Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
-
Upload
igneel-creed -
Category
Documents
-
view
236 -
download
1
Transcript of Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
1/627
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
2/627
nical Chemistry
venth Edition
lliam J Marshall, MA PhD MSc MB BS FRCP FRCPath FRCPEdin FSB FRSC
nical Director of Pathology, The London Clinic, London, UK
meritus Reader in Clinical Biochemistry, King’s College London, London, UK
phen K Bangert, MA MB BChir MSc MBA FRCPath
nsultant Chemical Pathologist, East Sussex Healthcare NHS Trust, Eastbourne, UK
arta Lapsley, MB BCh BAO MD FRCPath
nsultant Chemical Pathologist, Epsom and St Helier University Hospitals NHS Trust, Epsom,
osby Ltd.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
3/627
pyright
2012 Elsevier Ltd All rights reserved.
part of this publication may be reproduced or transmitted in any form or by any means, elect
mechanical, including photocopying, recording, or any information storage and retrieval sy
thout permission in writing from the publisher. Details on how to seek permission, fuormation about the Publisher’s permissions policies and our arrangements with organizations
the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our web
ww.elsevier.com/permissions.
is book and the individual contributions contained in it are protected under copyright b
blisher (other than as may be noted herein).
st published 1988 by Gower Medical Publishing
cond edition 1992 by Gower Medical Publishing
ird edition 1995 by Mosby
urth edition 2000 by Harcourt Publishers Limited
fth edition 2004
xth edition 2008
venth edition 2012
BN 9780723437031
ernational ISBN 9780723437048
itish Library Cataloguing in Publication Data
catalogue record for this book is available from the British Library
brary of Congress Cataloging in Publication Data
catalog record for this book is available from the Library of Congress
tices
owledge and best practice in this field are constantly changing. As new research and exper
oaden our understanding, changes in research methods, professional practices, or medical trea
http://www.elsevier.com/permissions
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
4/627
y become necessary.
actitioners and researchers must always rely on their own experience and knowledge in evalu
d using any information, methods, compounds, or experiments described herein. In using
ormation or methods they should be mindful of their own safety and the safety of others, incl
rties for whom they have a professional responsibility.
th respect to any drug or pharmaceutical products identified, readers are advised to check the
rrent information provided (i) on procedures featured or (ii) by the manufacturer of each produadministered, to verify the recommended dose or formula, the method and duratio
ministration, and contraindications. It is the responsibility of practitioners, relying on their
perience and knowledge of their patients, to make diagnoses, to determine dosages and the
atment for each individual patient, and to take all appropriate safety precautions.
the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, as
y liability for any injury and/or damage to persons or property as a matter of products liab
gligence or otherwise, or from any use or operation of any methods, products, instructions, or
ntained in the material herein.
nted in China
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
5/627
eface to the seventh edition
spite the wealth of information now available to students and practitioners of healthcare o
ernet, printed books still provide a valuable learning and reference resource, and, judging by
sixth edition of this book proved as popular as its predecessors: we hope that this seventh ed
ll be as well received.
e book was originally written primarily for medical students, but also proved popular with do
dying for postgraduate examinations and students and practitioners of clinical and biomeence. Each of these groups has differing requirements, and we have aimed to satisfy all these i
proach to the subject. In doing this, we have been helped by comments received from re
ound the world. Please continue to let us know where you think that we could further improv
ok.
ch chapter includes a summary of the basic biochemistry and physiology upon which understa
nical biochemistry depends. The nature, choice, use and limitations of laboratory investiga
urally comprises the bulk of each chapter, but clinical biochemistry is only one part of labor
dicine, and laboratory tests comprise only one group among the many types of investigailable to support diagnosis and management. We have therefore continued to outline the ro
her investigations, for example, imaging, and compare the type of information that they provide
t from laboratory tests. And because clinical biochemistry tests are widely used in asse
ients’ responses to treatment, we also provide summaries of treatment options, although we s
t this book is not, and is not intended to be, a textbook of metabolic medicine.
e Case histories, all drawn from the authors’ own experience, summarize the key points in
apter and may provide a useful starting point for examination revision. What we learn from
ients is often better remembered than what we learn from books.
e are aware that the book has a considerable overseas readership, greatly increased by
blication of the sixth edition in an international as well as standard edition. The publicati
idelines and recommendations for managing patients with particular conditions has been a fe
clinical practice in recent years. However, these are often country-specific, and we
phasized where such material has been developed for use in the UK.
e two original authors were delighted when Dr Marta Lapsley accepted their invitation to join
preparing this edition. Now that the senior author is semi-retired, this will provide continuit
future. We have enjoyed working together and learning from each other’s views and experien
ere have been no major changes in this edition, but the whole text has been carefully checked
here necessary, revised. In addition to reflecting advances in clinical biochemistry, we
minated material that we agreed was obsolete. One of the three of us took responsibility fo
ailed revision of each chapter and another then reviewed this material, but all us of have che
d agreed the final text. We hope that this approach has not only ensured the authority o
ormation, but has maintained the uniformity of the overall style.
Elsevier, Timothy Horne, our commissioning editor for several previous editions, has now re
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
6/627
d Jeremy Bowes has taken over from him. We have enjoyed working with Jeremy and with C
cMurray, the development editor and Anne Collett, the project manager. As always, we are gra
the designers, whose work results in such an attractive appearance of the book, complementin
t detracting from the text, and to the rest of the in-house team.
d at home, Wendy (Marshall), Lorraine (Bangert) and Michael (Lapsley) have been unstinti
ir support during our work on this book; we thank them all for their encouragement and for
bearance throughout.
lliam Marshall
phen Bangert
arta Lapsley
12
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
7/627
rther reading
ed references quickly become outdated. Readers seeking the most up-to-date information
pic are recommended to use one of the bibliographic databases specializing in medical
entific journals, for example Medline (the database of the National Library of Medicine i
ited States of America, which encompasses over nine million references to reviews and p
blished in nearly 4000 journals).
urnals that publish articles and reviews relating to clinical chemistry include Annals of Cliochemistry and Clinical Chemistry. Each issue of Endocrine and Metabolism Clinics of N
merica comprises sets of reviews on related topics, most of which are of direct relevan
nical chemistry. General medical journals such as the British Medical Journal, Lancet and
gland Journal of Medicine carry editorials and reviews of topics related to clinical chemistry
me to time. The monthly issues of Medicine together comprise a textbook of medicine, wh
dated on a three-year cycle and is highly recommended.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
8/627
ble of Contents
tructions for online access
pyright
eface to the seventh edition
rther reading
apter 1: Biochemical investigations in clinical medicine
apter 2: Water, sodium and potassiumapter 3: Hydrogen ion homoeostasis and blood gases
apter 4: The kidneys
apter 5: The liver
apter 6: The gastrointestinal tract
apter 7: The hypothalamus and the pituitary gland
apter 8: The adrenal glands
apter 9: The thyroid gland
apter 10: The gonads
apter 11: Disorders of carbohydrate metabolismapter 12: Calcium, phosphate and magnesium
apter 13: Plasma proteins and enzymes
apter 14: Lipids, lipoproteins and cardiovascular disease
apter 15: The locomotor and nervous systems
apter 16: Inherited metabolic diseases
apter 17: Disorders of haemoproteins, porphyrins and iron
apter 18: Metabolic aspects of malignant disease
apter 19: Therapeutic drug monitoring and chemical aspects of toxicology
apter 20: Clinical nutritionapter 21: Clinical chemistry at the extremes of age
pendix Adult reference ranges
dex
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
9/627
apter 1
ochemical investigations in clinical medicine
ntroduction
central function of the chemical pathology or clinical chemistry laboratory is to pr
ochemical information for the management of patients. Such information will be of value only i
curate and relevant, and if its significance is appreciated by the clinician so that it can be
propriately to guide clinical decision-making. This chapter examines how biochemical dat
quired and how they should be used.
Use of biochemical investigations
ochemical investigations are used extensively in medicine, both in relation to diseases that ha
vious metabolic basis (e.g. diabetes mellitus, hypothyroidism) and those in which bioche
anges are a consequence of the disease (e.g. kidney failure, malabsorption). The principal usochemical investigations are for diagnosis, prognosis, monitoring and screening (Fig. 1.1).
gure 1.1 The principal functions of biochemical tests.
iagnosis
edical diagnosis is based on the patient’s history, if available, the clinical signs founamination, the results of investigations and sometimes, retrospectively, on the respon
atment. Frequently, a confident diagnosis can be made on the basis of the history combined wi
dings on examination. Failing this, it is usually possible to formulate a differential diagnos
ect a short list of possible diagnoses. Biochemical and other investigations can then be us
tinguish between them.
vestigations may be selected to help either confirm or refute a diagnosis, and it is important th
nician appreciates how useful the chosen investigations are for these purposes. Maki
gnosis, even if incomplete, such as a diagnosis of hypoglycaemia without knowing its causeow treatment to be initiated.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
10/627
rognosis
vestigations used primarily for diagnosis may also provide prognostic information, while othe
ed specifically for this purpose. For example, serial measurements of plasma creat
ncentration in progressive kidney disease are used to indicate when dialysis may be requ
vestigations can also indicate the risk of developing a particular condition. For example, the ri
ronary artery disease increases with increasing plasma cholesterol concentration. However,
ks are calculated from epidemiological data and cannot give a precise prediction for a part
dividual.
Monitoring
major use of biochemical investigations is to follow the course of an illness and to monito
ects of treatment. To do this, there must be a suitable analyte, for instance glycated haemoglob
ients with diabetes mellitus. Biochemical investigations can also be used to detect complica
treatment, such as hypokalaemia during treatment with diuretics, and are extensively used to s
possible drug toxicity, particularly in trials, but also in some cases when a drug is in establ
e.
creening
ochemical investigations are widely used to determine whether a condition is present subclini
e best-known example is the mass screening of all newborn babies for phenylketonuria (P
ngenital hypothyroidism and some other conditions that is carried out in many countries, incl
UK and the USA. This is an example of population screening: other types include sele
reening (e.g. of older people for carcinoma of colon using the detection of faecal occult bldividual screening (e.g. as part of a ‘health check-up’) and opportunistic screening (e.g
percholesterolaemia in people found to have hypertension). The use of ‘biochemical pro
mbinations of biochemical tests performed on automated analysers, is discussed later in
apter.
pecimen collection
he test request
e specimen for analysis must be collected and transported to the laboratory according
ecified procedure if the data are to be of clinical value. This procedure begins with the clin
king a test request, either on paper or, increasingly, electronically. The completed request sh
lude:
atient’s name, sex and date of birth
ospital or other identification number
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
11/627
ward/clinic/address
ame of requesting doctor (telephone/pager number for urgent requests)
linical diagnosis/problem
est(s) requested
ype of specimen
ate and time of sampling
elevant treatment (e.g. drugs).
e provision of sufficient information reliably to identify the patient is self-evidently essentia
omission of any of the above may either cause delay in analysis and reporting or ma
possible to interpret the results. Many laboratories publish a minimum data set without which
ll refuse to analyse samples.
levant clinical information and details of treatment, especially with drugs, are necessary to
oratory staff to assess the results in their clinical context. Drugs may interfere with anal
thods in vitro or may cause changes in vivo that suggest a pathological process; for instance,
ychotropic drugs increase plasma prolactin concentration.
l laboratories should publish user guides, preferably available online. These should pr
ormation including the test repertoire, specimen requirements (see below), turnaround
otocols for dynamic function testing and local or national guidelines for the investigati
onitoring of particular conditions, together with contact details for making enquiries toratory.
he patient
me analytes are affected by variables such as posture, time of day, etc., and it may be necessa
ndardize the conditions under which the specimen is obtained. Factors of importance in
pect are listed in Figure 1.2 and are discussed further in subsequent chapters.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
12/627
gure 1.2 Examples of important factors that influence biochemical variables; these and other
cussed elsewhere in this book.
en when standardized conditions are used for sampling, the results of repeated quantitative
g. daily measurements of fasting blood glucose concentration) will themselves show a Gau
tribution, clustering about the ‘usual’ value for the individual. Typically, the scatter, which ca
essed by determining the standard deviation (SD), is less for analytes subject to strict regu
g. fasting blood glucose and plasma calcium concentrations) than for others (e.g. plasma en
ivities). This biological variation can be expressed as the coefficient of variation (CVpeated tests, where CV = SD × 100/mean value.
he specimen
e specimen provided must be appropriate for the test requested. Most biochemical analyse
de on serum or plasma, but occasionally whole blood is required (e.g. for ‘blood gases’)
alyses of urine, cerebrospinal fluid, pleural fluid, etc., can also be valuable. For most analys
um or plasma either fluid is acceptable but in some instances it is of critical importance whi
se is used; for example, serum is required for protein electrophoresis and plasma for measurerenin activity. Haemolysis must be avoided when blood is drawn and, if the patient is rece
ravenous therapy, blood must be drawn from a remote site (e.g. the opposite arm) to
ntamination. Haemolysis causes increases in plasma potassium and phosphate concentration
partate aminotransferase activity, owing to leakage from red cells. If haemolysis is a consequ
a delay in centrifugation to separate blood cells from plasma, glucose concentration can fall. O
alytes may also be affected by haemolysis, depending on the analytical method used. The labor
ould always draw attention to potentially spurious results. It should be noted that leakage from
vitro can cause increases in plasma potassium and phosphate concentrations even in the absen
vious haemolysis, particularly in patients with high white blood cell or platelet counts.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
13/627
llecting a blood specimen into the wrong container can lead to (usually obviously) erron
ults (Case history 1.1): citrate and EDTA, which are used as anticoagulants in containers use
me haematological tests, combine with calcium and cause low measured concentrations i
sma; so does oxalate (the anticoagulant in containers for blood glucose measurement, which
ntain fluoride to inhibit glycolysis), and it is clearly inappropriate to collect blood for li
asurement into a container with lithium heparin as an anticoagulant. Laboratory user guides s
ovide clear guidance on the types of specimen, and, where appropriate, the sampling condition
laboratory tests. This should include guidance on the sequence in which individual specimen
filled to avoid any possibility of contamination; for example, blood should be collectedain tubes’ (not containing an anticoagulant or other additive) before being collected into a
ntaining, for example, EDTA.
Case history 1.1
e laboratory staff were concerned when a serum specimen from an outpatient due to atten
betic clinic was analysed and the following results were found:
vestigations
erum: potassium 12.2 mmol/L
dium 140 mmol/L
eatinine 84 µmol/L
lcium 0.34 mmol/L
hosphate 1.22 mmol/L
omment
e potassium and calcium concentrations are not compatible with life. Investigation disclosed
um phlebotomist who had taken the blood had collected the original specimen into a
ntaining (potassium) fluoride and oxalate, the correct container for an accurate blood glu
asurement, but had then concealed his error by transferring the sample to a plain tube. Oxalat
an anticoagulant by binding to calcium ions (cofactors in several of the reactions in the cl
scade) to form insoluble calcium oxalate.
l specimens must be correctly labelled and transported to the laboratory without delay. T
ould be a written protocol for discarding incorrectly collected or labelled specimens. For tes
um or plasma, the fluid is then separated from blood cells by centrifugation and then anal
hen analysis is delayed, or when specimens are sent to distant laboratories for ana
gradation of labile analytes must be prevented by refrigerating or freezing the serum or plasma
ual care is needed with the collection and transportation of other specimens, such as urin
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
14/627
ebrospinal fluid. All specimens should be regarded as potentially infectious and handled
propriate precautions.
rgent requests
though laboratories should endeavour to generate results as quickly as possible, some request
urgent in that their results may have an immediate bearing on the management of the pa
amples include the measurement of serum paracetamol concentration in a patient who has ta
ug overdose, measurement of serum troponin concentration in a patient with chest painasurement of serum potassium concentration in a patient with acute kidney injury (renal fai
ecial provision must be made for such samples to be ‘fast-tracked’ through the analytical pro
eit in full accordance with procedures to ensure quality, and the results reported to the reque
nician as soon as they have been validated.
epeat requesting
hen biochemical investigations are being used to monitor the progress of a patient’s cond
ial analyses will be required, and the question arises how frequently these should be perfor
is will depend on both physiological and pathological factors. For example, in patients
ated with thyroxine for hypothyroidism, it can take several weeks for the plasma concentrati
yroid stimulating hormone (TSH) to stabilize at a new value after a change in the dose of thyro
peating thyroid function tests in a patient whose dose of thyroxine has been changed at an inter
month may therefore provide misleading information, and could prompt a doctor who i
gnizant with the rate of response to make a further change of dose prematurely. In contrast, pl
ucose and potassium concentrations can change very rapidly in patients being treated for dia
oacidosis, and it may be appropriate to make measurements as frequently as every 1–2 h, at
tially. Laboratory user guides may include guidance on repeat testing, based on local
ionally agreed protocols.
ample analysis and reporting of results
nalysis
e ideal analytical method is accurate, precise, sensitive and specific. It gives a correct curate: Fig. 1.3) that is the same if repeated (precision: Fig. 1.3). It measures low concentra
the analyte (sensitive) and is not subject to interference by other substances (specific). In add
should preferably be cheap, simple and quick to perform. In practice, no test is ideal, bu
hologist must ensure that the results are sufficiently reliable to be clinically useful. Labor
ff make considerable efforts to achieve this, and analytical methods are subject to rigorous q
ntrol and quality assurance procedures.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
15/627
gure 1.3 Precision and accuracy of biochemical tests. Both graphs show the distribution of r
repeated analysis of the same sample by different methods. Precision: the mean value is the
each case, but the scatter about the mean is less in method A than in method B. Method
refore, more precise. Accuracy: both are equally precise, but in method D, the mean value d
m the true value. The mean for method C is equal to the true value. Both methods are eq
ecise, but method C is more accurate.
vertheless, there will always be a potential for some degree of imprecision or analytical vari
a result. The extent of this can be assessed by making repeated analyses (using exactly the
thod) on the same sample (cf. biological variation, above). The results will cluster about a
which the SD can be calculated. The imprecision of the analysis can be expressed as the
here CV = SD × 100/mean result. As will be discussed later in this chapter, an understanding o
ncepts of both analytical and biological variation is essential to the informed interpretati
oratory data.
s important to appreciate that results obtained using different methods may not be interchange
hen a comparison between two results is being made for clinical purposes, the same anal
thod should be used on both occasions.
s often appropriate to perform a group of related tests on a specimen. For example, plasma ca
d phosphate concentrations and alkaline phosphatase activity all provide information that m
eful in the diagnosis of bone disease; several liver ‘function’ tests may usefully be gro
gether. Such groupings are sometimes referred to as ‘biochemical profiles’. Many currailable analysers can perform numerous assays simultaneously on a single specimen. How
hough it may be tempting to perform all the assays on every specimen, this approach generat
ormous amount of information, some of which may be unwanted, ignored or misinterpreted (e
vated creatine kinase (CK) activity in someone who has recently undertaken severe exercise
nstrued as evidence of myocardial damage). Worst of all, it may actually divert the clinic
ention from important results. Discrete analysis, that is, performing only the necessary
quired to answer the clinical question (e.g. ‘Is this patient’s jaundice cholestatic or du
patocellular disease?’), is to be preferred.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
16/627
eporting results
ce analysis has been completed and the necessary quality control checks made and found
isfactory, a report can be issued. Cumulative reports, which show previous as well as cu
ults, allow trends in data to be picked out at a glance. It may be appropriate to add a commen
port to assist the clinician with its interpretation. Results that indicate a need for rapid cl
ervention should be communicated to the requesting clinician as a matter of urgency.
oint of care testing
t all analyses need to be performed in a central laboratory. Reagent sticks for testing urine
dside or in the clinic have long been available. Various substances, including glucose, pro
irubin, ketones and nitrites (indicative of urinary tract infection), can be tested for using
cks.
sting of blood for analytes, such as glucose, and hydrogen ion and ‘blood gases’ at point of
s also been available for some time. Indeed, the availability of easily used instruments to me
ucose allows patients with diabetes to monitor their blood glucose concentrations at home. In r
ars, manufacturers have developed instruments that can perform a wide range of tests suitab
e at the point of care. Such instruments allow the more rapid provision of analytical resul
ients in whom they are required urgently (e.g. in intensive therapy units) but may also be use
nvenience (e.g. in doctors’ offices (surgeries)). It is clearly desirable that such instruments sh
capable of providing results that are as robust with regard to accuracy and precision as
ovided by the main laboratory. These instruments are designed to be very simple to operate bu
vertheless essential that individuals using them, who will usually not be laboratory staf
operly trained in their use. They should adhere to protocols designed to ensure the quality of r
d to provide a robust audit trail so that, for example, should a manufacturer report a problem w
rticular test, patients whose results may have been affected can be identified. Both the trainin
ality issues should be supervised from the laboratory.
me analyses can be performed outside traditional healthcare settings and the results given dir
patients. An example is the measurement of plasma cholesterol concentration in retail pharm
ch analyses should be subject to appropriate quality assurance procedures and trained pers
ould be available to advise patients on the significance of the results.
ources of error
roneous results are at best a nuisance; at worst, they have potential for causing considerable h
rors can be minimized by scrupulous adherence to robust, agreed protocols at every stage o
ting process: this means a lot more than ensuring that the analysis is performed correctly. E
n occur at various stages in the process:
pre-analytical, occurring outside the laboratory (e.g. the wrong specimen being colle
slabelling, incorrect preservation, etc.)
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
17/627
nalytical, occurring within the laboratory (e.g. human or instrumental error)
ost-analytical, whereby a correct result is generated but is incorrectly recorded in the pat
ord (e.g. because of a transcription error).
alytical errors can be systematic (also known as bias: different analytical methods may pro
ults that are higher or lower (it is to be hoped only slightly so) than the definitive or refe
thod) or random. Many of the few errors that do occur even in good laboratories are detect
ality control procedures, including data-handling software or personal scrutiny of repororatory staff. Some are so bizarre that they are easily recognized for what they are. More s
es are more likely to go undetected. Unfortunately, the risk of errors occurring can never be en
minated.
nterpretation of results
hen the result of a biochemical test is obtained, the following points must be taken
nsideration:
s it normal?
s it significantly different from any previous results?
s it consistent with the clinical findings?
it normal?
e use of the word ‘normal’ is fraught with difficulty. Statistically, it refers to a distributio
ues from repeated measurement of the same quantity and is described by the bell-shaped Gau
rve (Fig. 1.4). Many biological variables show a Gaussian distribution: the majority of indivi
thin a population will have a value approximating to the mean for the whole, and the frequ
th which any value occurs decreases with increasing distance from the mean.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
18/627
gure 1.4 Gaussian distribution. The range of the mean ± 2 standard deviations (SDs) encomp
5% of the total number of test results. The range of the mean ± 3 SDs encompasses 99.7% o
al number.
r some analytes, the distribution of values is skewed; an example is plasma bilirubin concentr
ch data can often be mathematically transformed to a normal distribution: data distributed w
ew to the right of the mean (as is the case with bilirubin) can often be transformed to a no
tribution if re-plotted on a semi-logarithmic scale.
the variable being measured has a normal (Gaussian) distribution in a population, statistical t
edicts that approximately 95% of the values in the population will lie within the range given b
an ± 2 SDs (Fig. 1.4); of the remaining 5%, half the values will be higher and half will be l
n the limits of this range.
hen establishing the range of values for a particular variable in healthy people, it is convention
st examine a representative sample of sufficient size to determine whether or not the values fal
ussian distribution. The range (mean ± 2 SDs) can then be calculated; this, in statistical term
‘normal range’. Several important points arise from this:
Although it is assumed that the population is healthy, values from 5% of individuals by definitio
tside the normal range. This suggests that, if the measurements were to be made in a gro
mparable individuals, 1 in 20 would have a value outside this range.
he specialized statistical use of the word ‘normal’ does not equate with what is generally mea
word, that is, ‘habitual’ or ‘usually encountered’.
The statistical ‘normal’ may not be related to another common use of the word, which is to i
edom from risk. For example, there is an association between increased risk of coronary
ease and plasma cholesterol concentrations even within the normal range as derived
asurements on apparently healthy men.
us, the normal range for an analyte, defined and calculated as described, has severe limitatio
ly identifies the range of values that can be expected to occur most often in individuals wh
mparable with those in the population for whom the range was derived. It is not necessarily no
terms of being ‘ideal’, nor is it associated with no risk of having or developing dis
rthermore, by definition it will exclude values from some healthy individuals. In all cases
ust be compared with like. When physiological factors affect the concentration of an analyte
g. 1.2), an individual’s result must be assessed by comparing it with the value expected
mparable healthy people. It may, therefore, be necessary to establish normal ranges for subs
population, such as various age groups, or males or females only.
alleviate the problems associated with the use of the word ‘normal’, the term reference int
I) (often called the ‘reference range’) has been widely adopted by laboratory staff, using num
ues (reference limits) generally based on the mean ± 2 SDs. Results can be compared with t
thout assumptions being made about the meaning of ‘normal’. In practice, the term ‘normal r
still in general use outside laboratories. It is used synonymously with ‘reference interval’ in
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
19/627
ok. Reference intervals for some common analytes are given in the Appendix: these are as us
e of the authors’ laboratories, and are appropriate for the case histories, but may not apply to
oratories because of differences in analytical methods and in the characteristics of the popu
which the data are based. Differences between reference ranges are a particular problem
munoassays, since different antibodies may vary in their specificity for the analyte and the ext
hich they exhibit cross-reactivity with other, similar molecular species. Nevertheless, effort
ng made in the UK to introduce common standards in various areas of pathology, including un
erence ranges.
using RIs to assess the significance of a particular result, the individual is being compared w
pulation. Some analytes show considerable biological variation, but the combined analytica
ological variations will usually be less for an individual than for a population. For exa
hough the reference interval for plasma creatinine concentration is 60–120 µmol/L, the day-t
riation in an individual is much less than this. Thus, it is possible for a test result to be abno
an individual, yet still be within the accepted ‘normal range’.
n abnormal result does not always indicate the presence of a pathological process, n
rmal result its absence. However, the more abnormal a result, that is, the greater its diffem the limits of the reference interval, the greater is the probability that it is related
hological process.
practice, there is rarely an absolute demarcation between normal values and those seen in dis
uivocal results must be investigated further. If an important decision in the management of a pa
to be based upon a single result, it is vital that the cut-off point, or ‘decision level’, is chos
sure that the test functions efficiently. In screening for PKU, for example, the blood concentrati
enylalanine selected to indicate a positive result must include all infants with the condition; in
rds, there must be no false negatives. Because there is some overlap in the values seen iesence and absence of PKU, this inevitably means that some normal children will test po
lse positives) and will be subjected to further investigation. Generally, it is unusual to ha
ermine a patient’s management on the basis of one result alone.
has been explained that 5% of healthy people will, by definition, have a value for a given var
t is outside the reference interval. If a second and independent variable is measured
obability that this result will be ‘abnormal’ is also 0.05 (5%). However, the abnormal results
t arise in the same individuals and the overall probability of an abnormal result from at leas
t will be >5%. It follows that the more tests that are performed on an individual, the greateobability that the result of one of them will be abnormal: for 10 independent variables
obability is 0.4; in other words, at least one abnormal result would be expected in 40% of h
ople. For 20 variables, the probability is 0.64.
though biochemical parameters are frequently, to some extent, interdependent (e.g. albumin
al protein), the use of multichannel analysers to produce ‘biochemical profiles’ inevitably
nerating a number of spuriously ‘abnormal’ results. Before any decision can be made on the
such results, some information is required about the probability that they are indicative
hological process. This topic is discussed on p. 8.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
20/627
it different?
the result of a previous test is available, the clinician will be able to compare the result
cide whether any difference between them is significant. This will depend on the precision o
ay itself (a measure of its reproducibility) and the natural biological variation. Some examp
riation in common analytes are given in Figure 1.5.
gure 1.5 Analytical and biological variation. Analytical variation: typical standard deviation
peated measurements made using a multichannel analyser on a single quality control serum
ncentrations in the normal range. Biological variation: means of standard deviations for rep
asurements made at weekly intervals in a group of healthy subjects over a period of ten w
rrected for analytical variation.
e probability that the difference between two results is analytically significant at a lev
=
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
21/627
here SDA and SDB are the SDs for the analytical and biological variation, respectively. I
ference between two test results exceeds 2.8 times the SD of the test, the difference ca
garded as of potential clinical significance: the probability of this difference being a resu
alytical and biological variation is
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
22/627
using the result of a test, it is important to know how reliable the test is and how suitable it
intended purpose. Thus, the laboratory personnel must ensure, as far as is practicable, that the
accurate and precise, and the clinician should appreciate how useful the test is in the conte
hich it is used. Various properties of a test can be calculated to provide this information.
pecificity and sensitivity
rlier in the chapter, the terms ‘sensitivity’ and ‘specificity’ were used to describe characterist
alytical methods. The terms are also widely used in the context of the utility of laboratory testsecificity of a test is a measure of the incidence of negative results in persons known to be fre
ease, that is, ‘true negative’ (TN). Sensitivity is a measure of the incidence of positive resu
ients known to have a condition, that is, ‘true positive’ (TP). A specificity of 90% implies
% of disease-free people would be classified as having the disease on the basis of the test r
y would have a ‘false positive’ (FP) result. A sensitivity of 90% implies that only 90% of pe
own to have the disease would be diagnosed as having it on the basis of that test alone: 10% w
‘false negatives’ (FN).
ecificity and sensitivity are calculated as follows:
ideal diagnostic test would be 100% sensitive, giving positive results in all subjects w
rticular disease, and also 100% specific, giving negative results in all subjects free of the dis
cause the ranges of results in quantitative tests that can occur in health and in disease a
ways show some overlap, individual tests do not achieve such high standards. Factors that inc
specificity of a test tend to decrease the sensitivity, and vice versa. To take an extreme examp
were decided to diagnose hyperthyroidism only if the plasma free thyroxine concentration w
st 32 pmol/L (the upper limit of the reference range is 26 pmol/L), the test would have effec
0% specificity: positive results (>32 pmol/L) would only be seen in thyrotoxicosis (an except
ery rare condition in which patients are resistant to thyroid hormones). On the other hand, th
uld have a low sensitivity in that many patients with mild hyperthyroidism woul
sdiagnosed. If a concentration of 20 pmol/L were used, the test would be very sensitive (all
th hyperthyroidism would be correctly assigned) but have low specificity, because many no
ople would also be diagnosed as having the condition. These concepts are illustrated in Figure
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
23/627
gure 1.6 Because the ranges of values for a test result in health and disease overlap (A),
ients with disease will have results within the reference range (false negatives), while
dividuals free of disease will have results outside this range (false positives). If the diagnostic
f value for a test is set too high (B), there will be no false positives, but many false nega
ecificity is increased but sensitivity decreases. If the diagnostic cut-off value is set too low (C
mber of false positives, and sensitivity, increases, at the expense of a decrease in specificity.
hether it is desirable to maximize specificity or sensitivity depends on the nature of the con
t the test is used to diagnose and the consequences of making an incorrect diagnosis. For exam
nsitivity is paramount in a screening test for a harmful condition, but the inevitable false po
ults mean that all positive results will have to be investigated further. However, in sele
ients for a trial of a new treatment, a highly specific test is more appropriate to ensure th
atment is being given only to patients who have a particular condition. In some cases, this dec
y not be straightforward, for example in the context of chest pain and suspected acute myoc
arction, where the possible options are to identify all those who have had a myocardial infar
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
24/627
ule in’) or to identify all those who have definitely not (‘rule out’). The preferred option s
pend on the relative outcomes of treatment and non-treatment for patients in the two groups.
e way of comparing the sensitivity and specificity of different tests is to construct rec
erating characteristic curves (ROC curves). Each test is performed in each of a seri
propriate individuals. The specificity and sensitivity are calculated using different cut-off valu
ermine whether a given result is positive or negative (Fig. 1.7). The curves can then be assess
ermine which test performs best in the specific circumstances for which it is required.
gure 1.7 ROC curves for three hypothetical tests, A, B and C. Examination of the curves show
t A performs less well in terms of both sensitivity and specificity than tests B and C. Test B
ter specificity than C, but C has better sensitivity.
e specialized use of the terms ‘sensitivity’ and ‘specificity’ that has been discussed here inntext of the utility of laboratory tests sometimes causes confusion, as these terms are also us
scribe purely analytical properties of tests. Readers should appreciate that, in this latter con
nsitivity’ relates to the ability of a test to detect low concentrations of an analyte and ‘specif
its ability to measure the analyte of interest and not some other (usually similar) substance.
fficiency
e efficiency of a test is the number of correct results divided by the total number of tests.
iciency is given by:
hen sensitivity and specificity are equally important, the test with the greatest efficiency shou
ed.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
25/627
redictive values
en a highly specific and sensitive test may not necessarily perform well in a clinical context.
because the ability of a test to diagnose disease depends on the prevalence of the condition i
pulation being studied (prevalence is the number of people with the condition in relation
pulation). This ability is given by the ‘predictive value’ (PV). PV+ve, the PV for a positive r
he percentage of all positive results that are true positives, that is:
a condition has a low prevalence and the test is
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
26/627
e concept of predictive values is an unfamiliar one for many people: it has no obvious paral
r everyday lives. The concept of odds is a more familiar one. Likelihood ratios (LRs) expres
ds that a given finding (e.g. a particular result) would occur in a person with, as oppos
thout, a particular condition. The LR for a positive result is given by:
e LR –ve (the odds that a negative test result would occur in a person with, as opposed to with
rticular condition) is given by:
Rs can be used to convert the probability of a condition being present before the test was doncase of a screening test, this is the prevalence) to the post-test probability of its being present
eater the value of the LR, the more useful the test will have been.
vidence-based clinical biochemistry
ost clinicians use laboratory tests primarily on the basis of their own clinical expertise
erpret results intuitively. Ideally, tests should be chosen on the basis of evidence of their u
d their results used on the basis of outcome measures. Such an approach is advocated as part oactice of evidence-based medicine, and could be facilitated by the use of test characteristics
have been discussed above. However, it remains the case that many well-established tests
en introduced into clinical practice without being properly evaluated, and few systematic rev
existing tests have been performed. Furthermore, new tests are often introduced into laborato
pertoires without a systematic assessment of their utility having been made, and their valu
mitations may only become apparent in the light of experience of their day-to-day use.
Clinical audit
nical audit is part of the process of ensuring quality—in this context, of ensuring the provisio
gh quality laboratory service. In this respect, it is complementary to the other techniques of q
urance, which in the main concentrate on the analytical aspects of the service, that is, the prov
precise and accurate results. Clinical audit is the process of systematically examining pract
der to ensure that it is efficient and beneficial to patients. It involves identifying an area of pra
ting standards or guidelines (e.g. a protocol for investigation of patients suspected of hav
rticular condition), implementing changes designed to achieve these and then examining compl
th them and the effects on patient care. The cycle is completed by review of the standards i
ht of this analysis and their modification as required. It should be followed by re-audit aft
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
27/627
propriate interval. Whether undertaken in the context of formal audit or not, ongoing li
ween the providers and users of laboratory services is essential to ensure that the service
latter’s needs. It also provides a forum for laboratory staff to educate users about chang
actice designed to improve the service.
e term ‘audit’ is also applied to procedures used by some laboratory accreditation bodi
amine the internal functioning of laboratories. It is beyond the scope of this book to describe
ocedures.
creening
reening tests are used to detect disease in groups of apparently healthy individuals. Such tests
applied to whole populations (e.g. the detection of PKU and other inherited metabolic disord
newborn), to groups known to be at risk (the detection of hypercholesterolaemia in the relativ
ople with premature coronary heart disease), or to groups of people selected for other re
ochemical profiling of preoperative patients, health screening for business executives
eening for common conditions in the elderly).
previously discussed, high sensitivity is particularly desirable for screening tests but, to
necessary further tests of normal people, high specificity is also an important consider
reening tests for PKU are designed to maximize sensitivity but are also highly specific. How
KU has a low incidence so that even with a sensitivity of 100% and specificity of 99.9%
edictive value of a positive test is only 10%, that is, 9 out of 10 positive tests will be show
ther investigation to be false positives. These calculations are made as follows:
ncidence of PKU = 1 in 10 000 live births
ensitivity = 100% or
pecificity = 99.9% or
number of positive tests per 10 000 infants
numbers of TP and FP results: TP = 1, FP = 9
redictive value of a positive test
the other hand, the predictive value of a negative test will be 100%, confirming that no case
missed using the screening test.
reening for specific conditions is discussed in other chapters of this book. Such screening is
sed on the use of considerably less specific or sensitive tests, and therefore has a low effic
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
28/627
detecting disease. Indiscriminate biochemical profiling is also inefficient. The more tests th
rformed, the greater is the probability that an apparently abnormal result will arise that is n
ult of a pathological process.
hen multiple analyses are performed and an unexpected abnormality is found, a decision mu
de as to what action to take. The abnormality may be considered insignificant in some cl
cumstances, but, if it is not, further investigations must be made. Although these may be of ult
nefit to the patient, he or she may suffer anxiety in the short term, and their cost and econ
nsequences may be considerable. At the very least, the tests should be repeated to ensure thanormality was not due to analytical error.
e ready availability of an investigation often leads to its being used unnecessaril
ppropriately. Doctors should be encouraged to be selective in making test requests. B
questing a test, a doctor should know how the result will influence the management of the patie
will not have an influence, it should not be requested.
mmary
Biochemical investigations are used for diagnosis, monitoring, screening and in prognosis
Specimens for analysis must be collected and transported to the laboratory under approp
nditions
Analytical results are affected by both analytical and biological variation
Results can be compared either with reference intervals or with the results of previous tests
The utility of test results depends on many factors: an ‘abnormal’ result should not be assum
dicate a pathological process, nor a ‘normal’ one to exclude disease or potential disease
The utility of tests can be measured and described mathematically: applying this informatio
nsiderably enhance the value of laboratory test results in clinical practice.
Plasma and serum
asma is the aqueous phase of blood and can be obtained by removal of blood cells from blo
hich an anticoagulant has been added. Serum is the aqueous phase of blood that has been allow
t. For technical reasons, many biochemical measurements are more conveniently made on se
t the concentrations of most analytes are effectively the same in both fluids. In this book, the
rum’ is used only where actual measurements made in serum are referred to (e.g. in the
tories) and in the few instances where serum must be used for analysis.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
29/627
apter 2
ater, sodium and potassium
ntroduction
Water distribution
ater accounts for approximately 60% of body weight in men and 55% in women, the diffe
lecting the typically greater body fat content in women. Approximately 66% of this water is
racellular fluid (ICF) and 33% in the extracellular fluid (ECF); only 8% of body water is i
sma (Fig. 2.1). Water is not actively transported in the body. It is, in general, freely perm
ough the ICF and ECF and its distribution is determined by the osmotic contents of
mpartments. Except in the kidneys, the osmotic concentrations, or osmolalities, of
mpartments are always equal: they are isotonic. Any change in the solute content of a compart
genders a shift of water, which restores isotonicity.
gure 2.1 Distribution of water, sodium and potassium in the body of a 70 kg man. The distributmilar in women, although the amount of water as a percentage of body weight is less. In ch
d infants, total body water is 75–80% of body weight, with a higher ECF : ICF volume ratio th
ults, but the proportion of the total body water contained in the plasma is the same. Note
hough plasma volume is approximately 3.5 L, blood volume in a 70 kg man is approximately 5
e major contributors to the osmolality of the ECF are sodium and its associated anions, m
oride and bicarbonate; in the ICF the predominant cation is potassium. Other determinants of
molality include glucose and urea. Protein makes a numerically small contribution of approxim
%. This is because osmolality is dependent on the molar concentrations of solutes: althoug
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
30/627
al concentration of plasma proteins is approximately 70 g/L, their high molecular weight resu
ir combined molar concentrations being
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
31/627
s important to appreciate that there is a massive internal turnover of sodium. Sodium is sec
o the gut at a rate of approximately 1000 mmol/24 h and filtered by the kidneys at a ra
000 mmol/24 h, the vast majority being regained by reabsorption in the gut and renal tub
pectively. If there is even a partial failure of this reabsorption, sodium homoeostasis wi
mpromised.
otassium distribution
tassium is the predominant intracellular cation. Some 90% of the total body potassium id therefore exchangeable, while the remainder is bound in red blood cells, bone and brain ti
wever, only approximately 2% (50–60 mmol) of the total is located in the extrace
mpartment (see Fig. 2.1), where it is readily accessible for measurement. Plasma pota
ncentration is not, therefore, an accurate index of total body potassium status, but, because o
ect of potassium on membrane excitability, is important in its own right. The potas
ncentration of serum is 0.2–0.3 mmol/L higher than that of plasma, owing to the relea
tassium from platelets during clot formation, but this difference is not usually of pra
nificance.
ere is a constant tendency for potassium to diffuse down its concentration gradient from the I
ECF, opposed by the action of Na +,K +-ATPase (the sodium pump), which transports potas
o cells. Potassium homoeostasis and its disorders are described later in this chapter.
ater and sodium homoeostasis
Water and ECF osmolality
anges in body water content independent of the amount of solute will alter the osmolality
). The osmolality of the ECF is normally maintained in the range 282–295 mmol/kg of water
s of water from the ECF, such as occurs with water deprivation, will increase its osmolality
ult in movement of water from the ICF to the ECF. However, a slight increase in ECF osmo
ll still occur, stimulating the hypothalamic thirst centre, causing thirst and thus promoting a d
drink, and stimulation of the hypothalamic osmoreceptors, which causes the release of vasopr
ntidiuretic hormone, ADH).
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
32/627
gure 2.3 Physiological responses to water loss.
sopressin renders the renal collecting ducts permeable to water (its combination with V2 rece
ults in the insertion of aquaporins (water channels) into the normally impermeable a
mbrane of the cells of the collecting tubules), permitting water reabsorption and concentratiurine; the maximum urine concentration that can be achieved in humans is about 1200 mmo
e osmoreceptors are highly sensitive to osmolality, responding to a change of as little as
asma vasopressin concentration falls to very low values at an osmolality of 282 mmol/kg, but
arply if osmolality increases above this level (Fig. 2.4A). However, if an increase in
molality occurs as a result of the presence of a solute such as urea that diffuses readily acros
mbranes, ICF osmolality also increases and osmoreceptors are not stimulated.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
33/627
gure 2.4 (A) Vasopressin secretion is stimulated by a rise in ECF osmolality above a thresho
proximately 282 mmol/kg; in hypotension (blue line), this threshold is reduced and the respon
eater. (B) Vasopressin secretion is stimulated exponentially by hypotension. Note the differen
scales of the vertical axes.
ECF osmolality falls, there is no sensation of thirst and vasopressin secretion is inhibited. D
ne is produced, allowing water excretion and restoration of ECF osmolality to normal.
sopressin responses to changes in osmolality occur rapidly. In health, the ingestion of water su
requirements leads to a rapid diuresis, and water depletion to a rapid increase in the concentr
the urine.
her stimuli affecting vasopressin secretion (Fig. 2.5) include angiotensin II, arterial and ve
roreceptors and volume receptors (which sense blood pressure and volume, respectiv
povolaemia and hypotension increase the slope of the vasopressin response to an increa
molality (see Fig. 2.4A) and lower the threshold osmolality for vasopressin secretion
sopressin response to a fall in blood pressure is exponential: it is relatively small with
creases in plasma volume, but greater falls cause a massive increase in vasopressin secretio
g. 2.4B). Osmolar controls are overridden, so that ECF volume is defended (by stimulating
ention) at the expense of a decrease in osmolality.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
34/627
gure 2.5 Factors affecting vasopressin secretion. ECF osmolality is normally the most importase.
odium and ECF volume
e volume of the ECF is directly dependent on the total body sodium content as water i
d loss are regulated to maintain a constant ECF osmolality, and hence sodium concentration
cause sodium is virtually confined to the ECF.
etary sodium intake is highly variable, being of the order of 100–200 mmol/day on a tystern diet. Much of this is the result of salt added to food during manufacturing or prepar
dium balance is maintained by regulation of its renal excretion. Sodium excretion is depende
omerular filtration, but the glomerular filtration rate (GFR) appears to become an important lim
tor in sodium excretion only at extremely low rates of filtration (sodium retention is a late fe
chronic kidney disease). Normally, approximately 70% of filtered sodium is actively reabso
the proximal convoluted tubules, with further reabsorption in the loops of Henle. So
bsorption is reduced in the proximal convoluted tubules if blood volume increases (because t
ociated with a fall in the oncotic pressure in peritubular capillary blood) and if sympat
ivity decreases (such as also tends to occur with an increase in blood volume). However, alth
ost sodium reabsorption occurs in the proximal nephron, and
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
35/627
ood volume (effective blood volume, see below).
triuretic peptide hormones also have a role in controlling sodium excretion. Atrial natriu
ptide (ANP) is a 28 amino acid peptide, one of a family of similar peptides, secreted b
rdiac atria in response to atrial stretch following a rise in atrial pressure (e.g. due to ECF vo
pansion). ANP acts both directly by inhibiting distal tubular sodium reabsorption and thr
creasing renin (and hence aldosterone) secretion. It also antagonizes the pressor effec
repinephrine (noradrenaline) and angiotensin II (and thus tends to increase GFR) and
temic vasodilatory effect. It appears to provide ‘fine tuning’ of sodium homoeostasis bobably more important in pathological states than physiologically. Two other structurally si
ptides have been identified: one (brain natriuretic peptide, BNP) is secreted by the ca
ntricles in response to ventricular stretching and has similar properties to ANP; the other (C
riuretic peptide, CNP) is present in high concentrations in vascular endothelium and
sodilator. Measurement of BNP is of value in the management of patients with suspected ca
lure (see Chapter 14). Increased secretion of natriuretic peptides has been postulated to be at
part responsible for the natriuresis seen in cerebral salt-wasting (see p. 27).
summary, in health, the response to an increase in ECF volume (e.g. as a result of infusing isoine) is a decrease in the secretion of aldosterone (with no change in that of vasopressin), lead
riuresis. In sodium depletion, an increase in aldosterone secretion leads to sodium retention
ter retention is only stimulated in severe sodium depletion. In general, the control mechanism
CF volume respond less rapidly and are less precise than the control mechanisms for
molality. Unless hypovolaemia is severe, maintenance of osmolality takes precedence.
ysiological responses to a decrease in ECF volume are illustrated in Figure 2.6. In addition
anges in sodium excretion, these also involve changes in the tone of arteriolar smooth muscle
nce peripheral vascular resistance and blood pressure.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
36/627
gure 2.6 Physiological responses to a decrease in plasma volume. These involve respons
tore plasma volume and to maintain blood pressure.
ater and sodium depletion
ater depletion, or combined water and sodium depletion, will occur if losses are greater
ake. Depletion of water alone is seen much less frequently than depletion of both water
dium. As sodium cannot be excreted from the body without water, sodium loss never occurs
t is always accompanied by some loss of water. The fluid may be isotonic or hypotonic
pect to ECF.
e clinical and biochemical features of water depletion and of isotonic sodium and water los
ite different, as are the physiological responses, and it is helpful to consider them separate
nical practice, however, states of fluid depletion encompass the whole spectrum between o extremes and the clinical and biochemical features will reflect this. Furthermore, it shou
preciated that they may have been modified by previous treatment.
Water depletion
ater depletion will occur if water intake is inadequate or if losses are excessive . Exce
s of water without any sodium loss is unusual, except in diabetes insipidus, but, even if th
s of sodium as well, provided that this is small, the clinical consequences will be re
marily to the water depletion (Fig. 2.7).
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
37/627
gure 2.7 Water depletion. Causes and clinical features of predominant water depletion. In inf
stroenteritis and in acclimatization to high temperatures, some sodium is lost from the gut and
pectively, but the effects of water loss may predominate. *Unless due to renal water loss.
ss of water from the ECF causes an increase in osmolality, which in turn causes movemeter from the ICF to the ECF, thus lessening the increase. Nevertheless, the increase in
molality will be sufficient to stimulate the thirst centre and vasopressin secretion. Plasma so
ncentration is increased; plasma protein concentration and the haematocrit are usually only sli
vated. Unless water depletion is due to uncontrolled loss through the kidneys, the urine bec
ghly concentrated and there is a rapid decrease in its volume. Because water loss is borne b
al body water pool, and not just the ECF (Fig. 2.8) , signs of a reduced ECF volume ar
ually present. Furthermore, the increased colloid osmotic pressure of the plasma tends to
racellular water in the vascular compartment. Circulatory failure is a very late feature of
pletion: it is much more likely to occur if sodium depletion is also present.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
38/627
gure 2.8 Comparison of the effects of water loss and isotonic fluid loss from the extrace
mpartment. When only water is lost from the ECF, the increase in osmolality causes water to mm the ICF, which minimizes the decrease in plasma volume. When isotonic fluid is lost from
CF, no osmotic imbalance is produced, there is no movement of water from the ICF and the effe
sma volume is, therefore, much greater. Similarly, excess isotonic fluid is confined to the ECF
excess of water is shared by the whole body water compartment and the effect on the ECF is
ch less.
vere water depletion induces cerebral dehydration, which may cause cerebral haemorrhage th
ring of blood vessels. In the short term, cerebral shrinkage is mitigated somewhat by moveme
racellular ions into cerebral cells, causing an osmotic intracellular shift of water. If dehydrrsists, brain cells adapt by synthesizing osmotically active organic compounds (‘osmolytes’)
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
39/627
rebral oedema may then follow rapid fluid replacement (see Fig. 2.9B).
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
40/627
gure 2.9 The effects of hyponatraemia (panel A) and hypernatraemia (panel B) on the b
aptive changes and the effects of rapid correction. T = tonicity; N = normal.
e management of water depletion requires treatment of the underlying cause and replaceme
fluid deficit. Water should preferably be given either orally or via a nasogastric tube. If this
ssible, either 5% dextrose or, if there is also some sodium depletion, ‘dextrose–saline’
xtrose, 0.18% sodium chloride) should be given intravenously. As a general guide, the aim sh
to correct approximately two-thirds of the deficit in the first 24 h and the remainder in the
h while avoiding a decrease in sodium concentration of more than 10 mmol/L in the first w rapidly the sodium concentration should be normalized depends on how quickly i
veloped. If it is long-standing (as it often is in the elderly) an increase of no more
mmol/L/h is recommended, but initially more rapid correction (1 mmol/L/h) may be approp
acute water depletion (more common in children).
odium depletion
dium depletion is seldom due to inadequate oral intake alone, but sometimes inadequate pare
put is responsible. More often, it is a consequence of excessive sodium loss (Fig. 2.10). Son be lost from the body either isotonically (e.g. in plasma) or hypotonically (e.g. in sweat or d
ne). In each case, there will be a decrease in ECF volume (see Fig. 2.8), but this will be l
fluid lost is hypotonic than if it is isotonic, as some of the water loss will be shared with the
e clinical features of sodium depletion (see Fig. 2.10) are primarily a result of the decrease in
ume.
gure 2.10 Causes and clinical features of predominant sodium depletion. The clinical signs arhypovolaemia. Oliguria develops gradually: it is primarily due to the decrease in GFR, rather
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
41/627
the effects of vasopressin. Thirst is usually a late feature.
e normal responses to hypovolaemia are an increase in aldosterone secretion, stimulating
dium reabsorption in the distal convoluted tubules and collecting ducts, and a fall in urine vo
a consequence of a decreased GFR. Significantly increased vasopressin secretion, w
mulates the production of a highly concentrated urine, only occurs with more severe ECF vo
pletion (see Fig. 2.4).
e decrease in GFR may lead to pre-renal uraemia (see Case history 4.1). In contrast to the epure water depletion, plasma protein concentration and the haematocrit are usually c
reased in sodium depletion, unless this is a result of the loss of plasma or blood. Furtherm
cause the fluid loss is borne mainly by the ECF, signs of a reduced ECF volume are us
esent, and there is a greater risk of peripheral circulatory failure than in water depletion
tures of sodium and water depletion are compared in Figure 2.11.
gure 2.11 Clinical and laboratory findings in sodium and water depletion. *Unless due to lo
ood.
e plasma sodium concentration can give an indication of the relative amounts of water and so
t have been lost: plasma sodium will be normal if the fluid lost is isotonic with respect to the
d increased if it is hypotonic. With severe sodium depletion, increased vasopressin secrcondary to the resulting hypovolaemia may cause water retention; plasma volume is then maint
the expense of osmolality and hyponatraemia develops. Thus the plasma sodium concentration
dium-depleted patient may be low, normal or high (Fig. 2.12).
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
42/627
gure 2.12 Plasma sodium concentration with various causes of sodium depletion. The p
dium concentration alone is a poor guide to ECF sodium status.
e management of sodium depletion involves treatment of the underlying cause and, if neces
toration of the intravascular volume by giving isotonic fluid (‘normal saline’ (0.9% so
oride) or colloid (plasma expanders or albumin)) by intravenous infusion. This can usualne rapidly, but any associated free water deficit requires more cautious correction.
ater and sodium excess
cess of water and sodium can result from a failure of normal excretion or from exce
ake. The latter is often iatrogenic. As with the syndromes of depletion, it is helpful to consid
uses and consequences of excess water alone and of sodium excess with isotonic retention of
parately, although in practice there is often a degree of overlap.
Water excess
is is usually related to an impairment of water excretion (Fig. 2.13). However, the limit t
lity of the healthy kidneys to excrete water is about 20 mL/min and, occasionally, excessive i
alone sufficient to cause water intoxication. This can sometimes occur in patients with psych
orders. It has also been described in people drinking large amounts of beer with a low s
ntent, because this results in a low osmotic load for excretion and there is a minimum osmo
ow which the urine cannot be diluted further. Increased thirst can occur in organic brain di
articularly trauma, and following surgery), although decreased thirst is more com
ponatraemia is invariably present in water overload. The increased water load is shared b
F and ECF.
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
43/627
gure 2.13 Causes and clinical features of excess body water.
e clinical features of water overload (see Fig. 2.13) are related to cerebral overhydration;
idence and severity depend upon the extent of the water excess and its time course. Thus, a pth a plasma sodium concentration of 120 mmol/L, in whom water retention has occurred grad
er several days, may be asymptomatic, while one in whom this is an acute phenomenon may
ns of severe water intoxication. In the short term, the effects of hypotonicity are mitigated to
ent by a movement of ions out of cerebral glial cells; more chronically (days), a decrea
racellular organic ‘osmolytes’ further reduces intracellular water content (see Fig 2.9A). As
e with water depletion, this adaptation necessitates a cautious approach to treatment, particu
chronic water overload. The management of water overload is discussed together with th
ponatraemia on p. 28.
odium excess
dium excess can result from increased intake or decreased excretion. The clinical feature
ated primarily to expansion of ECF volume (Fig. 2.14). When related to excessive intake (e.
ppropriate use of hypertonic saline), a rapid shift of water from the intracellular compartmen
o cause cerebral dehydration. When sodium overload is due to excessive intake, hypernatraem
ual (see Case history 2.5).
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
44/627
gure 2.14 Causes and clinical features of predominant sodium excess.
dium overload is more usually due to impaired excretion than to excessive intake . The
quent cause is secondary aldosteronism. This is seen in patients who, despite clinical evid
increased ECF volume (e.g. peripheral oedema), appear to have a decreased effective ar
ood volume, due, for example, to venous pooling or a disturbance in the normal distribution of
ween the vascular and extravascular compartments. This phenomenon is particularly assoc
th cardiac failure, hypoalbuminaemia and hepatic cirrhosis. Many such patients with sodium e
, paradoxically, hyponatraemic, implying the coexistence of a defect in free water excretion.
probably in part due to an increase in vasopressin secretion as a result of the decreased effe
ood volume. Also, the decrease in GFR and consequent increase in proximal tubular so
bsorption decreases the delivery of sodium and chloride to the loops of Henle and
nvoluted tubules. This reduces the diluting capacity of the kidneys, thereby compromising
cretion. Renal disease is a relatively uncommon cause of sodium excess, as is incr
neralocorticoid secretion due to primary adrenal disease (as in Conn’s syndrome, see p. 147).
s noteworthy, however, that oedema is not a feature of Conn’s syndrome: furthermore, in no
dividuals the administration of high doses of mineralocorticoids initially leads to sodium rete
d modest expansion of the ECF volume (but to an insufficient extent to cause oedema), but so
ance is then restored and a new steady state achieved. It is thought that the increased arterial f
ds to a decrease in sympathetic activity and secretion of angiotensin II, with a consequent inc
renal perfusion and GFR, together with the increased secretion of ANP. The net result
rease in the delivery of sodium to the distal nephron, which, together with ANP, counter
dium-retaining action of aldosterone. In oedematous states, relative arterial underfilling lead
l in GFR, increased proximal tubular sodium reabsorption, and decreased delivery of sodiu
distal nephron. Even though there is increased secretion of ANP, its ability to cause natriure
mited by this decreased sodium delivery.
e management of sodium excess should be directed towards the cause, where possib
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
45/627
dition, diuretics can be used to promote sodium excretion, and sodium intake must be contr
alysis may be necessary if renal function is poor and is occasionally necessary in acute so
erload associated with the use of hypertonic fluids.
aboratory assessment of water and sodium statu
asma sodium concentration is dependent on the relative amounts of sodium and water
sma. In isolation, therefore, plasma sodium concentration provides no information aboudium content of the ECF. It may be raised, normal or low, in states of sodium excess or deple
cording to the amount of water in the ECF.
e plasma sodium concentration is one of the most frequent measurements made in cl
emistry laboratories (largely for historical reasons), but definite indications for its measure
few and results are often misinterpreted. Plasma sodium concentration should be measured i
lowing:
atients with dehydration or excessive fluid loss, as a guide to appropriate replacement
atients on parenteral fluid replacement who are unable to indicate or respond to thirst (e.g
matose, infants and the elderly)
atients with unexplained confusion, abnormal behaviour or signs of CNS irritability.
the assessment of a patient’s water and sodium status, clinical observations, such as measurem
central venous pressure, fluid balance and body weight, may all provide vital information
rease in the concentration of plasma proteins or in the haematocrit suggests haemoconcentr
her abnormal results may suggest specific conditions; for example, hyperkalaemia ponatraemic patient with clinical evidence of sodium depletion suggests adrenal failure.
alysis of urine can provide valuable information, but results may be misleading. It shou
ablished whether the urine volume and composition are physiologically appropriate fo
ient’s water and sodium status. If they are not, the reason should be sought. For example, a
nary sodium excretion is to be expected in a patient with hyponatraemia who is sodium dep
nless this is due to renal sodium loss: natriuresis in such a patient would imply either a failu
osterone secretion or a failure of the kidney to respond to the hormone (Case history 2.1)).
odium measurement
dium concentration used to be measured by flame photometry, which determines the numb
dium atoms in a defined volume of solution. Sodium is now usually measured by ion-sele
ctrodes, which determine the activity of sodium; that is, the number of atoms that act as true io
efined volume of water.
der most circumstances, the two techniques give results that are, for practical clinical purpose
me. However, as activity is a measure of sodium in the water fraction of plasma (normally 93
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
46/627
lume), significant discrepancies between activity and concentration may arise if the frac
sma water content is decreased, such as in severe hyperlipidaemia or hyperproteinaemia
dium concentration, measured by flame photometry in mmol/L of plasma, will be less tha
ncentration inferred from the activity. This is because, although the concentration of sodiu
sma water is unchanged, there is less water and thus less sodium in a given volume of pl
alysers employing electrodes for which the plasma is diluted before measurement also g
uriously low result. This effect, known as pseudohyponatraemia, is only seen with s
perlipidaemia (when the plasma will usually appear turbid to the naked eye) (see Case h
2) and with large increases in total protein due to paraproteinaemia. If it is suspected, plmolality should be measured: it is osmolality that is regulated by the hypothalamus throug
ease of vasopressin. Plasma osmolality should be normal in a patient with pseudohyponatraem
Measurement of osmolality
ven that it is osmolality, rather than sodium concentration, that is controlled by the hypothalam
ght appear logical to measure plasma osmolality rather than sodium concentration.
asurement of osmolality is, however, less precise than that of sodium and is not easily autom
s nevertheless useful under certain circumstances.
easurement of osmolality may help in the interpretation of a low plasma sodium concentratio
necessary in water deprivation tests. It can also be useful in the investigation of patients susp
having ingested substances such as ethanol or ethylene glycol (see Case history 19.3) becau
esent, these increase the plasma osmolality. This can be revealed by comparing the mea
molality with the approximate expected value calculated using the formula:
here all concentrations are measured in mmol/L. The factor 2 is to allow for the major a
hloride and bicarbonate); other formulae include potassium and a slightly smaller multipli
ow for the fact that some anions and cations act as ion pairs rather than as individual moieties
easured osmolality (units: mmol/kg of water) and calculated osmolarity (mmol/L of solution
rmally numerically very similar. Significant discrepancies (an osmolar gap) occur when abno
motically active species are present in plasma (as may occur in poisoning) and when the fractter content of plasma is reduced, as in severe hyperlipidaemia or hyperproteinaemia.
Measurement of anions (bicarbonate and chloride)
change in plasma sodium concentration must be matched by a change in anion concentration
jor anions of the ECF are chloride and bicarbonate. Bicarbonate (strictly, total carbon dioxid
s mostly comprises bicarbonate ions) is frequently measured because it reflects the extrace
ffering capacity (note that the measurement must be made on a fresh sample to obtain an accult, owing to the loss of carbon dioxide to the atmosphere on standing), but the measureme
-
8/17/2019 Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta
47/627
sma chloride rarely adds to the information that can be derived from knowledge of the so
ncentration alone, and few laboratories in the UK measure plasma chloride concentration rout
wever, it may occasionally be helpful in the diagnosis of patients with non-respiratory acidos
e chloride-losing states.
Hyponatraemia
slightly low plasma sodium concentration is a frequent finding. The mean plasma so
ncentration of hospital inpatients is ~5 mmol/L lower than in healthy controls. Mild hyponatra
seen with a wide variety of illnesses and may be multifactorial in origin (see the ‘sick
ndrome’, p. 30). It is essentially a secondary phenomenon that merely reflects the presen
ease; treatment should be directed at the underlying cause and not at the hyponatra
ponatraemia itself may warrant primary treatment, but usually only when it is severe or assoc
th clinical features of water intoxication (see Fig. 2.13).
auses
has been emphasized that plasma sodium concentration depends on the amounts of both sodium
ter in the plasma, and so a low sodium concentration does not necessarily imply sodium depl
deed, hyponatraemia is more frequently a result of a defect in water homoeostasis that causes w
ention and hence dilution of plasma sodium. One of three mechanisms is usually prim
ponsible for the development and maintenance of hyponatraemia, although in individual pa
ore than one factor may be involved. These are:
odium depletion (hypovolaemic hyponatraemia)
water excess (