Mathematics and transfusion medicine - Blood Transfusion

EDITORIALMathematics and transfusion medicineRoberto Reverberi"La filosofia * è scritta in questo grandissimo libro che continuamente ci sta aperto innanzi a gli occhi(io dico l'universo), ma non si può intendere se prima non s'impara a intender la lingua, e conoscer icaratteri, ne' quali è scritto. Egli è scritto in lingua matematica …"("Philosophy * is written in this grand book – I mean the Universe – which stands continually open toour gaze, but it cannot be understood unless one first learns to comprehend the language and interpret thecharacters in which it is written. It is written in the language of mathematics …")(Galileo Galilei, Il Saggiatore, 1623)This issue of Blood Transfusion contains the first articleof a short series on mathematical aspects of transfusionmedicine.In more than three decades of professional activity, Ihave noticed that there are some topics of transfusionmedicine in which a mathematical treatment is important oreven essential.I am not referring to theoretical questions, but topractical applications: in fact, in the course of the series, Ihope the reader will appreciate that understanding themathematical aspects is invaluable as a guide to practicaldecisions.With this objective in mind, I have chosen a few topics,such as the kinetics of removal in exchange transfusionand therapeutic apheresis, the factors influencing theantigen-antibody reaction in immunohaematology, theconfidence limits of the leucocyte count in leucoreducedblood components.As my objective is essentially practical, I decidedto give readers the possibility to perform the calculationsthemselves, using common computer programmes, suchas Excel and the like.Therefore, each article is accompanied by instructionson how to programme a spreadsheet so as to obtain thedesired results.Although the instructions are, in most cases, elementary,my experience suggests that they will not be useless: thosewho are not particularly versed in mathematics orinformation sciences often get lost, when confronting suchproblems, and a certain uneasiness about these issues isvery common among doctors.In order not to discourage those readers, themathematical details are reduced to a minimum or confinedto within appendices.Lastly, this is an open series: interested readers arefree to propose new topics or, even better, to send theircontributions.* The "philosophy" that Galileo is talking about is the "philosophia naturalis",i.e. the study of nature.The same terminology was used by Newton in the title of his most famous book("Philosophiae naturalis principia mathematica", 1687), in which he also referredto mathematics.Blood Transfus 2007; 5: 49© SIMTI Servizi Srl49049_Editoriale_reverberi.p65 4905/07/2007, 10.16

REVIEWThe genetics of the Rhesus blood group system*Willy A. FlegelInstitut für Klinische Transfusionsmedizin und Immungenetik Ulm und Institut für Transfusionsmedizin,Universitätsklinikum Ulm, GermanyThe Rhesus factor is clinically the most importantprotein-based blood group system. With 49 antigens sofar described, it is the largest of all 29 blood group systems.The unusually large number of Rhesus antigens isattributable to its complex genetic basis. The antigens arelocated on two Rhesus proteins - RhD and RhCE - and areproduced by differences in their protein sequences. In CDnomenclature, they are termed CD240D and CD240CE.Unlike proteins of other blood groups, Rhesus proteinsare expressed only in the membranes of red blood cells andtheir immediate precursors 1 .Rhesus is second in its clinical importance only to theABO blood group. Since the introduction of postpartumanti-D prophylaxis in the late 1960s, and combined preandpostpartum anti-D prophylaxis in the early 1990s, theincidence of haemolytic disease in newborns due toalloimmunization has been reduced by more than 90%. Upto 1% of all pregnant women have clinically significantanti-erythrocyte antibodies 2,3 .Anti-D remains the main indication for phototherapyor exchange transfusions in newborns 2,4 , and pregnantwomen who are D negative show an above averageincidence.The five most important Rhesus antigens are the causeof most alloimmunizations following blood transfusion.According to the German haemotherapy guidelines[Richtlinien zur Gewinnung von Blut undBlutbestandteilen und zur Anwendung vonBlutprodukten] 5 , D negative transfusion recipients mustalways be given D negative erythrocyte products.Since 2000, women of reproductive age and girls havealso received transfusions compatible for further Rhesusantigens such as C, c, E and e in addition to the K antigenof the Kell blood group 5 .This procedure also applies to patients who receiveregular transfusions or have immunohaematological50problems, like anti-erythrocyte allo- and autoantibodies.In the case of autoantibodies, their exact specificity is notusually determined. Although one thirds of suchautoantibodies are directed at Rhesus proteins, this hasvirtually no practical consequences for treatment 1 .The D antigen, discovered in 1939, was the first Rhesusantigen to be described. D positive patients were termedRhesus-positive. In 1946, a quantitative variant with aweakly expressed D antigen was discovered and termed"D u ". This variant, now called "weak D", is of clinical anddiagnostic importance.Since 1953, is has been clear that there are alsoqualitative variants of the D antigen. Although patientswith this partial D variant are positive for the D antigen,they can also form anti-D.The genetic basisIn order to understand the genetic basis of diseases, itis important to understand individual differences in geneticvariability, as well as their frequency and distribution inthe population 6 . There is usually a close correlation betweenthe genotype and the expressed phenotype. Thus, takinga change in the RHD gene as an example, it is possible tomake inferences about the expression of the RhD proteinin the erythrocyte membrane. As is the case with many Dvariants, modified RhD protein can have importantimplications for transfusion related antigenicity.The molecular basis of the RH allelesThe first Rhesus gene, the RHCE gene, was discoveredin 1990. The RHD gene was found two years later, and thetotal deletion of this gene ascertained as the cause of theEuropean D negative phenotype.*Part of this review was presented by the Author during the XXXIX SIMTI Congress(Paestum, SA, 4-7 October, 2006)Blood Transfus 2007; 5: 50-57© SIMTI Servizi Srl050-57_flegel.p65 5005/07/2007, 10.34

Genetics of Rhesus systemMore than 170 alleles have been found on the RHDgene since. The site has still not been explored fully, even15 years after the first RH gene was cloned. DNB, thecommonest of all European partial D alleles, was describedas recently as 2002 7 .In 2002, comparisons between the Human GenomeProject and the Mammal Genome Project increasedunderstanding of the formation of the two RH genes onchromosome 1 (figure 1) 8 .Most mammals only have one RH gene, whose positioncorresponds to the human RHCE gene. The RHD genearose from the duplication of an ancestral RH gene duringmammalian evolution. An RHD deletion occurred 9 duringthe evolution of hominids, so that many modern humanscompletely lack the RHD gene. This haplotype (glossary)is the leading cause of the D negative phenotypeworldwide.The RH alleles can be grouped according to theirmolecular structure. For the most part, these groups showpoint mutations (SNP, single nucleotide polymorphisms)which cause missense, nonsense, frame shift or splice sitemutations (glossary). RHD-CE-D hybrid alleles are oftenformed by gene conversion.The examples of molecular changes and their effectson the D antigen (Table I) show how the D antigenphenotype correlates with the molecular structure.The molecular basis of the Rhesus phenotypesThe two Rhesus proteins, RhD and RhCE, are verysimilar, differing in only 36 of the 417 amino acids, whichthey each comprise. Each has twelve segments within theerythrocyte membrane and six extracellular loops (figure2). Both the amino (NH 3) and the carboxyl (COOH) terminalare located within the cell.D negative phenotypeThe clinically essential difference between Rhesuspositive and Rhesus negative hinges on the presence orTable I - Molecular changes in RHD alleles and their correlation with phenotypes of the D antigenClassification D antigen Molecular basis Representative example New rhesusof antigen phenotype Protein alteration Mechanism* Description of the Common antigenchange RHD allele namePartial D Qualitatively Amino acid substitution on Missense mutation RHD(G355S) DNB Unknownaltered on the external surfacehybrid protein: protein Gene conversion RHD-CE(3-6)-D DVI type 3 BARCsegment exchangeon the outer surfaceWeak D Quantitatively Substitution of amino Missense mutation RHD(V270G) Weak D type 1 Unknownattenuated acids in the membraneor extracellularlyDEL Quantitatively Strongly reduced Missense mutation RHD (M2951) in C De n/a** Unknownmarkedly translation or proteinat RHD (K409K) n/a**attenuated the site expression of splicingD negative D negative Absent protein Gene deletion RHD deletion D negative Impossibleexpression Nonsense mutation RHD (Y330X) n/a**Frame shift mutation RHD (488 del 4) n/a**Modifying gene Defect in theRHAG geneRh nullHybridprotein: Gene conversion RHD-CE(4-7)-Dexchange of proteinCde Ssegment on the externalsurfaceAntithetical Presence Missense mutation in Missense mutation RHCE allele: n/a** E versus eRHCE of antigen E amino acid position in amino acid position Ala 226 codes forprotein antigen or e 226 codes for antigen E 226 in RHCE antigen e, Pro 226codes for antigen E* see glossary; ** not assignedBlood Transfus 2007; 5: 50-5751050-57_flegel.p65 5105/07/2007, 10.34

Flegel W AFigure 1 - Duplication of the RH gene and deletion of the RHD gene. The ancestral conditionis shown as the RH gene locus in the mouse. The single RH gene is adjacent to thethree genes SMP1, P29-associated protein (P) and NPD014 (N). Duplication createda second, reversed RH gene in humans, which is located between N and SMP1. At theinsertion points before and after the RHD gene is a DNA segment about 9,000nucleotides or base pairs (bp) long. The two DNA segments flank the RHD gene andare termed the upstream or downstream Rhesus box. In the RHD positive haplotype,the RHD gene could be lost again through recombination (figure 3). The scale givesthe approximate length of 50,000 nucleotides in the genomic DNA.absence of the RhD protein in the erythrocyte membrane(D positive resp. D negative).It is unusual for erythrocyte or other cell proteins to belacking entirely in many humans. This particular geneticfeature contributes to the strong antigenicity of the RhDprotein. During duplication of the ancestral RH gene, twoDNA segments were formed, known as the Rhesus box(Figure 1) 9 . The RHD deletion resulted from an unequalcrossover (figure 3), which occurs when two DNAsegments are highly homologous, such as those of theRhesus box. The RHD negative haplotype commonestamong Europeans is characterized by a hybrid Rhesus box.Subtle molecular differences between the various forms ofthe Rhesus box are used for genetic testing.The molecular basis of D antigen variantsAside from lack of the RhD protein, the D negativephenotype is caused mainly by a series of changes in theRhD protein, which in turn change the phenotype of the Dantigen.Depending on the phenotype and their molecularstructure, these RHD alleles are classified as either partialD, weak D or DEL.Partial DThe RhD protein traverses the erythrocyte membraneseveral times, leaving only part of the protein exposed atthe surface (Figure 2). If an amino acid is substituted in aportion of the RhD protein which is located at the outersurface of the erythrocyte membrane, single epitopes ofthe D antigen can be lost or new antigens can be formed.DNB is the commonest European partial D (Table I).D categories are a subgroup of partial D. The structureof the RH gene site facilitates gene conversions (figure4) 10 . In the RHD gene some homologous exons of the RHCEgene will be inserted, forming a hybrid Rhesus allele whichexpresses a corresponding hybrid protein. This is how theD categories III to VI arose. The changes usually affect along string of amino acids, which is always located on theerythrocyte surface.52Blood Transfus 2007; 5: 50-57050-57_flegel.p65 5205/07/2007, 10.34

Genetics of Rhesus systemFigure 2 - The Rhesus protein in the erythrocyte membrane. Both Rhesus proteins show 417 amino acids, shown here as circles.Mature proteins in the membrane lack the first amino acid. The amino acid substitutions which distinguish the RhD fromthe RhCE protein are shown in yellow, with the four amino acids which code for the C antigen in green and the one whichcodes for the E antigen in black. The single amino acid substitutions which code for partial D are in blue, those which codefor weak D are in red. The mutations, identified by the Ulm group, are in light blue and orange.Figure 3 - Deletion of the RHD gene. Deletion of the RHD gene resulted fromrecombination between an upstream and a downstream Rhesus box on two differentchromosomes. This is termed an unequal crossover. When the two crossed strandsseparate (from A over the recombination site to B), the DNA at the RH gene sitecompletely lacks the RHD gene (C). This haplotype (C) occurs in about 41% ofthe population. An individual homozygous for this haplotype (about 17% are) isD-negative.Blood Transfus 2007; 5:50-5753050-57_flegel.p65 5305/07/2007, 10.34

Flegel W AFigure 4 - Category DVI as a result of gene conversion. The two RH genes lie on their chromosome pointing in opposite directions(i.e., a cluster). When the chromosome folds, the two RH genes are adjacent, now pointing in the same direction. Thisconfiguration allows gene conversion in cis, whereby a DNA segment is transferred from one gene to another. The middlesection of the RHD gene (yellow) is substituted by the corresponding homologous section of the RHCE gene (green) (A).This type of gene conversion is responsible for the RHD-CE(3-6)-D allele, which codes for the D category VI of themolecular type 3 (DVI type 3) (B). Exons 1 to 10 are drawn on both RH genes (C). Due to the contrary directions, theterminal exons of the two RH genes (RHD and RHCE exon 10) lie closest to each other. On the RHD gene, exons 3 to 6 aresubstituted by the homologous exons of the RHCE gene.Weak DIf an amino acid substitution is located within theerythrocyte membrane or the cytoplasm, this will result in aweak D phenotype (figure 2) 11 . Integration of the RhDprotein into the membrane will be hindered, leading toquantitative weakening of the D antigen. There is usuallyno qualitative change, and hence no anti-D immunization.The weak D type 1 is the commonest in Europe (Table I).DELA particularly weakly expressed D antigen is termedDEL (earlier Del), because it could only be demonstratedusing elution. In elution, antibodies are separated fromerythrocytes to demonstrate them in the eluate. Themolecular changes are more severe than those seen withweak D, considerably hindering but not completelypreventing integration into the cell membrane. All DELalleles are rare in Europe, but up to 30% of all apparently Dnegative individuals in East Asia are bearers of the DELallele RHD (K409K) 10,12 .The C/c and E/e antigensThe clinically important Rhesus antigens C, c, E and eare the result of RhCE protein changes at only five aminoacid locations (figure 2). Antigens are termed antithetical ifa protein can present only one of them. They are causedby protein polymorphisms. Often there are two variants ofa protein, which differ at only one amino acid location,such as the Rhesus antigens E and e. RHCE alleles showingthe amino acid proline at position 226 express the E antigen,whereas RHCE alleles showing the amino acid alanine atthis position express the e antigen (Table I) 1 . Similardifferences between two RHCE alleles account for theantithetical C and c antigens. The antigen pairs C/c andE/e are not antithetical, however, because they result fromsubstitutions at different locations. The four possiblecombinations occur at different frequencies (amongEuropeans: Ce > ce > cE > CE) and are inherited ashaplotypes.Clinical applicationsGenetic investigation, like all investigations in medicine,should only be carried out in the context of a clear aim 13 . Asfar as transfusions are concerned, molecular biologicaltechniques already are being used to provide cost-effectiveanswers to a number of clinically important questions.Methods used include polymerase chain reaction (PCR)for gene amplification and subsequent identification byelectrophoresis, nucleotide sequencing and hybridizationon biochips 14 .Anti-D in patientsThe clinical problems encountered are caused by a smallnumber of RHD alleles. Patients usually show partial D, insome rare cases weak D immunized by normal D antigen.Since category VI (DVI) is the most important of these, theAuthors recommend using monoclonal anti-D antibodiesfor typing, which do not react with DVI 15,16 .54Blood Transfus 2007; 5: 50-57050-57_flegel.p65 5405/07/2007, 10.34

Genetics of Rhesus systemThis procedure was included in the Germanhaemotherapy guidelines in 1996 and has not been changedsince. DVI carriers are therefore deliberately typed as falsenegatives to prevent transfusions with D positive bloodand likely anti-D immunization 17 .After these precautions were built into the Germanguidelines, they also were adopted by other Europeancountries. Unlike partial D, no anti-D alloimmunization hasyet been reported for weak D type 1, 2 or 3 18 . From a clinicalperspective, it is helpful that this involves the commonestD alleles, which make up almost 90% of all weak D types inGermany 19 , since these patients can receive D positiveblood transfusions and do not require D negative products.This procedure saves up to 5% of all D negativeerythrocyte products, since they can perfectly well bereplaced by D positive products 11 , thus avoidingbottlenecks in the supply of D negative blood products 20 .Pregnant women and anti-D prophylaxisPregnant women with weak D types 1 to 3 can also begiven D positive blood transfusions, and require no anti-Dprophylaxis. Each year, a one-off genetic test helps avoidrepeated administration of anti-D to 3,500 pregnant womenin Germany alone (up to 5% of all D negative pregnancies),and with it, all possible side effects of this prophylaxis,which these women do not require. Up to 5% of all anti-Dinjections are therefore unnecessary.A one-off genetic test is more cost-effective thanrepeated administration of anti-D products. In order toimplement this approach, the guidelines for medical careduring pregnancy and after birth (motherhood guidelines),issued by the Federal Committee of Physicians and HealthInsurance Funds [Bundesausschuss der Ärzte undKrankenkassen], would need to be adapted accordingly 21 .All pregnant women with rare weak D types would receivethe necessary prophylaxis, which they would notautomatically receive under the current state ofhaemotherapy 5 and motherhood guidelines 21 .A foetus can be shown to be D positive bydemonstrating foetal DNA in the plasma of peripheralmaternal blood 22 . Anti-D prophylaxis is unnecessary if thefoetus is D-negative. This could save about 40% of all theanti-D prophylaxis currently given during pregnancy. Thismethod was developed in countries bordering on Germany,where intensive efforts are under way to implement thisapproach to genetic diagnosis 23 .Prenatal diagnosisIf the foetus needs to be tested for D antigen,amniocentesis or sampling from the trophoblast is themethod of choice 14 . Cordocentesis is no longer performed.As already mentioned, maternal plasma may be able to beused in future.Having a child and anti-D antibodiesIf the father is heterozygous for the RHD deletion, thereis a 50% chance of the foetus being D-negative, in whichcase the pregnancy essentially free of any haematologicalrisk. If the father is homozygous for the RHD gene, thefoetus will definitely inherit the D antigen, which couldinfluence the couple's decision on whether to have a childor not.For several decades, it was impossible to determinewhether an individual is heterozygous or homozygous forRHD because serological methods are unsuitable. Withthe advent of the genetic diagnosis of the hybrid Rhesusbox, however, the possibilities have been expandedconsiderably 9 . If the father is D-positive, it is now sufficientto test him for the RHD deletion.Use in other diseasesIf the standard serological methods fail, geneticdiagnosis is the method of choice for a reliable blood grouptyping of patients after a transfusion and those with autooralloimmune haematological anaemias. Althoughtransfused leucocytes can under certain circumstancespersist for years, they will not interfere with routine geneticdiagnosis.Blood donorsAppropriate investigation for the RHD gene canidentify apparently D negative donors, who in reality areweak D or DEL, thus ensuring that their blood will be givenonly to D positive recipients 18 . Without genetic diagnostics,D negative blood transfusion recipients will continue to beimmunized by the D antigen contained in such blood 24-27 .Donors who so far were misidentified as being Dnegative and whose erythrocytes are D-/D+ chimeras cannow be identified correctly. Lifelong chimerism can resultfrom monochorionic twin pregnancies. Any transfusionfrom donor sources such as these can result in anti-Dimmunization, because they also contain several millilitresof erythrocytes with a perfectly normal D positivephenotype. This D positive blood can only be detectedusing genetic investigation, not with routine serologicalmethods 10,27 . Any case of anti-D immunization is ofconsiderable clinical importance for girls and women ofreproductive age. In the case of a D positive pregnancy,Blood Transfus 2007; 5:50-5755050-57_flegel.p65 5505/07/2007, 10.34

Flegel W Athis would be likely to result in Rhesus haemolytic diseaseof the newborn.The function of Rhesus proteinsMost blood group proteins have a known function.While purifying human Rhesus proteins, Americanphysician Peter Agre discovered a water transporterprotein 28 . This discovery earned him the 2003 Nobel Prizefor Chemistry. Despite intensive efforts, however, nofunction has been found for the RhD and RhCE proteins.Although the Rhesus associated antigen (RhAG), a Rhesushomologue contained in erythrocytes, can transportammonium ions 29 , the Rhesus proteins themselves couldnot be shown to have any such function. One possiblefunction under investigation involves the exchange of CO 2and even O 2. Other information on the RH alleles will onlybe gained from the everyday clinical application of geneticdiagnostics, which could thus contribute to identifyingtheir function.From the perspective of basic research, wheretransfusion medicine will continue to make a contribution,scientific work on Rhesus 30 and other blood groups hasbeen quite productive, and is anything but finished.OutlookGenetic diagnosis has been used for blood group typingin clinical transfusion medicine ever since 2000 31,32 . Asantenatal care has shown, genetic blood group typing hasled to a better quality of care, by helping to avoid potentialside effects and reducing costs. This is a rare combination,and justifies the extra costs involved in optimizing care viathe use of genetic diagnostic techniques. As well asimproving patient care, these methods can fuel thedevelopment of new methods 14 , which will also be used forhealth care outside of Germany. European departments oftransfusion medicine are leading the field in molecular bloodgroup diagnostics and applications, and will continue tocontribute to improving patient care.Common Genome Variability Terms 4,6SNP (single nucleotide polymorphism)Point mutation. Variability in a nucleotide sequence dueto change of a single nucleotide.AlleleThe expression of a coding or non coding nucleotidesequence (the exon resp. intron of a gene) with two ormore variants, often differing by only a point mutation.GenotypeA pair of alleles or variants of a nucleotide sequenceoccurring at homologous sites on paired chromosomes.HaplotypeA combination of alleles or variants of a nucleotidesequence located close together on the samechromosome, and usually inherited together.Missense mutationAmino acid substitution in a protein caused by a pointmutation. It can alter the function or antigenicity of aprotein.Nonsense mutationA stop codon caused by a point mutation whichprematurely stops synthesis of the amino acid chain,leading to loss of protein function of its expression.Silent mutationA point mutation which does not change the aminoacid at the site. Although the protein is unchanged, itstill can be associated with a clinically relevantphenotype and be used diagnostically.Frame shift mutationThe loss or insertion of one or two nucleotides whichshifts the reading frame and prematurely stops proteinsynthesis (or extends it in some rare cases), resulting inloss of protein function or expression.Splice site mutationA point mutation at a splice site (the exon-intronjunction), causing faulty splicing of messenger RNA(mRNA) and skipping an exon, thus changing the aminoacid sequence. Leads to loss of protein function orexpression.Gene conversionThe non-reciprocal exchange between two or morehomologous genes, whereby a certain nucleotidesequence on a gene is substituted by a sequence onanother gene, which is located on the same chromosome(conversion in cis).AcknowledgementsThis article is published by the courtesy of ChristopherBaethge, M.D., Editor in Chief of the journal DeutschesÄrzteblatt. This English translation was provided by thejournal Deutsches Ärzteblatt.Key Word: Rhesus, blood group, molecular diagnostic,transfusion, pregnancy.References1) Flegel WA, Wagner FF. Blutgruppen: Alloantigene aufErythrozyten. In: Mueller-Eckhardt C, Kiefel V, eds.:Transfusionsmedizin. Berlin, Berlin Springer; 2003: p. 145-85.2) Howard H, Martlew V, McFadyen I, et al. Consequences for56Blood Transfus 2007; 5: 50-57050-57_flegel.p65 5605/07/2007, 10.34

Genetics of Rhesus systemfetus and neonate of maternal red cell allo-immunisation.Arch Dis Child Fetal Neonatal Ed 1998; 78: F62–6.3) Filbey D, Hanson U,Wesstrom G. The prevalence of red cellantibodies in pregnancy correlated to the outcome of thenewborn: a 12 year study in central Sweden. Acta ObstetGynecol Scand 1995; 74: 687–92.4) Cheong YC, Goodrick J, Kyle PM, Soothill P. Managementof anti-Rhesus-D antibodies in pregnancy: a review from1994 to 1998. Fetal Diagn Ther 2001; 16: 294–8.5) Bundesärztekammer, Paul-Ehrlich-Institut. Richtlinien zurGewinnung von Blut und Blutbestandteilen und zurAnwendung von Blutprodukten (Hämotherapie) –Gesamtnovelle 2005. Bundesanzeiger 2005; 57(209a): 4–35.6) Cichon S, Freudenberg J, Propping P, Nöthen MM.Variabilität im menschlichen Genom. Dtsch Arztebl 2002;99: 3091–101.7) Wagner FF, Eicher NI, Jorgensen JR, et al. DNB: a partial Dwith anti-D frequent in Central Europe. Blood 2002; 100:2253–6. .8) Wagner FF, Flegel WA. RHCE represents the ancestral RHposition, while RHD is the duplicated gene. Blood 2002; 99:2272–3.9) Wagner FF, Flegel WA. RHD gene deletion occurred in theRhesus box. Blood 2000; 95: 3662–8.10) Wagner FF, Frohmajer A, Flegel WA. RHD positive haplotypesin D negative Europeans. BMC Genet 2001; 2: 10.11) Wagner FF, Gassner C, Müller TH, et al. Molecular basis ofweak D phenotypes. Blood 1999; 93: 385–93.12) Shao CP, Maas JH, Su YQ, et al. Molecular background ofRh D-positive, D-negative, D(el) and weak D phenotypes inChinese. Vox Sang 2002; 83: 156–61.13) Propping P. Genetische Diagnostik vor dem Hintergrund vonMillionen Polymorphismen. Dtsch Arztebl 2004; 101: 3100–1.14) Flegel WA,Wagner FF, Müller TH, Gassner C. Rh phenotypeprediction by DNA typing and its application to practice.Transfus Med 1998; 8: 281–302.15) Wagner FF, Kasulke D, Kerowgan M, Flegel WA. Frequenciesof the blood groups ABO, Rhesus, D category VI, Kell, andof clinically relevant high-frequency antigens in South-Western Germany. Infusionsther Transfusionsmed 1995; 22:285–90.16) Wagner FF, Gassner C, Müller TH, et al. Three molecularstructures cause Rhesus D category VI phenotypes with distinctimmunohematologic features. Blood 1998; 91: 2157–68.17) Lippert H-D, Flegel WA. Kommentar zumTransfusionsgesetz (TFG) und den Hämotherapie-Richtlinien. Heidelberg: Springer 2002.18) Flegel WA. How I manage donors and patients with a weakD phenotype. Curr Opin Hematol 2006; 13: 476–83.19) Wagner FF, Frohmajer A, Ladewig B, et al. Weak D allelesexpress distinct phenotypes. Blood 2000; 95: 2699–708.20) Garratty G. Do we need to be more concerned about weak Dantigens? Transfusion 2005; 45: 1547–51.21) Gemeinsamer Bundesausschuss. Richtlinien desBundesausschusses der Ärzte und Krankenkassen über dieärztliche Betreuung während der Schwangerschaft und nachder Entbindung. Bundesanzeiger 1986; 60 a (Beilage), zuletztgeändert: Bundesanzeiger 2003; 126: 14.906.22) Lo YMD, Hjelm NM, Fidler C, et al. Prenatal diagnosis offetal RhD status by molecular analysis of maternal plasma.N Engl J Med 1998; 339: 1734–8.23) Bianchi DW, Avent ND, Costa JM, van der Schoot CE.Noninvasive prenatal diagnosis of fetal Rhesus D: ready forprime(r) time. Obstet Gynecol 2005; 106: 841–4.24) Wagner T, Körmöczi GF, Buchta C, et al. Anti-Dimmunization by DEL red blood cells. Transfusion 2005;45: 520–6.25) Gassner C, Doescher A, Drnovsek TD, et al. Presence ofRHD in serologically D-, C/E+ individuals: a Europeanmulticenter study. Transfusion 2005; 45: 527–38.26) Yasuda H, Ohto H, Sakuma S, Ishikawa Y. Secondary anti-Dimmunization by DEL red blood cells. Transfusion 2005;45: 1581–4.27) Flegel WA. Homing in on D antigen immunogenicity.Transfusion 2005; 45: 466–8.28) Agre P, Saboori AM, Asimos A, Smith BL. Purification andpartial characterization of the Mr 30,000 integral membraneprotein associated with the erythrocyte Rh(D) antigen. JBiol Chem 1987; 262: 17497–503.29) Marini AM, Matassi G, Raynal V, et al. The human RhesusassociatedRhAG protein and a kidney homologue promoteammonium transport in yeast. Nat Genet 2000; 26: 341–4.30) Flegel WA. The Rhesus Site. DRK-Blutspendedienst Baden-Württemberg-Hessen. 1998–2007. www.uniulm. de/~wflegel/RH/31) Müller TH, Hallensleben M, Schunter F, Blasczyk R.Molekulargenetische Blutgruppendiagnostik. Dtsch Arztebl2001; 98: A317–22.32) Northoff H, Flegel WA. Genotyping and phenotyping: thetwo sides of one coin. Infusionsther Transfusionsmed 1999;26: 5.Correspondence: Prof. Dr. med. Willy A. FlegelInstitut für TransfusionsmedizinUniversitätsklinikum UlmHelmholtzstrasse 10 - 89081 Ulm - Germanye-mail: willy.flegel@uni-ulm.deBlood Transfus 2007; 5:50-5757050-57_flegel.p65 5705/07/2007, 10.34

REVIEWNew technologies in immunohaematologyFernanda Morelati 1 , Wilma Barcellini 2 , Maria Cristina Manera 1 , Cinzia Paccapelo 1 ,Nicoletta Revelli 1 , Maria Antonietta Villa 1 , Maurizio Marconi 11Centro Trasfusionale e di Immunoematologia, Dipartimento di Medicina Rigenerativa, Fondazione OspedaleMaggiore Policlinico, Mangiagalli e Regina Elena, Istituto di Ricovero e Cura a Carattere Scientifico, Milano, Italy2U.O. Ematologia 2, Dipartimento di Medicina e Medicina Specialistica, Fondazione Ospedale MaggiorePoliclinico, Mangiagalli e Regina Elena, Istituto di Ricovero e Cura a Carattere Scientifico, Milano, ItalyIntroductionSince the discovery of the ABO system, numerousimportant innovations have contributed to a continuous,rapid evolution in the diagnostic methods for in vitromeasurements of the antigen-antibody reaction, allowinga significant improvement in the compatibility betweenblood from donors and the recipients. Apart from theintroduction of ABO typing, these methods include thedetermination of Rh type and phenotype, the direct andindirect antiglobulin tests, cross-matching and consequentidentification of antigens and antibodies of clinicalrelevance, the use of low ionic strength additives andenzyme treatments, the development of monoclonalreagents and solid-phase and microcolumn platforms forperforming the pre-transfusion tests.Since transfusion safety depends on a series of strictlyinter-related processes 1 , among which pre-transfusion testshave a predominant role, in recent years some of the newtechnologies that integrate the classical techniques inimmunohaematology have become valid instruments forimproving the safety of transfusions. The aim of this reviewis to illustrate the principles and practical applications ofthese emerging techniques used in our laboratory to identifyantigens and antibodies, in cases of red cell or plateletimmunisation.Automation for complex casesThe most recent data in the literature 2 indicate that, stillnowadays, incorrect identification of samples and errors inperforming tests are the most frequent causes of transfusionreactions and complications, with sometimes dramaticconsequences 3 for health.The use of completely automated systems, indivisiblefrom the use of information technology, is the most efficientstrategy for achieving two main goals in the field ofimmunohaematology:58- reducing transfusion risks related to human errors, byautomating the stages related to identifying samples,selecting reagents, performing and interpreting resultsand transferring data to the laboratory's informationmanagement system;- guaranteeing the traceability of all the elementsinvolved in the analytic process, which can be archivedand remain accessible after the test has been performed.Following the 1990s the use of automated systemsincreased in all industrialised countries in parallel with thedevelopment and marketing of new technologies; thesesystems have increased the objectivity and stability of theresults and the standardisation of the process with respectto the traditional liquid phase methods.The most widely used systems are based on the use of:- microcolumns, with different types of commercialproducts, which enable the results to be seen after thepassage of red blood cells through a matrix containingthe reagents; the main advantage of this technology,which has led to its widespread use, is mainly related tothe fact that the antiglobulin test can be carried outwithout washing steps;- polystyrene microplates with wells pre-coated withlyophilised red bloods or platelets, or anti-erythrocyteor anti-platelet antibodies: the antibodies present arerevealed by immuno-adherence after addition of redblood cells coated with an anti-IgG human antiglobulin;a more recent system, based on the use of microplatessensitised by an anti-IgG human antiglobulin, enablesthe reaction to be visualised through magnetised redcells and for the antiglobulin test to be carried outwithout washing steps.The combined use of these techniques and latestgeneration, completely automated instruments hasenabled automation of even more sophisticatedimmunohaematology tests. These tests can be used inBlood Transfus 2007; 5: 58-65 DOI 10.2450/2007.0006-07© SIMTI Servizi Srl058-65_Morelati.p65 5805/07/2007, 10.39

New immunohaematological technologiesparticular conditions to resolve the most complex cases. Inour Centre, full automation has been efficiently applied inthe following conditions.1) Large-scale cell phenotypingA fully automated high output system based on solid -phase technology 4-7 is currently used for the red cellextended phenotype.The system enables typing of 14 red blood cell antigensof the greatest transfusional relevance, using samples ofblood in anticoagulated (EDTA) blood, processed within3-6 days of collection, and a combination of:a. polyclonal antisera (anti-Fy a , anti-Fy b , anti-Jk a , anti-Jk b ,anti-S, anti-s, anti-Co a , anti-Js b , anti-Lu b , anti-Kp bImmucor, Norcross, GA, USA) prepared for use with anautomated instrument and the solid phase method; theresults are confirmed using the same working conditionsand polyclonal antisera of the same specificitiesprepared for the test-tube method;b. plasma from immunised donors (anti-Ge2, anti-PP 1P k ,anti-U, anti-Vel), diluted 1:5 in saline and stored at+4 °C until use.The instrument processes samples in batches of 50-100, dispensing 12 samples, 7 typing reagents and 1negative control/sample for each plate.Over a period of 12 months, this procedure was used tocarry out 134,129 typings on 12,644 blood donors attendingthe 'Rare Blood Group Bank – Reference Centre, Region ofLombardy'. In 1% of the cases (1,339 typings) the resultwas not conclusive (indeterminate/doubtful/invalid) at thefirst test.The commercial antisera were the cause of inconclusiveresults in 156 (0.12%) typings and human plasma in 1,183(0.9%) typings.No inconclusive results were observed with anti-Fy b ,anti-K, and anti-k specificities. A high percentage ofrepetitions were required after the first test with the anti-Vel and anti-PP 1P k plasma samples (803 tests for anti-Vel,161 for anti-PP 1P k ), related to the peculiarity of the reactionsof the antibodies themselves. In 233 (0.17%) of these casesa manual method had to be used in order to identify theantigens.2) Identification of red blood cell antibodiesThe possibility of automating this complex process wasevaluated in a pilot study 8,9 carried out in 2004 using acompletely automated instrument based on standardcommercial panels and microcolumn technology.One of the most important difficulties in theidentification of the red cell antibodies was related to theantigen profile of the commercial panels, which was scarcelyuseful when mixtures of antibodies were present.In theses cases, further extensively typed red bloodcells are necessary to achieve complete identification ofthe specificities involved.Two new solid-phase panels 4-6, 10 , selected for Rhphenotype and also prepared for use in a completelyautomated instrument, recently became available. Thesewere evaluated for their performance in an automatedprocess when mixtures of red cell antibodies are present,that include also Rh specificities.Two 14-cell panels were used for this evaluation: thefirst panel consisted of homozygous cells for the C and Eantigens and the second comprised 13 cells negative forthe Rh(D) antigen and a control, Rh(D) positive cell.The panels were used to test samples from 61 nonimmunisedsubjects and 104 immunised subjects, who hadundergone complete immunohaematologicalinvestigations, prior to the evaluation.Among the subjects investigated, 75 had singleantibody (28 anti-D, 7 anti-CD, 2 anti-CDE, 26 anti-E, 1 antie,1 anti-C, 6 anti-c, 1 anti-K, 2 anti-Jk a , and 1 anti-M) whilethe other 29 patients had mixtures of antibodies (4 anti-D,20 anti-E, 3 anti-C, 8 anti-c, 5 anti-C w , 11 anti-K, 1 anti-Kp a ,3 anti-Jk a , 1 anti-Jk b , 2 anti-Fy a , 1 anti-Fy b , and 3 anti-S).The negative samples were evaluated with the two panelsand the positive samples with one of the panels accordingto the known specificity (57 with the first panel and 47 withthe second). The tests were carried out using the methoddefined by the instrument, which required interpretation ofthe results by the operator. In the case of discrepancy withthe previous result, the specificity involved was verifiedwith the manual methods used in the laboratory. Completecorrespondence with known results (Table I) was observedin non-immunised subjects and in 91 (87.5%) of the 104samples with red cell antibodies. In 11 of these, 12additional antibodies were found, of which 9 were identifiedwith the first panel (2 anti-c, 1 anti-E, 1 anti-Jk a , 1 anti-K, 1anti-C w , 1 anti-N and 2 autoantibodies reacting only in thesolid-phase) and three with the second panel (2 antibodiesagainst low incidence antigens and 1 autoantibody reactingonly in the solid-phase).Two class IgM antibodies were not detected by thefirst panel (1 anti-K) and by the second panel (1 anti-M).3) Selecting platelet concentrates for patientswith immunological refractorinessThe condition known as ‘immunological refractorinessBlood Transfus 2007; 5: 58-65 DOI 10.2450/2007.0006-0759058-65_Morelati.p65 5905/07/2007, 10.39

Morelati F et al.Table I - Results of the identification of red cell antibodies carried out with a completely automated instrumentType of sample Complete agreement Additional antibodies Antibodies not detected by(n.) detected by the instrument (n.) the instrument (n.)Immunised subjects 91 11 samples 2 IgM antibodies12 antibodies* (anti-K, anti-M)Non-immunised subjects 61 0 0(*) 2 anti-c, 1 anti-E, 1 anti-Jk a , 1 anti-K, 1 anti-C w , 1 anti-N, 3 autoantibodies, 2 antibodies against low incidence antigensto transfusion of standard platelet concentrates' is one ofthe most important complications in subjects requiringtransfusion support and indicates repeated, poorincrements in post-transfusion platelet count (threeconsecutive platelet counts below 5 x 10 9 /L). This condition,which can be associated with severe clinical complications,is caused by the presence of class I antibodies againsthuman leucocyte antigens (HLA) and is observed in 13-14% of patients with leukaemia transfused with standardblood components and in 3-4% of subjects transfused withleucocyte-depleted blood components. The traditionaltransfusion approach is based on choosing HLA identicalor compatible donors or on selecting appropriate donorsthrough tests of platelet compatibility. In our laboratorywe have chosen the latter approach by using the plateletconcentrates (from buffy-coats or obtained throughapheresis products) present in the daily inventory of theunits and a completely automated instrument based on thesolid-phase technology.Over a period of 33 months, post-transfusion plateletcount increments were evaluated in 40 refractory subjects(27 women, 13 men) transfused with platelet concentratesselected using this procedure 11,12 and the increments wererelated to known detrimental clinical factors and to the posttransfusionincrements (at 1 hour after the transfusion),observed after the last three transfusions with standardplatelet concentrates (not selected by platelet cross-match).Within 48 hours from starting the selection procedure, thesubjects under consideration had been transfused with569 platelet concentrates (median value 8 concentrates/patient, containing 202 ± 71 x 10 9 platelets), obtained frombuffy-coats or by apheresis procedures. The median pretransfusionplatelet count was 7.7 ± 5.5 10 9 /L and the posttransfusionplatelet increments exceeded 10,000 platelets/µL in 68% of the cases (Figure 1).The post-transfusion counts in subjects withdetrimental factors were lower (28.9 ± 20.3 x 10 3 platelets/µL at 1 hour after the end of the transfusion) than those60observed in subjects without such factors (35.9 ± 21.2 x10 3 platelets/µL at 1 hour after the end of the transfusion).The investigation of autoimmune haemolyticdisordersAmong the many red cell immunhaematology problems,one of the most difficult to manage is autoimmunehaemolytic disease with a negative direct antiglobulin test(DAT). In order to resolve the diagnostic problem in thesecases, a battery of investigations must be used, carried outwith different methods (agglutination tests, solid phasetests, ELISA, flow cytometry, immunoradiometric tests,evaluation of complement consumption). One particularlyuseful test in the study of these complications is the mitogenstimulation test (MS-DAT), designed by Barcellini 13, 14 andcolleagues, which is used to evaluate the prevalence ofpositive results in subjects with autoimmune haemolyticanaemia (AIHA) in clinical remission or in an active phaseof the disease. The MS-DAT test is carried out bystimulating whole blood cultures from the investigatedsubjects with mitogen (phytohaemagglutinin – PHA;phorbol-12-myristate-13-acetate – PMA; or pokeweedmitogen-PWM); the production of antibodies in the cultureafter stimulation is evaluated by a competitive ELISA insolid phase. An agglutination DAT, using the standardtest-tube method, a DAT with red cells washed in low ionicstrength solution or cold physiological saline, and a solidphase test for immunoadherence were carried out in parallelin all subjects 4-6 . Using this technique, 33 subjects withAIHA were studied (of whom 27 in an active phase ofdisease with a positive DAT, and 6 subjects with previousAIHA who had become DAT negative) and 7 subjects withDAT-negative AIHA, whose disease had been diagnosedon the basis of exclusion of all other causes of haemolysisand on the response to steroid therapy. Furthermore, westudied 69 subjects with chronic B-cell lymphocyticleukaemia (B-CLL), a disease associated with a highprevalence of autoimmunity against red blood cells, withBlood Transfus 2007; 5: 58-65 DOI 10.2450/2007.0006-07058-65_Morelati.p65 6005/07/2007, 10.39

PLT x 10 3 /µLNew immunohaematological technologies5045Transfusion with platelet concentrates from buffy coatsTransfusion with apheresis concenentrateses4035302520151050pre- pre- post-1 hourpost-1 hourpost-post-24 hours24 hourspre- pre- post-1 hourpost-1 hourpost-post –24 hours24 hoursTransfusion with standard platelet concentratesTransfusion with cross-match negative platelet concentratesFigure 1 - Pre- and post-transfusion platelet counts in 40 refractory patients (median, 25 th and 75 thpercentiles)Table II - The effect of mitogenic stimulation in the presence of antibodies against autologous red blood cells in whole bloodcultureNot stimulated PHA (#) PMA ($) PWM (&)Patients with AIHA 322±49* 623±122** 465±55*** 635±134**Patients with B-CLL 134±15** 207±29* 182±37** 183±25*Controls 75±7 75±9 70±6 76±14(#) Phytohaemagglutinin; ($) Phorbol-12-myristate-13-acetate; (&) Pokeweed mitogen; * p

Morelati F et al.Table III - Clinical and laboratory characteristics of patients with DAT-negative and MS-DAT positive AIHAN. S e x Hb g/dL Total bilirubin Reticulocytes Haptoglobin LDH Steorid DAT MS-DAT(indirect) µmol/L % mg/L U/L therapy (*) IgG ng/mL1 F 13.3 8 (5) 1.4 1,200 420 Yes neg 2792 F 12.4 13.7 (12) 0.8 2,290 296 Yes neg 3213 F 12.3 10.3 (7) 0.7 1,830 377 No neg 4334 M 13.1 10 (9) 1.8 1,200 472 No neg 3025 M 12 10 (8) 0.9 690 303 Yes neg 2566 M 15 13.7 (8) 0.2 870 291 No neg 3227 M 12 27 (22) 3.7 200 480 No neg 8138 F 10.9 28 (22) 4.4 200 420 No neg 4339 M 11.1 20 (18) 1.1 200 550 No neg 1,23010 F 5.3 20 (19) 6.9 200 232 No neg 1,66011 M 8.8 n.d. 4.3 200 520 Yes neg 51612 M 10.8 40 (27) 15.9 200 470 No neg 85613 F 11.1 9 (8) 0.9 500 338 No neg 314Normal Women 0-17 (0-12)

New immunohaematological technologies- the risk of haemolytic disease of the newborn,- the risk of neonatal alloimmune thrombocytopenia,- RHD zygosity;c) large scale typing of blood donors for red cell andplatelet antigens, when typing antisera are not availableor difficult to obtain.Molecular characterisation of antigens is essential intransfused immunised patients, in order to select compatibleunits of red blood cells and in pregnant women, in order todecide whether to administer RhD prophylaxis.Over the course of about 2 years, molecular techniqueswere used to study 28 blood donors (5 suspected ABOvariants, 5 discrepancies in Rh determination, 3discrepancies in the typing of other red cell antigens and15 donors negative for high incidence antigens) and 40patients sent for immunohaematological investigations (6with ABO discrepancies and 4 with Rh discrepancies in theagglutination techniques, 11 transfused subjects for bonemarrow transplantation or thalassaemia or malignancy, 10DAT-positive transfused subjects, 3 immunised pregnantwomen and their partners, whose offspring were suspectedto be affected by foetal or neonatal alloimmunethrombocytopenia, 1 foetus with suspected haemolyticdisease, 4 subjects with haemolytic transfusion reactionsand 1 with platelet-specific alloimmunisation). Moleculartyping was carried out using DNA, at a concentration of 5-10 µg, extracted by from peripheral blood in EDTA, usingthe salting out method 22 . The commercial kits used werebased on the Polymerase Chain Reaction Sequence-Specific Primers (PCR-SSP) method, and prepared accordingto international knowledge in the relevant field 23-42 . Thefindings were 3 ABO variants (A el), 4 D variants (1 DFR, 1Rh33, 1 Dweak type 1 and 1 Dweak type 5), 14 cases ofabsence of a high incidence antigen (3 k, 2 Lu b , 1 Co a , 7 Fynull, 1 HPA-1a); in another 6 cases (5 donors and 1 patient)genomic typing revealed the presence of antigens that theserological techniques had not detected (3 Fy b weak, 2 Lu band 1 antigen of the KEL system).A second study was carried out to identify donors ofplatelets negative for human platelet antigens (HPA).The anti-HPA alloantibodies and relative antigens wereinvolved in cases of post-transfusion purpura, inimmunological refractoriness to standard platelettransfusions and in cases of foeto-neonatal alloimmunethrombocytopenia. The availability of donors of knownplatelet type is essential in order to ensure effectivetransfusions of platelet concentrates in these subjects.However, it is difficult to identify such subjects because ofthe scarcity of specific typing antisera to use in the classicalmethods (in ELISA, in flow cytometry or in solid phase).Genomic DNA from 149 Caucasian, group O blood donorswas analysed in order to determine the genotype of theHPA-1a, HPA-1b, HPA-2a, HPA-2b, HPA-3a, HPA-3b, HPA-5a and HPA-5b antigens. The PCR-SSP method and acommercial kit were used. The allelic, phenotypic andgenotypic frequencies observed in this population of blooddonors were compatible with those reported in otherEuropean studies 43-48 (Table IV).ConclusionsIn this article we have presented our recent experiencewith some of the new technologies internationally appliedto the field of red cell and platelet immunohaematology 49-54to reduce the risk of incompatibility between donors andrecipients and, therefore, to improve transfusion safety.The use of automation, in particular, seems to be a validapproach for reducing the risk related to human error andguaranteeing the traceability of all the operative phases inall critical processes involving large numbers, such asextended red cell typing, identification of antibodies inmultiply immunised subjects and in the transfusion therapyof patients affected by immunological refractoriness tostandard platelet concentrates.The use of molecular techniques has becomeindispensable, in combination with agglutination methods,for defining the red cell antigen profile in recently transfusedand immunised subjects or in cases in which discrepanciescannot be resolved with traditional serological techniquesand in the consequent selection of compatible red bloodcell units. These techniques represent the only methodavailable for characterising the platelet antigen profile,which cannot be determined otherwise because of the lackof suitable typing reagents, and in antenatal investigationsto evaluate the risk of haemolytic disease of the newborn,the risk of alloimmune thrombocytopenia and the antigenprofile of the foetus. The mitogen stimulation test is aclinically important assay in the management of cases ofsuspected AIHA, that gives negative results in thetraditional test.Finally, it should be emphasised that transfusion safetydepends on a series of processes which must be improvedand monitored over time and that the use of newtechnologies is only one element in the transfusionprocedure.It is, therefore, essential to use new technologies withina carefully defined process including all the phases betweenselection of the donor, the transfusion of bloodcomponents and the follow-up of the patients.Blood Transfus 2007; 5: 58-65 DOI 10.2450/2007.0006-0763058-65_Morelati.p65 6309/07/2007, 9.20

Morelati F et al.Table IV - Results of HPA typing and a comparison with frequencies reported in other European countriesAllele frequency (%)HPA antigen Observed (1) Italy the Netherlands Finland Wales Germany Austria(Fratellanza 2005) (Simsek 1993) (Kekomaki 1995) (Sellers 1999) (Kiefel 1993) (Holensteiner 1995)HPA-1a 0.832 0.838 0.846 0.86 0.825 0.834 0.852HPA-1b 0.167 0.162 0.154 0.14 0.175 0.166 0.148HPA-2a 0.879 0.85 0.934 0.91 0.902 0.940 0.918HPA-2b 0.12 0.15 0.066 0.09 0.098 0.060 0.082HPA-3a 0.701 0.658 0.555 0.59 0.607 0.616 0.612HPA-3b 0.298 0.342 0.445 0.41 0.393 0.384 0.388HPA-5a 0.859 0.920 0.902 0.95 0.903 0.889 0.892HPA-5b 0.14 0.080 0.098 0.05 0.097 0.111 0.108(1) Centro Trasfusionale e di Immunoematologia, Dipartimento di Medicina Rigenerativa, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena,Istituto di Ricovero e Cura a Carattere Scientifico, Milano (Italy)References1) Dzik WH. New technology for transfusion safety. Br JHaematol 2007; 136: 181-90.2) Serious Hazards of Transfusion (SHOT), Report 2005;http://www.shotuk.org/index.htm3) Sazama K. Reports of 335 transfusion-associated deaths:1976 through 1985. Transfusion 1990; 30: 583-90.4) Plapp FV, Sinor LT, Rachel JM, et al. A solid-phase antibodyscreen. Am J Clin Pathol 1984; 82: 719-21.5) Beck ML, Rachel JM, Sinor LT, et al. Semiautomated solidphaseadherence assays for pretransfusion testing. Med LabSci 1984; 41:374-81.6) Rolih SD, Eisinger RW, Moheng MC, et al. Solid-phaseadherence assays: alternative to conventional blood banktests. Lab Med 1985; 16:766.7) Morelati F, Revelli N, Paccapelo C, et al. An automatedinstrument for mass red blood cell antigen typing. Transfusion2006; 46 (Suppl.): 125A (Abstract SP271)8) Morelati F, Paccapelo C, Villa MA, et al. Antibodyidentification tests performed by a software and a walkawayinstrument. Transfusion 2004; 44 (Suppl.):SP323F.9) Paccapelo C, Morelati F, Villa MA, et al. Identificazione dianticorpi eritrocitari con uno strumento automatico. XXXVIConvegno Nazionale di Studi di Medicina Trasfusionale dellaSocietà Italiana di Medicina Trasfusionale eImmunoematologia 2004 (Abstract COM 083).10) Morelati F, Paccapelo C, Revelli N, et al. The identificationof red cell antibodies by an automated system. Transfusion2006; 46(Suppl.):146A (Abstract SP341).11) Rachel JM, Sinor LT, Tawfik OW, et al. A solid-phase redcell adherence test for platelet cross-matching. Med Lab Sci1985; 42:194-5.12) Rebulla P, Morelati F, Revelli N, et al. Outcomes of anautomated procedure for the selection of effective plateletsfor patients refractory to random donors based on crossmatchinglocally available platelet products. Br J Haematol2004; 125: 83-9.13) Barcellini W, Clerici G, Montesano R, et al. In vitroquantification of anti-red blood cell antibody production inidiopathic autoimmune haemolytic anaemia: effect of mitogenand cytokine stimulation. Br J Haematol 2000; 111: 452-60.14) Barcellini W, Montesano R, Clerici G, et al. In vitro productionof anti-RBC antibodies and cytokines in chronic lymphocyticleukemia. Am J Hematol 2002; 71:177-83.15) Daniels G. Human Blood Groups (2nd ed.), Blackwell Science,Oxford (2002).16) Daniels G. Molecular blood grouping. Vox Sang 2004;87(Suppl. 1): S63–S66.17) Yamamoto F. Molecular genetics of ABO. Vox Sang 2000;78 (Suppl. 2): 91-103.18) Avent ND, Reid ME. The Rh blood group system: a review.Blood 2000; 95:375-87.19) Reid ME, Rios M, Yazdanbakhsh K. Application of molecularbiology techniques to transfusion medicine. Semin Hematol2003; 37:166-176.20) Cartron JP. Molecular basis of red cell protein antigendeficiencies. Vox Sang 2000; 78: 7-23.21) Denomme GA. The structure and function of the moleculesthat carry human red blood cell and platelet antigens. TranfMed Rev 2004; 18: 203-31.22) Miller SA, Dykes DD, Polesky HF. A simple salting outprocedure for extracting DNA from human nucleated cells.Nucleic Acid Res 1988; 16:1215.23) Chester MA, Olsson ML. The ABO blood group gene: alocus of considerable genetic diversity. Transfus Med Rev2001;15:177–200.24) Yamamoto F, McNeill PD, Hakomori S. Human histo-bloodgroup A2 transferase coded by A2 allele, one of the Asubtypes, is characterized by a single base deletion in thecoding sequence, which results in an additional domain at thecarboxyl terminal. Biochem Biophys Res Commun 1992;187: 366–74.64Blood Transfus 2007; 5: 58-65 DOI 10.2450/2007.0006-07058-65_Morelati.p65 6409/07/2007, 9.20

New immunohaematological technologies25) Wagner FF, Flegel WA. RHD gene deletion occurred in theRhesus box. Blood 2000; 95: 3662–8.26) Mouro I, Colin Y, Chérif-Zahar B, et al. Molecular geneticbasis of the human Rhesus blood group system. Nat Genet1993; 5: 62–5.27) Wagner FF, Gassner C, Müller TH, et al. Molecular basis ofweak D phenotypes. Blood 1999; 93:385–93.28) Westhoff CM. The Rh blood group system in review: a newface for the next decade. Transfusion 2004; 44:1663–73.29) Lee S. Molecular basis of Kell blood group phenotypes. VoxSang 1997; 73:1–11.30) Hadley TJ, Peiper SC. From malaria to chemokine receptor:the emerging physiologic role of the Duffy blood groupantigen. Blood 1997; 89: 3077–91.31) Tournamille C, Colin Y, Cartron JP, et al. Disruption of aGATA motif in the Duffy gene promoter abolishes erythroidgene expression in Duffy-negative individuals. Nat Genet1995; 10:224–8.32) Lo Y, Corbetta N, Chamberlain PF, et al. Presence of fetalDNA in maternal plasma and serum. Lancet 1997; 350:485–7.33) van der Schoot CE, Tax GH, Rijnders RJ, et al. Prenataltyping of Rh and Kell blood group system antigens: the edgeof a watershed. Transfus Med Rev 2003; 17: 31-44.34) Daniels G, Finning K, Martin P, et al. Fetal blood groupgenotyping from DNA from maternal plasma: an importantadvance in the management and prevention of haemolyticdisease of the fetus and newborn. Vox Sang 2004; 87: 225–32.35) van der Schoot CE, Soussan AA, Dee R, et al. Screening forfoetal RHD-genotype by plasma PCR in all D-negativepregnant women is feasible. Vox Sang 2004; 87(Suppl. 3): 9(Abstract).36) Finning KM, Martin PG, Soothill PW, et al. Prediction offetal D status from maternal plasma: introduction of a newnon-invasive fetal RHD genotyping service. Transfusion2002; 42:1079-85.37) DuPont BR, Grant SG, Oto SH, et al. Molecularcharacterization of glycophorin A transcripts in humanerythroid cell using RT-PCR, allele-specific restriction andsequencing. Vox Sang 1995; 68:121-9.38) Lucien N, Sidoux-Walter F, Olives B, et al. Characterizationof the gene encoding the human Kidd blood group-ureatransporter protein. Evidence for splice site mutations inJknull individuals. J Biol Chem 1998; 273:12793-80.39) Chartron JP, Ripoche P. Urea transport and Kidd bloodgroups. Transfus Clin Biol 1995;2:309-15.40) Olsson ML, Hansson C, Avent ND, et al. A clinicallyapplicable method for determination of the three major allelesat the Duffy (FY) blood group locus using PCR with allelespecificprimers. Transfusion 1998; 38 :168-73.41) Wagner FF, Frohmajer A, Flegel WA. RHD positivehaplotypes in D negative Europeans. BMC Genetic 2001;2:10.42) Storry JR, Charles-Pierre D, Rios M, et al. Use of DNAanalysis as an alternative donor screening strategy forDombrock blood group antigens. Transfusion 2002: 42(Suppl): 25s:(Abstract).43) Fratellanza G, Scarpato N, D'Agostino E, et al. Genomicrestriction fragment length polymorphism typing of humanplatelet-specific antigens in blood donors from southern Italyand in alloimmunised pregnancies. Blood Transfusion 2005;3: 222-30.44) Simsek S, Faber NM, Bleeker PM, et al. Determination ofhuman platelet antigen frequencies in the Dutch populationby immunophenotyping and DNA (allele-specific restrictionenzyme) analysis. Blood 1993; 81: 835-40.45) Kekomaki S, Partanen J, Kekomaki R. Platelet alloantigensHPA-1, -2, -3, -5 and -6b in Finns. Trans Medicine 1995;5:193-8.46) Sellers J, Thomposn J, Guttridge MG, et al. Human plateletantigens: typing by PCR using sequence-specific primersand their distribution in blood donors resident in Wales. EurJ Immunogenet 1999; 26: 393-7.47) Kiefel V, Kroll H, Bonnert J, et al. Platelet alloantigensfrequencies in Caucasians: a serological study. TransfMedicine 1993; 3: 237-42.48) Holensteiner A, Walchshofer S, Adler A, et al. Human plateletantigen gene frequencies in the Austrian population.Haemostasis 1995; 25:133-6.49) Olsson ML. New developments in immunohematology. VoxSang 2004; 87 (Suppl. 2): 566-71.50) Petrik J: Microarray technology. The future of blood testing?Vox Sang 2001; 80:1-11.51) van Drunen J, Beckers EAM, Sintnicolaas K, et al. Rapidgenotyping of blood group systems using the Pyrosequencingtechnique. Vox Sang 2002; 83:104 (Abstract 315).52) Van der Shoot CE. Molecular diagnostics inimmunohaematology. Vox Sang 2004; 87(Suppl.2):S189-S192.53) Reid ME. Applications of DNA-based assays in blood groupantigen and antibody identification. Transfusion 2003;43:1748-57.54) Stowell C, Dzik W. Emerging Technologies in TransfusionMedicine. AABB Press 2003.Received: 28 February 2007 – Accepted: 24 April 2007Correspondence: Fernanda MorelatiCentro Trasfusionale e di Immunoematologia, Dipartimento di MedicinaRigenerativa - Fondazione Ospedale Maggiore Policlinico, Mangiagalli eRegina Elena, Istituto di Ricovero e Cura a Carattere ScientificoVia Francesco Sforza, 3520122 Milano - Italye-mail: f.morelati@policlinico.mi.itBlood Transfus 2007; 5: 58-65 DOI 10.2450/2007.0006-0765058-65_Morelati.p65 6509/07/2007, 9.20

ORIGINAL ARTICLEThe first data from the haemovigilance system in ItalyAdele Giampaolo, Vanessa Piccinini, Liviana Catalano, Francesca Abbonizio,Hamisa Jane HassanReparto di Metodologie Transfusionali, Dipartimento di Ematologia, Oncologia e Medicina Molecolare,Istituto Superiore di Sanità, Roma, ItalyBackground. Haemovigilance is defined as the surveillance of adverse reactions occurringin donors and in recipients of blood components and as epidemiological surveillance of donors.The ultimate purpose of haemovigilance is to prevent the repetition of adverse events and reactions.Since the 2002/98/EC Directive came into force, the introduction of haemovigilance systemshas become a priority for all countries in the European Community. The Italian haemovigilancesystem is essentially in line with the Directive, although it does not include surveillance ofadverse events in donors and does not have a national level of registration of severe incidentsconnected with the collection, processing and storage of blood and blood components.Epidemiological surveillance of donors has been performed nationally since 1989 for HIV andsince 1999 for HBV, HCV and Treponema pallidum. Surveillance of adverse events in recipientswas started at the end of 2004.Materials and methods. The national form proposed for notifying adverse reactions (PETRA)was prepared by the National Institute of Health and distributed to all Transfusion Structures.Results. The data collected (adverse reactions, errors, and near miss errors) came from 21.0%of the Transfusion Structures in 2004 and 38.4% in 2005. The system monitored 49.6 % of allthe units distributed in Italy. Overall 1,495 adverse reactions were reported, which is equivalentto 0.8 reactions/1,000 units of blood components distributed. There were 16 reports of errorsinvolving transfusions to the wrong patient. Not all the Transfusion Structures sent their datausing the PETRA form. From the 986 PETRA forms received, it was possible to analyse therelevance of the transfusion, the outcome of the patient, the type of blood component involved,the type of error and the type of near miss error.Conclusions. This study is the first Italian report on transfusion errors and adverse reactions.Key words: haemovigilance, transfusion, PETRA, adverse reactions, errors, near miss errors.IntroductionThe clinical risk of transfusions is perceivedpredominantly as the risk of acquiring infectiousdiseases. In reality, over the last 20 years, the incidenceof transfusion-related transmission of diseases hasdecreased significantly, thanks to ever greater attentiongiven to the stages of selecting donors and screeningthe units collected. The real transfusion process,mostly carried out in hospital wards and operatingtheatres, tends to be less considered, but now needsto be monitored to increase the safety of the wholeprocess. Errors related to the identification of thepatient, of the sample test-tube and of the bloodcomponent expose patients to risk and, in some cases,increase the risk of mortality.From the monitoring of the adverse reactions dueto transfusions, reported by countries in which ahaemovigilance system has been active for some time,66Blood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0007-07© SIMTI Servizi Srl066-074_hassan.p65 6605/07/2007, 10.44

Haemovigilance in Italyit can be deduced that immunological adverse events,transfusion-related acute lung injury (TRALI) anderrors in the transfusion process are much more likelythan infections transmitted by transfusion of bloodcomponents 1,2 .The ultimate purpose of haemovigilance, definedas the surveillance of unexpected or adverse reactionsin donors and recipients and as epidemiologicalsurveillance in donors, is to prevent the repetition ofadverse events and reactions 3 . In fact, the informationobtained from haemovigilance systems can contributeto improving the safety of blood collection andtransfusion by: a) supplying the medical communitywith a valid source of information about the risksrelated to transfusion; b) indicating correctivemeasures to prevent the repetition of some accidentsor dysfunctions of the transfusion process, includingparticularly significant ones, such as samples takenfrom the wrong person, mistaken identification of thesample, errors in the request form, and bloodtransfused to the wrong person; c) alerting the hospitalwards and Transfusion Structures (TS) about adverseevents that could involve several patients, such asthose related to the transmission of infectious diseasesand to the collection and processing of the blood.In Europe, the first haemovigilance systems wereactivated in France in 1994 4 and in the UnitedKingdom in 1996 2 , although these two systems differgreatly. The former is obligatory and requiresnotification of adverse events of all degrees ofseverity; notification of near miss errors (that is, errorsrecognised before the transfusion) is voluntary. Thelatter system is voluntary and collects information onsevere adverse reactions, transfusion errors and nearmiss errors. Since Directive 2002/98/EC came intoforce, the introduction of haemovigilance systems hasbecome a priority for all countries in the EuropeanCommunity.In Italy, the surveillance of adverse events inrecipients was activated at the end of 2004 by theNational Institute of Health [Istituto Superiore diSanità - ISS] 5 . There were already systems formonitoring adverse reactions in some Regions 6 ; at anational level, efforts were made to guarantee thehomogeneity of the data collected by using the sameforms. The proposed form for national suveillance wasdesigned by Transfusion Medicine specialists andagreed upon through consultatory meetings withrepresentatives of the Regional Centres of Coordinationand Compensation [Centri Regionali diCoordinamento and Compensazione – CRCC].Subsequently, dedicated software was developed,based on the paper form. This software, called PETRA(Programma degli Errori Transfusionali e delleReazioni Avverse – Programme for Transfusion Errorsand Adverse Reactions), was distributed by the ISSto all TS. Participation in the haemovigilance systemwas not obligatory, but was strongly recommendedby the institutions (ISS and CRCC).The assimilation of Directive 2002/98/EC of theEuropean Parliament and Council of Europe, during2005, has made the notification of severe unwantedreactions and incidents obligatory: "Whatever severeincident, whether due to an accident or an error, relatedto the collection, control, processing, storage,distribution or assignation of blood or bloodcomponents, which could influence their quality andsafety, as well as any severe unwanted reactionobserved during or after a transfusion which could berelated to the quality and safety of the blood and itscomponents, or to human error, is notified to the regionor autonomous province involved, which, in its turn,notifies the ISS" (article 13 in 7 ).The Italian system of haemovigilance issubstantially in line with the European Directive,although it lacks the surveillance of adverse orunexpected events in donors and registration at anational level of severe incidents related to thecollection, processing and storage of blood and bloodcomponents, which could have effects on the qualityand safety of the blood component. In the future,changes should be made to the current nationalhaemovigilance system in order to render it conformwith the new requisites. Epidemiological surveillanceof donors was started at a national level in 1989 forHIV and in 1999 for HBV, HCV and Treponemapallidum. This monitoring enables estimations ofregional and national prevalences and incidences ofthe major transfusion-transmissible infections and alsoenables evaluation of the residual risk of suchinfections 8-10 .The collection and analysis of data on undesiredeffects of transfusion rely on a close collaborationbetween the TS, which supply the blood components,and the hospital wards. This collaboration is essential,in order to ensure complete investigations of everyunfavourable event. The Committee for the Good Useof Blood, by involving all the professional figuresBlood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-0767066-074_hassan.p65 6705/07/2007, 10.44

Giampaolo A et al."dealing with blood", could represent the context inwhich to spread the culture of haemovigilance, makingcollaboration between TS and hospital wards morepossible 11 .MethodsThe collection of haemovigilance data is theresponsibility of the 326 Italian TS distributedthroughout the country and located within hospitals 12 .Inside the TS, blood is donated, tested and processedto produce the blood components, that are then sentto the wards; the TS must then receive, from the doctorwho uses the transfusion therapy, documentation onevery transfusion and any adverse reactions (article15, paragraph 3 in 13 ).The TS store and manage the information, fillingin the computerised PETRA form for every caseidentified. The records are periodically sent to theCRCC, which then transmit the regional data to theISS. The data collected can be used at local, regional,national and international levels.The haemovigilance system calls for theregistration of immediate transfusion reactions(haemolysis, TRALI, bacterial contamination,anaphylactic shock, etc.), late effects [haemolysis,graft-versus-host disease (GvHD), post-transfusionalpurpura, etc.] and transfusion of wrong bloodcomponents.The system is also designed to collect informationon near miss errors, that is, mistakes recognised beforethe transfusion, which could have led to thetransfusion of a mistaken blood group or theregistration, collection or management of a wrong,inappropriate or unusable blood component.The information required by the PETRAnotification form are those considered minimum byRecommendation R(95)15 14 ; that is: (a) date of birth,sex and identification code of the patient transfused;(b) number of the units and identification codes ofthe blood components involved in the adverse event;(c) description of the type of blood component, itsmethod of preparation and the conditions and durationof storage of the blood component prior to itstransfusion; (d) severity of the event, reportedaccording to a graded scale (mild symptoms, longtermmorbidity, immediate threat to life, death); (e)imputability, that is, the relation between theunfavourable effects observed and the bloodcomponent transfused, using a graded scale.The PETRA forms sent by the Region of Piemonteare filled in electronically, interfacing the PETRAsoftware with that of the "Form for recording adverseevents of transfusion therapy", used by the Region.The incidence of adverse events was calculatedbased on the number of units of blood componentsdistributed, which is monitored by the National andRegional Register of Blood and Plasma 15 .ResultsSurvey of adverse events in 2005In 2005, the percentage of the 326 Italian TS thatparticipated in the haemovigilance survey was 38.4%,which was almost double that in 2004 (21.0%). Thispercentage also includes the TS that did not usePETRA and those that stated the absence of adversereactions.Table I reports the percentage participation in2004-2005. The regions in which all the TSparticipated in 2005 survey were Friuli-VeneziaGiulia, Liguria, Lombardia, the Marche, Piemonte,the Autonomous Province of Bolzano and Valled'Aosta; the coverage in military TS was 100%. Therate of responses was high in Lazio and Veneto.Considering the number of units of bloodcomponents distributed in 2005 by the TS, whichwere included in survey (1,834,474), the systemmonitored 49.6% of the units distributed in Italy(3,701,724) (Figure 1).In 2005, there were reports of 1,495 adversereactions, 823 of which were reported using asummary data-sheet other than PETRA, such that, inmost cases of events notified, a description of thecausative role of the transfusion in the adverse reactionwas not given.Overall, 0.8 reactions were reported every 1,000units of blood components distributed.Almost all the adverse events reported were acute:46.9% of the reactions were febrile type reactions and38.7% were of an allergic nature (urticaria andanaphylactic reactions). Only six delayed type adversereactions were reported: four late haemolyticreactions, one case of GvHD and one HIV infection(Figure 2).The collection of data carried out with differentmethods by the structures that participated in thehaemovigilance system in 2005 was not homogenousand did not enable further analysis.68Blood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-07066-074_hassan.p65 6805/07/2007, 10.44

Haemovigilance in ItalyTable I - Participating Transfusion Structures2004 (%) 2005 (%)Abruzzo 7.7 0.0Basilicata 0.0 0.0Calabria 0.0 0.0Campania 0.0 4.5Emilia Romagna 0.0 23.1Friuli Venezia Giulia 0.0 100.0Lazio 16.7 58.3Liguria 90.0 100.0Lombardia 100.0 100.0Marche 100.0 100.0Molise 0.0 0.0Piemonte 0.0 100.0Puglia 8.0 4.0Sardegna 7.7 7.7Sicilia 36.4 33.3Toscana 0.0 0.0Autonomous Province of Bolzano 25.0 100.0Autonomous Province of Trento 0.0 0.0Umbria 0.0 0.0Valle d’Aosta 100.0 100.0Veneto 31.6 63.2Military structures 0.0 100.0100 %75 - 50%64%26%50 - 25 %

Giampaolo A et al.late haemolysis0.3%fever46.9%urticaria36.5%HIV0.1%Volume overload2.1%GVHD0.1%TRALI0.1%other4.1%shivers4.6%nausea, vomiting2.1%anaphylaxis2.2%acute haemolysis0.9%Figure 2 - Adverse reactions 2005Table II - PETRA forms receivedAbruzzo 6Campania 4Emilia Romagna 65Friuli Venezia Giulia 13Lazio 25Liguria 18Marche 57Piemonte 483*Puglia 8Sardegna 4Sicilia 44Autonomous Province of Bolzano 1Veneto 257Military structures 1(*) See "Methods" sectionPETRA 2004-2005Adverse reactionsOverall, 986 forms were returned (Table II), ofwhich 871 reported adverse reactions and 63 reportednear miss errors or errors; 52 forms could not beevaluated (Figure 3A).In 848 of the forms, the adverse reaction wasattributed to the transfusion; in 65% of these (n=555)the association with the transfusion was stated to bestrong (certain or probable cause); only these weretaken into consideration for the analysis (Figure 3B).As far as regards the severity of the reactions, thesecaused mild symptoms in 65.8% of the reported casesand long-term morbidity in 31.9%, were lifethreateningin 2.1%, and led to death in 0.2%. Figure4 shows the severity of the reactions in the events inwhich the transfusion was probably or certainly thecause of the reaction. The only death was due to anacute haemolytic reaction caused by ABOincompatibility. Life-threatening events wereanaphylactic reactions (n=5), volume overloadsyndrome (n=2), TRALI (n=1), febrile reaction (n=1),urticarial reaction (n=1) and reactions defined as"other" (n=2).Almost all the adverse reactions reported wereacute; only three were delayed (late haemolysis). Asregards the type of reaction, 46.4% were of an allergicnature (urticaria and anaphylactic reactions) and33.9% of a febrile type (Figure 5).The type of blood component involved in theadverse reactions was reported in 92.9% of the cases70Blood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-07066-074_hassan.p65 7005/07/2007, 10.44

Haemovigilance in ItalyAB10001000800800600400adverse reactionserrors, near miss errorsnot evalu able600400ce rta inprobableimprobableexclu ded20020000Figure 3 - PETRA reports received350300250200150100urticaria43.1%volume overload1.8%TRALI0.2%other4.0%shivers8.8%500ProbableCertainlate haemolysis0.6%nausea, vomiting3.3%anaphylaxis3.3%acute haemolysis0.9%Mild symptomsLong-term morbidityLife- threate ning in theshort-termDeathfever33.9%Figure 4 - Severity of the adverse reactionsFigure 5 - Adverse reactions associated with transfusions(PETRA 2004-2005)(n=516), in which the transfusion was stated to beprobably or certainly the cause of the reaction: redblood cells in 374 reactions, platelets in 73, freshfrozenplasma in 53, whole blood in 5 and stem cellsin 11.ErrorsThere were 16 notifications of errors, concerningtransfusions given to the wrong patient: thetransfusions were ABO incompatible in 56% of thecases, Rh incompatible in 6% and not specified in theother 38% of the cases. The cases notified were errorsof identification of the patient and/or sample test-tube:in one case, the sample used for the request was takenfrom a different patient; in another case, the error wascaused by the patient having the same name as anotherpatient, already registered in the TS archives; in oneother case, the transfusion request, related to anotherpatient, was not checked; in the other cases, the errorwas caused by not correctly identifying the patient inBlood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-0771066-074_hassan.p65 7105/07/2007, 11.06

Giampaolo A et al.Table III - Description of near miss errorsIn the wardSample taken from wrong person 36%Errors in data identifying the test-tube 12%Error in request form 17%Incongruities between data on request form and on test-tube 10%In the Transfusion StructureExchange of samples 10%Error during the stage of accepting the request in the management system 2.5%Mistaking one patient for another 2.5%Error in serological test in the laboratory 5%Issue of wrong units 2.5%Predeposited unit belonging to another patient sent to another TS 2.5%the ward at the time of the transfusion. In the sixreports in which the patients' outcome was described,one patient died, three had no consequences and twohad mild symptoms.Near miss errorsThere were 47 reports of near miss errors, that is,errors recognised before the transfusion. Six of theforms did not report the type of error that had occurred,the other forms described, for the most part (75%),errors occurring in the wards at the time of taking thesamples, errors in filling in the identifying data onthe test-tube, and errors in completing the requestform.The other 25% of the reports concerned errorsoccurring in the TS at the time of accepting the requestand the samples, when issuing the units, and whenconducting serological tests in the laboratory. TableIII presents the various types of near miss errors indetail.Discussion and conclusionsThis study, the first Italian report on the notificationof adverse reactions to transfusions, refers to years2004 and 2005 and concerns about 50% of all bloodcomponents distributed in the nation.In this first analysis of national haemovigilancedata it was considered useful to analyse both thePETRA forms that were received and the summarydata-sheets (concerning only the type of adversereaction) supplied by those regions, which, incompliance with Italian legislation, had already setup a system for recording transfusions and notifyingadverse reactions. These different systems of reportingdid, however, lead to a lack of homogeneity in thenotifications, thus enabling only a partial analysis ofthe events notified. A more detailed analysis waspossible only of the PETRA forms.The notification form was designed by transfusionmedicine specialist and agreed upon consultatorymeetings with representatives of CRCC and othercomponents of the transfusion system. Health caremanagers were informed of the new haemovigilancesystem by the ISS and invited, as presidents of theCommittees on the Good Use of Blood, to spread theculture of haemovigilance and to encourage wards,in which blood products are used, to notify an adverseevents occurring after a transfusion.In 2005, 1,495 adverse reactions were notified,including those reported by TS not using PETRA.The TS participating, including those which declaredno transfusion reactions, had distributed 1,834,474units of blood components. Therefore, the reportedrate of reactions was 0.8/1,000 units. This isundoubtedly an underestimate, given that the Frenchsystem has consistently recorded about 3 adversereactions/1,000 units of blood components since1998 1 .The percentage of notifications varied greatly fromregion to region. The scarcity of notifications (underreporting)is a problem in all haemovigilance systemsand can have various causes: the adverse reaction and/or its relation with the transfusion is not alwaysrecognised; when it is recognised, the importance ofnotifying it is not accepted as a tool to improvetransfusion safety; staff may be afraid of disciplinaryaction in the case of notified errors.Almost all the adverse events notified were acutereactions, since only five delayed reactions werenotified. An almost complete lack of recording of latereactions (delayed haemolytic reactions, GvHD, posttransfusionalpurpura, viral, bacterial and parasitic72Blood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-07066-074_hassan.p65 7205/07/2007, 11.05

Haemovigilance in Italyinfections, haemochromatosis) was reported.InFrance, these reactions represent 35% of all theadverse events reported 1 .The lack of agreed definitions negatively affectsdata collection. For example, it would be useful tohave shared diagnostic criteria for TRALI, which is atransfusion reaction that is difficult to diagnose,because the symptoms can be very different and thereis not a specific diagnostic test. TRALI is included asan adverse reaction in the United Kingdomhaemovigilance system [Serious Hazards ofTransfusion (SHOT)] which, between 1996 and 2004,reported 162 cases, of which 36 were fatal; of these,13 were considered probably or certainly due to thetransfusion. In France, this adverse reaction was notdiagnosed until 2002 (although it was, perhaps,sometimes reported as volume overload). After aneffort of the haemovigilance system to improve thediagnosis of this reaction, TRALI is now betteridentified and studied. In 2003, 15 cases probably orcertainly due to transfusions were identified, of whichthree were fatal. Only one case of TRALI was reportedin the Italian system, leading to the suspicion that thereis underreporting of this reaction, perhaps due in partto the difficulty in its diagnosis.The types of transfusion reactions were febrile in46.9% of the cases and allergic (urticaria andanaphylactic reactions) in 38.7%. The analysis of onlythose cases, in which the transfusion was certainly orprobably causal, notified with the PETRA system,showed opposite proportions of these two pathologies:46.5% of the reactions were of an allergic nature andabout 34% were febrile-type reactions. Thesepercentages are in line with data reported by otherhaemovigilance systems.The promotion of a different culture, in whichnotifying errors is in some way fostered, is afundamental step in any attempt to identify and tacklesystem defects in the transfusion chain. Errors andnear miss errors (the latter being more frequent)represent failures in the system and by analysing them,critical points to be kept under control can beidentified.It has emerged from the reports on haemovigilancein France and the United Kingdom that transfusionalerrors leading to ABO-incompatible transfusionsrequire great attention. Up to 2002, the Frenchhaemovigilance system only recorded ABOincompatibilities that had a clinical effect; thenotification of clinical grade 0 errors was introducedonly subsequently, as an instrument to achieveimprovements, even in the absence of adversereactions.In Italy, as in other countries, a transfusion reactiondue to ABO incompatibility is considered a sentinelevent, which can and must be prevented."Recommendations for the prevention of transfusionreactions due to ABO incompatibility" has been issuedby the Ministry of Health (Office III – Quality ofactivities and services – General Management ofHealth Care Planning, levels of care and system ethicalprinciples).The regions of Veneto and Emilia Romagna wereparticularly active in notifying errors and near misserrors. Previous implementation of systems for clinicalrisk management and a culture of reporting adverseevents, with the aim of improving patient safety,increased participation to haemovigilance and enablebetter and greater integration of monitoring systems.Overall, there were nine reported cases of ABOincompatibility, of which one was fatal; all these caseswere due to an error in one of the critical steps ofidentifying the patient. These nine cases account for0.6% of the events notified.SHOT found that the overall number of errors wasequivalent to 70% of the events notified and that casesof ABO incompatibility accounted for 10% of theevents notified 2 . The need to educate and train staff isconsidered fundamental for the function of ahaemovigilance system. The data on errors should bemonitored for longer, in order that awareness of theimportance of notifying errors, by staff, can make thehaemovigilance system fully effective.Estimating the incidence of errors is essential tomonitor blood safety and to help health care managersto make informed decisions on systematicallyintroducing instruments for the identification ofpatients/units of blood components. These instrumentsare commercially available and already used in somehospitals.Compared to the organization of the transfusionsystem in other countries, in which blood banks arecompletely separate from hospitals, Italian TS arelocated within hospitals leading to a differentmanagement of blood units, presumably with a moreeffective traceability.Having overcome the problem of under-reporting,given that the incidence of errors is lower than thatBlood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-0773066-074_hassan.p65 7305/07/2007, 11.05

eported by other countries, it could be concluded thatthe organisation of the Italian transfusion system isbetter able to guarantee the correct identification ofpatients/units of blood components.AcknowledgementsWe thank the CRCC that participated in thenational haemovigilance system and, in particular, allthe Transfusion Structures that used the PETRAsoftware.We also thank Dr. Roberto Baroni for his preciouscontribution in standardising and analysing the clinicaldata.References1) Rebibo D, Hauser L, Slimani A, et al. The Frenchhaemovigilance system: organization and results for2003. Transfus Apher Sci 2004; 31: 145-53.2) Stainsby D, Jones H, Asher D, et al. Serious hazards oftransfusion: a decade of hemovigilance in the UK.Transfus Med Rev 2006; 20: 273-82.3) Directive n. 98/2002 of the European Parliament andthe Council of Europe, 27 January 2003, OfficialGazzette of the European Union L 33/30 of 08/02/2003.4) Andreu G, Morel P, Forestier F, et al. Hemovigilancenetwork in France: organization and analysis ofimmediate transfusion incident reports from 1994 to1998. Transfusion 2002; 42: 1356-64.5) Grazzini G, Hassan HJ, Aprili G. Haemovigilance.International Forum. Vox Sang 2006; 90: 207-41.6) Ministerial Decree, 1 March 2000. Adozione del progettorelativo al piano nazionale sangue e plasma per iltriennio 1999-2001. Ordinary supplement n. 52 to theGazzetta Ufficiale della Repubblica Italiana – Generalseries n. 73 of 28 March 2000.7) Legislative Decree 19 August 2005. Attuazione dellaDirettiva 98/2002 che stabilisce norme di qualità e disicurezza per la raccolta, il controllo, la lavorazione, laconservazione e la distribuzione del sangue umano e deisuoi componenti. Gazzetta Ufficiale della RepubblicaItaliana – General series n. 221 of 22/09/2005.8) Piccinini V, Vulcano F, Catalano L, et al. Sorveglianzadelle malattie infettive trasmissibili con la trasfusione(SMITT) nell'anno 2004. Notiziario ISS 2006; 19: 11-7.9) Gonzalez M, Règine V, Piccinini V, et al. Residual riskof transfusion-transmitted human immunodeficiencyvirus, hepatitis C virus, and hepatitis B virus infectionsin Italy. Transfusion 2005; 45: 1670-6.10) Velati C, Fomiatti L, Baruffi L, et al. Impact of nucleicacid amplification technology (NAT) in Italy in the threeyears following implementation (2001-2003).EuroSurveill 2005; 10: 12-4.11) Hassan HJ. Towards haemovigilance. BloodTransfusion 2004; 2, 155-9.12) Piccinini V, Abbonizio F, Catalano L, et al. Mappa delleStrutture Trasfusionali esistenti sul territorio nazionale(aggiornamento 2005). Notiziario ISS - Strumenti diRiferimento. 2006; 03 (S1): 1-75.13) Ministerial Decree 3 March 2005. Caratteristiche emodalità per la donazione del sangue e diemocomponenti. Gazzetta Ufficiale della RepubblicaItaliana – General Series n. 85 of 13/04/2005.14) Council of Europe. Committee of Ministers.Recommendation N. R(95) 15. Preparation, use andguarantee of the quality of blood components. Adoptedby the Committee of Ministers on 12 October 1995 atthe 545 th Meeting (issued 10 June 1996)15) Catalano L, Abbonizio F, Giampaolo A, et al. Registronazionale e regionale del sangue e del plasma. Rapporto2005. Rapporti ISTISAN 2006; 06 (30).Received: 10 January 2007 – Accepted: 24 April 2007Correspondence: Dr. Hamisa Jane HassanReparto di Metodologie TrasfusionaliDipartimento di Ematologia, Oncologia e Medicina MolecolareIstituto Superiore di SanitàViale Regina Elena, 299 - 00161 Roma - Italye-mail: j.hassan@iss.it74Blood Transfus 2007; 5: 66-74 DOI 10.2450/2007.0001-07066-074_hassan.p65 7405/07/2007, 11.05

ORIGINAL ARTICLETuscan Study on the Appropriateness of Fresh-frozen Plasma Transfusion(TuSAPlaT)Giancarlo Maria Liumbruno 1 , Maria Laura Sodini 1 , Giuliano Grazzini 21Servizio di Immunoematologia and Medicina Trasfusionale, AUSL N. 6 – Livorno, Italy2CRCC Regione Tuscany, ItalyIntroduction. The considerable increase in the consumption of fresh-frozen plasma (FFP) recordedin 2002 in the Region of Tuscany made it necessary to check the appropriateness of the use of this bloodcomponent in all transfusion facilities in the Tuscan network.Materials and methods. From July 1, 2003 to December 31, 2005, the Regional Blood TransfusionCo-ordinating Centre carried out an audit on the clinical use of FFP in the 40 structures included in theTuscan transfusion network. The study had two complementary parts: a review of guidelines on the useof FFP and the involvement of Hospital Transfusion Committees in evaluating the outcome of the auditand in the consequent local policy decisions and in educating clinicians.Results. The data from all 40 of the regional transfusion structures were analysed. The audit, whichwas initially retrospective, gradually became prospective. The percentage clinical use of FFP decreased,compared to 2002, in each of the 3 years of the study: a) 2003: - 8.92%; b) 2004: - 2.11%; c) 2005: -1.97%. The inappropriate requests for plasma decreased from 27% to 22.7% of the total. It was possibleto classify the inappropriate requests for plasma on the basis of homogeneous, regionally defined criteria.The most frequent inappropriate indication (60.7% of the total) was the use of plasma in the case ofhaemorrhage in patients with a normal PT and/or PTT or unavailable results. Each hospital revised its ownguidelines between 2004 and 2005 and the Hospital Transfusion Committees set up appropriate educationaland behavioural interventions.Conclusions. The capacity of transfusion facilities to make data on the use of blood componentsavailable systematically and continuously is an essential feature of clinical governance; systematic clinicalauditing increases the level of appropriate behaviours in the transfusion sector, contemporaneouslycontributing to self-sufficiency in transfusion products, and may direct research towards those clinicalsettings at greatest risk of inappropriate use of transfusion therapy with FFP.Key words: fresh-frozen plasma, transfusion, audit, appropriateness.IntroductionThe transfusion of blood components plays afundamental role in the management of various pathologiesand is, sometimes, a life-saving treatment. Blood is,however, an expensive and limited resource; over the last20 years there has been a progressive increase in demandsfor this product, mainly as a result of the advances inoncohaematological therapies and the increase in majorsurgery. The achievement of self-sufficiency in blood andits derivatives, a goal destined to be affected also bychanges in the demographic characteristics of thepopulation, is a priority for all health care systems, notonly for those in western societies 1-6 .Presented in part at the 39 th SIMTI National Congress of Transfusion Medicine(Paestum, SA, Italy, October, 4-7, 2006) and the 59 th Meeting of the AABB(Miami Beach, FL, USA, October 21-24, 2006).Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-07© SIMTI Servizi Srl75075-84_Liumbruno.p65 7505/07/2007, 11.01

Liumbruno GM et al.In Italy, in the 5 years from 2001-2005, the clinical use offresh-frozen plasma (FFP) has tended to increase, despite2% reductions in 2002 and in 2005 compared to theconsumption in the respective preceding years, 2001 and2004. In fact, the consumption of FFP increased from 146,844L in 2001 to 153,493 L in 2005 1,7-10 .In 2002, the Tuscany Regional Blood Transfusion CoordinatingCentre (RBTCC) for transfusion services foundthat the use of FFP in the Region clearly contrasted withthe national trend; indeed, that year, the use of FFP, reachedits peak of 10,625 L (equivalent to about 3L/1,000 residents/year), with a increase of 4% with respect to 2001. Therefore,in 2003, the RBTCC, in the context of a regionalbenchmarking project lasting several years, that was aimedat continuous improvement of quality and self-sufficiencyof the Tuscan transfusion system, introduced a systematicaudit of the appropriateness of the clinical use of FFP inthe hospitals making up the regional transfusion network.Inappropriate transfusion therapy with FFP is probablyone of the main avoidable risks for patients and it is knownthat guidelines are not, per se, able to influence clinicalpractice 11,12 , because without the help of other instruments,such as auditing and educational interventions, they maynot be able to reach a level of diffusion and acceptancesuch as to be able to direct clinical practice towards therecommended behaviours.There is, however, evidence showing that clinicalauditing is an educational and behavioural strategy that iseffective in altering the level of inappropriate transfusiontreatment and is able to contribute to self-sufficiency inblood products, to a reduction in health care costs and tothe delivery of better health care 13 .The aim of this study was to verify the impact of anauditing process on the appropriateness of the clinical useof FFP and to classify the main causes of the inappropriateuse of this product within the regional transfusion network.Materials and methodsThe transfusion system of the Region of Tuscany isspread throughout the region and comprises 15 Servicesof Immunohaematology and Transfusion Medicine (SIMT)and 25 depending on them 14 ; the Tuscan transfusionnetwork is incorporated in regional hospitals, managed by12 Health Care Companies and by 4 University HospitalCompanies, which deliver health care services to apopulation of about 3,600,000 people.The study was co-ordinated by the RBTCC and carriedout over a period of 3 years, that is, the second semestersof 2003 and 2004 and throughout 2005; in this period all theregional transfusion structures audited the requestsreceived for FFP and, at the same time, undertook twocomplementary actions: updating their guidelines andinvolving the Hospital Transfusion Committees (ComitatiOspedalieri per il Buon Uso del Sangue, COBUS) in theevaluation of the outcome of the audit and in theconsequent implications for local policy and education ofclinicians.The audit was conducted by manual review oftransfusion requests by medical staff of the SIMT and theTransfusion Sections. Each transfusion structure was askedfor the following data for each year: a) method of carryingout the appropriateness audit (retrospective orprospective) of the FFP transfusion requests received; b)total number of requests for FFP received; c) number ofFFP requests audited; d) classification of the requests, onthe basis of literature data and specific hospital guidelines,into: 1) appropriate; 2) inappropriate and, from 2004, 3) intoa grey zone, with the aim of dividing the definitelyinappropriate requests from those that were potentiallyinappropriate; e) classification of the inappropriaterequests, from July 2004, according to the following,predefined, homogeneous regional criteria: 1) haemorrhagein a surgical context with PT and/or PTT not available; 2)haemorrhage in a non-surgical context with PT and/or PTTnot available; 3) haemorrhage with PT and/or PTT normal;4) prophylaxis of haemorrhagic events with PT and/or PTTnormal or not performed; 5) prophylaxis of haemorrhagicevents in the presence of altered PT and/or PTT; 6)hypoalbuminaemia and/or nutritional purposes; 7) requestsbased on predefined formulas; 8) concomitance of morethan one of the preceding causes of inappropriateness; 9)other causes.The data collected by the SIMT and the TransfusionSections were subsequently sent at regular intervals to theRBTCC, which processed them, and reported them togetherwith information on the regional production andconsumption of FFP from 2002 to 2005; furthermore, theclinical use of FFP from 2002 to 2005 was also comparedwith the consumption of red blood cells (RBCs), a parameterused by the RBTCC as a regional indicator of the intensityof health care activities.The statistical analysis (Student's t test) was carriedout with GraphPad InStat software (version 3.00, GraphPadSoftware, San Diego, CA – USA).ResultsData were received from all 40 regional transfusionstructures and analysed. Each hospital revised its own76Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-07075-84_Liumbruno.p65 7605/07/2007, 11.01

Appropriateness of FFP transfusionguidelines between 2004 and 2005, also in the light of theguidelines published in July 2004 in the British Journal ofHaematology by the British Committee for Standards inHaematology 15 .The indications, for which the clinical use of FFP isconsidered appropriate, are reported below:- correction of a congenital deficiency of a clotting factorfor which the specific factor concentrate is lacking;- acquired deficiency of multiple clotting factors, whenPT and aPTT are > 1.5 normal levels, in the presence ofan ongoing haemorrhage or severe risk of haemorrhage;- use as a fluid replacement in the apheretic treatment ofthrombotic microangiopathies;- reconstitution of whole blood for exchangetransfusions;- hereditary angioedema due to C 1esterase deficiency,in the absence of a specific plasma-derivative.In the same period, the COBUS of all the structuresinvolved in the study evaluated the results of the auditscarried out and involved clinicians in a shared review ofthe guidelines on transfusion therapy with FFP and in theirdiffusion within the hospitals.The overall data for the 3 years examined on the waythe auditing was carried out showed that, on average, 56%of the transfusion requests underwent systematic,prospective auditing.The mean percentage of prospective audits of therequests increased in a statistically significant manner from30% in 2003 to 65% in 2004 (p = 0.0001) and to 72% in 2005(p < 0.0001); the increase between 2004 and 2005 was notstatistically significant (p = 0.26) (Table I).The total percentage of requests subjected toappropriateness auditing increased from 39.2%, in 2003, to99.1%, in 2005; that of the inappropriate requests changedfrom 27% in 2003 to 26.7% in 2004, and to 22.7% in 2005;furthermore, in the period of 18 months between July 2004and December 2005, 5.6% of the requests checked (900/15,996) were classified in the grey zone (Table II).In the 24 months of the study, the Tuscan transfusionstructures received a total of 24,918 requests for FFP andaudited 19,070 (76.5%); 13,506 (70.8%) of the requests wereconsidered appropriate, 4,664 (24.5%) inappropriate and900 (4.7% of the total in the 3 years) fell in the grey zone.The analysis of the percentage of requests for FFP,which were audited by the individual transfusion structures(data not shown), revealed increases that were statisticallysignificant for the comparison between 2003 and 2004 (p =0.03) and between 2003 and 2005 (p = 0.02), but not for thecomparison between 2004 and 2005 (p = 0.2).The trends in the percentages of appropriate, notappropriate and grey zone requests, analysed for theindividual hospitals for the period under consideration,did not show statistically significant changes (data notshown), although there was a general tendency to areduction in the total percentage of inappropriate and greyzone requests for FFP.Compared with the level of use in 2002, the clinical useof FFP (in litres) decreased in each of the 3 years of thestudy: a) 2003: - 8.92%; b) 2004: - 2.11%; c) 2005: - 1.97%;similarly, there was a reduction in the litres of FFP derivedfrom separation that were used clinically: a) 2003: -13.99%,b) 2004: - 10.13%, c) 2005: - 17.91%, and a steady increasein the percentage of FFP from apheresis used clinically(Table III) (Figure 1).Table I - Region of Tuscany: trend in the percentage of prospective audits of Fresh Frozen Plasma (FFP) requests from 2003 to 2005Prospective audits of FFP requests in the period 2003-2005 (%)2003 2004 2005 2003-2005Mean 30 65 72 56Standard deviation 16 36 33 34Median 37 77 85 45Range 0-45 0-100 0-100 0-10095% Confidence Interval 20.3 – 39.7 43.2 – 86.6 52.2 – 91.7 45 - 67p2003 vs 2004 2003 vs 2005 2004 vs 20050.0001 < 0.0001 0.26Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-0777075-84_Liumbruno.p65 7705/07/2007, 11.01

Liumbruno GM et al.Table II -Region of Tuscany: classification of the transfusion requests for fresh-frozen plasma (FFP)received and audited from 2003 to 2005FFP requests 2003 2004 2005 2004-2005(6 months) (6 months) (12 months) (18 months)Received 7,841 5,978 11,099 17,077Audited 3,074 4,997 10,999 15,996Audited % 39.2 83.5 99.1 93.7Appropriate 2,244 3,263 7,999 11,262Appropriate % 73 65.3 72.7 70.4Not appropriate 830 1,335 2,499 3,834Not appropriate % 27 26,7 22,7 24Grey zone 0 399 501 900Grey zone % 0 8 4,6 5,6100%90%80%5,2214,491 4,6924,28670%49.14% 46.41% 45.12% 41.15%60%50%40%30%20%6,1305,4045,186 5,70850.86% 53.59% 54.88% 58.85%10%0%2002 2003 2004 2005FFP from separation (L)FFP from apheresis (L)Figure 1 –Region of Tuscany: fresh-frozen plasma (FFP) from apheresis and from separation used for clinicalpurposes from 2002 to 2005. The values are expressed in litres and as a percentage of the total.78Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-07075-84_Liumbruno.p65 7805/07/2007, 11.01

Appropriateness of FFP transfusionTable III - Region of Tuscany: consumption of fresh-frozen plasma (FFP) for clinical use from 2002 to 2005 and production ofthe FFP in the same period.2002 2003 2004 2005FFP produced (L) 57,400 59,036 62,626 65,979Difference in production compared to 2002 (%) + 2.85 + 9.1 + 14.95FFP distributed for clinical use (L) from apheresis 5,404 5,186 5,708 6,130from separation 5,221 4,491 4,692 4,286Total 10,625 9,677 10,400 10,416Difference in the distribution of FFP from apheresis - 4.35 + 5.62 + 13.43for clinical use compared to 2002 (%) from separation - 13.99 - 10.13 - 17.91Total - 8.92 - 2.11 - 1.97The year in which the clinical use of FFP was lowestwas 2003; there were then moderate increases in 2004 and2005, although the levels of 2002 were not reached. Theincreased use of FFP (in litres) from 2003 to 2005 was,nevertheless, percentage-wise, less than that expected onthe basis of the increased consumption of RBCs in thesame period. RBC use is an indicator of the intensity ofhealth care services recorded by the RBTCC throughoutthe region; in fact, the consumption of RBC units was145,498 units in 2002, decreased in 2003 to 144,921 units (-0.39 %) then rose in 2004 to 151,947 units (+ 4.43 %) and in2005 to 154,080 units (+ 5.98 %) (Figure 2).The regional production of plasma changed from 57,400L in 2002 to 65,979 L in 2005 (Table III) (Figure 3). Therelationship (in litres) between FFP used clinically and thetotal amount of FFP produced showed a tendency todecrease with respect to that in 2002: a) 2002: 18.51%; b)2003: 16.39%; c) 2004: 16.6%; and d) 2005: 15.78%.In order to classify the types of inappropriate requestsaccording to predefined criteria established by the Region,an analysis was conducted on 15,996 requests (93.7% ofthe total 17,077 received) for FFP during 18 consecutivemonths from July 1, 2004 to December 31, 2005; 3,834 (24%)of the requests were considered inappropriate, 11,262(70.4%) appropriate and 900 (5.6%) fell in the grey zone(Table II).Table IV reports the types of inappropriate requestsand their proportional incidence in decreasing order.The analysis of the types of inappropriate requestsrecorded by the individual hospitals (data not shown)revealed the following: (i) there were 2,328 inappropriaterequests for FFP (60.7% of all inappropriate requests) forthe treatment of haemorrhagic events in patients with PTand/or PTT values unavailable or normal): 38.7% of theserequests (n=901) came from three University hospitals; (ii)1,853 requests (48.3% of the total) were definedinappropriate because they were made without the resultsof screening coagulation tests: 59.2% of these (n=758) werefrom the same three University hospitals; (iii) 906 of theinappropriate requests (23.7% of the total) were made forprophylaxis of haemorrhagic events (with PT and/or PTTnot available, normal or altered): 29.6% of these requests(n=268) were made by a single University hospital; 3)treatment of hypoalbuminaemia and/or nutritionalpurposes were the reasons for 212 inappropriate requestsfor FFP (5.5% of the total): 56.6% of these requests (n=120)came from one University hospital.DiscussionThe systematic review of 19,070 requests for FFP,equivalent to 76.5% of all requests for the period 2003-2005, showed that the inappropriate use of FFP in Tuscanydecreased from 27% in 2003 to 22.7%, in 2005. The meanreduction over the whole 3-year period was 24.5%; thisresult, if transferred into daily clinical practice, implies thatof every four requests for FFP, one is inappropriate. Thepercentage of inappropriate clinical use of FFP recorded inthis regional study is however, in the lower range (10-74%)reported in previous studies on the same subject carriedout between 1986 and 2006, although these were conductedin heterogeneous health care systems, including some nonwesterncountries 6,16-37 . The percentage reduction ininappropriate use (4.3%) was detectable, although relativelysmall; it is compatible with results obtained in other studiescarried out in health care systems comparable with theItalian system. In these studies, prospective auditing wascombined with corrective actions (such as improving bloodcomponent request form) and educational interventionsaimed at improving clinical practice 17,29 . The decrease ininappropriate requests for FFP during the course of theBlood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-0779075-84_Liumbruno.p65 7905/07/2007, 11.01

Liumbruno GM et al.8644.435.9820-22003 vs 2002-0.39-2.112004 vs 2002-1.972005 vs 2002-4-6-8Clinical use of FFP (%)Clinical use of RBCs (%)-10-8.92Figure 2 - Region of Tuscany: comparison between the changes in percentage consumption of fresh-frozen plasma (FFP)and red blood cells (RBC) in the years 2003-2005 compared to in 200270,00060,00062,62665,97957,40059,03650,00040,00030,00020,000• Production of FFP (L)• Clinical use of FFP (L)10,00010,6259,677 10,400 10,41602002 2003 2004 2005Figure 3 - Region of Tuscany: production of fresh-frozen plasma (FFP) and clinical use of FFP from 2002 in litres80Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-07075-84_Liumbruno.p65 8005/07/2007, 11.01

Appropriateness of FFP transfusionTable IV - Region of Tuscany: causes of inappropriate use of fresh-frozen plasma as revealed by the audit ofrequests from July 1, 2004 to December 31, 2005Cause of inappropriate use Number of requests %Haemorrhage in a surgical setting with PT and/or PTT not available 1,281 33.4Haemorrhage in non-surgical setting with PT and/or PTT not available 572 14.9Haemorrhage with PT and/or PTT normal 475 12.4Prophylaxis against bleeding with PT and/or PTT normal or not carried out 470 12.3Prophylaxis against bleeding with PT and/or PTT altered 436 11.4Hypoalbuminaemia and/or nutritional purposes 212 5.5Concomitance of more than one inappropriateness 199 5.2Other causes 97 2.5Requests based on a predefined formula 92 2.4Total 3,834 100study was attributed to various factors, probably actingtogether: a) the increase in audits of appropriateness (+44.3% in 2004 and + 15.6% in 2005) and, in particular, thecontinuous increase of prospective auditing from 2003,even if this did not reach statistical significance in 2005; b)the interventions carried out between 2004 and 2005, suchas the divulgation of the updated guidelines within thehospitals and the education of clinicians promoted by theCOBUS. In fact, the COBUS are able to influence theappropriateness of the use of blood componentssignificantly 38 . The results of the present study confirmalready existing evidence on the capacity of behaviouralinterventions to modify the level of inappropriatetransfusion therapy 13 . Moreover, this regional study seemsto indicate that auditing and educational interventions,particularly if combined with a review of the guidelines ontransfusion therapy, can enhance their diffusion andacceptance, improving the appropriateness of bloodproduct use. The analysis of the regional consumption ofplasma, during the period of the study, confirmed theincrease of appropriateness of FFP therapy. Other indicesof the more appropriate use of FFP are the considerablereduction in the clinical use of plasma obtained byseparation in favour of that from apheresis (withconsequent less exposure of recipients to donors) and thelesser percentage increase in FFP use compared to that ofRBC use. The increased use of RBCs in the same periodreflects the considerable intensification of health care, thatoccurred in the Region from 2003 onwards, particularly intransplant medicine 39 . This could explain why, in the sameperiod that prospective auditing increased significantly,there was also a moderate increase in the consumption ofFFP, although this never reached the levels of 2002.The possibility of classifying requests in a grey zone,introduced in the audit from 2004, enabled 900 potentiallyinappropriate requests to be picked up in the following 18months.These requests were separated from the definitelyinappropriate requests, thus reducing the "bias" ofmistaken classification and undoubtedly contributing toreducing the percentage of inappropriate requests 40,41 .Furthermore, the use of the grey zone was attributed thecapacity to increase the robustness of the studies on theinappropriateness of transfusion therapy, precisely becauseof the separation that this zone enables between potentiallyand truly inappropriate transfusions 42 .The analysis of about 16,000 requests made in 18consecutive months, showing that 24% of these wereinappropriate, enabled the main causes of theinappropriateness to be determined and grouped inpredefined classes. The highest percentage of inappropriaterequests (60.7%) was for the treatment of haemorrhagicevents, both in medical and surgical settings, withoutinformation on baseline haemostatic parameters (PT andPTT) or in the presence of normal values for theseparameters. It, therefore, appears that there are situationsin which the therapeutic choice seems to be dictatedpredominantly by the clinical evaluation, independently oflaboratory information, since the data were missingcompletely in 48.3% of the cases. This last figure probablyunderestimates the real percentage of requests made in theabsence of laboratory data, because it does not take intoaccount the 12.3% of inappropriate requests for theprophylaxis of haemorrhagic events in which the absenceof PT and/or PTT was combined with normal results forthese parameters.Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-0781075-84_Liumbruno.p65 8105/07/2007, 11.01

Liumbruno GM et al.Thus, the lack of laboratory data we found seems to begreater than the 34.1% reported by other Authors 16 ; thisdifference is probably due to the different design of thestudy mentioned, which also evaluated the laboratory dataof patients transfused with FFP, with the aid of a specificdatabase. Both results were, however, encouraging in thelight of the data from the Sanguis study in 1994, whichfound that PT data were missing in 84% of the transfusionrecipients of FFP in 43 European hospitals 43 .In second place as a cause of inappropriate use of FFP,was prophylaxis of bleeding (23.7% of inappropriaterequests). Transfusion therapy with FFP to correctcoagulation test abnormalities before carrying out invasiveprocedures is a common clinical practice, even if there areno "evidence-based" guidelines on this issue. Haemostasisis a complex process of interactions between pro-coagulantmolecules, platelets, natural anticoagulants, the fibrinolyticsystem and the endothelium. The screening tests ofhaemostasis prior to surgery are usually PT/INR and PTT;these laboratory tests have been developed to identifycauses of bleeding in patients with a demonstratedhaemorrhagic diathesis (high pre-test probability) and notto evaluate haemostasis in patients with a negative historyfor haemorrhage, nor has it been demonstrated that theyare able to do so. The transfusion of FFP before an invasiveprocedure, in order to correct mild to moderate changes incoagulation tests, is not, therefore, able to correct theanomaly or to reduce the perceived haemorrhagic risk 44-54 .The use of FFP to correct hypoproteinaemia or fornutritional purposes is greater in developing countries, butit is difficult to find a justification for this practice in theItalian health care system 19 .The present study also reveals that inappropriate useof FFP is more common in University hospitals than insmaller, peripheral hospitals, as already reported in theliterature 18 ; this could be due to the greater complexity ofcases treated in the University hospitals and the greateramount of blood components distributed by thesehospitals, both factors which could limit the possibility ofcarrying out audits and adhering to guidelines.Possible limitations of this study are: a) that the auditdid not use information technology, which, with a specificdatabase, would have enabled a whole set of diagnostic,clinical and laboratory data to be processed and comparedat national and international levels 16,41,55-57 ; b) the definitionof inappropriateness did not take into account the dose ofFFP used and, therefore, the possible administration ofsubtherapeutic doses 19,41,46,58,59 ; c) the comparison betweenthe years was not homogeneous, because the data for 2005were not divided by semester, but referred to the wholecalendar year.The capacity of transfusion structures to make dataavailable systematically and continuously on the use ofblood components is an essential feature of clinicalgovernance 60 ; systematic clinical auditing, particularly ifperformed at a large scale such as regionally, thus enablingthe acquisition of a substantial quantity of data, increasesthe degree of appropriate behaviours in the field oftransfusion therapy, contributes to self-sufficiency, andcan direct clinical research towards those sectors in whichinappropriate treatment is greatest. Transfusion medicineis, in fact, a transversal medical discipline in which thereare few adequately structured trials that provide evidenceon the indications and efficacy of transfusion therapy.Appropriately designed clinical trials could modify somedecisions, so that they are made on "evidence-based"criteria 61 . This is what is happening in the USA, where amulticentre clinical trial should provide evidence on whichto decide on the clinical use of FFP in patients with liverdisease and INR values of 1.3 - 2, who are candidates forinvasive procedures 61 ; this multicentre study was triggeredby a retrospective audit 53 , followed by a prospective audit 48 ,conducted in the same hospital, and a systematic review ofthe literature on the subject 49 .References1) Catalano L, Abbonizio F, Giampaolo A, et al. Registronazionale e regionale del sangue e del plasma. Rapporto 2005.Istituto Superiore di Sanità. Available at: http://www.iss.it/binary/publ/cont/06-30.1163668346.pdf.2) Sullivan MT, Cotten R, Read EJ, et al. Blood collection andtransfusion in the United States in 2001. Transfusion 2007;47: 385-94.3) Greinacher A, Fendrich K, Alpen U, et al. Impact ofdemographic changes on the blood supply: Mecklenburg-West Pomerania as a model region for Europe. Transfusion2007; 47: 395-401.4) Vamvakas EC. Epidemiology of blood transfusion andforecasts of demand for blood. In: Vamvakas EC, editor.Evidence-Based Practice of Transfusion Medicine, Bethesda:American Association of Blood Banks; 2001. p. 177-99.5) Currie CJ, Patel TC, McEvan P, et al. Evaluation of thefuture supply and demand for blood products in the UnitedKingdom National Health Service. Transfus Med 2004; 14:19-24.6) Yeh CJ, Wu CF, Hsu WT, et al. Transfusion audit of freshfrozenplasma in southern Taiwan. Vox Sang 2006; 91: 270-4.7) Catalano L, Abbonizio F, Hassan JH. Registro nazionale eregionale del sangue e del plasma. Rapporto 2001. IstitutoSuperiore di Sanità. Available at: http://www.iss.it/binary/publ/publi/0315.1109149613.pdf.8) Catalano L, Abbonizio F, Giampaolo A, et al. Registronazionale e regionale del sangue e del plasma. Rapporto 2002.82Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-07075-84_Liumbruno.p65 8205/07/2007, 11.01

Appropriateness of FFP transfusionIstituto Superiore di Sanità. Available at: http://www.iss.it/binary/publ/publi/0341.1109236509.pdf.9) Catalano L, Abbonizio F, Giampaolo A, et al. Registronazionale e regionale del sangue e del plasma. Rapporto 2003.Istituto Superiore di Sanità. Available at: http://www.iss.it/binary/publ2/cont/04-36.1140439702.pdf.10) Catalano L, Abbonizio F, Giampaolo A, et al. Registronazionale e regionale del sangue e del plasma. Rapporto 2004.Istituto Superiore di Sanità. Available at: http://www.iss.it/binary/publ/cont/05-45.1141828723.pdf.11) Lomas J, Anderson GM, Domnick-Pierre K, et al. Do practiceguidelines guide practice? The effect of a consensus statement onthe practice of physicians. N Engl J Med 1989; 321: 1306-11.12) McClelland B. Effective use of blood components. In:Murphy MF, Pamphilon DH, editors. Practical TransfusionMedicine, Oxford: Blackwell Science; 2001. p. 65-76.13) Wilson K, MacDougall L, Ferguson D, et al. The effectivenessof interventions to reduce physician's levels of inappropriatetransfusion: what can be learned from a systematic review ofthe literature. Transfusion 2002; 42: 1224-9.14) Picinini V, Abbonizio F, Catalano L, et al. Mappa dellestrutture trasfusionali esistenti sul territorio nazionale(aggiornamento 2005). Istituto Superiore di Sanità. Availableat: http://www.iss.it/binary/publ/cont/0394_9303_2006_I_06-S1.1145354064.pdf.15) O'Shaughnessy DF, Atterbury C, Bolton Maggs P, et al.Guidelines for the use of fresh-frozen plasma, cryoprecipitateand cryosupernatant. Br J Haematol 2004; 126: 11-28.16) Palo R, Capraro L, Hovilehto S, et al. Population-based auditof fresh frozen plasma transfusion practices. Transfusion2006; 46: 1921-5.17) Hui CH, Williams I, Davis K. Clinical audit of the use offresh frozen plasma and platelets in a tertiary teachinghospital and the impact of a new transfusion request form.Intern Med J 2004; 35: 283-8.18) Soutar RL, Jobanputra S, Tait RC. A two-phase audit offresh frozen plasma: a regional approach [letter]. TransfusMed 2004; 14: 75-6.19) Kakkar N, Kaurt R, Dhanoa J. Improvement in fresh frozenplasma transfusion practice: results of an outcome audit.Transfus Med 2004; 14: 231-5.20) Schofield WN, Rubin GL, Dean MG. Appropriateness ofplatelet, fresh frozen plasma and cryoprecipitate transfusionin New South Wales public hospitals. Med J Aust 2003;178: 117-21.21) Chng WJ, Tan MK, Kuperan P. An audit of fresh frozenplasma usage in an acute general hospital in Singapore.Singapore Med J 2003; 44: 574-8.22) Stainsby D, Burrowes-King V. Audits of the appropriate useof fresh frozen plasma. Blood Matters 2002; 10: 7-9.Available at: http://www.blood.co.uk/hospitals/library/pdf/bm10.pdf.23) Luk C, Eckert KM, Barr RM, et al. Prospective audit of theuse of fresh-frozen plasma, based on Canadian MedicalAssociation transfusion guidelines [letter]. Can Med AssocJ 2002; 166: 1539-40.24) Prathiba R, Jayaranee S, Ramesh JC, et al. An audit of freshfrozen plasma usage in a tertiary referral centre in a developingcountry. Malays J Pathol 2001; 23: 41-6.25) Beloeil H, Brosseau M, Benhamou D. Transfusion of freshfrozen plasma (FFP): audit of prescriptions. Ann Fr AnesthReanim 2001; 20: 686-92.26) Hameedullah, Khan FA, Kamal RS. Improvement ofintraoperative fresh frozen plasma transfusion practice -impact of medical audits and provider education. J Pak MedAssoc 2000; 50: 253-6.27) Marti-Carvajal AJ, Munoz-Navarro SR, Pena-Marti GE, etal. An audit of appropriate use of blood products in adultpatients in a Venezuelan general university hospital. Int JQual Health Care 1999; 11: 391-5.28) Jones HP, Jones J, Napier JA, et al. Clinical use of FFP:results of a retrospective process and audit outcome. TransfusMed 1998; 8: 37-41.29) Tuckfield A, Haeusler MN, Grigg AP, et al. Reduction ofinappropriate use of blood products by prospectivemonitoring of transfusion request forms. Med J Aust 1997;167: 473-6.30) Marconi M, Almini D, Pizzi MN, et al. Quality assurance ofclinical transfusion practice by implementation of the privilegeof blood prescription and computerized prospective auditof blood requests. Transfus Med 1996; 6: 11-9.31) Cheng G, Wong HF, Chan A, et al. The effects of a selfeducating blood component request form and enforcementsof transfusion guidelines on FFP and platelet usage, QueenMary Hospital, Hong Kong. British Committee for Standardsin Hematology (BCSH). Clin Lab Haematol 1996; 18: 83-7.32) Schots J, Steenssens L. Blood usage review in a Belgianuniversity hospital. Int J Qual Health Care 1994; 6: 41-5.33) Thomson A, Contreras M, Knowles S. Blood componenttreatment: a retrospective audit in five major Londonhospitals. J Clin Pathol 1991; 44: 734-7.34) Barnette RE, Fish DJ, Eisenstaedt RS. Modification of freshfrozenplasma transfusion practices through educationalintervention. Transfusion 1990; 30: 253-7.35) Brien WF, Butler RJ, Inwood MJ. An audit of bloodcomponent therapy in a Canadian general teaching hospital.CMAJ 1989; 140: 812-5.36) Mozes B, Epstein M, Ben-Bassat I, et al. Evaluation of theappropriateness of blood and blood product transfusion usingpreset criteria. Transfusion 1989; 29: 473-6.37) Blumberg N, Laczin J, McMican A, et al. A critical survey offresh-frozen plasma use. Transfusion 1986; 26: 511-3.38) Hayes SL, Torella F. The role of hospital transfusioncommittees in blood product conservation. Transfus MedRev 2004; 18: 93-104.39) Organizzazione Toscana Trapianti (OTT). Attività didonazione e trapianto di organi, tessuti e cellule nella RegioneToscana. OTT Report 2005. Available at: http://www.salute.toscana.it/sst/ott/pdf/report_ott_2005.pdf .40) Hasley PB, Lave J, Kapoor WN. The necessary and theunnecessary transfusion: a critical review of reportedappropriateness rates and criteria for red cell transfusions.Transfusion 1994; 34: 110-5.41) Rotschild JM, McGurk S, Honour M, et al. Assessment ofeducation and computerized decision support interventions forimproving transfusion practice. Transfusion 2007; 47: 228-39.42) Kanter MH. The transfusion audit as a tool to improve transfusionpractice: a critical appraisal. Transfus Sci 1998; 19: 69-81.Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-0783075-84_Liumbruno.p65 8305/07/2007, 11.01

Liumbruno GM et al.43) Use of blood products for elective surgery in 43 Europeanhospitals. The Sanguis Study Group. Transfus Med 1994;4: 251-68.44) Holand L, Sarode R. Should plasma be transfusedprophylactically before invasive procedures? Curr OpinHematol 2006; 13: 447-51.45) Gajic O, Dzik WH, Toy P. Fresh frozen plasma and platelettransfusion for non bleeding patients in the intensive careunit: benefit or harm? Crit Care Med 2006; 34 (5 Suppl):S170-3.46) Holland LL, Brooks JP, Toward rational fresh frozen plasmatransfusion. Am J Clin Pathol 2006; 126: 133-9.47) Triulzi DJ. The art of plasma transfusion therapy [editorial].Transfusion 2006; 46: 1268-70.48) Adel-Wahab OI, Healy B, Dzick WH. Effect of fresh frozenplasma transfusion on prothrombin time and bleeding inpatients with mild coagulation abnormalities. Transfusion2006; 46: 1279-85.49) Segal JB, Dzick WH; Transfusion Medicine/HemostasisClinical Trials Network. Paucity of studies to support thatabnormal coagulation test results predict bleeding in the settingof invasive procedures: an evidence-based review.Transfusion 2005; 45: 1413-25.50) Holland LL, Foster TM, Marlar RA, et al. Fresh frozenplasma is ineffective for correcting minimally elevatedinternational normalized ratios [letter]. Transfusion 2005;45: 1234-5.51) Dara SI, Rana R, Afessa B, et al. Fresh frozen plasmatransfusion in critically ill medical patients with coagulopathy.Crit Care Med 2005; 33: 2667-71.52) Wallis JP, Dzik S. Is fresh frozen plasma overtransfused inthe United States? [letter]. Transfusion 2004; 44: 1674-5.53) Dzick W, Rao A. Why do physicians request fresh frozenplasma? [letter]. Transfusion 2004; 44: 1393-4.54) Stanworth SJ, Brunskill SJ, Hyde CJ, et al. Is fresh frozenplasma clinically effective? A systematic review of randomisedcontrolled trials. Br J Haematol 2004; 126: 139-52.55) Zimmermann M, Büscher M, Linhardt C, et al. A survey ofblood component use in a German university hospital.Transfusion 1997; 37: 1075-83.56) Titlestad K, Kristensen T, Jorgensen J, et al. Monitoringtransfusion practice – a computerized procedure. TransfusMed 2002; 12: 25-34.57) Titlestad K, Georgsen J, Jorgensen J, et al. Monitoringtransfusion practices at two university hospitals. Vox Sang2001; 80: 40-7.58) Santagostino E, Mancuso ME, Morfini M, et al. Solvent/detergent plasma for prevention of bleeding in recessivelyinherited coagulation disorders: dosing, pharmacokinetics andclinical efficacy. Haematologica 2006; 91: 634-9.59) Chowdhury P, Saayman UP, Paulus U, et al. Efficacy ofstandard dose and 30 ml/kg fresh frozen plasma in correctinglaboratory parameters of haemostasis in critically ill patients.Br J Haematol 2004; 125: 69-73.60) Casati G. Clinical governance: an ethical appeal or amanagement programme? [editorial]. Blood Transfus 2006;4: 251-64.61) Dzick WH. The NHLBI Clinical Trials Network intransfusion medicine and hemostasis: an overview. J ClinApher 2006; 21: 57-9.Received: 24 April 2007 – Revision accepted: 4 June 2007Correspondence: Dr. Giancarlo Maria Liumbruno,Viale Italia, 19 – 57126 Livorno, Italye-mail: giancarlo@liumbruno.it84Blood Transfus 2007; 5: 75-84 DOI 10.2450/2007.0015-07075-84_Liumbruno.p65 8405/07/2007, 11.01

Plasma derivatives and strategies for reaching self-sufficiencyin Liguria: the role of the Transfusion Medicine Serviceof the Gaslini InstituteLaura Bocciardo 1 , Marina Martinengo 1 , Diego Ardenghi 1 , Tullia Emanueli 2 ,Enrica Oliva 3 , Gino Tripodi 11Servizio di Immunoematologia e Medicina Trasfusionale, Istituto Giannina Gaslini,Genova2Servizio di Farmacia, Istituto Giannina Gaslini, Genova3CRCC Liguria, Genova, ItalyORIGINAL ARTICLEBackground. Since 2002, Liguria has been part of the Interregional Agreement on Plasma Derivatives(AIP) stipulated among some Regions of north Italy with the aim of contributing to self-sufficiency of theinterregional system through exchanges between the facilities lacking products and those with an excess.In Liguria , the management of plasma derivates is entrusted to the Regional Centre for Co-ordination andCompensation (CRCC) which, with strategies of compensation, tries to guarantee that the needs forplasma derivates are covered in the hospitals in its territory.The Services of Immunohaematology and Transfusion Medicine (SIMT) have a goal of increasing theproduction of plasma in order to participate actively in achieving regional self-sufficiency.Methods. The SIMT of the G. Gaslini Institute introduced some strategies aimed at reaching this goal.The increase in the number of donations made with a cell separator, the introduction of multicomponentdonations of plasma and platelets and the collection of high concentration platelet concentrates led to aconsiderable increase category A plasma sent for fractioning. Finally, the implementation of shared guidelineson the use of blood components enabled the clinical use of the plasma collected to be kept under control.Results and conclusions. The analysis of the trends of consumption of the most widely used plasmaderivatives showed an increase in the overall demands, which can be attributed to the paediatric focus ofour hospital and to its highly specialised wards.On the basis of the industrial technical yield, it was possible to calculate the theoretical coverage of therequirements for plasma: this highlighted a better theoretical coverage for albumin but a shortfall ofintravenous immunoglobulins. The amount of plasma necessary to meet the theoretical needs was calculatedfor each plasma derivative, revealing that the derivative requiring the greatest volume of plasma is intravenousimmunoglobulins. This finding confirms the change in the “driving product”: it is now the consumption ofintravenous immunoglobulins that determines the amount of plasma that is sent for industrial processing.Key words: plasma production; blood derivatives.IntroductionSince 2002, the Region of Liguria has been part of anInterregional Agreement on Plasma Derivatives (AIP) 1stipulated in 1998 among some Regions of north Italy(Veneto, Abruzzo, Emilia Romagna, Friuli Venezia Giulia)and the Autonomous Provinces of Trento and Bolzano,and subsequently joined by Valle d'Aosta, Tuscany,Basilicata and Umbria.Veneto, as the leading Region, manages the processingof plasma and the production of blood derivatives of allthe Regions belonging to the AIP, through the RegionalCo-ordination for Transfusion Activities (CRAT).It stipulates a single interregional agreement with thepharmaceutical industry.The Region of Veneto is also in charge of thedistribution of therapeutic blood derivatives to the otherBlood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-06© SIMTI Servizi Srl85085-92_bocciardi.p65 8505/07/2007, 11.20

Bocciardo L et alRegions on the basis of the amount of plasma supplied byeach one of them.The primary purpose of the Agreement is to combinethe efforts of the individual members, in a unanimous andsynergistic way, in order to contribute to reaching thepriority goal, established by legislators with law n. 219/05and its subsequent modifications and integrations, that is,national self-sufficiency. Once internal self-sufficiency hasbeen guaranteed, each participating region is committed tocontributing to the self-sufficiency of the interregionalsystem, through exchanges between the facilities lackingproducts and those with excesses.Being part of the AIP carries economic andorganizational advantages both in relations with thepharmaceutical industry and in interregional relationsamong the Regions adhering to the Agreement. The AIP:- guarantees greater negotiating power and, therefore,better conditions, with respect to single Regions, withthe pharmaceutical industry, that produces drugsderived from the industrial processing of plasma;- encourages exchanges of blood derivatives betweenRegions with excesses and those with shortfalls ofproducts at costs lower than market prices;- guarantees, through the contribution of plasmaconferred by each Region, the constant availability ofblood derivatives also for Regions that are not able toconfer sufficient volumes of plasma for autonomousformation of batches for industrial transformation,thereby limiting these Regions' recourse to thecommercial market;- imposes the standardisation and use of the sameproduction processes (selection of donors, stages ofprocessing and storage, tracking procedures),contributing to the definition of the Plasma Master File(PMF), which is obligatory for safer and more reliableproduction of blood derivatives;- may foster a more extensive sharing adoption ofguidelines on the use of blood derivatives, also at asupraregional level, particularly for limited products,such as intravenous immunoglobulins (IgG).The management of plasma derivates, obtained fromplasma produced by centres in the Region of Liguria, isentrusted to the Regional Centre for Co-ordination andCompensation (CRCC). On the basis of the different needsthat the Ligurian hospitals have for plasma derivatives, theCRCC tries to guarantee that these needs are met by usinga mechanism of compensation. In this context, every Serviceof Immunohaematology and Transfusion Medicine (SIMT)has the aim of optimising the types of donations to increasethe production of plasma and, therefore, the amount of thisblood component sent for industrial processing to obtainthe plasma-derived drugs, that are widely used in a clinicalcontext.An increased production of plasma is fundamental forparticipating actively in the achievement of regional selfsufficiencyin the supply of plasma derivatives. The SIMTof the Gaslini Institute has, therefore, introduced somestrategies aimed at reaching this objective.In this study we compare the data on the production ofplasma by the SIMT of the Gaslini Institute and those onthe requirements for blood derivatives by the Instituteitself, in order to analyse the theoretical coverage of theneeds for plasma derivates, in particular albumin,intravenous IgG and antithrombin III.MethodsStrategies for increasing the production ofplasma to send for industrial fractionationUntil 2002, plasma in our SIMT was obtained exclusivelyfrom two sources: productive plasmapheresis (for categoryA plasma) and whole blood (for category B plasma and aminimal proportion of category C plasma). Three possibleareas of intervention were identified:a) an increase in the plasmapheresis procedures;b) changing the collection of platelets intomulticomponent donations of plasma and platelets and,subsequently, the introduction of procedures to collecthigh-concentration platelet concentrates to increase thevolumes of plasma yielded in each procedure;c) the introduction of guidelines on the good use of bloodcomponents and close monitoring of the requests fromwards, to limit the clinical use of plasma.Analysis of the productionand internal consumption of plasmaThe data on the production of plasma in the SIMT ofthe Gaslini Institute in the period 2001-2005 were analysed(volume of each category of plasma), as were the datarelating to the proportion of plasma sent for industrialprocessing and the proportion retained for clinical use.Analysis of the consumption of plasmaderivatives and of the percentages oftheoretical cover of the requirementsThe data relating to the consumption of the plasmaderivatives most widely used in the wards in our hospital(albumin, intravenous IgG and antithrombin III) wereanalysed.86Blood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-06085-92_bocciardi.p65 8605/07/2007, 11.20

Plasma derivative self-sufficency in LiguriaOn the basis of the yields of each derivative producedper litre of plasma, defined by the pharmaceutical industryin agreement with the AIP (Table I), we calculated theamounts of plasma derivates obtained from the volume ofplasma sent for industrial processing. We, therefore,obtained the percentage theoretical coverage ofrequirements for each blood derivative guaranteed by theplasma production of the SIMT and identified the derivativewhose requirements define the amount of plasma that mustbe sent to the industry to guarantee internal selfsufficiency.Litres of plasma20001800160014001200100080060040020001720 17301830 189015602001 2002 2003 2004 2005Table I - Industrial technical yields of plasma derivatesYearAlbumin I.V. IgG Antithrombin IIIPlasma A 25 g 2.9 g 250 UIPlasma B 25 g 2.9 g -Plasma C 25 g 2.9 g -ResultsStrategies introducedStarting from the second half of 2002, attempts weremade to increase the number of plasmapheresis proceduresthrough a reorganisation of the system in the blooddonation room; after 6 months, this organisational modelwas suspended because of the negative repercussionsobserved in the management of the blood donations. Otherstrategies of this type are currently not possible.However, thanks to new cell separators, sinceSeptember 2003 has been possible to replace the collectionof platelets by apheresis by multicomponent donations ofplasma and platelets, without this causing significantreductions in the yields of platelet collections.In the two years 2004-2005, the mean volume of plasmafrom plasmaplateletapheresis was increased slightly, bymodifying the programme of the cell separators. InSeptember 2006, the collection of high-concentrationplatelet concentrates began, leading to a significantincrease in the volume of plasma collected.In December 2004, the Committee for the Good Use ofBlood approved guidelines on the use of bloodcomponents. Thanks to these guidelines, a more rigorouscontrol of the requests for plasma for clinical use waspossible. In contrast, internal guidelines, approved byCommittee for the Good Use of Blood, on the use ofalbumin, intravenous immunoglobulins and antithrombinIII are still not available, although they have been thesubject of discussion and debate for some time now.Figure 1 - Distribution of the production of plasma (in litres) inthe period from 2001-2005Plasma productionThe analysis of the data for the period 2001-2005showed a considerable increase in the volume of plasmaproduced (Figure 1): the production increased from 1,560litres in 2001 to 1,720 and 1,730 litres in 2002 and 2003,respectively; there were further increases in production in2004 and 2005, reaching 1,830 and 1,890 litres of plasma,respectively. From the destination of the plasma that, afterthe validation process, enters the distribution circuit, itcan be seen that during the 5 years considered, the volumeof category B and C plasma did not vary significantly,ranging between 1,008 and 1,096 litres a year. In contrast,the increase in plasma from apheresis (plasma A) sent forindustrial fractionation was constant and appreciable,passing from 267 litres in 2001 to 508 litres in 2005.The observed increase was not, overall, related to theincrease in the number of apheresis procedures which,particularly in the three years 2003-2005, was basicallystable and lower than the values in the preceding two years(Table II), but rather, could be attributed to the optimisationof the collection procedures carried out using cellseparators. Finally, it can be seen that during the periodstudied, the volume of plasma used for clinical purposesremained relatively steady (Table III and Figure 2).This evaluation of the production of plasma in theRegion of Liguria enabled a quantification of the role playedby the SIMT of the Gaslini Institute within the Region:during 2005, our Institute collected 8% of the total regionalcollection (Figure 3) of plasma from ordinary donationsand 16% (Figure 4) of the plasma collected by cellseparators.Blood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-0687085-92_bocciardi.p65 8705/07/2007, 11.20

Bocciardo L et alTable II - Productive apheresis procedures carried out in the SIMTN. of plasmapheresis N. of plateletphersis N. of plasma-platelet Total n. of productiveprocedures procedures pheresis procedures apheresis procedures2001 882 922 0 1,8042002 1,041 874 0 1,9152003 727 288 671 1,6862004 808 0 975 1,7832005 866 3 887 1,756Table III - Distribution of plasma (in litres) in the period 2001-2005Total plasma Plasma for Class A plasma Class B+C plasma Totale plasmadistributed clinical use sent for industrial sent for industrial sent for industrialfractionation fractionation fractionation2001 1,520 245 267 1,008 1,2752002 1,619 286 308 1,025 1,3332003 1,621 250 360 1,011 1,3712004 1,774 290 461 1,023 1,4842005 1,864 260 508 1,096 1,604Litres of plasma200018001600140012001000800600400200024510082672861025308Figure 2 - Graphical representation of the distribution of theproduction of plasma in litresRequirements for blood derivativesTable IV reports the data on the consumption ofalbumin, intravenous IgG and antithrombin III in the yearsstudied. The graph in figure 5 illustrates the trends in thesame data, after normalisation to the values in 2001.250101136029010232001 2002 2003 2004 2005YearPlasma for clinical usePlasma B+C sent for industrial fractionationPlasma A sent for industrial fractionation4612601096508The overall normalised needs, obtained by summingthe specific needs, are also represented graphically.As can be seen from this latter graph, the overallrequirements are increasing constantly (with a mean annualincrease of about 6%), even if the specific requirements foreach plasma derivative are obviously dependent on thepopulation of patients in the corresponding year.In particular, it can be seen that in 2002, the maximumrelative need for intravenous IgG (8,445 g) correspondedto the minimum requirements for both albumin (34,190 g)and antithrombin III (386,000 U.I.); similarly, in 2004, aminimum relative need for intravenous IgG (6,476 g)corresponded to maximum requirements for both albumin(44,000 g) and antithrombin III (550,000 U.I.).Covering the needsStarting from the industrial technical yields (Table I), itis clear that the increased production of plasma in our SIMTimproved the theoretical coverage of the need for albumin,which increased from 82.6 % in 2001 to 91.6 % in 2005.The data on intravenous IgG were less comforting: thetheoretical coverage of needs, which in 2001 was 61.4 %,decreased to 54.2 % in 2005 (Table V).As far as concerns antithrombin III, we calculated that23.9% of the requirements were covered in 2005.It should be noted that this is a partial result in that88Blood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-06085-92_bocciardi.p65 8805/07/2007, 11.20

Plasma derivative self-sufficency in LiguriaGENOVA GALLIERAGENOVA GASLINIPIETRALIGURE6%SAVONA12%GENOVA GALLIERA17%GENOVA GASLINI8%GENOVA RIVAROLOGENOVA S.MARTINOGENOVASAMPIERDARENAGENOVA SESTRILAVAGNA13%GENOVA RIVAROLO2%GENOVA VOLTRIIMPERIAGENOVA S.MARTINO9%LA SPEZIALA SPEZIA13%IMPERIA8%GENOVA VOLTRI3%GENOVASAMPIERDARENA6%GENOVA SESTRI3%LAVAGNAPIETRALIGURESAVONAFigure 3 - Pie chart of the production of plasma from ordinary blood donations in the Region of Liguria in 2005GENOVA GALLIERAPIETRALIGURE1%LAVAGNA9%LA SPEZIA 3%IMPERIA 2%GENOVA VOLTRI 3%SAVONA21%GENOVA SESTRI 8%GENOVA GALLIERA15%GENOVA RIVAROLO1%GENOVA S.MARTINO17%GENOVA GASLINI16%GENOVA GASLINIGENOVA RIVAROLOGENOVA S.MARTINOGENOVASAMPIERDARENAGENOVA SESTRIGENOVA VOLTRIIMPERIALA SPEZIALAVAGNAGENOVA SAMPIERDARENA 4%PIETRALIGURESAVONAFigure 4 - Pie chart of the production of class A plasma from apheresis in the Region of Liguria in 2005Blood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-0689085-92_bocciardi.p65 8909/07/2007, 9.47

Bocciardo L et alTable IV - Consumption of plasma derivatives by the Gaslini InstituteConsumption of Consumption of Consumption ofalbumin intravenous IgG antithrombin III2001 38,580 g 6,024 g 405,000 UI2002 34,190 g 8,445 g 386,000 UI2003 43,290 g 7,830 g 388,000 UI2004 44,000 g 6,476 g 550,000 UI2005 43,762 g 8,579 g 530,500 UITable V - Theoretical coverage of the requirements of plasma derivatives2001 2002 2003 2004 2005Albumin 82.62% 97.47% 79.18% 84.32% 91.63%Intravenous IgG 61.38% 45.78% 50.78% 66.45% 54.22%Antithrombin III 23.94%% (with reference to 2001)1601401201008060402002001 2002 2003 2004 2005YearAlbuminIntravenous IgGAntithrombin IIITotalFigure 5 - Requirements for albumin, intravenous IgG and antithrombin III, normalised to 2001 levels. Thevirtual overall requirements, calculated as the sum of the three products, is also reported, againnormalised to 2001 levelsantithrombin III was distributed after third party processingstarting from March of the same year.Finally, for each blood derivative considered, weevaluated the amount of plasma necessary to meet thetheoretical requirements.The graph in figure 6 shows the trends in the amountsneeded for each of these derivatives each year.It is clear from this that the blood derivative giving riseto the greatest request for plasma is intravenous IgG.DiscussionIn line with the national plan for reaching selfsufficiencyin plasma derivatives, the SIMT of the GasliniInstitute for children defined a plan to increase the volumeof plasma to send for industrial fractionation as an objectiveof its Quality System. This led to the definition of newprocedures for both production and clinical use of plasma.The increase in plasma production was obtainedwithout compromising the primary objective of the SIMT90Blood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-06085-92_bocciardi.p65 9009/07/2007, 9.47

Plasma derivative self-sufficency in Liguria35003000Litres of plasma2500200015001000500AlbuminIntravenous IgGAntithrombin III02001 2002 2003 2004 2005YearFigura 6 - The number of litres of plasma necessary to meet the demands for albumin, intravenous IgG andantithrombin III over the years studiedof maintaining self-sufficiency to meet needs for red cellconcentrates, but rather by optimising the type of donationon the basis of the users' characteristics. The increase inplasma from apheresis in 2002 was due to the increase indonations made with a cell separator. In 2003, theintroduction of multicomponent donations of plasma andplatelets led to a further increase in the volume of class Aplasma, without having to increase the number ofdonations.In September 2006, the technique of collecting highconcentrationplatelet concentrates during multicomponentdonations was introduced. It is estimated that thistechnique will further increase the production of categoryA plasma by about 400 litres/year. In addition, consideringthat about 25% of the units of plasma are assigned to lowweight patients, the use of paediatric units of 50 mL (unitsfor clinical use are currently fractioned into subunits of100-150 mL) will not only further reduce the risk oftransfusions (by enabling plasma from the same donor tobe used for more than one transfusion) but also save anestimated volume of more than 40 litres of plasma eachyear.The implementation of shared guidelines on the use ofthe blood components has, finally, allowed the clinical useof the plasma collected to be kept under control.Although the production of plasma has increasedconsiderably in the last few years, it is not yet sufficient tomeet the Institute's requirements for blood derivativeswhich are, in fact, guaranteed by the policy of the CRCC.The consumption of blood derivatives in our Instituteis high because the patients are children and because itcontains numerous highly specialised wards. The highconsumption of plasma derivatives is only partly undercontrol, because of the lack of agreed guidelines within theInstitute. However, the introduction of regional guidelineson the use of albumin in 2002 may have contributed to thefact that there was a relatively limited increase in theconsumption of albumin, in the three years 2003-2005. Onthe other hand, the use of intravenous IgG has alwaysbeen such that the theoretical coverage of needs is poor. Arecent review of the literature 2 showed that the number ofpublications on the use of intravenous IgG continues toincrease rapidly, although in reality there is little evidenceobtained from randomised, controlled clinical trials. Thishas led to a notable increase in off-label prescriptions,estimated, by the Food and Drug Administration, to accountfor between 50% and 70% of all prescriptions, even inorganisational settings such as in America, in which theuse of guidelines is consolidated 2 . Although the increasein the consumption of intravenous IgG is also due to theexpanded indications for use in a certain number of clinicalconditions, it is clear that the Institute must implement anagreed document regulating the use of this product. Theincreased consumption of immunoglobulins is aphenomenon present nationwide; our data do not, therefore,reflect the fact that our Institute cares for paediatricpatients, but rather confirm the change in the 'drivingproduct', already noted by other researchers 3 . Theconsumption of albumin is now no longer significantlygreater than that of the other plasma derivatives 4-6 , suchBlood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-0691085-92_bocciardi.p65 9109/07/2007, 9.47

Bocciardo L et althat it is mainly the consumption of intravenous IgG thatdetermines the amount of plasma to be sent for industrialprocessing.Precisely because of its widespread clinical use, thegoal of meeting demands for intravenous IgG is difficult toachieve, since it would require an annual increase of about1,350 litres in the production of plasma, which is equivalentto about 84% of the current production. If such an aimwere to be reached, it would be of great financial importance,given that the cost of intravenous IgG is much higher thanthat of albumin. It should be remembered that the greaterproductive efficiency, achieved in 2005-2006, by thecompany that processes the plasma, with the current yieldof intravenous IgG being 3.1 g/L, could contribute to greateravailability of the product, without lowering the level ofpurity of the drug.As far as concerns antithrombin III, the problem of thelow theoretical coverage of requirements (23.94%) derivesfrom the fact that this drug is produced only from class Aplasma. At a regional level, this makes Liguria dependenton the production in other regions belonging to the AIP,which collect a larger number of donations with cellseparators.In order to contribute to self-sufficiency in plasmaderivatives, it is essential to implement regionally-basedstrategies and, in some circumstances, interregional ones.These strategies must take into account factors related tothe type of population of the donors, logistical aspectsand agreements between the Associations that promotethe donation of blood. For this reason the role of the intraand extra-regional co-ordination, defined by law (LegislativeDecree. 219/05), is fundamental.ConclusionsThe analysis of the data collected leads to theconclusion that, although the production of plasma in thelast four years has increased and the consumption forclinical use has remained roughly stable, the volume ofplasma sent for industrial processing is insufficient to meetthe requirements for plasma derivatives: this is particularlyobvious for intravenous IgG, which are heavily requested.The quantification of this lack of coverage enabled usto plan procedures to reduce the shortfalls: the strategiesdecided were to further increase the production of plasmaand in particular the volume of plasma collected using acell separator, and to optimise the clinical use of plasmaand plasma derivatives.Our analysis also led us to conclude that theoptimisation of plasma production will never be sufficientto cover the notable increase in the consumption ofintravenous IgG in our Institute: it is, therefore, extremelyimportant to evaluate whether the increased requests aredue to an increase or improvement in health care servicesdelivered, or whether they are due to inappropriateconsumption of these plasma derivatives. In the light ofthe most authoritative international recommendations,collaboration with clinicians is essential in order to optimisethe use of plasma products and must have the concreteoutcome of the production of guidelines and the control ofthe appropriateness of the requests 2 .In fact, the good use of plasma derivatives, achievedby following appropriate guidelines, has obviousimplications for the safety of transfusions, but also hasclear financial consequences 7 .The SIMT of the Gaslini Institute, thanks to itsrationalisation of plasma production and optimisation ofthe consumption of plasma derivatives and plasma forclinical use, is participating actively in the attempt to reachregional self-sufficiency. Nevertheless, the CRCC isessential in order to evaluate the various realties in theregional context.References1) Delibera Giunta Regionale del Veneto n.3305, 15 September 1998.2) Darabi K, Abdel-Wahab O, Dzik WH. Current usage ofintravenous immune globulin and the rationale behind it: theMassachusetts General Hospital data and a review of theliterature. Transfusion 2006; 46: 741-53.3) Zucchelli P. I consorzi in convenzione con l'industria per lalavorazione del plasma. Convegno Interregionale dei ServiziTrasfusionali del Nord. Vol. Abstract. 2005; 96.4) Sala ML, Isacchini E, Busi E, et al. Autosufficienza di plasmae plasmaderivati in Regione Lombardia. La Trasf del Sangue2001; 46:130-6.5) Rossi D. Raccolta di plasma e consumo di plasmaderivati. IlServizio Trasfusionale 2004; 3: 32-3.6) Catalano L, Orlando M. Indagine conoscitiva sul consumo diplasmaderivati nelle regioni italiane: anno 1994. La Trasfusdel Sangue 1997; 4: 189-94.7) Gessi AM, Fabbri M, Farinato MA, Di Gregorio P. Usoimproprio di plasma fresco congelato ed emoderivati:incidenza sui DRG. Convegno Interregionale dei ServiziTrasfusionali del Nord. Vol. Abstract. 2005; 66.Received: 30 November 2006 – Revision accepted: 18 December 2006Correspondence: Dr. Diego ArdenghiSIMT, Istituto Gaslini, Genova, Italye-mail: diego.ardenghi@libero.it92Blood Transfus 2007; 5: xx-xx DOI 10.2450/2007.0033-06085-92_bocciardi.p65 9209/07/2007, 9.47

MATHEMATICAL TOOLSRemoval kinetics of exchange transfusionRoberto Reverberi, Lorenzo ReverberiServizio di Immunoematologia e Trasfusionale, Arcispedale S.Anna, Azienda Ospedaliera -Universitaria di Ferrara,ItalyExchange transfusion has just a few elective indications:the therapy of haemolytic disease of the newborn (HDN)and cases in which red cell exchange is appropriate but thechild is too small to use a cell separator.Exchange transfusion has several positive effects inthe therapy of HDN: it removes the neonatal red cells, whichare destined to be destroyed and hence to generatebilirubin, the already produced bilirubin, and the offendingantibody 1 . Moreover, exchange transfusion also correctsanaemia, if present 1 .Physiological bases of exchange transfusionin HDNThe offending antibody is an IgG and has a substantialextravascular distribution (about 60%). As the equilibrationbetween the intravascular and extravascular compartmentsis slow (≈ 1-3%/hour) 2 , removal of the antibody throughthe exchange may only be partial and the residual or newlyproduced neonatal red cells are still exposed to the risk ofhaemolysis. It is, therefore, justified to remove as manyneonatal red cells as possible and to maintain an elevatedtotal haemoglobin concentration with compatible adult redcells, so as to inhibit the autologous production from thebone marrow. The bilirubin in the intravascular space isonly a small part of the total and, although equilibration isfast, exchange transfusion is not particularly successful:rebound is immediate 1 , so that at the end of an exchangeduring which 87% of neonatal red cells are removed, theconcentration of bilirubin is still 60% of that at the start ofthe exchange 3 . It is, therefore, advisable to prevent theproduction of bilirubin by removing the "doomed" red cellsin time 1 .The removal of red cellsAlexander S. Wiener (1907-1976), co-worker of KarlLandsteiner in the discovery of the Rh factor (1940) andBlood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-07© SIMTI Servizi Srlpioneer of the therapy of HDN, asserted 4 that he and Wexlerhad been the first to derive the formula which describesthe removal of red cells during exchange transfusion 5 :e −Residual fraction ,=where v is the total exchanged volume, V is the patient'sblood volume and e is the transcendental number, base ofthe natural logarithm. However, he later stated that hisformula was not correct because it did not take intoconsideration the difference between the patient'shaematocrit and that of the transfused blood 6 . Moreover,according to others, even when the two haematocritshappen to be the same, the formula should be modified toaccount for the difference between venous haematocritand body haematocrit 7 . Wiener also stated that it wasobvious that, for an equal exchanged volume, exchangetransfusion would be more efficient, the more the neonatewas anaemic 6 . Taken together, these opinions riskmisleading the reader seriously.Wiener's formula describes a continuous flow exchange.However, during exchange transfusion, neonatal blood isreplaced by donor blood, withdrawing and transfusingalternatively 5-20 mL of blood. The exchange is, therefore,intermittent. It can be demonstrated (see Appendix A) thatwhen blood is first withdrawn and then transfused, at eachcycle:Residual fraction , (1)where s is the volume exchanged at each cycle (cyclevolume), V is the patient's blood volume, k is the ratiobetween body and venous haematocrit (= 0.91) 7 and n isthe number of exchange cycles.vV⎛= ⎜1 −⎝s ⎞⎟kV ⎠n93093-101_reverberi.p65 9305/07/2007, 11.24

Reverberi R , Reverberi LAnalogously, when blood is first transfused and thenwithdrawn:⎛= ⎜1 −⎝sk(V +⎞⎟s)⎠Residual fraction , (2)In Appendix B, the reader will find instructions forprogramming a spreadsheet to calculate the residual fractionaccording to the cycle volume and the neonate's bloodvolume. Some exemplary data are shown in table I.As can be easily noted, the efficiency of the exchangeis greater when blood is first withdrawn. The difference isproportional to the volume exchanged at each cycle. Thecycle volume influences the efficiency of the exchange in apeculiar way: when blood is first withdrawn, efficiency isproportional to the cycle volume, whereas when blood isfirst transfused, efficiency is inversely proportional to thecycle volume. However, as the cycle volume rarely exceeds10% of the total blood volume, the differences are limitedto a few percentage points.Formulae (1) and (2) do not include references to thepatient's haematocrit, nor to the haematocrit of thetransfused blood: contrary to the above cited opinions,the residual fraction only depends on the terms representedin the two formulae.Absolute amount of neonatal red cellsHowever, we should be more interested in the absoluteamount of the residual neonatal red cells, rather than thefraction. In this case it is clear that, if the neonate is anaemic,during the exchange it is possible to fall under a minimumntarget level of neonatal red cells earlier.Incidentally, this was the probable meaning of Wiener'ssentence cited above.Figure 1 compares the cases of two hypotheticalneonates, with the same weight (blood volume) but initialhaematocrits of 25% and 50%. If our aim is to leave no morethan 9 mL of neonatal red cells (about 10% of the normalinitial volume), the target is met in the first case after 14cycles of exchange and in the second after 20 cycles.The neonate's venous haematocrit at the endof the procedureDuring the exchange transfusion, neonatal red cells areprogressively substituted by the transfused cells. Thevenous haematocrit after n cycles of exchange (H n)principally depends on the haematocrit of the transfusedblood (H D) and, moreover, the volume exchanged at eachcycle (s), the number of cycles (n), and the neonate's initialvenous haematocrit (H V), according to the following formula(see Appendix A):nns ⎞ ⎛ s ⎞HnH ⎜ ⎛ ⎞⎜⎛ = 1 − ⎟ ×⎟V + 1 −×kV⎜ 1 − ⎟⎝ ⎠kV⎝ ⎝ ⎠ ⎠(3)The above formula is appropriate when the exchangebegins withdrawal of blood. The analogous formula for theopposite case is given in Appendix A.Figure 2 shows the kinetics of the substitution of thetransfused blood for the neonatal red cells and the changesof the patient's venous haematocrit during the exchangetransfusion.HDTable I -Fraction of neonatal red cells remaining after 1, 2, 5, 10, and 20 cycles during the exchangetransfusion. It is assumed that the volume exchanged at each cycle is 10% of the total bloodvolume. Columns "Withdrawal first" and "Transfusion first" show the results when the exchangebegins with withdrawal or transfusion, respectively. For the sake of comparison, the column on thefar right shows the results of an isovolumetric continuous flow exchange. Its efficiency is intermediatebetween the other twoCycle Total exchanged volume Residual fraction (%) of the neonatal red cells(% of the blood volume) Intermittent flow Continuous flowWithdrawal first Transfusion first1 10 89.0 90.0 89.52 20 79.2 81.0 80.15 50 55.9 59.1 57.410 100 31.2 34.9 32.920 200 9.8 12.2 10.894Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-07093-101_reverberi.p65 9405/07/2007, 11.24

Removal kinetics of exchange transfusionResidual neonatal RBC(mL)10090807060504030201000 20 40 60 80 100 120 140 160 180 200 220 240Total exchanged volume (%)Figure 1 - Effect of the neonate's initial venous haematocrit on the residual quantity of neonatal red cellsduring the exchange transfusion. It is assumed that the blood volume of the neonate is 200 mL,the volume of the exchange at each cycle is 20 mL (10%), and blood is first withdrawn and thentransfused. The solid curve with closed symbols refers to an initial venous haematocrit of25%. The dotted curve with open symbols refers to a haematocrit of 50%Circulating RBC (mL)10090807060504030201000 20 40 60 80 100 120 140 160 180 200 220 240Total exchanged volume (%)Figure 2 - Kinetics of the substitution of donor red cells for the neonatal red cells and of the venoushaematocrit during the exchange transfusion. The conditions of the procedure are thesame as those imagined in figure 1 (initial venous haematocrit of the neonate: 25%). Thehaematocrit of the transfused blood is 50%. The solid curve with closed symbols indicatesthe residual volume of neonatal red cells. The dotted curve with closed symbols indicatesthe transfused red cells circulating in the patient. The solid curve with open symbolsindicates the patient's venous haematocrit.Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-0795093-101_reverberi.p65 9509/07/2007, 9.53

Reverberi R , Reverberi LChoosing the haematocrit of the blood to betransfusedIt has been noted that the haematocrit of the transfusedblood does not influence the efficiency of the procedureand, therefore, it may be chosen on the basis of otherconsiderations: the haematocrit should not be too low, inorder not to induce a state of anaemia and to correct it, ifpresent; an elevated haematocrit at the end of the procedureexerts an inhibitory effect on autologous marrowproduction, which is in any case destined to precociousdestruction. On the other hand, an excessive haematocritof the transfused blood limits the total exchanged volume,lest the neonate be made too polycythaemic 6 . This entailsthe risk of an insufficient exchange 6 . Since the 1950s, it hasbeen customary to concentrate the reconstituted wholeblood up to a haematocrit of about 50% 6 . This practice wasconfirmed by the recent Italian guidelines 8 . Probably, thischoice stems from the fact that the normal neonatalhaematocrit is about 50%. Such an elevated haematocrit isdue to the greater affinity for oxygen by the foetalhaemoglobin that, therefore, has a lower oxygen carryingcapacity to the tissues. In fact, haematocrit rapidlydecreases in the weeks following birth. After about 2months, even though the substitution by adulthaemoglobin is not complete, the haematocrit has fallen toabout 35% 9 . Therefore, a haematocrit of 50% is notnecessary and it could be decreased to 45% or less, so asto profit from the greater volume of reconstituted wholeblood available to increase the removal of neonatal redcells or bilirubin.Exchange transfusion in the therapyof sickle cell anaemiaExchange transfusion is used preferentially or as analternative to simple transfusion in a number of indicationsin children with sickle cell anaemia 10 . In adults or biggerchildren, automated red cell exchange is the preferredprocedure (this topic will be dealt with in the next article ofthis series), but the manual exchange of whole blood maybe necessary for smaller children, given the limitations inthe extracorporeal volume. In this case, the removal kineticsis the same as that described above. The difference lies inthe target, which is not to remove as many autologous redcells as possible, but to reduce the percentage ofhaemoglobin S without increasing the haematocritexcessively, in order not to cause a high blood viscosity 11 .The relevant parameters are, therefore, the percentageof haemoglobin S and the venous haematocrit. This lattercan be calculated from (3). The first part of equation (3)represents the contribution (H p) of the residual autologousred cells (see Appendix A):⎛ s ⎞Hp = ⎜ 1 − ⎟ × HV⎝ kV ⎠Therefore, the percentage of haemoglobin S after ncycles (HbS n) will be:HpHbSn = × HbS0Hwhere HbS 0is the initial percentage of haemoglobin S.Limitations of the formulaeMost of the terms contained in the formulae can bemeasured with precision. However, blood volume is, usually,only estimated with the aid of a formula 12 and the value ofthe correction for the body haematocrit is taken from theliterature 7 . In this regard, however, it should be noted thatthe value may differ from the normal significantly, in caseof splenomegaly 13 .Moreover, the formulae assume that:- there is no dead space in the transfusion set used forthe exchange;- the transfused blood mixes immediately in thecirculation;- the blood volume does not change during theprocedure;- all transfused red cells survive in the circulation (atleast till the end of the exchange);- the patient's red cells are not destroyed during theprocedure.As regards the first assumption, after the transfusionstep, before blood is again withdrawn, it is advisable toaspirate a small volume of blood from the patient and toreinfuse it. The second and third assumptions are plausible.The fourth is also plausible, in the case of HDN, becausethe neonate's reticuloendothelial system shouldpreferentially remove the sensitised autologous red cells.Finally, the fifth assumption depends on the duration ofthe procedure.ConclusionsThe exchange is more efficient if blood is first withdrawnand if the volume exchanged at each cycle is large. However,when the cycle volume does not exceed 10% of the patient'snn96Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-07093-101_reverberi.p65 9609/07/2007, 9.53

Removal kinetics of exchange transfusionblood volume, the difference remains within a fewpercentage points.The removed fraction of the patient's red cells onlydepends on the cycle volume and the number of cycles.However, the absolute amount of residual red cells alsodepends on the patient's initial venous haematocrit. Thehaematocrit of the transfused blood influences the finalhaematocrit of the patient, but not the removal efficiency.Keywords: exchange transfusion, exchange kinetics, haemolyticdisease of the newborn (HDN), sickle cell anaemia.References1. Nathan DG, Oski FA. Hematology of Infancy and Childhood.3rd ed, Philadelphia, PA, WB Saunders, 1987.2. Kaplan AA. A Practical Guide to Therapeutic PlasmaExchange. Malden, MA, Blackwell Science, 1999.3. Greenough A. Rhesus disease: postnatal management andoutcome. Eur J Pediatr 1999; 158: 689-93.4. Wiener AS. Blood group museum. II. Some mathematicalcuriosities and amusements. Am J Clin Pathol 1964; 3:307-8.5. Wiener AS, Wexler IB. The use of heparin when performingexchange blood transfusions in newborn infants. J Lab ClinMed 1946; 31:1016-9.6. Wiener AS, Wexler IB, Brancato GJ. Treatment oferythroblastosis fetalis by exchange transfusion. Statisticalanalysis of results. J Pediatr 1954; 5:546-68.7. Mollison PL, Engelfriet CP, Contreras M. Blood Transfusionin Clinical Medicine. 10th ed, Oxford, Great Britain, BlackwellScience, 1997.8. SIMTI and SIN Task Force. Recommendations on transfusiontherapy in neonatology. Blood Transfus 2006; 4:158-80.9. Greer JP, Foerster J, Lukens JN, et al. (Eds). Wintrobe'sClinical Hematology. 11th ed. Philadelphia, Pa, LippincottWilliams & Wilkins, 2003.10. Amrolia PJ, Almeida A, Halsey C, et al. Therapeutic challengesin childhood sickle cell disease. Part 1: Current and futuretreatment options. Br J Haematol 2003; 120: 725–36.11. Davies SC, Roberts-Harewood M. Blood transfusion in sicklecell disease. Blood Reviews 1997; 11: 57-71.12. Lentner C (ed). Geigy Scientific Tables. Vol. 2. 8th ed, Basle,Switzerland, Ciba-Geigy, 1987.13. Mollison PL, Engelfriet CP, Contreras M. Blood Transfusionin Clinical Medicine. 8th ed, Oxford, Great Britain, BlackwellScientific Publications, 1987.Appendix A – Derivation of the formulaeThe fraction of neonatal red cells removed inthe exchange transfusionIf we denote the volume exchanged at each cycle with sand the blood volume of the patient with V, then the fractionremoved at each cycle, when blood is first withdrawn,sshould be . VHowever, the red cells are more concentrated in thevenous blood than in the whole vascular system, so that:HB= k × H Vwhere H Bis the body haematocrit, H Vthe venoushaematocrit and k is the correction factor (= 0.91) 7 .Let R Bbe the total volume of the patient's red cells:HBHVRB = × V = × k × V100 100(1).Let R Sbe the volume of the patient's red cells in thevolume exchanged at each cycle:RHVHB= × s =100 1001×kS ×Then the removed fraction will be equal toRemoved fractionand the residual fraction will be:Residual fractionss= ,kV=1 −skV., or:At the next cycle, the volumes exchanged will be thesame, but the removal will concern the residual amountonly s − kV 1 , therefore the residual fraction after the secondcycle will be:R SR BReceived: 10 January 2007 – Accepted: 24 April 2007Correspondence: Dr. Roberto ReverberiServizio di Immunoematologia e Trasfusionale Arcispedale S.AnnaCorso Giovecca 20344100 Ferrara - Italye-mail: sitfe@ospfe.its s − × 1− kV kV 1 , or1− kVs2.Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-0797093-101_reverberi.p65 9709/07/2007, 9.53

Reverberi R , Reverberi LIt is easy to see that after n cycles:Residual fraction s = 1 − (2). kV When blood is first transfused, on the basis of verysimilar considerations to those above, it can be derivedthat:Residual fraction= ns − k(V + s)1 (3).In the above formulae, only the correction for thedifference between venous and body haematocrit was used.However, a further correction is advisable when thehaematocrit is measured by centrifugation because of thesmall amount of plasma (≈ 2%) that remains entrapped inthe red cell column 8 . In this case k = 0.91 x 0.98 = 0.89.Absolute amount of neonatal red cellsremaining after the exchange transfusionThe volume of the neonatal red cells remaining afterthe exchange cycles (R R) can be derived from formula (2):nR Bcan be derived from (1). The removed volume can beeasily calculated by difference.Venous haematocrit after the exchangetransfusionLet us consider the case in which blood is firstswithdrawn. Let a = and b =1 − a .kVAt each cycle, a fraction equal to a of the donor's redcells present in the circulation will be removed, whilst avolume D Swill be added.Table II shows the situation in the first cycles of anexchange transfusion.As can be seen from the far right column, at each cycle,a term to the power n – 1 is added in parentheses.By definition, we know that:HD=DsSDS× 100 = × 100kV( 1 − b)where H Dis the haematocrit of the transfused blood, andthat:2n−1DSHd= ( 1+b + b + ... + b ) × × 100 ,kVRR s = RB× 1 − , kV nwhere H dis the venous haematocrit (in the patient’scirculation) of the transfused red cells after cycle n.Therefore:or formula (3), as appropriate:d2n−1n( 1 − b) × ( 1 + b+b + ... + b ) × HD= ( 1 − b ) HDH =×RR=RB× s − k(V + s)1 ,nTable II - Volumes of donor red cells (V GR) exchanged during the exchange transfusion. At each cycle, blood is first withdrawn andthen transfused. In the first cycle, no donor red cells are in the circulation during the withdrawal step. D Sis the volumeof donor red cells transfused at each cycleWithdrawal stepTransfusion stepCYCLE Removed V GRRemaining V GRFinal V GR1 0 0 D S2 a × D Sb × D SD S+ b × D S3 a × (1 + b) × D S(b + b 2 ) × D S(1 + b + b 2 ) × D S4 a × (1 + b + b 2 ) × D S(b + b 2 + b 3 ) × D S(1 + b + b 2 + b 3 ) × D S… … … …n a × (1 + b + b 2 +…+ b n-2 ) × D S(b + b 2 +…+ b n-1 ) × D S(1 + b + b 2 +…+ b n-1 ) × D S98Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-07093-101_reverberi.p65 9809/07/2007, 9.53

Removal kinetics of exchange transfusionBy substitution, the following formula is obtained:Hd = 1 ×kV n s 1 − − HD .In the same way, we can derive the venous haematocritof the patient’s red cells after the exchange transfusion(H p): s Hp = 1 − × HV. kV Therefore, at the end of n cycles of exchange transfusion,the total venous haematocrit will be the sum of the twopartial ones 7 :nAppendix B – Programming a spreadsheet*Spreadsheets are programmes used to performcalculations, especially on large data series.The most renowned is Excel, which is part of the MSOffice (Microsoft) package.However, the following instructions are equally validfor other similar programmes, such as OpenOffice.org Calcand Gnumeric. As these are less well known, here is a briefdescription.OpenOffice.org CalcThis is part of the OpenOffice.org suite. It is freelydownloadable from http://www.openoffice.org/ and isavailable for Windows (98 and later), Mac Os X, Linux andother operating systems. Calc can open, modify, and savefiles in Excel and other formats, besides its own native one.The available functions are similar to those of Excel. Hn = − 1 × kV kV nns sH 1 ×V + 1 − − HD.This formula allows the calculation of the venoushaematocrit of the patient during the exchange transfusion(H n), starting from the patient’s venous haematocrit beforethe procedure (H V) and the haematocrit of the transfusedblood (H D).Analogously, when blood is first transfused, thefollowing formula can be derived:nn s s H nH = 1− V × + 1 − 1− × H D k( V + s) k( V s) + .GnumericGnumeric is part of the GNU project. It can be freelydownloaded from http://www.gnome.org/projects/gnumeric/ and is available in versions for Windows (XP,2000 or later) and Linux. Gnumeric is able to open, modifyand save files in Excel and other formats. Availablefunctions are similar to those of Excel but, in addition,Gnumeric offers a series of statistical tests.The following instructions only presuppose a basicknowledge of how to use a personal computer.InstructionsOpen the spreadsheet. The programme automaticallyproposes a new sheet (table), with rows progressivelynumbered starting from 1 and columns alphabeticallylabelled starting from A.Each cell is identified by a couple of column and rowidentifiers: e.g., the top leftmost cell is A1; the fourth fromtop in the B column is B4 and so on. Each cell may containnumerical values, formulae or explanatory text. This last isuseful for reminding the user of the meaning of the adjacentvalues or formulae.Calculation of the residual fraction andabsolute quantity of the patient's red cellsEnter the text listed in Table III into the appropriatecells. Enter the values listed in Table IV.* Italian readers using the localized (Italian) versions of the spreadsheets shouldfollow the instructions in the Italian translation of this paper, which is availableon line at http://www.transfusionmedicine.org/. Briefly, "potenza" should besubstituted for "power" and ";" should be substituted for ",".Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-0799093-101_reverberi.p65 9909/07/2007, 9.53

Reverberi R , Reverberi LTable III - Explanatory text ("labels") to be entered into thespecified cellCellText to be enteredA1 Kinetics of exchange tansfusionA3 Patient's blood volume (mL)A4 Volume of each exchange cycle (mL)A5 Patient's initial venous haematocrit (%)A6 Haematocrit of the blood to be transfused (%)A7 Correction factor for the body haematocritA9 CycleB9 Exchanged volume (% of the blood volume)C9 Residual fraction (%)D9 Residual quantity (mL)formulae in an appropriate way. Calculated values appearimmediately. Row 10 contains the starting values, beforethe exchange. Cells B10-B35, C10-C35 and D10-D35 containthe total exchanged volume, expressed as a percentage ofthe blood volume, the residual fraction (%) and the residualabsolute amount (mL) of the patient's red cells, respectively.If the user changes one or more values in cells B3-B7, theresults in the cells B10-B35, C10-C35 and D10-D35 areimmediately recalculated by the programme.When blood is first transfused, the formula to be enteredinto cell B10 is:=100*POWER(1-$B$4/($B$7*($B$3+$B$4)),A10) .Table IV - Values to be entered into the specified cells. Pleaseenter the numerical values only, not the textCellValue to be enteredB3 Blood volume of the patient (mL), e.g. 200B4 Volume of each exchange cycle (mL), e.g. 20B5 Initial haematocrit of the patient, e.g. 25B6 Haematocrit of the transfused blood, e.g. 50B7 Correction factor for body haematocrit: 0.91If the cell seems to be covered by the text of the cell atits left, resize the adjacent cell or begin to write in thedesired cell: this will prevent the text in the left cell fromexceeding the border.Enter the numbers from 0 to 25 into the cells from A10to A35.These numbers represent the cycles of the exchangetransfusion. The number of cycles can be increased furtherat will. Enter the formulae listed in Table V.The other instructions are the same.Calculation of the patient's venoushaematocrit after the exchange transfusion(The following instructions presuppose that the previousones have been correctly performed).Enter text and formulae contained in Table VI.Table VI - Instructions to calculate the venous haematocrit ofthe patientCellE9Text to be enteredPatient's venous haematocritFormula to be enteredE10 =$B$5* POWER(1-$B$4/$B$3/$B$7,A10) +$B$6*(1- POWER(1-$B$4/$B$3/$B$7,A10))Table V - Formulae to be entered into the specified cells.Formulae should be entered as plain text, withoutforgetting the initial "=" signCellB10C10D10Formula to be entered=$B$4/$B$3*A10*100=100* POWER(1-$B$4/$B$3/$B$7,A10)=$B$5/100*$B$7*$B$3*C10/100Right click on cell B10; choose "Copy" from the popupmenu; paste the cell ("Paste") on cells B11 to B35 (orfurther on, if numbering continues past A35).In the same way, copy the formula in C10 on cells C11to C35 and that in D10 on cells D11 to D35. The spreadsheettransposes the references to the cells contained in theCopy cell E10 into cells E11-E35 (or further on, ifnumbering continues past A35). Cells E10-E35 will showthe venous haematocrit of the patient after each exchangecycle.If blood is first transfused, the formula to be enteredinto E10 is:=$B$5*POWER(1-$B$4/($B$3+$B$4)/$B$7,A10)+$B$6*(1-POWER(1-$B$4/($B$3+$B$4)/$B$7,A10)) .Calculation of the percentageof haemoglobin S(The following instructions presuppose that the previousones have been correctly performed).100Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-07093-101_reverberi.p65 10009/07/2007, 9.53

Removal kinetics of exchange transfusionEnter the text, the value, and the formula contained inTable VII.Table VII - Instructions to calculate the residual percentage ofhaemoglobin SCellA8F9Text to be enteredInitial percentage of HbSResidual percentage of HbSValue to be enteredB8 Initial percentage of HbS, e.g. 70Formula to be enteredF10=$B$5* POWER(1-$B$4/$B$3/$B$7,A10) /E10*$B$8Copy cell F10 into cells F11-F35 (or further on, ifnumbering continues past A35). Cells F10-F35 will showthe percentage of haemoglobin S after each exchange cycle.If blood is first transfused, the formula to be enteredinto F10 is:=$B$5* POWER (1-$B$4/($B$3+$B$4)/$B$7,A10) /E10*$B$8 .Blood Transfus 2007; 5: 93-101 DOI 10.2450/2007.0018-07101093-101_reverberi.p65 10109/07/2007, 9.53

CASE REPORTAn acute haemolytic transfusion reaction due to anti-Jk aMaria Antonietta Villa 1 , Marilyn Moulds 2 , Elena Beatrice Coluccio 1 , Mara NicolettaPizzi 1 , Cinzia Paccapelo 1 , Nicoletta Revelli 1 , Fernanda Morelati 1 , Francesca Truglio 1 ,Maria Cristina Manera 1 , Alberto Tedeschi 3 , Maurizio Marconi 11U.O. Centro Trasfusionale e di Immunoematologia, Dip. di Medicina Rigenerativa, Fond. Osp. MaggiorePoliclinico, Mangiagalli e Regina Elena, Ist. di Ricovero e Cura a Carattere Scientifico, Milan, Italy2Education Services, Immucor Inc., Norcross, GA, USA3U.O. Medicina Interna II, Fond. Osp. Maggiore Policlinico, Mangiagalli e Regina Elena, Ist. di Ricovero e Curaa Carattere Scientifico, Milan, ItalyIntroductionThe Kidd system antibodies are characteristicallydifficult to detect. They show variability in immunoglobulinclass, subclass and serological characteristics. They aregenerally detected by an antiglobulin test, using apolyspecific antiglobulin or complement antiserum. Often,the antibodies are only detected using cells with a doubledose (homozygous) expression of Kidd antigens, enzymetreatedcells or by using sensitive immunohaematologicaltechniques.Case reportA 73-year old woman, with a history of two pregnanciesand no red cell transfusions, was admitted to our hospital.She had severe anaemia, cirrhosis related to hepatitis Cvirus infection, cryoglobulins, mild ascites infection andmild renal failure. There were no reported incidents of redcell immunisation.On admission, her haemoglobin concentration was7.5 g/dL and her haematocrit 23%.On day 3 of hospitalisation, the first red blood cell (RBC)transfusion was required and performed (day 0).The patient's group was A 1B, Rh+.The antibody screening was negative, using ourstandard automated method for pre-transfusion testing(AutoVue System with Ortho BioVue microcolumn, Ortho-Clinical Diagnostics, Inc., Raritan, New Jersey, USA) withpolyspecific anti-human globulin (anti-IgG+C3d) and anEDTA plasma sample. She received an ABO/Rh compatiblestandard packed red cell unit (about 170 mL of packed redcells) with a Type & Screen procedure, as indicated bynational legislation.The post-transfusion level of haemoglobin was 8.4 g/dL.Additional pre-transfusion tests were performed on day14, although no transfusion of RBC units was performed.On day 19 after the first RBC transfusion, two RBCunits were requested because the woman's haemoglobinhad decreased to 7.2 g/dL.The antibody screening tests were still negative,according to our standard automated method. The sameday of the request, she received one AB, CCDee standardpacked red cell unit, with the Type & Screen procedure.The transfusion was interrupted 2.5 hours after beingstarted because of a transfusion reaction: chills, lumbarpain and dark red urine.The pre-transfusion and post-transfusion data indicatethat there was no significant change in body temperature(pre-transfusion +37.2 °C - post-transfusion +37.5 °C) andminor modifications of blood pressure (pre-transfusion 130/65 - post-transfusion 140/85) and heart rate (pre-transfusion78 beats per min - post-transfusion 88 beats per minutes).Dark red urine was still observed 24 hours after thereaction.To determine the cause of the post-transfusionhaemolysis, immediately after the reaction, the patient'spost-transfusion serum and plasma samples were inspectedfor evidence of haemolysis and compared with the plasmapre-transfusion sample, using the scale of values proposedby Elliot 1 .The post-transfusion samples were grosslyhaemolysed, with a dark red hue similar to haemolysis of200 mL of RBC in 3,000 mL of plasma.The post-transfusion biochemical values (Figure 1)revealed an increase in free plasma haemoglobin from 18.0mg/dL to 260 mg/dL (reference value

Acute reaction due to anti-Jk a300Free plasmatic hemoglobin (mg/dL)2502001501005000 14 19 19 20 21 32Day from first transfusionFigure 1 - Free plasmatic haemoglobin levels. The clinical symptoms of an acute haemolytic post-transfusionreaction were confirmed by increases of free plasmatic haemoglobin from 18 mg/dL to 260 mg/dL(reference value

Villa MA et al.stabilised ficin-solution (Ortho-Clinical Diagnostics) assuggested by the manufacturer.5) Saline 20 °C: 100µL of plasma or serum, one drop of a3% red cell suspension, 30 minutes of incubation at20 °C, centrifugation for 1 minute and macroscopicreading.6) Solid-phase IAT using a commercial panel foridentification (Ready-ID ® , Extend I and II, Immucor),100µL of LISS additive (Capture ® LISS, Immucor), 50µLof plasma or serum, 20 minutes of incubation at 37 °C,six washes with saline, addition of one drop of indicatorcells (Capture-R ® Indicator Red Cells, Immucor),centrifugation and macroscopic reading.Instead, for cross-matching (Capture-R ® Select,Immucor), we prepared a 0.3-0.5% suspension of wellwashed red blood cells in saline, 50µL of the red bloodcell suspension, centrifuged the strip, carried out sixwashes with saline, and proceeded with LISS andplasma or serum as described for identification.7) Erythrocytes Magnetized ® technology 3 (ScreenLys,Diagast, France), which uses IgG-coated plates. Toprevent the neutralisation of anti-IgG, 60µL of low-ionicand diluent (NanoLys and Screen Diluent, Diagast)solutions were dispensed before the addition of samplesand RBC (12 µL plasma, 15µL of a 1% three-cell panel,Hemascreen, Diagast).After incubation, without any washing/centrifugation,magnetisation was performed and the plates were placedon a magnetic workstation (FreeslyNano, Diagast) thatallowed the adherence of sensitised-magnetic RBC.8) Typing of the red cells of the patient and transfusedunits was performed using the standard agglutinationtube methods.ResultsDuring our investigations, at the time of collectingsamples, no haemodilution was performed in the patientby infusion of saline or other liquid.Table I - Results of serological investigationsTime IAT*-column Cross-match DAT† Identification No. of AHTR Eluate(standard screening) (tube) of antibodies transfused in IATin serum RBC units testDay 0 Negative No, T&S Negative Nd 1 (Packed RBC) No ndDay 14 Negative No Negative anti-Jk a only ndin the SP (‡)and EMT(#)Day 19 Negative No, T&S Negative anti-Jk a 1 (Packed RBC) Yes ndPre-transfusiononly in the SP(‡)and EMT(#)Day 19 Negative No, T&S Negative anti-Jk a anti-Jk a onlyPost-transfusion - Positive in SP(‡) only in the SP(‡) in SP(‡)performed afterand EMT(#)AHTR($)Day 20 Negative No Negative anti-Jk a NegativePost-transfusiononly in the SP(‡)and EMT(#)Day 21 Weak (score ±) No Negative anti-Jk a ndPost-transfusion only with Jk(a+b-) reactive in IAT-columnonly with 2 out of 4 Jk(a+b+)red cell suspensions testedDay 25 Negative No Negative ndPost-transfusionDay 32 Negative No Negative ndPost-transfusion(*) IAT: indirect antiglobulin test;(†) DAT: direct antiglobulin test;(‡) SP: solid-phase method;(#) EMT: erythrocyte magnetised technology;($) AHTR: Acute Haemolytic Transfusion Reaction.104Blood Transfus 2007; 5: 102-106 DOI 10.2450/2007.0007-07102-106_villa.p65 10405/07/2007, 11.27

Acute reaction due to anti-Jk aThe results of our investigation are reported in table I.An anti-Jk a was detected only by the solid-phase methodand Erythrocytes Magnetized ® technology in the sampleof day 14 (after the first transfusion) while all other methodsgave negative results. No differences were detectedbetween serum and EDTA plasma. Two out of four cellsJk(a+b-) show weak positive results for the presence ofanti-Jk a by the column agglutination method, using theplasma sample of day 21 (2 days after the reaction).These two cells had a double dose expression of thetarget antigen. This result was not confirmed by furtherinvestigations on the samples drawn on day 25 and day 32after the first transfusion.The tube DAT was negative for all samples tested.Despite this, we performed an elution from pre-transfusionand post-reaction blood samples (day 19) and tested themby tube PEG-IAT and by the solid phase method.The anti-Jk a was detected in the post-transfusion eluate,using only the solid phase method.No antibodies were detected in any samples bymicrocolumn agglutination, using serum with freshcomplement added.The patient's RBC phenotype, determined on pretransfusionand on all post-transfusion samples, was A 1BCCDee; Jk(a-b+); Fy(a+b+); M+N+S+s+.Samples from the first RBC unit transfused and the unitinvolved in the reaction were available and typed asJk(a+b+) and Jk(a+b-), respectively.The antibody screening test was negative in both blooddonors using the solid-phase method and positive resultsin cross-matches were detected only in solid-phase, usingthe pre and post-reaction samples (day 19), while resultswere negative in solid-phase using selected AB, CCDee,Jk(a-b+) units.After our investigations, the patient recovered andneeded no further transfusions.DiscussionAfter the discovery of the first anti-Jk a , many casesinvolving Kidd system antibodies have been reported anda substantial number have been implicated in AHTR. Kiddantibodies are usually IgG or a combination of IgG andIgM; pure IgM examples are rare.IgM Kidd antibodies can bind the complement and areconsequently detected only by broad spectrum antiglobulinreagents, containing an anti-complement component.It has been generally considered that the major antibodycomponent of Kidd antibodies must be IgG withcomplement-fixing ability.Current guidelines for pre-transfusion testing 4 indicatethat antiglobulin reagents with a potent anti-IgG caneffectively detect such antibodies, without the need of ananti-complement component.This has encouraged Blood Transfusion Centres to usetechniques which compromise the detection of thecomplement, on the basis that the antibodies will bedetected by the anti-IgG component of the antiglobulinreagent.We report a case in which the IgG component of analloantibody-Jk a was demonstrated only in solid-phase andby Erythrocytes Magnetized ® technology, but not detectedin the patient's plasma or serum, when tested bymicrocolumn agglutination with broad spectrumantiglobulin or in tubes with anti-IgG antiglobulin reagent.The antibody was the cause of an AHTR associatedwith the transfusion of Jk(a+) blood units.The use of plasma rather than serum was not the causeof false negative results in the initial screening tests,although the column agglutination tests (using geltechnology) can compromise the detection of Kiddantibodies 5 .This problem may sometimes be associated with shearforces 6 .Because the primary immune response could havealready occurred with a previous pregnancy, the firsttransfusion of random Jk(a+b+) could have caused asecondary immune response, that resulted in thehaemolysis of the second Jk(a+b-) transfused unit.As recently reported by the results of a largeretrospective multicentre study 7 , we determined that theanti-Jk a was the cause of a secondary response in a shortinterval of time after transfusion (14 days after the firsttransfusion); the level of antibody was such that serologicaldetection was only possible using the most sensitivetechnique.This same circumstance was reported by Callahanet al. 8 : the Authors described an anti-Jk b , detected only insolid-phase and not by PEG, gel or LISS techniques, whichcaused a delayed haemolytic transfusion reaction in apatient affected by sickle cell disease.Naturally occurring anti-Jk a antibodies, detectable onlyin solid-phase, were previously reported by Rumsey et al. 9and the use of manual polybrene 10 was the method of choicefor the detection of weak anti-Jk a .During the investigations, 2 days after our patient'sreaction, the antibody increased to levels detectable byour standard method and then immediately declined to apoint at which antibody was no longer detectable inBlood Transfus 2007; 5: 102-106 DOI 10.2450/2007.0007-07105102-106_villa.p65 10505/07/2007, 11.27

Villa MA et al.microcolumn agglutination.This case demonstrates one example of a haemolytictransfusion reaction due to anti-Jk a , which was notdetectable with a highly sensitive routine test method(microcolumn agglutination).Our data document the well-known difficulties in thedetection of Kidd antibodies and highlight the importanceof using additional sensitive techniques and multiplemethods, particularly in cases of haemolytic transfusionreactions.In this case, significant haemolysis caused by anti-Jk awas detected only in solid-phase and by ErythrocytesMagnetized ® technology and not by column or liquid-tubetechnologies.References1) Elliot K, Sanders J, Brecher ME. Visualizing the haemolytictransfusion reaction. Transfusion 2003; 43: 297.2) American Red Cross Immunohematology. Methods andProcedures. 1993 American Red Cross, Rockville.3) Boulet A, Petit S, Desmet P, et al. Antibody screeningperformance study using erythrocytes magnetizedtechnology (E.M. Technology) on a fully automated system.Transfusion 2005; 45: (Suppl) SP326, 124A.4) Working Party of the British Committee for Standards inHaematology Blood Transfusion Task Force. Guidelines forcompatibility procedures in blood transfusion laboratories.Transfus Med 2004; 14: 59-73.5) Yates J, Howell P, Overfield J, et al. IgG anti-Jk a /Jk bantibodies are unlikely to fix complement. Transfus Med1998; 8: 133-40.6) Phillips P, Voak D, Knowles S, et al. An explanation and theclinical significance of the failure of microcolumn test to detectweak ABO and other antibodies. Transfus Med 1997; 7: 47-53.7) Schonewille H, van de Watering LMG, Loomans DSE, et al.Red blood cell alloantibodies after transfusion: factorsinfluencing incidence and specificity. Transfusion 2006; 46:250-6.8) Callahan DL, Kennedy MS, Ramalli MA, et al. (2000)Delayed haemolytic transfusion reaction caused by Jkbantibody detected by only solid phase technique. Transfusion2005; 40: (Suppl),SP290, 113S.9) Rumsey DH, Nace SJ, Rubino M, et al. Naturally-occurringanti-Jk a in infant twins. Immunohematology 1999; 15: 159-62.10) Maynard BA, Smith DS, Farrar RP, et al. Anti-Jk a , -C, and -E in a single patient, initially demonstrable only by the manualhexadimethrine bromide (Polybvrene) test, withincompatibilities confirmed by 51Cr-labeled red cells studies.Transfusion 1968; 28: 302-6.Received: 19 March 2007 – Accepted: 24 April 2007Correspondence: Dr. Maria Antonietta VillaU.O. Centro Trasfusionale e di Immunoematologia, Dipartimento di MedicinaRigenerativa - Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena,Fondazione IRCCS di Natura PubblicaVia Francesco Sforza, 28 20122 Milan - Italye-mail: a.villa@policlinico.mi.it106Blood Transfus 2007; 5: 102-106 DOI 10.2450/2007.0007-07102-106_villa.p65 10605/07/2007, 11.27

LETTER TO THE EDITORProject to promote blood donation in a multicultural societyVincenzo SaturniPresident of AVIS, Region of Lombardy, ItalyDear Editor,the main priority of voluntary blood donor associationsis to spread the practice of donating blood and itscomponents, thus guaranteeing self-sufficiency in theseproducts; this can be achieved principally throughincreasing the number of organised and periodic donors.According to some statistical estimates, the populationof over 60-year olds will increase from 998,142 in 2006 to1,611,447 in 2046 (+613,305), while the population between18 and 60 years old will decrease from 5,418,693 to 4,045,176(- 1,373,517). On the basis of these demographic projections,we considered a more significant involvement of peoplefrom other parts of the world, with different cultures,religions and ethnicities.In our opinion, the pathway of intercultural harmony,integration and interactions should undoubtedly start fromaspects of health care, which must be based on scientificanalysis and careful evaluation of the existing literature,but must have a strong social value.Although the opinions above are not exhaustive andworthy of further analysis, they do stimulate reflection andallow us to make some considerations on a subject, that isperhaps a novelty only for us, and examine it withoutpreconceptions or prejudices. Where the development ofa culture of donation passes through the awareness andthe responsible involvement of the individual, health careissues constitute an important but not unique criterion,that must guide us.For example, if we take risk behaviours, as weunderstand them, these can have completely different facetsdepending on the person's culture and/or religion, whichare not necessarily linked to those of the country of origin.This problem is also clearly present in already existingdonors.Finally, participation in the development of voluntaryactivities and unpaid, periodic donations is a significantBlood Transfus 2007; 5:107-108© SIMTI Servizi Srland concrete form of interaction, efficiently promoted byvoluntary blood donor associations.On this background and based on a predominantlyanthropological approach, we set up the "Project to promoteblood donation in a multicultural society". The stages ofthis projects were as follows:- the establishment of a scientific committee (seeappendix), co-ordinated by the author of this letter andsupervised by the anthropologist Annamaria Fantauzzi,with the active participation of associations of the Latin-American, Romanian, Moroccan and Sikh communities,as well as the Union of Islamic Communities in Italy(UCOI);- the first meeting, held in January of this year, duringwhich the following aims and strategies were decided;- organisation of a meeting: "Blood donation in amultiethnic society", held on March 17 th in Milan. Thismeeting was aimed at representatives of groups andassociations of foreigners and voluntary blood donorassociations, community leaders and/or referencefigures for foreign citizens. Some of these communityfigures were speakers at the meeting. There were 119participants [52 from the Italian Association of VoluntaryBlood Donors (AVIS), 10 from the Policlinico Hospitalin Milan, and 57 from the different communitiesinvolved, including the Philippine community].The meeting was followed by a multiethnic buffet.The following aims have been defined for this project.1. Promote the culture of solidarity and donation.2. Promote the practice of donating blood within variouscommunities of foreigners, while respecting theirdifferent cultures and religions.3. Spread clear information and indications on how to getto know AVIS and other donor associations and theiractions in local, national and international settings.4. Set up and promote activities aimed at creating a107107-108_saturni.p65 10705/07/2007, 11.27

Saturni Vnetwork between the people involved in order to:a) exchange information, experiences, strategies andto enhance awareness;b) facilitate involvement and meetings;c) promote joining AVIS and other associations;d) participate in meetings.5. Create the instruments for carrying out campaigns toincrease awareness within reference points for foreignfamilies and in meeting places.6. Facilitate the availability of information about rare bloodgroups.7. Create networks for the development of cultural andscientific exchanges with the countries of origin of theforeign communities (for example, reciprocal visits,exchange of informative material, relations between rareblood group banks, invitations to particularly significantmoments of our association, such as assemblies,meetings, exhibitions and celebrations).At the end of the meeting, we set up a mixed workinggroup, that will work to produce material, in differentlanguages, to promote blood donation and to organisevarious manifestations, one of which could take placeconcomitantly with the World Blood Donor Day on June14 th , 2007.AppendixProject leader- Lombardy and Piedmont Regional Sections of AVIS,in collaboration with the Lombardy Regional Bank ofRare Blood Groups and the Association of Donors ofthe Policlinico Hospital, Milan.Scientific Committee- Sebastian Asaftei, Orthodox Church (Turin)- Mohamed Boussetta, AMECE Association – Morocco- Giorgio Dulio, President of AVIS, Piedmont Region- Annamaria Fantauzzi, (anthropologist, La SapienzaUniversity of Rome - EHESS Paris, Head of theObservatory on the culture of blood donation byimmigrants of the National AVIS),- Stelian Ionescu, Association of Romanians in Italy(Milan)- Yassin Lafram, Young Muslims in Italy – UCOII- Vincenzo Saturni, President of AVIS, Lombardy Region(co-ordinator)- Sukhdev Sing, Sikh community (Cremona)- Maria Pilar Yenque, Latin-America Association- Giorgio Marmiroli, Blood Donors' Association of thePoliclinico Hospital (Milan)AVIS representatives- Igino Arboatti, Health Care Director – AVIS, PiedmontRegion- Gianpietro Briola, Health Care Director - AVIS,Lombardy Region- Aldo Cardani, Councillor - AVIS, Lombardy Region –School representative- Claudio Crocco, Councillor - AVIS, Lombardy Region– Youth policies representative- Sandro Fisso, Administrator - AVIS, Piedmont Region- Felice Majori, Vice-President - AVIS, Lombardy Region- Luisella Peano, Councillor - AVIS, Piedmont Region –Head of Training.- Tiziana Tacchini, Vice-President - AVIS, PiedmontRegionRepresentative of the Regional Rare Blood GroupsBank:- Fernanda Morelati, Chief of the Regional Rare BloodGroups Bank- Maurizio Marconi, Director of SIMT, PoliclinicoHospital, MilanRepresentative of the Blood Donors' Association ofthe Policlinico Hospital, Milan- Giorgio Marmiroli, Vice-President- Giovanna Cremonesi, Health Care DirectorReceived: 23 April 2007 – Accepted: 21 May 2007Correspondence: Dr. Vincenzo SaturniAVIS Regione LombardiaLargo Donatori di Sangue, 120122 Milan, Italye-mail: avis.lombardia@avis.itOrganising Secretariat- Laura Cubeddu - Danila Fiume, AVIS, LombardyRegion108Blood Transfus 2007; 5: 107-108107-108_saturni.p65 10805/07/2007, 11.27

PRESS REVIEWPress Reviewby Giorgio RealiLerut E, Van-Damme B, Nouizat-Pirenne F,et al.Duffy and Kidd blood group antigens: minorhistocompatibility antigens involved in renal allograftrejection?Transfusion 2007; 47: 28-40There is increasing evidence that minorhistocompatibility antigens play a role in the graft rejectionprocess. Their characterisation is, therefore, of importance,but very complex and they are still poorly defined. Lerutand colleagues from the University of Leuven (Belgium),from the University of Paris (France), and from the NationalInstitute of Blood Transfusion in Paris investigated whetherthe Duffy (FY) and Kidd (JK) blood groups antigens couldplay a role as minor histocompatibility antigens in renalallograft rejection. These polymorphic and immunogenicantigens are widely distributed in human organs, areexpressed and functional in the kidney, and combineparticular proprieties, strongly suggesting that they couldindeed act as minor histocompatibility antigens in renalallograft rejection. Three hundred and seventy cadaverrenal transplants (342 first, 28 non-first transplants),performed in the Renal Transplant Programme at theCatholic University of Leuven between January 1, 1991and March 31, 1995, with an immunosuppressive regimencontaining cyclosporine A and corticosteroids, werestudied retrospectively; the recipients were followed-upfor more than 7 years after transplantation. All donors andrecipients were Caucasian. All relevant data (demographic,clinical, biological and immunohaematological) wereprospectively collected and entered in the University ofLeuven renal transplant database. When included in theprogramme, each recipient received transfusion with twounits of concentrated, non-leucodepleted red cells. Thefollowing characteristics were studied: age and sex of donorBlood Transfus 2007; 5: xx-xx© SIMTI Servizi Srland recipient; donor/recipient sex mismatch; number ofHLA–A, -B and DR mismatches; cold and warm ischaemiatime; mean (mg/kg/day) dose and total dose (mean dose xnumber of days administered/10 3 ) of cyclosporine A andits level at 1 and 5 years after transplantation; maximalpercent of panel reactive antibodies; presence of anti-FYand anti-JK antibodies at day 0 and after transplantation.FY and JK polymorphisms were identified in all donor/recipient pairs by real time polymerase chain reaction(genotyping) and in all pairs the matching or mismatchingstatus was defined for both systems. All biopsies werereviewed and historical screening results for FY and JKalloantibodies and graft survival were retrospectivelyanalysed. The results strongly suggest an immunologicalrole for the FY polymorphism in renal transplantation. Itwas, in fact, observed that FY mismatched grafts hadsignificantly more chronic lesions than FY matched grafts,therefore supporting a minor histocompatibility status forFY. HLA-DR11 was more frequent in both recipients anddonors of FY mismatched pairs without chronic allograftnephropathy, suggesting a protective effect of thismolecule.The results regarding JK polymorphisms are less clear,although JK mismatched grafts had higher scores forinterstitial inflammation than JK matched grafts. FY and JKmatching had no influence on graft survival. As well known,most of blood group antigens are not solely confined toerythrocytes but are widely expressed through the humanorgans, and, therefore, the term "histo-blood groupantigens" has been correctly introduced. In summary, theobservations and the results of this important retrospectivestudy strongly support the idea that FY could play a roleas a minor histocompatibility antigen. Although histologicaldifferences were detected between matched andmismatched JK, the data were less strong than those forthe FY system.109109-116_reali_press r.p65 10905/07/2007, 11.28

Reali GCastro V, Kroll H, Origa AF, et al.A prospective study on the prevalence and risk factors forneonatal thrombocytopenia and plateletalloimmunization among 9332 unselected Braziliannewborns.Transfusion 2007; 47: 59-66.The article presents a screening study on theprevalence of neonatal alloimmune thrombocytopenia(NAIT) in Brazilian newborns, a population characterisedby an admixture of indigenous, African, and Caucasiansubjects. The study was performed at the State Universityof Campinas (Brazil), with the collaboration of colleaguesfrom Germany and the United States. NAIT develops as aresult of an interaction between human platelet antigen(HPA) HLA allotypes and the ethnic background of thepopulation. NAIT is present in 0.5% to 0.9% of unselectedCaucasian newborns. Its prevalence in other populationsis unknown. A series of studies among Caucasians revealedthat 80% of NAIT results from disparity in the biallelicHPA-1 system: approximately 2% of Caucasians lacks thecommon HPA-1a allele and are at risk of NAIT. However,the disease occurs in only 10% of these people and isstrongly associated with the maternal HLA-DRw52a type.The HPA-5 system is also commonly involved in NAITamong Caucasians, but presents with milder clinicalpictures, compared with the HPA-1-mediated disease.The study was carried out, prospectively, over a 3-yearperiod, to determine the prevalence and the causes ofthrombocytopenia in cord blood samples. The plateletcount was performed on cord blood, using a standardautomatic blood cell counter. Genotyping for HPA-1, -2, -3,and –5 was performed (in pairs of thrombocytopenicneonates and their mothers) by polymerase chain reaction(PCR)-restriction fragment length polymorphism, and forHPA-4 by PCR with allele-specific primers. Sera of allmothers with genotypic mismatch from each group weretested for the presence of anti-HPA antibodies, using themonoclonal antibody immobilisation of platelet antigen(MAIPA) technique. Thrombocytopenia (defined as aplatelet count below 100 x 10 9 /L) was detected in 142 out of9,332 newborns studied (1.5%). The majority ofthrombocytopenic neonates (128 cases, 90.1%) had plateletcounts between 50-100x10 9 /L; moderate thrombocytopenia(

Press Reviewresponse, compared to the polyclonal anti-D, was shownbetween mixtures containing 75% rIgG1 and 25% rIgG3and those containing 50% rIgG1 and rIgG3. Mixtures ofrIgA1 and either rIgG1 or rIgG3 showed no synergic effect.The inability of rIgG2 anti-D to mediate a response inboth assays was ascribed to the topography, surfacedensity, and restricted mobility of the D antigen. Anti-D isbound near the lipid bilayer, surrounded by glycocalyx,and this may sterically hinder the access of the relativelyinflexible IgG2 anti-D, but not that of IgG1 or IgG3 anti-D.In conclusion, in vitro functional studies onrecombinant anti-D antibodies (rIgG and rIgA), performedat Copenhagen University Hospital, showed that in theADCC and CL assays no single antibody was comparableto a currently used prophylactic polyclonal anti-D in theability to induce a respiratory burst and to promote lysis ofred blood cells. A combination of rIgG1 and rIgG3 did,however, demonstrate similar activity in CL and ADCCassays as polyclonal anti-D, and could be of value in futureimmunoprophylaxis for prevention of HDN due to D.Grassineau D, Papa K, Ducourneau A, et al.Improving minority blood donation; anthropologicapproach in a migrant communityTransfusion 2007; 47: 402-9.In countries with large migrant populations, the supplyof compatible blood for transfusion in recipients fromdistant geographic regions may be a problem. Indeed, redblood cell (RBC) polymorphisms can lead to significantphenotypic discrepancies between patients from migrantcommunities and the local donor pool, and somepopulations can be considered as high-risk groups foralloimmunisation. This is especially true for sub-SaharanAfricans, living in countries with predominant Caucasian(European) populations. Haematological hereditarydiseases, such as sickle cell anaemia, and traditional culturalvalues that discourage blood donation further complicatethe problem. Marseilles, in France, where this study wasperformed, has shown a strong capacity for integratingimmigrant populations. While migrants have traditionallybeen from bordering countries (Italy, Spain, North Africa),recent immigration has been more diverse, with a growingnumber of people from sub-Saharan Africa, Asia, and EastEurope. Already numbering 70,000 persons, one of thefastest growing and largest minorities in the south ofFrance, are immigrants from the Comoros Islands (locatedbetween northern Madagascar and northern Mozambique).It has been found that Comorian immigrants have a typicalsub-Saharan African phenotypes, characterised by theabsence of a high frequency antigen in the Duffy system(F3 and F5), by the Fy(a-b-) phenotype (also in the Duffysystem), U var or S-s- in the MNS system, and by thepresence of (C)ce s allele in the Rhesus system.The purpose of the report was to present the methodused in a drive to promote blood collection in the communityformed by Comorian immigrants living in Marseilles. Witha culturally adapted message, developed on ananthropologic approach, and working in close collaborationwith scientific and medical members of the Comoriancommunity, this population has been sensitised to blooddonation. An association of voluntary blood donors ofComorian origin was also founded. Involvement of religiousand political leaders was essential to support theintervention and to dispel any ambiguity about objectionsto blood donation.As result, a targeted collection of specific sub-Saharanblood phenotypes was achieved, recruiting mostly secondgenerationdonors. At first, a total of 92 people of Comorianorigin volunteered to donate blood. Based on the medicalpre-donation interview, only 59 individuals were selectedfor collection: 33 candidate donors were excluded, mostbecause of a history of malaria or exposure to infectiousdisease. The first collection was effective in providingspecific red blood cell phenotypes, present in 7 donors(nearly 12%), a very high percentage compared with only0.56% in the previously existing donor pool. Subsequently,it was decided to promote regular blood donation byaddressing people directly towards the principal BloodTransfusion Centre in MarseillesIn conclusion, the approach led to an increase of new,regular and voluntary Comorian blood donors. This methodcould also be applied to improving donation of peripheralblood stem cells or organs from the same population.Wagner SJ, Eder AFA model to predict the improvement of automated bloodculture bacterial detection by doubling platelet samplevolumeTransfusion 2007; 47: 430-33.In the USA, blood banks initiated automated culturesfor detecting bacteria in apheresis platelet concentrates(PC), in response to the voluntary standards of the Collegeof American Pathologists and American Association ofBlood Banks (AABB). Some bacterial infections have beentransmitted through apheresis PC. Results from testing1,277,508 platelet apheresis collections from March 2004 toFebruary 2006 were reported by the American Red Cross(ARC). The results were: a true culture-positive rate ofBlood Transfus 2007; 5: xx-xx111109-116_reali_press r.p65 11105/07/2007, 11.28

Reali Gapproximately 1 in 5,400; a confirmed false-positive rate ofapproximately 1 in 2,900 units, and an unconfirmed andindeterminate rate of approximately 1 in 12,900. During thisperiod of time there were, however, 17 septic transfusionreactions from PC with false-negative results, includingtwo fatalities. Thirteen of these 17 cases involvedStaphylococcus species, which are known to be slowgrowingorganisms, that exhibit lag times when deliberatelyinoculated into platelet components.In this report, Wagner and Eder from the ARC BloodDepartment (Rockville, MD) and the ARC Blood Services(Washington, DC) explored the impact of doubling thesample volume (from 4 to 8 mL) for culture detection ofslow growing bacterial organisms. Bacterial growth wasmodelled by varying the initial inoculum, doubling time,and lag time. The numbers of organisms present at a 24-hour sampling time were calculated. Limitation in samplingunits with low bacterial load and the possibility of selfsterilisationof units during processing time complicatedthe direct assessment of bacterial growth in the spikingstudies. The Authors, consequently, used a mathematicmodel coupled with Poisson analysis to predict theimprovement of bacterial detection by doubling the samplevolume. From each inoculum, the percentage of improveddetection was defined as delta, i.e. the difference betweenthe percentage of detection in 8- and 4-mL samples.The results suggest that: 1) the maximum improvementof detection was independent of bacterial doubling time orlag time; 2) the maximum improvement occurred with inocula

Press ReviewSanders RP, Geiger TL, Heddle N, et al.A revised classification scheme for acute transfusionreactionsTransfusion 2007; 47: 621-8.The classification of transfusion reactions issummarised in the AABB Technical Manual, which servesas a standard reference for Transfusion Medicine Services.In the AABB classification the acute reactions, occurringwithin 24 hours after transfusion, are divided by presumedaetiology into immunological and non-immunologicalsubtypes. The immunological types include the followingreactions: haemolytic, fever and/or chill non-haemolytic,urticarial, anaphylactic and transfusion-related acute lunginjury (TRALI); the non-immunological reactions include:hypotension associated with ACE inhibition, circulatoryoverload, non-immune haemolysis, air embolism,hypocalcaemia, hypothermia and transfusion-associatedsepsis. Although this system adequately describes thegeneral features associated with the various types of acutetransfusion reactions, it was not designed to provide strictcriteria for diagnosis and classification. Consequently, itsuse to classify individual reactions can result in significantinter- and intra-observer variability, leading to unreliablecomparisons of research results between different Centresand between studies from the same Centre.Sanders and colleagues from the Department ofOncology at St Jude Children's Hospital in Memphis (TN,USA) reviewed blood bank and medical records to identifyall suspected transfusion reactions that occurred in theirInstitute from January 1, 1996 until December 31, 2003.Initially classified according to the AABB system, eachreaction was then reclassified with a revised model, whichrefines and clarifies various categories, and adds severitygrades in the format of the Common Terminology Criteriafor Adverse Events (CTCAE), a system created by USANational Cancer Institute, to standardise the reporting ofadverse events in cancer therapy. The revised systemincludes the following categories: febrile asymptomatic,allergic, inflammatory, mixed allergic and inflammatory,TRALI, hypotension, haemolysis (immune or non-immune),sepsis, circulatory overload. It also incorporates a severitygrade (mild, moderate, severe, life-threatening, fatal). Furthergroups of the revised system consider the likelihood ofrelation to transfusion: unrelated, unlikely, possible,probable, definite. All blood components administered tothe patients were irradiated and leucoreduced; plateletconcentrates were all single donors apheresis units.During the 8-year study period, 627 suspectedtransfusion reactions were reported and 595 of these hadcomplete records suitable for analysis. Three hundred andseventy-two occurred with transfusion of plateletconcentrates, 213 in red blood cell transfusions, 5 in freshfrozenplasma transfusions and 3 in transfusions ofgranulocyte concentrates. In additional two cases, patientreceiving both platelet concentrates and erythrocyteconcentrates within a short interval and the bloodcomponent that elicited the reaction could not be clearlydistinguished.A comparison of the classification approaches appliedto these 595 transfusion reactions showed clear advantagesof the revised system. Of 128 reactions classified by AABBcriteria as inconclusive (mixture of reaction types orotherwise qualified), all but 5 could be classified intodiscrete categories in the new scheme. The revised systemintroduces more specific criteria for certain reactions andpermits the grading of severity and attribution (likelihoodthat the adverse event is related to transfusion), leading tomore thorough and more flexible categorisation oftransfusion reactions.Tobler LH, Stramer SL, Chien DY, et al.Antibodies to a novel antigen in acute hepatitis C virusinfectionVox Sang 2007; 92: 1-7.Hepatitis C virus (HCV) is the aetiological agent of aliver infection that affects, worldwide, about 170 millionpeople. Commercially licensed HCV enzyme immunoassays(EIA), which are used to screen for HCV exposure, userecombinant proteins, presenting linear epitopes. It is,however, recognized that conformational antigens mayplay an important role in the immunobiology of acute HCVinfection and may enable earlier antibody detection. In thelate 1990s, a new donor-screening method, nucleic acidamplification technology (NAT), was introduced: thistechnology enables the detection of HCV RNA positive,but antibody non-reactive, infected units of blood duringroutine donor screening. Very rarely, some HCV-infectedpeople apparently fail to produce antibodies, but mostsamples of this type (RNA positive/antibody negative)simple represent very early HCV infection.Tobler and co-workers, from the Blood System ResearchInstitute in San Francisco (CA, USA), evaluated earlyantibody production, using three new EIA, containingantigenic proteins not present in licensed assays. The firstnew EIA contains two proteins: a conformational antigenthat retains both protease and helicase enzymatic activities(NS3/4a), and a single multiple-epitope fusion antigen(MEFA 7.1). MEFA 7.1 represents an enhanced antigen,incorporating known major epitopes of HCV protein. TheBlood Transfus 2007; 5: xx-xx113109-116_reali_press r.p65 11305/07/2007, 11.28

Reali Gsecond EIA used the F (frameshift) and core proteins ofhepatitis virus as the anti-HCV detection antigens. Thethird research EIA used E1/E2 protein generated in Chinesehamster ovary (CHO) cells as the anti-HCV detectionantigen. Forty-two RNA positive/EIA 3.0 (latestimmunoassay licensed) negative samples, including twopersistently serosilent cases, were used to evaluate theseresearch EIA. Fifty-four EIA 3.0 negative/RNA negativeand three HCV RNA positive/antibody positive specimenswere included, as controls. Only the MEFA 7.1 NS3/4a waspositive in seven (17%) of the 42 HCV RNA positive samples,in all three positive controls but in none of 54 EIA 3.0negative/RNA negative controls. Thirty-five specimens,including the two serosilent case were negative. Notably,six out of the seven (86%) MEFA 7.1 NS3/4a positivesamples had evidence of active hepatitis (ALT >210/IU/L).Conversely, all 42 NAT yield specimens, including theserosilent cases, were non-reactive in the research EIAassays, using both the F/core protein combination and theCHO E1/E2 antigens. All 54 negative specimens used ascontrols were non-reactive with the three research EIAs,while the RNA+/antibody+ controls were positive.In conclusion, a novel research EIA, usingconformational and enhanced linear epitopes (NS3/4a andMEFA 7.1 antigens, respectively) detected HCV antibodiesin 17% of viraemic donations missed by the HCV 3.0 EIA.This research EIA appears to detect HCV antibodies closerto the initiation of acute hepatitis. Given that the averageRNA-positive/antibody negative window period is ~56.4days, this 17% yield would translate into a ~10 day earlierdetection of antibodies.Ansart-Pirenne H, Martin-Blanc S, Le PennecPY, et al.FY*X real-time polymerase chain reaction with meltingcurve analysis associated with a complete one-step realtimeFY genotypingVox Sang 2007; 92: 142-7.The Duffy (FY) blood group antigens have been known,for many years, to be a receptor for malaria parasitesPlasmodium vivax and Plasmodium knowlesi. Morerecently, the Duffy protein was found to be a receptor forvarious classes of chemokines and was renamed DARC(Duffy antigen receptor for chemokines), a seventransmembrane glycoprotein present on erythroid and nonerythroidtissues. FY is a clinically significant blood groupsystem, involved in haemolytic transfusion reactions andin haemolytic disease of the newborn. The system iscontrolled by four major alleles: FY*A and FY*B, theCaucasian common alleles, encoding for Fy a and Fy bantigens; FY*X allele responsible for reduced expressionof Fy b antigen; and FY*Fy, a silent gene. Homozygosityfor FY*Fy gives rise to the Fy(a-b-) phenotype, common inBlack people but extremely rare in Caucasians. In Fy(a-b-)Black individuals the FY*Fy gene is a T46C mutation in thepromoter region, which abolishes erythroid expression ofthe FY gene, that is, in contrast, normally expressed innon-erythroid tissues. The FY*X allele, found morefrequently among Caucasians, is caused by a C286Tmutation in the FY*B gene, resulting in an Arg to Cyssubstitution at position 89 of the Duffy protein.Despite the recent development of a real-timefluorescent polymerase chain reaction (PCR) method forFY genotyping, FY*X genotyping has not been describedby this method. The study by Ansart-Pirenne et al. focusedon the real-time FY*X genotype development, associatedwith a complete, one-step real-time FY genotyping, basedon fluorescence resonance energy transfer (FRET)technology. The study was carried out at the InstitutNational de la Transfusion Sanguine in Paris (France).Two series of blood samples were studied: i) 72 bloodspecimens from Fy(a+b-) Caucasian blood donors, studiedby real-time PCR only; ii) 47 blood samples from Caucasiansand Black patients referred to the Laboratory of theInstitute, and studied both by real-time PCR and by PCRrestrictionfragment length polymorphism (RFLP). TheDuffy phenotypes were determined by an agglutinationmethod using an indirect antiglobulin gel test with anti-Fy aand anti-Fy b human polyclonal sera. When necessary, weakexpression of Fy b was confirmed by adsorption/elutionstudies.Two mutations (C286T and A319G) were identified inthe FY*X gene: only the C286T single nucleotidepolymorphism (SNP) is associated with the weak Fy bphenotype. The frequency of FY*X allele in the Fy(a+b-)donors was 0.014, identical to the frequency previouslydescribed (0.015). In 68 out of 72 donors the genotype wasconfirmed to be FY*A/*A, but in two cases the genotypewas FY*A/*X and in two FY*A/Fy. With the Caucasianand Black patients, a complete correlation between PCR-RFLP and real-time PCR was found, whatever the allelecombination tested. When the known phenotype was notcorrelated to FY*A genotype, the presence of the Fy b antigenwas always confirmed by absorption/elution studies.In conclusion, the real-time technology method is rapidand accurate for FY genotyping and able to detect theFY*X gene in all the allele combinations studied. Thedetection of the FY*X allele is useful for determining the114Blood Transfus 2007; 5: xx-xx109-116_reali_press r.p65 11405/07/2007, 11.28

Press ReviewFy b antigen status both in blood donors and in recipients,considering the transfusion use of blood units and thepreparation of panels for antibody screening.Vamvakas ECWhite-blood-cell-containing allogeneic blood transfusionand postoperative infection or mortality: an updated metaanalysisVox Sang 2007; 92: 224-32Vamvakas, one of the researchers most interested intransfusion-related immunomodulatory (TRIM) effects ofallogeneic blood transfusion (ABT), tried to undertake afinal meta-analysis that should generate conclusions, basedon all randomised controlled trials (RCT) published on thematter. In fact, no further RCT are presently either plannedor underway. The purported TRIM effects of ABT remaindebated and elusive, but most evidence from animal modelssuggests that TRIM is mediated by white blood cells(WBC). The donor's WBC may either directly down-regulaterecipient immune functions or indirectly mediate the TRIMeffects by releasing soluble mediators into the supernatantfluid of red blood cells (RBC), during storage. As mentionedabove, new studies are not underway and, given that mosttransfusions currently administrated in North America andin Western Europe are WBC-reduced, it will be difficult (ifnot impossible) to undertake further RCT comparingrecipients of non WBC-reduced vs. WBC-reduced RBC.RCT reporting on the association of ABT with postoperativeinfections and/or short term (up to 3 months aftertransfusion) all-cause mortality were retrieved using acomputerised (MedLine) search of the English literature, amanual search of the bibliographies of all pertinent articles,and a manual search of abstracts appearing in thesupplements of scientific journals (of transfusion medicine,haematology, anaesthesiology/critical care). Twelve RCTreporting on infections and 11 RCT reporting on mortalitywere eligible for meta-analysis. Information was extractedfrom the reports of the retrieved studies, using astandardised data abstraction form. An odds ratio (OR) ofpostoperative infections or mortality in recipients of non-WBC-reduced RBC or whole blood vs. WBC-reduced RBCor whole blood and a 95% confidence interval (CI) of theOR were calculated for each RCT.An association with post-operative infection wasdemonstrated across RCT transfusing RBC, WBC-reducedafter storage (summary OR = 2.25; 95% CI = 1.12-4.25) butnot before storage (summary OR = 1.06; 95% CI = 0.91-1.24). An association of ABT with mortality wasdemonstrated across RCT conducted in cardiac surgery(summary OR = 1.72; 95% CI = 1.05-2.81) and across RCTtransfusing buffy-coat-reduced RBC vs. RBC, WBCreducedbefore storage (summary OR = 1.6; 95% CI = 1.14-2.24) but not across RCT transfusing non-buffy-coatreducedRBC vs. RBC, WBC-reduced before storage(summary OR = 1.01; 95% CI = 0.73-1.40).The meta-analysis presented in this article did notdemonstrate an overall deleterious TRIM effect of ABT,although it did detect a TRIM effect across three specificsubsets of RCT: studies transfusing to the control armWBC-reduced RBC or whole blood filtered after storageand reporting on post-operative infections; studiesconducted in cardiac surgery and reporting on short-term,all-cause mortality; and studies transfusing buffy-coatreducedRBC to the treatment arm and WBC-reduced RBC,filtered before storage, to the control arm and reporting onshort-term, all-cause mortality. Other associations foundin subgroup analyses appear to contradict current theoriesabout the pathogenesis of TRIM.van Burik J-A, Brunstein CGInfectious complications following unrelated cord bloodtransplantation [review]Vox Sang 2007; 92: 289-96.Umbilical cord blood (UCB) is an increasingly usedsource of haematopoietic stem cells (HSC) fortransplantation of patients with advanced and high-riskhaematological diseases. UCB has been shown to be aneffective alternative for patients who lack a suitable adultdonor, but has largely been limited by the ability of findingan UCB unit with an adequate cell dose, which has beenshown to be an important determinant of outcomes. Newerstrategies, however, such as double UCB grafts and nonmyeloablativeconditioning regimens, are expanding UCBtransplantation (UCBT) to larger and older patientpopulations. The review by van Burik and Brunsteinoutlines infections present at death, the epidemiology ofindividual infections, and the impact of UCB as HSC source.The Authors, from the University of Minneapolis (MN,USA), reviewed the most important studies published onthe matter since 2000. Most studies found an approximaterate of infection at death of 30-40% among UCBT recipients.Bacterial infections are the most frequent source of seriousinfections after UCBT and the risk is higher for patientswith graft failure. This kind of infection often occurs priorto the engraftment. In addition, there is a delayed recoveryof the immune response among patients with graft-versushostdisease (GvHD). Commonly identified bacteria includeGram-negative bacilli, such Pseudomonas, Acinitobacter,Blood Transfus 2007; 5: xx-xx115109-116_reali_press r.p65 11505/07/2007, 11.28

Reali Gand Escherichia coli, and Gram-positive organisms suchas coagulase-negative Staphylococci, Enterococcus, andStaphylococcus aureus. As far as fungal infections areconcerned, Aspergillus is the most frequent agent,accounting for about 50-70% of infections, and Candidaspecies are the cause of most of the remaining fungalinfections. Viral infections or reactivations (herpes viruses,cytomegalovirus, Epstein-Barr virus, respiratory viruses)tend to be observed starting after engraftment until immunereconstitution and their incidence is likely to be influencedby whether GvHD does or does not develop.Toxoplasmosis, as parasitic infection, has been reportedfollowing UCBT and appears to be associated with pretransplantationseropositivity and high mortality.The risk of serious infections among children receivingUCBT is, however, comparable to that of children receivingmarrow HSC and is lower than that of recipients of T-celldepletedstem cells. Among adults, despite an overall higherincidence of severe infections after UCBT that afterunrelated donor grafts, non-relapse mortality and overallsurvival were not significantly different between recipientsof HSC from the different sources.The Authors conclude that further studies are neededto confirm these observations and correlate them withanalyses of markers of immune reconstitution.Maury S, Balère-Appert ML, Chir Z, et al.Unrelated stem cell transplantation for severe acquiredaplastic anemia: improved outcome in the era of highresolutionHLA matching between donor and recipientHaematologica 2007; 92: 589-96.Severe acquired aplastic anaemia (SAA) is a potentialfatal bone marrow failure syndrome, occurring mainly inchildren and young adults. Haematopoietic stem celltransplantation (HSCT) from an HLA-matched family donoris treatment of choice for patient affected with SAA, butthis approach is limited by the availability of HLA-matcheddonors. Immunosuppressive therapy (a combination ofanti-thymocyte globulin and cyclosporine A) is thealternative treatment for patients without a suitable relateddonor. HSCT from a phenotypically HLA-matchedunrelated donor is indicated as salvage treatment forpatients who fall to respond to one or more courses ofimmunosuppressive therapy and for patients whoexperience relapse of the disease. Historically, HSCT usingunrelated donors for patients affected with SAA wasassociated with high mortality rates. The Authors of thisreport, on behalf of the French Society of Bone MarrowTransplantation and Cellular Therapy (SFGM-TC), haveanalysed the outcome of 89 patients (median age 17 years,range 0-52) with acquired SAA, undergoing HSCT from anunrelated donor between 1989 and 2004, with the aim toverify whether, in recent years, the outcome of thesetransplantations improved, and, if so, to determine whetherthe improvement resulted from changes in selection ofpatients, changes in transplantation technique, or both.Reviewed cases were consecutively reported to the FrenchRegistry by 25 transplant centres.The HLA matching between recipient and donor wasbased on HLA search determinants according to thestandard techniques: serology and DNA typing. Two timeperiods were considered: 1989-1998 (37 transplants) and1999-2004 (52 transplants). In the first period, most donors(91%) were matched at a generic level (or antigen level) forHLA-A, -B, -DR. In the second period, most of the pairsrecipient-donor was matched at allelic level, with molecularanalysis of HLA-A, -B, -C, -DRB1 and –DQB1 loci. Fiveyearssurvival probabilities (95% confidence interval) weredifferent in the two groups: 29±7% and 50±7%, respectively(p