r-1. a systematic approach to assesment of erythropoiesis

Clin. Lab. Haem.2001, 23,271-283 REVIEW

A systematic approach to the assessmentof erythropoiesis

H.M. WATERS*. *Utiiversity Department of Haematology, Manchester Royal Infirmary and 1-Department of

L.H. SEALf Biological Sciences, Manchester Metropolitan University, Manchester, UK

Summary The pathogenesis of anaemia may be simple or complex and the differential diagnosiscan be difiicult. An appreciation of the erythropoietic processes is required, togetherwith regular review of investigations, to ensure that appropriate protocols are adopted.The application of tests, which define different facets of erythropoiesis, should beappropriate to the clinical circumstances. In some situations, such as the anaemia ofchronic disorders, pregnancy and chronic renal failure, a detailed analysis oferythropoiesis is often required.

Guidelines for investigating anaemia due to megaloblastosis or haemoglobinopathy arewell established, whereas disturbances of iron metabolism are often difficult to classify.These require a clear distinction between storage and functional iron to differentiatewhether the defect is due to readily treatable simple iron deficiency or more complexmechanisms, which do not respond to iron supplementation. Determination of red cellhaemoglobin content, reticulocyte analysis and the assay of serum transferrin recep-tors are new generation parameters developed to address this.

Practice pressures and new treatment options have contributed to investigationsbecoming more complex, especially those of the secondary anaemias, as new tests havebecome more readily available and often automated. This has resulted in reducedturnaround times and clinical demand has driven request patterns. Initiatives todevelop evidence-based anaemia management protocols are welcomed but, whereverpossible, should he developed through collahoration hetween the haematologydepartment and the user unit, and hased on available guidelines.

Keywords Erythropoiesis, review, anaemia, chronic, parameters, secondary

Introduction

Anaemia can be a feature ofthe disease processes in nearly180 disorders. Whilst, in some of these, the pathogenesismay be simple, in others it can be complex, which canmai e differential diagnosis more difficult. An appreciationof the cell biology, physiology and pathophysiology oferythropoiesis is essential to ensure that appropriateinvestigation protocols are adopted in individual cases.This approach cuts across traditional ciinicai and

Accepted for publication 7 August 2001Correspondence: H. M. Waters. University Department ofHaematology. Manchester Royal Infirmary, Oxford Road, ManchesterM13 9WL. UK. E-mail: [email protected]

© 2001 Blackwell Science Limited

laboratory classifications. As improvements in the under-standing of erythropoiesis continue, it is also important tore-evaiuate approaches and protocols periodically toincorporate new ideas and evaluate any new or improvedanalytical methods for the investigation of anaemia.

Normal erythropoiesis is the result of a steady-statebaiance between red cell mass (RCM) and erythropoietin(EPO), with each influencing the other (Figure 1). Inuncomplicated sudden blood loss, RCM falls, EPO rises andRCM is restored by a phase of hyperproliferative erythro-poiesis. Three main variations from this EPO/RCM rela-tionship have been identified in anaemias.

1. Hypoproliferative erythropoiesis in which either thebone marrow is unabie to respond to an enhanced EPOsignal, as seen in iron deficiency - sideropenic, or where

A systematic approach to the assessment of erythropoiesis

BoneMarrow

Red Cell Mass

Kidney

Erythropoietin

Figure 1, Components of erythiropoiesis.

the EPO control mechanism or response to it isdiminished, as seen in chronic renal failure (CRE) andthe anaemia of chronic disease (ACD) - non-sideropenic.The latter two are complex conditions and are expandedupon later.

2. Ineffective erythropoiesis in which an increasedresponse is found but the output of cells is inadequate,as seen in megaloblastic anaemia, aplastic anaemia,thalassaemia, abnormal haemoglobins and iron defici-ency.

3. Hyperproliferative erythropoiesis in which an increasedresponse occurs but the demand for cell replacementexceeds the compensatory capacity of the bone mar-row, as seen in haemolytic anaemias.

Some types of anaemia show more than one type ofpathogenesis, for example, thalassaemia can have bothhyperproliferative and ineffective erythropoiesis, whereasiron deficiency can have hypoproliferative and ineffectiveelements. Analyses are available which define differentfacets of erythropoiesis. Some are basic descriptors, forexample, the red cell count or haemoglobin concentration,whereas others are functional response markers, forexample, reticulocytes or soluble transferrin receptors.When investigating anaemia, these analyses must be usedin an appropriate sequence or algorithm. Initially it issufficient to establish whether anaemia is due to an inabilityto make red cells or an inability to make haemoglobin. Insome instances this approach will be all that is required.However, in others, a full profile of erythropoiesis isessential and may require quite different thinking, inclu-ding using different reference ranges for some investiga-tions. The approach will also difi'er depending on whetherthe investigation is trying to establish a diagnosis or lookingfor a transition point during the course of a disorder and itstreatment, for example, checking iron availability whenpatients with CRF are treated with EPO. Table 1 summar-izes the application of parameters described below.

In erythrocytosis, although the outcome is different,the approach is similar. There may be an appropriate

relationship between EPO and RCM but, because offailure of O2 delivery to the kidney, RCM increases as inthe secondary polycythaemias. Alternatively there maybe a hyperproliferative response because of loss of thephysiological link between RCM and EPO, as seen inprimary polycythaemia when there is an erythro-poietic clone with a reduced or absent dependency onEPO,

Traditional parameters

Investigations: past

Clinical skills and observation originally formed the basisof a patient profile. These, together with a limited, manual,blood count, assessment of the reticulocyte response andexamination of blood film and bone marrow morphologyall played major roles. Diagnosis was presumptive andoften unreliable in differentiating anaemias with differentpathophysiologies.

Investigations: present

The differential diagnosis now includes a much extendedblood count profile which includes a readily available,accurate measurement of the mean cell volume (MCV).Assay of circulating levels of ferritin, vitamin B12 andfolate are carried out by most laboratories. Differentialstudies into the patient's vitamin B12 absorptive capacityare undertaken routinely in most large hospitals. Testsfor serum autoantibodies are offered by fewer centres, butare also readily available. Some functional tests tend tobe available only from a few specialized laboratories.

Blood film

Automated erythrocyte analysis has reached the stagewhere examination of the peripheral blood film is usedmainly to verify indices. This is particularly important inthe case of mixed cell populations and for red cell changeswhich do not affect blood count parameters.

Bone marrow examination

This provides a general evaluation of erythropoiesis andhas been largely superseded by non-invasive methods.Sufficient information can usually be derived from per-ipheral erythrocyte analysis to indicate erythropoieticdrive. The estimation of stainable iron in bone marrowbiopsies is a time-honoured method for assessing ironstores, although the procedure is traumatic for the patient

© 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23. 271-283

Table

Test

1, Summary of parameters

Sample

useful in evaluating erythropoiesis

Application

H. M. Waters & L H. Seal

Blood fllm

Bone marrow

MCV

Hypochromia

Reticulocytes

Ferritin

sTfK

Serum iron/TIBC

ZPP

B]2 assay

Folate assay

Peripheral blood

Marrow aspirate

Peripheral blood

Peripheral blood

Peripheral blood

Serum

Serum

Serum

Peripheral blood

Serum

Serum/red cell

IFAB Serum

Bi2 absorption studies /n vivo

du suppression test Marrow aspirate

Methylmaionic acid Serum

Homocysteine Serum

Important for mixed cell populations and for red cell changes which do not affectblood count parameters.

General evaluation of erythropoiesis. Gold standard for iron stores.

General indication of the size of circulating red cells. Non-speciflc. Cut-off valuesneed to be reconsidered in some situations.

Reflned MCH measurement indicating haemoglobin content of individualerythrocytes only available with some analysers. Provides a direct indicationof iron delivery to the bone marrow.

Manual methods of limited value. Automated analysis provides a range ofparameters including reticulocyte haemoglobin content, a sensitive indicatorof functional iron deflciency.

Reliable estimate of iron stores but can be disproportionately increased dueto acute phase response. Used in haematinic screen.

Generally considered a sensitive indicator of iron deflcient erythropoiesisunaffected by acute phase response. Results more meaningful when expressedas a ratio with serum ferritin.

Only reliable application is in screening for idiopathic haemochromatosis.

Rapid analysis indicative of both absolute and functional iron deflciency.Not always reliable in monitoring iron supplementation.

Differential diagnosis of macrocytosis. Reference ranges derived from whitepopulations may not be generally applicable. Used in haematinic screen.

DiiTerential diagnosis of macrocytosis. Serum and red cell assays each havelimitations therefore both should be assayed. Used in haematinic screen.

Diagnostic test for pernicious anaemia but only 70% sensitivity.

Differential diagnosis of B12 malabsorption. Does not exclude malabsorptionof food-bound vitamin.

May be useful to conflrm mild B]2 or folate deficiency.

Increased in B12 deflciency.

Increased in B12 and folate deflciency.

and the method only semiquantitative. Observing iron in

developing cells provides direct information to differentiate

between reduced stores in iron deficiency and impaired

functional release from reticuloendothelial cells in the

chronic anaemias (Baynes, 1996; Worwood, 1997). For

this reason most investigations, which attempt to evaluate

methods for assessing iron stores and functional iron, have

used bone marrow iron observation as the reference

method.

Mean cell volume and mean cell haemoglobin

A low MCV indicates a general decrease in the size of

circulating red cells and this effect may be seen in more

than one condition. In contrast, although discovery of a

high MCV can facilitate the early diagnosis of megalob-

lastic anaemia, since this will often precede clinical

anaemia by months or even years, up to 4% of all

patients may have abnormally high MCVs in the

absence of any other abnormality (Lindenbaum,

1983), Natural and biological variations, sometimes

compounded by coexisting pathological conditions often

result in great variation in the MCV rendering it an

unreliable, if not misleading, measurement on occasion.

Thus, a normal, lowered or raised MCV cannot always

be equated to normo-, micro- or macrocytosis, respect-

ively, often making examination of the blood film

essential. Cavill (1997) describes the MCV as an 'elastic

parameter' and suggests that, since the haemoglobin

content of red cells is fixed throughout their lifespan, the

2001 Blackwell Science Ltd.. Clin. lab. Haem., 23, 271-283


mean cell haemoglobin (MCH) provides a more reliable,and standardizable, index of cell size.

Reticulocytes

Reticulocyte analysis is undergoing fundamental change.Subjective visual investigations have restricted its useful-ness in the past to situations where marked changes inbone marrow response may be occurring such as hae-morrhage, haemolysis or the response to treatment. Highlevels of imprecision, with a coefficient of variation (CV)typically in excess of 25%, have severely limited theapplicability. In an attempt to reduce method imprecision,the International Council for Standardization in Haema-tology (ICSH) has proposed a manual reference methodbased on the determination of the reticulocyte to red cellratio (International Council for Standardization in Hae-matology, 1998). It is suggested that this method, whichstandardizes specimen collection, staining and countingprocedures, should be used to determine the accuracy ofautomated counting systems and to provide a means forinstrument calibration. The advent of analysis by flowcytometry has encouraged a re-evaluation of the role ofreticuloctye analysis. This is discussed further, in 'NewParameters', below.

Ferritin assay

Body iron in excess of requirements comprises ferritin andhaemosiderin. Most is in the form of ferritin, which iswater-soluble. Iron may be mobilized from both forms, butis more accessible from ferritin. Found chiefly in thecytoplasm of cells of the reticuloendothelial system, it wasonce thought that ferritin did not appear in plasma orextracellular fluid under normal conditions but develop-ment of a sensitive immunoradiometric technique byAddison et al. (1972) enabled the demonstration thatferritin is a constituent of all normal human serum. Incontrast to tissue ferritin, ferritin present in serum is lowin iron content and glycosylated. However, the concen-tration in serum may be directly related to body iron storesand the assay has become the most widely appliedevaluation of iron status. The original radioassay tech-niques have been largely superseded by automatedenzyme-labelled techniques.

In most normal adults, concentrations range from 15 to300 f.ig/\, but these vary with age and sex (Worwood,1982, 1997: Baynes, 1996). Serum ferritin concentra-tions are relatively high at birth. As fetal red cells areremoved from the circulation during the first few weeks oflife, the haemoglobin concentration falls to 10-11 g/dl. At

this time the rate of erythropoiesis is relatively low andiron released from the destruction of fetal cells is stored inthe tissues. This process is accompanied by an increase inserum ferritin concentration. Synthesis of adult haemo-globin commences during the following 2 months, caus-ing a rapid fall in both iron stores and the serum ferritinconcentration. Both remain relatively low throughoutchildhood and adolescence.

Serum ferritin levels are higher in men than in womenof childbearing age, refiecting difierences in storage iron.Levels in postmenopausal women are closer to thosefound in men. The concentration falls during pregnancy.The initial reduction is thought to reflect increasedmaternal erythropoiesis, whereas in the latter stagestransfer of iron to the fetus is the main cause. From35 weeks, serum ferritin concentrations are lower inwomen who have not received iron therapy than inthose who have.

Addison et al. (1972) found that patients with irondeficiency anaemia have serum ferritin levels approxi-mately one-tenth of normal subjects, while patients withiron overload (haemochromatosis, haemosiderosis) haveserum ferritin levels much higher than normal. Serumferritin levels may serve as a tool to monitor the effects ofiron therapy. However, raised concentrations must beinterpreted with caution. Ferritin behaves as an acutephase protein and in some circumstances the serum levelmay not illustrate the true state of iron stores sinceimmunoassay techniques refiect ferritin protein ratherthan iron. In both adults and children, chronic infiamma-tion results in a disproportionate increase in ferritin levels inrelation to iron reserves. Elevated ferritin levels, not directlyrelated to iron stores, are also observed in acute and chronicliver disease, chronic renal failure and in some types ofneoplastic disease (Baynes, 1996; Worwood, 1997).

Serum iron and total iron binding capacity

Measurement of the serum iron and total iron bindingcapacity (TIBC) can provide information relating totransport iron. Recommendations on methodology havebeen made by the ICSH (International Committee forStandardization in Haematology, 1978, 1990), Irondeficiency results in low serum iron concentrations,whereas high levels are found in iron overload conditions.Unfortunately this is not always the case. In the presenceof inflammation or infection the serum iron is lowirrespective of the level of storage iron present. Highconcentrations may also be encountered in liver disease,hypoplastic anaemias and in conditions where erythro-poiesis is inelTective (Worwood, 1997).

© 2001 Blackwell Science Ltd.. Clin. Lab. Haem.. 23. 271-283

H. M. Waters & L, H. Seal

Isolated sampling from the transport iron compartmentcan provide unreliable, if not misleading, information.Circulating iron, bound to transferrin, comprises only avery small part (0.01%) of the total body iron and has avery high turnover rate of 10-20 times per day in normalsubjects (Cavill etal., 1977). These factors contribute tothe variability of measurement encountered and severelylimit the diagnostic usefulness of an individual serum ironmeasurement. The TIBC is indicative of the circulatingtransferrin concentration and is a more stable measure-ment than the serum iron (Cavill, 1982). Measurement ofboth the serum iron and TIBC provides more informationthan individual measurements (Worwood, 1997), Trans-ferrin normally circulates about one-third saturated.Calculation of the percentage transferrin saturation fromthe serum iron and TIBC (serum iron/TIBC x 100) reflectsthe transport compartment but, for reasons oudinedabove, is also too variable as a single measurement(Baynes, 1996).

In iron deficiency, a low serum iron is associated with araised TIBC, resulting in a low percentage transferrinsaturation. A saturation level of 16% or less refiectsinadequate transport iron to sustain erythropoiesis. Preg-nancy and anaemia secondary to chronic disease are alsoassociated with low transferrin saturation despite thepresence of adequate iron stores. Iron overload results inhigh transferrin saturation levels due to the combinationof an elevated serum iron and a normal or depressed TIBC.

The only viable clinical role for serum iron and TIBC is inscreening for idiopathic haemochromatosis. The diagnosisof this condition, one of the most common inheritedconditions in populations of European origin, shouldinclude the demonstration of a raised transferrin saturation(Males > 60%, Females > 50%) on at least two bloodsamples (Worwood, 1997), In the early stages ofthe diseasethis abnormality will be encountered before excessive ironaccumulation and the resulting raised ferritin.

Zinc protoporphyrin

Zinc protoporphyrin (ZPP) is synthesized by developingerythrocytes when there is insulBcient iron available toferrochetalase for haem production. It can be quantifiedby spectroscopy and instruments are available for rapidanalysis. Values are standardized by expressing the resultas a ratio to haem. In healthy blood donors, a goodinverse correlation between serum ferritin and ZPP hasbeen found. Studies in a range of disorders suggest thatZPP is indicative of both absolute and functional irondeficiency and may substitute for percentage hypochro-mia where the instrumentation required for this analysis

© 2001 Blackwell Science Ltd.. CUn. hah. Haem.. 23, 271-283

is unavailable. However, ZPP does not predict a responseto iron supplementation in some renal patients. There aresome instances when fluorescent substances in samplesinterfere with the assay, giving falsely elevated results.Washing the erythrocytes before analysis can reduce thisbut some of the test's facility is lost (Braun, 1999),

Vitamin B^ and folate assays

Vitamin B12 and folate act as coenzymes in two importantmetabolic functions vital to normal cell growth and DNAsynthesis: (1) the synthesis of methionine, and (2) theconversion of methylmalonyl CoA to succinyl CoA (Weir &Scott, 1983; Chanarin et al, 1992; Hoflbrand & Jackson,1993). In the first of these, vitamin Bj2 deficiency interfereswith the regulation of folate metabolism, resulting inimpaired DNA synthesis and consequently megaloblasticanaemia. Because both vitamins are linked by the reactionpathway for methionine synthesis, a deficiency in eitherwill disrupt this and produce the same symptoms. How-ever, the causal mechanism remains a topic of somediscussion (Chanarin et al, 1992; Hoffbrand & Jackson,1993). Much ofthe research on this pathway resulted fromthe discovery that nitrous oxide inhibited thymidylatesynthesis through inactivation of vitamin B12 (Amesset al, 1975, 1978). In the second reaction, vitamin B12deficiency results in impaired fatty acid metabolism whichmay in turn lead to abnormal myeiin lipid formation,contributing to neurological lesions (Weir & Scott, 1983).

In general, the serum or plasma vitamin B]2 concentra-tion is related to the degree of tissue depletion, lowest levelsoccurring in the most severe cases (Chanarin, 1990; BritishCommittee for Standards in Haematology General Haema-tology Task Force, 1994a). Normal levels vary dependingon the method used and range between 120 and 1100 ng/l.Results falling within plus or minus 20% ofthe lower end ofthe reference interval should be regarded as Indeterminate',and the sample repeated (British Committee for Standards inHaematology General Haematology Task Force 1994a),Carmel (1999) suggests that ethnic and racial factorsshould be considered when investigating vitamin B12disorders and metabolism, and questions the adequacy ofreference ranges derived largely from white populations.Elevated vitamin B12 levels have been associated with theuse of oral contraceptives and multivitamins, and inmyeloproliferative diseases such as chronic granulocyticleukaemia and myelomonocytic leukaemia. An elevatedvitamin Bj 2 level in itself has not been known to causeclinical problems (Chanarin, 1979, 1990),

Folate levels in both serum and red cells are used toassess folate status. Normal serum concentrations range


between 3.0 and 20.0 fig/l In contrast, red cells havemuch higher folate levels, i.e. between 150 and 600 fig/l.A low serum folate is indicative of deficiency in theabsence of recent folate intake. Red cell folate is the bestindicator of long-term folate stores (Chanarin, 1979,1990). A low red cell folate value is suggestive of aprolonged folate deficiency. However, a consequence of thecommon metabolic pathway is that B12 deficiency disruptsthe uptake of folate into red cells. This can lead to a lowred cell folate value even with adequate folate intake. Forthis reason, it is often necessary to measure both vitaminsin a clinical workup. A case can be made for either theserum or red cell folate to be assayed but, because of thelimitations of each, it would seem reasonable to assay bothor to assay the red cell concentration whenever the serumfolate is low. The true incidence of folate deficiency is notwell known and most figures are derived from thefrequency of anaemia in pregnancy. Globally, malnutri-tion is the most common cause of both folate and vitaminBj2 deficiency (Dawson & Waters, 1994),

Both vitamins can be measured in serum by competitiveprotein binding assay. Most vitamin B12 methods useintrinsic factor (IF) as the specific binding protein,although radioimmunoassays have been described(O'Sullivan et al, 1992). ^-Lactoglobulin from cow's milkis the binding agent commonly used in folate assays.During the past 10 years, fully automated non-radioiso-topic systems have largely replaced radioisotopic manualmethods. In these systems, monoclonal antibodies areoften used to set up the conditions whereby the specificbinder is attached to a solid phase.

Intrinsic factor antibodies and assay

The demonstration of circulating intrinsic factor antibod-ies (IFAB) is virtually diagnostic of pernicious anaemia(Lindenbaum, 1983), They are rarely found in otherautoimmune diseases such as disorders of the thyroid,diabetes mellitus and myasthenia gravis. Two types ofIFAB are known to exist. Type I blocks the IF binding sitefor B12, preventing uptake of the vitamin, whereas type IIreacts with an alternative antigen site on the IF moleculeand may prevent attachment ofthe IF-B12 complex to theileal binding sites (Roitt et al, 1964; Chanarin, 1979).Traditionally, techniques of testing for IFABs have beenfunctional methods for the detection of type I antibody(Chanarin, 1979). Immunoassay methods have also beendescribed which are capable of detecting both type I andtype II IFAB (Conn, 1986; Waters et al, 1989), or type IIalone (Sourial, 1988). Subsequent studies using thismethodology have indicated that type II IFAB occur.

alone and in combination with type I, in a much higherproportion of pernicious anaemia samples than waspreviously recognized (Conn, 1986; Waters et al, 1993),Assay of gastric IF is a direct measurement of the defect inpernicious anaemia but this requires an invasive tech-nique and is rarely used.

Absorption studies

The urinary excretion method of Schilling (1953) remainsthe most commonly used and best standardized test forabsorption of free vitamin B12, The recommended proce-dure (International Committee for Standardization inHaematology, 1981) is based on the Schilling test andemploys a 1.0-jUg aqueous oral dose of cyanocobalaminradiolabeiled with cobalt-57. In this form, the test is carriedout as a two-part procedure, without IF (part I) and with IF(part II) added to the oral dose. Dual-labelled methods,based on the Schilling test, where the two stages arecombined by simultaneously administering two radiola-beiled versions of cyanocobalamin, one bound to IF, theother unbound, were developed in the 1960s (Katz, DiMase& Donaldson, 1963; Bell, Bridges & Nelson, 1965).Although the original adaptations held potential advan-tages over the two-stage approach, particularly in terms ofpatient attendance and sample collection, doubts have beenexpressed over a later commercial modification (Payne &Finney, 1972; McDonald, Barr & Barton, 1975; Fairbanks,Wahner&Phyliky, 1983;Zuckier&Chervu, 1984; Atrah&Davidson, 1989). This test has recently been withdrawnfrom the market.

Demonstration of normal absorption of free vitamin B12does not exclude malabsorption of the food-bound vita-min. In patients with a subnormal serum B]2. a normalSchilling test and satisfactory dietary intake, a food-boundabsorption test should be considered (British Committeefor Standards in Haematology General Haematology TaskForce, 1994a). Vitamin B12 deficiency due to food-boundmalabsorption does not reach the same degree of severityas that which may occur in pernicious anaemia becausethe enterohepatic circulation of the vitamin remains intactwhilst IF is available. The incidence of the condition isunknown but Carmel (1995, 1996) believes it to be atleast as common as pernicious anaemia and a major causeof vitamin B12 deficiency.

Deoxyuridine suppression test

The deoxyuridine suppression test can be used to confirm

mild vitamin B12 or folate deficiency, although in some

deficient patients the test can be normal. Bone marrow is

© 2001 Blackwell Science Ltd., Ctin. Lab. Haem., 23, 271-283

H. M. Waters & L. H. Seal

the recommended sample as opposed to peripheral bloodlymphocytes. The test is labour-intensive and expensive,and is only available in a few specialized centres (BritishCommittee for Standards in Haematology General Hae-matology Task Force, 1994a),

type of disorder the pathogenesis is typically morecomplex. The main situations where this arises are inthe ACD in rheumatoid patients, pregnancy and chronicrenal failure patients being treated with recombinanthuman erythropoietin (r-HuEPO),

Methylmaionic acid and homocysteine

Both methylmaionic acid (MMA) and homocysteine (Hey)are increased in the serum of patients with vitamin B12deficiency, whereas only Hey is increased in folatedeficiency. Measurement of MMA in a random urinesample can be used as a screening test, although the 24-hexcretion level is more sensitive. Serum is the sample ofchoice. These metabolite assays can be automated usingcapillary gas chromatography-mass spectrometry, highperformance liquid chromatography-mass spectrometry orimmunoassay, but the equipment required is expensiveand currently only available in a few specialized centres.However, new generation, integrated, systems are likely tobecome more available to a larger number of laboratories.Thin-layer chromatography is recommended as an inex-pensive but sensitive screen for urinary MMA (BritishCommittee for Standards in Haematology General Hae-matology Task Force, 1994a).

Guidelines

A clinical history and blood count with appropriatefollow-up tests usually leads to a rapid laboratorydiagnosis and protocols for investigating anaemia due tomegaloblastosis (British Committee for Standards inHaematology General Haematology Task Force, 1994a)or haemoglobinopathy (British Committee for Standardsin Haematology General Haematology Task Force, 1994b,1994c, 1998) are well established. The pathogenesis ofdisturbances of iron metabolism is more difBcult toclassify, and requires a clear distinction between storageand functional iron. Most of the newer generation testshave been developed to address this. Therefore theremainder of this article will concentrate on hypoprolif-erative variants around functional iron deficiency andanaemia of chronic disorders.

Problem areas

The mechanism underlying some anaemias is relativelysimple and the diagnosis straightforward, e.g. simple irondeficiency. In some situations a more detailed analysis oferythropoiesis may be necessary to ensure that anadequate response to treatment may be achieved. In this

© 2001 Blackwell Science Ltd.. CUn. Lab. Haem.. 23. 271-283

Anaemia of chronic disorders

Means (1999) has reviewed the pathophysiology of thissyndrome. Often confused with and sometimes wronglyreferred to as 'iron deficiency anaemias', ACD has a muchmore complex pathogenesis than simple iron-deficienterythropoiesis, ACD is associated with infiammatorystates, including chronic infection, rheumatoid arthritisand neoplasia. It is one of the most common disordersfound in medicine and accounts for a significant propor-tion of anaemia investigations. There is evidence that thecommon factor causing anaemia is the change in immuneand inflammatory mediators common to all of thesedisorders. Increased concentrations of tumour necrosisfactor, interleukin-1 and interferons have been implicatedand the mechanisms by which they disrupt erythropoiesisexplored. The effect on erythropoiesis is threefold. Ashortened red cell survival producing a slight increase inred cell production, impaired erythropoiesis due to reducederythropoietin production and reduced progenator sensi-tivity, and impaired release of reticuloendothelial iron. Theresulting hypoproliferative erythropoiesis has in the pastbeen characterized by a low serum iron and a normal orraised serum ferritin. This confiicting information hasmade it difficult to draw conclusions about true iron statusin relation to the anaemia. Recombinant erythropoietinhas already been used to correct the anaemia in somecases of ACD. Better measures of erythropoiesis will haveto be used to ensure adequate iron is available to supportthe erythropoietic drive and for monitoring treatmentresponse. A fundamentally different approach is thereforerequired to differentiate whether the defective red celloutput is due to simple iron deficiency, which is readilytreatable, or more complex mechanisms which do notrespond to iron supplementation.

Pregnancy

A number of physiological changes occur in pregnancy,including expansion of plasma volume, an EPO-drivenincrease in erythropoiesis to provide an expansion in RCMand increasing demands of the fetus for haematinics.These adaptations complicate the assessment of haemat-inic deficiency; for example, a reduction in haemoglobinconcentration cannot be relied upon as a first-line


screening test. Iron stores are challenged by normalpregnancy. Research, by tracking EPO concentration, hasshown that some benefit is likely to be gained by ironsupplementation (Milman et al, 1997), Better criteria aretherefore needed to provide a reliable indication of ironstores and functional iron in this situation. Using a studypopulation which included anaemia, chronic inflamma-tion and some thalassaemias, van den Broek et al. (1998)explored measures of iron status referenced against bonemarrow-stainable iron. They confirmed that the MCV isunreliable as a screen due to the erythropoietic drivecausing shortened transit times. Using a typical cut-offvalue of 1.2 ^mol/1, these authors showed ZPP to havevery poor specificity and that this could lead to significantover-diagnosis of iron deficiency. When referenced againstbone marrow iron they suggest that serum ferritin can beused as a reliable indicator of stores provided that thecut-off value is raised to 30 /ig/l. They also suggestspecificity can be further improved (to 89%) by usingmultiple parameters, e.g. by combining serum ferritin withserum transferrin receptor (sTfR) and C-reactive proteinmeasurements.

Similar questions have been raised about the signifi-cance of estimates of B12 and folate status in pregnancy(House et al, 2000). Reliability of single estimates isquestionable. It may be necessary to follow serumvitamin assays with biochemical indices (serum Heyand MMA) before advising supplementation. In anotherstudy, Pardo et al. (2000) confirmed that serum B12assay alone in non-anaemic patients may be unreliable.Hey and MMA values indicated there was no differencebetween a group with reduced B12 and one with normallevels. More extensive studies are needed to definevariability in all of these parameters in pregnancy beforea protocol can be recommended.

Renal failure

Decreased production of erythropoietin is the main causeof anaemia in chronic renal failure. Introduction ofr-HuEPO provided a major advance in the treatment ofthis form of anaemia. To achieve the optimum responsefrom r-HuEPO therapy, the supply of iron to sites oferythropoiesis must be adequate to maintain the result-ing increase in red cell mass. Failure to provide sufficientiron is considered to be the single most important reasonfor r-HuEPO hyporesponsiveness, as iron supply limits thestimulation rate of erythropoiesis (Horl et al, 1996;Cavill etal, 1997). A small proportion of patients mayrespond poorly to r-HuEPO for reasons other than lack ofiron. These may be infection, infiammatory disease.

malignancy, blood loss or hyperparathyroidism. Alumin-ium overload can produce a functional deficiency thatdoes not respond to iron therapy (Cavill et al, 1997;Tarng et al, 1999). Best practice guidelines on thestrategies for managing anaemia in renal failure areavailable (Cameron, 1999).

The iron status of patients on r-HuEPO therapy mayhave to be assessed by a variety of methods at differentstages of the treatment. Haemoglobin and ferritin con-centrations may be used to assess iron stores. Transportiron is reflected by transferrin saturation but is variable.An assessment of the erythron iron supply can be obtainedfrom the measurement of hypochromic cells (Horl et al,1996; Cavill et al, 1997). The serum ferritin level shouldbe measured prior to r-HuEPO treatment and, wherenecessary, iron supplementation should be provided untilthe ferritin concentration is raised to more than 200 ^ig/\(Cavill etal, 1997).

New parameters

Hypochromic cell analysis

Functional iron deficiency arises when the rate oferythropoiesis exceeds the supply of iron to the erythron,irrespective ofthe level of storage iron present (Cavill et al,1997). With the appropriate analyser it is possible todetermine the haemoglobin content of individual erythro-cytes. This measurement refiects the iron content ofcirculating red cells and provides a direct indication of irondelivery to the bone marrow (Macdougall et al. 1992).The emergence of a subpopulation of hypochromicerythrocytes is an indicator of lack of functional ironavailabifity to developing cells. Normally less than 2.5% oferythrocytes have a haemoglobin concentration of lessthan 28 g/dL. In studies on renal patients this cut-offvalue gave a sensitivity in detecting iron deficiency of 91%but a specificity of only 54%. The specificity can beincreased by raising the cut-off level but then sensitivityfalls (Cullen et al, 1999). Baumann Kurer et al (1995)used a cut-off of 11% in a study on iron deficiency inchronic infiammatory disease, giving a specificity of 90%but sensitivity was less than 77%. More studies arerequired on this parameter but at present availability is alimiting factor.

Reticulocyte analysis

Flow cytometry, in addition to providing information onreticulocyte numbers, also indicates the amount of RNApresent on a cell-to-cell basis and hence their maturity.

© 2001 Blackwell Science Ltd., Clin. Uib. Haem.. 23, 271-283

H. M. Waters & L. H. Sea/

This shows potential as a measure of erythropoiesis. Dueto the much larger number of cells analysed, flowcytometry has made this investigation viable when theproportion of reticulocytes is around 1% of the red cellpopulation, often where the interest is greatest. Somestandardization of methodology is still required and astandard material for method comparison is needed (Cavill& Rowan, 1996). From this new analysis, two parametersare available to investigate anaemia. The reticulocytecount, which indicates effective erythropoiesis, and theproportion of reticulocytes with a high RNA content,which indicates the intensity of reticulocyte stimulation(d'Onofrio et ah, 1996).

The clinical value of these parameters has been studiedin a wide range of haematological disorders. Whilst theyare useful, there is some question whether, for example,they give any better indication than white cell changes ina bone marrow transplant setting. Once reticulocytescould be isolated in flow cytometry, furtber analysis couldbe made on individual cells and population characteristicssucb as reticulocyte volume, reticulocyte haemoglobinconcentration and reticulocyte haemoglobin content couldbe measured. It would appear that reticulocyte haemoglo-bin content, in particular, shows promise as a sensitiveindicator of functional iron deficiency (Brugnara, 1998).

Cullen et al (1999) have shown that reticulocytehaemoglobin content may have better diagnostic meritthan hypochromic cell analysis. Tbese authors used a cut-off for reticulocyte haemoglobin content of < 26 pg in theirrenal study. This produced a sensitivity and specificity fordetecting iron deficiency of 100% and 73%, respectively.Brugnara et al. (1999) have also found reticulocytehaemoglobin content to be a strong predictor of irondeficiency in children. It may also be useful for treatmentmonitoring as early changes are seen in response totherapy. Where the appropriate instrumentation is avail-able this could be a valuable parameter for assessingeffective iron availability to developing erythrocytes.

Serutn transferrin receptor assay

Recent study of the control of iron metabolism at thecellular level, particularly the relationship between apo-ferritin and transferrin receptor (TfR) synthesis, hassignificantly improved understanding of iron metabolismin health and disease. Transferrin receptor molecules areeventually cleaved from the cell surface and form a freepool in plasma TfR (measured as sTfR). As erythropoieticcells have the highest concentration of receptors, they arethe main contributors to sTfR concentration. In negativeiron balance, serum ferritin concentration falls with iron

© 2001 Blackwell Science Ltd.. Clin. Uib. Haem., 23, 271-283

stores. No significant change occurs in sTfR concentrationuntil iron stores are exhausted or availability is restricted.At this point the sTfR concentration rises in response totbe lack of iron. Unlike serum ferritin, sTfR levels are notinfluenced by infection or infiammation and have there-fore been monitored in a wide range of anaemias includingthose with more complex pathogenesis, e.g. anaemias ofchronic disease. The assay is generally considered to be asensitive indicator of iron-deficient erythropoiesis which isnot affected by the acute phase response. Reference valuesand changes in sideropenia and anaemia have beenexplored but the parameter has not gained favour as astandalone measure of iron status. It has been suggestedthat combination with ferritin is more informative.

Punnonen, Irjala & Rajmaki (1997) have suggestedcalculation ofthe ratio sTfR/log ferritin (sTfR-F Index) as away of combining sTfR and ferritin results. They foundthis ratio was higher in iron-depleted patients with orwithout an accompanying infectious or infiammatorycondition compared with iron replete ACD patients andprovided a useful parameter for the identification ofpatients with depleted iron stores. Means et al. (1999)studied sTfR in patients with a wide range of anaemiaaetiologies. Their study was referenced against bonemarrow aspirate iron and found sTfR assay significantlymore sensitive but less specific than other iron statusassays in identifying the absence of stainable iron. Incomparison, serum ferritin alone was highly specific butinsensitive. Serum ferritin predicted aspirate iron storesaccurately when values were above or below the referencerange. They suggested a decision algorithm combiningserum ferritin and sTfR sequentially. sTfR should beassayed when serum ferritin falls in the range of 25-300 /ig/1. This improved sensitivity and specificity.Flowers & Cook (1999) used the ratio of sTfR/ferritinobtained from dried plasma spots to identify iron defici-ency. These authors suggest that dried plasma spots maybe used to study the prevalence of anaemia in epidemi-ological studies. However, they acknowledge that ACDpatients were not included in the study.

In a study using response to iron supplementation asthe main indicator, Suominen et al. (2000) comparedbone marrow iron level, sTfR concentration, serum ferritinand sTfR-F index as indicators of functional iron deficiencyin ACD patients. The sTfR concentration and sTfR-F indexwere botb more effective indicators of patients who wouldbenefit from supplementation tban either bone marrowiron or serum ferritin, STfR and sTfR-F index bothreturned to normal after supplementation. One patientwith zero bone marrow iron levels bad normal baselinevalues and did not respond to supplementation. These


authors concluded that sTfR and sTfR-F index are usefulparameters in detecting functional iron deficiency irres-pective of storage iron level. At present, tbe assay is notwidely used and it remains to be seen whether the sTfRassay, alone or in combination with the serum ferritin,gains wide acceptance.

Practice pressures

Non-invasive analyses, which describe different facets ofthe erythropoietic process, should be applied in anappropriate, sequential, manner to obtain maximuminformation. This investigative approach can be modifieddepending on whether a diagnosis is to be established orthe patient's condition or treatment are being monitored.Periodic re-assessment of existing procedures and evalu-ation of new, or modified, methods is necessary to improvethe investigation and understanding of erythropoiesis.Since so many disorders have anaemia as a complication,investigation of these anaemias exerts significant demandson pathology services. With new treatment optionsmaking it possible to improve the quality of life (Florencio,1981; Scholz et al, 1997; Silverberg et al, 2001), theinvestigation has become more complex, especially that ofthe secondary anaemias. However, current trends towardsnon-investigative profiling can work against this. As newtests for the investigation of anaemia were developed, thetraditional approach was to use tbese as follow-up testsand to treat them as 'specials' with particular attentionfrom the haematologist guiding the investigation toconclusion. This practice is changing and, as tests havebecome more readily available, clinical demand has drivenrequesting patterns. Now results are often reported directto the requesting clinician who will then make tbeir owninterpretation. The use (and abuse) of any test is directlyrelated to availability. Establishment of a method, withincreased confidence in the relevance of the results,usually leads to more requests, with a corresponding risein laboratory workload.

In some respects a non-investigative strategy may beviewed as a positive factor. From tbe clinicians viewpoint ascreening approacb where most laboratory testing iscarried out at tbe first visit, whether known at the timeto be relevant or not, is efficient. In primary care this resultsin fewer journeys to the general practice surgery for thepatient. In the outpatient clinic, or on the ward, fewerepisodes will be required, resulting in fewer appointmentsor faster bed turnover, respectively. However, the trans-ition to sucb a service also has drawbacks. In anuncontrolled service, relevant abnormal results will con-stitute an even smaller minority ofthe total produced, most

ofthe remainder being unnecessary to the well-being ofthepatient. Because time spent examining data is inverselyrelated to its volume there is a danger that tbe relevantmay be obscured by the irrelevant, making it even moredifficult to define the pathophysiology of anaemia (orany otber disorder) in individual cases. Production ofredundant information also bas cost implications.

Advances in instrumentation have resulted in automa-tion of many routine assays. Available systems fall intotwo main categories; 'closed' which restrict the user to tbeinstrument manufacturer's reagents and preclude devel-opment or running of in-house assays; and 'open' whichperform liquid handling robotic functions, allowing anyreagents to be used. Non-isotopic assay systems areavailable for use with random access automated analysers,offering the main advantages of increased throughput ofsamples and reduced turnaround times of patient results.The choice of system will depend upon the investigationscarried out together witb the circumstances witbin theindividual department. Adequate independent evaluationsof available systems are required, particularly for tbe largevolume measurements such as vitamin B12. folate andferritin. Precision for within-batch assay should be < 5%CV and for between-batcbes < 10% CV. The best precisionis usually required at the lower end of tbe referenceinterval. Accuracy should be controlled by the stan-dardization of calibrators against international standards[National Institute for Biological Standards and Control(NIBSC), Potters Bar. UK, http://www.nibsc.ac.uk].Reagent stability over a reasonable period is importantfor the laboratory carrying out small numbers of tests.Systems that can accept a variety of anticoagulatedspecimens as well as serum are beneficial.

Witb fully automated systems most errors are likely tobe limited to the pre- and postanalytical stages. These maybe dilutional, for example in the preparation of red celllysate, or transcription errors. Consideration thereforeneeds to be given to sample labelling and tracing features.The use of barcoded labels reduces tbe possibility oftranscription faults in addition to speeding up patient datainput. Adequate interfacing to a host computer shouldallow two-way transfer and matching of patient identifi-cation and results. Linkage to an electronic transfersystem reduces the chance of error at tbe request receiptand result delivery stages. Significant laboratory invest-ment, eitber in terms of equipment capital costs, reagentrental arrangements or electronic data processing andtransfer of results is required to maximize efficiency. Amultidisciplinary or multicentre approach may have to beadopted to guarantee a sufficiently high throughput ofsamples to ensure tbat associated costs are maintained at

© 2001 Blackwell Science Ltd.. Clin. lah. Eaem., 23, 271-283

H. M. Waters & L. H. Seal

an acceptable level and the full potential of the instru-

mentation is realized.

In the competitive world of automated analysers,

manufacturers must keep pace with service developments

by offering, and evolving, relevant test repertoires. As new

tests are automated, turnaround times for results gener-

ally decrease and this tends to be accompanied by a

further rise in workload. For example, this may yet prove

to be the case with sTfR estimations. Should more

manufacturers of automated immunoassay systems offer

this assay, the use will be facilitated and is likely to rise.

Developments in instrument software systems also allow

manufacturers to provide the option of linking results, for

example the sTfR to the serum ferritin level. This should

encourage the adoption of an algorithm approach such as

for a hypoproliferative-renal profile.

Clinical acceptability

Cinical governance quality initiatives include the devel-

opment of evidence-based clinical guidelines aimed at

agreement on local clinical activity, resulting in improved

patient outcomes (Department of Health, 1999). The non-

investigative approach outlined above cannot guarantee

this without the laboratory becoming more involved in the

interpretation of results and provision of advice on any

further testing required. Experience of ACD, particularly in

rheumatoid patients, and CRF patients being treated with

r-HuEPO, has led to a growth of literature on anaemia

pathogenesis in rheumatoid and renal journals. It has

been suggested that renal physicians are now challenging

the haematologists' understanding of iron metabolism

(Cavill, 1998). To standardize and optimize treatment of

renal disease in patients with anaemia, some renal centres

have developed anaemia management protocols (Senger,

Trenkle & St lohn, 1998). Specialist clinical or nursing

staff are also being employed to coordinate the assessment

and treatment of anaemia in their patients. These

initiatives should be welcomed but, wherever possible,

should be developed through collaboration between the

haematology department and the user unit, and based on

available guidelines.

References

Addison G.M., Beamish M.R., Hales C.N., Hodgkins M., Jacobs A.& Llewellin P. (1972) An immunoradiometric assay for ferritinin tlie serum of normal subjects and patients with iron defici-ency and iron overload. Journal of Clinical Pathology 25,326-329.

Amess J.A.L, Burman J.F., Greany M. & Nancekievili D.G. (1975)The eiTect of nitrous oxide on haematopoiesis. Proceedings ofthe

European African Division of the International Society of Haema-tology, Istanbul, 614.

Amess J.A.L., Burman J.F., Rees G.M., Nancekievill D.G. & MollinD.L. (1978) Megaloblastic haemopoiesis in patients receivingnitrous oxide. Lancet ii, 339-342.

Atrah H.L & Davidson R.J.L. (1989) A survey and critical eval-uation of a dual isotope (Dicopac) vitamin B12 absorption test.European Journal of Nuclear Medicine 15, 57-60.

Baumann Kurer S., Seifert B., Michel B., Ruegg R. & Fehr J.(1995) Prediction of iron deficiency in chronic inflammatoryrheumatic disease. British Jourtial of Haematology 91 (4).820-826.

Baynes R.D. (1996) Assessment of iron status. Clinical Biochem-istry 29 (3), 209-215.

Bell T.K., Bridges J.M. & Nelson M.G. (1965) Simultaneous freeand bound radioactive vitamin B12 urinary excretion test.Journal of CUnical Pathology 18, 611-613.

Braun J. (1999) Erythrocyte zinc protoporphyrin. Kidney Inter-national 55, S57-S60.

British Committee for Standards in Haematology General Hae-matology Task Force (1994a) Guidelines on the investigationand diagnosis of cobalamin and folate deficiencies. Clinical andLaboratory Haematology 16, 101-115.

British Committee for Standards in Haematology General Hae-matology Task Force (1994b) Guidelines for the fetal diagnosisof globin gene disorders. Journal of Clinical Pathology 47,199-204.

British Committee for Standards in Haematology General Hae-matology Task Force (1994c) Guidelines for the investigationof alpha and beta thalassaemia traits. Journal of ClinicalPathology 47, 289-295.

British Committee for Standards in Haematology General Hae-matology Task Force (1998) Guideline: the laboratory diag-nosis of haemoglobinopathies. British Journal of Haematology101, 783-792.

van den Broek N.R., Letsky E.A., White S.A. & Shenkin A. (1998)Iron status in pregnant women: which measurements arevalid.' British Journal of Haematology 103 (3), 817-824.

Brugnara C. (1998) Use of reticulocyte cellular indices in thediagnosis and treatment of haematological disorders. Interna-tional Journal of Clinical and Laboratory Research 28 (1), 1-11.

Brugnara C, Zurakowski D., DiCanzio J.. Boyd T. & Platt 0.(1999) Reticulocyte hemoglobin content to diagnose irondeficiency in children. Journal of American Medical Association281 (23). 2225-2230.

Cameron J.S. (1999) European best practice guidelines for themanagement of anaemia in patients with chronic renal failure.Nephrology, Dialysis, Transplantation 14 (Suppl. 2). 61-65.

Carmel R. (1995) Malabsorption of food cobalamin. In: Bailliere'sClinical Haematology, Megaloblastic Anaemia, Vol 8/3. (ed. S.N.Wickramasinghe), pp. 639-655. Bailliere Tindall, London.

Carmel R. (1996) Prevalence of undiagnosed pernicious anaemiain the elderly. Archives of Internal Medicine 156, 1097-1100.

Carmel R. (1999) Ethnic and racial factors in cobalamin meta-bolism and its disorders. Seminars in Hematology 36 (1).88-100.

Cavill I. (1982) Diagnostic methods. Clinics in Haematology 11/2,259-273.

Cavill I. (1997) Diagnosis of cobalamin deficiency: the old andthe new. British Journal of Haematology 99, 238-239.

' 2001 Blackwell Science Ltd.. Clin. Lab. Haem., 23, 271-283


Cavill I. (1998) Is the practice of haematology evidence-based.'Journal of Clinical Pathology 51, 944.

Cavill I. & Rowan M. (1996) Preface. Clinical and LaboratoryHaematology 18 (Suppl. 1). 1.

Cavill I., Macdougall I.C, Gokal R., Beguin Y., Peters T,, PippardM.. Schaefer R. & Winearls C. (1997) Iron management inpatients on rHuEPO. British Journal of Renal Medicine 2 (2).6-8.

Cavill I., Ricketts C, Napier J.A.F. & Jacobs A. (1977) Ferroki-netics and erythropoiesis in man: red cell production anddestruction in normal and anaemic subjects. British Journal ofHaematology 35, 33-40.

Chanarin I. (1979) The Megaloblastic Anaemias, 2nd edn. Black-well Publishing, Oxford.

Chanarin I. (1990) The Megaloblastic Anaemias, 3rd edn. Black-well Publishing. Oxford.

Chanarin I., Deacon R., Lumb M. & Perry J. (1992) Cobalaminand folate: recent developments. Journal of Clinical Pathology45, 277-283.

Conn D.A. (1986) Intrinsic factor antibody detection andquantitation. Medical Laboratory Sciences 43, 48-52.

Cullen P., Sofiker J.. Hopfli M., Bremner C, Schalaghecken R.,Mehrens T., Assmann G. & Schaefer R.M. (1999) Hypochromicred cells and reticulocyte haemoglobin content as markers ofiron-deficient er3^hropoiesis in patients undergoing chronic hae-modialysis. Nephrology, Dialysis, Transplantation 14-, 659-665.

Dawson D.W. & Waters H.M. (1994) Malnutrition: folate andcobalamin deficiency. British Journal of Biomedical Science 51,221-227.

Department of Health (1999) Clinical Governance. Quality in theNew NHS. Health Service Circular HSC/065. Department ofHealth. London.

Fairbanks V.F., Wahner H.W. & Phyliky R.L. (1983) Tests forpernicious anemia: the 'Schilling Test'. Mayo Clinic Proceedings58, 541-544.

Florencio C.A. (1981) Effects of iron and ascorbic acid supple-mentation on haemoglobin level and work efficiency ofanemic women. Journal of Occupational Medicine 23 (10).699-704.

Flowers C.H. & Cook J.D. (1999) Dried plasma spot measure-ments for ferritin and transferrin receptor for assessing ironstatus. Clinical Chemistry 45 (10), 1826-1832.

Hoflbrand A.V. & Jackson B.F.A. (1993) Correction of the DNAsynthesis defect in vitamin B12 deficiency by tetrahydrofolate:evidence in favour of the methyl-folate trap hypothesis as thecause of megaloblastic anaemia in vitamin Bi2 deficiency.British Journal of Haematology 83, 643-647.

Horl W.H.. Cavill I., Macdougall I.C. Schaefer R.M. & Sunder-Plassmann G. (1996) How to diagnose and correct irondeficiency during r-huEPO therapy - a consensus report.Nephrology, Dialysis, Transplantation 11, 246-250.

House J.D., March S.B., Ratnam S., Ives E., Brosnan J.T. & FrielJ.K. (2000) Folate and vitamin B12 status of women in New-foundland at their first prenatal visit. Canadian Medical Associ-ation Journal 162 (11), 1557-1559.

International Committee for Standardization in Haematology(1978) Recommendations for measurement of serum iron inhuman blood. British Journal of Haematology 38, 291-294.

International Committee for Standardization in Haematology(1981) Recommended methods for the measurement of vita-

min B]2 absorption. Journal of Nuclear Medicine 22,1091-1093.

International Committee for Standardization in Haematology(1990) Revised recommendations for measurement of serumiron in human blood. British Journal of Haematology 75,615-616.

International Council for Standardization in Haematology (1998)Proposed reference method for reticulocyte counting based onthe determination of the reticulocyte to red-cell ratio. Clinicaland Laboratory Haematology 20, 77-79.

Katz J.H.. DiMase J. & Donaldson Jr R.M. (1963) Simultaneousadministration of gastric juice-bound and free radioactivecyanocobalamin: rapid procedure for differentiating betweenintrinsic factor deficiency and other causes of vitamin B12malabsorption. Journal of Laboratory and Clinical Medtcitie 61 ,266-271.

Lindenbaum J. (1983) Status of laboratory testing in the diag-nosis of megaloblastic anemia. Blood 61 (4), 624-627.

Macdougall LC, Cavill I.. Hulme B., Bain B., McGregor E., McKayP., Sanders E.. Coles G.A. & Williams J.D. (1992) Detection offunctional iron deficiency during erythropoietin treatment: anew approach. British Medical Journal 304, 225-226.

McDonald J.W.. Barr R.M. & Barton W.B. (1975) SpuriousSchilling test results obtained with intrinsic factor enclosed incapsules. Annals of Internal Medicine 83, 827-829.

Means R.T. Jr (1999) Advances in the anemia of chronic disease.International Journal of Hematology 70, 7-12.

Means R.T. Jr, Allen J.. Sears D.A.. Schuster S.J. (1999) Serumsoluble transferrin receptor and the prediction of marrowaspirate iron results in a heterogeneous group of patients.Clinical and Laboratory Haematology 21 , 161-167.

Milman N., Graudal N., Nielsen O.J. & Agger A.O. (1997) Serumerythropoietin during normal pregnancy: relationship tohaemoglobin and iron status markers and impact of ironsupplementation in a longitudinal, placebo-controlled studyon 118 women. International Journal of Hematology 66 (2).159-168.

O'Sullivan J.J., Leeming R.J., Lynch S.S. & Pollock A. (1992)Radioimmunoassay that measures serum vitamin Bi2. Journalof Clmical Pathology 45, 328-331.

d'Onofrio G., Kuse R., Foures C, Jou J.M.. Pradella M. & Zini G.(1996) Reticulocytes in haematological disorders. Clinical andLaboratory Haematology 18 (Suppl. 1), 29-34.

Pardo J., Peled Y.. Bar J., Hod M., Sela B.A.. Rafael Z.B. & Orvieto R.(2000) Evaluation of low serum vitamin Bi2-in the non-anaemicpregnant patient. Human Reproduction 15 (1), 224-226.

Payne R.W. & Finney R.D. (1972) An evaluation ofthe doubleisotope test in the diagnosis of pernicious anaemia. ScottishMedical Journal 17, 359-363.

Punnonen K., Irjala K. & Rajamaki A. (1997) Serum transferrinreceptor and its ratio to serum ferritin in the diagnosis of irondeficiency. Blood 89 (3), 1052-1057.

Roitt LM., Doniach D. & Shapland C (1964) Intrinsic factor auto-antibodies. Lancet ii, 469-470.

Schilling R.F. (1953) Intrinsic factor studies II. The effect ofgastric juice on the urinary excretion of radioactivity after theoral administration of radioactive vitamin B12. Journal ofLaboratory and Clinical Medicine 42, 860-866.

Scholz B.D.. Gross R.. Schultink W. & Sastroamidjojo S. (1997)Anaemia is associated with reduced productivity of women

I 2001 Blackwell Science Ltd.. Clin. Lab. Haem., 23. 271-283

H. M, Waters & L H. Seal

workers even in Iess-physically-strenuous tasks. British Journalof Nutrition 77 (1), 47-57.

Senger J.M., Trenkle J.A. & St. John W.D. (1998) How to designan anemia management protocol. ANNA Journal 25 (2),235-240.

Silverberg D.S., laina A., Wexler D. & Blum M. (2001) Thepathological consequences of anaemia. Clinical and LaboratoryHaematology 2 3 , 1-6.

Sourial N.A. (1988) Rapid protein A assay for intrinsic factor andits binding antibody. Journal of Clinical Pathology 41 , 568-572.

Suominen P., Mottonen T., Rajamaki A. & Irjala K. (2000) Singlevalues of serum transferrin receptor and transferrin receptorferritin index can be used to detect true and functional irondeficiency in rheumatoid arthritis patients with anemia.Arthritis & Rheumatism 43 (5), 1016-1020.

Tarng D.C, Huang T.P., Chen T.W. & Yang W.C. (1999) Ery-thropoietin hyporesponsiveness: from iron deficiency to ironoverload. Kidney International 55 (Suppl. 69), S107-S118.

Waters H.M., Dawson D.W., Howarth J.E. & Geary CG. (1993)High incidence of type II autoantibodies in pernicious anaemia.Journal of Clinical Pathology 46 , 45 -47 .

Waters H.M., Smith C, Howarth J.E., Dawson D.W. & DelamoreLW. (1989) New enzyme immunoassay for detecting total,type I, and type II intrinsic factor antibodies. Journal of ClinicalPathology 42, 307-312.

Weir D.G. & Scott J.M. (1983) Interrelationships of folates andcobalamins. In: Nutrition in Haematology, Contemporary Issuesin Clinical Nutrition 5. (ed. J. Lindenbaum). pp. 121-141.Churchill Livingstone, New York.

Worwood M. (1982) Ferritin in human tissues and serum. Clinicsin Haematology 11 (2), 275-307.

Worwood M. (1997) The laboratory assessment of iron status -an update. Clinica Chimica Acta 259, 3-23.

Zuckier L.S. & Chervu L.R. (1984) Schilling evaluation of per-nicious anaemia: current status. Journal of Nuclear Medicine 25,1032-1039.

© 2001 Blackwell Science Ltd., Clin. Lab. Haem., 23, 271-283

r-1. a systematic approach to assesment of erythropoiesis

Documents