Advances in Clinical Chemistry 1st Edition Gregory S. Makowski (Eds.)

Visit https://ebookultra.com to download the full version and
explore more ebooks
Advances in Clinical Chemistry 1st Edition
Gregory S. Makowski (Eds.)
_____ Click the link below to download _____
https://ebookultra.com/download/advances-in-clinical-
chemistry-1st-edition-gregory-s-makowski-eds/
Explore and download more ebooks at ebookultra.com

Here are some suggested products you might be interested in.
Click the link to download
Advances in Insect Physiology Vol 29 1st Edition Peter
Evans
https://ebookultra.com/download/advances-in-insect-physiology-
vol-29-1st-edition-peter-evans/
Advances in Gas Phase Ion Chemistry Volume 4 Advances in
Gas Phase Ion Chemistry 1st Edition L.M. Babcock
https://ebookultra.com/download/advances-in-gas-phase-ion-chemistry-
volume-4-advances-in-gas-phase-ion-chemistry-1st-edition-l-m-babcock/
Advances in Carbene Chemistry 1st Edition U.H. Brinker
https://ebookultra.com/download/advances-in-carbene-chemistry-1st-
edition-u-h-brinker/
Advances in Mergers and Acquisitions Volume 2 Advances in
Mergers and Acquisitions 1st Edition A. Gregory
https://ebookultra.com/download/advances-in-mergers-and-acquisitions-
volume-2-advances-in-mergers-and-acquisitions-1st-edition-a-gregory/

Synthetic Peptides A User s Guide Advances in Molecular
Biology 2nd Edition Gregory Grant
https://ebookultra.com/download/synthetic-peptides-a-user-s-guide-
advances-in-molecular-biology-2nd-edition-gregory-grant/
Advances in Heterocyclic Chemistry 87 1st Edition Alan R.
Katritzky
https://ebookultra.com/download/advances-in-heterocyclic-
chemistry-87-1st-edition-alan-r-katritzky/
Advances in Physical Organic Chemistry 37 1st Edition John
Richard
https://ebookultra.com/download/advances-in-physical-organic-
chemistry-37-1st-edition-john-richard/
Advances in Physical Organic Chemistry 38 1st Edition John
Richard
https://ebookultra.com/download/advances-in-physical-organic-
chemistry-38-1st-edition-john-richard/
Advances in Heterocyclic Chemistry Volume 114 First
Edition Ramsden
https://ebookultra.com/download/advances-in-heterocyclic-chemistry-
volume-114-first-edition-ramsden/

Advances in Clinical Chemistry 1st Edition Gregory S.
Makowski (Eds.) Digital Instant Download
Author(s): Gregory S. Makowski (Eds.)
ISBN(s): 9780123747976
Edition: 1
File Details: PDF, 1.21 MB
Year: 2009
Language: english

ADVANCES IN CLINICAL CHEMISTRY
VOLUME 48

This page intentionally left blank

Advancesin
CLINICAL
CHEMISTRY
Edited by
GREGORY S. MAKOWSKI
Department of Laboratory Medicine
University of Connecticut Health Center
Farmington, CT, USA
VOLUME 48
AMSTERDAM • BOSTON HEIDELBERG LONDON
NEW YORK OXFORD PARIS SAN DIEGO
SAN FRANCISCO SINGAPORE SYDNEY TOKYO
Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier
32 Jamestown Road, London NW1 7BY, UK
30 Corporate Drive, Suite 400, Burlington, MA 01803, USA
525 B Street, Suite 1900, San Diego, CA 92101-4495, USA
Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands
This book is printed on acid-free paper.

Copyright2009, Elsevier Inc. All rights reserved
No part of this publication may be reproduced, stored in a retrieval system
or transmitted in any form or by any means electronic, mechanical, photocopying,
recording or otherwise without the prior written permission of the publisher
Permissions may be sought directly from Elsevier’s Science & Technology Rights
Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333;
email: [email protected]. Alternatively you can submit your request online by
visiting the Elsevier web site at http://www.elsevier.com/locate/permissions, and selecting
Obtaining permission to use Elsevier material
Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons
or property as a matter of products liability, negligence or otherwise, or from any use
or operation of any methods, products, instructions or ideas contained in the material
herein. Because of rapid advances in the medical sciences, in particular, independent
verification of diagnoses and drug dosages should be made
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-12-374797-6
ISSN: 0065-2423
For information on all Academic Press publications
visit our website at www.elsevierdirect.com
Printed and bound in USA
09101112 10987654321

CONTENTS
CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix
P
REFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
Clinical Validation of Biomarkers for Predicting Risk
S
TANLEYS. LEVINSON
1. Abstract ...................................................................................... 1
2. Introduction................................................................................. 2
3. RR/OR Ratios as Diagnostic Tools...................................................... 3
4. ROC Plots/Curves as a Diagnostic Tool................................................. 4
5. Comparison of RR/OR with ROC curves............................................... 7
6. Distributions................................................................................. 9
7. Bayesian Principles.......................................................................... 9
8. Weaknesses of ROC Analysis............................................................. 12
9. Weaknesses of RR/OR..................................................................... 13
10. Stand-Alone versus Synergic Biomarkers................................................ 15
11. Techniques for Improving Stratiﬁcation of Synergic Biomarkers..................... 15
12. Criteria for Identifying Testing of Clinical Consequence............................... 16
13. Discussion.................................................................................... 19
14. Conclusions.................................................................................. 20
Glossary of Expressions and Explanations.............................................. 21
References.................................................................................... 22
The Potential Role of Heat Shock Proteins in Cardiovascular Disease:
Evidence fromIn VitroandIn VivoStudies
M. G
HAYOUR-MOBARHAN, A.A. RAHSEPAR,S.TAVALLAIE,
S. R
AHSEPAR,ANDG.A.A. FERNS
1. Abstract ...................................................................................... 28
2. Introduction................................................................................. 28
3. HSPs and Atherogenesis................................................................... 34
4. HSPs and Autoimmunity in Atherogenesis.............................................. 45
5. Therapeutic Implications................................................................... 58
6. Conclusions.................................................................................. 59
References.................................................................................... 59
v

The Emerging Role of Symmetric Dimethylarginine in Vascular Disease
A
RDUINOA. MANGONI
1. Abstract....................................................................................... 73
2. Introduction.................................................................................. 74
3. Synthesis, Transport, and Metabolism of ADMA and SDMA........................ 75
4. ADMA and the Cardiovascular System.................................................. 78
5. SDMA and the Cardiovascular System................................................... 79
6. Discussion.................................................................................... 88
References.................................................................................... 89
Melanocortin-4 Receptor Mutations in Obesity
F
ERRUCCIOSANTINI,MARGHERITAMAFFEI,CATERINAPELOSINI,
G
UIDOSALVETTI,GIOVANNASCARTABELLI,ANDALDOPINCHERA
1. Abstract....................................................................................... 95
2. Introduction.................................................................................. 96
3. The Melanocortin System.................................................................. 97
4. The MC4R ................................................................................... 99
5. Mutations in the MC4R.................................................................... 100
6. Functional Alterations of MC4R.......................................................... 102
7. Clinical Phenotype of MC4R-Mutated Individuals..................................... 102
8. Implications of MC4R Mutations in the Clinical Management of Obesity........... 103
9. Conclusions.................................................................................. 103
References.................................................................................... 104
Proinflammatory Cytokines in CRP Baseline Regulation
C
ARITAM. EKLUND
1. Abstract....................................................................................... 111
2. C-Reactive Protein and Inﬂammation.................................................... 112
3. Demographic, Metabolic, and Socioeconomic Factors ................................. 114
4. Proinﬂammatory Cytokines................................................................ 118
5. Signaling Through IL Receptors.......................................................... 123
6. Genetic Polymorphisms..................................................................... 124
7. Conclusions.................................................................................. 124
References.................................................................................... 126
Fetal Skin Wound Healing
E
DWARDP. BUCHANAN,MICHAELT. LONGAKER,ANDH. PETERLORENZ
1. Abstract....................................................................................... 138
2. Introduction.................................................................................. 138
3. Development................................................................................. 140
vi CONTENTS

4. Scarless Fetal Wound Repair Speciﬁcity................................................. 141
5. Stem Cells .................................................................................... 147
6. Cellular Inﬂammatory Mediators ......................................................... 149
7. Cytokines .................................................................................... 151
References.................................................................................... 155
Clinical Relevance of BNP Measurement in the Follow-Up
of Patients with Chronic Heart Failure
A
LDOCLERICO,MARIANNAFONTANA,ANDREARIPOLI,ANDMICHELEEMDIN
1. Abstract ...................................................................................... 163
2. Background and Aim of the Study........................................................ 164
3. Biochemical and Physiological Properties of B-Type Natriuretic Peptides........... 165
4. Circulating Levels of B-Type Natriuretic Peptides...................................... 167
5. Variations of Plasma B-Type Natriuretic Peptides, Dependent on
Pharmacological Treatment, as Surrogate End-Point for Treatment
of Patients with HF......................................................................... 168
6. Prognostic Relevance of Plasma BNP/NT-proBNP Variations After Treatment ... 169
7. Meta-Analysis for Overall Mortality Including All Randomized Clinical Trials.... 174
8. BNP-Guided Therapy in Chronic Heart Failure: Instructions for Use............... 175
References.................................................................................... 176
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
CONTENTS
vii

This page intentionally left blank

CONTRIBUTORS
Numbers in parentheses indicate the pages on which the authors’ contributions begin.
EDWARDP. BUCHANAN (137),Division Plastic Surgery, Department of
Surgery, Stanford University School of Medicine, Stanford, California
94305, USA
A
LDOCLERICO(163),Scuola Superiore Sant’Anna, 56126 Pisa, Italy; and
Gabriele Monasterio Foundation CNR-Regione Toscana, 56126 Pisa, Italy
C
ARITAM. EKLUND(111),Department of Microbiology and Immunology,
University of Tampere, Medical School, 33520 Tampere, Finland
M
ICHELEEMDIN(163),Gabriele Monasterio Foundation CNR-Regione
Toscana, 56126 Pisa, Italy
G.A.A. F
ERNS(27),Postgraduate Medical School, University of Surrey,
Guildford, Surrey GU2 7WG, UK
M
ARIANNAFONTANA(163),Gabriele Monasterio Foundation CNR-Regione
Toscana, 56126 Pisa, Italy
M. G
HAYOUR-MOBARHAN(27),Cardiovascular Research Center, Avicenna
Research Institute, Mashhad University of Medical Science (MUMS),
Mashhad 91376-73119, Iran; and Department of Nutrition and Biochemistry,
Faculty of Medicine, MUMS, Mashhad 91376-73119, Iran
S
TANLEYS. LEVINSON(1),Laboratory Service, Department of Veterans AVairs
Medical Center, Louisville, Kentucky 40206, USA; and Department
of Pathology and Laboratory Medicine, School of Medicine, University
of Louisville, Louisville, Kentucky 40292, USA
M
ICHAELT. LONGAKER(137),Division Plastic Surgery, Department of
Surgery, Pediatric Surgical Research Laboratory, Stanford University School
of Medicine, Stanford, California 94305-5148, USA
ix

H. PETERLORENZ(137),Division Plastic Surgery, Department of Surgery,
Pediatric Surgical Research Laboratory, Stanford University School of
Medicine, Stanford, California 94305-5148, USA
M
ARGHERITAMAFFEI(95),Dulbecco Telethon Institute at the Department of
Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy
A
RDUINOA. MANGONI(73),Department of Clinical Pharmacology, School of
Medicine, Flinders University, Adelaide 5001, Australia
C
ATERINAPELOSINI(95),Department of Endocrinology and Kidney, University
Hospital of Pisa, 56124 Pisa, Italy
A
LDOPINCHERA(95),Department of Endocrinology and Kidney, University
Hospital of Pisa, 56124 Pisa, Italy
A.A. R
AHSEPAR, (27),Cardiovascular Research Center, Avicenna Research
Institute, Mashhad University of Medical Science (MUMS), Mashhad
91376-73119, Iran; and Department of Nutrition and Biochemistry, Faculty
of Medicine, MUMS, Mashhad 91376-73119, Iran
S. R
AHSEPAR(27),Cardiovascular Research Center, Avicenna Research
Institute, Mashhad University of Medical Science (MUMS), Mashhad
91376-73119, Iran; and Department of Nutrition and Biochemistry, Faculty
of Medicine, MUMS, Mashhad 91376-73119, Iran
A
NDREARIPOLI(163),Gabriele Monasterio Foundation CNR-Regione
Toscana, 56126 Pisa, Italy
G
UIDOSALVETTI(95),Department of Endocrinology and Kidney, University
Hospital of Pisa, 56124 Pisa, Italy
F
ERRUCCIOSANTINI(95),Department of Endocrinology and Kidney, University
Hospital of Pisa, 56124 Pisa, Italy
G
IOVANNASCARTABELLI(95),Department of Endocrinology and Kidney,
University Hospital of Pisa, 56124 Pisa, Italy
S. T
AVALLAIE(27),Department of Nutrition and Biochemistry, Faculty of
Medicine, MUMS, Mashhad 91376-73119, Iran
x
CONTRIBUTORS

PREFACE
I am pleased to present volume forty-eight ofAdvances in Clinical
Chemistryseries.
In this second volume for 2009, the lead chapter explores the fundamental
importance of receiver operator curves as the gold standard for statistical
analysis of test diagnostic performance. The next chapter probes the signiﬁ-
cance of heat shock proteins, a group of highly conserved proteins expressed
under stress, as risk factors for development of cardiovascular disease. The
role of symmetric and asymmetric dimethylarginine in nitric oxide synthesis
is explored in the following chapter which discusses potential impact on
vascular homeostasis and vascular disease. An interesting chapter on obesity
seeks to explore the involvement of the melanocortin receptor system as an
important mediator of leptin eVect on body weight and metabolism. The next
chapter investigates the impact of low-grade inﬂammation and proinﬂam-
matory cytokines on C reactive protein. An interesting chapter investigates
the unique restoration of extracellular matrix architecture, strength, and
function in fetal wound healing. The role of the inﬂammatory response,
cellular mediators, cytokines, and growth factors are elucidated in this
fascinating process. We conclude volume forty-eight with a manuscript on
the usefulness of BNP to biochemically monitor patients with chronic heart
failure.
I extend my appreciation to each contributor of volume forty-eight and
thank colleagues who participated in the peer-review process. I extend thanks
to my Elsevier editorial liaison, Gayathri Venkatasamy.
I sincerely hope the second volume of 2009 will be enjoyed by our readers.
As always, I invite comments and suggestions for future review articles for
theAdvances in Clinical Chemistryseries.
In keeping with the tradition of the series, I would like to dedicate volume
forty-eight to my brother Keith.
G
REGORYS. MAKOWSKI
xi

This page intentionally left blank

CLINICAL VALIDATION OF BIOMARKERS
FOR PREDICTING RISK
Stanley S. Levinson*
,†,1
*Laboratory Service, Department of Veterans Affairs
Medical Center, Louisville, Kentucky 40206, USA
†
Department of Pathology and Laboratory Medicine,
School of Medicine, University of Louisville,
Louisville, Kentucky 40292, USA
1. Abstract . ....................................................................... 1
2. Introduction. . . . ................................................................ 2
3. RR/OR Ratios as Diagnostic Tools . . .......................................... 3
4. ROC Plots/Curves as a Diagnostic Tool . . ...................................... 4
5. Comparison of RR/OR with ROC curves ...................................... 7
6. Distributions . . . ................................................................ 9
7. Bayesian Principles . ............................................................ 9
8. Weaknesses of ROC Analysis . ................................................. 12
8.1. Diagnostic Models........................................................ 12
8.2. Prognostic Models........................................................ 13
9. Weaknesses of RR/OR . . . . ..................................................... 13
10. Stand‐Alone versus Synergic Biomarkers . ...................................... 15
11. Techniques for Improving Stratiﬁcation of Synergic Biomarkers . . . ............. 15
12. Criteria for Identifying Testing of Clinical Consequence ........................ 16
13. Discussion...................................................................... 19
14. Conclusions . . . . ................................................................ 20
Glossary of Expressions and Explanations...................................... 21
References...................................................................... 22
1. Abstract
Background: A useful biomarker should improve clinical management in
an economically reasonable way. This should be determined from well‐
designed outcome studies that show clinical management can be altered on
1
Corresponding author: Stanley S. Levinson, e-mail: [email protected]
1
0065-2423/09 $35.00 Copyright 2009, Elsevier Inc.
DOI: 10.1016/S0065-2423(09)48001-6 All rights reserved.
ADVANCES IN CLINICAL CHEMISTRY, VOL.
48

the basis of the biomarker. It is important not to confuse results from testing
prior to outcome with outcome studies.
Content: This chapter reviews statistical tests used to evaluate studies
performed prior to ﬁnal outcome studies and criteria that assess whether or
not a biomarker should be considered for outcome studies at each step.
I review how relative risk and odds ratios are related to receiver operator
characteristic (ROC) plot analysis. Other statistical techniques such as re-
classiﬁcation and the Hosmer Lemeshow test that have been suggested for
evaluation of diagnostic usefulness are considered. Weaknesses of each
technique are discussed.
Summary: I consider ROC analysis to be a mainstay against which other
statistical tests of diagnostic performance should be compared. The impor-
tance of expressing data in terms of predictive values is emphasized. Tests
showing weak diagnostic associations with a disease are diYcult to evaluate
for outcome study application, because there is usually great diVerence in
between‐study variance so that the true relationship between the biomarker,
its diagnostic ability, and predictive capability are unclear.
2. Introduction
The ﬁnal criterion for deﬁning biomarkers for clinical prognosis should be:
can clinical management be altered on the basis of test results leading to
improved care in an economically reasonable way while reaching a dimen-
sion of quality that is critical for preventing wrong results and wrongful
treatment of patients?[1]. This information should be obtained from well‐
designed prospective outcome studies[2, 3]. A recent editorial questioned
the overzealous emphasis onpvalues of 0.05[4].But,forclinicalstudies,
even well‐deﬁned statistical signiﬁcance is not suYcient to impart useful-
ness, since there is a breach between statistical signiﬁcance and meaningful
diagnostic discrimination.
In this chapter, I postulate that the statistical tests used to evaluate studies
performed prior to ﬁnal outcome studies (preliminary studies) help in point-
ing toward which biomarkers might show suYcient diagnostic discrimination
to be evaluated in large expensive prospective studies, not to determine
whether these tests should be put into clinical service. In recent years, a
plethora of new biomarkers have been proposed, most for predicting risk
or progression of disease [5–8], some of which, it seems to me, have been
suggested for clinical use on the basis of preliminary studies alone [6, 9–11].
This chapter reviews strengths and weaknesses of statistical techniques
that help us decide which biomarkers might be appropriate for application
to outcome studies and reviews diagnostic criteria that can be applied at each
2
STANLEY S. LEVINSON

step of the evaluation. Central to these preliminary decisions is an estimation
of positive predictive value (PPV).
Traditionally, receiver operator characteristic (ROC) analysis has been
used to describe and compare the clinical accuracy of biomarkers. ROC
analysis expresses data as diagnostic sensitivity and speciﬁcity that can
easily be translated into useful predictive values (PPV and negative predic-
tive value, NPV). Often, the performance of new markers are gauged by
relative risk (RR) or odds (OR) ratios. I will review how RR/OR are related
to ROC analysis. I will also review other approaches for assessing discrimi-
nation that have been proposed to be more sensitive than ROC analysis
that may have merit [12]. In spite of several weaknesses that I will discuss,
I am of the opinion that illustration of data in terms of ROC analysis is a
good way to appreciate the strength of the diagnostic relationship (or
association) between the biomarker and the disease. I will discuss how
weak diagnostic associations between disease and a biomarker, that fall
into the area of low accuracy by ROC analysis, usually show great inter-
study variability so that the true relationships are unclear and predictive
capability is poor. Such inconsistent behavior produces a great challenge in
terms of cost and likely results for prospective outcome studies. I hope this
chapter will provide a better insight for the clinical laboratory practitioner
as to how the statistics used for preliminary evaluation of biomarkers
translate into meaningful clinical discrimination. Although, in most exam-
ples, the focus is on inﬂammatory biomarkers, these principles apply to
biomarkers in general.
3. RR/OR Ratios as Diagnostic Tools
In a cohort study, the association between a factor and the occurrence of
an event is often depicted as the RR (see Glossary) that requires a known
incidence. OR (see Glossary) is the odds that a case is exposed divided by the
odds that a control is exposed. OR are usually applied to studies where
incidence rates cannot be established, especially cross‐sectional studies. Usu-
ally, OR are derived from multivariate logistic type computations, where
adjustment of risk for covariant markers (possible confounders) can be made
[13]. When determining the usefulness of new biomarkers, it is important to
adjust risk for established markers (covariant or possible confounders) to
decide if the new marker is independent of the established markers and,
therefore, whether or not it adds additional diagnostic discrimination (not
necessarily useful diagnostic discrimination). In cohort studies, the adjusted
RR can be calculated from the OR [14]. An adjusted RR, known as the
hazard ratio, can be obtained for survival analysis censored outcomes [15].
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK 3

4. ROC Plots/Curves as a Diagnostic Tool
ROC curves compare diagnostic accuracy of methods over all possible
sensitivity/speciﬁcity pairs by plotting the true positive rate (TPR) which is
the diagnostic sensitivity, against that of the false positive rate (FPR) which is
1‐(diagnostic speciﬁcity). Figure 1 illustrates examples of an idealized series
of ROC curves [16]. Each point on the curve/plot corresponds to a diagnostic
sensitivity/speciﬁcity associated with a speciﬁc concentration of biomarker so
that a best reference cutoVcan easily be determined. ROC analysis can be
statistically adjusted for possible confounders just as RR/OR can [17–19].
Each point on the ROC curve also represents a likelihood ratio (see Glossary)
[19, 20].
The ROC curve is obtained by plotting the cumulative frequency of the
true positive results (TP) which is the proportion of diseased subjects against
the false positive results (FP) which is the proportion of nondiseased subjects
each correctly diagnosed at various cutoVpoints. In essence, the number of
subjects that are TP and FP are cumulatively added at each concentration of
the biomarker and the cumulative sum at each concentration is divided by
the total number of TP and FP to give the cumulative frequency. As such the
concentration at each sensitivity/speciﬁcity pair is known. An example for
constructing a ROC curve is shown in Table 1 and Fig. 2.
As shown in Fig. 2, although the scale reﬂecting concentration is not
linear, the concentrations for important cutoVs (sometimes called decision
levels) can be denoted on a third axis. Generally, the emphasis is on the
sensitivity/speciﬁcity pair that best ﬁts the clinical paradigm—diagnostically
how many FP can we tolerate to reach an optimal number of TP. For any
biomarker in which the disease and nondisease distributions overlap, the
sensitivity and speciﬁcity move in opposite directions over the span of the
ROC curve, so there are always trade‐oVs between the two requiring a
decision‐level selection that must depend on the clinical circumstances. Nev-
ertheless, since the biomarker concentrations have been used to generate the
ROC graph once a sensitivity/speciﬁcity pair that best ﬁts the clinical situa-
tion is identiﬁed the corresponding biomarker concentrations or decision
levels are known.
Unfortunately, graphs in many articles do not show the biomarker con-
centrations on a third axis. Some have been concerned that the decision level
may not be obvious from an inspection of the ROC plot. An interesting
alternative is to plot the sensitivity and speciﬁcity on separate curves but on
the same graph against concentration—called cumulative distribution analy-
sis [21]. Complicated neural networks, discriminant analysis, and logistic
regression techniques have also been used to compute appropriate decision
levels [22]. Nevertheless, although theoretically interesting, it is the clinical
4
STANLEY S. LEVINSON

paradigm applied to the appropriate sensitivity/speciﬁcity pair that must
ultimately determine the adequacy of an optimal cutoV(s) and no mathemat-
ical approach can be used alone to deﬁne a decision level.
The area under the ROC curve, which for binary outcomes is called the
c‐statistic, is equivalent to the nonparametric Mann–Whitney statistic and
is not aV ected by skewness of the underlying data distribution [23]. The
c‐statistic can be used to globally compare tests. The test with the higher
c‐statistic is considered the better test. A test with high accuracy generally
shows ac‐statistic of0.9, intermediate accuracy; 0.7–0.9, useful for some
Concentration
0[X+
–
10] [X +
–
20] [X +
–
30] [X +
–
40]
171
36
16
9
3
2
1.5
1.0
0.5
0.75
0.85
0.95
1.0
0.8
0.6
0.4
0.2
0
False positive rate (1–specificity)
True positive rate (sensitivity)
0 0.2 0.4 0.6 0.8 1.0
OR
AUROC
or
c-statistic
FIG. 1. Series of idealized ROC curves showing the relationship between the area under the
ROC curve (AUROC), also called thec‐statistic and ORs. If the AUROC curve is 1.0, the
method is perfectly accurate. In this case, the plot follows they‐axis into the left corner of
the plot. If the area under the ROC curve is 0.5, the method shows no discrimination. The
diagonal line with an AUROC of 0.5 illustrates no discrimination. Each point on the curve
corresponds to a biomarker concentration (indicated on the top horizontal axis). If the
biomarker concentration is increasing with the disease, the concentration on the top axis is
decreasing [X‐ ] from left to right. If the biomarker concentration is decreasing with disease,
the concentration on the top axis is increasing [Xþ] from left to right. Above, as indicated
by the arrow are odds ratios (OR) corresponding to AUROC indicated by the arrow just
below. The true positive rate is the same as the diagnostic sensitivity, while the false positive
rate is equal to 1‐(diagnostic speciﬁcity). The relationship between AUROC and OR was
modiﬁed from Pepeet al. [16] with permission.
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK
5

purposes; and low accuracy of 0.5–0.7 [18]. Nevertheless, ROC curves are
more than just thec‐statistic; they are a graphic depiction of all sensitivity/
speciﬁcity pairs [18]. Moreover, the curve does not always show an idealized
shape [24], so that a cutoVpoint may show much more discrimination (or
less) than implied by thec‐statistic.
For example, it was shown that the replacement of low density lipoprotein
(LDL) cholesterol (C) by apo B or nonhigh‐density lipoprotein cholesterol
(non‐HDLC) caused an adjusted ROC curve to change its area from 0.7 to
0.73–0.74 (Fig. 3) with an increased TPR (diagnostic sensitivity) from about
0.4 to about 0.47 at a diagnostic speciﬁcity of 83% [25] which is the cutoV
point that equals the National Cholesterol Education Program (NCEP)
guideline [26] for an elevated LDLC (about 1300 mg/L) and non‐ HDLC
(about 1600 mg/L). If such an improvement be accurate, replacement of
TABLE 1
E
XAMPLE OFROC CURVECALCULATION FOR A BIOMARKER THAT IS INCREASING IN
CONCENTRATION WITH DISEASE
Biomarker
concentration
a
(mg/dL)
Number
of
subjects
b
Number
of
normals
c
Number
of
diseased
CUM
SUM
of normals
CUM
SUM
of diseased FPR
d
TPR
d
350 101 1 100 1 100 0.012 0.299
300 80 5 75 6 175 0.071 0.522
250 56 6 50 12 225 0.143 0.672
200 45 5 40 17 265 0.202 0.791
150 40 10 30 27 295 0.321 0.881
125 37 17 20 44 315 0.534 0.940
100 20 10 10 55 325 0.655 0.970
75 15 10 5 65 330 0.774 0.985
50 9 5 4 70 334 0.833 0.997
25 15 14 1 84 335 1.000 1.000
a
The data is ranked into concentrations of the biomarker with decreasing or increasing
concentration (the highest or lowest concentration ﬁrst), depending on whether or not the disease
increases or decreases with concentration, respectively. For example, LDLC increases with
disease while HDLC decreases with disease.
b
The number of subject with disease and without disease at each concentration is counted.
This begins a transformation from actual data to a binary‐type classiﬁcation based on frequency.
Transformation introduces uncertainty and is a weakness of the method.
c
The number of normal subjects at each concentration is added cumulatively to the next to
obtain a series of cumulative (CUM) sums. The diseased subjects are likewise cumulatively
summed.
d
The TPR and FPR is obtained by dividing the CUM SUM at each concentration by the
maximum cumulative sum for diseased and normals, respectively, to derive the TPR and FPR
for each concentration, each of which is plotted on a separate axis (as shown in Fig. 2).
6 STANLEY S. LEVINSON

LDLC by apo B or non‐ HDLC in the routine lipid screen would identify 7%
more high‐risk persons. Notice, the improvement in diagnostic discrimina-
tion of 7% at the critical cutoVvalue is apparent from the graphic depiction
of the curve while the global estimate from thec‐statistic is only 3–4%.
5. Comparison of RR/OR with ROC curves
Like ROC curves, OR can be related to TPR and FPR, in that the OR¼
[TPR/(1‐TPR)][(1‐FPR)/FPR] [16]. Using this information, an approxi-
mation as to the relationship between OR andc‐statistic has been estimated
as shown in Fig. 1 [16]. Based on the ﬁgure, an OR of about 3.0 would be
required to reach borderline intermediate accuracy (c‐statistic0.7) and
about 36 (c‐statistic0.95) to reach high accuracy levels. It was suggested
that an OR less than 3.0, corresponding to ac‐statistic of about 0.65, would
not be adequate for individual classiﬁcation [16]. Most of the tests whose
Biomarker concentration (mg/dL)
1.0
0.8
0.6
0.4
0.2
0
0
350 300 250 200 150 125 100 75 50 25
0.2 0.4
False positive rate (1–specificity)
True positive rate (sensitivity)
0.6 0.8 1.0
FIG. 2. Illustrative plot of a ROC curve. Table 1 shows hypothetical concentrations for a
biomarker and describes how to arrange the ranked concentrations, how to generate the
cumulative sums and calculate the TPR and FPR from the frequency coincident with (generated
by) concentrations of the biomarker.
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK
7

discriminations are described in terms of RR/OR fall in the range between
1.0 and 2.0 [6]. These are equivalent toc‐statistics of<0.6 or poor diagnostic
discrimination (Fig. 1). Discrimination that would produce ORs of 1:16 or
greater are rarely seen in the literature since this would usually be expressed
by ROC analysis withc‐statistics greater than 0.80.
LDLC (mg/dL) concentration
After correction for:
age, BMI, smoking, BP
0.40
Auroc curve
0.47
LDLC
Apo B or
NonHDLC
LDLC = 0.70
= 0.74
= 0.73
Apo B
NonHDLC
0 130 100 50
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
False positive rate (1–specificity)
True positive rate (sensitivity)
0 0.1 0.2 0.3 0.4 0.5 0.6
FIG. 3. ROC analysis of apo B and lipoprotein lipids after adjustment for the traditional risk
factors of age, smoking (S), hypertension (BP), and body mass index (BMI). The vertical line at
about a FPR of 17% (speciﬁcity 87%) represents the sensitivity–speciﬁcity points that correspond
to the NCEP’s LDLC and non‐HDLC recommended cutoVpoint of 1300 and 1600 mg/L,
respectively. At this point, LDLC curve shows a TPR of about 0.4 (40%) while the apo B and
non‐HDLC curves, that are nearly superimposed, show a TPR of about 0.47 (47%). This is a 7%
diVerence while the AUROC, shown in the block, shows only a 3–4% diVerence (0.7 vs. 0.73 or
0.74). AUROC curve is a global measure that is equivalent to thec‐statistic. Modiﬁed from
reference [25], with permission. The adjusted ROC curves were generated from logistic regression
equations by the method described in Ref. [18].
8 STANLEY S. LEVINSON

6. Distributions
It is important to remember that the distribution (see Glossary) over which
the biomarker data are spread is also important in evaluation of discrimina-
tion. Too few values usually cause statistical signiﬁcance to be insigniﬁcant,
but when many values are presented statistical signiﬁcance may be reached,
but with poor diagnostic discrimination. Even an apparently strong bio-
marker can be a poor screening test. This is because both RR/OR and
ROC analysis often evaluate data that is lumped together on each end of
the distribution for the biomarker, so that the eVect of being highly exposed
to a biomarker is compared with being slightly exposed while most people in
the middle of the distribution are being ignored [27]. This occurs either
because there are a few values in the middle or because a few extreme values
at the end are causing the biomarker to appear statistically strong when, in
fact, diVerentiation in the important area where most of the results fall is
poor. For these reasons, it is important to be able to examine the distribution
of the data, so that one can assess if it is uniformly distributed. This can be
accomplished by displaying the data as scattergrams (plots) or box and
whisker plots.
Figure 4 illustrates a case in which when the data is expressed in ﬁve
intervals, it appears that the relationship between CRP concentration and
risk of disease is linear, as denoted by the dotted line. But when the data are
expressed in 10 intervals (Fig 4, below), a plateau, as indicated by the dotted
line, is observed between concentrations of CRP from about 0.64 to 5.17 mg/L.
This discrepancy occurs because the extreme values on either end are causing the
RR to reach signiﬁcance, so that when the data are plotted in too few intervals it
appears linear, but when more intervals are used, it becomes apparent that there
is little discrimination in the important middle where most of the women’s
results fall.
7. Bayesian Principles
The PPV is the probability of disease in a patient with an abnormal
biomarker. The NPV is the probability of not having the disease when the
biomarker is normal. Predictive values are sometimes called posterior (or
posttest) probability [28]. Prevalence is the pretest probability or proportion
of persons in a population at any given point having the disease. Prevalence is
also called the prior probability [28]. In the eighteenth century, the Reverend
Thomas Bayes developed an equation to assess the probability that an event
will actually occur on the basis of prior probability. For medical decision
making, Bayesian statistics relate the pretest probability to the posttest
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK 9

probability. With prognostic testing, the prevalence is often not known since
a disease has not yet developed, so the incidence may be substituted.
The TPR or diagnostic sensitivity and FPR or 1‐(diagnostic speciﬁcity)
are characteristics of the test. Whether drawn from ROC analysis or other
means, these parameters are not Bayesian in their own right. When
< 0.49
< 0.36
< 0.64 < 2.02 < 2.74 < 3.71 < 5.17 7.73< 0.1.0 < 1.46
0.36 – 0.64 – 1.00 – 1.46 – 2.02– 2.74 – 3.71 – 5.17–>
–
7.73
0.49 – 1.08 > 1.08 – 2.09 > 2.09 – 4.19 > 4.19
3.0
2.5
2.0
1.5
1.0
3.0
2.5
1.5
1.0
0.5
2.0
1
12345678910
2 345
Range of hs-CRP concentrations (mg/L)
Quintiles of hs-CRP
Decile of hs-CRP
Range of hs-CRP concentrations (mg/L)
Relative risk (RR)
FIG. 4. Framingham risk scores adjusted RR for high sensitivity (hs)‐ CRP (from the Women
Health Study of 27,939) binned into quintiles (above) and deciles (below). The numbers 1–5
(above) and 1–10 (below) indicate each quintile and decile, respectively. The range of hs‐CRP
concentrations within each quintile or decile are listed just above or below the appropriate
interval, respectively. The dotted line, above, illustrates a linear trend over the entire range of
values with quintiles. The dotted line below for deciles illustrates a plateau between about 0.64
and 5.17 mg/L of hs‐CRP, a range that includes about 50% of women between ages 30 and 49
[50]. Sixty percent of women show hs‐CRP concentrations>1 mg/L [51]. Figures were con-
structed from data expressed in tables [46, 49] modiﬁed from Ref. [6], with permission.
10 STANLEY S. LEVINSON

diagnostic sensitivity and speciﬁcity values are used to calculate predictive
values from population prevalence, they enter the realm of Bayesian
statistics.
As illustrated in Table 2, the following relationships allow classiﬁcation
between prevalence and diagnostic testing [29]: (1) true positives (TP), the
number of diseased patients correctly classiﬁed by the tests; (2) false positives
(FP), the number of patients without the disease misclassiﬁed by the test;
(3) false negatives (FN), the number of diseased patients misclassiﬁed by the
test; and (4) true negatives (TN), the number of patients without the disease
correctly classiﬁed by the test. From these deﬁnitions, the diagnostic relation-
ships shown in Table 2 can be derived.
It is important to remember, as show in Table 2, that two types of
predictive values can be calculated from preliminary testing: a conditional
predictive value and a revised or actual predictive value. The diagnostic
sensitivity and speciﬁcity are inherent properties of the test, while the revised
PPV and NPV are dependent on the prevalence of the disease in the popula-
tion. When a test sample is evaluated, the conditional predictive values will
denote the PPV and NPV in the test sample only. Unless a cohort represen-
tative of the actual population to be tested is studied, it is necessary to
calculate a predictive value that is revised to ﬁt the actual prevalence of the
population to be tested. For a disease with a prevalence of 2%, a biomarker
with a sensitivity of 99% and a speciﬁcity of 99% will give a PPV of only
66.9%, and a biomarker with a speciﬁcity and sensitivity of only 50% will give
a NPV of 98% [30]. As the prevalence of a disease approaches zero, the PPV
of a biomarker approaches zero, while as prevalence approaches 100%, the
NPV approaches zero.
Generally, the prevalence of disease in population screening is low so that
the NPV is very high while the PPV tends to be low. For this reason, even a
TABLE 2
D
EFINITIONS OFDIAGNOSTICRELATIONSHIPS
Diagnostic sensitivity or TPR¼TP/(TPþFN)
Diagnostic speciﬁcity or (1FPR)¼TN/(FPþTN)
a
Conditional predictive values:
Predictive value of a positive test result (PPV)¼TP/(TPþFP)
Predictive value of a negative test result (NPV)¼TN/(TNþFN)
Revised PPV¼
ðdiagnostic sensitivityprevalenceÞ
ðdiagnostic sensitivityprevalenceÞþ ð1specificityÞð1prevalenceÞ
a
Unless a cohort representative of the actual population to be
tested is studied, it is necessary to calculate a revised predictive
value to ﬁt the actual prevalence of the population tested and since
the prevalence of disease in populations is usually low, it is the revised
PPV that is all important.
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK
11

test with a good sensitivity and speciﬁcity of 95% will yield a low PPV and
most positive tests will be false positives. Recall, if a test has a diagnostic
sensitivity of 95% and a speciﬁcity of 95%, and there is a prevalence of 1%,
there will be 100 people with disease in 10,000 and the test can identify 95 of
them correctly. But, 5% or 495 people of the 9900 without disease will also
have positive test results, so the PPV will only be 95/590100¼16.1% or
83.9% of the positive results will be false. Thus, in general screening, the
revised PPV that is calculated as shown in Table 1 is of utmost importance.
As a result, weak diagnostic relationships between a biomarker and disease
give rise to low diagnostic sensitivities and thus very low revised PPV.
8. Weaknesses of ROC Analysis
Although the focus of this chapter is on prognosis, it is worthwhile to
brieﬂy compare diagnostic models with prognostic models since a compari-
son helps one to better understand how various degrees of accuracy ﬁt into
the overall spectrum of clinical discrimination and predictability.
8.1. D
IAGNOSTICMODELS
There are some accurate biomarker tests; nevertheless, even those in the
range of high accuracy must be carefully evaluated as to its ability for
deﬁnitive diagnosis. Troponin I for diagnosis of myocardial infarction
showed ac‐statistic of 0.99 and a diagnostic sensitivity of 96% at a speciﬁcity
of>99% [31]. This was conﬁrmed and positive troponins are now required
for deﬁnitive diagnosis of myocardial infarction [32].
B‐type naturetic peptide (BNP) was shown to have ac‐statistic of about
0.91 for diagnosis of congestive heart failure [33]. At the recommended cutoV
point of 100 pg/mL BNP shows a diagnostic sensitivity of about 90% but a
diagnostic speciﬁcity of only about 75% (FPR of 25% at TPR of 90% midway
betweenc‐statistic of 0.85 and 0.95, Fig. 1). Depending on the prevalence in a
selected population, this would result in a large number of false positive
results, but more importantly about 10% of people with the disease would be
missed. For this reason, this test was considered inaccurate for initial diag-
nosis of heart failure where echocardiography remains the test of choice [3],
but BNP has application in the emergency department [34]. Thus, in spite of
ac‐statistic in the high accuracy range, BNP did not meet the requirements
for deﬁnitive diagnosis, illustrating again the importance of examining
the ROC curve for sensitivity/speciﬁcity pairs rather than relying on the
c‐statistic alone.
12
STANLEY S. LEVINSON

8.2. PROGNOSTICMODELS
In the above examples, ROC analysis was used for diagnostic purposes.
Prognosis is more problematic than diagnosis and ROC analysis shows
various weaknesses. It has been estimated that prognostic models usually
cannot reach a maximumc‐statistic of 1.0 [12]. This is not surprising since in
disease a trait is already present, while in prognosis, the condition may not
have evolved suYciently to yet express the trait. Moreover, even genetic traits
leading to disease remain unexpressed or partially expressed.
Calibration compares the actually observed and predicted probabilities.
For tests with high accuracy, thec‐statistic for perfectly calibrated models
has been estimated to be only between 0.75 and 0.9 [12].
Thec‐statistic may be insensitive for detecting small but clinically useful
discrimination. Thec‐statistic is a ranked nonparametric score that is mini-
mally aVected by the shape of the distribution. A weakness of this approach
is that the actual scores are usually transformed into a binary‐type classiﬁca-
tion (see Table 1 and Fig. 2). If a small number of persons in a cohort exhibit
high risk while the preponderance of individuals are at low risk, binary rank‐
based measure do not take this distribution diVerence into account. More-
over, the inﬂuence of two pairs on thec‐statistic would be the same although
one pair might have a much larger diVerence than the other [12, 35].
It was noted that if, as suggested [16], an RR/OR of 3.0 was required as a strict
criterion for inclusion of each additional biomarker in risk prediction, then,
most components of the Framingham risk score would be ineligible for inclu-
sion [12]. None of the traditional risk markers of blood pressure, smoking, or
lipids achieved a RR3.0 [12], although modiﬁcation of each reduces heart
disease [26].
ROC analysis does not easily summarize survival relationships as does a
hazard ratio [15]. Nevertheless, ROC plots can be constructed from Kaplan–
Meier plots [36] or other time to event analysis [37, 38] which has become
common practice [5, 39–44].
9. Weaknesses of RR/OR
When the relationship between the consequence and the biomarker is
weak, the RR/OR is small and the between study variance is large so that
diVerent studies may show a wide range of RR/ORs and the true value
is unclear. For example, an earlier study showed an adjusted RR for CRP
in the fourth quartile of 4.1[45]. This was a nested case–control study that
only measured 366 samples. When all 27,000 samples were measured
for CRP, the adjusted RR in the ﬁfth quintile was a much lower 2.3 [46].
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK 13

Fig. 1 shows that this diVerence would be reﬂected by ac‐statistic of about
0.6 at a RR of 2.3 and 0.75 at a RR of 4.1, the diVerence between low and
intermediate accuracy, respectively. A later metastudy examining the rela-
tionship between CRP and coronary disease found a combined adjusted OR
of 1.49 when limiting the data to later studies and concluded that higher RR/
ORs found in earlier study was due to publication bias [47]. Figure 1 indi-
cates an OR of 1.49 would be equivalent to ac‐statistic of about 0.53 which
shows little or no useful diagnostic discrimination.
Very large samples are currently encouraged for clinical studies. Since the
conﬁdence level or conﬁdence interval (see Glossary) [48] is dependent on the
sample size, studies of new biomarkers using very large samples often show
statistically signiﬁcant diVerences between disease and nondisease persons with
narrow conﬁdence intervals that do not overlap after adjustment in one study
but show narrow conﬁdence intervals with no statistically signiﬁcant relation-
ship in another study. If the relationship between the new biomarker and the
disease is weak, this ambiguity can be explained in part by small diVerences in
the samples—such as diVerences in sample selection, methods bias, and con-
founding bias[28]. When the relationship is weak, these covariants may actually
show more interstudy variation than the association between the biomarker and
the disease. Also, persons at high risk may be on more medications than those at
lower risk that might further confound weak relationships.
Moreover, it is important to remember that like the ROC curve an OR/RR
represents a spectrum of cutoVvalues. For example, if a biomarker has an
OR of 3.0 at a 10% FPR (a good 90% diagnostic speciﬁcity), Fig. 1 indicates
it would only identify about 25% of the positive cases. On the other hand, if
the same biomarker identiﬁed 80% of the cases then it would have a FPR of
about 60% or speciﬁcity of only 40%. Even at an OR of 36, Fig. 1 indicates
that if a cutoVwith a TPR of about 0.95 was chosen so that almost all
positive cases would be identiﬁed, the FPR would be very large, about 0.5
(speciﬁcity only 50%). This type of exercise illustrates the advantage of
examining a spectrum of cutoVs.
Often, the RR/OR data are broken into intervals—tertiles, quartiles, or
quintiles. The lowest interval is considered normal and given a RR/OR of 1.
If there is a proportional association between the disease and the biomarker,
the higher the interval the greater the risk. This allows assessment of risk at
various concentrations (sort of a poor man’s ROC curve). Results are often
expressed in only a few intervals. Expression of data in less than 10 intervals
may lead to misinterpretation.
This problem is illustrated in Fig. 4, where data showing the relationship
between CRP and coronary disease from a cohort of 27,939 women after
adjustment from two diVerent publications are shown. When the data are
expressed in ﬁve intervals [46], it appears that the relationship between CRP
14
STANLEY S. LEVINSON

concentration and risk of disease is linear, as denoted by the dotted line
(Fig 3, above). But when the data are expressed in 10 intervals [49](Fig 3,
below), a plateau [6], as indicated by the dotted line, is observed between
concentrations of CRP from about 0.64 to about 5.17 mg/L. More the 60% of
women have CRP concentrations greater than 1 mg/L [50, 51]. Since the
prevalence of coronary disease is very low in these women, examination of
the data expressed as deciles would suggest that the biomarker would have a
very low PPV which was demonstrated as being less than 1% [52], but would
not be as apparent from the data expressed as quintiles.
Misinterpretation due to too few intervals is not limited to RR/OR, but
also applies to ROC curves. The diVerence is that programs that are now
available for ROC analysis are usually expressed as a continuous plot of all
of the data and intervals are rarely used.
10. Stand‐Alone versus Synergic Biomarkers
Generally, a stand‐alone biomarker must have a very high level of discrimi-
nation to accurately detect a disease with ac‐statistic greater than 90% such as
troponins. Since this level of discrimination is unusual for prognosis, these risk
markers are generally synergistic in that several markers are combined to
achieve a high discrimination for total risk. Thus, risk for coronary disease is
assessed using the established risk factors of age, hypertension, smoking
history, lipid status, and body mass index (BMI) [26]. For example, a recent
test for polymorphism and progression to type 2 diabetes showed an overall
RR of 1.54 [53]. Clearly, as a stand‐alone marker, it would show little or no
clinical usefulness since Fig. 1 indicates that an OR of 1.5 is about equivalent
to ac‐statistic of about 0.55. This means that at a 15% FPR the TPR would be
about 20%. Assuming a prevalence of disease of 1%, this would translate into a
PPV of 1.3% ([true positives/true positivesþfalse positives]100) (98.7 of
every 100 persons identiﬁed would be false positive). Nevertheless, such a
marker could be useful if it added additional real clinical prediction value to
existing markers for diabetes such as BMI or borderline elevated glucose.
11. Techniques for Improving Stratification of Synergic
Biomarkers
One problem with analyzing associations between genes and other global
approaches to identifying relevant biomarkers (i.e., Proteomics) is that these
approaches are not hypothesis driven but rather the associations are deﬁned
from the data and the hypothesis formulated later [7]. This means the
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK 15

relationship is more likely to be fortuitous. On the other hand, the relation-
ship between many new inﬂammatory markers and disease are hypothesis
driven. There is substantial evidence that atherosclerosis is in part an inﬂam-
matory disease [54]. As a result, it is not surprising that investigators exam-
ined inﬂammatory markers to determine if these might add additional
discrimination to traditional markers for coronary disease [6].
Although hypothesis driven, generally, new inﬂammatory biomarkers
have shown weak incremental diagnostic relationships with coronary disease
when added to existing markers (RR/OR<1.5) [5, 41, 42, 44] that places
them in the region of poor accuracy by ROC analysis (Fig. 1). Nevertheless,
the RR/ORs may show statistically signiﬁcant increase.
Clinical reclassiﬁcation has been used to assess the clinical value of syner-
gic prognostic biomarkers that show statistical signiﬁcance by RR/OR but
little diVerentiation according to thec‐statistic [12, 35]. Clinically relevant
risk categories are deﬁned and the ability of a new marker to correctly
reclassify patients who have been assigned on the basis of an old marker(s)
alone is evaluated. The change in estimated risk can then be compared for ﬁt
using the Hosmer–Lemeshow test (see Glossary) [35].
Although this approach may have merit, when diagnostic associations
with the biomarker are weak, it is limited by the same types of interstudy
variability that aVects the RR/OR. For example, in one article, reclassiﬁca-
tion of persons at intermediate risk for coronary disease to high risk on the
basis of CRP was limited to 2.7% [55], whereas, another study showed a
reclassiﬁcation of 12% for those at intermediate risk [12]. Moreover, when
the prevalence of a disease is low, reclassiﬁcation may cause many more
persons without the disease to be reclassiﬁed into the high‐risk group than
persons with the disease, giving rise to worse prediction [56]. Also, the
Hosmser–Lemeshow test shows no evidence of lack of ﬁt if the test statistic
isp0.05, with evidence of ﬁt above the 95 percentile. This means there is a
good probability of a Type II error (see Glossary) that, according to some,
makes this test unsuitable as a mean to assess precise model‐ﬁt [57].
12. Criteria for Identifying Testing of Clinical Consequence
It is important to remember that the conclusions drawn from any study are
dependent on the samples being studied and many assumptions made regard-
ing the design and appropriateness of the study [28]. Questions that should be
asked include: Does the sample of patients, controls, or other comparison
groups truly represent the population that one wishes to study? Is the study
well designed? and Are those factors that may confound the study appropri-
ately included as covariant markers or otherwise controlled for? If a
16
STANLEY S. LEVINSON

treatment is involved there may be a need for randomization. Outcome
studies are expensive and biomarkers selected to be tested must be carefully
considered after examination of all of the preliminary data. The weaker the
diagnostic relationship between the biomarker and the disease, the larger will
be the sample needed and the more expensive the outcome study. If there is
no clear treatment based on the concentration of the biomarker, continued
evaluation of the test may be pointless.
Assuming that these crucial factors are appropriate, criteria for determin-
ing the practical usefulness for biomarkers have been published in the form
of questions [3], upon which Table 3 is based. Question 1 is usually answered
by a preliminary study conducted among a group of patients with the disease
and a group without disease. Generally, if the average value of those with
disease or the RR/OR is statistically signiﬁcantly diVerent from those with-
out disease, it is concluded the answer is yes. Nevertheless, before going on to
assess question 2, it is important to consider the overlap between the two
groups. If there is a great deal of overlap between the groups, as illustrated in
Fig 4, below, the test may not have much practical value, an instance of the
breach between statistical signiﬁcance and diagnostic discrimination.
If question 1 seems true, question 2 may be tested by examining some
patients who have the disease and others who do not [3]. Again, if there is a
statistically signiﬁcant diVerence, it may be concluded that the answer to
question 2 is yes, but before going on to test question 3, cutoVvalues should
be drawn from ROC curves or other means so that improvements in diag-
nostic sensitivity and speciﬁcity or reclassiﬁcation can be evaluated. Revised
positive predictive assessment (PPV and NPV) should be calculated based on
the actual prevalence [20].
If the results from question 2 suggest the test may be clinically useful,
question 3 can be tested in a cohort study, usually retrospective studies on
stored samples. At this point, it is important to have the data evaluated with
TABLE 3
Q
UESTIONS(CRITERIA)FORDETERMININGCLINICALCONSEQUENCE
1. Do test results in patients with the target disorder diVer from those in normal people or do the
results identify those at increased risk?
2. Does the test identify patients who are suspected of having the disorder or synergistically add
discrimination to other testing (can the test be used to make or improve a diagnosis or risk
assessment)?
3. Does the test result improve the diagnostic eYciency or risk assessment beyond current testing
in what appears to be a clinically useful way?
4. Do patients who undergo this diagnostic test fare better (in their ultimate health outcomes) than
similar patients who are not tested and is the test economically reasonable?
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK
17

respect to population norms (reference range) for nonaVected people [58].
Great overlap between the normal reference range and the disease group as
shown in Fig. 4, below, may reduce the predictive values, producing a test of
questionable usefulness.
The purpose of the testing and statistical analysis to this point is to decide if
it is clinically and economically reasonable to address question 4 that can only
be answered by prospective outcome studies. These are expensive studies and
biomarkers selected must be carefully considered after examination of all of
the preliminary data. Besides the questions proposed in Table 3, another
suggestion in helping to decide which tests should be examined in outcome
studies is whether or not the test is speciﬁc for the condition targeted [58].Itis
not necessarily required that the biomarker be speciﬁc (plays an etiologic or
causal roll in the disease), called a risk factor [59], but, if it is highly nonspeciﬁc,
it is likely that covariate biomarkers, especially true risk factors, will reduce its
eVect. For example, traditional biomarkers for coronary disease include blood
pressure and measures of cholesterol [26]. Although the exact mechanisms are
not known and synergic increases inc‐statistic may be small [12, 35], evidence
indicates that these are, not only predictors, but true risk factors [7, 59, 60].
Moreover, abundant evidence indicates therapeutic moderation of these
biomarkers reduces the risk [26, 60, 61].
Clearly, it would be good to have a new test with speciﬁcity that adds to or
even more reliably predicts coronary disease than the traditional tests but, in
spite of evidence that inﬂammation is an important factor in atherosclerosis,
inﬂammatory biomarkers examined to date do not seem to meet this stan-
dard. It is unclear that the markers being measured have any causal relation-
ship. This would classify them as risk markers rather than risk factors [59].
They show little speciﬁcity, since they correlate with cancer, liver disease,
hormone therapy, various other heart conditions, and chronic and acute
diseases [6]. Besides, because of weak associations, they show RR/ORs
between 4.1 and 1.0 after adjustment for traditional risk markers [5, 6, 41,
42, 45, 47], so that the true relationship is unclear.
A related question that is important is that of absolute versus relative risk
(RR). The NCEP guidelines use absolute risk. This risk varies up to>20%
per 10‐year period. The use of RR is much more tenuous. For example, in the
Woman’s Health Study there were 121 events in about 28,000 women that
would result in about 0.4% of women having an event [46]. This is a very low
incidence that would not be apparent from the RR data. Thus, it is important
in evaluating the meaning of the results to consider the prevalence so that
accurate predictive assessments can be made. Such a low prevalence is apt to
give rise to many false positives results. The safety‐to‐beneﬁt ratio becomes a
greater issue when false positive results that represent low‐risk persons are
treated [60].
18
STANLEY S. LEVINSON

Frequentists have been contrasted with Bayesians [62]. Frequentists are
considered objectivists who in essence are classical statisticians who do not
wish to make statistical inferences beyond which the parameters of the
experiment prescribed such as the standard error of the mean, conﬁdence
interval, etc. Bayesians have been called subjectivists since they are prepared
to make inferences based on prior events or probability. It has been said that
in the extreme Frequentists must repeat an experiment an inﬁnite number of
times to deﬁne exact parameters, which cannot be achieved. Others have said
that Bayesians learn from experience, but what makes us think the future will
be like the past.
What is of interest to us is that classical statistics was intended to apply to a
few hundred data point and only to compare a few parameters [62] which
means that the classical conﬁdence level of 0.05 was certainly not intended to
deﬁne diagnostic discrimination when in some cases thousands of subjects
are tested for multiple parameters. Yet, Bayesian principles are only needed
for preliminary experiments when appropriate cohorts are not available. In
appropriate cohort studies and well‐designed outcome studies, the results are
in accord with classical objective statistics in that inferences are only from
within the experiment. In fact, since very large numbers of subjects are often
tested in clinical trials, the estimated parameters should be very tight and, if
the entire trial was repeated, although the repeat result cannot be exactly the
same as the ﬁrst, the results should be very similar. If in cohort studies, the
results disagree, it must be that the cohorts were suYciently diVerent such
that weak relationships between the biomarker and disease discrimination
was altered as a result of confounding parameters.
13. Discussion
It is important not to confuse the means with the end. That is to say, results
from testing prior to outcome studies should not be used to determine which
biomarkers are diagnostically useful, but only to determine which biomar-
kers appear to lend themselves for outcome studies. This chapter has focused
only on clinical criteria and has not considered the reliability of the measur-
ing process where it has been shown that common biomarkers may not
always be analytically reliable [1, 8].
Some have referred to guidelines similar to those listed in Table 3 as the
truth, the whole truth, and nothing but the truth [1]. Many biomarkers can
meet the ﬁrst two criteria. On the average, they diVerentiate between aVected
and unaV ected persons, even in a sample of persons suspected of having the
disease. This may be called the truth. Other tests may appear to meet
the criteria listed in question 3 in preliminary studies. This may be considered
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK 19

the whole truth. But the acid test is whether the biomarker not only improves
the eYciency beyond current testing, but also whether patients who undergo
this biomarker test fare better (in their ultimate health outcomes) than
similar patients who are not tested. This is nothing but the truth. Few recent
biomarkers have been shown to reach this standard.
14. Conclusions
Especially important in preliminary analysis are predictive assessments
which reach beyond statistical signiﬁcance. Whether ROC analysis is used
or reclassiﬁcation, predictive estimates based on the actual prevalence or
incidence can be made [56]. It has long been recognized that whenever the
number of unaVected persons is very large and the test is poorly accurate,
PPV will be very poor [30].
In spite of some weaknesses, my view is that ROC analysis is fundamental
for assessing prognostic discrimination, and ROC graphs denoting pertinent
cutoVs should be published with all papers. With survival studies, although
ROC analysis may not summarize the data in a single term as can hazard
ratio, for prognosis ROC plots can be developed for important times from
the survival analysis [5, 39, 40, 42] that for coronary disease may be about 1,
5, or 10 years.
By providing the maximal amount of information in a report, and espe-
cially information on predictive assessment, the experimenter is producing
the ammunition needed for convincing others. If it can be shown that,
although there is little improvement in the ROC curve, there is diVerentiation
between groups as deﬁned by RR/OR and reclassiﬁcation shows improved
predictions with results that are consistent between studies, this may be
acceptable for initiating an outcome study. Moreover, even if a ROC analysis
is not provided in a report, a rough assessment of the discrimination and
predictive values can be made by referring to plots similar to those shown in
Fig. 1.
These conclusions beg the question: would it be more reasonable to test for
a disease in an outcome study that shows a 7% increase in diagnostic
sensitivity at a critical cutoVvalue as shown in Fig. 2 for non‐ HDLC and
apo B, if such a relationship has been conﬁrmed in several studies, or to test
for a disease with a weak relationship deﬁned by an increase in RR/OR less
than 1.5 but little or no increase inc‐statistic and whose increment in
predictive beneﬁt is unclear. This does not mean that the latter test may
not have clinical usefulness, but when diagnostic speciﬁcity and sensitivity
are this poor, the possible usefulness of the test must be very carefully
examined. The weaker the relationship, the greater the challenge.
20
STANLEY S. LEVINSON

Glossary of Expressions and Explanations
Conﬁdence intervals (CI), distributions of data and sample size: CI contain
information similar to conﬁdence levels. They deﬁne statistical signiﬁcance.
CI represent a range of values for each variable of interest. If the CI for the
diseased group overlaps the average value reﬂecting no disease, this repre-
sents a statistically not signiﬁcant result. If there is no overlap, the results are
signiﬁcantly diVerent. Thus, the 95% CI is similar to the 0.05p‐value conﬁ-
dence level obtained from hypothesis testing. CI have an advantage that they
emphasize the size of the eVect. If the data distribution is Gaussian, the 95%
width of the CI for a two‐sided test is calculated as 1.96the standard error
of the mean (SEM), where the SEM¼standard deviation for the distribu-
tion/square root of the number of observations (n) [48] and, after calculation,
this value is added to and subtracted from the mean to produce the CI. For a
ROC curve, a one sided calculation is used so the width of the CI is 1.64
SEM added to and subtracted from thec‐statistic.
As with conﬁdence levels, CI is an inverse function ofn, becoming very
small as the number of observations become very large. If there are too few
observations, the CI will be large and the CI for the average eVect for the
disease and control groups may overlap one another. As the sample increases
in size both the conﬁdence level and CI become smaller. If the distributions
are truly diVerent, a statistically signiﬁcant diVerence will become apparent
with suYcient size. A sample of appropriate size can be calculated from the
power formula (see Type I and Type II Errors, below). It is important to
remember that as the size of the sample grows, if the average eVects stay the
same and are truly diVerent, the conﬁdence levels and CI become smaller and
eventually show a statistically signiﬁcant diVerence, but the diagnostic dis-
crimination stays the same so that thec‐statistic and RR/OR do not change.
For RR/OR, if the biomarker shows a signiﬁcant diVerence between
disease and no disease, the CI should not overlap 1.0. CI should also
accompanyc‐statistics when biomarkers are being compared. When data
are displayed as RR/OR in binned groups, CI for each bin should be
presented and CI can be obtained for sensitivity/speciﬁcity points of interest
on a ROC curve. This allows accuracy at selected cutoVs to be better
evaluated.
Hosmer Lemeshow test: A goodness of ﬁt test in which observations are
sorted and binned into about 10 groups. Within each group, the estimated
observed proportion and average expected frequencies are compared. The
statistic has aw
2
distribution withg‐2 degrees of freedom, wheregis the
number of bins (24). Like all goodness of ﬁt tests, there is a high chance of
making a Type II error.
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK 21

Likelihood ratio: Likelihood ratios can be deﬁned as the ratio between
the probability of a test result in persons that have the disease and the
probability of that result in persons without the disease. Likelihood ratios
correspond to slopes on the ROC curve, as such they are deﬁned by the
(sensitivity)/(1sensitivity) at any point[19,20]. Likelihood ratios can be a
powerful tool in conﬁrming a diagnosis because the joint likelihood ratio is
the product of each individual biomarker. Thus, the likelihood ratios for several
diVerent independent biomarkers can be combined to help in conﬁrming or
refuting a disease. It is important to remember that like other parameters of
ROC analysis, likelihood ratios are not Bayesian until adjusted for prevalence.
Thus, there is a conditional and revised likelihood and the revised likelihood can
be expressed as a percent [20].
Odds ratio: Odds is the probability of those with disease divided by those
without disease, expressed as: subjects with disease/(1subjects with dis-
ease). The odds ratio is the odds that the cases have particular test results
divided by the odds that the controls have the particular test result.
Relative risk: RR is measured as the ratio of disease incidence in those
positive for a particular test (above some cutoVvalue) and the incidence in
those negative for the test (below the cutoVvalue). Expressed as: Incidence of
disease in exposed/Incidence of disease in unexposed.
Type I and Type II Error:Type I or alpha (a) error is expressed by the
conventionalpvalue. This is the 95% conﬁdence level (p0.05). This means
there is one chance in 20 of making a Type I error. On the other hand, a study
might conclude there is no diVerence between the disease group and the
control group when, in fact, there is a diVerence. This is a type II or beta
(b) error—often referred to as the false negative rate. This usually occurs
because the samples were too small. The probability for an adequate sample
size is determined by calculating the power: Power¼(1b).
A
CKNOWLEDGMENT
This work was supported by the Department of Veteran AVairs, Louisville, Kentucky, USA.
REFERENCES
[1] J.O. Westgard, T. Darcy, The truth about quality: medical usefulness and analytical
reliability of laboratory tests, Clin. Chim. Acta 346 (2004) 3–11.
[2] B. Bozkurt, D.L. Mann, Use of biomarkers in the management of heart failure: are we there
yet? Circulation 107 (2003) 1231–1233.
[3] D.L. Sackett, R.B. Haynes, The architecture of diagnostic research, BMJ 324 (2002)
539–541.
[4] N. Rifai, D.G. Altman, P.M. Bossuyt, Reporting bias in diagnostic and prognostic studies:
time for action, Clin. Chem. 54 (2008) 1101–1103.
22 STANLEY S. LEVINSON

[5] A.R. Folsom, L.E. Chambless, C.M. Ballantyne, J. Coresh, G. Heiss, K.K. Wu, et al.
An assessment of incremental coronary risk prediction using C‐ reactive protein and other
novel risk markers: the atherosclerosis risk in communities study, Arch. Intern. Med. 166
(2006) 1368–1373.
[6] S.S. Levinson, Inﬂammatory and long‐term risk markers, Clin. Lab. Med. 26 (2006)
553–570.
[7] J. Loscalzo, Association studies in an era of too much information: clinical analysis of new
biomarker and genetic data, Circulation 116 (2007) 1866–1870.
[8] D.A. Morrow, J.A. de Lemos, Benchmarks for the assessment of novel cardiovascular
biomarkers, Circulation 115 (2007) 949–952.
[9] P.M. Ridker, P.W. Wilson, S.M. Grundy, Should C‐reactive protein be added to metabolic
syndrome and to assessment of global cardiovascular risk? Circulation 109 (2004)
2818–2825.
[10] N. Rifai, P.M. Ridker, Proposed cardiovascular risk assessment algorithm using high‐
sensitivity C‐reactive protein and lipid screening, Clin. Chem. 47 (2001) 28–30.
[11] Y.M. Smulders, A. Thijs, J.W. Twisk, New cardiovascular risk determinants do exist and
are clinically useful, Eur. Heart J. 29 (2008) 436–440.
[12] N. Cook, Use and misuse of the receiver operator characteristic curve in risk prediction,
Circulation 115 (2007) 928–935.
[13] B. Dawson‐Saunders, R.G. Trapp, Statistical Methods for Multiple Variables in Basic and
Clinical Biostatistics. Appleton and Lange, Norwalk, CN, 1990.
[14] J. Zhang, K.F. Yu, What’s the relative risk? A method of correcting the odds ratio in cohort
studies of common outcomes, JAMA 280 (1998) 1690–1691.
[15] S.R. Rao, D.A. Schoenfeld, Survival methods, Circulation 115 (2007) 109–113.
[16] M.S. Pepe, H. Janes, G. Longton, W. Leisenring, P. Newcomb, Limitations of the odds
ratio in gauging the performance of a diagnostic, prognostic, or screening marker, Am.
J. Epidemiol. 159 (2004) 882–890.
[17] E.K. Shultz, Multivariate receiver‐operating characteristic curve analysis: prostate cancer
screening as an example, Clin. Chem. 41 (1995) 1248–1255.
[18] M.H. Zweig, S.K. Broste, R.A. Reinhart, ROC curve analysis: an example showing the
relationships among serum lipid and apolipoprotein concentrations in identifying patients
with coronary artery disease, Clin. Chem. 38 (1992) 1425–1428.
[19] M.H. Zweig, G. Campbell, Receiver‐operating characteristic (ROC) plots: a fundamental
evaluation tool in clinical medicine, Clin. Chem. 39 (1993) 561–577.
[20] S.G. Pauker, R.I. Kopelman, Interpreting hoofbeats: can Bayes help clear the haze?
N. Engl. J. Med. 327 (1992) 1009–1013.
[21] J.S. Krouwer, Cumulative distribution analysis graphs—an alternative to ROC curves,
Clin. Chem. 33 (1987) 2305–2306.
[22] V.A. DeBari, Computation of decision levels from diVerentiated logistic regression proba-
bility curves, Ann. Clin. Lab. Sci. 36 (2006) 194–200.
[23] J.A. Hanley, B.J. McNeil, The meaning and use of the area under a receiver operating
characteristic (ROC) curve, Radiology 143 (1982) 29–36.
[24] J.A. Swets, Measuring the accuracy of diagnostic systems, Science 240 (1988) 1285–1293.
[25] S.S. Levinson, Comparison of apolipoprotein B and non‐high‐density lipoprotein choles-
terol for identifying coronary artery disease risk based on receiver operating curve analysis,
Am. J. Clin. Pathol. 127 (2007) 449–455.
[26] Expert panel on detection, evaluation, and treatment of high blood cholesterol in adults,
Executive summary of the third report of the National Cholesterol Education Program
(NCEP). Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol
in Adults (Adult Treatment Panel III, JAMA 285 (2001) 2486–2497.
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK
23

[27] N.J. Wald, A.K. Hackshaw, C.D. Frost, When can a risk factor be used as a worthwhile
screening test? BMJ 319 (1999) 1562–1565.
[28] R. Fletcher, S.W. Fletcher, Clinical Epidemiology, Lippincott Williams and Wilkins, Balti-
more, 2005.
[29] R. John, M.S. Lifshitz, J. Hang, D. Fink, Post‐analysis: medical decision‐making, Henry’s
Clinical Diagnosis and Management by Laboratory Methods, 21 ed Saunders Elsevier,
Philadelpia, 2007, 69–75.
[30] T.J. Vecchio, Predictive value of a single diagnostic test in unselected populations, N. Engl.
J. Med. 274 (1966) 1171–1173.
[31] J.E. Adams, K.B. Schechtman, Y. Landt, J.H. Ladenson, S. JA, Comparable detection of
acute myocardial infarction by creatine kinase MB isoenzyme and cardiac troponin I, Clin.
Chem. 40 (1994) 1291–1295.
[32] K. Thygesen, J.S. Alpert, H.D. White, A.S. Jaffe, F.S. Apple, M. Galvani, et al. Universal
deﬁnition of myocardial infarction, Circulation 116 (2007) 2634–2653.
[33] A.S. Maisel, P. Krishnaswamy, R.M. Nowak, J. McCord, J.E. Hollander, P. Duc, et al.
Rapid measurement of B‐type natriuretic peptide in the emergency diagnosis of heart
failure, N. Engl. J. Med. 347 (2002) 161–167.
[34] R. Pﬁster, C.A. Schneider, Natriuretic peptides BNP and NT‐ pro‐BNP: established labora-
tory markers in clinical practice or just perspectives? Clin. Chim. Acta 349 (2004) 25–38.
[35] N. Cook, Statistical evaluation of prognostic versus diagnostic models: beyond the ROC
curve, Clin. Chem. 54 (2008) 17–23.
[36] J.M. Bland, D.G. Altman, Survival probabilities (the Kaplan–Meier method), BMJ 317
(1998) 1572.
[37] D.G. Altman, J.M. Bland, Time to event (survival) data, BMJ 317 (1998) 468–469.
[38] P.J. Heagerty, T. Lumley, M.S. Pepe, Time‐dependent ROC curves for censored survival
data and a diagnostic marker, Biometrics 56 (2000) 337–344.
[39] M.J. Bos, C.M. Schipper, P.J. Koudstaal, J.C. Witteman, A. Hofman, M.M. Breteler, High
serum C‐reactive protein level is not an independent predictor for stroke: the Rotterdam
Study, Circulation 114 (2006) 1591–1598.
[40] J.J. Cao, A.M. Arnold, T.A. Manolio, J.F. Polak, B.M. Psaty, C.H. Hirsch, et al. Associa-
tion of carotid artery intima‐media thickness, plaques, and C‐reactive protein with future
cardiovascular disease and all‐ cause mortality: the Cardiovascular Health Study, Circula-
tion 116 (2007) 32–38.
[41] A.C. St-Pierre, B. Cantin, J. Bergeron, M. Pirro, G.R. Dagenais, J.P. Despres, et al.
Inﬂammatory markers and long‐term risk of ischemic heart disease in men. A 13‐year
follow‐up of the Quebec Cardiovascular Study, Atherosclerosis 182 (2005) 315–321.
[42] T.J. Wang, P. Gona, M.G. Larson, G.H. Toﬂer, D. Levy, C. Newton-Cheh, et al. Multiple
biomarkers for the prediction of ﬁrst major cardiovascular events and death, N. Engl.
J. Med. 355 (2006) 2631–2639.
[43] S. Blankenberg, M.J. McQueen, M. Smieja, J. Pogue, C. Balion, E. Lonn, et al. Compara-
tive impact of multiple biomarkers and N‐Terminal pro‐brain natriuretic peptide in the
context of conventional risk factors for the prediction of recurrent cardiovascular events in
the Heart Outcomes Prevention Evaluation (HOPE) Study, Circulation 114 (2006) 201–208.
[44] I. Tzoulaki, G.D. Murray, A.J. Lee, A. Rumley, G.D. Lowe, F.G. Fowkes, Relative value
of inﬂammatory, hemostatic, and rheological factors for incident myocardial infarction and
stroke: the Edinburgh Artery Study, Circulation 115 (2007) 2119–2127.
[45] P.M. Ridker, J.E. Buring, J. Shih, M. Matias, C.H. Hennekens, Prospective study of
C‐reactive protein and the risk of future cardiovascular events among apparently healthy
women, Circulation 98 (1998) 731–733.
24 STANLEY S. LEVINSON

[46] P.M. Ridker, N. Rifai, L. Rose, J.E. Buring, N.R. Cook, Comparison of C‐reactive protein
and low‐density lipoprotein cholesterol levels in the prediction of ﬁrst cardiovascular events,
N. Engl. J. Med. 347 (2002) 1557–1565.
[47] J. Danesh, J.G. Wheeler, G.M. Hirschﬁeld, S. Eda, G. Eiriksdottir, A. Rumley, et al.
C‐reactive protein and other circulating markers of inﬂammation in the prediction of
coronary heart disease, N. Engl. J. Med. 350 (2004) 1387–1397.
[48] A.R. Henderson, Chemistry with conﬁdence: should clinical chemistry require conﬁdence
intervals for analytical and other data? Clin. Chem. 39 (1993) 929–935.
[49] P.M. Ridker, N. Cook, Clinical usefulness of very high and very low levels of C‐reactive
protein across the full range of Framingham Risk Scores, Circulation 109 (2004) 1955–1959.
[50] E.S. Ford, W.H. Giles, A.H. Mokdad, G.L. Myers, Distribution and correlates of
C‐reactive protein concentrations among adult US women, Clin. Chem. 50 (2004) 574–581.
[51] S. Woloshin, L.M. Schwartz, Distribution of C‐reactive protein values in the United States,
N. Engl. J. Med. 352 (2005) 1611–1613.
[52] S.S. Levinson, R.J. Elin, What is C‐reactive protein telling us about coronary artery disease?
Arch. Intern. Med. 162 (2002) 389–392.
[53] J.C. Florez, K.A. Jablonski, N. Bayley, T.I. Pollin, P.I. deBakker, A.R. Shuldiner, et al.
TCF7L2 polymorphisms and progression to diabetes in the Diabetes Prevention Program,
N. Engl. J. Med. 355 (2006) 241–250.
[54] R. Ross, Atherosclerosis—an inﬂammatory disease, N. Engl. J. Med. 340 (1999) 115–126.
[55] T.J. Wang, M.G. Larson, R.S. Vasan, Biomarkers for prediction of cardiovascular events,
N. Engl. J. Med. 356 (2007) 1472–1475.
[56] M.S. Pepe, H. Janes, J.W. Gu, Use and misuse of the receiver operating characteristic curve
in risk prediction, Letter by Pepe et al. regarding article. Circulation 116 (2007) e132;
author reply e134.
[57] S.J. Janket, Y. Shen, A.E. Baird, Why must new cardiovascular risk factors be carefully re‐
assessed prior to clinical application? Eur. Heart J. 29 (2008) 1336–1337; author reply 1337.
[58] L. Mosca, C‐reactive protein—to screen or not to screen? N. Engl. J. Med. 347 (2002)
1615–1617.
[59] T.J. Wang, New cardiovascular risk factors exist, but are they clinically useful? Eur. Heart J.
29 (2008) 441–444.
[60] S.M. Grundy, Promise of low‐density lipoprotein‐lowering therapy for primary and sec-
ondary prevention, Circulation 117 (2008) 569–573; discussion 573.
[61] S.L. Hardoon, P.H. Whincup, L.T. Lennon, S.G. Wannamethee, S. Capewell, R.W. Morris,
How much of the recent decline in the incidence of myocardial infarction in British men can
be explained by changes in cardiovascular risk factors? Evidence from a prospective
population‐based study, Circulation 117 (2008) 598–604.
[62] E. Bradley, Bayesians, frequentists, and scientists, J. Am. Stat. Assoc. 100 (2005) 1–5.
CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK
25

This page intentionally left blank

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS IN
CARDIOVASCULAR DISEASE: EVIDENCE FROM IN VITRO
ANDIN VIVOSTUDIES
M. Ghayour-Mobarhan,*
,†
A.A. Rahsepar,*
,†
S. Tavallaie,
†
S. Rahsepar,*
,†
and G.A.A. Ferns
‡,1
*Cardiovascular Research Center, Avicenna Research Institute,
Mashhad University of Medical Science (MUMS),
Mashhad 91376-73119, Iran
†
Department of Nutrition and Biochemistry, Faculty
of Medicine, MUMS, Mashhad 91376-73119, Iran
‡
Postgraduate Medical School, University of Surrey,
Guildford, Surrey GU2 7WG, UK
1. Abstract . ........................................................................ 28
2. Introduction. .................................................................... 28
2.1. Discovery of the HSPs, Their Classiﬁcation and Their Functions ........... 28
2.2. Atherosclerosis............................................................ 28
3. HSPs and Atherogenesis......................................................... 34
3.1. HSPs and Animal Models of Atherogenesis................................ 36
3.2. Modulation of HSP Expression in Cells Involved
in AtherogenesisIn Vitro.................................................. 38
3.3. Soluble or Circulating HSPs............................................... 44
4. HSPs and Autoimmunity in Atherogenesis...................................... 45
4.1. General Consideration. . ................................................... 45
4.2. Molecular Mimicry and Relation to Infection.............................. 47
4.3. Antibodies to HSPs and Infections......................................... 47
4.4. Antibodies to HSPs and Cardiovascular Risk Factors...................... 48
4.5. Antibody Titers to HSPs and Their Relationship to CVD Burden.......... 54
4.6. Changes in Titers of HSP Antibodies During Acute
Coronary Syndromes . ................................................... 56
5. Therapeutic Implications........................................................ 58
6. Conclusions . .................................................................... 59
References....................................................................... 59
1
Corresponding author: GAA Ferns, e-mail: [email protected]
27
0065-2423/09 $35.00 Copyright 2009, Elsevier Inc.
DOI: 10.1016/S0065-2423(09)48002-8 All rights reserved.
ADVANCES IN CLINICAL CHEMISTRY, VOL.
48

1. Abstract
The heat shock proteins (HSPs) are highly conserved families of proteins
expressed by a number of cell types following exposure to stressful environ-
mental conditions. These conditions include several known risk factors for
cardiovascular disease. A number of the HSPs have been shown to be
molecular chaperones that are involved in the refolding of other damaged
protein molecules. Over the past two decades there has been an increasing
interest in the relationship between HSPs and cardiovascular disease, and
particularly whether an autoimmune response may be implicated. The fact
that microorganisms also produce HSPs, and that these are homologous to
human HSPs has given rise to concept of molecular mimicry. While most of
the past studies have focused on HSP 65 and 70, there has been recent interest
and investigations of the possible role of the smaller HSPs, such as HSP27,
in atherogenesis. Furthermore, the possibility that autoimmunity may be
mediating the deleterious eVects of HSPs has led some investigators to
explore tolerization as a potential therapeutic approach.
2. Introduction
2.1. D
ISCOVERY OF THEHSPS,THEIRCLASSIFICATION ANDTHEIRFUNCTIONS
Approximately four decades ago, Ritossa and colleagues [1] observed that
exposing larval salivary glands from Drosophila to heat induced speciﬁc genes
in the giant chromosomes of the gland cells; it is now known that these genes
encode proteins called HSPs. The HSPs are highly conserved families of
proteins found in the cells of all organisms and several of them are known to
function as molecular chaperones. The HSPs may be divided into seven major
families according to their molecular weights: HSP10, small HSPs (15–30
kDa), HSP40, HSP60, HSP70, HSP90, and HSP100 (Table 1). HSP expression
is increased in response to several environmental stresses in addition to heat
stress; these include: certain forms of nutritional deﬁciency, oxidative stress,
and ultraviolet radiation. This is mediated by the release of heat shock factor 1
and its binding to heat shock elements in the ﬂanking regions of the HSP genes
[2](Fig. 1). Moreover, in addition to their role as chaperones, HSP have other
putative roles [3–6]. Table 1 shows a summary of their functions.
2.2. A
THEROSCLEROSIS
Atherosclerosis is a chronic multifactorial disease that underlies the path-
ophysiology of cardiovascular disease (CVD), stroke and peripheral vascular
disease (PVD), and is the major cause of mortality worldwide [7, 8]. It is
28
GHAYOUR-MOBARHAN ET AL.

TABLE 1
S
UMMARY OF THE
N
OMENCLATURE
,L
OCATION
,
AND
F
UNCTION OF THE
M
AJOR
H
EAT
S
HOCK
P
ROTEIN
F
AMILIES
FamilyOrganism HSP ‐related proteinsLocationFunctions
Small HSPsE. coliLbp A and BCytosolSuppresses aggregation and heat inactivation of proteins
in vitro; confers thermotolerance through stabilization
of microﬁlaments; antiapoptotic activity
S. cerevisiaeHSP27Cytosol
A and B crystallin Cytosol
HSP27Cytosol
Hsp40E. coliDnaJCytosolEssential cochaperone activity with Hsp70 proteins to
enhance rate of adenosine triphosphatease activity and
substrate release
S. cerevisiaeYdj 1Cytosol/nucleus
MammalsHdj 1 and Hdj 2
Hsp60E. coliGroELCytosol
mitochondria
Refolds and prevents aggregation of denatured proteins
in vitro; may facilitate protein degradation by acting as
a cofactor in proteolytic system; role in the assembly of
bacteriophages and Rubisco (an abundant protein in
the chloroplast)
S. cerevisiaeHSP60Chloroplasts
mitochondria
PlantsCpn60
MammalsHSP60
Hsp70E. coliDnaKCytosolRoles in lambda phage replication; autoregulation of the
heat shock response; interaction with nascent chain
polypeptides; functions in interorganellar transport;
roles in signal transduction; refolds and maintains
denatured proteinsin vitro; role in cell cycle and
proliferation; antiapoptotic activity; potential antigen ‐
presenting molecule in tumor cells
S. cerevisiaeSsa 1–4Cytosol
Ssb 1,2Cytosol
Kar2ER mitochondria
Ssc1
MammalsHSC70Cytosol/nucleus
Cytosol/nucleus
ER mitochondria
HSP70
BIP
MHSP70
Hsp90E. coliHtpGCytosolRole in signal transduction (e.g., interaction with steroid
hormone receptors, tyrosine kinases, serine/threonine
kinases); refolds and maintains proteins in vitro;
autoregulation of the heat shock response; role in cell
cycle and proliferation
S. cerevisiaeHSP83Cytosol
MammalsHSP90Cytosol
GRP94ER
Hsp100E. coliCytosolRole in stress tolerance; helps the solubilization of heat ‐
inactivated proteins from insoluble aggregates S. cerevisiaeCytosol
HSP, heat shock protein;E. coli,Escherichia coli;S. cerevisiae,Saccharomyces cerevisiae; ER, endoplasmic reticulum.
Modiﬁed from Lambet al. [2]. Publisher and year of copyright: Elsevier, 2002.

characterized by the accumulation of lipids and extracellular matrix in the
intima of large and medium sized arteries. It is associated with mononuclear
cell inﬁltration, and smooth muscle proliferation [9]. Risk factors for CVD
include: age, male sex, family history of CVD, hypertension, hypercholester-
olemia, smoking, diabetes mellitus, socioeconomic status, and obesity [9].
There are several emerging risk factors for CVD including markers of
oxidative stress, inﬂammation, and autoimmunity [10].
2.2.1.Atherosclerosis and the Role of Inﬂammation
The inﬂammatory nature of atherosclerosis was ﬁrst described in the 1850s
[11], however, more recent interest has developed because immunocytochem-
ical studies have allowed the cellular composition of atherosclerotic plaques
to be determined and related to the onset of clinical events, such as plaque
rupture [12]. Furthermore, inﬂammatory processes also appear to be involved
in atherogenesis [13]. The earliest lesions in atherogenesis, are fatty streaks,
and these are commonly found in infants and young children [14]. They are
characterized by a relative paucity of lipid accumulation and comparative
abundance of intimal inﬂammatory cells that include activated T lymphocytes
(helper, suppressor, and regulator), mast cells, macrophages, dendritic
cells [15], and less commonly granulocytes and NK cells [16–18].
Heat shock
HSF1 monomer
HSF
trimerisation
Nuclear
translocation
Binding of HSF to
heat shock concensus
element (HSE)
TATA A
GAAnnTTCnnGAA
Regulatory protein
complex
FIG. 1. Schematic representation of the regulation of mammalian heat shock protein expres-
sion. HSF, heat shock transcription factors; HSE, heat shock consensus element; TATAA,
DNA sequence containing TATAA repeats. Reference: [2]. Publisher and year of Copyright:
Elsevier, 2002. Permission for reproduction/adaptation was granted by the copyright holder.
30 GHAYOUR-MOBARHAN ET AL.

Epidemiological studies have supported the role of inﬂammation in CVD.
Serum C-reactive protein (CRP) concentrations have been reported to be a
stronger independent predictor of coronary events than low density lipopro-
tein (LDL) cholesterol levels [19–22]. It has also been reported that elevated
levels of soluble intercellular adhesion molecule (ICAM)-I, a marker of
endothelial cell activation, are associated with increased coronary risk [23]
and its expression is increased in human atherosclerotic lesions [24]. Comple-
ment activation [25–27] may play a role in endothelial injury during athero-
genesis and may be a consequence of autoimmune responses to modiﬁed
LDL [28] or denatured HSPs [29]. The expression of human lymphocytic
antigen (HLA) class II antigen and secretion of several cytokines, within
atherosclerotic lesions supports the involvement of inﬂammation in athero-
sclerosis [30]. Advanced atheromatous lesions also contain large numbers
of T lymphocytes [30], most of which are T helper (h) type 1 cells bearing
alpha/beta receptor [17]. Furthermore, activated T cells bearing gamma/delta
receptors are abundant at the earliest stages of atherogenesis [31] and
atherosclerosis can be inhibited by depletion of T lymphocytes [32].
Xuet al.[33] have suggested that CD4+ cells predominate within the T cell
population in early lesions, while Van Der Walet al.[18] have reported an
increased CD8/CD4 ratio in both early and late lesions. There is a prepon-
derance of pro-inﬂammatory Th1 cells expressing IFN-gand IL-2 compared
to Th2 cells producing interleukin (IL)-4, IL-5, and IL-10 [34, 35]. In apoli-
poprotein E deﬁcient mice it has been reported that Th1-inhibition is asso-
ciated with a 60% reduction in atherosclerotic lesion area [36]. Regulatory
T cells (Treg) are a subpopulation of T cells which exert important regulatory
eVects on immune function [37–39], Type 1 Treg cells can inhibit immune
responses by secreting TGF-band IL-10 [40, 41], while Th2 cells suppress
inﬂammation and dampen macrophage activity via a broader spectrum of
anti-inﬂammatory cytokines, and may have protective eVects against athero-
genesis [35, 42–44]. Switching the balance of activity from Th1 to Th2 may
therefore be protective in atherogenesis [45]. Depletion of CD4+ and CD8+
T cells has been reported to reduce the formation of fatty streaks in C57BL/
6J mice [46], which supports the importance of T cells in atherogenesis.
However, there remains controversy about the precise role of cellular
immunity in atherogenesis as some studies have shown that immune-suppres-
sion may result in enhanced atherogenesis in experimental models [47, 48].
2.2.2.Atherosclerosis and the Role of Infection
Several studies have shown a positive association between the degree of
atherosclerosis burden and presence of chronic infectious microorganisms
[19, 49], these include: the Herpes group of viruses, notablyCytomegalovirus
(CMV) andherpes simplex virus type 1(HSV-1) [50],Helicobacter (H) pylori
THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS 31

[51],Chlamydia (C) pneumonia[52],Hepatitis A virus(HAV) [53], and
infectious organisms that give rise to gingivitis [54]. These infective processes
may exert a pro-atherogenic eVect in early life. Pesonen and coworkers [55]
have shown that in young children, the presence of antibodies to several
microorganisms was positively associated with carotid intimal thickening,
a marker of atherosclerosis. It has been proposed that infection acquired
during childhood may lead to atherosclerosis in later life [56]; and it has been
reported that there is a positive association between the number of infectious
organisms a person has been exposed to and the extent of CVD [57](Fig. 2).
Splenectomy is associated with an increased susceptibility to both infection
by organisms such asC. pneumoniaand more severe atherosclerosis [58–60].
Individuals with chronic infections have high serum levels of HSP60, which
are also associated with severity of atherosclerosis [61]. The potential
mechanisms by which infections may induce atherosclerosis and their inter-
action with other pro-inﬂammatory processes is shown in Figs. 3 and 4.
Pathogen burden
Number of seropositivities
3.1
9.8
P < 0.001
Control
Control Limited disease
Limited
disease
Advanced
disease
Advanced disease
15.0
20
15
10
5
0
20
25
15
10
5
0
000
1.4
5.9
7.7
7.0
14.9
20.0
0–3 4–5 6–8
Mortality (%) Mortality (%)
Extent of disease
0
3.5
13.9
20
15
10
5
0
Mortality (%)
6–8
4–5
0–3
Number of
seropositivities
FIG. 2. Cardiovascular mortality rate according to pathogen burden and extent of atheroscle-
rosis. Reference: [57]. Publisher and year of Copyright: American Heart Association, 2002.
Permission for reproduction/adaptation was granted by the copyright holder.
32 GHAYOUR-MOBARHAN ET AL.

Another Random Scribd Document
with Unrelated Content

lobbantották gyulékony szenvedélyét. A leány Zeráhnak nevezte
magát, s a királyi park kertészének szülötte volt.
A válasz, a mivel a leány Mahadi szerelmi ostromát fogadta,
egészen foglyul ejté az ostromlót.
– Inkább akarok egy koldusnak a felesége lenni, mint egy
királynak a rabnője.
Az ilyen felelet megérdemli azt a viszonzást hogy:
– Légy tehát egy királynak a felesége!
Erre aztán nehezebb volt azt mondani, hogy «nem».
Zeráh próbálgatta azt egy ideig: gondolkozott rajta, hogy a
kezében tartott rózsát összetépje-e s odadobja a király lábaihoz?
Végre győzött a nagyravágyás s oda nyujtá a rózsát a királynak.
Mahadi lelke megtelt a nagy örömmel. Mikor egy király a porból
emel föl egy igazgyöngyöt, mely az ő koronájába foglaltatni
érdemes, még nagyobbra van vele, mint ha egy másik király
diadémjából törte volna ki azt. Mahadi menyegzőjére összehivta
minden vazalljait és vezéreit, a szövetséges kalifákat, s azok meg is
jelentek a menyegzőre mind, egynek kivételével: ez volt Ilhámi.
Annak elébb a maroccoi szultán ostromzároló hajó-hadát kellett
elverni Cadix alól; addig nem jöhetett.
A menyegző hét napig tartott, és a hét napot összekötő
éjszakáknak sem volt álomhozó pihenése, a fényt, mámort, pazarlást
a nappal és az éjjel versenyezve fokozta egymás fölé.
A hetedik napon megjött Ilhami.
Győztesen érkezett meg; az ellenség elsülyesztett hajóiról
letépett zászlókkal. Azokat hozta nászajándékul a bátyjának.
Mikor aztán Mahadi odavezette őt menyasszonya elé, Ilhami
eltakarta arczát a kezével, s elkezdett zokogni:

– Mit tettél? Elraboltad az egyetlen kincsemet a világon!
Zeráh Ilhami herczegnek a menyasszonya volt már akkor, mikor a
király őt meglátta és megszerette.
És így az első, a mitől Almansor óvta a fiát, már megtörtént. S
azon többé segíteni nem lehetett.
Mahadi nagyon szerette Zeráht, bár az a tudás, hogy már elébb
az öcscsének volt eljegyezve s azt «talán» szerette is, örökké kihajtó
tövis-gyökér maradt a szivében.
Zerah boldog apává tette a királyt. Első szülöttét Muzának
nevezték.
Mahadinak azonban nem csak arra volt gondja, hogy életet
adjon, királynak az életelvevés is nehéz kötelességei közé tartozik.
Nagy szigorral bánt el a martalócz néppel. A belső ellenséget irtotta
fegyverrel, s megtisztítá a rablóktól a tengert s az erdőt.
Egyszer kézrekerült maga a sierra nevadai hegyszakadékok réme:
a sarczoló Toma rablófőnök is. Felhozták őt lánczraverten Cordovába,
s törvényt tartottak fölötte. Tengernyi gonosztett derült ki rá. Ha
annyi feje lett volna is neki, mint a tündérországban lakó Majmuna
dzsinnek, bizony egy sem maradt volna meg neki belőle. A birák
kifáradtak annak a feljegyzésébe, a mit ő maga bevallott, s a mit a
minden vidékről betóduló károsultak siettek felőle beárulni.
És mégis volt bátorsága Tomának, mikor már büneinek
sokaságából az egész oszlop fel volt rakva, megkegyelmeztetéseért
folyamodni a kalifához.
Ő maga megirta levélben az indokot, a miért azt meri kivánni a
királytól, hogy minden borzasztó bünlajstroma daczára,
kegyelmezzen meg az életének, sőt bocsássa őt szabadon.
Az alkáde, a kire rá volt bizva, hogy e kegyelemkérő levelet
átadja a kalifának, Zeráhnál találta őt.

– Hagyj békét mostan, mondá a király. A feleségem mesél a kis
fiamnak a «kigyókirálynéról», azt nem szakíthatom félbe.
Mikor később visszajött az alkáde, akkor azt mondá neki a király
az ebédnél ülve:
– Bizony te is okosabbat gondolhatnál ki, mint az embernek az
emésztését zavarni azzal, hogy gonosztevők dolgaival keverd föl az
epéjét.
Megint eljött az alkáde a kegyelemkérő levéllel. Akkor a király a
fürdőben ült:
– Csak nem kivánod, hogy az alatt olvassak zsiványrimánkodást,
míg a borbély a fejemet szappanozza!
Ismét előjött estefelé az alkáde az irással, s megvárta, míg a
király a vacsorától fölkel.
– Hidd el nekem: annyit ittam, hogy kettőst látok; tánczol előttem
minden betű: el nem tudom olvasni a leveledet most! Jöjj később.
Az alkáde még egyszer eljött, most álmából költé fel a királyt.
– Át kell venned a kegyelemkérő levelet, mert magad
parancsoltad, hogy addig ki ne végezzenek semmi elitéltet, míg kérő
levelét el nem fogadtad.
– Nagyon édes álmomból háborítottál fel, mondá a király. No hát
ad ide azt a levelet!
S hogy az édes álma folytatását el ne mulaszsza, a vánkosa alá
tette az elitélt könyörgő levelét s aludt tovább.
A rablóvezért pedig, miután éjfélig nem jött vissza a kérő levél a
király kegyelmével, éjfél után lenyakazták törvény és igazság szerint.
Csak másnap reggel, mikorra kialudta mámorát, vette elő a rabló
kegyelemkérő levelét Mahadi s azt olvasta belőle:

– Mikor nagy apádat Omár szultánt, a szaszszanidák ostromolták
Granadában s ő előre látta, hogy elfogják nyomni, az omaridák
minden kincsét összehordatta és befalaztatá palotája boltozatai közé.
E kincsek százötven teveterühre mentek s közöttük az arany-asztal,
mely körül a kalifák valamennyi vezérekkel együtt étkeztek, a Tarik
vezér viselte paizs, csupa gyémántokkal kirakva, hogy azon kard
nem fogott, s minden ellenség, a ki belenézett, elvakult tőle, a
gyászruha, a mit Ayesha özvegységében viselt, csupa fekete igaz
gyöngyökkel kivarrva; tizenkét ezüst láda, mind tele Lizimachus-
aranyakkal, minden ládának a fölemeléséhez nyolcz erős férfi kellett.
Ott volt a csodálatos alhambrai arany-hattyú, mely magától úszott,
repdesett és énekelt. És az arany-oroszlánok, tojás nagyságú rubin
szemekkel, a mik az elfogadási terem bejáratát őrizték, s mikor a
kalifa belépett, felálltak és a farkaikat elkezdték csóválni. Ott volt az
arany trón, melyről a törvényt kihirdették, és az egy darabból
faragott hegykristály láda, melyben a proféta kegylevele őriztetett,
melyet a legelső omaridának, saját bőrköpenyéből lehasítva, átadott.
Ott volt a gyémántok legnagyobbika, az Aldebaran, mely az
oroszlánok termében felakasztva éjjel is oly világot terjesztett abban,
mintha nappal volna. De a mi mind ezen kincseknél becsesebb, ott
volt a «bölcsek köve», melyet Paracelsus ajándékozott, s melynek
birtokosa minden jövendőt előre megtud, valamint, hogy Omár kalifa
is megtudta előre, hogy az ostromlók legelébb a cadixi toronynál
fognak betörni. Mindezeket a kincseket a kalifa ötszáz épitészszel
hatvan nap alatt úgy elrejteté, hogy azoknak nyomára nem lehetett
akadni. A palotának egy termét belül egészen lebontották s a mint a
kincsek el voltak rejtve a falak közé, ismét mindent úgy
helyreállítottak, márvány faragványok, oszlopok visszakerültek a
helyeikre, hogy többé se ismerős, se idegen meg nem tudhatta,
hová vannak a kincsek elrejtve.
Mikor a mű be volt végezve, a kalifa nagy lakomát adott az
ötszáz épitésznek; hanem a bor meg volt mérgezve «borasz»-szal.
Mind meghaltak tőle, egynek kivételével, a ki olyan okos volt, hogy a
poharába töltött bor közé tengeri sót öntött, a mi a «borasz» növény
mérgét megöli. Ez az egy én voltam. Én is halottnak tettetém

magamat, nehogy erőszakkal végezzenek ki, ha észreveszik, hogy
élve maradtam, s csak arra volt gondom, hogy elrejtsem magam a
pillérek közé. Így mikor az ötszázat egy ó-kutba belehányták, én
maradtam legfelül, s éjszaka kiszöktem a sirból.
Most már csak ketten tudtuk, hogy hol vannak az omaridák
kincsei: a kalifa, meg én. A kalifa, mikor az ostromlók betörtek a
városba, harczban elesett. Bizott a bölcsek azon jóslatában, hogy se
kard, se nyil, se dárda vasa őt meg nem ölheti: – réznyillal lőtték
sziven.
Én pedig elmenekültem a hegyek közé.
Minden ezután elkövetett vérengzéseim, gyilkos pusztításaim
egyedüli oka az volt, hogy tudtam, hol vannak az omaridák kincsei.
Azért gyüjtöttem magam körül a vidék minden rablóit, hogy egyszer
rajta törjek Cordován, elfoglaljam a palotát, kibontsam a kincset
rejtő boltozatokat; a mikor aztán a kincsek birtokában az egész mór
nemzetet régi fényében helyreállítom, s hatalmasabb kalifává teszem
magamat, mint te vagy, vagy akár az apád volt a hatalmas
Almansor; a ki maga is hasztalan kutatta egész életén keresztül a
kincseket: rajtuk járt, közöttök élt, de nem birt ráakadni.
Most már vége van. Harámjaimat elpusztítottad, magamat
meglánczoltattál; fejemet karóra ítélted: kénytelen vagyok veled
megalkudni.
Ha te én nekem, királyi szavadra s a próféta szakállára
megesküszöl, hogy életemet és szabadságomat visszaadod, s aztán
minden font arany után, a mit neked átadok, egy uncziát adsz
nekem; – a megkövült csillagért, az Aldebaránért megteszesz Malaga
kormányzójának; – a bölcsek kövéért felcsapsz nemes lovagnak; –
akkor én felfedezem te neked, hogy hová vannak elrejtve az
omaridák kincsei!
Mahadi rögtön futtatá ajtónállóit az alkadéhoz, hogy függeszszék
fel a rabló Toma kivégeztetését. De biz annak a feje már akkor fel is
volt tűzve a karóra: azzal nem beszélhetett többet.

Most jutott már – nagy későn – eszébe haldokló apjának, a bölcs
Almansornak jó tanácsa: «ne hagyj senkit kivégeztetni addig, a míg
kegyelemkérő levelét el nem olvastad.»
(«De te azt azért mégis meg fogod tenni!»)
Most már késő volt bánkódni rajta.
S az omaridák kincséről szóló hagyomány nem volt mese. Mahadi
apjától, a megholt kalifától sokszor hallott felőlük; ki azokat ifjú
korában sokszor látta. A műremekek le voltak irva a krónikákban.
Létezésükről egész odáig szóltak beirott pergamenek, a mikor az
ostromolt kalifa elrejtésüket rábízta egy «Vocair» nevü mór
épitészre; s aztán akkor az épitészt és valamennyi segédét megöleté.
Mahadi annyit tudott meg, hogy azok a kincsek itt vannak az ő
palotájában elrejtve.
Ha itt vannak, akkor megvannak.
A palota sok milliókat ér: de a minek temetőjévé lett, az omár-
kincs huszszor annyit.
Mahadi parancsot adott, hogy bontsák le a királyi palotát.
Bontsák addig, a míg a befalazott kincsekre rá nem találnak.
És azzal elkezdték az építészet remekét szétbontani; a gyönyörű
oszlopsorok, százféle szinű márványból faragva, ledöntettek, a
pompás boltozatok, aranyozott iveikkel, lazurkő burkolataikkal,
drágakő-mozaikok letépettek, halomra hányattak, a tündérpalota
lassankint eltünt a helyéről, már csak a fundamentoma volt meg: azt
is szétbontották; a kincsek még sem kerültek elő. Utoljára a legalsó
pinczebolt alatt, a mint felszakgatták a nehéz kőlapokat, megtalálták
a lépcsőzetet, mely azon üregbe levitt, a hová a kincsek el voltak
rejtve.
Ha Mahadi a harámvezérnek megkegyelmezett volna, egyetlen
oszlopát sem kellett volna ledöntetni a palotájának: az egyenesen

rávezette volna őt arra a rejtett üregre, a mit keresett.
Így pedig Almansor harmadik végrendeletét is megszegte
Mahadi, lebontatva a feje fölül azt a gyönyörű királyi palotát, mely az
Alhambrával vetekedett.
Bizonyos, hogy a rabló azért irta a kegyelemkérő levelében azt a
királynak, hogy a roppant kincsek a palota termeiben vannak
elfalazva, hogy ha őt mégis kivégezteti, holta után boszut álljon
rajta; a boszuja sikerült, a király lerontatta a palotáját.
De legalább rátalált a tündérregei kincsekre.
Ha rátalált.
A mint az első rés volt ütve a rejtett boltozaton hirül adták a
királynak.
Belépni a kincses boltba még nem volt szabad; mert ahogy az
akkori bölcsek állíták, a rejtett kincsek ártó kigőzölgésekkel töltik
meg az ilyen üreget; az arany «materia urens»-t s az ezüst
«selenium»-ot, a drágakő pedig «materia detonans»-t párolog ki
hosszú idő jártával, melyek közül az első elégeti a tüdőt, a második
holdkórosságot idéz elő, a harmadik pedig azt, ki által beszivatik, a
villámütés tárgyának teszi meg. Azonkívül a papok is itt voltak,
felvilágosítani mindenkit, hogy a kincsek fölött ott ül a sárkány, a
baziliszk és a kigyókirályné, melyek közül az első a vaktában
közeledőt elégeti, a másik kővé változtatja, a harmadik pedig pokolra
viszi. Tehát azt az alchymistáknak a rossz gőzöktől «caput
mortuum»-mal és «flos coeli»-vel meg kell tisztítani, azután a
máguszoknak bűvtörő ceremoniákkal a szörnyektől megszabadítani,
csak azután lehet belemenni.
Addig csak az üregbe előre tolt fáklya fényénél lehetett a
királynak megtekinteni a mesés barlang tündér kincseit.
Ez nem volt álom. Ott volt egymásra halmozva az omaridák rég
keresett hires kincstára: a ragyogó érczszekrények, a művészi
remekek, a drága kövektől tündöklő edény, fegyverzet, ékszerhalom,

s a mindannyiból kimagasló arany trón. Egy ezüst szekrénynek a
födele fel is volt nyitva, s az szinültig volt tele sárga pénzhalommal.
Mahadi tánczolni kezdett örömében, a mit király és igazhivő soha
sem szokott tenni. Ime tehát jó volt meg nem tartani Almansor
végtanácsát: «le ne bontasd a palotádat!» A megtett áldozat
sokszorosan meg lett jutalmazva a nyereség által. Most már az
azután következő végtanácsot is meg lehet törni: «ne avasd
államügyekbe az asszonyokat.»
Mahadi sietett a hölgyeihez; legelébb is azoknak mondta el az
örömhirt. Moslim szokás szerint több feleséget tartott, de azok között
most is Zerah volt a kegyencz szultána. A legfiatalabb szultána pedig
volt Alibeh, a fezi szultán leánya, kinek atyját a marokkói császár a
trónjától megfosztotta.
– Most mondjátok meg ti, hölgyeim, mit kezdjek annyi
mérhetetlen kincscsel? szólt a kalifa a szép asszonyokhoz.
Legelőször Zerah felelt:
– Engedd el egy évi adóját a népednek. Építtes menházat
azoknak a számára, a kiknek nincs hajlékuk. Ruházd fel azokat az
árvákat, a kiknek atyái a te atyád harczaiban elestek. Végy
magadnak kincseiden áldást!
– S mit mondasz te? kérdé a kalifa Alibehtől.
– Szereltess fel hajóhadat, rakd meg hadsereggel, köss ki az
afrikai partokon, izenj hadat a maroccoi császárnak. Végy magadnak
kincseiden dicsőséget.
– Te vagy az igazi! kiáltá fel Mahadi, keblére vonva a válasz-adót,
s attól a naptól fogva Alibeh szultána volt a kalifa kegyencz
asszonya, Zerah volt az elmellőzött.
Mahadi rögtön foganatba is vette kegyencznője tanácsát. A
marokkói császár követét mindjárt maga elé hozatá, levágatta a
fejét, azt küldte vissza a császárnak; hadvezéreinek pedig parancsot

adott, hogy rögtön gyüjtsenek nagy hadsereget s roppant hajóhadat
az afrikai hadjárathoz.
Ez alatt a máguszok és alchymisták elkészültek a kincses boltnak
rossz szellemektől és gyilkos kigőzölgésektől megszabadításával, s
akkor aztán a bolt bejáratát egészen felnyittatva, leszállt a király
egész kiséretével a kincsek barlangjába.
Legelső kivánsága volt magát a bölcsek kövét meglátni.
Mindeneknél elébb azt óhajtá megtudni, hogy mi lesz a marokkói
hadjáratának kimenetele? A bölcsek köve a jövendők tudójává
teszen.
A bölcsek ráismertek a kristály szekrényre, a melyben az a kő
tartatott s odahozták azt a király elé, hogy nyissa fel saját kezével.
A kő egészen fekete volt, a minő a próbakő, s egy papyrus-
levélbe volt takarva, mely tele volt jegyezve chaldaei irással. A király
oda nyujtá azt egyik tudósának, s maga az alatt kezébe fogta a
követ.
A tudós pedig azt olvasta le a papyrus levélről.
«Óh te, ki e követ valaha kezedbe fogod venni, tudd meg, hogy
nem a bölcsek kövét találtad meg, hanem a bolondok kövét. Én, Ben
Ali Vocair, Omár khalifa épitésze, a ki az igazi bölcsek kövét bírom:
előre tudom, hogy a rám bizott kincsek elrejtéseért mi jutalom vár
rám? De te nem tudod meg soha, hogy én a kincseket hova
rejtettem el? mert én az igazi omár-kincstárt éjszaka, előre készített
föld alatti rejtek uton át egy sziklabarlangba szállítottam el s a mit
ide, e boltozat alá elrejték, az csak utánzata az igazi kincseknek. A
mi aranyat látsz itt, az csak «similor», az ezüst csak czink, a
gyémánt csak hegykristály, a rubin csak almandin, s a kivert arany és
ezüstpénz csak megaranyozott, ezüstözött ólom. Egész kincstárad
hiúság és nevetség, magad pedig bolond vagy és szerencsétlen s a
mihez kezdesz, el van veszve, mielőtt belefognál.»
Ú

Úgy volt. A mesés kincstár ki volt cserélve. Ez, a mire rátaláltak,
csak csufondáros másolata volt az igazinak.
Mahadi így jött rá, hogy mégis igaza volt Almansornak: «ne
bontsd le a királyi palotádat.» De ő azt azért mégis megtette.
Most azután az is következett még, hogy Almansornak a
negyedik végtanácsában is igaza volt: «ne avasd az asszonyaidat az
ország dolgaiba.» Mivelhogy a bizonyosnak hitt kincsek helyén nem
találva mást, mint hitvány érczet, a marokkói hadjáratra kiindított
hajóhad és hadsereg költségeit nem tudta sehonnan előteremteni s
így elszéledt egész hada, mielőtt ellenséget látott volna, nem maradt
neki belőle egyéb, mint a szégyen.
Ezzel a szégyennel Mahadi még jobban kerülte Zeraht; nem akart
vele találkozni, nehogy annak tekintetéből olvassa azt a
szemrehányást: «lásd! ha az én tanácsomra hallgattál volna, nem ért
volna ez a gyalázat!»
Annál inkább szeretett Mahadi Alibeh karjai közt elrejtőzni a
gúny-kaczagó világ elől. Ez az egy nő nem vethette neki szemére
csufos kudarczát. Ő volt ennek az oka. Annál inkább igyekezett a
gyönyör mámorával, az élvezetek változatosságával eltompítani
Mahadi lelkében azt az újra felsajgó fájdalmat.
Mahadi szerelmi gyöngédsége, édes szenvedélye pedig az által is
nőtt Alibeh szultána iránt, hogy az omár királyi törzset egy új
sarjadékkal ajándékozza meg. Most már féltékenyen őrizte
kegyenczét a kalifa, mintha az volna az országa, a kincse, dicsősége.
Volt rá oka, hogy úgy őrizte, mert az elhagyott asszony mindig
ébren levő veszedelem a boldog vetélytársnéra nézve. Zerah
szultána gyűlöletét nem lehetett hazug mosolygás alá elrejteni.
Mahadi óvta Alibeht Zerah ajándékaitól.
De egytől még sem tudta megóvni s az volt egy gyönyörű
porczellán-cserépben virító narancsfa, melyet Zerah ajándékozott

Alibehnek. Az a narancsfa mikor gyümölcsöt fog érlelni, azokban az
arany almákban halálos mérget fog rejtegetni.
A középkori alchymisták titka volt (nem kár a titokért, hogy
elveszett), hogyan lehet egy egész fának a gyümölcsét úgy
megmérgezni, hogy a ki azt az ágról leszakítja, és megizleli,
menthetetlenül meghaljon tőle? Így ölték meg egy fügével Hunyady
Mátyás királyt, így egy baraczkkal Estrelles Gabriellát.
Mert a növényekben rejlő méreg kitalálhatatlan csodája a
természetnek. A burgonya-növénynek a gyümölcsében méreg van s
a gumója eledel, a spárgának a gyönge hajtása étek, a gyümölcse
veszedelem. A megynek a gyümölcse csemege, a magja ölő méreg.
A legöldöklőbb méregnövény, az upasfa gyilkoló nedve csak a
gyökereiben van meg, a gyümölcse nem árt senkinek. Csak egy
növény van, melynek a gyökerében is, a gyümölcsében is halálos
méreg van. Ez az «atropa mandragora.» A gyökere hasonlít az
emberi alakhoz, kétágú szárával, Theophrastus azt irja róla, hogy ez
a gyökér sír, mint egy ember, mikor a földből kiszakítják, hogy a
sirástól nyavalyássá lesz, a ki meghallja: azért kutyákkal szokták azt
kikapartatni a földből; méreg gyümölcsei pedig hasonlítanak a
szilvához. Hannibal római hadjáratában ezzel mérgezte meg a
borokat, miket az ellenségnek hátrahagyott, s mikor azok elbódulva
hevertek, akkor ütött rajtok. Ennek a mandragorának a gyökerét
tudták a régi bűvészek úgy beleojtani a fa gyökereibe, hogy az
azokkal összeforrt. Az ilyen gyümölcsfával aztán közölte a beleojtott
mandragora gyökér méregtámasztó erejét: narancs, füge, baraczk
olyan halálos mérget szívott magába, a minő a mandragora
gyümölcsében rejlik.
Mikor a Zerahtól ajándékozott narancsfa első arany almáját
megérlelé, a boldog Alibeh leszakította azt az ágról s így szólt
Mahadinak:
– Lásd: én ezt a gyümölcsöt megkívántam, s áldott asszonynak, a
mit megkíván, azt meg is kell ennie, külömben a természet rende

össze-vissza bomlik rá nézve. De én ezt az arany almát mégis
megosztom veled; hogy tudjad, milyen nagy a szerelmem hozzád.
És Mahadinak nem jutott eszébe Almansor utolsó végtanácsa,
«ne egyél olyan gyümölcsből, a mit egy áldott állapotban levő
asszony megkiván.» Elfogadta a megosztott narancsot. S még az
nap Alibehhel együtt halottak lettek – mind a ketten – mind a
hárman.
Így jár az, ki az apja végtanácsát meg nem tartja.

TISZTELT HÁZ, KARZAT ÉS BUFFET.
EGY SZERENCSÉTLEN KÉPVISELŐ ELBESZÉLÉSE UTÁN.
No gyere ide, te Kakas Márton, hadd mondjak el neked egy
siralmas történetet; aztán nem bánom, ha kiirod is az ujságba; –
velem történt.
Azt tudod, hogy én micsoda intrikával kerültem be ide ebbe az
emberaszalóba? Valóságos áldozatképen. Én igen tisztességes
szolgabiró voltam odahaza: nagyon szerettek; nem is tudom, mi
kifogásuk lehetett volna ellenem, mikor soha se voltam otthon, ha
kerestek! Miattam azt csinálhatta minden ember, a mit akart.
Barátom: még a tolvajok se apprehendálhattak rám. Nem volt
senkinek olyan szép gyüjteménye elintézetlen felsőbb rendeletekből,
mint nekem. Hanem hát tudod, hogy fiatal embernek sok az
ellensége; addig ármánykodtak ellenem, míg hirem tudtom nélkül
megválasztottak képviselőnek, hogy eltuszkoljanak az én jó
kényelmes szolgabirói székemből. – Én, mondhatom, hogy még csak
a kis ujjamat se mozditottam meg, hogy képviselő legyek; (a mutató
ujjamat meg a hüvelykujjamat, az igaz, hogy megmozdítottam egy
ötezer forintos váltó aláirására, a választási költségek fedezése
végett; hanem hát ezt a rezón megkivánja). Bizony Isten, a
választási programmom is oly szerény, hogy én abban a
választóimnak nem igértem semmit: csak épen, hogy az adójukat
erre a három esztendőre elengedtetem, a helyett inkább három
vasutat szavazok meg, a minek a választási központ lesz a találkozó
helye; katonát többet nem állítunk, nem is tartunk, hanem azért a
muszkát megverjük; azután a németnek fordulunk, azt is megverjük;
az urbéri pöröket harmincz esztendőre visszaigazítjuk, a nagy
pusztákat felosztjuk, a helyett inkább a kényszeriskolázást törüljük

el; és végtére olyan pénzt találunk fel, a milyennel a régi időben
éltek az ezermesterek, hogy mikor az ember a piaczon egy tallért
kiád, mire hazamegy, már megint a zsebében van a kiadott tallér. –
Egy szóval csupa észszerü dolgokat igértem nekik, barátom.
Hanem hát ismered ezeket a mi nyomoruságos pártviszonyainkat.
Hát lehet itt valamit csinálni? Hiszen láthatod, hogy mi történik a
klubban, mikor valami szőnyegre kerül. A miniszter feláll, beszél
valamit. Erre feláll Zsedényi, elmondja, hogy neki ellenészrevételei
vannak, de nagyon jó néven fogja venni, ha Várady Gábor azokról
lebeszéli, Várady Gábor aztán megnyugtatja. Falk Miksának alapos
észrevételei vannak s keres magának egy süket embert, a kinek azt
a fülébe sughassa. Wahrman azalatt anecdotákat csinál, míg Szilágyi
Dezsőnek a vállára könyökölnek kétfelől, hogy fel ne kelhessen, a
tarokkozni óhajtókat egy hosszú beszéddel megkoplaltatni; végre
Gorove kimondja a határozatot s P. Szatmáry Károly egy akkora
darab papirosra, a mennyi a térdekalácsán elfér, megirja a referádát
az értekezletről. A független szabadelvü pártnak az inasa ez alatt ott
várja az előszobában, hogy mit végeztek a szabadelvü klubban? erre
Bánhidy Béla báró elnöklete alatt elhatározza az értekezlet, hogy
Simonyi Lajos báró tartson a holnapi gyülésben ezzel ellenkező
beszédet. Ugyanez történik a független klubban, meg a jobb-oldali
klubban. Másnap előadja a miniszter a javaslatát. Az elnök előbb
bejelenti, hogy érkezett három szekér petitio a bank, vám és keleti
ügyben; ez a kérvényi bizottságnak tüzelő anyagul deputatumképen
kiutalványoztatik; azután kiosztja a szerepeket: Tombor Iván viszi a
jegyzőkönyvet Molnár Aladár jegyzi a jobboldali, Orbán Balázs a
baloldali, Beöthy Algernon a karzati szónokokat; s arra Gullner Gyula
felveszi a mult ülés jegyzőkönyvét s az általános discursus közepett
leolvassa róla, hogy pernahajder volt az apánk s vigyen el bennünket
az ördög! mi pedig azt hitelesítjük. Akkor az elnök nagy nehezen
csendet harangoz. A miniszter előadja javaslatát. – Bukik vele! –
Előre látható. – Utána feláll Simonyi Lajos, a nadrágszíján egyet ránt
jobbra; a mi annak a jele, hogy ma kegyetlen lesz. Megesik rajta
néha, hogy ahhoz a törvényjavaslathoz szól, a melyik a szőnyegen
van; ritkábban, hogy olyan anecdotákat bocsát közre, a mik

odatartoznak; felszolgáltat nagyszerü gorombaságokat s bámul rajta,
hogy senki se nevet neki; nyájaskodik, s mosolyog rajta, hogy senki
sem akar érte haragudni. Mikor elvégezte: a miniszter kilátásai ötven
perczenttel javultak a törvényjavaslata elfogadtatására. Utána feláll
egy frondeur a miniszterek mögötti padokról: minden ember azt
várja, hogy no most hátulról hogy röpíti levegőbe a veres karszéket
miniszterestül (azok után, a miket a folyosón beszélt), dehogy teszi:
még megdicséri, (mind a kettőt.) Most rágyujt Helfy Náczi: elkezdi
48-nál s folytatja egész a hordár uralomig s addig beszél, míg
minden ellenséget ki nem űz a házból. A kik odabenn maradnak,
végzik azalatt, míg ő beszél, a napi dolgaikat, úgy hogy az elnök azt
sem tudja már, hogy a discurálókat intse-e rendre, vagy a szónokot,
hogy mit zavarja a kedélyes beszélgetést a maga lármázásával? A
fáczitja az a monolognak, hogy Helfy nem ért egyet sem a
miniszterrel, sem báró Simonyival. Azután feláll Kállay Béni,
elmondani, hogy ő nem ért egyet az egész tisztelt házzal; ezután
feláll Apponyi Albert s felvilágosítja a félreértést, mintha ő az előtte
szólottak valamelyikével, (odaértve tisztelt elvtársát, Kállay Bénit is)
egyetértene. Az elnök azt hiszi, hogy már egy szónok sincs fölirva, a
ki elfojthatatlan vágyat érez az egyetnemértését kinyilatkoztatni s ki
akarja tűzni a kérdést. – Ekkor megnyílik a gyászos kriptaajtó, s
miként a Krisztus által föltámasztott Lázár, felmerül a szélsőbali
katakombákból a tisztelt ház bányaréme, Simonyi Ernő. Egy kisértet,
a ki mosolyog. Egyike azoknak a bronzoroszlánoknak, a kiknek a
szájukból szakadatlanul csörgedez elő a szökőkút víz-sugara. Erre a
tisztelt ház menekül minden ajtón kifelé. A képviselők mennek
szanaszét ebédelni. Aztán még vannak olyan fanatikus fakirjai is a
kötelességteljesítésnek, a kik ebéd után megint visszajönnek.
Simonyi Ernő még mindig csorgatja a csapon. Végre a tiszta bor
lefolytával odaönti a seprüjét is a tisztelt majoritás szeme, szája
közé. Erre a majoritás megharagszik, s frondeurök és mamelukok
egyetértőleg megszavazzák a törvényt, a negyvennégyféle tisztelt
egyetnemértők pedig nem szavazzák meg, de kisebbségben
maradnak.
S ez így ment folyvást.

Te persze mindebből nem tudsz semmit; mert azalatt rendesen
correcturát csinálsz a helyeden kussadva. A szegény laikus karzati
nép azt hiszi, hogy a legszorgalmasabban tanulmányozod az előttünk
fekvő törvényjavaslatot.
Hát aztán hogy jusson itt szóhoz az ember?
Barátom! Már a harmadik év is végére jár s én még mindig szűz
vagyok. – Honny soit, qui mal y pense.
Még mindig adós vagyok a szűz-beszédemmel.
A választóim egyre szekíroznak, hogy hát mi lett belőlem, néma
barátnak álltam be? Silentiumot adtak talán, vagy mi? Az «átkos
kaurmány» vesztegetett meg, hogy tétlen hallgatásban fojtom el
legszentebb meggyőződéseimet? Kerüljek csak haza! Hát a vasútból
mi lesz? – Az adót exequálják ám; s ugyan csak osztják a földet; de
nem nekünk, hanem a mienket. Hát már ha annyit sem tudtam
kivívni, hogy a megye székhelye megmaradjon a választó
kerületemben, de legalább Törökország integritásáért csak szót
emelhetnék. Hát az az autonom tarifa mi a fity-fránya lesz? Ha az
valami tüzelő fa, a mit az urasági erdőből lehet kapni, akkor meg ne
buktassam; de ha csak companistája annak a másik két átkozott
fának, Mustafa, Carafa, hát legyen velük a harmadik! Várják
azonban, hogy beszéljek, külömben a jövő választáskor a mostani
szolgabírót fogják képviselőnek elküldeni, mert már azzal is kutyául
meg vannak elégedve.
Ez azonban hát csak mellékes dolog. Hanem a mi fő szempont,
hát a szép «Leonie» is arra capricirozta magát, hogy akar tőlem egy
nagyhatású szónoklatot hallani.
No tudod, az a szép kis leány, a kinek az a nagyszerű három
emeletes háza van a belső körúton. Tehermentes! Már tudniillik a
ház. Ezt meg lehet tudni a telekkönyvből. Már azt, hogy a leány
szívére nincs-e már valami intabulálva? azt nem lehet olyan könnyen
megtudni.

Csak azt nem szeretem, hogy az én kedves collegám és druszám
is odajárt a házhoz. Hiszen ismered? Nem-e? Még megérem, hogy
engem sem ismersz? Mondom! pedig hányszor tarokkoztunk együtt!
Igaz, hogy csak a keresztnevemen hivogatnak: «Czenczi.» S ebből
sok ember azt hiszi, hogy az nekem vezetéknevem, mert olyan nevű
familia is van. Igen mulatságos, mikor a kisasszonyok úgy
szólongatnak, hogy «kérem, Czenczi úr.» A druszám persze
«Vincze.» Egy vármegyéből vagyunk ideküldve. – Pedig hát az csak
egy «kukkó.»
Tudod, hogy mi az a «kukkó?»
Az, a ki nem ismer mást, csak a maga országát; nem szeret
mást, csak a maga feleségét; a ki addig, a míg egy dolgot el nem
végez, addig máshoz nem kezd; a ki azt várja, hogy hijják valahová,
addig oda nem megy; a ki pénzt kölcsön se nem kér, se nem ád;
kártyázni le nem ül; egyszóval: a ki nagy kukkó; – hát az a kukkó.
Hanem már rájöttem, hogy az én druszám csak teszi magát
annak.
Hát nem rég együtt voltunk az én szép imádottamnál. Az épen
egy tárczába való tuffot himezett zseniliával. Egy rózsa- és nefelejts-
koszorúból összefont V betűt ábrázolt az.
– Hej ha én azt tudnám, hogy ki lesz az a boldog Vincze, a kinek
ez a kezdőbetűjét képezi? – flótáztam én epedő hangon.
– Hát ez attól függ, felelé a kis nagyravágyó, hogy melyik fog
önök közül hatásosabb beszédet tartani a legközelebbi ülésszak
alatt?
Láttál ilyet? Ehhez képest Schillernek az a kisasszonya, a ki azt
kivánta a Seladonjától, hogy hozza fel a tigrisek közé leejtett
keztyűjét, még irgalmas nénének járja.
S még a kukkó mondja rá, hogy próbáljuk meg!

Pedig otthon még csak a bizottsági ülésben sem meri felnyitni a
száját.
No itt aztán én sem vagyok az az ember, a ki egy parit el szokott
utasítani. Ugrottam én már le az emeletből is fogadásból.
Áll a fogadás. Tehát a legközelebbi ülésszakban, a mint
elkezdődik «a bank- és vámügygyel kapcsolatos török-orosz
háborúnak» a tárgyalása, feliratjuk magunkat, egyikünk pro, a
másikunk contra.
Ráállt a kukkó; csak azt kötötte ki, hogy a beszédet nem szabad
mással készíttetni.
Van is nekem arra szükségem!
Tudom én, hogy hogyan kell egy országgyülési dikcziót készíteni.
Van a nemzeti casinónak egy gyönyörű szép könyvtára.
Abban van egy igen érdekes könyv, a minek az a czíme, hogy
«catalogus.»
Ha azt magam elé veszem, nincs én nekem szükségem
Csávolszkyra, hogy az készítse el a beszédemet; csak felütöm az
illető rovatokat: «hadászat,» – «nemzetgazdászat,» «pénzügy,» s a
melyik auctornak a neve legjobban megtetszik, azt kiírom
magamnak, utána írom a saját geniális ötleteimet, mintha citatumok
volnának. Ugyan kinek fog eszébe jutni az országgyülésen, hogy
utána keresgéljen, valjon csakugyan mondta-e azt ez meg amaz a
híres ember, avagy nem mondta?
Például ezt:
«A nagy savojai Eugen szerint minden hadviseléshez három
dolog kell; sok beszéd, – még több beszéd, – és igen sok beszéd.»
No vagy ezt:

«A nagy Julius Cæsar de Bello Gallico czímű munkájában azt írja,
hogy mikor az ember egy nagy hadjáratot meg akar indítani, akkor
nagyon jó, ha volt az embernek egy előrelátó apja, a ki annak
idejében az egyezer forintok letételével a fiát megváltotta a
katonáskodás alól.»
De halld meg ezt:
– A dicső Liszt, a ki mikor nem zongorázik, olyankor nemzet-
gazdászattal foglalkozik, azt írja a Nationale system der politischen
oeconomie czímű munkájában, hogy «ha egy garast az egyik
zsebemből a másikba teszek, akkor abból két garas lesz;» –
továbbá:
– A hirhedett Stuart Mill – a hires Stuart Máriának az unokája, azt
írja a «Principies of political oeconomie» czímű remekművében: hogy
ha az emberek négykézláb járnának, az a csizmadiákra nézve
nagyon előnyös volna, mert akkor két pár csizmát szakgatnának egy
helyett.
Hasonlóul:
– A hirhedett két nemzetgazdász, Kasper et Braun az ő
«Fliegende Blätter» czím alatt kiadott reáltudományos folyóiratukban
bebizonyítják, hogy ha két rongyos ember a szennyes ingét
egymással megcseréli, akkor mind a ketten tisztát váltottak.
Azután:
– A nagynevű Verulami Baco, «Novum organon Scientiarum»
czímű zsebkönyvében megmondja, hogy ha két ember nyír egy
birkát, vagy egy ember nyír két birkát, az a birkára nézve
tökéletesen mindegy.
Ezt is hallgasd meg:
– Machiavelli szerint minden monarchiának az alapja az anarchia.
Meg ezt:

– Bell Lancaster iskolája szerint minden deficit megszüntethető az
által, ha az ember megszünik fizetni.
De ez se rosz:
– Strousberg feljegyzései szerint financialis kérdésekben azé a
győzelem, a ki legtöbbet igér s a legkevesebbet ád.
De valamennyinél jobb az:
– A mint ünnepelt pénzköltőnk Rothschild Salamon mondja, hogy
legnehezebb egy milliomosnak az első százezer forintot elkölteni: az
utolsót már nagyon könnyű.
Végül, hogy egy hazai celebritást is idézzek:
– Azt írja gróf Lónyai Menyhért – tisztelt barátom és
képviselőtársam s e téren a legismertebb tekintély és
legszakavatottabb férfiu, a Bankügyről írott munkájában, hogy a
bankot úgy csinálják, mint az ágyut, hogy vesz az ember egy nagy
lyukat, s ezt körülönti ezüsttel, aranynyal; – lyukunk már van.
No már most képzelheted, hogy ha én ezekkel az ágyúkkal
kirukkolok, hát micsoda strages lesz ott a képviselőházban?
Két hétig mindenki arról beszélt a casinóban: «hol van Czenczi?»
«Odafenn van a könyvtárban, a speechét csinálja.»
Mikor feliratkoztunk a jegyzőknél, úgy kalkuláltam ki a dolgot,
hogy másodiknak következzem a kukkó után: neki is adhassak egy
pár oldaldöfést.
– Hát te készen vagy-e már a dikczióddal? kérdém tőle.
– Dehogy vagyok.
– Hát mit fogsz csinálni? Csak úgy a hasadból beszélsz?
– De majd úgy teszek, mint Bibliás Kis János: felviszem
magammal az ó-testamentumot s majd abból felolvasok egy pár

caputot.
Az én beszédem már akkor a zsebemben volt. Be is volt tanulva
jól.
Az ujságírókkal mind végigveszekedtem; mert a nevemet nem jól
nyomatták a szónoklatok lajstromában, azt mind helyre kellett nekik
igazítani.
A választóimnak táviratoztam, hogy küldjenek egy deputatiót ide
fel. A költséget én viselem.
Mikor elkezdődött a nagy vita, alig fértem a bőrömbe. Szidtam
azt a sok hosszú unalmas szónokot, mind elpusztítják a közönség
türelmét, mire rám kerül a sor, semmi sem marad belőle.
Végre ki lehetett számítani azt a napot, melyen rám kerül a sor.
Előttevaló nap Alter és Kissnél egész értekezletet tartottunk, hogy
milyen színű nyakravaló fog legjobban illeni a holnapi
debütirozásomhoz?
– Attól függ, hogy milyen színű lesz a szónoklat!
– A szónoklat veres, még pedig veres-sipka-veres.
– Akkor a hozzávaló nyakravaló: «Bismarck en-colère!»
Kónyi Manóval is barátságot kötöttem.
– Kérem gyorsírótábornok úr, csak jól figyeljenek a beszédemre!
El ne szalaszszanak egy közbeeső éljenzést, helyeslést is.
– Nem tehetné meg képviszelő úr azt a barátszágot, hogy adná
ide a szónoklatát irászban? Monda a tisztelt honatyamegörökítő,
ismeretes csángó dialectusában.
– Mit gondol barátom! Én egy betűt sem írok le előre abból, a mit
mondani akarok; csak úgy ex tripode beszélek, invita Minerva; a mi
épen eszméim tárházából kapóra jön! Van is nekem írott dikczióm!

– No majd gondoszan lemászoljuk.
– Lemázolnak!
Ominosus csángókiejtés!
Egy igen nagy hibája van az országháznak. Hogy nincs benne
egy olyan csendes, nyugalmas hely; olyan «Asyl für Herrn, die für
einige Zeit in stiller Zurückgezogenheit zu leben wünschen,» a hol a
szónoklataik agybeli terhét viselő képviselők addig, a míg rájuk kerül
a sor, csendes ruminálásban tölthetnék el a hosszú negyedórákat.
Nem képzeled, hogy micsoda tortura az! Odahaza szépen elkészül az
ember a beszédével: úgy tudja, mint a karikacsapás; szépen
összerakja, mint a kártyavárat, a mit minden szélfujástól őrizni kell,
hogy össze ne dőljön. Aztán mikor felmegy az ember a házba a
halálos itéletet váró rabnak lelkifurdalásával, s szemközt találkozik a
jegyzővel, a ki a szónokokat felhívja, abban a perczben úgy tetszik
az embernek, mintha egy betűt se tudna többé abból, a mit el akar
mondani.
Ha bemegy a tisztelt házba, ott egy másik szónok beszél egy
másik speechet, a míg az ember a magáét forgatja a fejében, ezét
meg hallgatja a fülével, olyan valami kedveset érez, mint mikor
odakün a zöldben egyszerre két kintornán hall két külömböző nótát
muzsikálni.
Hova meneküljön? ki a folyosóra? A jobb oldali folyosón van
három rekesz. Az elsőt teleülik vidéki deputatiók, a kik megfogják az
embert, s azt akarják, hogy legyen szószólójuk a miniszternél nem
tudom én miféle vízigátnak, vagy törvényszéknek a reperálása
végett. Itt nem lehet maradni. Az üveges rekeszben ott conspirálnak
a frondeurök a szélbalokkal és végjobbokkal; ott csinálják a kisebb
országgyülést, ott leskelődnek az intimusok, a kőnyomatú
lapszerkesztők az érkező miniszterekre, hogy a ki boldogabb,
elfoghassa őket. Ott nem lehet megmaradni. A harmadik rekeszték a
vidékről feljött főispánoké, kik a vármegye baját hozták fel, a
képviselőházon kivüli celebritásoké, a főhivatalnokoké, a kiket

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!
ebookultra.com

Advances in Clinical Chemistry 1st Edition Gregory S. Makowski (Eds.)

About This Presentation

Slide Content

Tags

Categories

Download

Quick Actions

Statistics

Related Slideshows

Advances in Clinical Chemistry 1st Edition Gregory S. Makowski (Eds.)

About This Presentation

Slide Content

Slide 1

Slide 2

Slide 3

Slide 5

Slide 7

Slide 8

Slide 9

Slide 10

Slide 11

Slide 12

Slide 13

Slide 14

Slide 15

Slide 16

Slide 17

Slide 18

Slide 19

Slide 20

Slide 21

Slide 22

Slide 23

Slide 24

Slide 25

Slide 26

Slide 27

Slide 28

Slide 29

Slide 30

Slide 31

Slide 32

Slide 33

Slide 34

Slide 35

Slide 36

Slide 37

Slide 38

Slide 39

Slide 40

Slide 41

Slide 42

Slide 43

Slide 44

Slide 45

Slide 46

Slide 47

Slide 48

Slide 49

Slide 50

Slide 51

Slide 52

Slide 53

Slide 54

Slide 55

Slide 56

Slide 57

Slide 58

Slide 59

Slide 60

Slide 61

Slide 62

Slide 63

Slide 64

Slide 65

Slide 66

Slide 67

Slide 68

Slide 69

Slide 70

Slide 71

Slide 72

Slide 73

Slide 74

Tags

Categories

Download

Quick Actions

Statistics