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


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.


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


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


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


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


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.


8/627

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


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.


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


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.


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.


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


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.


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.


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.)


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.


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


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.


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

=


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


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


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


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.


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


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


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


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.


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


30/627

al concentration of plasma proteins is approximately 70 g/L, their high molecular weight resu

ir combined molar concentrations being


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).


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.


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.


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


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.


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).


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.


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’)


39/627

rebral oedema may then follow rapid fluid replacement (see Fig. 2.9B).


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


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).


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.


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).


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


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


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


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 (

Documents

Clinical Chemistry, Seventh Edition - Marshall, William J. & Bangert, Stephen K & Lapsley, Marta