajotte RV, et al. Rat heart fatty acid-binding protein content is increased in experimental diabetes. Biochem Biophys Res Commun 1994 ; 199 : 639 -646. 31 Wodzig KW, Pelsers MM, van der Vusse GJ, Roos W, Glatz JF. One-step enzyme linked immunosorbent assay (ELISA) for plasma fatty acid-binding protein. Ann Clin Biochem 1997 ; 34 : 263 -268. 32 de Groot MJ, Wodzig KW, Simoons ML, Glatz JF, Hermens WT. Measurement of myocardial infarct size from plasma fatty acid-binding protein or myoglobin, using individually estimated clearance rates. Cardiovasc Res 1999 ; 44 : 315 -324. 33 Ohkaru Y, Asayama K, Ishii H, Nishimura S, Sunahara N, Tanaka T, et al. Development of a sandwich enzyme-linked immunosorbent assay for the determination of human heart type fatty acid-binding protein in plasma and urine by using two different monoclonal antibodies specific for human heart fatty acid-binding protein. J Immunol Methods 1995 ; 178 : 99 -111. 34 Sanders GT, Schouten Y, de Winter RJ. Evaluation of human heart type fatty acid-binding protein assay for the early detection of myocardial infarction [Abstract]. Clin Chem 1998 ; 44 (Suppl 6): A132 . 35 Fiebig M, Dodge M, Mangion J. Characterization and development of a COBAS EIA to quantitate human heart fatty acid-binding protein for the early detection of acute myocardial infarction [Abstract]. Clin Chem 1997 ; 43 (Suppl 6): S158 . 36 Speijer H, Laterveer-Vreeswijk RH, Glatz JF, Nieuwenhuizen W, Hermens WT. One-step immunoassay for measuring protein concentrations in plasma, based on precipitate-enhanced ellipsometry. Anal Biochem 2004 ; 326 : 257 -261. 37 Chan CP, Sum KW, Cheung KY, Glatz JF, Sanderson JE, Hempel A, et al. Development of a quantitative lateral-flow assay for rapid detection of fatty acid-binding protein. J Immunol Methods 2003 ; 279 : 91 -100. 38 Chan CPY, Cheng WS, Glatz JFC. Early diagnosis of acute myocardial infarction using immunosensors and immunotests. Anal Lett 2003 ; 36 : 1987 -2004. 39 Watanabe T, Ohkubo Y, Matsuoka H, Kimura H, Sakai Y, Ohkaru Y, et al. Development of a simple whole blood panel test for detection of human heart-type fatty acid-binding protein. Clin Biochem 2001 ; 34 : 257 -263. 40 Siegmann-Thoss C, Renneberg R, Glatz JFC, Spener F. Enzyme immunosensors for diagnosis of myocardial infarction. Sensors Actuators 1996 ; B30 : 71 -76. 41 Kunz U, Kasterlamp A, Renneberg R, Spener F, Cammann K. Sensing fatty acid-binding protein with planar and fiber-optical surface plasmon resonance spectroscopy devices. Sensors Actuators 1996 ; 32 : 129 -155. 42 Schreiber A, Feldbrügge R, Key G, Glatz JF, Spener F. An immunosensor based on disposable electrodes for rapid estimation of fatty acid-binding protein, an early marker of myocardial infarction. Biosens Bioelectron 1997 ; 12 : 1131 -1137. 43 Key G, Schreiber A, Feldbrugge R, McNeil CJ, Jorgensen P, Pelsers MM, et al. Multicenter evaluation of an amperometric immunosensor for plasma fatty acid-binding protein: an early marker for acute myocardial infarction. Clin Biochem 1999 ; 32 : 229 -231. 44 O’Regan TM, Pravda M, O’Sullivan CK, Guilbault GG. Development of a disposable immunosensor for the detection of human heart-type fatty acid-binding protein in whole blood using screen-printed carbon electrodes. Talanta 2002 ; 57 : 501 -510. 45 Van der Voort D, Pelsers MM, Korf J, Hermens WT, Glatz JFC. Development of a displacement immunoassay for human heart-type fatty acid-binding protein in plasma: the basic conditions. Biosens Bioelectron 2003 ; 19 : 465 -471. 46 Glatz JF, Kleine AH, van Nieuwenhoven FA, Hermens WT, van Dieijen-Visser MP, van der Vusse GJ. Fatty-acid-binding protein as a plasma marker for the estimation of myocardial infarct size in humans. Br Heart J 1994 ; 71 : 135 -140. 47 Pelsers MM, Chapelle JP, Knapen M, Vermeer C, Muijtjens AM, Hermens WT, et al. Influence of age and sex and day-to-day and within-day biological variation on plasma concentrations of fatty acid-binding protein and myoglobin in healthy subjects. Clin Chem 1999 ; 45 : 441 -443. 48 Chen IW, David R, Maxon HR, Sperling M, Stein EA. Age-, sex- and race-related differences in myoglobin concentrations in the serum of healthy persons. Clin Chem 1980 ; 26 : 1864 -1868. 49 Pagani F, Bonora R, Bonetti G, Panteghini M. Evaluation of a sandwich enzyme-linked immunosorbent assay for the measurement of serum heart fatty acid-binding protein. Ann Clin Biochem 2002 ; 39 : 404 -405. 50 Glatz JF, van Bilsen M, Paulussen RJ, Veerkamp J, van der Vusse GJ, Reneman RS. Release of fatty acid-binding protein from isolated rat heart subjected to ischemia and reperfusion or to the calcium paradox. Biochim Biophys Acta 1988 ; 961 : 148 -152. 51 Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction redefined—a consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction. J Am Coll Cardiol 2000 ; 36 : 959 -969. 52 Hermens WT. Mechanisms of protein release from injured heart muscle. Dev Cardiovasc Med 1998 ; 205 : 85 -98. 53 Pelsers MM, Hermens WT, Glatz JF. Fatty acid-binding proteins as plasma markers of tissue injury. Clin Chim Acta 2005 ; 352 : 15 -35. 54 Okamoto F, Sohmiya K, Ohkaru Y, Kawamura K, Asayama K, Kimura H, et al. Human heart-type cytoplasmic fatty acid-binding protein (H-FABP) for the diagnosis of acute myocardial infarction. Clinical evaluation of H-FABP in comparison with myoglobin and creatine kinase isoenzyme MB. Clin Chem Lab Med 2000 ; 38 : 231 -238. 55 Chan CP, Sanderson JE, Glatz JF, Cheng WS, Hempel A, Renneberg R. A superior early myocardial infarction marker—human heart-type fatty acid-binding protein. Z Kardiol 2004 ; 93 : 388 -397. 56 Seino Y, Tomita Y, Takano T, Ohbayashi K, . Tokyo Rapid-Test Office Cardiologists (Tokyo-ROC) Study. Office cardiologists cooperative study on whole blood rapid panel tests in patients with suspicious acute myocardial infarction: comparison between heart-type fatty acid-binding protein and troponin T tests. Circ J 2004 ; 68 : 144 -148. 57 Haastrup B, Gill S, Kristensen SR, Jorgensen PJ, Glatz JF, Haghfelt T, et al. Biochemical markers of ischemia for the early identification of acute myocardial infarction without ST segment elevation. Cardiology 2000 ; 94 : 254 -261. 58 Van Nieuwenhoven FA, Kleine AH, Wodzig WH, Hermens WT, Kragten HA, Maessen JG, et al. Discrimination between myocardial and skeletal muscle injury by assessment of the plasma ratio of myoglobin over fatty acid-binding protein. Circulation 1995 ; 92 : 2848 -2854. 59 Sorichter S, Mair J, Koller A, Pelsers MM, Puschendorf B, Glatz JF. Early assessment of exercise induced skeletal muscle injury using plasma fatty acid binding protein. Br J Sports Med 1998 ; 32 : 121 -124. 60 Yuan Y, Kwong AW, Kaptein WA, Fong C, Tse M, Glatz JF, et al. The responses of fatty acid-binding protein and creatine kinase to acute and chronic exercise in junior rowers. Res Q Exerc Sport 2003 ; 74 : 277 -283. 61 Kragten JA, Hermens WT, van Dieijen-Visser MP. Cardiac troponin T release into plasma after acute myocardial infarction: only fractional recovery compared with enzymes. Ann Clin Biochem 1996 ; 33 : 314 -323. 62 Wodzig KW, Kragten JA, Modrzejewski W, Gorski J, van Dieijen-Visser MP, Glatz JF, et al. Thrombolytic therapy does not change the release ratios of enzymatic and non-enzymatic myocardial marker proteins. Clin Chim Acta 1998 ; 272 : 209 -223. 63 Ishii J, Nagamura Y, Nomura M, Wang JH, Taga S, Kinoshita M, et al. Early detection of successful coronary reperfusion based on serum concentration of human heart-type cytoplasmic fatty acid-binding protein. Clin Chim Acta 1997 ; 262 : 13 -27. 64 de Lemos JA, Antman EM, Morrow DA, Llevadot J, Giugliano RP, Coulter SA, et al. Heart-type fatty acid binding protein as a marker of reperfusion after thrombolytic therapy. Clin Chim Acta 2000 ; 298 : 85 -97. 65 De Groot MJ, Muijtjens AM, Simoons ML, Hermens WT, Glatz JFC. Assessment of coronary reperfusion in patients with myocardial infarction using fatty acid-binding proteins concentrations in plasma. Heart 2001 ; 85 : 278 -285. 66 Hayashida N, Chihara S, Akasu K, Oda T, Tayama E, Kai E, et al. Plasma and urinary levels of heart fatty acid-binding protein in patients undergoing cardiac surgery. Jpn Circ J 2000 ; 64 : 18 -22. 67 Setsuta K, Seino Y, Ogawa T, Arao M, Miyatake Y, Takano T. Use of cytosolic and myofibril markers in the detection of ongoing myocardial damage in patients with chronic heart failure. Am J Med 2002 ; 113 : 717 -722. 68 Pelsers MMAL. Fatty acid-binding protein as plasma marker for tissue injury [PhD thesis] 2004 Maastricht University Maastricht, The Netherlands. . 69 Goto T, Takase H, Toriyama T, Sugiura T, Sato K, Ueda R, et al. Circulating concentrations of cardiac proteins indicate the severity of congestive heart failure. Heart 2003 ; 89 : 1303 -1307. 70 Ishii J, Nakamura Y, Naruse H. Heart-type fatty acid-binding protein is a useful marker for diagnosis of acute myocardial infarction and risk stratification of cardiac mortality and cardiac events in patients with early hours of acute coronary syndrome. Circulation 2003 ; 18 : IV-316 . 71 Erlikh A, Trifonov I, Katrukha A, Gratsianskii N. Prognostic value of serum heart fatty acid-binding protein in patients with non-ST elevation acute coronary syndrome: results of follow-up for 6 months. Circulation 2003 ; 18 : IV-648 . 72 Tambara K, Fujita M, Miyamoto S, Doi K, Nishimura K, Komeda M. Pericardial fluid level of heart-type cytoplasmic fatty acid-binding protein (H-FABP) is an indicator of severe myocardial ischemia. Int J Cardiol 2004 ; 93 : 281 -284. 73 Katrukha A, Bereznikova A, Filatov V. Development of sandwich time-resolved immuno-fluorometric assay for the quantitative determination of fatty acid-binding protein (FABP) [Abstract]. Clin Chem 1997 ; 43 (Suppl 6): S106 . 74 Robers M, Van der Hulst FF, Fischer MA, Roos W, Salud CE, Eisenwiener HG, et al. Development of a rapid microparticle-enhanced turbidimetric immunoassay for plasma fatty acid-binding protein, an early marker of acute myocardial infarction. Clin Chem 1998 ; 44 : 1564 -1567. 75 Koelemay MJ, Gorgels JP, van Vlies B, Smits R, Tijssen JG, Haagen FD. Failure of new biochemical markers to exclude acute myocardial infarction at admission. Lancet 1993 ; 342 : 1220 -1222. 76 Kleine AH, Glatz JF, Van Nieuwenhoven FA, Van der Vusse GJ. Release of heart fatty acid-binding protein into plasma after acute myocardial infarction in man. Mol Cell Biochem 1992 ; 116 : 155 -162. 77 Wodzig KW, Kragten JA, Hermens WT, Glatz JF, van Dieijen-Visser MP. Estimation of myocardial infarct size from plasma myoglobin or fatty acid-binding protein. Influence of renal function. Eur J Clin Chem Clin Biochem 1997 ; 35 : 191 -198. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Willemse, Johan, L;Hendriks, Dirk, F

doi: 10.1373/clinchem.2005.055814pmid: 16299049

Abstract Background: The importance of carboxypeptidase U (CPU) as a novel regulator of the fibrinolytic rate has attracted much interest during recent years. CPU circulates in plasma as a zymogen, proCPU, that can be activated by thrombin, thrombin-thrombomodulin (T-Tm), or plasmin. Given that the proCPU concentration in plasma is far below its Km for activation by the T-Tm complex, the formation of CPU will be directly proportional to the proCPU concentration. A low or high proCPU plasma concentration might therefore tip the balance between profibrinolytic and antifibrinolytic pathways and thereby cause a predisposition to bleeding or thrombosis. Content: To measure plasma proCPU concentrations, different methods have been developed based on 2 different principles: antigen determination and measurement of CPU activity after quantitative conversion of the proenzyme to its active form by addition of T-Tm. The major drawbacks that should be kept in mind when analyzing clinical samples by both principles are reviewed. Conclusions: proCPU is a potential prothrombotic risk factor. Evaluation of its relationship with thrombosis requires accurate assays. Many assays used in different clinical settings are inadequately validated, forcing reconsideration of conclusions made in these reports. Carboxypeptidase U (CPU;1 activated thrombin-activable fibrinolysis inhibitor; EC 3.4.17.20) is a recently discovered enzyme present in the circulation as its zymogen, procarboxypeptidase U [proCPU; thrombin-activable fibrinolysis inhibitor (TAFI)]. CPU was first described by Hendriks et al. (1)(2)(3) as a labile enzyme that interferes with the assay of carboxypeptidase N (CPN), a constitutively active carboxypeptidase in plasma. They named the enzyme carboxypeptidase U, the U standing for unstable. At that time, CPU was assumed to participate in the processing of bioactive peptides such bradykinin and anaphylatoxins by analogy with CPN, and a possible role in fibrinolysis was suggested (2)(4)(5)(6). In the following years the proCPU/CPU pathway was identified as an important coagulation-dependent pathway in the control of the fibrinolytic rate. By removing the C-terminal lysine residues on partially degraded fibrin, CPU prevents the acceleration of plasminogen binding and activation and promotes the inhibition of plasmin by α2-antiplasmin (7)(8)(9)(10). As long as the plasma CPU activity remains above a certain threshold value, fibrinolysis does not accelerate but stays in its initial phase (11)(12). It has also been suggested that CPU is an acute-phase protein (13)(14) and that, in addition to its role in fibrinolysis, it plays a role in regulating inflammation (15)(16)(17). Indeed, recent reports have suggested that CPU can cleave bradykinin and C5a—and thereby modulate their proinflammatory functions—as efficiently as plasmin-cleaved fibrin peptides (17). Thrombin is believed to be the activator of proCPU in vivo, albeit with low catalytic efficiency in vitro. Thrombomodulin stimulates the rate of proCPU activation by thrombin 1250-fold (18)(19). proCPU circulates in plasma at a concentration of 73–250 nmol/L (4.4–15.0 μg/mL) (20)(21)(22)(23). Given that the proCPU concentration in plasma is far below its Km for activation by the thrombin–thrombomodulin (T-Tm) complex, the formation of CPU will be dependent on both the T-Tm concentration and the proCPU concentration (18)(20)(21)(22)(23). A low or high proCPU plasma concentration might therefore tip the balance between profibrinolytic and antifibrinolytic pathways and thereby cause a predisposition to bleeding or thrombosis. Plasma proCPU concentrations were therefore analyzed in different clinical settings. The proCPU in plasma can be measured by both immunologic assays (sandwich ELISA) or activity assays. In this brief review we will discuss the advantages and disadvantages of immunologic and activity-based assays and focus on the major drawbacks of those assays that should be kept in mind when clinical samples are analyzed. proCPU is a difficult protein to analyze in a laboratory. Research groups who are planning to analyze proCPU in clinical studies should be aware of those difficulties and pitfalls before they choose a analytical method to perform the proCPU measurements. Unfortunately, many clinical studies were performed in the past with inadequately validated assays, forcing reconsideration of conclusions made in those reports. Antigen Assays The most common way of measuring antigen concentrations is by immunoassay. Different research groups have developed in-house ELISAs, and currently, several commercial ELISAs are available (20)(21)(22)(23)(24). A rocket immunoelectrophoresis method was also described to measure proCPU antigen by use of rabbit anti-proCPU IgG (25). advantages ELISAs are very easy to perform: no preliminary activation of the zymogen is required, and no interference of the constitutively active CPN is seen, which explain the widespread use of this assay format. drawbacks To understand the major pitfalls of the antigen assays, a few recently discovered characteristics should first be reviewed. Two functional proCPU variants have been described: the first variant differed from the published proCPU sequence at nucleotide 505 (G-to-A substitution producing an Ala-to-Thr substitution at position 147 in the protein) (26). Brouwers et al.(27) described a second functional polymorphism at nucleotide 1040 (C-to-T substitution producing a Thr-to-Ile substitution at position 325 in the protein). This Thr325Ile polymorphism is of particular interest because the proCPU Ile325 variant has an extended functional half-life (28). Because increased antifibrinolytic activity and, theoretically, a higher risk of thrombosis can be attributable to higher proCPU concentrations or to increased CPU stability (related to the Thr325Ile polymorphism), both polymorphisms have been investigated in different clinical studies. In addition to these polymorphisms, several polymorphisms in the 5′ regulatory region, the promoter region, and the 3′ untranslated region have been described (29)(30). Some of these polymorphisms appear to be in strong association with plasma proCPU concentrations, which could imply differences in proCPU gene expression. The polymorphisms associated with proCPU plasma concentrations, however, are in strong linkage dysequilibrium with each other and with the 505A/G single-nucleotide polymorphism in the coding region (29)(30). This Ala147Thr polymorphism has the strongest association with antigen concentrations but is also in strong linkage dysequilibrium with the Thr325Ile polymorphism (27). Gils et al. (24), however, recently demonstrated that the substitution of Ile for Thr at position 325 can have a large impact on the immunoreactivity of the isoform. From 144 monoclonal antibody combinations (which were tested according to differences in immunoreactivity), they developed 2 distinct ELISAs. The first ELISA was independent of the 325 genotype, but the second showed a complete lack of reactivity with the Ile/Ile genotype. Partial dependency of the 325 genotype was also seen with most of the commercially available ELISAs used in many clinical studies [e.g., 44 (8.9)% and 100 (30)% for the Ile/Ile and Thr/Thr isoforms, respectively, expressed as the mean (SD) percentage vs normal pooled plasma as measured by the commercially available Milan Analytica assay]. Thus, the association reported in several studies between the polymorphisms and proCPU antigen concentrations may reflect differences in assay sensitivity between isoforms rather than differences in expression. Although ELISAs are very attractive and very easy to use, great care should be taken. Assays should be very well characterized with respect to genotype-dependent reactivity. In addition to the polymorphism issue, various proCPU and CPU forms are present in plasma (Fig. 11 ). It is likely that antibodies raised against the zymogen can have different reactivities toward proCPU, proCPU bound to plasminogen, CPU, and conformationally and proteolytically inactivated forms of proCPU and CPU (23)(24)(31). Although in plasma the protein will be present predominantly in its zymogen form, relevant amounts of active and subsequently inactivated enzyme can be expected in plasma from patients in whom the coagulation and fibrinolysis cascades are activated. Gils et al. (24) found different reactivities toward blotted proCPU fragments when they used 27 different monoclonal antibodies. For example, some antibodies reacted only with the 55-kDa proCPU, some with both the 55- and 35-kDa proCPU fragments, and others with the 55-, 35-, and 25-kDa proCPU fragments. Standardization of the various assays is another critical issue. Different preparations of pooled plasma and purified proCPU used as calibrators in the different assays are probably one of the main reasons for the large variations in the mean plasma concentrations and ranges reported (32). proCPU concentrations in pooled plasma varied from 73 nmol/L (4.4 μg/mL) to 250 nmol/L (15 μg/mL) (20)(21)(22)(25). Strömqvist et al. (22) measured the plasma proCPU concentration in 479 apparently healthy volunteers and found a mean (SD) of 13.4 (2.5) μg/mL, or 223 nmol/L. Activity-Based Assays Activity assays can be used as an alternative for antigen assays. Most of the assays are based on the fact that CPU can cleave a C-terminal arginine or lysine from a synthetic substrate. The cleaved arginine/lysine or the other fragment can be measured by different analytical methods. We will discuss the advantages and drawbacks of activity-based assays and will give an overview of the most important currently used methods. advantages A major advantage is that only the enzymatically active CPU is measured (after quantitative conversion of proCPU) without interference from inactivated CPU, the propeptide, and proteolytic fragments of CPU. Another important advantage of activity assays compared with antigen assays is that the 2 described polymorphisms in the coding region of proCPU (Thr147Ala and Thr325Ile) have similar activation kinetics and the activated enzymes have similar carboxypeptidase B-like activity toward small synthetic substrates (26)(28). Activity assays can therefore be used to study whether the reported association between the functional polymorphisms and proCPU concentrations is just a laboratory artifact attributable to different reactivities, as suggested by Gils et al. (24), or whether there is indeed a significant association (32). drawbacks proCPU is an inactive zymogen; therefore, accurate measurement of proCPU concentrations by activity-based assays requires quantitative conversion of the zymogen to CPU. Thus, a very well standardized activation protocol is required (18)(19). Because most of the activity assays use synthetic substrates that are not selective for CPU, the interfering activity of plasma CPN must also be accounted for (33). This makes activity-based assays more time-consuming than antigen assays. Major drawbacks of this kind of assay include the intrinsic instability of CPU and the differences in stability among isoforms. CPU has a temperature-dependent instability with a half-life of 8–10 min at 37 °C in both serum and in the purified form, except for the recently described Ile325Ile polymorphism, which has a half-life of 15 min (28)(34). Much evidence points to a structural change that accounts for the irreversible loss of activity (34). Because of this instability, samples should be kept on ice after activation. CPU can be stabilized by the presence of excess substrate. At low substrate concentrations, however, the stabilizing effect of the substrate becomes negligible, and linear substrate conversion occurs during a very short time interval. When the incubation interval of an endpoint method is longer than this linear time interval, a significant fraction of the CPU will decay and the enzymatic velocity and, hence, proCPU concentration will be underestimated (35)(36). These stability problems have been solved in the past by use of very high substrate concentrations (which stabilize the active enzyme) and by validating the linear time interval of substrate conversion, by incubating for short periods of time, or by incubating at room temperature (35)(36). Another possibility for solving this problem is use of a continuously monitored reaction, which enables the determination of accurate initial velocities even at very low substrate concentrations (36)(37). Overview of Currently Used Activity-Based Assays hplc-assisted assay The most sensitive activity-based assay for measuring proCPU is based on the HPLC-assisted quantification of hippuric acid liberated from hippuryl-l-arginine with o-methylhippuric acid as an internal standard (Fig. 22 ) (35). On activation of the plasma samples by T-Tm, hippuryl-l-arginine is added in high excess (24 mmol/L; Km = 0.84 mmol/L) to stabilize the active enzyme during the 30-min incubation at 37 °C, after which the reaction is stopped with 1 mol/L HCl. After addition of the internal standard o-methylhippuric acid, the hippuric acid and internal standard are extracted with ethyl acetate, evaporated to dryness, and measured by HPLC using a C18 column. Although this method is very sensitive, with a detection limit as low as 1 U/L, throughput is lower because it is very time-consuming. However, it is a method of first choice when high sensitivity is required, e.g., for checking the degree of depletion in proCPU-depleted plasma or measuring the generation of CPU during in vitro clot lysis in the absence of Tm. Using this method, Schatteman et al. (35) established a reference interval based on 490 healthy individuals: the mean (SD) proCPU concentration, expressed as CPU activity, was 964 (155) U/L (1 U of enzyme activity was defined as the amount of enzyme required to hydrolyze 1 μmol of hippuryl-l-arginine per minute at 37 °C under the conditions described). As discussed above, the 964 U/L corresponds to a proCPU concentration of 223 nmol/L (13.4 μg/mL). spectrophotometric endpoint activity assays Typical assays for carboxypeptidase activity rely on a change of absorbance, and sensitivity is always an important issue. A widely used activity measurement is carboxypeptidase-mediated release of hippuric acid from either hippuryl-l-arginine or hippuryl-l-lysine with a concomitant increase in absorbance at 254 nm (Fig. 22 ). In an alternative homogeneous assay, the released hippuric acid is chemically converted by cyanuric acid chloride and the absorbance is measured at 382 nm (Fig. 22 ) (38). Because the cyanuric acid chloride reagent is toxic, this method is not widely used. A coupled enzymatic assay has also been described that uses hippuricase as auxiliary enzyme, providing a more sensitive absorbance measurement at 506 nm (Fig. 22 ). The assay is based on the conversion of the zymogen by T-Tm, followed by measurement of the activity with p-hydroxy-hippuryl-l-arginine as substrate. The released p-hydroxyhippuric acid is quantified colorimetrically at 506 nm after enzymatic conversion by hippuricase and oxidation to the quinoneimine dye (39)(40). The authors correlated the assay with the HPLC-assisted assay (r = 0.979; n = 25) and validated the linear interval of substrate conversion. This assay offers an alternative to the precise, but more cumbersome, HPLC-assisted assay in clinical laboratory settings. kinetic spectrophotometric assays Hydrolysis of the substrates hippuryl-l-arginine and hippuryl-l-lysine can be monitored continuously at 254 nm. The extinction coefficients for the 2 substrates and the change in absorptivity that occurs on hydrolysis of the substrates were determined by Boffa et al. (41) and were as follows: ε(Hip-Arg) = 2.149 L · mmol−1 · cm−1; Δε(Hip-Arg→Hippuric acid) = 0.524 L · mmol−1 · cm−1; ε(Hip-Lys) = 2.156 L · mmol−1 · cm−1; Δε(Hip-Lys→Hippuric acid) = 0.517 L · mmol−1 · cm−1. A kinetic spectrophotometric method for the determination of basic carboxypeptidases was described using furyl-acrolyl peptides (42). The principle of this method is that the absorbance maxima of these synthetic peptide substrates shift to the left (ultraviolet) when the C-terminal arginine or lysine is cleaved by the carboxypeptidase. However, the strong background absorbance of the furyl-acrolyl peptide products compromises the sensitivity and accuracy of the assay. An anisylazoformyl chromophore that showed low product absorbance and improved sensitivity has been described (43)(44). Recently, Willemse et al. (36)(37) published a rapid coupled enzymatic assay for measuring CPU activity with C-terminal arginine–containing peptides as substrate. After activation of proCPU by T-Tm, the resulting CPU activity cleaves a substrate with a C-terminal arginine. Arginine kinase phosphorylates the generated arginine, using ATP as a cofactor. The ATP used in this reaction is regenerated by pyruvate kinase, which catalyzes the transfer of the reactive phosphorus in phosphoenolpyruvate to ADP. The final step in this coupled enzymatic assay is the reduction of pyruvate to lactate by lactate dehydrogenase with concomitant oxidation of NADH, which can be monitored continuously at 340 nm. The assay was validated against the HPLC-assisted assay (r = 0.956; n = 56) and showed comparable reproducibility (within- and between day CVs, 2% and 3.9%, respectively) (37). The advantage of kinetic determination of CPU activity is that there is no need to use an excess of substrate. A specific advantage of the kinetic assay described by Willemse et al. (36) is that it can be modified easily to allow the use of other C-terminal-arginine–containing substrates, and thus, alternative substrates (even physiologic) can be used. With this method it is possible to search for very selective CPU substrates so that correction for the interfering CPN activity—an important drawback of activity assays—could be avoided in the future (work in progress). Conclusions In recent years, considerable debate has arisen on the correct measurement of proCPU in plasma; this problem is discussed annually in the Scientific Standardisation Committee Meeting on fibrinolysis of the International Society of Thrombosis and Hemostasis. Plasma proCPU can be measured by antigen (ELISA) or by functional (activity) assays. Activity assays are considered by many groups as a secondary concern because of the prerequisite for quantitative activation of the zymogen and the need to correct for the interference by plasma CPN. In addition, the intrinsic instability of CPU complicates activity assays: samples should be kept on ice after activation, high substrate concentrations should be used to stabilize the enzyme during the assay, and the time interval of linear substrate conversion should be very well validated. Because of its instability, underestimation of CPU activity, and hence the proCPU concentration, is a major concern when using assays that lack adequate validation. It is therefore understandable that many clinical research groups who are not familiar with enzymatic assays prefer the use of ELISAs. Several ELISAs entered the market, and those commercially available ELISAs became the most important tool to analyze proCPU concentrations in clinical samples. Recently however, Gils et al. (24) reported that most of the currently available commercial antigen assays seem to be partially dependent on the naturally occurring Thr325Ile genotype. The findings of Gils et al. (24) compromise the interpretation of several epidemiologic studies evaluating proCPU antigen concentrations as a risk marker for cardiovascular events and force reconsideration of previously made conclusions. For example, a case–control study (using an activity-based assay) of patients with stable angina pectoris showed significantly higher proCPU concentrations in patients compared with controls (45). In contrast, the HIFMECH study group reported that high (above the 90th percentile) proCPU antigen concentrations (measured with a commercial ELISA) may be protective for myocardial infarction (46). Because there is a difference in sensitivity for the various isoforms among assays, ELISAs should be used very carefully and should be very well characterized with respect to genotype-dependent reactivity before they enter the market. In addition, conclusions regarding association of single-nucleotide polymorphisms with proCPU antigen concentrations in plasma should be reconsidered. Standardization of the various assays is another critical issue. The use of different preparations of pooled plasma and purified proCPU as calibrators in the various assays probably contributes considerably to the large variations in the mean plasma concentrations and ranges reported. proCPU concentrations in pooled plasma varied from 73 nmol/L (4.4 μg/mL) to 250 nmol/L (15 μg/mL) (20)(21)(22)(25). In summary, we can state that antigen assays—when they are very well characterized regarding genotype-dependent reactivity—are an easy way to measure proCPU concentrations, whereas activity assays—although they are more time-consuming—are a very good alternative for measuring proCPU. Some research groups showed a good correlation between ELISAs and activity assays (21)(22)(24). The most important advantage of activity assays compared with antigen assays is that the 2 described polymorphisms in the coding region of proCPU (Thr147Ala and Thr325Ile) have similar activation kinetics and the activated enzymes have similar carboxypeptidase B-like activity toward small synthetic substrates (26)(28). In addition, inactivated CPU and enzymatically inactive fragments (e.g., the propeptide and proteolytic fragments) do not interfere in activity assays. A well-standardized activation protocol should be used, and the intrinsic instability of the enzyme should be taken in account. Recently, a kinetic assay was described by our group that allows the use of many different C-terminal-arginine substrates even at very low substrate concentrations (36)(37). This assay could be used to search for more sensitive and selective CPU substrates that may eliminate the need to correct for interfering CPN activity, making activity-based assays more attractive and less time-consuming (36). Figure 1. Open in new tabDownload slide Major proCPU fragments produced by activation. On activation, CPU can be inactivated by a conformational change (t1/2 = 8–15 min at 37 °C). This 35-kDa fragment (iCPU) can then be cleaved at Arg-302 by thrombin (T) or plasmin (Pl), producing 24- and 11-kDa fragments. Antibodies used in ELISAs may not only react with proCPU, but can also recognize CPU, inactivated CPU, fragments of inactivated CPU (24- and 11-kDa), the 20-kDa propeptide (Pro), and some minor cleavage fragments. Figure 1. Open in new tabDownload slide Major proCPU fragments produced by activation. On activation, CPU can be inactivated by a conformational change (t1/2 = 8–15 min at 37 °C). This 35-kDa fragment (iCPU) can then be cleaved at Arg-302 by thrombin (T) or plasmin (Pl), producing 24- and 11-kDa fragments. Antibodies used in ELISAs may not only react with proCPU, but can also recognize CPU, inactivated CPU, fragments of inactivated CPU (24- and 11-kDa), the 20-kDa propeptide (Pro), and some minor cleavage fragments. Figure 2. Open in new tabDownload slide Schematic representation of the different methods to analyze proCPU. proCPU can be measured directly by antigen assays (ELISAs) or, after quantitative activation with T-Tm, with an activity-based assay. Hippuryl-l-arginine is the substrate most commonly used. Activity assays are based on quantification of hippuric acid or arginine by HPLC or spectrophotometry. The instability of CPU and the influence of the Thr325Ile polymorphism on ELISAs and activity assays are also indicated. CPUi, inactivated CPU; Int st, internal standard; Lactate dehydr, lactate dehydrogenase; p-OH, p-hydroxy. Figure 2. Open in new tabDownload slide Schematic representation of the different methods to analyze proCPU. proCPU can be measured directly by antigen assays (ELISAs) or, after quantitative activation with T-Tm, with an activity-based assay. Hippuryl-l-arginine is the substrate most commonly used. Activity assays are based on quantification of hippuric acid or arginine by HPLC or spectrophotometry. The instability of CPU and the influence of the Thr325Ile polymorphism on ELISAs and activity assays are also indicated. CPUi, inactivated CPU; Int st, internal standard; Lactate dehydr, lactate dehydrogenase; p-OH, p-hydroxy. 1 Nonstandard abbreviations: CPU, carboxypeptidase U; proCPU, procarboxypeptidase U; TAFI, thrombin-activable fibrinolysis inhibitor; CPN, carboxypeptidase N; and T-Tm, thrombin–thrombomodulin complex. We gratefully acknowledge Yani Sim for excellent technical assistance and Robert Verkerk for many fruitful discussions. J.W. is a research assistant of the Fund for Scientific Research Flanders (FWO-Vlaanderen). 1 Hendriks D, Scharpé S, van Sande M. Partial characterization of a new arginine carboxypeptidase from human serum. J Clin Chem Clin Biochem 1988 ; 26 : 305 . 2 Hendriks D, Scharpé S, van Sande M, Lommaert MP. Characterisation of a carboxypeptidase in human serum distinct from carboxypeptidase N. J Clin Chem Clin Biochem 1989 ; 27 : 277 -285. 3 Hendriks D, Scharpé S, van Sande M, Lommaert MP. A labile enzyme in fresh human serum interferes with the assay of carboxypeptidase N. [Technical Brief] Clin Chem 1989 ; 35 : 177 . 4 Hendriks D, Wang W, van Sande M, Scharpé S. Human serum carboxypeptidase U: a new kininase?. Agents Actions Suppl 1992 ; 38 : 407 -413. 5 Shinohara T, Sakurada C, Suzuki T, Takeuchi O, Campbell W, Ikeda S, et al. Pro-carboxypeptidase R cleaves bradykinin following activation. Int Arch Allergy Immunol 1994 ; 103 : 400 -404. 6 Kato K, Shinagawa N, Hayakawa T, Takeyama H, Campbell W, Okada H. Changes in arginine carboxypeptidase (CPR) activity in stressed rats. Pathophysiology 1994 ; 1 : 131 -136. 7 Wang W, Boffa MB, Bajzar L, Walker J, Nesheim ME. A study of the mechanism of inhibition of fibrinolysis by activated thrombin-activable fibrinolysis inhibitor. J Biol Chem 1998 ; 273 : 27176 -27181. 8 Sakharov DV, Plow EF, Rijken DC. On the mechanism of the antifibrinolytic activity of plasma carboxypeptidase B. J Biol Chem 1997 ; 272 : 14477 -14482. 9 Schneider M, Nesheim M. A study of the protection of plasmin from antiplasmin inhibition within an intact fibrin clot during the course of clot lysis. J Biol Chem 2004 ; 279 : 13333 -13339. 10 Schneider M, Brufatto N, Neill E, Nesheim M. Activated thrombin-activatable fibrinolysis inhibitor reduces the ability of high molecular weight fibrin degradation products to protect plasmin from antiplasmin. J Biol Chem 2004 ; 279 : 13340 -13345. 11 Leurs J, Nerme V, Sim Y, Hendriks D. Carboxypeptidase U (TAFIa) prevents lysis from proceeding into the propagation phase through a threshold-dependent mechanism. J Thromb Haemost 2004 ; 2 : 416 -423. 12 Walker JB, Bajzar L. The intrinsic threshold of the fibrinolytic system is modulated by basic carboxypeptidases, but the magnitude of the antifibrinolytic effect of activated thrombin activable fibrinolyis inhibitor is masked by its instability. J Biol Chem 2004 ; 279 : 27896 -27904. 13 Sato T, Miwa T, Akatsu H, Matsukawa N, Obata K, Okada N, et al. Procarboxypeptidase R is an acute phase protein in mouse, whereas carboxypeptidase N is not. J Immunol 2000 ; 165 : 1053 -1058. 14 Boffa MD, Hamill JD, Maret D, Brown D, Scott ML, Nesheim ME, et al. Acute phase mediators modulate thrombin-activable fibrinolysis inhibitor (TAFI) gene expression in Hep G2 cells. J Biol Chem 2003 ; 278 : 9250 -9257. 15 Campbell WD, Lazoura E, Okada N, Okada H. Inactivation of C3a and C5a octapeptides by carboxypeptidase R and carboxypeptidase N. Microbiol Immunol 2002 ; 46 : 131 -134. 16 Campbell W, Okada N, Okada H. Carboxypeptidase R is an inactivator of complement-derived inflammatory peptides and an inhibitor of fibrinolysis. Immunol Rev 2001 ; 180 : 162 -166. 17 Myles T, Nishimura T, Yun TH, Nagashima M, Morser J, Patterson AJ, et al. Thrombin activatable fibrinolysis inhibitor, a potential regulator of vascular inflammation. J Biol Chem 2003 ; 278 : 51059 -51067. 18 Bajzar L, Morser J, Nesheim M. TAFI, or plasma procarboxypeptidase B, couples the coagulation and fibrinolytic cascades through the thrombin-thrombomodulin complex. J Biol Chem 1996 ; 271 : 16603 -16608. 19 Schatteman K, Goossens F, Scharpé S, Hendriks D. Proteolytic activation of purified procarboxypeptidase U. Clin Chim Acta 2000 ; 292 : 25 -40. 20 Bajzar L, Nesheim ME, Tracy PB. The profibrinolytic effect of activated protein C in clots formed from plasma is TAFI dependent. Blood 1996 ; 88 : 2093 -2100. 21 Mosnier LO, von-dem-Borne PA, Meijers JC, Bouma BN. Plasma TAFI levels influence the clot lysis time in healthy individuals in the presence of an intact intrinsic pathway of coagulation. Thromb Haemost 1998 ; 80 : 829 -835. 22 Stromqvist M, Schatteman K, Leurs J, Verkerk R, Andersson JO, Johansson T, et al. Immunological assay for the determination of procarboxypeptidase U antigen levels in human plasma. Thromb Haemost 2001 ; 85 : 12 -17. 23 Guo X, Morioka A, Kaneko Y, Okada N, Obata K, Nomura T, et al. Arginine carboxypeptidase (CPR) in human plasma determined with sandwich ELISA. Microbiol Immunol 1999 ; 43 : 691 -698. 24 Gils A, Alessi MC, Brouwers E, Peeters M, Marx P, Leurs J, et al. Development of a genotype 325-specific proCPU/TAFI ELISA. Arterioscler Thromb Vasc Biol 2003 ; 23 : 1122 -1127. 25 Van Tilburg NH, Rosendaal FR, Bertina RM. Thrombin activatable fibrinolysis inhibitor and the risk for deep vein thrombosis. Blood 2000 ; 95 : 2855 -2859. 26 Zhao L, Morser J, Bajzar L, Nesheim M, Nagashima M. Identification and characterization of two thrombin activatable fibrinolysis inhibitor isoforms. Thromb Haemost 1998 ; 80 : 949 -955. 27 Brouwers GJ, Vos HL, Leebeek FW, Bulk S, Schneider M, Boffa M, et al. A novel, possibly functional, single nucleotide polymorphism in the coding region of the thrombin-activatable fibrinolysis inhibitor (TAFI) gene is also associated with TAFI levels. Blood 2001 ; 98 : 1992 -1993. 28 Schneider M, Boffa M, Stewart R, Rahman M, Koschinsky M, Nesheim M. Two naturally occurring variants of TAFI (Thr-325 and Ile-325) differ substantially with respect to thermal stability and antifibrinolytic activity of the enzyme. J Biol Chem 2002 ; 277 : 1021 -1030. 29 Franco RF, Fagundes MG, Meijers JM, Reitsma PH, Lourenco D, Morelli V, et al. Identification of polymorphisms in the 5′ untranslated region of the TAFI gene: relationship with TAFI levels and risk of venous thrombosis. Haematologica 2001 ; 86 : 510 -517. 30 Henry M, Aubert H, Morange PE, Nanni I, Alessi MC, Tiret L, et al. Identification of polymorphisms in the promoter and the 3′ region of the TAFI gene: evidence that plasma TAFI antigen levels are strongly genetically controlled. Blood 2001 ; 97 : 2053 -2058. 31 Tani S, Akatsu H, Ishikawa Y, Okada N, Okada H. Preferential detection of pro-carboxypeptidase R by enzyme-linked immunosorbent assay. Microbiol Immunol 2003 ; 47 : 295 -300. 32 Guimaraes AHC, van Tilburg NH, Vos HL, Bertina RM, Rijken DC. Association between thrombin activatable fibrinolysis inhibitor genotype and levels in plasma: comparison of different assays. Br J Haematol 2004 ; 124 : 659 -665. 33 Hendriks D, Scharpé S, van Sande M. Assay of carboxypeptidase N activity in serum by liquid chromatographic detection of hippuric acid. Clin Chem 1985 ; 12 : 1936 -1939. 34 Boffa MB, Bell R, Stevens WK, Nesheim ME. Roles of thermal instability and proteolytic cleavage in regulation of thrombin-activable fibrinolysis inhibitor. J Biol Chem 2000 ; 275 : 12868 -12878. 35 Schatteman K, Goossens F, Scharpé S, Hendriks D. Assay of procarboxypeptidase U, a novel determinant of the fibrinolytic cascade, in human plasma. Clin Chem 1999 ; 45 : 807 -813. 36 Willemse J, Leurs J, Verkerk R, Hendriks D. Development of a fast kinetic method for the determination of carboxypeptidase U using C-terminal arginine containing peptides as substrate. Anal Biochem 2005 ; 340 : 106 -112. 37 Willemse JL, Leurs JR, Hendriks DF. Fast kinetic assay for the determination of procarboxypeptidase U (TAFI) in human plasma. J Thromb Haemost 2005 ; 3 : 2353 -2355. 38 Hendriks D, Van Sande M, Scharpé S. Colorimetric assay for carboxypeptidase N in serum. Clin Chim Acta 1986 ; 157 : 103 -108. 39 Hendriks D, Scharpé S, van Sande M, Lommaert MP, Kasahara Y. Carboxypeptidase N: colorimetric assay using a new substrate. Anal Biochem 1987 ; 164 : 90 -95. 40 Schatteman K, Goossens F, Leurs J, Kasahara Y, Scharpé S, Hendriks D. Fast and homogeneous assay for plasma procarboxypeptidase U. Clin Chem Lab Med 2001 ; 39 : 806 -810. 41 Boffa MB, Wang W, Bajzar L, Nesheim ME. Plasma and recombinant thrombin activable fibrinolysis inhibitor (TAFI) and activated TAFI compared with respect to glycosylation, thrombin/thrombomodulin dependent activation, thermal stability and enzymatic properties. J Biol Chem 1998 ; 273 : 2127 -2135. 42 Plummer TH, Kimmel MT. An improved spectrophotometric assay for human carboxypeptidase N. Anal Biochem 1980 ; 108 : 348 -353. 43 Mock WL, Xu D. Catalytic activity of carboxypeptidase B and of carboxypeptidase Y with anisylazoformyl substrates. Bioorg Med Chem Lett 1999 ; 9 : 187 -192. 44 Mock WL, Stanford DJ. Anisylazoformylarginine: a superior assay substrate for carboxypeptidase B type enzymes. Bioorg Med Chem Lett 2002 ; 12 : 1193 -1194. 45 Silveira A, Schatteman K, Goossens F, Moor E, Scharpé S, Stromqvist M, et al. Plasma procarboxypeptidase U in men with symptomatic coronary artery disease. Thromb Haemost 2000 ; 84 : 364 -368. 46 Juhan-Vague I, Morange PE, Aubert H, Henry M, Aillaud MF, Alessi MC, et al. Plasma thrombin-activatable fibrinolysis inhibitor antigen concentration and genotype in relation to myocardial infarction in the north and south of Europe. Arterioscler Thromb Vasc Biol 2002 ; 22 : 867 -873. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Li, Sam, FY;Kricka, Larry, J

doi: 10.1373/clinchem.2005.059600pmid: 16299048

Abstract Clinical analysis often requires rapid, automated, and high-throughput analytical systems. Microchip capillary electrophoresis (CE) has the potential to achieve very rapid analysis (typically seconds), easy integration of multiple analytical steps, and parallel operation. Although it is currently still in an early stage of development, there are already many reports in the literature describing the applications of microchip CE in clinical analysis. At the same time, more fully automated and higher throughput commercial instruments for microchip CE are becoming available and are expected to further enhance the development of applications of microchip CE in routine clinical testing. To put into perspective its potential, we briefly compare microchip CE with conventional CE and review developments in this technique that may be useful in diagnosis of major diseases. Since the development of capillary electrophoresis on a chip (microchip CE)1 in the early 1990s (1)(2), this technology has been a focus of research in chemical and biochemical analysis and has been reviewed extensively (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32). Potential advantages of microchip CE include miniaturization, integration, high speed, and reduced reagent consumption. We present an overview of the approaches and selected applications of microchip CE devices in the diagnosis of major diseases, including cancer (33)(34)(35)(36)(37)(38)(39)(40)(41); cardiovascular (42)(43)(44)(45), renal (46)(47), neurologic (48)(49)(50), thyroid (51)(52), and infectious diseases (53)(54)(55)(56)(57)(58)(59)(60)(61)(62)(63)(64); immune disorders(65)(66)(67)(68)(69); diabetes (70)(71)(72); and hereditary diseases (73)(74)(75)(76)(77). We also briefly compare microchip CE with conventional CE and speculate on the future role of microchip CE in the clinical laboratory. Commercial systems for microchip CE analysis have recently become available and are increasingly used for routine analysis. The introduction of more automated and higher throughput systems will likely make microchip CE technology more widely accepted in clinical analysis laboratories. Several companies are now supplying ready-made and/or custom-fabricated microchips, which will facilitate application of microchip CE devices. Some suppliers of commercial microchip CE systems and foundries for fabricating microchip CE devices are listed in Table 11 . Approaches and selected applications of microchip CE for diagnosis of major diseases are described in the following sections. Cancer Several approaches have been developed for the analysis of cancer susceptibility genes by microchip CE, including single-strand conformation polymorphism analysis (33) and a combination of allele-specific DNA amplification with heteroduplex analysis (34)(35)(36)(37). Common mutations in the BRCA1 and BRCA2 genes show strong correlations with breast cancer, particularly in the Ashkenazi Jewish population (33). Using microchip CE, Landers’ group decreased single-strand conformation polymorphism analysis time to ∼130 s, at least a 100-fold improvement over conventional methods (33)(34)(35)(36)(37). Another approach developed by the same group combines allele-specific DNA amplification with heteroduplex analysis for the detection of each mutation in the BRCA1 and BRCA2 genes (34)(35)(36)(37). A schematic diagram of the system used is shown in Fig. 11 , and results obtained for detection of mutations of BRCA1 are shown in Fig. 22 . The system was also used in assays for T- and B-cell lymphoproliferative disorders (38), and microchip CE was found to provide the same information as slab gel electrophoresis and conventional CE, but with much shorter analysis time (160 s vs 2.5 h on slab gel and 15 min on capillary CE). Thomas et al. (39) performed microchip CE in uncoated polymer-based microchannels filled with various separation matrices for rapid analysis of ligase detection reaction products of low-abundance point mutations in genomic DNA. Diagnostic testing for point mutations in the human k-ras oncogene was performed on DNA obtained from colorectal tumors. The ligase detection reaction products were resolved from unligated primers in <120 s, ∼17 times faster than capillary gel electrophoresis, with only a slight decrease in resolution. Cantafora et al. (40) evaluated RNA messengers involved in lipid trafficking of human intestinal cells by reverse transcription-PCR with competimer technology and microchip CE. Their results showed that analysis of specific RNA messengers allows reliable evaluation of relative gene expression in CaCo-2 cells and confirmed the role of cholesterol as a positive inducer of specific factors such as LXR-α and FXR. In a microchip-based enzyme assay for protein kinase A (41), the assay reagents were mixed in etched channels by electroosmotic flow. A fluorescein-labeled heptapeptide (LeuArgArgAlaSerLeuGly) was used as substrate for the enzymatic reaction. Substrate and the products of the reaction were separated electrophoretically, and the results demonstrated the usefulness of microchip CE for performing enzymatic assays for which a fluorogenic substrate is not readily available (41). Cardiovascular and Related Diseases Microchip CE has been investigated as a diagnostic tool for assessment of arteriosclerosis via analysis of LDL (42)(43) and homocysteine (44). In addition, microchip CE has been used for the analysis of DNA fragmentation in cells to evaluate apoptosis in individual cardiomyocytes for the diagnosis of doxorubicin-induced cardiomyopathy, a life-threatening condition in patients undergoing chemotherapy (45). For LDL analysis, Ceriotti et al. (42) used microchip CE with either glass microchannels dynamically coated with 40 mmol/L methylglucamine to prevent lipoprotein adsorption or uncoated glass microchannels with 0.3 mmol/L sodium dodecyl sulfate added to the sample buffer to sharpen the LDL peak. A microchip CE–based method with electrochemical detection has been developed for the analysis of total and protein-bound homocysteine (44), which should be routinely tested in patients at risk for cardiovascular disease, according to the American Heart Association (44). Typical results obtained for the detection and separation of homocysteine and reduced glutathione by microchip CE are shown in Fig. 33 . Kleparnik et al. (45) used a compact disk–like microfluidic device to perform cell lysis, electrophoresis, and laser-induced fluorescence detection of doxorubicin-induced apoptosis of individual cardiomyocytes in the microchip device. Apoptosis was detected by analysis of DNA fragmentation in cells treated with doxorubicin for different durations. The results showed that prolonged exposure of cardiomyocytes to doxorubicin was associated with cellular necrosis (45). Renal Markers Analysis of renal markers such as creatinine, creatine, uric acid, urea, and p-aminohippuric acid in biological fluids allows evaluation of renal and muscular functions (46)(47). Microchip CE analysis of renal markers has been integrated with electrochemical detection (46)(47). One method was based on coupling of enzymatic bioassays using creatininase, creatinase, and sarcosine oxidase and electrophoretic separation and amperometric detection of reaction products (46). Gold-coated thick-film electrodes were used for the amperometric detection, and detection limits of 2 × 10−5 to 4 × 10−5 mol/L (signal-to-noise ratio = 3) were reported. An alternative method enabled direct measurement of the renal markers by pulsed amperometry for the detection of nitrogen-containing compounds and more easily oxidized uric acid (47). Four renal markers (creatine, creatinine, p-aminohippuric acid, and uric acid) were readily measured within 5 min. A creatinine:creatine ratio was determined in <2 min (compared with ∼20 min with conventional creatinine analysis methods). The microchip CE device allowed rapid, simple, and economical renal function testing. These devices have substantial potential advantages for decentralized clinical testing and point-of-care testing. Further developments to integrate additional functions (e.g., on-chip filtering of biological fluids) to analyze serum samples, and to perform parallel and multiple runs on a single-chip platform could enhance the acceptance of microchip CE for use in high-throughput clinical microanalyzers. Neurologic Diseases Microchip CE has been used to analyze cerebral spinal fluid (CSF) samples for inflammatory cytokines and inhibitors of an enzyme for nerve function (48)(49)(50). Lapos et al.(48) demonstrated the applicability of a microchip CE device with dual laser-induced fluorescence (LIF) and electrochemical detectors for the analysis of CSF samples from patients with multiple sclerosis (48). Simultaneous LIF detection of 4-chloro-7-nitrobenzofurazan–derivatized amino acids (Arg, Phe, and Glu) and electrochemical detection of dopamine and catechol were demonstrated. Microfluidic assays for inhibitors of acetylcholinesterase, an enzyme essential for proper nerve function, have also been developed (49). Separation and detection of a mixture of 4 cationic inhibitors—tacrine, edrophonium, and tetramethyl- and tetraethyl-ammonium chloride—was accomplished within 70 s with a multiplexed screening device combining flow injection analysis and microchip CE separation. Chip-based immunoaffinity isolation has been combined with CE for the rapid analysis of inflammatory cytokines in CSF (50). A panel of 6 immobilized antibodies was attached to the injection port of the chip to isolate the reactive cytokines from CSF samples obtained from patients with traumatic head injury. Thyroid Function Schmalzing et al. (51) described a competitive immunoassay for the determination of serum thyroxine (3,5,3′,5′-tetraiodo-l-thyronine) based on electrophoretic separation in fused-silica microchips and LIF detection. Analysis speed was substantially faster than conventional immunoassay and electrophoresis in capillaries, with separation of free from bound labeled thyroxine in ∼15 s for serum samples. A microchip-based amperometric competitive immunoassay using a ferrocene redox label has been developed for the determination of 3,3′,5-triiodo-l-thyronine (52). The assay consisted of on-chip, precolumn reaction of the labeled antigen and the target antigen with the antibody, electrophoretic separation of the free and bound labeled antigen, and amperometric detection of the redox tag. A minimum detectable concentration of 1000 μg/L and an analysis time of 130 s were achieved. Although the above protocols were developed for the testing of thyroid function, integration of immunologic reactions, electrophoretic separation, and a sensitive detection system on a microchip platform enables rapid immunoassay for a wide range of analytes. Infectious Diseases/Pathogens High-speed, high-efficiency microchip CE devices are potentially powerful tools for the detection of pathogens and diagnosis of infectious diseases, an important function because of the acute nature of certain infectious diseases and the growing threat of bioterrorism. Analysis of specimens for the presence of viruses (53)(54)(55)(56)(57)(58) and bacteria (59)(60)(61)(62)(63)(64) by microchip CE has been demonstrated. To date, most of these studies have focused on the analysis of extracted and PCR-amplified nucleic acid fragments rather than on intact viral or bacterial particles. viral infections Zhou et al. (53) developed a microfluidic system for the determination of severe acute respiratory syndrome coronavirus. For nasopharyngeal swabs from patients with a clinical diagnosis of severe acute respiratory syndrome, this system provided higher positivity rates more rapidly than conventional reverse transcription PCR (17 of 18 vs 12 of 18 positive identifications). Chen’s group analyzed hepatitis C virus qualitatively (54) and quantitatively (55). For qualitative analysis, they used a plastic (polymethylmethacrylate) microchip for the analysis of products from a 2-stage PCR with 2 pairs of primers designed to amplify the 5′ noncoding region. The microchip CE device could resolve the 145-bp amplicon of hepatitis C virus in <1.5 min (54). For quantitative analysis, reverse transcription–competitive PCR was used (55). Co-reverse transcription and coamplification were performed for wild-type RNA extracted from serum with a constant amount of recombinant internal standard RNA with the same primer binding region as the target template except for the removal of a centrally located 25-bp segment specific to the target RNA. Results were comparable to those obtained with a commercial hybridization assay. The major advantage of the microchip CE assay was that it was less labor-intensive than the hybridization-based detection method. Diagnosis of herpes simplex encephalitis can be performed by analysis of PCR products of DNA extracted from CSF. Hofgartner et al. (56) used microchip CE to analyze archival DNA from 33 CSF specimens submitted for herpes simplex virus PCR testing. Microchip CE achieved 100% sensitivity and specificity with much shorter total analysis time than established methods (<100 s/sample vs 18 h for routine clinical liquid hybridization/gel retardation assay). Because the hybridization step is eliminated, the real-time, quantitative, fluorescence-based PCR assay also has the advantage of rapid analysis, but expensive instrumentation is required for this approach. Microchip CE-based technology has the advantage of versatility because it can detect a variety of fluorescently labeled clinical markers and integrate additional laboratory functions in addition to separation and detection steps. Vegvari et al. (57) developed a hybrid microdevice based on a combination of a polyvinylchloride supporting plate and a fused-silica capillary fitted into a U-groove on the plate for the analysis of various samples, including peptides, proteins, DNA, viruses, and bacteria (e.g., Semliki Forest virus). Because of the high optical transparency of the fused-silica capillary, ultraviolet detection can be used, and this type of device is expected to be useful for a broad range of applications (57). CE devices for DNA analysis have also been microfabricated by compression molding of polycarbonate (58). DNA separation in these devices provided good resolution and run-to-run reproducibility. In addition, on-chip PCR-CE of a 500-bp region of bacteriophage lambda DNA was demonstrated by thermally cycling the entire chip, with the sample reservoir of the CE device serving as the PCR chamber (58). bacterial infections Microchip devices integrating PCR with CE have been used for the analysis of bacteria such as Escherichia coli (59)(60)(61)(62)(63), Staphylococcus(59), Salmonella (62), and Streptococcus (64). Lagally et al.(59) developed an integrated portable genetic analysis system for detection of pathogens. The results obtained for the analysis of DNA from intact cells of different strains of E. coli are shown in Fig. 44 . The system could be used to detect the serotype and pathogenic status of a given cell population simultaneously and to detect antibiotic resistance (59). Other integrated microdevices have been developed for cell lysis, multiplex PCR amplification, and electrophoretic sizing of PCR products with a marker (60). For example, plastic microfluidic devices have been fabricated that integrate PCR, microfluidic valving, and electrophoresis for bacterial detection and identification (62). Amplicons generated in the plastic device from PCR reactions with genomic DNA from E. coli (232 bp) and Salmonella (429 and 539 bp) were successfully analyzed. Woolley et al. (63) used an integrated PCR-CE device for rapid assay of genomic Salmonella DNA. The entire assay was accomplished in <45 min, including both the PCR and the CE separation steps. Microchip CE detection of cariogenic bacterial genes has also been achieved (64). Allele-specific PCR primers were designed based on the dextranase gene to identify Streptococcus mutans and Streptococcus sorbrinus in dental plaque. A polymer mixture consisting of hydroxypropyl methylcellulose and polyethylene oxide served as the separation medium for microchip CE. Rapid (85 s), precise (CV = 0.3%; n = 7), high-resolution (resolution = 2.67 for 226 bp/202 bp), and sensitive (10- to 100-fold better than agarose gel electrophoresis) analysis was achieved. Immune Disorders Several microchip CE methods have been developed to detect increased concentrations of IgG, which may be associated with chronic infection (polyclonal increases) or cancer (monoclonal increase) (65)(66)(67)(68). Linder et al.(65) developed a heterogeneous competitive immunoassay of human IgG that used Cy5-human IgG as tracer and Cy3-mouse IgG as internal standard. Quantification of human IgG in serum was difficult because of the high relative SD at a low human IgG concentration and the weak concentration dependence at high IgG concentrations. Nevertheless, it was possible to unambiguously distinguish patient serum samples with increased IgG (35.5 g/L in one patient with chronic infection and 64.7 g/L in another patient with myeloma) from those with IgG concentrations within the reference interval (8–16 g/L). Conventional CE with a short capillary and a glass microchip CE device has been used to analyze fluorescein isothiocyanate–labeled anti-human IgG (66). The microchip device had several advantages, including high efficiency, fast analysis time, and low requirements for samples and solvents. Wang et al. (67) described an electrochemical enzyme immunoassay on microchip platforms. Precolumn reactions of alkaline phosphatase–labeled antibody with antigen, electrophoretic separation of the free antibody and antibody-antigen complex, postcolumn reaction of the enzyme tracer with the 4-aminophenyl phosphate substrate, and amperometric detection of the liberated 4-aminophenol product were performed on the same device. A very low detection limit [1.7 × 10−18 mol/L (2.5 × 10−16 g/mL)] was reported for the model analyte (mouse IgG). In another study, direct measurement of antibodies and monitoring of immunologic interactions was achieved with a microchip CE system with contactless conductivity detection (68). With a glass microchip CE device, separation and detection of IgG (anti-IgM), IgM, and the complex were performed simultaneously. Diabetes Insulin and glucose analyses are important in the diagnosis of several conditions, including diabetes, pancreatic islet cell malfunction, hypoglycemia, and insulinoma. Several studies have been devoted to the use of microchip CE devices for the measurement of insulin (69)(70) and glucose (69)(71). Dual immunologic and enzymatic microchip-based assays for simultaneous measurements of insulin and glucose have been described by Wang et al. (69). Insulin immunodetection was performed with alkaline phosphatase–labeled antibody with postcolumn addition of p-nitrophenylphosphate substrate, and glucose analysis with glucose dehydrogenase and NAD+. A microfluidic chip has been developed for continuous electrophoresis-based immunoassay monitoring of hormone secretion from live cells (70). Insulin secreted from islets of Langerhans was detected in a CE competitive immunoassay. Insulin secretion profiles could be obtained, and characteristics of first- and second-phase insulin secretion could be observed. Microchip CE systems are highly suited for monitoring the chemical environment of live cells with high temporal resolution, and such devices may be used for cell-based sensing and diagnostic systems in routine clinical laboratories. Du et al. (71) described a microchip CE device for electrophoretic detection of glucose in human plasma. Separation and injection channels were fabricated on a poly(dimethylsiloxane) layer. Copper microelectrodes were fabricated on the electrode plate by selective electrodeless deposition. In pilot studies, glucose in human plasma from 3 healthy individuals and 2 individuals with diabetes was successfully determined. Hereditary Diseases Microchip CE has been used for the analysis of genetic diseases such as Duchenne muscular dystrophy (DMD) (72)(73)(74) and hemochromatosis (75), and for genomic DNA analysis (76). DMD is caused by mutations in the dystrophin gene on the X chromosome. Carriers of the disease are identified by the detection of duplicated or deleted exons in the gene (72)(73)(74). The deletions/duplications associated with DMD tend to be located at certain regions of the gene, making the diagnosis of DMD (and the related, less severe Becker muscular dystrophy) a straightforward process involving analysis of a limited number of PCR-amplified DNA fragments (72)(73)(74). An integrated microdevice for infrared-mediated PCR amplifications directly coupled to microchip CE DNA separations was developed for this purpose. In this approach, infrared radiation directly heats the PCR mixture rather than heating the microchip (72), allowing fast temperature cycling and shorter analysis time. Unfortunately, separation efficiency was hampered by heating of the microchip sieving matrix buffer solution during PCR cycling. Because heating is integral to PCR, this problem must be solved before this approach can become useful. Nevertheless, PCR-CE provided considerable savings in time, labor, and materials compared with traditional methods (e.g., Southern blot). Solution of the heating problem could make feasible a fully integrated diagnostic device for genetic diseases based on microchip CE. Ertl et al. (75) developed a sheath-flow supported electrochemical detection system for microchip CE analysis of an allele-specific, PCR-based single-nucleotide polymorphism typing assay and demonstrated the application of the system for diagnosis of hereditary hemochromatosis by detection of the C282Y polymorphism. The sheath-flow design minimized interferences of the detection system from electrophoresis potentials. Unlike optical detection systems, the electrochemical detector is easier to miniaturize because bulky optical components (i.e., light source, lens, and filters) are not required. The use of multiple electrodes may make this system a portable micro total-analysis system for analyzing complex analyte mixtures. A semiautomated sample preparation, amplification, and electrophoretic separation platform has also been developed for analysis of human genomic DNA to detect hereditary and infectious diseases based on microchannel CE with 2-color optical detection (excitation wavelengths at 488 and 532 nm). Two-base pair resolution of single-stranded DNA was achieved in the analysis of PCR products from leukocyte lysates (76). Comparison of Microchip CE and Conventional CE Conventional CE revolutionized DNA analysis and was vital to the success of the Human Genome Project (77)(78). Introduction of microchip CE technology represents another major step in the development of miniaturized, rapid, automated, and integrated analytical systems with potential to meet the requirements of lower cost, faster, more sensitive, and more selective analytical systems to solve complex clinical analysis problems. The most commonly available features of conventional and microchip CE are compared in Table 22 . Microchip CE has some disadvantages compared with conventional CE, such as lower peak capacity because of the shorter separation channels and a lack of compatibility with versatile ultraviolet detectors because microchip fabrication materials are typically not transparent to ultraviolet light. Important advantages of microchip CE, however, include rapid analysis (typically ∼10 times faster) and easy integration of sample preparation and derivatization steps, which allow further reduction of analysis time and labor costs. Because cost and speed are crucial concerns for modern clinical analysis, these advantages are expected to become more important when new generations of commercial microchip CE systems become available, and eventually microchip CE may overtake conventional CE as the technology continues to mature. Future Prospects Although microchip CE is still in the early stages of development, the technique has already shown applicability in many areas of clinical analysis. Compared with conventional techniques, microchip CE–based molecular diagnostic methods have demonstrated advantages in terms of analysis speed, cost savings, and detection sensitivity. As this review has shown, there is already a rapidly growing collection of new applications based on microchip CE. At the same time, many on-column preconcentration methods have been developed to improve detection sensitivity in CE separations that are readily transferable to the microchip CE format (79)(80). With the introduction of highly automated, high-throughput commercial instrumentation, microchip CE is likely to replace many of the complex and slower analytical systems used in routine analyses. Further improvements in automation and an increase in sample throughput along with development of new testing protocols and enhancement of detection sensitivity for certain analytes could make microchip CE systems key instruments for clinical analyses. Table 1. Suppliers of microchip CE instrumentation and foundries for microchip CE devices. Suppliers/Foundries . Website . Microchip CE instrumentation . Suppliers  Agilent (Palo Alto, CA) www.agilent.com Bioanalyzer 2100: microchip CE platform for analysis of DNA, RNA, proteins, and cells 5100 ALP: automated lab-on-a-chip platform handling up to 12-well plates  Caliper Life Science (Hopkinton, MA) www.caliperls.com Labchip automated electrophoresis systems  CE Resources (Singapore, Republic of Singapore) www.ce-resources.com Modular systems for microchip CE: multichannel HV power supply (HVPS) and different detectors  Hitachi (Tokyo, Japan) www.hitachi.com.jp SV110: microchip CE with LED confocal fluorescence detector  Shimadzu (Kyoto, Japan) www.shimadzu.com.jp MCE-2010: quartz microchip CE–based system with photodiode array detector Foundries  AMIC (Uppsala, Sweden) www.amic.se  Bartels Mikrotechnik (Dortmund, Germany) www.bartels-mikrotechnik.de  Epigem (Cleveland, UK) http://www.epigem.co.uk  Micralyne (Edmonton, Canada) http://www.micralyne.com/index.html  Microfluidic Chip Shop (Jena, Germany) http://www.microfluidic-chipshop.com  Scandinavian Micro Biodevices (Lyngby, Denmark) http://www.smb.dk Suppliers/Foundries . Website . Microchip CE instrumentation . Suppliers  Agilent (Palo Alto, CA) www.agilent.com Bioanalyzer 2100: microchip CE platform for analysis of DNA, RNA, proteins, and cells 5100 ALP: automated lab-on-a-chip platform handling up to 12-well plates  Caliper Life Science (Hopkinton, MA) www.caliperls.com Labchip automated electrophoresis systems  CE Resources (Singapore, Republic of Singapore) www.ce-resources.com Modular systems for microchip CE: multichannel HV power supply (HVPS) and different detectors  Hitachi (Tokyo, Japan) www.hitachi.com.jp SV110: microchip CE with LED confocal fluorescence detector  Shimadzu (Kyoto, Japan) www.shimadzu.com.jp MCE-2010: quartz microchip CE–based system with photodiode array detector Foundries  AMIC (Uppsala, Sweden) www.amic.se  Bartels Mikrotechnik (Dortmund, Germany) www.bartels-mikrotechnik.de  Epigem (Cleveland, UK) http://www.epigem.co.uk  Micralyne (Edmonton, Canada) http://www.micralyne.com/index.html  Microfluidic Chip Shop (Jena, Germany) http://www.microfluidic-chipshop.com  Scandinavian Micro Biodevices (Lyngby, Denmark) http://www.smb.dk Table 1. Suppliers of microchip CE instrumentation and foundries for microchip CE devices. Suppliers/Foundries . Website . Microchip CE instrumentation . Suppliers  Agilent (Palo Alto, CA) www.agilent.com Bioanalyzer 2100: microchip CE platform for analysis of DNA, RNA, proteins, and cells 5100 ALP: automated lab-on-a-chip platform handling up to 12-well plates  Caliper Life Science (Hopkinton, MA) www.caliperls.com Labchip automated electrophoresis systems  CE Resources (Singapore, Republic of Singapore) www.ce-resources.com Modular systems for microchip CE: multichannel HV power supply (HVPS) and different detectors  Hitachi (Tokyo, Japan) www.hitachi.com.jp SV110: microchip CE with LED confocal fluorescence detector  Shimadzu (Kyoto, Japan) www.shimadzu.com.jp MCE-2010: quartz microchip CE–based system with photodiode array detector Foundries  AMIC (Uppsala, Sweden) www.amic.se  Bartels Mikrotechnik (Dortmund, Germany) www.bartels-mikrotechnik.de  Epigem (Cleveland, UK) http://www.epigem.co.uk  Micralyne (Edmonton, Canada) http://www.micralyne.com/index.html  Microfluidic Chip Shop (Jena, Germany) http://www.microfluidic-chipshop.com  Scandinavian Micro Biodevices (Lyngby, Denmark) http://www.smb.dk Suppliers/Foundries . Website . Microchip CE instrumentation . Suppliers  Agilent (Palo Alto, CA) www.agilent.com Bioanalyzer 2100: microchip CE platform for analysis of DNA, RNA, proteins, and cells 5100 ALP: automated lab-on-a-chip platform handling up to 12-well plates  Caliper Life Science (Hopkinton, MA) www.caliperls.com Labchip automated electrophoresis systems  CE Resources (Singapore, Republic of Singapore) www.ce-resources.com Modular systems for microchip CE: multichannel HV power supply (HVPS) and different detectors  Hitachi (Tokyo, Japan) www.hitachi.com.jp SV110: microchip CE with LED confocal fluorescence detector  Shimadzu (Kyoto, Japan) www.shimadzu.com.jp MCE-2010: quartz microchip CE–based system with photodiode array detector Foundries  AMIC (Uppsala, Sweden) www.amic.se  Bartels Mikrotechnik (Dortmund, Germany) www.bartels-mikrotechnik.de  Epigem (Cleveland, UK) http://www.epigem.co.uk  Micralyne (Edmonton, Canada) http://www.micralyne.com/index.html  Microfluidic Chip Shop (Jena, Germany) http://www.microfluidic-chipshop.com  Scandinavian Micro Biodevices (Lyngby, Denmark) http://www.smb.dk Figure 1. Open in new tabDownload slide LIF detection system for the electrophoretic microchip, based on an epifluorescent design (A), and single-channel chip, showing reservoirs, injection channel, and separation channel (B). Reprinted with permission from Ferrance and Landers (36). Figure 1. Open in new tabDownload slide LIF detection system for the electrophoretic microchip, based on an epifluorescent design (A), and single-channel chip, showing reservoirs, injection channel, and separation channel (B). Reprinted with permission from Ferrance and Landers (36). Figure 2. Open in new tabDownload slide Diagnostic screening of DNA mutations in the BRCA1 and BRCA2 breast cancer susceptibility genes, using heteroduplex analysis on a microchip. The 3 types of DNA mutations—deletions, insertions, and substitutions—are all readily detected in separations taking <130 s. RFU, relative fluorescence units; hm, homozygote; ht, heterozygote. Reprinted with permission from Ferrance and Landers (36). Figure 2. Open in new tabDownload slide Diagnostic screening of DNA mutations in the BRCA1 and BRCA2 breast cancer susceptibility genes, using heteroduplex analysis on a microchip. The 3 types of DNA mutations—deletions, insertions, and substitutions—are all readily detected in separations taking <130 s. RFU, relative fluorescence units; hm, homozygote; ht, heterozygote. Reprinted with permission from Ferrance and Landers (36). Figure 3. Open in new tabDownload slide Detection and separation of 200 μmol/L homocysteine (Hcy) and reduced glutathione (GSH). Elution buffer, 20 mmol/L TES (pH 8.0), 400 μmol/L tris(2-carboxyethyl)phosphine (pH 7.5); applied potential, 1800 V (360 V/cm); injection, 4 s at 1200 V; E = +100 mV with Au/Hg amalgamated electrode. Reprinted with permission from Pasas et al. (44). Figure 3. Open in new tabDownload slide Detection and separation of 200 μmol/L homocysteine (Hcy) and reduced glutathione (GSH). Elution buffer, 20 mmol/L TES (pH 8.0), 400 μmol/L tris(2-carboxyethyl)phosphine (pH 7.5); applied potential, 1800 V (360 V/cm); injection, 4 s at 1200 V; E = +100 mV with Au/Hg amalgamated electrode. Reprinted with permission from Pasas et al. (44). Figure 4. Open in new tabDownload slide Pathogenic organism analysis conducted directly from intact E. coli cells on the portable PCR-CE microsystem. (A), analysis of E. coli K12 cells, showing only the presence of the coinjected DNA ladder and the 280-bp 16S species-specific amplicon. (B), analysis of E. coli O55:H7 cells, showing the ladder, the 280-bp 16S species-specific amplicon, and the 625-bp fliC amplicon characteristic of cells presenting the H7 surface antigen. (C), analysis of E. coli O157:H7 cells, showing the DNA ladder, the 16S species-specific amplicon, the 625-bp fliC amplicon, and the 348-bp sltI amplicon, characteristic of E. coli both possessing an H7 antigen and expressing shigatoxin. Each analysis was conducted with a starting concentration of 40 cells in the reactor in a time of 30 min. Reprinted with permission from Lagally et al. (59). Figure 4. Open in new tabDownload slide Pathogenic organism analysis conducted directly from intact E. coli cells on the portable PCR-CE microsystem. (A), analysis of E. coli K12 cells, showing only the presence of the coinjected DNA ladder and the 280-bp 16S species-specific amplicon. (B), analysis of E. coli O55:H7 cells, showing the ladder, the 280-bp 16S species-specific amplicon, and the 625-bp fliC amplicon characteristic of cells presenting the H7 surface antigen. (C), analysis of E. coli O157:H7 cells, showing the DNA ladder, the 16S species-specific amplicon, the 625-bp fliC amplicon, and the 348-bp sltI amplicon, characteristic of E. coli both possessing an H7 antigen and expressing shigatoxin. Each analysis was conducted with a starting concentration of 40 cells in the reactor in a time of 30 min. Reprinted with permission from Lagally et al. (59). Table 2. Comparison of conventional and microchip CE. Feature . Conventional CE . Microchip CE . Injection Hydrodynamic; electrokinetic Mainly electrokinetic Detection Mainly ultraviolet and LIF Mainly LIF Separation channels Mainly silica; single capillary or capillary array Glass or polymer Separation media Buffers, gels, sieving polymers, microparticles Buffers, sieving polymers, microparticles Analysis speed Fast (typically minutes) Very fast (typically seconds) Peak capacity More peaks because of longer capillaries Fewer peaks because of short channels Integration Hard to connect capillaries without dead volume Easy to integrate multiple functions, e.g., PCR-CE Automation Highly automated Highly automated in some commercial systems Throughput Very high for multicapillary systems Very high for multichannel systems Sample amount Very small (nanoliters to microliters) Very small (nanoliters to microliters) Reagent usage Very small (typically microliters to milliliters/day) Very small (typically microliters to milliliters/day) Potential for growth Relatively mature Emerging technology with potential for novel microchip designs and new applications Feature . Conventional CE . Microchip CE . Injection Hydrodynamic; electrokinetic Mainly electrokinetic Detection Mainly ultraviolet and LIF Mainly LIF Separation channels Mainly silica; single capillary or capillary array Glass or polymer Separation media Buffers, gels, sieving polymers, microparticles Buffers, sieving polymers, microparticles Analysis speed Fast (typically minutes) Very fast (typically seconds) Peak capacity More peaks because of longer capillaries Fewer peaks because of short channels Integration Hard to connect capillaries without dead volume Easy to integrate multiple functions, e.g., PCR-CE Automation Highly automated Highly automated in some commercial systems Throughput Very high for multicapillary systems Very high for multichannel systems Sample amount Very small (nanoliters to microliters) Very small (nanoliters to microliters) Reagent usage Very small (typically microliters to milliliters/day) Very small (typically microliters to milliliters/day) Potential for growth Relatively mature Emerging technology with potential for novel microchip designs and new applications Table 2. Comparison of conventional and microchip CE. Feature . Conventional CE . Microchip CE . Injection Hydrodynamic; electrokinetic Mainly electrokinetic Detection Mainly ultraviolet and LIF Mainly LIF Separation channels Mainly silica; single capillary or capillary array Glass or polymer Separation media Buffers, gels, sieving polymers, microparticles Buffers, sieving polymers, microparticles Analysis speed Fast (typically minutes) Very fast (typically seconds) Peak capacity More peaks because of longer capillaries Fewer peaks because of short channels Integration Hard to connect capillaries without dead volume Easy to integrate multiple functions, e.g., PCR-CE Automation Highly automated Highly automated in some commercial systems Throughput Very high for multicapillary systems Very high for multichannel systems Sample amount Very small (nanoliters to microliters) Very small (nanoliters to microliters) Reagent usage Very small (typically microliters to milliliters/day) Very small (typically microliters to milliliters/day) Potential for growth Relatively mature Emerging technology with potential for novel microchip designs and new applications Feature . Conventional CE . Microchip CE . Injection Hydrodynamic; electrokinetic Mainly electrokinetic Detection Mainly ultraviolet and LIF Mainly LIF Separation channels Mainly silica; single capillary or capillary array Glass or polymer Separation media Buffers, gels, sieving polymers, microparticles Buffers, sieving polymers, microparticles Analysis speed Fast (typically minutes) Very fast (typically seconds) Peak capacity More peaks because of longer capillaries Fewer peaks because of short channels Integration Hard to connect capillaries without dead volume Easy to integrate multiple functions, e.g., PCR-CE Automation Highly automated Highly automated in some commercial systems Throughput Very high for multicapillary systems Very high for multichannel systems Sample amount Very small (nanoliters to microliters) Very small (nanoliters to microliters) Reagent usage Very small (typically microliters to milliliters/day) Very small (typically microliters to milliliters/day) Potential for growth Relatively mature Emerging technology with potential for novel microchip designs and new applications 1 Nonstandard abbreviations: CE, capillary electrophoresis; CSF, cerebral spinal fluid; LIF, laser-induced fluorescence; and DMD, Duchenne muscular dystrophy. S.F.Y. Li, is a founder and shareholder of CE Resources Pte Ltd. L.J. Kricka is a scientific cofounder of ChemCore, a company acquired by Caliper Technologies Corporation, but he has no financial interest in any of the companies mentioned in this review. 1 Harrison DJ, Manz A, Fan ZH, Lido H, Widmer HM. Capillary electrophoresis and sample injection systems integrated on a planar glass chip. Anal Chem 1992 ; 64 : 1926 -1932. 2 Manz A, Harrison DJ, Verpoorte EMJ, Fettinger JC, Paulus A, Ludi H, et al. Planar chips technology for miniaturization and integration of separation techniques into monitoring systems—capillary electrophoresis on a chip. J Chromatogr 1992 ; 593 : 253 -258. 3 Tullo A, D’Erchia AM, Sbisa E. Methods for screening tumors for p53 status and therapeutic exploitation. Expert Rev Mol Diagn 2003 ; 3 : 289 -301. 4 Verpoorte E. Microfluidic chips for clinical and forensic analysis. Electrophoresis 2002 ; 23 : 677 -712. 5 Gross W, Marz W. Application of electrophoretic techniques to the diagnosis of disorders of lipoprotein metabolism. Examples at the levels of lipoproteins and apolipoproteins. Anal Chim Acta 1999 ; 383 : 169 -184. 6 Fodinger M, Sunder-Plassmann G, Wagner OF. Trends in molecular diagnostics. Wien Klin Wochenschr 1999 ; 111 : 315 -319. 7 McGlennen RC. Miniaturization technologies for molecular diagnostics. Clin Chem 2001 ; 47 : 393 -402. 8 Vo-Dinh T, Cullum BM, Stokes DL. Nanosensors and biochips: frontiers in biomolecular diagnostics. Sensor Actuat B-Chem 2001 ; 74 : 2 -11. 9 McKenzie SE, Mansfield E, Rappaport E, Surrey S, Fortina P. Parallel molecular genetic analysis. Eur J Hum Genet 1998 ; 6 : 417 -429. 10 Kricka LJ. Prospects for microchips in immunoassays and immunoassay tests at the point-of-care. J Clin Ligand Assay 2002 ; 25 : 317 -324. 11 Jin LJ, Ferrance J, Landers JP. Miniaturized electrophoresis: an evolving role in laboratory medicine. Biotechniques 2001 ; 31 : 1332 -1353. 12 Kricka LJ. Microchips, microarrays, biochips and nanochips: personal laboratories for the 21st century. Clin Chim Acta 2001 ; 307 : 219 -223. 13 Galvin P. A nanobiotechnology roadmap for high-throughput single nucleotide polymorphism analysis. Psychiatr Genet 2002 ; 12 : 75 -82. 14 Viovy JL. Electrophoresis of DNA and other polyelectrolytes: physical mechanisms. Rev Mod Phys 2000 ; 72 : 813 -872. 15 Oin JH, Fung YS, Lin BC. DNA diagnosis by capillary electrophoresis and microfabricated electrophoretic devices. Expert Rev Mol Diagn 2003 ; 3 : 387 -394. 16 Mitchelson KR. The use of capillary electrophoresis for DNA polymorphism analysis. Mol Biotechnol 2003 ; 24 : 41 -68. 17 Andersen PS, Jespersgaard C, Vuust J, Christiansen M, Larsen LA. Capillary electrophoresis-based single strand DNA conformation analysis in high-throughput mutation screening. Hum Mutat 2003 ; 21 : 455 -465. 18 Chen L, Ren JC. High-throughput DNA analysis by microchip electrophoresis. Comb Chem High Throughput Screen 2004 ; 7 : 29 -43. 19 Ng JH, Ilag LL. Biomedical applications of protein chips. J Cell Mol Med 2002 ; 6 : 329 -340. 20 Le Bricon T. Laboratory identification and measurement of urinary proteins. Ann Biol Clin Paris 2002 ; 60 : 525 -540. 21 Chen YL, Lu ZH. Microchip: novel tool in the field of life science. Prog Biochem Biophys 1998 ; 25 : 517 -522. 22 Kricka LJ. Miniaturization of analytical systems. Clin Chem 1998 ; 44 : 2008 -2014. 23 Xu JQ, He XZ, Zhou YX, Liu LT, Cheng J. Research and applications of biochip technologies. Chin Sci Bull 2000 ; 45 : 101 -108. 24 Wang J. From DNA biosensors to gene chips. Nucleic Acids Res 2000 ; 28 : 3011 -3016. 25 Madou M, Florkey J. From batch to continuous manufacturing of microbiomedical devices. Chem Rev 2000 ; 100 : 2679 -2691. 26 Cunningham DD. Fluidics and sample handling in clinical chemical analysis. Anal Chim Acta 2001 ; 429 : 1 -18. 27 Rossier J, Reymond F, Michel PE. Polymer microfluidic chips for electrochemical and biochemical analyses. Electrophoresis 2002 ; 23 : 858 -867. 28 Chen SH. Microchip electrophoresis and the analysis of polymerase chain reaction products. LC-GC North Amer 2002 ; 20 : 164 -173. 29 Cooper JW, Wang YJ, Lee CS. Recent advances in capillary separations for proteomics. Electrophoresis 2004 ; 25 : 3913 -3926. 30 Cai XX, Cui DF. The development of biosensors and biochips in IECAS. Netw Parallel Comput Proc 2004 ; 3222 : 522 -525. 31 Bashir R. BioMEMS: state-of-the-art in detection, opportunities and prospects. Adv Drug Deliv Rev 2004 ; 56 : 1565 -1586. 32 Minc N, Viovy JL. Microfluidics and biological applications: the stakes and trends. C R Physique 2004 ; 5 : 565 -575. 33 Tian HJ, Jaquins-Gerstl A, Munro N, Trucco M, Brody LC, Landers JP. Single-strand conformation polymorphism analysis by capillary and microchip electrophoresis: a fast, simple method for detection of common mutations in BRCA1 and BRCA2. Genomics 2000 ; 63 : 25 -34. 34 Tian HJ, Brody LC, Fan SJ, Huang ZL, Landers JP. Capillary and microchip electrophoresis for rapid detection of known mutations by combining allele-specific DNA amplification with heteroduplex analysis. Clin Chem 2001 ; 47 : 173 -185. 35 Tian HJ, Brody LC, Landers JP. Rapid detection of deletion, insertion, and substitution mutations via heteroduplex analysis using capillary- and microchip-based electrophoresis. Genome Res 2000 ; 10 : 1403 -1413. 36 Ferrance J, Landers JP. Exploiting sensitive laser-induced fluorescence detection on electrophoretic microchips for executing rapid clinical diagnostics. Luminescence 2001 ; 16 : 79 -88. 37 Tian HJ, Landers JP. Hydroxyethylcellulose as an effective polymer network for DNA analysis in uncoated glass microchips: optimization and application to mutation detection via heteroduplex analysis. Anal Biochem 2002 ; 309 : 212 -223. 38 Munro NJ, Snow K, Kant EA, Landers JP. Molecular diagnostics on microfabricated electrophoretic devices: from slab gel- to capillary- to microchip-based assays for T- and B-cell lymphoproliferative disorders. Clin Chem 1999 ; 45 : 1906 -1917. 39 Thomas G, Sinville R, Sutton S, Farquar H, Hammer RP, Soper SA, et al. Capillary and microelectrophoretic separations of ligase detection reaction products produced from low-abundant point mutations in genomic DNA. Electrophoresis 2004 ; 25 : 1668 -1677. 40 Cantafora A, Blotta I, Rivabene R, Pisciotta L, Bertolini S. Evaluation of RNA messengers involved in lipid trafficking of human intestinal cells by reverse-transcription polymerase chain reaction with competimer technology and microchip electrophoresis. Electrophoresis 2003 ; 24 : 3748 -3754. 41 Cohen CB, Chin-Dixon E, Jeong S, Nikiforov TT. A microchip-based enzyme assay for protein kinase. Anal Biochem 1999 ; 273 : 89 -97. 42 Ceriotti L, Shibata T, Folmer B, Weiller BH, Roberts MA, de Rooij NF, et al. Low-density lipoprotein analysis in microchip capillary electrophoresis systems. Electrophoresis 2002 ; 23 : 3615 -3622. 43 Weiller BH, Ceriotti L, Shibata T, Rein D, Roberts MA, Lichtenberg J, et al. Analysis of lipoproteins by capillary zone electrophoresis in microfluidic devices: assay development and surface roughness measurements. Anal Chem 2002 ; 74 : 1702 -1711. 44 Pasas SA, Lacher NA, Davies MI, Lunte SM. Detection of homocysteine by conventional and microchip capillary electrophoresis/electrochemistry. Electrophoresis 2002 ; 23 : 759 -766. 45 Kleparnik K, Horky M. Detection of DNA fragmentation in a single apoptotic cardiomyocyte by electrophoresis on a microfluidic device. Electrophoresis 2003 ; 24 : 3778 -3783. 46 Wang J, Chatrathi MP. Microfabricated electrophoresis chip for bioassay of renal markers. Anal Chem 2003 ; 75 : 525 -529. 47 Garcia CD, Henry CS. Direct detection of renal function markers using microchip CE with pulsed electrochemical detection. Analyst 2004 ; 129 : 579 -584. 48 Lapos JA, Manica DP, Ewing AG. Dual fluorescence and electrochemical detection on an electrophoresis microchip. Anal Chem 2002 ; 74 : 3348 -3353. 49 Hadd AG, Jacobson SC, Ramsey JM. Microfluidic assays of acetylcholinesterase inhibitors. Anal Chem 1999 ; 71 : 5206 -5212. 50 Phillips TM. Rapid analysis of inflammatory cytokines in cerebrospinal fluid using chip-based immunoaffinity electrophoresis. Electrophoresis 2004 ; 25 : 1652 -1659. 51 Schmalzing D, Koutny LB, Taylor TA, Nashabeh W, Fuchs M. Immunoassay for thyroxine (T4) in serum using capillary electrophoresis and micromachined devices. J Chromatogr B 1997 ; 697 : 175 -180. 52 Wang J, Ibanez A, Chatrathi MP. Microchip-based amperometric immunoassays using redox tracers. Electrophoresis 2002 ; 23 : 3744 -3749. 53 Zhou ZM, Liu DY, Zhong RT, Dai ZP, Wu DP, Wang H, et al. Determination of SARS-coronavirus by a microfluidic chip system. Electrophoresis 2004 ; 25 : 3032 -3039. 54 Chen YH, Wang WC, Young KC, Chang TT, Chen SH. Plastic microchip electrophoresis for analysis of PCR products of hepatitis C virus. Clin Chem 1999 ; 45 : 1938 -1943. 55 Young KC, Lien HM, Lin CC, Chang TT, Lee GB, Chen SH. Microchip and capillary electrophoresis for quantitative analysis of hepatitis C virus based on RT-competitive PCR. Talanta 2002 ; 56 : 323 -330. 56 Hofgartner WT, Huhmer AFR, Landers JP, Kant JA. Rapid diagnosis of herpes simplex encephalitis using microchip electrophoresis of PCR products. Clin Chem 1999 ; 45 : 2120 -2128. 57 Vegvari A, Hjerten S. Hybrid microdevice electrophoresis of peptides, proteins, DNA, viruses, and bacteria in various separation media, using UV-detection. Electrophoresis 2003 ; 24 : 3815 -3820. 58 Liu YJ, Ganser D, Schneider A, Liu R, Grodzinski P, Kroutchinina N. Microfabricated polycarbonate CE devices for DNA analysis. Anal Chem 2001 ; 73 : 4196 -4201. 59 Lagally ET, Scherer JR, Blazej RG, Toriello NM, Diep BA, Ramchandani M, Set al. Integrated portable genetic analysis microsystem for pathogen/infectious disease detection. Anal Chem 2004 ; 76 : 3162 -3170. 60 Waters LC, Jacobson SC, Kroutchinina N, Khandurina J, Foote RS, Ramsey JM. Microchip device for cell lysis, multiplex PCR amplification, and electrophoretic sizing. Anal Chem 1998 ; 70 : 158 -162. 61 Waters LC, Jacobson SC, Kroutchinina N, Khandurina J, Foote RS, Ramsey JM. Multiple sample PCR amplification and electrophoretic analysis on a microchip. Anal Chem 1998 ; 70 : 5172 -5176. 62 Koh CG, Tan W, Zhao MQ, Ricco AJ, Fan ZH. Integrating polymerase chain reaction, valving, and electrophoresis in a plastic device for bacterial detection. Anal Chem 2003 ; 75 : 4591 -4598. 63 Woolley AT, Hadley D, Landre P, deMello AJ, Mathies RA, Northrup MA. Functional integration of PCR amplification and capillary electrophoresis in a microfabricated DNA analysis device. Anal Chem 1996 ; 68 : 4081 -4086. 64 Karasawa K, Arakawa H, Igarashi T, Goto N, Maeda M. Detection of cariogenic bacterial genes by microchip electrophoresis. J Chromatogr B-Biomed 2004 ; 810 : 41 -44. 65 Linder V, Verpoorte E, de Rooij NF, Sigris H, Thormann W. Application of surface biopassivated disposable poly(dimethylsiloxane)/glass chips to a heterogeneous competitive human serum immunoglobulin G immunoassay with incorporated internal standard. Electrophoresis 2002 ; 23 : 740 -749. 66 Rodriguez I, Zhang Y, Lee HK, Li SFY. Conventional capillary electrophoresis in comparison with short-capillary electrophoresis and microfabricated glass chip capillary electrophoresis for the analysis of anti-human immunoglobulin G. J Chromatogr A 1997 ; 781 : 287 -293. 67 Wang J, Ibanez A, Chatrathi MP, Escarpa A. Electrochemical enzyme immunoassays on microchip platforms. Anal Chem 2001 ; 73 : 5323 -5327. 68 Abad-Viller EM, Tanyanyiwa J, Fernandez-Abedul MT, Costa-Garcia A, Hauser PC. Detection of human immunoglobulin in microchip and conventional capillary electrophoresis with contactless conductivity measurements. Anal Chem 2004 ; 76 : 1282 -1288. 69 Wang J, Ibanez A, Chatrathi MP. On-chip integration of enzyme and immunoassays: simultaneous measurements of insulin and glucose. J Am Chem Soc 2003 ; 125 : 8444 -8445. 70 Roper MG, Shackman JG, Kahlgren GM, Kennedy RT. Microfluidic chip for continuous monitoring of hormone secretion from live cells using an electrophoresis-based immunoassay. Anal Chem 2003 ; 75 : 4711 -4717. 71 Du Y, Yan JL, Zhou WZ, Yang XY, Wang EK. Direct electrochemical detection of glucose in human plasma on capillary electrophoresis microchips. Electrophoresis 2004 ; 25 : 3853 -3859. 72 Ferrance JP, Wu QR, Giordano B, Hernandez C, Kwok Y, Snow K, et al. Developments toward a complete micro-total analysis system for Duchenne muscular dystrophy diagnosis. Anal Chim Acta 2003 ; 500 : 223 -236. 73 Ferrance J, Snow K, Landers JP. Evaluation of microchip electrophoresis as a molecular diagnostic method for Duchenne muscular dystrophy. Clin Chem 2002 ; 48 : 380 -383. 74 Sanders JC, Breadmore MC, Mitchell PS, Landers JP. A simple PDMS-based electro-fluidic interface for microchip electrophoretic separations. Analyst 2002 ; 127 : 1558 -1563. 75 Ertl P, Emrich CA, Singhal P, Mathies RA. Capillary electrophoresis chips with a sheath-flow supported electrochemical detection system. Anal Chem 2004 ; 76 : 3749 -3755. 76 Raisi F, Blizard BA, Shabari AR, Ching J, Kintz GJ, Mitchell J, et al. Human genomic DNA analysis using a semi-automated sample preparation, amplification, and electrophoresis separation platform. J Sep Sci 2004 ; 27 : 275 -283. 77 Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, et al. The sequence of the human genome. Science 2001 ; 291 : 1304 -1351. 78 Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, et al. Initial sequencing and analysis of the human genome. Nature 2001 ; 409 : 860 -921. 79 Britz-McKibbin P, Ichihashi T, Tsubota K, Chen DDY, Terabe S. Complementary on-line preconcentration strategies for steroids by capillary electrophoresis. J Chromatogr A 2003 ; 1013 : 65 -76. 80 Sera Y, Matsubara N, Otsuka K, Terabe S. Sweeping on a microchip: concentration profiles of the focused zone in micellar electrokinetic chromatography. Electrophoresis 2001 ; 22 : 3509 -3513. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Garcia-Barceló,, Mercè;So,, Man-ting;Lau, Danny, Ko-chun;Leon, Thomas, Yuk-yu;Yuan,, Zheng-wei;Cai,, Wei-song;Lui, Vincent, Chi-hang;Fu,, Ming;Herbrick,, Jo-Anne;Gutter,, Emily;Proud,, Virginia;Li,, Long;Pierre-Louis,, Jacqueline;Aleck,, Kirk;

Kerkeni,, Mohsen;Addad,, Faouzi;Chauffert,, Maryline;Myara,, Anne;Ben Farhat,, Mohamed;Miled,, Abdelhedi;Maaroufi,, Khira;Trivin,, François

doi: 10.1373/clinchem.2005.057950pmid: 16284093

Abstract Background: Hyperhomocysteinemia is an independent, graded risk factor for coronary artery disease (CAD). The G894T variant of endothelial nitric oxide synthase (eNOS) was postulated to be associated with hyperhomocysteinemia and could influence individual susceptibility to CAD. The aims of this study were to investigate (a) the relationship of the eNOS G894T polymorphism with the presence and the severity of CAD and (b) the possible relationship between hyperhomocysteinemia and the eNOS G894T variant for the risk of CAD severity in a Tunisian population. Methods: We used PCR with restriction fragment length polymorphism analysis to detect the G894T variant of the eNOS gene in 100 patients with CAD and 120 healthy controls. The severity of CAD was expressed by the number of affected vessels. Total plasma homocysteine concentrations were determined by direct chemiluminescence assay. Results: The frequencies of the eNOS GG, GT, and TT genotypes in the CAD group were significantly different from those in the control group (45%, 44%, and 11% vs 60%, 35.8% and 4.2%, respectively; P = 0.035). There was no association between the eNOS G894T genotype frequencies and the number of stenosed vessels (P = 0.149). In the CAD group, the coexistence of the 894 GT or TT genotypes and hyperhomocysteinemia led to an increased risk of CAD severity. Conclusion: The G894T polymorphism of the eNOS gene is associated with the presence of CAD, and in conjunction with hyperhomocysteinemia, increased the risk of CAD severity in a Tunisian population. Epidemiologic studies have shown that dyslipidemia, diabetes mellitus, obesity, hypertension, and cigarette smoking are risk factors for coronary artery disease (CAD)1 (1)(2)(3). Assessment of these metabolic or lifestyle risk factors has, however, been ineffective in completely predicting the development of the atherosclerotic process, suggesting that specific genetic predisposition should also be taken into account (4)(5). The vascular endothelium modulates blood vessel wall homeostasis through the production of factors regulating vessel tone, coagulation state, cell growth, cell death, and leukocyte trafficking (6). One of the most important endothelial cell products is nitric oxide (NO), which is synthesized from l-arginine by the enzyme endothelial nitric oxide synthase (eNOS) (7). NO plays a key role in the relaxation of vascular smooth muscle, inhibits platelet and leukocyte adhesion to the endothelium, reduces vascular smooth muscle cell migration and proliferation, and limits oxidation of the atherogenic LDLs (8). NO may modulate homocysteine (Hcy) concentration directly by inhibiting methionine synthase, the enzyme that synthesizes methionine from homocysteine and 5-methyltetrahydrofolate (9). Alternatively, NO may modulate Hcy concentrations indirectly via folate catabolism by inhibiting the synthesis of ferritin (10), a protein that promotes the irreversible oxidative cleavage of folate (11). Although low folate concentrations are associated with hyperhomocysteinemia (12), which is a risk factor for atherosclerosis (13)(14), the relative contributions of these potential mechanisms to Hcy modulation in vivo remain unclear. eNOS inhibition has also been shown to accelerate atherosclerosis in animal models, and abnormalities of the endothelial NO pathway are present in humans with atherosclerosis (15)(16). This evidence suggests that NO may inhibit several key steps in the atherosclerotic process and that alteration of NO production within the vascular endothelium could contribute to the pathogenesis of atherosclerosis. Thus, eNOS could be a candidate gene for atherosclerosis. Several polymorphisms have been identified in the eNOS gene, among which is one located in exon 7 (G984T), which modifies its coding sequence (Glu298→Asp). Associations between this variant and coronary spasm, CAD, and acute myocardial infarction have been reported, but data on its relationship with disease severity are lacking (17)(18). Here we describe the association between the G894T polymorphism of the eNOS gene and the presence and severity of CAD; we also evaluate the relationship between hyperhomocysteinemia, the eNOS G894T variant, and the risk of CAD severity in the Tunisian population. Materials and Methods study population The study population consisted of homogeneous Tunisian Arab descendents who resided in Tunisia and had no known Negroid or Mongoloid ancestry. The control group included 120 healthy volunteers (87 males) with no history of CAD, diabetes mellitus, or cerebrovascular diseases. Their mean (SD) age was 54 (10) years. One hundred consecutive patients (74 males) with angiographically documented CAD were enrolled from University Hospital Fattouma Bourguiba in Monastir (Cardiovascular Department). The mean age of this group was 59 (10) years. The number of significantly stenosed coronary arteries and lesions determined the severity of CAD (>50% luminal stenosis). The angiograms were assessed by 2 cardiologists who were unaware that the patients were to be included in the study and enabled patient subclassification as follow: group G0 (10%) had no stenosed vessels, group G1 (40%) had 1 stenosed vessel, group G2 (35%) had 2 stenosed vessels, and group G3 (15%) had severe CAD involving all 3 major coronary arteries. All participants were interviewed, and data on dyslipidemia, diabetes mellitus, hypertension, and smoking habits were recorded. Informed consent was obtained from each patient and control according to the guidelines of our ethics committee. For coronary risk factors, the following definitions were used: individuals were defined as hypertensive if their blood pressure was >140/90 mmHg or if they were receiving any antihypertensive treatment; individuals with a history of diabetes mellitus or those receiving any antidiabetic medication were considered to have diabetes; individuals were deemed dyslipidemic when their total cholesterol concentration was ≥5.68 mmol/L, their triglyceride concentration was ≥2.28 mmol/L, or they were receiving lipid-lowering drugs. Smoking history was coded as never or current smoker. measurement of total Hcy Plasma concentrations of total Hcy were measured by direct chemiluminescence using reagents and an ACS:180 automated analyzer from Bayer Vital GmbH. analysis of G894T polymorphism of eNOS gene Genomic DNA was extracted from whole blood samples by rapid methods (19). The coding sequence variant was a G→T substitution at position 894 in exon 7, which determines the Glu-to-Asp amino acid substitution (in codon 298) in the mature eNOS protein. Using a previously described procedure (18), we genotyped all participants by PCR amplification of exon 7 with the primers 5′-CATGAGGCTCAGCCCCAGAAC-3′ (sense) and 5′-AGTCAATCCCTTTGGTGCTCAC-3′ (antisense) followed by MboI restriction enzyme digestion for 16 h at 37 °C. In the presence of a T at nucleotide 894, which corresponds to Asp298, the 206-bp PCR product is cleaved into 2 fragments of 119 and 87 bp. The products of the digestion process were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (7%). statistical analyses Statistical analyses were performed with Statistical Package for Social Sciences for Windows, Ver. 10.0 (SPSS). Differences between the means of the 2 continuous variables were evaluated by Student t-test. Differences between noncontinuous variables, genotype distribution, and Hardy–Weinberg equilibrium were tested by χ2 analysis. One-way ANOVA was used to analyze the relationships between genotypes and the general characteristic and severity of CAD. Logistic regression analysis was used to assess the independent effect of each risk factor on the presence of CAD. Hyperhomocysteinemia was defined as a mildly increased Hcy (fasting total Hcy >15 μmol/L). Results are reported as the median and 25th and 75th percentiles, and P <0.05 was considered statistically significant. Results comparison of the 2 study groups The demographic and clinical characteristics of the CAD and control groups are given in Table 11 . The prevalence of atherogenic risk factors (including age, sex, hypertension, diabetes mellitus, and dyslipidemia) was significantly higher in the CAD group. Total Hcy was significantly higher in the CAD group than in the control group. distribution of the G894T polymorphism of the eNOS gene Although the distribution of genotypes in both the CAD and control groups satisfied the Hardy–Weinberg equilibrium, the G894T polymorphism of the eNOS gene was significantly associated with the presence of CAD in our patients (Table 22 ). The proportion of TT homozygotes was 11% in the CAD group and 4.2% in the control group (P = 0.035). G894T polymorphism of the eNOS gene and severity of cad Patients with CAD (n = 100) were subclassified into 4 subgroups (G0, G1, G2, and G3) according to the number of affected coronary arteries (Table 33 ). Statistically, the G894T polymorphism did not correlate with the extent of CAD (P = 0.149), but it tended to suggest a strong association with CAD severity. We found graded increased proportions of patients with the 894T allele presenting with 0- to 3-vessel stenosis (P = 0.056). association between eNOS G894T genotype and Hcy concentrations in study population The eNOS G894T genotype was significantly associated with Hcy in the CAD group (Table 44 ). The eNOS 984TT genotype was associated with increased Hcy in our patients (P <0.05). However, in the control group, the eNOS G894T genotype was not significantly associated with Hcy concentrations, probably because of the small number of individuals with the TT genotype (n = 5). Table 44 also shows the Hcy concentrations in patients presenting with 0- to 3-vessel stenosis according to eNOS gene polymorphism. We observed that patients with GT or TT genotypes and increased Hcy had more severe CAD. Interestingly, the coexistence of the 894GT or 894TT genotype and increased Hcy led to increased risk of CAD severity. results of multivariate analysis We used multiple logistic regression to test for independent correlates of the presence of CAD. Included in the model were age, sex, smoking, hypertension, diabetes mellitus, dyslipidemia, and eNOS G894T polymorphism. Age (P <0.001), sex (P <0.001), smoking (P = 0.031), hypertension (P = 0.018), diabetes mellitus (P = 0.01), and dyslipidemia (P = 0.026) were independent correlates of the presence of CAD, whereas the eNOS G894T polymorphism (P = 0.209) was not an independent predictor of CAD. Discussion We report an association between the common G894T polymorphism of the eNOS gene and the presence of CAD in the Tunisian population. We found an excess of homozygosity for the T894 variant among CAD cases compared with controls. Multivariate analysis showed that this association was not independent of other factors related to artery disease risk. Methylenetetrahydrofolate reductase and apolipoprotein E polymorphisms have been investigated as CAD risk factors (20). Gardemann et al. (21) showed that the TT genotype of the methylenetetrahydrofolate reductase C677T gene polymorphism is associated with extent of coronary atherosclerosis in patients at high risk for CAD. To date, the G894T polymorphism of the eNOS gene has been linked to increased risk of stroke, myocardial infarction, coronary atherosclerosis, and venous thrombosis (22)(23)(24)(25). Previous studies from Japan and the United Kingdom have already suggested a role for the G894T polymorphism in the development of coronary atherosclerosis, with the excess risk being confined to TT894 homozygosity (18)(26)(27), as in our study. These studies, however, also showed that the genotype frequency of the G894T polymorphism can vary considerably among different populations. Nevertheless, it is important to emphasize that some authors have failed to find any relationship between the G894T polymorphism and the risk of atherosclerosis (28)(29)(30)(31). In our study we investigated the association between the common G894T polymorphism of the eNOS gene and the severity of CAD. We found a strong, but not statistically significant, association. Controversial results regarding this association have been reported. Colombo and coworkers (32)(33) showed that the G894T polymorphism of the eNOS gene is associated with the severity of CAD, but authors of other studies found no association and no differences in the progression of atherosclerosis (34)(35). In our study we investigated whether a relationship between the eNOS G894T variant and increased Hcy concentrations increased the risk of CAD severity. This relationship had not been described previously. Heil et al. (25) showed that the G894T variant of eNOS increases the risk of recurrent venous thrombosis through interaction with increased Hcy concentrations. Brown et al. (36) postulated that the eNOS G894T variant influences plasma Hcy concentrations via folate catabolism. Our data suggest that the coexistence of the 894GT or 894TT genotype and hyperhomocysteinemia is associated with the number of vessels affected and increases the risk of CAD severity. Although this study was a preliminary study, these results are very interesting and could stimulate larger studies. Several studies have demonstrated that the bioavailability of NO is decreased in hyperhomocysteinemia (37)(38), which might be attributable to decreased NO production or to alternative mechanisms such oxidative stress or nitrosylation (37)(38)(39). NO can react with thiols such as Hcy to form S-nitrosothiols, which have vasodilatory and antiplatelet effects and are more stable than NO (40). Stamler et al. (40) hypothesized that under physiologic conditions, endothelial cells may modulate the deleterious effects of Hcy by releasing NO, which facilitates the formation of S-nitrosohomocysteine. When less NO is available, decreased amounts of S-nitrosohomocysteine will be formed, which would lead to exposure of the homeostatic system to the detrimental effects of Hcy (40). The eNOS G894T variant leads to an amino acid substitution of aspartate for glutamate at position 298 of the protein. An in vitro study showed that the eNOS protein with aspartate, but not glutamate, at position 298 is subject to cleavage by endogenous protease and produces 35-kDa amino-terminal and 100-kDa carboxyl-terminal fragments (41). The “degraded” 100-kDa eNOS is only a small fraction of the total. Nevertheless, Tesauro et al. (41) found significant potential structural changes in the Chou–Fasman secondary structure and pointed out that the coding region polymorphism has functional consequences. We therefore speculate that, in vivo, the eNOS 894TT genotype leads to altered NO production. In combination with the presence of hyperhomocysteinemia, this might lead to less capturing of Hcy in S-nitrosohomocysteine, with subsequent exposure of the homeostatic system to the toxic effects of Hcy. We hypothesize that interaction of the eNOS G894T variant and hyperhomocysteinemia induces less S-nitrosylation and could be one mechanism determining the extent of obstruction in CAD. In conclusion, the present study provides evidence that the G894T polymorphism of the eNOS gene is associated with the presence of CAD in the Tunisian population. The G894T polymorphism could represent a useful genetic marker to identify individuals prone to the development of atherosclerotic diseases. We found a relationship between hyperhomocysteinemia, the eNOS G894T variant, and the risk of CAD severity, which can be increased by the number of arteries stenosed. Further studies are needed to investigate the relationship of hyperhomocysteinemia associated with the eNOS polymorphism and the extent of obstructive CAD. Table 1. Demographic and clinical characteristics and Hcy concentrations of the study population. . CAD group . Control group . P . n 100 120 Mean (SD) age, years 59 (10) 54 (10) NS1 Male, (%) 74 (74) 87 (72.5) NS Smokers, n (%) 60 (60) 55 (45.8) <0.01 Hypertension, n (%) 41 (41) 7 (5.83) <0.001 Diabetes mellitus, n (%) 53 (53) 10 (8.33) <0.001 Dyslipidemia, n (%) 34 (34) 5 (4.16) <0.001 Median (25th–75th percentile) Hcy, μmol/L 13.45 (10.42–17.76) 10.86 (9.25–12) <0.001 . CAD group . Control group . P . n 100 120 Mean (SD) age, years 59 (10) 54 (10) NS1 Male, (%) 74 (74) 87 (72.5) NS Smokers, n (%) 60 (60) 55 (45.8) <0.01 Hypertension, n (%) 41 (41) 7 (5.83) <0.001 Diabetes mellitus, n (%) 53 (53) 10 (8.33) <0.001 Dyslipidemia, n (%) 34 (34) 5 (4.16) <0.001 Median (25th–75th percentile) Hcy, μmol/L 13.45 (10.42–17.76) 10.86 (9.25–12) <0.001 1 NS, not significant. Table 1. Demographic and clinical characteristics and Hcy concentrations of the study population. . CAD group . Control group . P . n 100 120 Mean (SD) age, years 59 (10) 54 (10) NS1 Male, (%) 74 (74) 87 (72.5) NS Smokers, n (%) 60 (60) 55 (45.8) <0.01 Hypertension, n (%) 41 (41) 7 (5.83) <0.001 Diabetes mellitus, n (%) 53 (53) 10 (8.33) <0.001 Dyslipidemia, n (%) 34 (34) 5 (4.16) <0.001 Median (25th–75th percentile) Hcy, μmol/L 13.45 (10.42–17.76) 10.86 (9.25–12) <0.001 . CAD group . Control group . P . n 100 120 Mean (SD) age, years 59 (10) 54 (10) NS1 Male, (%) 74 (74) 87 (72.5) NS Smokers, n (%) 60 (60) 55 (45.8) <0.01 Hypertension, n (%) 41 (41) 7 (5.83) <0.001 Diabetes mellitus, n (%) 53 (53) 10 (8.33) <0.001 Dyslipidemia, n (%) 34 (34) 5 (4.16) <0.001 Median (25th–75th percentile) Hcy, μmol/L 13.45 (10.42–17.76) 10.86 (9.25–12) <0.001 1 NS, not significant. Table 2. Genotype and allele frequencies of the G894T polymorphism of the eNOS gene in CAD and control groups. . CAD group . Control group . P . n 100 120 G894T polymorphism, n (%) 0.035  GG 45 (45) 72 (60)  GT 44 (44) 43 (35.8)  TT 11 (11) 5 (4.2) Allele, n (%) 0.026  G 134 (67) 187 (78)  T 66 (33) 53 (22) . CAD group . Control group . P . n 100 120 G894T polymorphism, n (%) 0.035  GG 45 (45) 72 (60)  GT 44 (44) 43 (35.8)  TT 11 (11) 5 (4.2) Allele, n (%) 0.026  G 134 (67) 187 (78)  T 66 (33) 53 (22) Table 2. Genotype and allele frequencies of the G894T polymorphism of the eNOS gene in CAD and control groups. . CAD group . Control group . P . n 100 120 G894T polymorphism, n (%) 0.035  GG 45 (45) 72 (60)  GT 44 (44) 43 (35.8)  TT 11 (11) 5 (4.2) Allele, n (%) 0.026  G 134 (67) 187 (78)  T 66 (33) 53 (22) . CAD group . Control group . P . n 100 120 G894T polymorphism, n (%) 0.035  GG 45 (45) 72 (60)  GT 44 (44) 43 (35.8)  TT 11 (11) 5 (4.2) Allele, n (%) 0.026  G 134 (67) 187 (78)  T 66 (33) 53 (22) Table 3. Genotype and allele frequencies of G894T polymorphism of the eNOS gene and CAD severity. . CAD group (n = 100) . . . . P . . G0 (n = 10) . G1 (n = 40) . G2 (n = 35) . G3 (n = 15) . . G894T polymorphism, n (%) 0.149  GG 7 (70) 21 (52.5) 14 (40) 3 (20)  GT 2 (20) 15 (37.5) 16 (45.7) 11 (73.3)  TT 1 (10) 4 (10) 5 (14.3) 1 (6.7) Allele, n (%) 0.056  G 16 (80) 57 (71.3) 44 (62.8) 17 (56.6)  T 4 (20) 23 (28.7) 26 (37.2) 13 (43.4) . CAD group (n = 100) . . . . P . . G0 (n = 10) . G1 (n = 40) . G2 (n = 35) . G3 (n = 15) . . G894T polymorphism, n (%) 0.149  GG 7 (70) 21 (52.5) 14 (40) 3 (20)  GT 2 (20) 15 (37.5) 16 (45.7) 11 (73.3)  TT 1 (10) 4 (10) 5 (14.3) 1 (6.7) Allele, n (%) 0.056  G 16 (80) 57 (71.3) 44 (62.8) 17 (56.6)  T 4 (20) 23 (28.7) 26 (37.2) 13 (43.4) Table 3. Genotype and allele frequencies of G894T polymorphism of the eNOS gene and CAD severity. . CAD group (n = 100) . . . . P . . G0 (n = 10) . G1 (n = 40) . G2 (n = 35) . G3 (n = 15) . . G894T polymorphism, n (%) 0.149  GG 7 (70) 21 (52.5) 14 (40) 3 (20)  GT 2 (20) 15 (37.5) 16 (45.7) 11 (73.3)  TT 1 (10) 4 (10) 5 (14.3) 1 (6.7) Allele, n (%) 0.056  G 16 (80) 57 (71.3) 44 (62.8) 17 (56.6)  T 4 (20) 23 (28.7) 26 (37.2) 13 (43.4) . CAD group (n = 100) . . . . P . . G0 (n = 10) . G1 (n = 40) . G2 (n = 35) . G3 (n = 15) . . G894T polymorphism, n (%) 0.149  GG 7 (70) 21 (52.5) 14 (40) 3 (20)  GT 2 (20) 15 (37.5) 16 (45.7) 11 (73.3)  TT 1 (10) 4 (10) 5 (14.3) 1 (6.7) Allele, n (%) 0.056  G 16 (80) 57 (71.3) 44 (62.8) 17 (56.6)  T 4 (20) 23 (28.7) 26 (37.2) 13 (43.4) Table 4. Association between eNOS G894T genotype and Hcy concentrations in control and CAD groups.1 . eNOS G894T genotype . . . . GG . GT . TT . Controls  n 72 43 5  Hcy, μmol/L 9.2 (8–12.5) 0.9 (9.2–14) 11.6 (9–15.3) CAD  n 44 45 11  Hcy, μmol/L 11.47 (10–14.1) 14.13 (10.7–20) 20.3 (14.5–32)2 Severity of CAD  0 vessels   n 7 2 1   Hcy, μmol/L 10 (8.7–11) 12.8 (10–15.8) 10.55  1 vessel   n 21 15 4   Hcy, μmol/L 11.6 (9.9–14) 13.6 (9.6–22.3) 14.13 (11.6–18)  2 vessels   n 13 17 5   Hcy, μmol/L 12.1 (10.2–16) 16.9 (12.5–23.7)3 18.9 (13.5–24)4  3 vessels   n 3 11 1   Hcy,μmol/L 13.5 (10.3–13.9) 19.2 (16.8–39.8)5 35.65 . eNOS G894T genotype . . . . GG . GT . TT . Controls  n 72 43 5  Hcy, μmol/L 9.2 (8–12.5) 0.9 (9.2–14) 11.6 (9–15.3) CAD  n 44 45 11  Hcy, μmol/L 11.47 (10–14.1) 14.13 (10.7–20) 20.3 (14.5–32)2 Severity of CAD  0 vessels   n 7 2 1   Hcy, μmol/L 10 (8.7–11) 12.8 (10–15.8) 10.55  1 vessel   n 21 15 4   Hcy, μmol/L 11.6 (9.9–14) 13.6 (9.6–22.3) 14.13 (11.6–18)  2 vessels   n 13 17 5   Hcy, μmol/L 12.1 (10.2–16) 16.9 (12.5–23.7)3 18.9 (13.5–24)4  3 vessels   n 3 11 1   Hcy,μmol/L 13.5 (10.3–13.9) 19.2 (16.8–39.8)5 35.65 1 Hcy concentrations are the median (25th–75th percentile). 2 Significantly different from patients or controls with GG genotype (P <0.05). 3 Significantly different from patients with GT genotype presenting with 1-vessel stenosis (P <0.05). 4 Significantly different from patients with TT genotype presenting with 1-vessel stenosis (P <0.01). 5 Significantly different from patients with GT genotype presenting with 2-vessel stenosis (P <0.05) and patients with GT genotype presenting with 1-vessel stenosis (P <0.01). Table 4. Association between eNOS G894T genotype and Hcy concentrations in control and CAD groups.1 . eNOS G894T genotype . . . . GG . GT . TT . Controls  n 72 43 5  Hcy, μmol/L 9.2 (8–12.5) 0.9 (9.2–14) 11.6 (9–15.3) CAD  n 44 45 11  Hcy, μmol/L 11.47 (10–14.1) 14.13 (10.7–20) 20.3 (14.5–32)2 Severity of CAD  0 vessels   n 7 2 1   Hcy, μmol/L 10 (8.7–11) 12.8 (10–15.8) 10.55  1 vessel   n 21 15 4   Hcy, μmol/L 11.6 (9.9–14) 13.6 (9.6–22.3) 14.13 (11.6–18)  2 vessels   n 13 17 5   Hcy, μmol/L 12.1 (10.2–16) 16.9 (12.5–23.7)3 18.9 (13.5–24)4  3 vessels   n 3 11 1   Hcy,μmol/L 13.5 (10.3–13.9) 19.2 (16.8–39.8)5 35.65 . eNOS G894T genotype . . . . GG . GT . TT . Controls  n 72 43 5  Hcy, μmol/L 9.2 (8–12.5) 0.9 (9.2–14) 11.6 (9–15.3) CAD  n 44 45 11  Hcy, μmol/L 11.47 (10–14.1) 14.13 (10.7–20) 20.3 (14.5–32)2 Severity of CAD  0 vessels   n 7 2 1   Hcy, μmol/L 10 (8.7–11) 12.8 (10–15.8) 10.55  1 vessel   n 21 15 4   Hcy, μmol/L 11.6 (9.9–14) 13.6 (9.6–22.3) 14.13 (11.6–18)  2 vessels   n 13 17 5   Hcy, μmol/L 12.1 (10.2–16) 16.9 (12.5–23.7)3 18.9 (13.5–24)4  3 vessels   n 3 11 1   Hcy,μmol/L 13.5 (10.3–13.9) 19.2 (16.8–39.8)5 35.65 1 Hcy concentrations are the median (25th–75th percentile). 2 Significantly different from patients or controls with GG genotype (P <0.05). 3 Significantly different from patients with GT genotype presenting with 1-vessel stenosis (P <0.05). 4 Significantly different from patients with TT genotype presenting with 1-vessel stenosis (P <0.01). 5 Significantly different from patients with GT genotype presenting with 2-vessel stenosis (P <0.05) and patients with GT genotype presenting with 1-vessel stenosis (P <0.01). 1 Nonstandard abbreviations: CAD, coronary artery disease; NO, nitric oxide; eNOS, endothelial nitric oxide synthase; and Hcy, homocysteine. We thank Dr. Mohamed Ben Farhat and his colleagues of the Cardiovascular Department at University Hospital Fattouma Bourguiba in Monastir (Tunisia) for providing samples. We also thank Drs. Salwa Yacoub and Monia Zaier of the regional Blood Transfusion Centre in Sousse (Tunisia) for recruiting the healthy controls. 1 Gotto AM. Interactions of the major risk factors for coronary heart disease. Am J Med 1986 ; 80 : 48 -55. 2 Kannel WB. Clinical misconceptions dispelled by epidemiological research. Circulation 1995 ; 92 : 3350 -3360. 3 Wilson PW, D’Agastino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB, et al. Prediction of coronary heart disease using risk factors categories. Circulation 1998 ; 97 : 1837 -1847. 4 Marenberg ME, Risch N, Berkman LF, Floderus B, de Faire U. Genetic susceptibility to death from coronary heart disease in a study of twins. N Engl J Med 1994 ; 330 : 1041 -1046. 5 Boerwinkle E, Ellsworth DL, Hallman DM, Biddinger A. Genetic analysis of atherosclerosis: a research paradigm for the common diseases. Hum Mol Genet 1996 ; 5 : 1405 -1410. 6 Vane JR, Anggard EE, Botting RM. Regulatory functions of the vascular endothelium. [Review] N Engl J Med 1990 ; 323 : 27 -36. 7 Moncada S, Higgs A. The l-arginine-nitric oxide pathway. [Review] N Engl Med 1993 ; 329 : 2002 -2012. 8 Schmidt HH, Walter U. NO at work. Cell 1994 ; 78 : 919 -925. 9 Danishpajooh IO, Gudi T, Chen Y, Kharitonov VG, Sharma VS, Boss GR. Nitric oxide inhibits methionine synthase activity in vivo and disrupts carbon flow through the folate pathway. J Biol Chem 2001 ; 276 : 27296 -27303. 10 Juckett MB, Weber M, Balla J, Jacob HS, Vercellotti GM. Nitric oxide donors modulate ferritin and protect endothelium from oxidative injury. Free Radic Biol Med 1996 ; 20 : 63 -73. 11 Suh JR, Oppenheim EW, Girgris S, Stover PJ. Purification and properties of folate-catabolizing enzyme. J Biol Chem 2000 ; 275 : 35646 -35655. 12 Durand P, Prost M, Loreau N, Lussier-Cacan S, Blache D. Impaired homocysteine metabolism and atherothrombotic disease. [Review] Lab Invest 2001 ; 81 : 645 -672. 13 Clarke R, Daly L, Robinson K, Naughten E, Cahalane S, Fowler B, et al. Hyperhomocysteinemia: an independent risk factor for vascular disease. N Engl J Med 1991 ; 324 : 1149 -1155. 14 . Homocysteine Studies Collaboration. Homocysteine and risk of ischemic heart disease and stroke: a meta-analysis. JAMA 2002 ; 288 : 2015 -2022. 15 Cayatte AJ, Polacino JJ, Horten K, Cohen RA. Chronic inhibition of nitric oxide production accelerates neointima formation and impairs endothelial function in hypercholesterolemic rabbits. Arterioscler Thromb 1994 ; 14 : 753 -759. 16 Ludmer PL, Selwyn AP, Shook TL, Wayne RR, Mudgo GH, Alexander RW, et al. Paradoxical vasoconstriction induced by acetylcholine in atherosclerotic coronary arteries. N Engl J Med 1986 ; 315 : 1046 -1051. 17 Yashimura M, Yassue H, Nakayama M, Shimaski Y, Sumida H, Sugiyama S, et al. A missense Glu298Asp variant in the endothelial nitric oxide synthase gene is associated with coronary spasm in the Japanese. Hum Genet 1998 ; 103 : 65 -69. 18 Hingorani AD, Liang CF, Fatibene J, Lyon A, Monteith S, Parsons A, et al. A common variant of the endothelial nitric oxide synthase (Glu298Asp) is a major risk for coronary artery disease in the UK. Circulation 1999 ; 100 : 1515 -1520. 19 Vallet-Colom I, Levy-Marchal C, Zarrouk D, Tichet J, Krishnamoorty R, Czermichow P, et al. HLA-DQ B1 codon 57 and genetic susceptibility to type I (insulin-dependent) diabetes mellitus in French children. Diabetologia 1990 ; 33 : 174 -176. 20 Ou T, Yamakawa-Kobayashi K, Arinami T, Amemiya H, Fujiwara H, Kawata K, et al. Methylenetetrahydrofolate reductase and apolipoprotein E polymorphisms are independent risk factors for coronary heart disease in Japanese: a case-control study. Atherosclerosis 1998 ; 137 : 23 -28. 21 Gardemann A, Weidemann H, Philipp M, Katz N, Tillmanns H, Wilhelm Hehrlein F, et al. The TT genotype of the methylenetetrahydrofolate reductase C677T gene polymorphism is associated with the extent of coronary atherosclerosis in patients at high risk for coronary artery disease. Eur Heart J 1999 ; 20 : 584 -592. 22 Lembo G, De Luca N, Battagli C, Iovino G, Aretini A, Musicco M, et al. A common variant of the endothelial nitric oxide synthase (Glu298Asp) is an independent risk factor for carotid atherosclerosis. Stroke 2001 ; 32 : 735 -740. 23 Szolnoki Z, Havasi V, Bene J, Komlosi K, Szoke D, Somogyvari F, et al. Endothelial nitric oxide synthase gene interactions and the risk of ischaemic stroke. Acta Neurol Scand 2005 ; 111 : 29 -33. 24 Shimasaki Y, Yasue H, Yashimura M, Nakayama M, Kugiyama K, Ogawa H, et al. Association of the missense Glu298Asp variant of the endothelial nitric oxide synthase gene with myocardial infarction. J Am Coll Cardiol 1998 ; 31 : 1506 -1510. 25 Heil SG, Den Hajer M, Van der rijt-pisa BJM, Kluijtmans LA, Blom HJ. The G894T variant of endothelial nitric oxide synthase (eNOS) increases the risk of recurrent venous thrombosis through interaction with elevated homocysteine levels. J Thromb Haemost 2004 ; 2 : 750 -753. 26 Hibi K, Lshigami T, Tamura K, Mizushima S, Nyui N, Fujita T, et al. Endothelial nitric oxide synthase gene polymorphism and acute myocardial infarction. Hypertension 1998 ; 32 : 521 -526. 27 Casas JP, Bautista LE, Humphries SE, Hingorani AD. Endothelial nitric oxide synthase genotype and ischemic heart disease: meta-analysis of 26 studies involving 23028 subjects. Circulation 2004 ; 109 : 1359 -1365. 28 Cai H, Wilken DE, Wang XL. The Glu298Asp (G894T) mutation at exon 7 of the endothelial nitric oxide synthase gene and coronary artery disease. J Mol Med 1999 ; 77 : 511 -514. 29 Granath B, Taylor RR, Van Bockxmeer FM, Mamotte CD. Lack of evidence for association between endothelial nitric oxide synthase gene polymorphisms and coronary artery disease in the Australian Caucasian population. J Cardiovasc Risk 2001 ; 8 : 235 -241. 30 Wang CL, Hsu LA, Ko YS, Ko YL, Lee YH. Lack of association between the Glu298Asp variant of the endothelial nitric oxide synthase gene and the risk of coronary artery disease among Taiwanese. J Formos Med Assoc 2001 ; 100 : 736 -740. 31 Schmoelzer I, Renner W, Paulweber B, Malaimare L, Iglseder B, Schmid P, et al. Lack of association between the Glu298Asp polymorphism of endothelial nitric oxide synthase with manifest coronary artery disease, carotid atherosclerosis and forearm vascular reactivity in two Austrian populations. Eur J Clin Invest 2003 ; 33 : 191 -198. 32 Colombo MG, Andreassi MG, Paradossi U, Botto N, Manfredi S, Masetti S, et al. Evidence for association of a common variant of the endothelial nitric oxide synthase gene (Glu298Asp polymorphism) to the presence, extent, and severity of coronary artery disease. Heart 2002 ; 87 : 525 -528. 33 Colombo MG, Paradossi U, Andreassi MG, Botto N, Manfredi S, Masetti S, et al. Endothelial nitric oxide synthase gene polymorphisms and risk of coronary artery disease. Clin Chem 2003 ; 49 : 389 -395. 34 Agema WRP, Maat MPM, Zwinderman AH, Kastelein JJP, Rabelink TJ, Boven AJ, et al. An integrated evaluation of endothelial constitutive nitric oxide synthase polymorphisms and coronary artery disease in men. Clin Sci 2004 ; 107 : 255 -261. 35 Yoon S, Shin C, Park HY, Moon J, Kim E, Kim HT, et al. Endothelial nitric oxide synthase gene is associated with vessel stenosis in Korean population. Clin Chim Acta 2005 ; 353 : 177 -185. 36 Brown KS, Kluijtmans LAJ, Young IS, Woodside J, Yarnell JWG, McMaster D, et al. Genetic evidence that nitric oxide modulates homocysteine, the NOS3 894TT genotype is a risk factor for hyperhomocysteinemia. Arterioscler Thromb Vasc Biol 2003 ; 23 : 1014 -1020. 37 Fisher PA, Dominguez GN, Cuniberti LA, Martinez V, Werba JP, Ramirez AJ, et al. Hyperhomocysteinemia induces renal hemodynamic dysfunction: is nitric oxide involved?. [Review] J Am Soc Nephrol 2003 ; 14 : 653 -660. 38 Stanger O, Weger M. Interactions of homocysteine, nitric oxide, folate and radicals in the progressively damaged endothelium. Clin Chem Lab Med 2003 ; 41 : 1444 -1454. 39 Dayal S, Arming E, Bottiglieri T, Boger RH, Sigmund CD, Faraci FM, et al. Cerebral vascular dysfunction mediated by superoxide in hyperhomocysteinemia mice. Stroke 2004 ; 35 : 1957 -1962. 40 Stamler JS, Osborne JA, Jaraki O, Rabbani LE, Mullins M, Singel D, et al. Adverse vascular effects of homocysteine are modulated by endothelium-derived relaxing factor and related oxides of nitrogen. J Clin Invest 1993 ; 91 : 308 -318. 41 Tesauro M, Thompson WC, Rogliani P, Qi L, Chaudhary PP, Moss J. Intracellular processing of endothelial nitric oxide synthase isoforms associated with differences in severity of cardiopulmonary diseases: cleavage of proteins with aspartate vs. glutamate at position 298. Proc Natl Acad Sci U S A 2000 ; 97 : 2832 -2835. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Stephan,, Carsten;Klaas,, Moritz;Müller,, Christian;Schnorr,, Dietmar;Loening, Stefan, A;Jung,, Klaus

doi: 10.1373/clinchem.2005.059170pmid: 16391327

Abstract Background: The comparability of total and free prostate-specific antigen (tPSA and fPSA) results among commercial PSA assays has been suggested to be improved by calibration to WHO PSA reference materials and the development of equimolar-response assays. To characterize the current situation, we assessed 5 frequently used commercial assay combinations for tPSA and fPSA regarding the interchangeability of the PSA values and the ratio of fPSA to tPSA (%fPSA), equimolar characteristics, and diagnostic accuracy. Methods: Sera from 314 patients with prostate cancer (PCa) and 282 men with no evidence of prostate cancer (NPCa) were measured with tPSA and fPSA assays from Abbott (AxSYM), Beckman Coulter (Access), Diagnostic Products Corporation (Immulite 2000), and Roche (Elecsys 2010) and with tPSA and complexed PSA (cPSA) assays from Bayer (ADVIA Centaur). Results: Method comparisons (Passing and Bablok regressions; Bland–Altman plots) showed assay-dependent results for tPSA, fPSA, and %fPSA. With the Access tPSA values taken as 100%, tPSA concentrations varied from 87% (AxSYM and ADVIA Centaur) to 115% (Immulite), leading to different numbers of patients classified according to the commonly recommended tPSA cutoffs for performing a biopsy. Different %fPSA values also led to assay-dependent ROC analysis results, a finding that shows the importance for the diagnostic accuracy. Conclusion: Interchangeability of tPSA, fPSA, and %fPSA values obtained by commercial PSA assays remains inadequate, but attention to this issue may minimize the misinterpretation of PSA results obtained by different assays. Different assays for concentrations of total and free prostate-specific antigen (tPSA and fPSA)2 provide discordant results [reviewed in Ref. (1)]. These assay-dependent variations could lead to misinterpretation of individual PSA values and erroneous clinical decisions about prostate carcinoma (PCa). Nonuniform assay calibration and nonequimolar detection of the various PSA forms are possible reasons for these discordant results (1)(2)(3)(4)(5)(6). PSA reference materials compiled by the WHO and equimolar-response assays were developed to adjust PSA calibration (7)(8)(9)(10)(11)(12), but since their introduction, several manufacturers (e.g., Beckman Coulter, Roche Diagnostics, and Diagnostic Products Corp.) have changed their assay platforms, and the few available comparison studies, with limited numbers of assays and small numbers of samples, have not fully characterized the current situation (13)(14). Therefore, focusing on the clinically important tPSA range up to 10 μg/L, we evaluated and characterized 5 frequently used fPSA and tPSA assay combinations with regard to the interchangeability of PSA values among the assays, their equimolar characteristics, and the diagnostic accuracy, estimated by ROC analyses, particularly of the percentage ratios of fPSA to tPSA (%fPSA). Materials and Methods study groups and samples We investigated archival sera collected between February 2001 and June 2004 from 596 untreated white men with tPSA concentrations of 0.49–10 μg/L (median, 5.14 μg/L) as determined with the Access Hybritech PSA system (Beckman Coulter). Only patients with at least 3 unthawed serum specimens were included in this retrospective study. The patients were classified into 2 groups: those with histologically confirmed PCa (314 men; median age, 66 years; range, 38–85 years) and those with no evidence of PCa on prostate biopsy (NPCa; 282 men; median age, 63 years; range, 43–79 years). Blood samples were collected in evacuated tubes (Sarstedt GmbH) in the Department of Urology or its outpatient division at the University Hospital Charité. The samples were taken before any diagnostic or therapeutic procedures involving the prostate and at least 4 weeks after digital rectal examination, prostatic biopsy, or transrectal ultrasound. Blood samples were allowed to clot for 1 h at room temperature and then were centrifuged at 1600g for 15 min at 4 °C. Sera were stored at −80 °C until analyzed. After thawing at room temperature, samples were processed within 3 h. The study was carried out in accordance with the standards of the local ethics board and the Helsinki Declaration of 1975 as revised in 1996, including informed consent obtained from all participants. psa assays and measurements We performed measurements in June to August 2004 in series of up to 30 samples, according to the manufacturers’ instructions on the following analyzers:, AxSYM (Abbott; tPSA, cat. no. 3C19-20; fPSA, cat. no. 3C 20-20), Elecsys 2010 (Roche Diagnostics; tPSA, cat. no. 11731262; fPSA, cat. no. 03289788), and ADVIA Centaur [Bayer Diagnostics; tPSA, cat. no. 118157; complexed PSA (cPSA), cat. no. 124830] on the same day as well as on the Access (Beckman Coulter; Hybritech PSA, cat. no. 37200; Hybritech fPSA, cat. no. 37210) and Immulite 2000 systems (Diagnostic Products Corp.; PSA, cat. no. L2KPS6; fPSA, cat. no. L2KPF2) on another day. The cPSA values were transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA, and the %fPSA values were calculated as percentage ratios of fPSA to tPSA. The between-run imprecision profiles of the measurements were estimated by use of control materials supplied by the manufacturers, commercial control materials, and in-house serum pools; all interassay CVs were <8% (n = 17–20 days). statistical analysis Data were analyzed with MedCalc (Ver. 8.1) and GraphPad Prism (Ver. 4.03 for Windows). Molar response plots to characterize the equimolarity of the assays were prepared according to Semjonow et al. (11). The Access Hybritech assay was used as the comparison method because its equimolarity is well characterized (11)(15). Significance was defined as P <0.05. Results and Discussion Calculations were made for all patients together and for the 2 groups (PCa and NPCa) separately to show potential differences depending on the clinical situation. The data are summarized in Tables 11 and 22 . The results of the method regression analyses according to Passing and Bablok (16) and the median values obtained with the various assays showed considerable differences among the assays (Table 11 ). These findings are illustrated in Figs. S1 and S2 [with the regression analyses and the Bland–Altman plots (17)], respectively, of the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol52/issue1/). Characteristic examples of the percentage difference plots of the %fPSA results of the various PSA assays are shown in Fig. 11 . When we set tPSA values obtained with the Access Hybritech as 100%, tPSA concentrations varied from 87% (AxSYM and ADVIA Centaur) to 115% (Immulite). When we used as cutoff a tPSA value of 4 μg/L, a limit commonly used for making the clinical decision to perform a prostate biopsy, the different assays classified distinctly different numbers of PCa or NPCa patients as either false negative or false positive (Table 22 ). For example, the Immulite and ADVIA Centaur systems did not give positive results for 46 and 93 of the 314 PCa patients, respectively, and falsely classified 144 and 120, respectively, of the 282 NPCa men as positive. If we had chosen threshold values lower than 4 μg/L, as recommended recently (18)(19)(20), similar differences would have occurred. ROC analyses for tPSA measured with the various assays (areas under the curves given in Table 22 ) did not reveal these clinically significant differences, suggesting that the differences were related to calibration differences that affect the cut point. fPSA and tPSA values differed inversely from the comparison method in some pairs of assays. For example, fPSA values obtained with the AxSYM were higher and Immulite fPSA values were lower than the Access Hybritech fPSA values, compared with the tPSA values. Thus, %fPSA values were clearly assay dependent, as indicated by the nonoverlapping 95% confidence intervals of the medians (Table 11 ), by the cutoffs at 90% sensitivity and specificity, and by the different ROC analysis results (Table 22 ). This finding is of particular clinical significance because the %fPSA values are generally used as tools to differentiate between patients with malignant and benign prostate diseases in the tPSA concentration gray zone of 2–10 μg/L. In addition, the differences in slopes between the PCa and NPCa patients for the calculated fPSA values and %fPSA obtained with the ADVIA Centaur assay (Table 11 ) are considerable. The molar response plots in Fig. 22 indicate that the AxSYM and ADVIA Centaur tPSA assays measured free and bound PSA in an equimolar fashion (Fig. 22 , A and B). The tPSA assays of Immulite and Elecsys showed nonequimolar characteristics with a slope significantly different from zero (Fig. 22 , C and D). These differences were not shown in previous studies with fewer serum samples (21) or with experiments using the WHO reference materials (13). For samples with %fPSA values <25%, the responses obtained with the Immulite and Elecsys tPSA assays were equimolar, as demonstrated by the slopes of the regression lines, which did not differ significantly from zero (P = 0.105 and 0.099 for Immulite and Elecsys). Our study documents the comparability among PSA assays from several manufacturers. The manufacturers generally claim that their PSA assays are calibrated against the WHO PSA reference material and that equimolar assays have been applied, procedures that should decrease the differences between assays (8)(10)(11)(12). Nevertheless, we found that the interassay variability could not be eliminated and that there was inadequate interchangeability of results. Although the study group was not representative for the prevalence of PCa in the population, the abbreviated diagnostic evaluation of the results also shows that this analytical limitation continues to have a severe impact on the clinical decision of whether to perform a prostate biopsy. Similar results were demonstrated in 2 recent studies (14)(22) and a population-based simulation study (23). To show the situation as it is, we did not recalibrate the assays using a common calibrant. According to our findings, the goal of assay-independent, interchangeable fPSA and tPSA results has not been achieved and may be unrealizable because of PSA heterogeneity, including structural diversity depending on malignant or benign origin, different analytical conditions based on the use of numerous antibodies with different epitope specificities and affinities (24), and the different technical principles underlying the various analyzers. To minimize the misinterpretation of PSA results obtained by different assays, clinical chemists should alert clinicians to the variation in assay results as well as the biological variation of PSA (25). Table 1. Method comparison of the various assays for tPSA, fPSA, and %fPSA with reference to the Access Hybritech assays characterized by the regression equations according to Passing and Bablok.1 . . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA, μg/L  All patients (n = 596) Median2 5.14 (4.88–5.54) 4.6 (4.26–4.89) 4.61 (4.30–4.84) 5.70 (5.40–6.06) 5.48 (5.10–5.71) Slope 1.00 0.87 (0.86–0.89) 0.87 (0.85–0.89) 1.15 (1.12–1.17) 1.01 (0.99–1.03) Intercept 0.00 0.03 (−0.04 to 0.11) 0.11 (0.05–0.17) −0.04 (−0.16 to 0.06) 0.15 (0.07–0.21)  NPCa (n = 282) Median 3.82 (3.08–4.46) 3.21 (2.80–3.88) 3.28 (2.88–3.78) 4.16 (3.40–5.03) 3.90 (3.49–4.65) Slope 1.00 0.84 (0.82–0.87) 0.85 (0.82–0.88) 1.11 (1.07–1.15) 1.03 (1.00–1.05) Intercept 0.00 0.07 (0.00–0.16) 0.14 (0.05–0.23) 0.07 (−0.07 to 0.17) 0.12 (0.03–0.20)  PCa (n = 314) Median 6.17 (5.83–6.59) 5.46 (5.02–5.74) 5.29 (4.98–5.56) 6.84 (6.50–7.27) 6.19 (5.84–6.55) Slope 1.00 0.88 (0.85–0.91) 0.89 (0.86–0.92) 1.16 (1.12–1.21) 1 (0.97–1.03) Intercept 0.00 0.05 (−0.11 to 0.18) 0.11 (−0.07 to 0.24) −0.12 (−0.33 to 0.10) 0.17 (0.02–0.31) fPSA, μg/L  All patients (n = 596) Median 0.65 (0.61–0.69) 0.70 (0.68–0.74) 0.683 (0.63–0.71) 0.58 (0.54–0.60) 0.63 (0.61–0.66) Slope 1.00 1.05 (1.02–1.08) 1.11 (1.05–1.18) 0.95 (0.93–0.98) 0.97 (0.95–1.00) Intercept 0.00 0.02 (0.00–0.03) −0.04 (−0.07 to 0.00) −0.02 (−0.04 to −0.01) 0.00 (−0.02 to 0.01)  NPCa (n = 282) Median 0.67 (0.60–0.73) 0.71 (0.64–0.78) 0.69 (0.62–0.73) 0.59 (0.53–0.63) 0.64 (0.58–0.70) Slope 1.00 1.06 (1.01–1.09) 1.01 (0.95–1.10) 0.95 (0.92–0.99) 0.95 (0.92–1.00) Intercept 0.00 0.01 (−0.02 to 0.03) 0.01 (−0.04 to 0.04) −0.02 (−0.04 to 0.00) 0.00 (−0.03 to 0.02)  PCa (n = 314) Median 0.64 (0.59–0.68) 0.69 (0.66–0.74) 0.67 (0.59–0.73) 0.57 (0.53–0.61) 0.63 (0.59–0.66) Slope 1.00 1.04 (1.00–1.08) 1.22 (1.12–1.34) 0.95 (0.92–1.00) 1.00 (0.95–1.03) Intercept 0.00 0.03 (0.01–0.05) −0.09 (−0.18 to −0.03) −0.03 (−0.06 to −0.01) −0.01 (−0.03 to 0.02) %fPSA4  All patients (n = 596) Median 14.5 (13.7–15.4) 17.9 (16.9–18.8) 17.7 (16.5–18.9) 11.4 (10.7–12.1) 13.8 (12.9–14.5) Slope 1.00 1.24 (1.20–1.27) 1.28 (1.21–1.35) 0.82 (0.80–0.85) 0.90 (0.87–0.93) Intercept 0.00 −0.17 (−0.64 to −0.26) −1.59 (−2.62 to −0.66) −0.27 (−0.62 to 0.07) 0.46 (0.12–0.86)  NPCa (n = 282) Median 18.9 (18.0–20.4) 23.7 (21.8–24.7) 23.4 (21.7–25.3) 15.0 (14.4–16.3) 17.7 (16.6–18.9) Slope 1.00 1.23 (1.17–1.29) 1.19 (1.10–1.28) 0.80 (0.76–0.85) 0.87 (0.83–0.91) Intercept 0.00 −0.14 (−1.13 to 0.72) −1.04 (−3.11 to 1.02) 0.06 (−0.75 to 0.81) 0.86 (0.06–1.59)  PCa (n = 314) Median 11.1 (10.5–12.0) 13.9 (13.0–14.9) 13.9 (12.8–14.6) 8.7 (8.07–9.30) 10.9 (10.2–11.5) Slope 1.00 1.27 (1.22–1.33) 1.57 (1.42–1.74) 0.85 (0.81–0.88) 0.99 (0.94–1.03) Intercept 0.00 −0.48 (−1.15 to 0.12) −4.05 (−6.08 to −2.72) −0.55 (−1.01 to −0.09) −0.35 (−0.88 to 0.13) . . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA, μg/L  All patients (n = 596) Median2 5.14 (4.88–5.54) 4.6 (4.26–4.89) 4.61 (4.30–4.84) 5.70 (5.40–6.06) 5.48 (5.10–5.71) Slope 1.00 0.87 (0.86–0.89) 0.87 (0.85–0.89) 1.15 (1.12–1.17) 1.01 (0.99–1.03) Intercept 0.00 0.03 (−0.04 to 0.11) 0.11 (0.05–0.17) −0.04 (−0.16 to 0.06) 0.15 (0.07–0.21)  NPCa (n = 282) Median 3.82 (3.08–4.46) 3.21 (2.80–3.88) 3.28 (2.88–3.78) 4.16 (3.40–5.03) 3.90 (3.49–4.65) Slope 1.00 0.84 (0.82–0.87) 0.85 (0.82–0.88) 1.11 (1.07–1.15) 1.03 (1.00–1.05) Intercept 0.00 0.07 (0.00–0.16) 0.14 (0.05–0.23) 0.07 (−0.07 to 0.17) 0.12 (0.03–0.20)  PCa (n = 314) Median 6.17 (5.83–6.59) 5.46 (5.02–5.74) 5.29 (4.98–5.56) 6.84 (6.50–7.27) 6.19 (5.84–6.55) Slope 1.00 0.88 (0.85–0.91) 0.89 (0.86–0.92) 1.16 (1.12–1.21) 1 (0.97–1.03) Intercept 0.00 0.05 (−0.11 to 0.18) 0.11 (−0.07 to 0.24) −0.12 (−0.33 to 0.10) 0.17 (0.02–0.31) fPSA, μg/L  All patients (n = 596) Median 0.65 (0.61–0.69) 0.70 (0.68–0.74) 0.683 (0.63–0.71) 0.58 (0.54–0.60) 0.63 (0.61–0.66) Slope 1.00 1.05 (1.02–1.08) 1.11 (1.05–1.18) 0.95 (0.93–0.98) 0.97 (0.95–1.00) Intercept 0.00 0.02 (0.00–0.03) −0.04 (−0.07 to 0.00) −0.02 (−0.04 to −0.01) 0.00 (−0.02 to 0.01)  NPCa (n = 282) Median 0.67 (0.60–0.73) 0.71 (0.64–0.78) 0.69 (0.62–0.73) 0.59 (0.53–0.63) 0.64 (0.58–0.70) Slope 1.00 1.06 (1.01–1.09) 1.01 (0.95–1.10) 0.95 (0.92–0.99) 0.95 (0.92–1.00) Intercept 0.00 0.01 (−0.02 to 0.03) 0.01 (−0.04 to 0.04) −0.02 (−0.04 to 0.00) 0.00 (−0.03 to 0.02)  PCa (n = 314) Median 0.64 (0.59–0.68) 0.69 (0.66–0.74) 0.67 (0.59–0.73) 0.57 (0.53–0.61) 0.63 (0.59–0.66) Slope 1.00 1.04 (1.00–1.08) 1.22 (1.12–1.34) 0.95 (0.92–1.00) 1.00 (0.95–1.03) Intercept 0.00 0.03 (0.01–0.05) −0.09 (−0.18 to −0.03) −0.03 (−0.06 to −0.01) −0.01 (−0.03 to 0.02) %fPSA4  All patients (n = 596) Median 14.5 (13.7–15.4) 17.9 (16.9–18.8) 17.7 (16.5–18.9) 11.4 (10.7–12.1) 13.8 (12.9–14.5) Slope 1.00 1.24 (1.20–1.27) 1.28 (1.21–1.35) 0.82 (0.80–0.85) 0.90 (0.87–0.93) Intercept 0.00 −0.17 (−0.64 to −0.26) −1.59 (−2.62 to −0.66) −0.27 (−0.62 to 0.07) 0.46 (0.12–0.86)  NPCa (n = 282) Median 18.9 (18.0–20.4) 23.7 (21.8–24.7) 23.4 (21.7–25.3) 15.0 (14.4–16.3) 17.7 (16.6–18.9) Slope 1.00 1.23 (1.17–1.29) 1.19 (1.10–1.28) 0.80 (0.76–0.85) 0.87 (0.83–0.91) Intercept 0.00 −0.14 (−1.13 to 0.72) −1.04 (−3.11 to 1.02) 0.06 (−0.75 to 0.81) 0.86 (0.06–1.59)  PCa (n = 314) Median 11.1 (10.5–12.0) 13.9 (13.0–14.9) 13.9 (12.8–14.6) 8.7 (8.07–9.30) 10.9 (10.2–11.5) Slope 1.00 1.27 (1.22–1.33) 1.57 (1.42–1.74) 0.85 (0.81–0.88) 0.99 (0.94–1.03) Intercept 0.00 −0.48 (−1.15 to 0.12) −4.05 (−6.08 to −2.72) −0.55 (−1.01 to −0.09) −0.35 (−0.88 to 0.13) 1 Values in parentheses are 95% confidence intervals. 2 Median values for all assays (tPSA, fPSA, %fPSA) were significantly different from the Access Hybritech assay values (Wilcoxon test of paired samples, P <0.0001) except for fPSA measured by ADVIA Centaur for all patients and for PCa patients (P <0.020 and <0.0287, respectively), whereas fPSA values for the NPCa patients measured with the ADVIA Centaur and with the Elecsys was not significantly different (P = 0.320 and 0.130, respectively). 3 The cPSA values were transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA. 4 %fPSA values were calculated as percentage ratios of fPSA to tPSA. Table 1. Method comparison of the various assays for tPSA, fPSA, and %fPSA with reference to the Access Hybritech assays characterized by the regression equations according to Passing and Bablok.1 . . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA, μg/L  All patients (n = 596) Median2 5.14 (4.88–5.54) 4.6 (4.26–4.89) 4.61 (4.30–4.84) 5.70 (5.40–6.06) 5.48 (5.10–5.71) Slope 1.00 0.87 (0.86–0.89) 0.87 (0.85–0.89) 1.15 (1.12–1.17) 1.01 (0.99–1.03) Intercept 0.00 0.03 (−0.04 to 0.11) 0.11 (0.05–0.17) −0.04 (−0.16 to 0.06) 0.15 (0.07–0.21)  NPCa (n = 282) Median 3.82 (3.08–4.46) 3.21 (2.80–3.88) 3.28 (2.88–3.78) 4.16 (3.40–5.03) 3.90 (3.49–4.65) Slope 1.00 0.84 (0.82–0.87) 0.85 (0.82–0.88) 1.11 (1.07–1.15) 1.03 (1.00–1.05) Intercept 0.00 0.07 (0.00–0.16) 0.14 (0.05–0.23) 0.07 (−0.07 to 0.17) 0.12 (0.03–0.20)  PCa (n = 314) Median 6.17 (5.83–6.59) 5.46 (5.02–5.74) 5.29 (4.98–5.56) 6.84 (6.50–7.27) 6.19 (5.84–6.55) Slope 1.00 0.88 (0.85–0.91) 0.89 (0.86–0.92) 1.16 (1.12–1.21) 1 (0.97–1.03) Intercept 0.00 0.05 (−0.11 to 0.18) 0.11 (−0.07 to 0.24) −0.12 (−0.33 to 0.10) 0.17 (0.02–0.31) fPSA, μg/L  All patients (n = 596) Median 0.65 (0.61–0.69) 0.70 (0.68–0.74) 0.683 (0.63–0.71) 0.58 (0.54–0.60) 0.63 (0.61–0.66) Slope 1.00 1.05 (1.02–1.08) 1.11 (1.05–1.18) 0.95 (0.93–0.98) 0.97 (0.95–1.00) Intercept 0.00 0.02 (0.00–0.03) −0.04 (−0.07 to 0.00) −0.02 (−0.04 to −0.01) 0.00 (−0.02 to 0.01)  NPCa (n = 282) Median 0.67 (0.60–0.73) 0.71 (0.64–0.78) 0.69 (0.62–0.73) 0.59 (0.53–0.63) 0.64 (0.58–0.70) Slope 1.00 1.06 (1.01–1.09) 1.01 (0.95–1.10) 0.95 (0.92–0.99) 0.95 (0.92–1.00) Intercept 0.00 0.01 (−0.02 to 0.03) 0.01 (−0.04 to 0.04) −0.02 (−0.04 to 0.00) 0.00 (−0.03 to 0.02)  PCa (n = 314) Median 0.64 (0.59–0.68) 0.69 (0.66–0.74) 0.67 (0.59–0.73) 0.57 (0.53–0.61) 0.63 (0.59–0.66) Slope 1.00 1.04 (1.00–1.08) 1.22 (1.12–1.34) 0.95 (0.92–1.00) 1.00 (0.95–1.03) Intercept 0.00 0.03 (0.01–0.05) −0.09 (−0.18 to −0.03) −0.03 (−0.06 to −0.01) −0.01 (−0.03 to 0.02) %fPSA4  All patients (n = 596) Median 14.5 (13.7–15.4) 17.9 (16.9–18.8) 17.7 (16.5–18.9) 11.4 (10.7–12.1) 13.8 (12.9–14.5) Slope 1.00 1.24 (1.20–1.27) 1.28 (1.21–1.35) 0.82 (0.80–0.85) 0.90 (0.87–0.93) Intercept 0.00 −0.17 (−0.64 to −0.26) −1.59 (−2.62 to −0.66) −0.27 (−0.62 to 0.07) 0.46 (0.12–0.86)  NPCa (n = 282) Median 18.9 (18.0–20.4) 23.7 (21.8–24.7) 23.4 (21.7–25.3) 15.0 (14.4–16.3) 17.7 (16.6–18.9) Slope 1.00 1.23 (1.17–1.29) 1.19 (1.10–1.28) 0.80 (0.76–0.85) 0.87 (0.83–0.91) Intercept 0.00 −0.14 (−1.13 to 0.72) −1.04 (−3.11 to 1.02) 0.06 (−0.75 to 0.81) 0.86 (0.06–1.59)  PCa (n = 314) Median 11.1 (10.5–12.0) 13.9 (13.0–14.9) 13.9 (12.8–14.6) 8.7 (8.07–9.30) 10.9 (10.2–11.5) Slope 1.00 1.27 (1.22–1.33) 1.57 (1.42–1.74) 0.85 (0.81–0.88) 0.99 (0.94–1.03) Intercept 0.00 −0.48 (−1.15 to 0.12) −4.05 (−6.08 to −2.72) −0.55 (−1.01 to −0.09) −0.35 (−0.88 to 0.13) . . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA, μg/L  All patients (n = 596) Median2 5.14 (4.88–5.54) 4.6 (4.26–4.89) 4.61 (4.30–4.84) 5.70 (5.40–6.06) 5.48 (5.10–5.71) Slope 1.00 0.87 (0.86–0.89) 0.87 (0.85–0.89) 1.15 (1.12–1.17) 1.01 (0.99–1.03) Intercept 0.00 0.03 (−0.04 to 0.11) 0.11 (0.05–0.17) −0.04 (−0.16 to 0.06) 0.15 (0.07–0.21)  NPCa (n = 282) Median 3.82 (3.08–4.46) 3.21 (2.80–3.88) 3.28 (2.88–3.78) 4.16 (3.40–5.03) 3.90 (3.49–4.65) Slope 1.00 0.84 (0.82–0.87) 0.85 (0.82–0.88) 1.11 (1.07–1.15) 1.03 (1.00–1.05) Intercept 0.00 0.07 (0.00–0.16) 0.14 (0.05–0.23) 0.07 (−0.07 to 0.17) 0.12 (0.03–0.20)  PCa (n = 314) Median 6.17 (5.83–6.59) 5.46 (5.02–5.74) 5.29 (4.98–5.56) 6.84 (6.50–7.27) 6.19 (5.84–6.55) Slope 1.00 0.88 (0.85–0.91) 0.89 (0.86–0.92) 1.16 (1.12–1.21) 1 (0.97–1.03) Intercept 0.00 0.05 (−0.11 to 0.18) 0.11 (−0.07 to 0.24) −0.12 (−0.33 to 0.10) 0.17 (0.02–0.31) fPSA, μg/L  All patients (n = 596) Median 0.65 (0.61–0.69) 0.70 (0.68–0.74) 0.683 (0.63–0.71) 0.58 (0.54–0.60) 0.63 (0.61–0.66) Slope 1.00 1.05 (1.02–1.08) 1.11 (1.05–1.18) 0.95 (0.93–0.98) 0.97 (0.95–1.00) Intercept 0.00 0.02 (0.00–0.03) −0.04 (−0.07 to 0.00) −0.02 (−0.04 to −0.01) 0.00 (−0.02 to 0.01)  NPCa (n = 282) Median 0.67 (0.60–0.73) 0.71 (0.64–0.78) 0.69 (0.62–0.73) 0.59 (0.53–0.63) 0.64 (0.58–0.70) Slope 1.00 1.06 (1.01–1.09) 1.01 (0.95–1.10) 0.95 (0.92–0.99) 0.95 (0.92–1.00) Intercept 0.00 0.01 (−0.02 to 0.03) 0.01 (−0.04 to 0.04) −0.02 (−0.04 to 0.00) 0.00 (−0.03 to 0.02)  PCa (n = 314) Median 0.64 (0.59–0.68) 0.69 (0.66–0.74) 0.67 (0.59–0.73) 0.57 (0.53–0.61) 0.63 (0.59–0.66) Slope 1.00 1.04 (1.00–1.08) 1.22 (1.12–1.34) 0.95 (0.92–1.00) 1.00 (0.95–1.03) Intercept 0.00 0.03 (0.01–0.05) −0.09 (−0.18 to −0.03) −0.03 (−0.06 to −0.01) −0.01 (−0.03 to 0.02) %fPSA4  All patients (n = 596) Median 14.5 (13.7–15.4) 17.9 (16.9–18.8) 17.7 (16.5–18.9) 11.4 (10.7–12.1) 13.8 (12.9–14.5) Slope 1.00 1.24 (1.20–1.27) 1.28 (1.21–1.35) 0.82 (0.80–0.85) 0.90 (0.87–0.93) Intercept 0.00 −0.17 (−0.64 to −0.26) −1.59 (−2.62 to −0.66) −0.27 (−0.62 to 0.07) 0.46 (0.12–0.86)  NPCa (n = 282) Median 18.9 (18.0–20.4) 23.7 (21.8–24.7) 23.4 (21.7–25.3) 15.0 (14.4–16.3) 17.7 (16.6–18.9) Slope 1.00 1.23 (1.17–1.29) 1.19 (1.10–1.28) 0.80 (0.76–0.85) 0.87 (0.83–0.91) Intercept 0.00 −0.14 (−1.13 to 0.72) −1.04 (−3.11 to 1.02) 0.06 (−0.75 to 0.81) 0.86 (0.06–1.59)  PCa (n = 314) Median 11.1 (10.5–12.0) 13.9 (13.0–14.9) 13.9 (12.8–14.6) 8.7 (8.07–9.30) 10.9 (10.2–11.5) Slope 1.00 1.27 (1.22–1.33) 1.57 (1.42–1.74) 0.85 (0.81–0.88) 0.99 (0.94–1.03) Intercept 0.00 −0.48 (−1.15 to 0.12) −4.05 (−6.08 to −2.72) −0.55 (−1.01 to −0.09) −0.35 (−0.88 to 0.13) 1 Values in parentheses are 95% confidence intervals. 2 Median values for all assays (tPSA, fPSA, %fPSA) were significantly different from the Access Hybritech assay values (Wilcoxon test of paired samples, P <0.0001) except for fPSA measured by ADVIA Centaur for all patients and for PCa patients (P <0.020 and <0.0287, respectively), whereas fPSA values for the NPCa patients measured with the ADVIA Centaur and with the Elecsys was not significantly different (P = 0.320 and 0.130, respectively). 3 The cPSA values were transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA. 4 %fPSA values were calculated as percentage ratios of fPSA to tPSA. Table 2. Diagnostic performance data given as numbers of patients in relation to the conventional tPSA threshold of 4 μg/L, cutoffs at 90% sensitivity and 90% specificity, and areas under the ROC curves for the various tPSA and %fPSA assays.1 . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA  NPCa >4 μg/L,2 n 139 117 120 144 139  PCa <4 μg/L,3 n 68 86 93 46 59  Cutoff at 90% sensitivity, μg/L 2.82 (2.40–3.40) 2.60 (2.12–3.04) 2.52 (2.12–3.07) 3.12 (2.68–3.75) 3.02 (2.48–3.54)  Cutoff at 90% specificity, μg/L 7.68 (7.21–8.21) 6.71 (6.20–7.05) 6.67 (6.26–7.32) 8.70 (8.33–9.15) 8.04 (7.38–8.60)  Area under the ROC curve 0.70 (0.66–0.74) 0.72 (0.68–0.75) 0.71 (0.67–0.75) 0.71 (0.68–0.75) 0.70 (0.66–0.74) %fPSA  Cutoff at 90% sensitivity, % 18.7 (17.2–20.0) 24.2 (22.0–25.7) 24.14 (21.8–25.8) 15.4 (14.5–17.1) 17.8 (17.0–19.1)  Cutoff at 90% specificity, % 10.2 (9.39–11.7) 12.8 (11.5–14.2) 9.94 (8.73–12.9) 8.50 (7.53–9.35) 9.94 (8.80–10.7)  Area under the ROC curve5 0.81 (0.78–0.84) 0.80 (0.77–0.83) 0.77 (0.74–0.81) 0.81 (0.77–0.84) 0.79 (0.75–0.82) . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA  NPCa >4 μg/L,2 n 139 117 120 144 139  PCa <4 μg/L,3 n 68 86 93 46 59  Cutoff at 90% sensitivity, μg/L 2.82 (2.40–3.40) 2.60 (2.12–3.04) 2.52 (2.12–3.07) 3.12 (2.68–3.75) 3.02 (2.48–3.54)  Cutoff at 90% specificity, μg/L 7.68 (7.21–8.21) 6.71 (6.20–7.05) 6.67 (6.26–7.32) 8.70 (8.33–9.15) 8.04 (7.38–8.60)  Area under the ROC curve 0.70 (0.66–0.74) 0.72 (0.68–0.75) 0.71 (0.67–0.75) 0.71 (0.68–0.75) 0.70 (0.66–0.74) %fPSA  Cutoff at 90% sensitivity, % 18.7 (17.2–20.0) 24.2 (22.0–25.7) 24.14 (21.8–25.8) 15.4 (14.5–17.1) 17.8 (17.0–19.1)  Cutoff at 90% specificity, % 10.2 (9.39–11.7) 12.8 (11.5–14.2) 9.94 (8.73–12.9) 8.50 (7.53–9.35) 9.94 (8.80–10.7)  Area under the ROC curve5 0.81 (0.78–0.84) 0.80 (0.77–0.83) 0.77 (0.74–0.81) 0.81 (0.77–0.84) 0.79 (0.75–0.82) 1 Values in parentheses are 95% confidence intervals. 2 Number of the 282 NPCa patients with tPSA values above the conventional threshold of 4 μg/L. 3 Number of the 314 PCa patients with tPSA values below the conventional threshold of 4 μg/L. 4 The cPSA values transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA were used to calculate %fPSA values. 5 Only the area obtained with the ADVIA Centaur assay was significantly different from the Access Hybritech value (pairwise comparison, P = 0.006). Table 2. Diagnostic performance data given as numbers of patients in relation to the conventional tPSA threshold of 4 μg/L, cutoffs at 90% sensitivity and 90% specificity, and areas under the ROC curves for the various tPSA and %fPSA assays.1 . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA  NPCa >4 μg/L,2 n 139 117 120 144 139  PCa <4 μg/L,3 n 68 86 93 46 59  Cutoff at 90% sensitivity, μg/L 2.82 (2.40–3.40) 2.60 (2.12–3.04) 2.52 (2.12–3.07) 3.12 (2.68–3.75) 3.02 (2.48–3.54)  Cutoff at 90% specificity, μg/L 7.68 (7.21–8.21) 6.71 (6.20–7.05) 6.67 (6.26–7.32) 8.70 (8.33–9.15) 8.04 (7.38–8.60)  Area under the ROC curve 0.70 (0.66–0.74) 0.72 (0.68–0.75) 0.71 (0.67–0.75) 0.71 (0.68–0.75) 0.70 (0.66–0.74) %fPSA  Cutoff at 90% sensitivity, % 18.7 (17.2–20.0) 24.2 (22.0–25.7) 24.14 (21.8–25.8) 15.4 (14.5–17.1) 17.8 (17.0–19.1)  Cutoff at 90% specificity, % 10.2 (9.39–11.7) 12.8 (11.5–14.2) 9.94 (8.73–12.9) 8.50 (7.53–9.35) 9.94 (8.80–10.7)  Area under the ROC curve5 0.81 (0.78–0.84) 0.80 (0.77–0.83) 0.77 (0.74–0.81) 0.81 (0.77–0.84) 0.79 (0.75–0.82) . Access . AxSYM . Centaur . Immulite . Elecsys . tPSA  NPCa >4 μg/L,2 n 139 117 120 144 139  PCa <4 μg/L,3 n 68 86 93 46 59  Cutoff at 90% sensitivity, μg/L 2.82 (2.40–3.40) 2.60 (2.12–3.04) 2.52 (2.12–3.07) 3.12 (2.68–3.75) 3.02 (2.48–3.54)  Cutoff at 90% specificity, μg/L 7.68 (7.21–8.21) 6.71 (6.20–7.05) 6.67 (6.26–7.32) 8.70 (8.33–9.15) 8.04 (7.38–8.60)  Area under the ROC curve 0.70 (0.66–0.74) 0.72 (0.68–0.75) 0.71 (0.67–0.75) 0.71 (0.68–0.75) 0.70 (0.66–0.74) %fPSA  Cutoff at 90% sensitivity, % 18.7 (17.2–20.0) 24.2 (22.0–25.7) 24.14 (21.8–25.8) 15.4 (14.5–17.1) 17.8 (17.0–19.1)  Cutoff at 90% specificity, % 10.2 (9.39–11.7) 12.8 (11.5–14.2) 9.94 (8.73–12.9) 8.50 (7.53–9.35) 9.94 (8.80–10.7)  Area under the ROC curve5 0.81 (0.78–0.84) 0.80 (0.77–0.83) 0.77 (0.74–0.81) 0.81 (0.77–0.84) 0.79 (0.75–0.82) 1 Values in parentheses are 95% confidence intervals. 2 Number of the 282 NPCa patients with tPSA values above the conventional threshold of 4 μg/L. 3 Number of the 314 PCa patients with tPSA values below the conventional threshold of 4 μg/L. 4 The cPSA values transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA were used to calculate %fPSA values. 5 Only the area obtained with the ADVIA Centaur assay was significantly different from the Access Hybritech value (pairwise comparison, P = 0.006). Figure 1. Open in new tabDownload slide Percentage difference plots of %fPSA values obtained with the various assays. Percentage differences between %fPSA values obtained with the AxSYM, ADVIA Centaur, Immulite 2000, or Elecsys 2010 assays and the Access Hybritech assay are plotted against the mean of the two assays according to Bland and Altman (17). The cPSA values obtained with the ADVIA Centaur analyzer (Bayer) were transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA, and the corresponding %fPSA values were calculated. Mean differences (solid lines) and upper and lower 1.96 × SD limits (dashed lines) are shown. Figure 1. Open in new tabDownload slide Percentage difference plots of %fPSA values obtained with the various assays. Percentage differences between %fPSA values obtained with the AxSYM, ADVIA Centaur, Immulite 2000, or Elecsys 2010 assays and the Access Hybritech assay are plotted against the mean of the two assays according to Bland and Altman (17). The cPSA values obtained with the ADVIA Centaur analyzer (Bayer) were transformed into fPSA concentrations by use of the equation fPSA = tPSA − cPSA, and the corresponding %fPSA values were calculated. Mean differences (solid lines) and upper and lower 1.96 × SD limits (dashed lines) are shown. Figure 2. Open in new tabDownload slide Molar response plots for total PSA assays. Percentage ratios of the total PSA concentrations obtained with the respective assays to the corresponding concentrations measured with the comparison method (Access Hybritech) are plotted against the %fPSA values obtained with the Access Hybritech assay. Assays with equimolar characteristics (A and B) show horizontal regression lines with slopes not significantly different from zero. Lines with positive slopes (C and D) indicate assays that overestimate fPSA. The equations of the regression lines (95% confidence intervals of the intercepts and slopes in parentheses) and the deviations of the slopes from zero with the corresponding P values are as follows: AxSYM (A), y = 0.04 (−0.08 to 0.15)x + 88.7 (86.5–90.9)% (P = 0.55); ADVIA Centaur (B), y = 0.11 (−0.01 to 0.24)x + 88.0 (85.7–90.2)% (P = 0.08); Immulite (C), y = 0.33 (0.16–0.51)x + 108 (105–111)% (P = 0.0002); Elecsys (D), y = 0.40 (0.26–0.53)x + 98.4 (95.9–100)% (P <0.0001). Figure 2. Open in new tabDownload slide Molar response plots for total PSA assays. Percentage ratios of the total PSA concentrations obtained with the respective assays to the corresponding concentrations measured with the comparison method (Access Hybritech) are plotted against the %fPSA values obtained with the Access Hybritech assay. Assays with equimolar characteristics (A and B) show horizontal regression lines with slopes not significantly different from zero. Lines with positive slopes (C and D) indicate assays that overestimate fPSA. The equations of the regression lines (95% confidence intervals of the intercepts and slopes in parentheses) and the deviations of the slopes from zero with the corresponding P values are as follows: AxSYM (A), y = 0.04 (−0.08 to 0.15)x + 88.7 (86.5–90.9)% (P = 0.55); ADVIA Centaur (B), y = 0.11 (−0.01 to 0.24)x + 88.0 (85.7–90.2)% (P = 0.08); Immulite (C), y = 0.33 (0.16–0.51)x + 108 (105–111)% (P = 0.0002); Elecsys (D), y = 0.40 (0.26–0.53)x + 98.4 (95.9–100)% (P <0.0001). 1 These authors contributed equally to this work. 2 Nonstandard abbreviations: tPSA, fPSA, and cPSA, total, free, and complexed prostate-specific antigen, respectively; PCa, prostate cancer; %fPSA, percentage ratio of fPSA to tPSA; and NPCa, no evidence of prostate cancer. We thank the manufacturers of the assays for free reagents provided with no further obligation. The work was also supported in parts by the Deutsche Forschungsgemeinschaft (Ju 365/5-1), SONNENFELD Foundation, Sparkassen Foundation, and Liselotte Beutel Foundation, Berlin. We thank Janett Reiche for excellent technical assistance. The study contains part of the doctoral thesis of M.K. 1 Semjonow A, De Angelis G, Oberpenning F, Schmid HP, Brandt B, Hertle L. The clinical impact of different assays for prostate specific antigen. BJU Int 2000 ; 86 : 590 -597. 2 Graves HC, Wehner N, Stamey TA. Comparison of a polyclonal and monoclonal immunoassay for PSA: need for an international antigen standard. J Urol 1990 ; 144 : 1516 -1522. 3 Zhou AM, Tewari PC, Bluestein BI, Caldwell GW, Larsen FL. Multiple forms of prostate-specific antigen in serum: differences in immunorecognition by monoclonal and polyclonal assays. Clin Chem 1993 ; 39 : 2483 -2491. 4 Stenman UH, Leinonen J, Zhang WM. Standardization of PSA determinations. Scand J Clin Lab Invest Suppl 1995 ; 221 : 45 -51. 5 Stamey TA. Second Stanford conference on international standardization of prostate-specific antigen immunoassays: September 1 and 2, 1994. Urology 1995 ; 45 : 173 -184. 6 Blijenberg BG, Storm BN, Kruger Boeken AE, Schröder FH. On the standardisation of total prostate-specific antigen: an exercise with two reference preparations. Clin Chem Lab Med 1999 ; 37 : 545 -552. 7 Sokoloff RL, Wolfert RL, Rittenhouse HG. Standardization of PSA immunoassays: proposals and practical limitations. J Clin Ligand Assay 1995 ; 18 : 86 -92. 8 Blijenberg BG, Storm BN, Van Zelst BD, Kruger AE, Schroder FH. New developments in the standardization of total prostate-specific antigen. Clin Biochem 1999 ; 32 : 627 -634. 9 Chan DW, Sokoll LJ. WHO first international standards for prostate-specific antigen: the beginning of the end for assay discrepancies?. Clin Chem 2000 ; 46 : 1291 -1292. 10 Rafferty B, Rigsby P, Rose M, Stamey T, Gaines DR. Reference reagents for prostate-specific antigen (PSA): establishment of the first international standards for free PSA and PSA (90:10). Clin Chem 2000 ; 46 : 1310 -1317. 11 Semjonow A, Oberpenning F, Weining C, Schon M, Brandt B, De Angelis G, et al. Do modifications of nonequimolar assays for total prostate-specific antigen improve detection of prostate cancer?. Clin Chem 2001 ; 47 : 1472 -1475. 12 Oberpenning F, Weining C, Brandt B, De Angelis G, Heinecke A, Hamm M, et al. A new modification of the Chiron ACS assay for total prostate-specific antigen achieves equimolar response characteristics and improves the detection of prostate cancer. Clin Chem Lab Med 2003 ; 41 : 90 -94. 13 Patel D, White PA, Milford WA. A comparison of six commercial assays for total and free prostate specific antigen (PSA): the predictive value of the ratio of free to total PSA. BJU Int 2000 ; 85 : 686 -689. 14 Link RE, Shariat SF, Nguyen CV, Farr A, Weinberg AD, Morton RA, et al. Variation in prostate specific antigen results from 2 different assay platforms: clinical impact on 2304 patients undergoing prostate cancer screening. J Urol 2004 ; 171 : 2234 -2238. 15 Laffin RJ, Chan DW, Tanasijevic MJ, Fischer GA, Markus W, Miller J, et al. Hybritech total and free prostate-specific antigen assays developed for the Beckman Coulter access automated chemiluminescent immunoassay system: a multicenter evaluation of analytical performance. Clin Chem 2001 ; 47 : 129 -132. 16 Passing H, Bablok W. A new biometrical procedure for testing the equality of measurements from two different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry, Part I. J Clin Chem Clin Biochem 1983 ; 21 : 709 -720. 17 Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999 ; 8 : 135 -160. 18 Törnblom M, Norming U, Adolfsson J, Becker C, Abrahamsson PA, Lilja H, et al. Diagnostic value of percent free prostate-specific antigen: retrospective analysis of a population-based screening study with emphasis on men with PSA levels less than 3.0 ng/mL. Urology 1999 ; 53 : 945 -950. 19 Catalona WJ, Ramos CG, Carvalhal GF, Yan Y. Lowering PSA cutoffs to enhance detection of curable prostate cancer. [Editorial] Urology 2000 ; 55 : 791 -795. 20 Parsons JK, Brawer MK, Cheli CD, Partin AW, Djavan R. Complexed prostate specific antigen (PSA) reduces unnecessary prostate biopsies in the 2.6-4. 0 ng/mL range of total PSA. BJU Int 2004 ; 94 : 47 -50. 21 Semjonow A, Oberpenning F, Brandt B, Zechel C, Brandau W, Hertle L. Impact of free-prostate specific antigen on discordant measurement results of assays for total prostate-specific antigen. Urology 1996 ; 48 (Suppl 6A): 10 -15. 22 Blijenberg BG, Yurdakul G, Van Zelst BD, Bangma CH, Wildhagen MF, Schroder FH. Discordant performance of assays for free and total prostate-specific antigen in relation to the early detection of prostate cancer. BJU Int 2001 ; 88 : 545 -550. 23 Roddam AW, Price CP, Allen NE, Ward AM. Assessing the clinical impact of prostate-specific antigen assay variability and nonequimolarity: a simulation study based on the population of the United Kingdom. Clin Chem 2004 ; 50 : 1012 -1016. 24 Stenman U-H, Paus E, Allard WJ, Andersson I, Andrès C, Barnett TR, et al. Summary report of the TD-3 workshop: characterization of 83 antibodies against prostate-specific antigen. Tumor Biol 1999 ; 20 (Suppl 1): 1 -12. 25 Soletormos G, Semjonow A, Sibley PE, Lamerz R, Petersen PH, Albrecht W, et al. Biological variation of total prostate-specific antigen: a survey of published estimates and consequences for clinical practice. Clin Chem 2005 ; 51 : 1342 -1351. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Watine,, Joseph;Friedberg,, Bruno;Nagy,, Eva;Onody,, Rita;Oosterhuis,, Wytze;Bunting, Peter, S;Charet,, Jean-Christophe;Horvath, Andrea, Rita

doi: 10.1373/clinchem.2005.056952pmid: 16391328

Abstract Background: It is not clear if good methodologic quality in current practice guidelines necessarily leads to more valid recommendations, i.e., those that are supported with consistent research evidence or, when evidence is conflicting or lacking, with sufficient consensus among the guideline development team. To help clarify this issue, we assessed whether there is a link between methodologic quality and recommendation validity in practice guidelines for the use of laboratory tests in the management of patients with non-small cell lung cancer (NSCLC). Methods: We conducted a systematic review of data on laboratory tests in NSCLC published in English or in French within the last 10 years and retrieved 11 practice guidelines for the use of these tests. The guidelines were critically appraised and scored for methodologic quality and recommendation validity based on the Appraisal of Guidelines Research and Evaluation (AGREE) criteria and on the systematic review. Results: Overall, these 11 guidelines had considerable shortcomings in methodologic quality and, to a lesser extent, in recommendation validity. Practice guidelines with the best methodologic quality were not necessarily the most valid in their recommendations, and conversely. Conclusions: Poor methodologic quality and lack of recommendation validity in laboratory medicine call for methodologic standards of guideline development and for international collaboration of guideline development agencies. We advise readers of guidelines to critically evaluate the methods used as well as the content of the recommendations before adopting them for use in practice. The methodologic quality of current practice guidelines must be improved (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). For many reasons, the methodologic quality of diagnostic guidelines is poorer than that of therapeutic guidelines, particularly in the field of laboratory medicine (10)(21)(22). It is not clear, however, whether these shortcomings in the methodology of guideline development necessarily lead to invalid recommendations, i.e., those that are not supported with consistent research evidence or sufficient consensus among the guideline development team when evidence is conflicting or lacking. We assessed to what extent methodologic quality is linked to recommendation validity in practice guidelines for the use of laboratory tests in the management of patients with non-small cell lung cancer (NSCLC),4 specifically laboratory tests measuring quantities in biological specimens, thus excluding tissue (or anatomic) pathology tests. Surgery performed at the early stages of NSCLC (I, II, or IIIA to a lesser extent) offers patients a reasonable chance of long-term survival, but this option is available to only a small minority of patients. In more advanced NSCLC (IIIB or higher), chemotherapy alone and chemotherapy with radiotherapy are options, but these therapies mostly aim to prolong patient survival, and the overwhelming majority of patients relapse. In many treatment facilities, no standard therapeutic schemes exist; therefore, controlled trials that include as many patients as possible are needed to assess the potential contribution of new drugs and new therapeutic schemes (23)(24)(25)(26). Demonstrating the superiority of a given protocol over another is difficult to accomplish if the prognostic features of different patient subgroups cannot be compared. Consequently, independent prognostic factors must be identified before valid therapeutic trials can be designed, conducted, and interpreted (27). The medical and scientific communities have developed methods for conducting and reporting such therapeutic trials (28). Materials and Methods search for and selection of guidelines The general principles of our manual and electronic search strategies have been described previously (21). All practice guidelines published in English or French within the last 10 years that provided advice for the use of laboratory tests in the management of NSCLC patients were selected for appraisal (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39). recommendations in practice guidelines Two of us (J.W. and B.F.) extracted all laboratory-related recommendations from the 11 guidelines selected for review (Table 11 ). In Table 11 , the term “unclear recommendation” indicates either that the clinical decisions to be made based on the results of the recommended laboratory investigations were not precisely specified or that the names of the recommended laboratory tests themselves were not specified (e.g., the general term “biochemistry tests” was used). Four organizations, the CIGNA HealthCare Medicare Administration (CIGNA), the European Group on Tumor Markers (EGTM), the National Academy of Clinical Biochemistry (NACB), and the Société de Pneumologie de Langue Française (SPLF), focused on tumor markers only. SPLF classifies 2 tumor markers in NSCLC, carcinoembryonic antigen (CEA) and cyfra 21-1, according to their “levels of scientific evidence”. SPLF considers the CEA level of scientific evidence as “not sufficient” for prognosis, staging, or surveillance, whereas the cyfra 21-1 level of evidence is considered sufficient for prognosis but not for staging or surveillance. EGTM recommends the measurement of cyfra 21-1 before therapy and during posttherapy follow-up in NSCLC patients, and of CEA in cases of adenocarcinoma or large cell carcinoma. EGTM also stresses the independent prognostic value of cyfra 21-1, CEA, and CA 125 and of cyfra 21-1, CEA, and tissue-polypeptide antigen for monitoring therapy efficacy in NSCLC. The NACB guidelines are very similar to those of EGTM except that they do not as clearly recommend that cyfra 21-1 or CEA be measured in NSCLC patients. CIGNA does not recommend the measurement of CEA, neuron-specific enolase (NSE), or cyfra 21-1. The 7 other organizations, the American College of Chest Physicians (ACCP), Agence Nationale pour le Développement de l’Evaluation Médicale (ANDEM), American Society of Clinical Oncology (ASCO), American Thoracic Society, and European Respiratory Society (ATS-ERS), British Thoracic Society and Society of Cardiothoracic Surgeons of Great Britain and Ireland (BTS-SCG), Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC), and the Scottish Intercollegiate Guidelines Network (SIGN) recommend the measurement of several different laboratory variables for the pretreatment evaluation of NSCLC patients (Table 11 ). ATS-ERS stresses the pretreatment prognostic significance of serum albumin and, to a lesser extent, that of serum calcium, particularly in case of advanced disease, whereas SIGN stresses the pretreatment prognostic significance of calcium, alkaline phosphatase (ALP), and liver function tests, and ASCO stresses the importance of lactate dehydrogenase (LD), hemoglobin, and leukocyte counts. ANDEM, ATS-ERS, and FNCLCC do not recommend the routine measurement of any laboratory variables other than those mentioned in Table 11 , including serum tumor markers, in NSCLC patients. systematic review of the evidence We previously carried out a systematic review of the evidence (40)(41), which we updated last year (42). Recommendations in the 11 guidelines about the use of laboratory tests in NSCLC (Table 11 ) were compared with the findings of our systematic reviews (40)(41)(42). In the management of NSCLC patients, laboratory tests can be useful in relation either to the disease itself or to the therapies administered. In relation to the therapies administered, if “routine chemistries and hematological tests” were to be taken into account, as summarized in Table 11 , virtually all 7 practice guidelines dealing with nontumor markers (ACCP, ANDEM, ASCO, ATS-ERS, BTS-SCG, FNCLCC, and SIGN) would probably agree that to evaluate toxicity or tolerance to the therapies administered to NSCLC patients, it may be necessary to measure hemoglobin, leukocyte counts, platelets, electrolytes, glucose, creatinine, transaminases, bilirubin, and albumin. On the basis of such a consensus, we have therefore considered that it is valid to recommend the measurements of these variables in NSCLC patients, particularly in patients suffering from advanced stages of disease. In relation to the disease itself, almost all authors agree that laboratory tests (excluding pathology tests) currently have no clinical utility for NSCLC screening or diagnosis. In addition, 2 prognostic covariables are universally used in NSCLC patients: disease stage and performance status (23)(24)(25). Among other prognostic covariables that can be used for the stratification of NSCLC patients in trials, some authors use patient age and sex. Our systematic reviews of the evidence (40)(41)(42) indicated that the pretreatment prognostic values of blood hemoglobin, leukocyte counts with differential, serum LD, albumin, calcium, and, to a lesser extent, NSE are very likely to be independent of the aforementioned other covariables. There thus is sufficient evidence for recommending the measurement of at least all of these laboratory variables in all NSCLC patients participating in therapeutic trials (42). We also consider it valid, based on the evidence, to recommend the measurement of blood hemoglobin in patients treated with radiotherapy (either inside or outside therapeutic trials), because the outcomes of patients with low blood hemoglobin are very likely to improve if they receive erythropoietin before radiation therapy (42)(43). These guidelines also suggest that laboratory tests might be useful for the staging (pretreatment prognostic evaluation) and surveillance (posttreatment prognostic evaluation) of NSCLC. Seven guidelines (ACCP, ANDEM, ASCO, ATS-ERS, BTS-SCG, FNCLCC, and SIGN) thus recommend the use of diverse biochemistry and/or hematology tests for the staging of NSCLC. Abnormal result(s) in a patient with otherwise resectable (stage IIIA, or lower) NSCLC might indicate the presence of metastases and unresectable disease; therefore, all putative metastatic sites must be carefully investigated. All of the guidelines do not agree, however, on which laboratory tests are necessary under these circumstances. Some recommend the use of calcium and/or albumin, whereas others recommend ALP and/or LD (Table 11 ). Other guidelines (e.g., those of EGTM, NACB, or SPLF to a lesser extent) extend this recommendation to cyfra 21-1 (44). According to our own systematic review, however, there is more evidence to support the use of routine biochemical and hematologic tests (i.e., leukocyte counts with differential, LD, albumin, and calcium), or even to a lesser extent NSE, than cyfra 21-1, thus confirming that some new laboratory tests (e.g., cyfra 21-1 or other tumor markers for the management of NSCLC) may be introduced into routine practice before they are demonstrated to have greater validity than older, less expensive tests (45)(46). We therefore consider there to be sufficient evidence for recommending the measurement of leukocyte counts with differential, serum LD, albumin, and calcium, but that it is not valid to recommend the measurement of tumor markers in NSCLC patients, except perhaps for NSE in patients in chemotherapy trials. In summary, available evidence suggests that the laboratory tests indicated in Table 22 should be performed for the pretreatment evaluation of NSCLC patients. Some laboratory tests recommended by one or several experts in the 11 practice guidelines are not mentioned in Table 22 , e.g., ALP for the staging of NSCLC, because according to the systematic review of the evidence, it is quite clear that the prognostic value of ALP is inferior to, and is not independent of, that of LD, as summarized in Table 33 . appraisal of guidelines Scores for methodologic quality were assigned to each of the 11 guidelines (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39), based on their critical appraisal using the Appraisal of Guidelines Research and Evaluation (AGREE) Instrument. The AGREE Instrument comprises 23 criteria, arranged in 6 domains (as shown in Table 44 ), covering the key elements of the guideline development process (47). Among the published appraisal checklists for practice guidelines, we chose the AGREE Instrument because it has shown the greatest potential as a tool for assessing recommendations for clinical pathways (48). This instrument has been endorsed by the WHO and the European Commission (21). In accordance with the AGREE recommendations, we assigned to each guideline 1 of 4 possible overall final scores: “very good” (the equivalent of “strongly recommend” of the AGREE Instrument) if the guideline rated high on the majority of items and most domain scores were >60%, indicating that the guideline had a high overall quality and could be considered for use in practice without alterations; “good” (the equivalent of “recommend with provisos or alterations” of the AGREE Instrument) if the guideline rated high or low on a similar number of items and most domain scores were 30%–60%, indicating that the guideline had a moderate overall quality; “not so good” (the equivalent of “would not recommend” of the AGREE Instrument) if the guideline rated low on the majority of items and most domain scores were <30%, indicating that the guideline had a low overall quality and serious shortcomings and thus should not be recommended for use in practice; and finally, “dubious” (the equivalent of “unsure” of the AGREE Instrument) if the guideline did not give sufficient information to enable us to assess its quality. We have chosen to use the scores “very good”, “good”, “not so good”, or “dubious”, rather than the original terminology of the AGREE Instrument, as indicated above, because we thought that this would lead to an easier understanding of our review. Scores for validity of recommendations were also assigned to each guideline, based on a systematic review of the evidence (40), which has been updated twice (41)(42), also taking into account the consensual opinions of the guideline development teams, as summarized above in the section Systematic Review of the Evidence. The scale of ratings that we used was the same as for methodologic quality, consisting of 4 possible scores (very good, good, not so good, or dubious), as explained in more detail in the Results section. Two scores for validity of recommendations were assigned to each guideline: one for recommendations regarding tumor markers and one for recommendations regarding other laboratory tests. how disagreements were resolved During the whole process of the study described in the 4 sections above (Search for and Selection of Guidelines, Recommendations in Practice Guidelines, Systematic Review of the Evidence, and Appraisal of Guidelines), disagreements between the 2 assessors (J.W. and B.F.) were resolved by consensus, and if necessary, a third person (J.C.C.) was available as a referee (this was never necessary). For methodologic quality, the consensual scores thus obtained were validated by an independent set of assessors (E.N. and R.O.), and when necessary a third person (A.R.H.) was used as a referee (this was necessary only once). Results scores for methodologic quality Each guideline was scored for each of the 6 domains of the AGREE Instrument, as shown in Table 44 . As can be seen in Table 44 , some guidelines are better in some aspects and others in other aspects of guideline development methodology. The overall final scores obtained for each guideline are indicated in Table 55 . scores for validity of recommendations The only guideline that clearly recommended the use of tumor markers (EGTM) was scored as not so good because there is no evidence that measurement of tumor markers in routine practice would improve NSCLC patient outcomes. The EGTM guideline was not attributed the worst possible score (dubious) because, as already stressed, tumor markers might be useful in therapeutic trials. The 4 guidelines that clearly did not recommend the use of tumor markers in routine practice (ANDEM, ATS-ERS, CIGNA, and FNCLCC) were scored as good. These 4 guidelines were not attributed the best possible score (very good) because they did not allude to the needs of patients in therapeutic trials. Guidelines that gave unclear recommendations regarding the use of tumor markers (NACB and SPLF) or that did not mention tumor markers at all (ACCP, ASCO, BTS-SCG, and SIGN) were also scored as not so good (Table 55 ). Regarding the other laboratory tests, the 5 guidelines (ACCP, ANDEM, ASCO, FNCLCC, and ATS-ERS) in which only a few laboratory tests were missing among those recommended (compared with the reference list of tests in Table 22 ) were scored as good. SIGN and BTS-SCG guidelines were scored as not so good because their lists of recommended laboratory tests were clearly less evidence based than those in the 5 other guidelines (as can be seen in Table 11 ). Discussion and Conclusions It is universally acknowledged that the methodologic quality of practice guidelines must improve (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22). For example, in a study that was probably one of the largest surveys ever done on the subject, Grilli et al. (7) assessed the quality of 431 practice guidelines produced by specialty societies and published in English-language peer-reviewed journals in 1988–1998. Their assessment included 3 criteria: whether the report provided information on the types of professionals and stakeholders involved in the development process; the strategy to identify primary evidence; and an explicit grading of recommendations according to the quality of supporting evidence (7). These 3 criteria were not met at all in 67%, 88%, and 82% of these 431 guidelines, respectively, and all 3 criteria for quality were met in only 22 guidelines (5%) (7). It is not clear whether this situation will improve in the near future (49). When we used the AGREE Instrument to assess the methodologic quality of 11 practice guidelines providing advice for the use of laboratory tests for the management of NSCLC, many fell short of basic quality criteria, a result that confirms the aforementioned observations (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22). Regarding our judgment that some of the recommendations made in the 11 guidelines are not entirely valid, it could be argued that our judgment might be incorrect because even evidence-based guidelines can contain different recommendations. Scientific evidence is only one of many factors that may influence the translation of research findings to the context of use. The process of considered judgment is essential in guideline development and often requires extensive discussions and consensus among experts (50). Availability of services, resources, and cost-effectiveness are important considerations (22) but do not apply to our study (Table 11 ). Recommendations about tumor markers are conflicting, and availability of services, resources, and cost-effectiveness do not support the use of any test that lacks evidence of clinical utility. In addition, some recommendations made in these 11 guidelines are unclear. Regarding these unclear recommendations, the 3 guidelines (ACCP, ASCO, and SIGN) in which the measurement of “other (routine) laboratory tests” was recommended (quite a vague recommendation indeed) offered both therapeutic and diagnostic recommendations; it is therefore possible that guideline development teams followed the right guideline development methods regarding therapeutic recommendations but did less well when formulating diagnostic ones. The AGREE Instrument, as a generic appraisal toolbox, did not allow us to investigate this possibility in more depth, and it was not possible for us to conclude whether the methodologic quality or validity of recommendations was better or worse in the 4 guidelines that offer both therapeutic and diagnostic recommendations than in the 7 other, purely diagnostic, guidelines, although the 3 guidelines that obtained the lowest possible scores regarding methodologic quality (dubious) were purely diagnostic guidelines (Table 55 ). In summary, the clinical validity of the recommendations regarding laboratory tests made in some of the 11 guidelines can be questioned (at least in those scoring not so good in Table 55 ). Some authors have shown that guidelines of poor methodologic quality are more likely to provide invalid diagnostic recommendations than guidelines of high methodologic quality (10)(18), but other authors have shown that guidelines of poor methodologic quality can provide diagnostic recommendations as valid as guidelines with high methodologic quality (2). In our study, practice guidelines with the best methodologic quality were not necessarily those that were the most valid in the content of their recommendations, and conversely. For example, ATS-ERS and CIGNA guidelines were valid in their recommendations, whereas their methodologic quality was poor, and SIGN and BTS-SCG guidelines were not valid in their recommendations, whereas their methodologic quality was good (Table 55 ). Because the AGREE Instrument does not involve the use of global scores to assess the methodologic quality of guidelines, we also looked at the individual scores obtained not only in each of the 6 domains (Table 44 ) but also in each of the 23 questions (data not shown), and again we were not able to establish any correlation between validity of content and methodologic quality in any of the 6 domains or in any of the 23 questions. This result is worrisome because the busy practitioner confronted with conflicting guideline recommendations has no easy means to help in deciding which guideline should be trusted. Conflicting recommendations on the use of laboratory tests are likely to lead to a waste of laboratory resources and might even cause harm to patients (51). Effective treatment depends on the effective use of diagnostic tests, and if diagnostic recommendations are not evidence based, it is reasonable to assume that therapeutic interventions will sometimes be initiated and monitored inappropriately. Fortunately, the shortcomings in methodologic quality seemed to be more frequent than those of the content validity (see Table 55 ). The discrepancy between methodologic guideline quality and clinical validity of recommendations is perhaps less obvious in therapeutic recommendations, in which the quality of evidence from randomized trials is higher, than in diagnostic recommendations, in which the level of evidence is generally much poorer. Another possibility is that in other areas of medicine, more valid laboratory recommendations in practice guidelines are available than in NSCLC. Whatever the true situation is, the results of our study call for the critical appraisal of guidelines providing both diagnostic and therapeutic recommendations in various medical areas. Such work is in progress within our team in the field of diabetes mellitus (21). On the basis of our studies and reports from the literature, however, we strongly advise colleagues to do similar studies in other areas of medicine before guideline recommendations are used in local practice. Because FNCLCC guidelines obtained the best scores in all items used for comparison (Table 55 ), one could argue that the French authors of the present review were biased toward guidelines in their own language. Taking into account the fact that we had the opportunity of expert discussions with some of the authors of the FNCLCC guidelines before their guidelines were published [these discussions have partly been published (52)], we rather believe that the authors of the FNCLCC guidelines are much more likely than the authors of the other guidelines [except perhaps for the authors of NACB guidelines who quoted us (37)] to have read and (partly) taken into account the results of systematic reviews available on this topic (40)(41)(42)(52). According to Shekelle et al. (53), the point at which no more than 90% of the guidelines published by the US Agency for Healthcare Research and Quality are still valid is 3.6 years (95% confidence interval, 2.6–4.6 years). To assess this hypothesis, we checked whether there was better correlation between methodologic quality and validity of recommendations in the most recently published guidelines. We failed to find any difference (Table 55 ), which suggests that the hypothesis of Shekelle et al. (53) may be valid for US Agency for Healthcare Research and Quality guidelines and similar sorts of guidelines, particularly in those areas in which the evidence base of recommendations develops faster than in the field we investigated. In conclusion, to overcome the methodologic shortcomings of current practice guidelines and to improve the validity of resulting recommendations, standardized methods for making evidence-based guideline recommendations in laboratory medicine must be disseminated. In particular, we need a unified system for grading diagnostic recommendations [such a work is in progress within the GRADE group (54)], as well as common standards for guideline reporting [such a work also seems to be in progress (55)], together with appropriate tools for guideline implementation. Finally, we need to educate our profession about the principles of evidence-based laboratory medicine and guideline development methods (22). We advise that guidelines be critically evaluated for methodology and content before recommendations are used in clinical practice. Table 1. Recommendations made in 11 clinical practice guidelines providing advice for the use of laboratory variables in the pretreatment management of NSCLC patients (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39). Guidelines . Recommended . Unclear recommendation . Not recommended . ACCP Hematocrit, ALP, calcium, electrolytes, glucose, GGT,1 SGOT Other routine laboratory tests None ANDEM Leukocyte count, albumin, SR, calcium, ALP, LD None Tumor markers ASCO Hemoglobin, leukocyte counts, LD, ALP, calcium Other routine chemistries; liver function tests LASA, CA 19-9, DNA index, DNA flow cytometric proliferation analysis, p53 tumor suppressor gene, ras oncogene ATS-ERS Blood counts, electrolytes, albumin, calcium, ALP, transaminases, bilirubin, creatinine None Tumor markers BTS-SCG Albumin, creatinine, glucose None None CIGNA2 None None CEA, NSE, cyfra 21-1 EGTM2 cyfra 21-1, CEA3 CA 125, TPA None FNCLCC Hemoglobin, leukocyte counts with differential, LD, albumin, calcium None Tumor markers NACB2 None cyfra 21-1, CEA, NSE None SIGN ALP, calcium Other biochemistry and hematology tests; liver function tests None SPLF2 None cyfra 21-1 CEA Guidelines . Recommended . Unclear recommendation . Not recommended . ACCP Hematocrit, ALP, calcium, electrolytes, glucose, GGT,1 SGOT Other routine laboratory tests None ANDEM Leukocyte count, albumin, SR, calcium, ALP, LD None Tumor markers ASCO Hemoglobin, leukocyte counts, LD, ALP, calcium Other routine chemistries; liver function tests LASA, CA 19-9, DNA index, DNA flow cytometric proliferation analysis, p53 tumor suppressor gene, ras oncogene ATS-ERS Blood counts, electrolytes, albumin, calcium, ALP, transaminases, bilirubin, creatinine None Tumor markers BTS-SCG Albumin, creatinine, glucose None None CIGNA2 None None CEA, NSE, cyfra 21-1 EGTM2 cyfra 21-1, CEA3 CA 125, TPA None FNCLCC Hemoglobin, leukocyte counts with differential, LD, albumin, calcium None Tumor markers NACB2 None cyfra 21-1, CEA, NSE None SIGN ALP, calcium Other biochemistry and hematology tests; liver function tests None SPLF2 None cyfra 21-1 CEA 1 GGT, γ-glutamyl transferase; SGOT, glutamic-oxaloacetic transpeptidase; SR, sedimentation rate; LASA, lipid-associated sialic acid; TPA, tissue-polypeptide antigen. 2 Guidelines intended for tumor markers only. 3 Only in cases of adenocarcinoma or large cell carcinoma. Table 1. Recommendations made in 11 clinical practice guidelines providing advice for the use of laboratory variables in the pretreatment management of NSCLC patients (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39). Guidelines . Recommended . Unclear recommendation . Not recommended . ACCP Hematocrit, ALP, calcium, electrolytes, glucose, GGT,1 SGOT Other routine laboratory tests None ANDEM Leukocyte count, albumin, SR, calcium, ALP, LD None Tumor markers ASCO Hemoglobin, leukocyte counts, LD, ALP, calcium Other routine chemistries; liver function tests LASA, CA 19-9, DNA index, DNA flow cytometric proliferation analysis, p53 tumor suppressor gene, ras oncogene ATS-ERS Blood counts, electrolytes, albumin, calcium, ALP, transaminases, bilirubin, creatinine None Tumor markers BTS-SCG Albumin, creatinine, glucose None None CIGNA2 None None CEA, NSE, cyfra 21-1 EGTM2 cyfra 21-1, CEA3 CA 125, TPA None FNCLCC Hemoglobin, leukocyte counts with differential, LD, albumin, calcium None Tumor markers NACB2 None cyfra 21-1, CEA, NSE None SIGN ALP, calcium Other biochemistry and hematology tests; liver function tests None SPLF2 None cyfra 21-1 CEA Guidelines . Recommended . Unclear recommendation . Not recommended . ACCP Hematocrit, ALP, calcium, electrolytes, glucose, GGT,1 SGOT Other routine laboratory tests None ANDEM Leukocyte count, albumin, SR, calcium, ALP, LD None Tumor markers ASCO Hemoglobin, leukocyte counts, LD, ALP, calcium Other routine chemistries; liver function tests LASA, CA 19-9, DNA index, DNA flow cytometric proliferation analysis, p53 tumor suppressor gene, ras oncogene ATS-ERS Blood counts, electrolytes, albumin, calcium, ALP, transaminases, bilirubin, creatinine None Tumor markers BTS-SCG Albumin, creatinine, glucose None None CIGNA2 None None CEA, NSE, cyfra 21-1 EGTM2 cyfra 21-1, CEA3 CA 125, TPA None FNCLCC Hemoglobin, leukocyte counts with differential, LD, albumin, calcium None Tumor markers NACB2 None cyfra 21-1, CEA, NSE None SIGN ALP, calcium Other biochemistry and hematology tests; liver function tests None SPLF2 None cyfra 21-1 CEA 1 GGT, γ-glutamyl transferase; SGOT, glutamic-oxaloacetic transpeptidase; SR, sedimentation rate; LASA, lipid-associated sialic acid; TPA, tissue-polypeptide antigen. 2 Guidelines intended for tumor markers only. 3 Only in cases of adenocarcinoma or large cell carcinoma. Table 2. Laboratory variables that should be measured for the pretreatment evaluation of NSCLC patients, taking into account both the systematic review of the evidence and the consensual opinions of the development teams of current practice guidelines (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39). Purpose of test . Variables to be measured . Evaluation of toxicity (or tolerance) to treatments In all patients: Hemoglobin,1 leukocyte counts with differential,1 platelets,1 electrolytes,1 glucose,1 creatinine,1 transaminases,1 bilirubin,1 albumin1 Pretreatment prognostic evaluation In all patients: Hemoglobin (if radiation therapy),2 leukocyte counts with differential,1 LD,1 albumin,1 calcium1 In patients participating in therapeutic trials: Hemoglobin,2 leukocyte counts with differential,2 LD,2 albumin,2 calcium,2 NSE2 Purpose of test . Variables to be measured . Evaluation of toxicity (or tolerance) to treatments In all patients: Hemoglobin,1 leukocyte counts with differential,1 platelets,1 electrolytes,1 glucose,1 creatinine,1 transaminases,1 bilirubin,1 albumin1 Pretreatment prognostic evaluation In all patients: Hemoglobin (if radiation therapy),2 leukocyte counts with differential,1 LD,1 albumin,1 calcium1 In patients participating in therapeutic trials: Hemoglobin,2 leukocyte counts with differential,2 LD,2 albumin,2 calcium,2 NSE2 1 Recommendation validated by most experts and not contradicted by systematic review of the evidence. 2 Recommendation based on systematic review of the evidence. Table 2. Laboratory variables that should be measured for the pretreatment evaluation of NSCLC patients, taking into account both the systematic review of the evidence and the consensual opinions of the development teams of current practice guidelines (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39). Purpose of test . Variables to be measured . Evaluation of toxicity (or tolerance) to treatments In all patients: Hemoglobin,1 leukocyte counts with differential,1 platelets,1 electrolytes,1 glucose,1 creatinine,1 transaminases,1 bilirubin,1 albumin1 Pretreatment prognostic evaluation In all patients: Hemoglobin (if radiation therapy),2 leukocyte counts with differential,1 LD,1 albumin,1 calcium1 In patients participating in therapeutic trials: Hemoglobin,2 leukocyte counts with differential,2 LD,2 albumin,2 calcium,2 NSE2 Purpose of test . Variables to be measured . Evaluation of toxicity (or tolerance) to treatments In all patients: Hemoglobin,1 leukocyte counts with differential,1 platelets,1 electrolytes,1 glucose,1 creatinine,1 transaminases,1 bilirubin,1 albumin1 Pretreatment prognostic evaluation In all patients: Hemoglobin (if radiation therapy),2 leukocyte counts with differential,1 LD,1 albumin,1 calcium1 In patients participating in therapeutic trials: Hemoglobin,2 leukocyte counts with differential,2 LD,2 albumin,2 calcium,2 NSE2 1 Recommendation validated by most experts and not contradicted by systematic review of the evidence. 2 Recommendation based on systematic review of the evidence. Table 3. Pretreatment prognostic value of LD vs that of ALP in NSCLC patients. . Results of multivariate analysis1 . . . . Not significant . Unspecified significance2 . Significant . LD3 16 (2450) 1 (189) 9 (4900) ALP 18 (7252) 0 1 (207)4 . Results of multivariate analysis1 . . . . Not significant . Unspecified significance2 . Significant . LD3 16 (2450) 1 (189) 9 (4900) ALP 18 (7252) 0 1 (207)4 1 Number of studies (number of patients included in the studies). 2 Significant in multivariate analysis that did not take into account performance status or disease stage. 3 The prognostic value of LD seems to be more obvious in unresected patients than in resected patients. 4 In this study, LD was not included in the multivariate analysis, whereas all studies that compared the prognostic value of LD with that of ALP concluded that LD performed better than ALP [for explanations and more details, see Refs. (40)(41)(42)]. Table 3. Pretreatment prognostic value of LD vs that of ALP in NSCLC patients. . Results of multivariate analysis1 . . . . Not significant . Unspecified significance2 . Significant . LD3 16 (2450) 1 (189) 9 (4900) ALP 18 (7252) 0 1 (207)4 . Results of multivariate analysis1 . . . . Not significant . Unspecified significance2 . Significant . LD3 16 (2450) 1 (189) 9 (4900) ALP 18 (7252) 0 1 (207)4 1 Number of studies (number of patients included in the studies). 2 Significant in multivariate analysis that did not take into account performance status or disease stage. 3 The prognostic value of LD seems to be more obvious in unresected patients than in resected patients. 4 In this study, LD was not included in the multivariate analysis, whereas all studies that compared the prognostic value of LD with that of ALP concluded that LD performed better than ALP [for explanations and more details, see Refs. (40)(41)(42)]. Table 4. AGREE scores for methodologic quality of NSCLC guidelines.1 Guidelines . Domain scores, % . . . . . . . Scope and purpose . Stakeholder involvement . Rigor of development . Clarity and presentation . Applicability . Editorial independence . ACCP 61 46 60 46 6 75 ANDEM 89 25 10 71 0 25 ASCO 94 50 71 67 17 75 ATS-ERS 44 4 5 29 0 8 BTS-SCG 100 33 60 79 6 83 CIGNA2 67 13 12 54 11 8 EGTM2 44 4 2 29 0 0 FNCLCC 94 54 57 79 17 33 NACB2 50 17 29 54 11 25 SIGN 89 75 76 75 33 25 SPLF2 61 46 48 38 17 8 Guidelines . Domain scores, % . . . . . . . Scope and purpose . Stakeholder involvement . Rigor of development . Clarity and presentation . Applicability . Editorial independence . ACCP 61 46 60 46 6 75 ANDEM 89 25 10 71 0 25 ASCO 94 50 71 67 17 75 ATS-ERS 44 4 5 29 0 8 BTS-SCG 100 33 60 79 6 83 CIGNA2 67 13 12 54 11 8 EGTM2 44 4 2 29 0 0 FNCLCC 94 54 57 79 17 33 NACB2 50 17 29 54 11 25 SIGN 89 75 76 75 33 25 SPLF2 61 46 48 38 17 8 1 Maximum score for best quality = 100%. 2 Guidelines intended for tumor markers only. Table 4. AGREE scores for methodologic quality of NSCLC guidelines.1 Guidelines . Domain scores, % . . . . . . . Scope and purpose . Stakeholder involvement . Rigor of development . Clarity and presentation . Applicability . Editorial independence . ACCP 61 46 60 46 6 75 ANDEM 89 25 10 71 0 25 ASCO 94 50 71 67 17 75 ATS-ERS 44 4 5 29 0 8 BTS-SCG 100 33 60 79 6 83 CIGNA2 67 13 12 54 11 8 EGTM2 44 4 2 29 0 0 FNCLCC 94 54 57 79 17 33 NACB2 50 17 29 54 11 25 SIGN 89 75 76 75 33 25 SPLF2 61 46 48 38 17 8 Guidelines . Domain scores, % . . . . . . . Scope and purpose . Stakeholder involvement . Rigor of development . Clarity and presentation . Applicability . Editorial independence . ACCP 61 46 60 46 6 75 ANDEM 89 25 10 71 0 25 ASCO 94 50 71 67 17 75 ATS-ERS 44 4 5 29 0 8 BTS-SCG 100 33 60 79 6 83 CIGNA2 67 13 12 54 11 8 EGTM2 44 4 2 29 0 0 FNCLCC 94 54 57 79 17 33 NACB2 50 17 29 54 11 25 SIGN 89 75 76 75 33 25 SPLF2 61 46 48 38 17 8 1 Maximum score for best quality = 100%. 2 Guidelines intended for tumor markers only. Table 5. Scores for methodologic quality and for validity of content of recommendations (scales adapted from the AGREE Instrument).1 Guidelines . Methodologic quality . Validity of content of recommendations . . . . Tumor markers . Other laboratory tests . ACCP2 Good Not so good Good ANDEM34 Not so good Good Good ASCO3 Good Not so good Good ATS-ERS34 Dubious Good Good BTS-SCG4 Good Not so good Not so good CIGNA24 Dubious Good EGTM34 Dubious Not so good FNCLCC Good Good Good NACB4 Not so good Not so good SIGN3 Good Not so good Not so good SPLF34 Not so good Not so good Guidelines . Methodologic quality . Validity of content of recommendations . . . . Tumor markers . Other laboratory tests . ACCP2 Good Not so good Good ANDEM34 Not so good Good Good ASCO3 Good Not so good Good ATS-ERS34 Dubious Good Good BTS-SCG4 Good Not so good Not so good CIGNA24 Dubious Good EGTM34 Dubious Not so good FNCLCC Good Good Good NACB4 Not so good Not so good SIGN3 Good Not so good Not so good SPLF34 Not so good Not so good 1 The best possible score (i.e., very good, the equivalent of strongly recommend of the AGREE instrument) was never attributed. 2 Guidelines published less than 3 years ago. 3 Guidelines published more than 5 years ago. 4 Diagnostic guidelines (the 4 other guidelines provide both diagnostic and therapeutic recommendations). Table 5. Scores for methodologic quality and for validity of content of recommendations (scales adapted from the AGREE Instrument).1 Guidelines . Methodologic quality . Validity of content of recommendations . . . . Tumor markers . Other laboratory tests . ACCP2 Good Not so good Good ANDEM34 Not so good Good Good ASCO3 Good Not so good Good ATS-ERS34 Dubious Good Good BTS-SCG4 Good Not so good Not so good CIGNA24 Dubious Good EGTM34 Dubious Not so good FNCLCC Good Good Good NACB4 Not so good Not so good SIGN3 Good Not so good Not so good SPLF34 Not so good Not so good Guidelines . Methodologic quality . Validity of content of recommendations . . . . Tumor markers . Other laboratory tests . ACCP2 Good Not so good Good ANDEM34 Not so good Good Good ASCO3 Good Not so good Good ATS-ERS34 Dubious Good Good BTS-SCG4 Good Not so good Not so good CIGNA24 Dubious Good EGTM34 Dubious Not so good FNCLCC Good Good Good NACB4 Not so good Not so good SIGN3 Good Not so good Not so good SPLF34 Not so good Not so good 1 The best possible score (i.e., very good, the equivalent of strongly recommend of the AGREE instrument) was never attributed. 2 Guidelines published less than 3 years ago. 3 Guidelines published more than 5 years ago. 4 Diagnostic guidelines (the 4 other guidelines provide both diagnostic and therapeutic recommendations). 1 Member of, 2 consultant to, and 3 Chair of the Committee on Evidence-Based Laboratory Medicine (C-EBLM) of the Education and Management Division (EMD) of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC-LM).http://www.ifcc.org/divisions/emd/c-eblm/aboutus.asp#2. 4 Nonstandard abbreviations: NSCLC, non-small cell lung cancer; EGTM, European Group on Tumor Markers; NACB, National Academy of Clinical Biochemistry; SPLF, Société de Pneumologie de Langue Française; CEA, carcinoembryonic antigen; NSE, neuron-specific enolase; ACCP, American College of Chest Physicians; ANDEM, Agence Nationale pour le Développement de l’Evaluation Médicale; ASCO, American Society of Clinical Oncology; ATS-ERS, American Thoracic Society and European Respiratory Society; BTS-SCG, British Thoracic Society and Society of Cardiothoracic Surgeons of Great Britain and Ireland; FNCLCC, Fédération Nationale des Centres de Lutte Contre le Cancer; SIGN, the Scottish Intercollegiate Guidelines Network; ALP, alkaline phosphatase; LD, lactate dehydrogenase; and AGREE, Appraisal of Guidelines Research and Evaluation. This work was presented at the Edutrak session entitled “The Role of Evidence-Based Laboratory Medicine in Guidelines”, developed with AACC’s EBM Group (R. Christenson, Chairman) and IFCC C-EBLM (A.R. Horvath, Chairwoman), at the XIXth International Congress on Clinical Chemistry (July 25–27, 2005, Orlando, FL). 1 Boluyt N, Lincke CR, Offringa M. Quality of evidence-based pediatric guidelines. Pediatrics 2005 ; 115 : 1378 -1391. 2 Burgers JS, Bailey JV, Klazinga NS, Van der Bij AK, Grol R, Feder G. Inside guidelines. Comparative analysis of recommendations and evidence in diabetes guidelines from 13 countries. Diabetes Care 2002 ; 25 : 1933 -1939. 3 Burgers JS, Fervers B, Haugh M, Brouwers M, Browman G, Philip T, et al. International assessment of the quality of clinical practice guidelines in oncology using the Appraisal of Guidelines and Research and Evaluation Instrument. J Clin Oncol 2004 ; 22 : 2000 -2007. 4 Cates JR, Young DN, Guerriero DJ, Jahn WT, Armine JP, Korbett AB, et al. An independent assessment of chiropractic practice guidelines. J Manipulative Physiol Ther 2003 ; 26 : 282 -286. 5 Cranney A, Waldegger L, Graham I, Man-Son-Hing M, Byszewski A, Ooi D. Systematic assessment of the quality of osteoporosis guidelines. BMC Musculoskelet Disord 2002 ; 3 : 20 (available free of charge at http://www.biomedcentral.com/content/pdf/1471-2474-3-20.pdf).. 6 Graham ID, Beardall S, Carter AO, Glennie J, Hebert PC, Tetroe JM, et al. What is the quality of drug therapy clinical practice guidelines in Canada?. CMAJ 2001 ; 165 : 157 -163(available free of charge at http://www.cmaj.ca/cgi/reprint/165/2/157.pdf).. 7 Grilli R, Magrini N, Penna A, Mura G, Liberati A. Practice guidelines developed by specialty societies: the need for a critical appraisal. Lancet 2000 ; 355 : 103 -106. 8 Grimshaw GM, Khunti K, Baker R. Diagnosis of heart failure in primary care: an assessment of international guidelines. Br J Gen Pract 2001 ; 51 : 384 -386. 9 Harpole LH, Kelley MJ, Schreiber G, Toloza EM, Kolimaga J, McCrory DC. Assessment of the scope and quality of clinical practice guidelines in lung cancer. Chest 2003 ; 123 (1 Suppl): 7S -20S. 10 Irani J, Brown CT, van der Meulen J, Emberton M. A review of guidelines on benign prostatic hyperplasia and lower urinary tract symptoms: are all guidelines the same?. BJU Int 2003 ; 92 : 937 -942. 11 Littlejohns P, Cluzeau F, Bale R, Grimshaw J, Feder G, Moran S. The quantity and quality of clinical practice guidelines for the management of depression in primary care in the UK. Br J Gen Pract 1999 ; 49 : 205 -210. 12 McAlister FA, Campbell NR, Zarnke K, Levine M, Graham ID. The management of hypertension in Canada: a review of current guidelines, their shortcomings and implications for the future. CMAJ 2001 ; 164 : 517 -522. 13 McCormack J, Perry T, Jr, Rangno R, van Breemen C, Wright JM, Baltzan M, et al. Assessing the quality of clinical practice guidelines. CMAJ 2002 ; 166 : 168 -169. 14 Shaneyfelt TM, Mayo-Smith MF, Rothwangl J. Are guidelines following guidelines? The methodological quality of clinical practice guidelines in the peer-reviewed medical literature. JAMA 1999 ; 281 : 1900 -1905. 15 Sudlow M, Thomson R. Clinical guidelines: quantity without quality. Qual Health Care 1997 ; 6 : 60 -61. 16 van Tulder MW, Tuut M, Pennick V, Bombardier C, Assendelft WJ. Quality of primary care guidelines for acute low back pain. Spine 2004 ; 29 : E357 -E362. 17 Varonen H, Makela M. Practice guidelines in Finland: availability and quality. Qual Health Care 1997 ; 6 : 75 -79. 18 Vogel N, Burnand B, Vial Y, Ruiz J, Paccaud F, Hohlfeld P. Screening for gestational diabetes: variation in guidelines. Eur J Obstet Gynecol Reprod Biol 2000 ; 91 : 29 -36. 19 Ward JE, Grieco V. Why we need guidelines for guidelines: a study of the quality of clinical practice guidelines in Australia. Med J Aust 1996 ; 165 : 574 -576. 20 Wegman A, van der Windt D, van Tulder M, Stalman W, de Vries T. Nonsteroidal antiinflammatory drugs or acetaminophen for osteoarthritis of the hip or knee? A systematic review of evidence and guidelines. J Rheumatol 2004 ; 31 : 344 -354. 21 Horvath AR, Nagy E, Watine J. Quality of guidelines for the laboratory management of diabetes mellitus. Scand J Clin Lab Invest Suppl 2005 ; 240 : 41 -50. 22 Oosterhuis WP, Bruns DE, Watine J, Sandberg S, Horvath AR. Evidence-based guidelines in laboratory medicine: principles and methods. Clin Chem 2004 ; 50 : 806 -818. 23 Bruderman I. Bronchogenic carcinoma. Baum GL Wolinsky E eds. Textbook of pulmonary diseases, 5th ed 1994 : 1345 -1392 Little, Brown and Company New York. . 24 Ginsberg RJ. Continuing controversies in staging NSCLC: an analysis of the revised 1997 staging system. Oncology 1998 ; 12 (Suppl 2): 51 -54. 25 Ihde DC, Pass HI, Glatstein EJ. Small cell lung cancer. De Vita VT Hellman S Rosenberg SA eds. Cancer. Principles and practice of oncology, 4th ed 1993 : 723 -758 JB Lippincott Philadelphia. . 26 Lebeau B. Analyse bibliographique quantitative de la recherche clinique française en thérapeutique des cancers broncho-pulmonaires. Rev Mal Respir 1998 ; 15 : 185 -189. 27 Fielding LP, Fenoglio-Preiser CM, Freedman LS. The future of prognostic factors in outcome prediction for patients with cancer. Cancer 1992 ; 70 : 2367 -2377. 28 Moher D, Schulz KF, Altman DG, . CONSORT GROUP (Consolidated Standards of Reporting Trials). The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. Ann Intern Med 2001 ; 134 : 657 -662. 29 . American College of Chest Physicians. Lung cancer guidelines. Chest 2003 ; 123 (1 Suppl): 1S -156S. 30 Agence Nationale pour le Développement de l’Evaluation Médicale. Conférence de Consensus. Bilan d’extension pré-thérapeutique du cancer bronchique non à petites cellules. http://www.uvp5.univ-paris5.fr/UV_MED/AC/Impression.asp?Nsuj=98.. 31 . American Society of Clinical Oncology. Clinical practice guidelines for the treatment of unresectable non-small-cell lung cancer. Adopted on May 16, 1997 by the American Society of Clinical Oncology. J Clin Oncol 1997 ; 15 : 2996 -3018. 32 . American Thoracic Society and European Respiratory Society. Pre-treatment evaluation of non-small-cell lung cancer. Am J Respir Crit Care Med 1997 ; 156 : 320 -332. 33 . British Thoracic Society and Society of Cardiothoracic Surgeons of Great Britain and Ireland. BTS guidelines: guidelines on the selection of patients with lung cancer for surgery. Thorax 2001 ; 56 : 89 -108http://thorax.bmjjournals.com/cgi/content/full/56/2/89 (accessed November 8, 2005).. 34 CIGNA HealthCare Medicare Administration. Tumor markers for diagnosis and management of cancer. Coverage position number 0172. http://www.cigna.com/health/provider/medical/procedural/coverage_positions/medical/mm_0172_coveragepositioncriteria_tumor_markers_for_diagnosis_mgmt_cancer.pdf (accessed November 8, 2005).. 35 . European Group on Tumour Markers. Tumour markers in lung cancer: EGTM recommendations. Anticancer Res 1999 ; 19 : 2817 -2819. 36 Fédération Nationale des Centres de Lutte Contre le Cancer. Standards, Options et Recommandations 2000 pour la prise en charge des patients atteints d’un cancer bronchopulmonaire non à petites cellules. http://www.fnclcc.fr/ (accessed November 8, 2005).. 37 National Academy of Clinical Biochemistry. Practice guidelines and recommendations for use of tumor markers in the clinic. Laboratory medicine practice guidelines. http://www.nacb.org/LMPG/Monograph_TumorMarkers.pdf (accessed November 8, 2005).. 38 Scottish Intercollegiate Guidelines Network. Management of lung cancer, pilot edition, February 1998. http://www.sign.ac.uk/ (this pilot edition is available from the corresponding author upon request).. 39 . Société de Pneumologie de Langue Française. Recommandations concernant l’utilisation des marqueurs tumoraux sériques dans la prise en charge des cancers bronchiques primitifs. Rev Mal Resp 1997 ; 14 (Suppl 3): 5 -39. 40 Watine J. Prognostic evaluation of primary non-small cell lung carcinoma patients using biological fluid variables. A systematic review. Scand J Clin Lab Invest 2000 ; 60 : 259 -274. 41 Watine J. Are laboratory investigations recommended in current medical practice guidelines supported by available evidence?. Clin Chem Lab Med 2002 ; 40 : 252 -255. 42 Watine J, Friedberg B, Charet JC. Laboratory variables and stratification of lung cancer patients: recommendations for the future (Therapeutic Trials and Clinical Practice Guidelines). Carafaro RL eds. Focus on lung cancer research 2004 : 173 -203 Nova Science Publishers Hauppauge, NY. . 43 Watine J, Bouarioua N. Anemia as an independent prognostic factor for survival in patients with cancer. Cancer 2002 ; 94 : 2793 -2796author reply 2796–7.. 44 Pujol JL, Molinier O, Ebert W, Daures JP, Barlesi F, Buccheri G, et al. CYFRA 21-1 is a prognostic determinant in non-small-cell lung cancer: results of a meta-analysis in 2063 patients. Br J Cancer 2004 ; 90 : 2097 -2105. 45 Sandberg S, Oosterhuis W, Freedman D, Kawai T. Systematic reviewing in laboratory medicine. Position paper from the IFCC committee on systematic reviewing in laboratory medicine. J Int Fed Clin Chem 1997 ; 9 : 154 -155. 46 Durand-Zalesky I, Rymer JC, Roudot-Thoraval F, Revuz J, Rosa J. Reducing unnecessary laboratory use with new test request form: example of tumour markers. Lancet 1993 ; 342 : 150 -153. 47 Appraisal of Guidelines for Research and Evaluation Collaboration. Development and validation of an international appraisal instrument for assessing the quality of clinical practice guidelines: the AGREE project. Qual Saf Health Care 2003 ; 12 : 18 -23(the AGREE Instrument is available at www.agreecollaboration.com).. 48 Vlayen J, Aertgeerts B, Hannes K, Sermeus W, Ramaekers D. A systematic review of appraisal tools for clinical practice guidelines: multiple similarities and one common deficit. Int J Qual Health Care 2005 ; 17 : 235 -242. 49 Hasenfeld R, Shekelle PG. Is the methodological quality of guidelines declining in the US? Comparison of the quality of US Agency for Health Care Policy and Research (AHCPR) guidelines with those published subsequently. Qual Saf Health Care 2003 ; 12 : 428 -434. 50 Burgers JS, van Everdingen JJ. Beyond the evidence in clinical guidelines. Lancet 2004 ; 364 : 392 -393. 51 Shekelle PG, Kravitz RL, Beart J, Marger M, Wang M, Lee M. Are nonspecific practice guidelines potentially harmful? A randomized comparison of the effect of nonspecific versus specific guidelines on physician decision making. Health Serv Res 2000 ; 34 : 1429 -1448. 52 Watine J, Charet JC. Faut-il suivre les sociétés savantes de pneumologie dans leurs recommandations concernant l’utilisation des marqueurs tumoraux sériques dans l’évaluation pronostique des cancers bronchiques primitifs non à petites cellules?. Rev Mal Respir 1999 ; 16 : 139 -149. 53 Shekelle PG, Ortiz E, Rhodes S, Morton SC, Eccles MP, Grimshaw JM, et al. Validity of the Agency for Healthcare Research and Quality clinical practice guidelines: how quickly do guidelines become outdated?. JAMA 2001 ; 286 : 1461 -1467. 54 Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S, et al. GRADE Working Group. Grading quality of evidence and strength of recommendations. BMJ 2004 ; 328 : 1490 . 55 Shiffman RN, Shekelle P, Overhage JM, Slutsky J, Grimshaw J, Deshpande AM. Standardized reporting of clinical practice guidelines: a proposal from the Conference on Guideline Standardization. Ann Intern Med 2003 ; 139 : 493 -498. © 2006 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)