EPO相关PRCA

Review articleRecommendations on haematological criteria for the diagnosis of epoetin-induced pure red cell aplasiaPure red cell aplasia(PRCA)is conventionally described as a syndrome characterised by normo-chromic normocytic anaemia,with a reticulocyte count below10·109/L and an almost complete absence of erythroblasts from the bone marrow (BM),without evidence of haemolysis(1,2).PRCA is generally considered to be selective for the erythroid cell lineage,resulting in severe anaemia with normal leucocyte and platelet counts.This presentation distinguishes PRCA from aplastic anaemia,which involves all three cell lineages(1). PRCA may be congenital or acquired,and a very large number of causes of the latter have been described in the literature(1).Many of these causes, however,have only been noted in a few patients, and there are few causes for the most common forms of PRCA(Table1)(3).The administration of epoetin permits the ter-minal maturation of erythroid precursors,which undergo apoptosis in its absence.Since its introduc-tion in1989,several hundreds of thousands of anaemic patients have received recombinant human erythropoietin(epoetin),increasing their haemoglo-bin levels safely and generally with few side-effects. In thefirst decade of its therapeutic use,however, three reports appeared of patients with chronic kidney disease(CKD)developing PRCA associated with autoantibodies against epoetin(4–6).There are very few reports available describing the develop-ment of PRCA in patients who developed autoan-tibodies against epoetin without having received epoetin therapy(7).However,from1998onwards, the incidence of epoetin-induced PRCA(epo-PRCA)increased considerably(3),reaching a peak in2002(8).The majority of these cases occurred in patients treated with epoetin-alpha in a human serum albumin(HSA)-free formulation marketed in Europe(EPREXÒ/ERYPOÒ;Ortho Biologics,Casadevall N,Cournoyer D,Marsh J,Messner H,Pallister C,Parker-Williams J,Rossert J.Recommendations on haematological criteria for the diagnosis of epoetin-induced pure red cell aplasia.Eur J Haematol2004:73:389–396.ÓBlackwell Munksgaard2004. Abstract:Pure red cell aplasia(PRCA)is a rare condition character-ised by an arrest in red blood cell production,which may be congenital or acquired.Recombinant human erythropoietin(epoetin)was intro-duced in1989for the treatment of anaemia of chronic kidney disease patients and has maintained an excellent therapeutic and safety record while treating hundreds of thousands of patients.A very rare,but serious adverse event associated with epoetin administration is a con-dition in which patients develop neutralising anti-erythropoietin anti-bodies and,consequently,PRCA.This condition is referred to as epoetin-induced PRCA(epo-PRCA).Since it is a rare condition,many haematologists and nephrologists around the world see the condition infrequently and may be uncertain about the diagnosis.For this reason, an ad hoc international working group of expert haematologists and nephrologists met together to derive new recommendations for the haematological diagnosis of epo-PRCA.These recommendations,which represent the consensus opinions of the working group,address haem-atological approaches to monitor and investigate suspected epo-PRCA and should help physicians differentiate between PRCA and other bone marrow diseases,as well as,between PRCA and epo-PRCA.Nicole Casadevall1,Denis Cournoyer2,Judith Marsh3,Hans Messner4,Chris Pallister5,John Parker-Williams3,JØrôme Rossert6 1Service d'Hematologie Biologique,Hôpital Hôtel-Dieu, Paris,France;2Division of Hematology,McGill University Health Centre,Montreal,Quebec,Canada;3Department of Haematology,St George's Hospital Medical School, London,UK;4Department of Medical Oncology, University of Toronto,Ontario,Canada;5Scientific Services,Ortho Biotech,High Wycombe,UK;6Service de NØphrologie,Hôpital Tenon,Paris,FranceKey words:erythropoietin;pure red cell aplasia; antibodies;reticulocytes;bone marrow Correspondence:Professor Nicole Casadevall,Service d'Hematologie Biologique,Hôpital Hôtel-Dieu,1Place du Parvis Notre Dame,75181Paris Cedex4,France Tel:+3342348266Fax:+3342348254e-mail:nicole.casadevall@htd.ap-hop-paris.fr Accepted for publication31August2004Eur J Haematol2004:73:389–396 All rights reserved CopyrightÓBlackwell Munksgaard2004EUROPEANJOURNAL OF HAEMATOLOGY389LLC,Manati,Puerto Rico),and in all reported cases epoetin was injected subcutaneously.All exogenous proteins,including therapeutic ones,are potentially immunogenic(9).The degree of immunogenicity is influenced by many factors, including the genetic background of the patient,the type of disease,the degree of similarity to the endogenous protein,the presence of fragments, aggregates or contaminants,the formulation,the route of administration,and the dosing interval and duration of treatment(9).As epoetin-alpha is identical to endogenous human erythropoietin in its physicochemical properties(10),it should not be recognised as foreign.As epo-PRCA causes severe anaemia and patients become heavily transfusion-dependent,it is unlikely that the condition was overlooked earlier in the history of epoetin use.A more probable explanation for the increased inci-dence rate is that changes in the European formu-lation of epoetin-alpha(EPREXÒ/ERYPOÒ),such as replacement of human serum albumin with polysorbate80/glycine to comply with new Euro-pean regulations(3),increased usage of pre-filled syringes,its storage and handling,and/or its administration resulted in increased immunogenic potential.Almost all patients who developed epo-PRCA received epoetin subcutaneously(there are only one or two cases with intravenous EPOGEN) (11).This route(which is known to be more immunogenic than the intravenous route)is now contraindicated in CKD patients treated with epoetin-alpha in Europe,and intravenous use is recommended in Canada and Australia. Although the vast majority of cases occurred in CKD patients treated with HSA-free epoetin-alpha,epo-PRCA is not uniquely linked to treat-ment with epoetin-alpha;some patients who developed epo-PRCA received only epoetin-beta (3,12–15).Epo-PRCA is also not limited to CKD patients,as there have also been two reports of epo-PRCA in epoetin-treated patients with mye-lodysplasia(16),and in rare instances anti-erythropoietin-mediated PRCA has occurred in patients who have never received epoetin(7).Current diagnostic criteria for epo-PRCAPure red cell aplasia caused by a complete arrest of erythropoiesis results,on average,in a1%daily reduction in haemoglobin level(1).The clinical presentation is thus a progressive,severe,isolated anaemia of sudden onset.According to current diagnostic criteria,the hallmark of PRCA is the absence of erythroblasts from an otherwise normal BM;in classic cases,the erythroblasts are either completely absent or constitute<5%on the BM differential count(Table2)(1).The cellularity of the BM is typically normal with normal myeloid cells and megakaryocytes.The peripheral blood is characterised by a low reticulocyte count of usually <1%,with a normal white cell and platelet count, and a normal differential white cell count.These features serve to distinguish PRCA from aplastic anaemia,which typically presents with pancytope-nia and severely decreased BM cellularity,and from myelodysplastic syndromes(MDS),which seldom show a reticulocyte count below10·109/L.In addition,in MDS,the red cells tend to be macr-ocytic,with dysplastic changes affecting more than one haematopoietic lineage.The BM is usually hypercellular,but it is hypocellular in about10%of cases.Although erythroblasts may only be present in small numbers,they are seldom<5%and may show megaloblastoid features in addition to other dysplastic changes.These standard diagnostic criteria have recently been refined to focus on epo-PRCA occurring as a consequence of treatment with epoetin(Table3) (3).In this proposal,major features are distin-guished from minor features that may provide suggestive evidence for the diagnosis of epo-PRCA, which should be confirmed by BM aspirate exam-ination and serum assays for antibodies.These authors stressed that although complete confirma-tional investigations are the aim of each suspected case of epo-PRCA,the condition should not be excluded because a BM examination and/or an assay for neutralising anti-erythropoietin anti-bodies have not or cannot be performed.Both of these sets of diagnostic criteria define a rate of decrease in red cell count or haemoglobinTable1.Most common causes of pure red cell aplasia•Lymphoproliferative disorders(e.g.chronic lymphocytic leukaemia,lymphoma,and T-large granular lymphocyte disorders)•Infections(e.g.parvovirus B19,Epstein–Barr virus)•Systemic autoimmune disease(e.g.systemic lupus erythematosus,rheumatoidarthritis)•Drugs(e.g.azathioprine,isoniazid,phenytoin)•Thymoma(in about5%of cases)•Idiopathic(in about50%of cases)•Myelodysplastic syndromeReproduced with permission from Eckardt and Casadevall(3),with some modifi-cation.Table2.Current standard diagnostic criteria for pure red cell aplasia•Fall in red cell count of about1%/d•Reticulocyte count below1%•No major changes in white cell count,platelet count,or differential leucocytecount•Normal cellularity of bone marrow•Less than1%erythroblasts(occasionally up to5%proerythroblasts or basophilicerythroblasts)•Normal myeloid cells and megakaryocytesBased on Dessypris(1).Casadevall et al.390level.In practice,such a decrease may be imposs-ible to document by the time a patient comes to clinical attention,as the relevant sequential data may not be available.Although the haemoglobin level falls in MDS,it seldom falls rapidly,but again this feature may be obscured by the lack of data. The two sets of criteria differ in threshold reticu-locyte count(see Tables2and3);clinical experi-ence with epo-PRCA suggests that in the vast majority of cases the reticulocyte count is below 10·109/L and this is reflected in the modified criteria(3,14).Based on the conventional haematological pres-entation of PRCA,both sets of diagnostic criteria emphasise that the cardinal haematological features of PRCA are the sole involvement of the erythroid lineage,normocellular BM and low reticulocyte count.Deriving new recommendations for haematological diagnosis of epo-PRCADespite the concern generated by the increased incidence of epo-PRCA in patients treated with epoetin,PRCA remains a very rare condition and systematic studies are lacking for this reason.It was felt that a new set of recommendations regarding diagnosis of PRCA should be drawn up to help nephrologists and haematologists worldwide,who may come across the condition only very occasion-ally,to diagnose epo-PRCA.For this purpose,a general questionnaire relating to diagnosis and monitoring of PRCA was circulated to all members of an ad hoc international working group of expert haematologists and nephrologists,and the re-sponses were collated for discussion at a later meeting.The questionnaire was accompanied by a set of photographs from BM slides relating to13 patients with a variety of conditions,which were evaluated blind by the individual members of the group in conjunction with a short clinical vignette.The original slides were provided at a working group meeting for further(unblinded)examination. During the meeting,the cases were reviewed and discussed to draw upon the personal experience of the participants,the available literature,and the current diagnostic criteria.Two haematologists (J.M.and J.P.-W.)performed a later review of BM slides from16patients in the UK with epo-PRCA;the results are presented in this report. Because of the paucity of studies,it is recognised that this process was not evidence-based,but reflected the opinions of those expert physicians who have had the most direct experience in the diagnosis of epo-PRCA.BM recommendationsConsensus statements•BM evaluation is essential to confirm the diag-nosis of PRCA•BM investigation should normally include both an aspirate and a trephine biopsy•If no BM examination is possible,a suspected diagnosis of PRCA may still be possible,but the level of confidence may be lower and further testing may be necessary•PRCA is characterised by an almost complete absence of red cell precursors(<5%erythro-blasts):–higher,but still depressed,percentages(e.g.6%)should be described as hypoplasia –in the presence of hypocellularity,the5% limit may be less reliable–the threshold should be lower in hypercellular BM•If some erythroid activity is persistent,matur-ation arrest is characteristic(absence of late normoblasts)•It is not possible to distinguish epo-PRCA from other forms of PRCA on the basis of BM examination aloneDiscussionWhy is a BM examination necessaryA BM examination is necessary to confirm the diagnosis of PRCA and exclude other causes of anaemia,whether or not information is available about the presence of neutralising anti-erythropoie-tin antibodies.Furthermore,BM examination is widely used,and the results will usually be obtained before those of antibody testing,which may not be available to all clinicians.Even without antibody testing,BM evaluation combined with the clinical picture will allow a presumptive diagnosis on whichTable3.Proposed diagnostic criteria for epoetin-induced pure red cell aplasiaMajor features•Treatment with epoetin for at least3wk•Drop of haemoglobin of about0.1g/dL/d without transfusions or transfusionneed of about1unit/wk to keep haemoglobin level stable•Reticulocytes below10·109/L•No major drop of leucocytes and plateletsMinor features•Skin and systemic allergic reactionsConfirmational investigations•Bone marrow aspirate shows normal cellularity and<5%erythroblasts,withevidence of a maturation block•Serum assay shows presence of anti-erythropoietin antibodies and evidence oftheir neutralising capacityReproduced with permission from Eckardt and Casadevall(3).Diagnosis of epoetin-induced PRCA391treatment decisions can be made.BM examination thus represents good clinical practice,and the reasons for not performing it in individual cases must be well justified.Without a BM examination, the outcome of the diagnosis will depend on other information.A BM examination should be trig-gered by a rapid decrease in the reticulocyte count, as discussed later.In addition to excluding other causes of anaemia,the BM examination may provide information relating to the underlying cause of the PRCA(see Table1).What should be performedBecause the diagnosis has therapeutic implications, a differential count should be performed on500 marrow cells.The normal range of erythroblasts ranges from10to30%.In contrast,in PRCA, erythroblasts constitute<5%of marrow cells as erythroid precursors areÔblockedÕin the matura-tion process and do not form mature erythro-blasts.This value of5%represents clinical experience that has not been validated on a large case series.Occasionally,a slightly higher compo-nent of erythroblasts may be observed,which could perhaps best be described as erythroid hypoplasia. In the presence of BM infiltration with lymphoid cells,as in chronic lymphocytic leukaemia(CLL), the myeloid:erythroid(M:E)ratio should be taken into account.Although some cliniciansfind it hard to justify an invasive trephine biopsy,it does constitute good clinical practice and has a high possibility of influencing the care of the patient.A trephine biopsy is essential in addition to an aspirate in order to exclude lymphoproliferative disorders,e.g. CLL or lymphoma,and to assess overall cellularity. BM appearance in epo-PRCAThe current diagnostic criteria(Table3)stress an almost complete absence of red cell precursors in the presence of normal cellularity and an otherwise normal BM.In the context of epo-PRCA,however, it is not clear whatÔan otherwise normal BMÕmeans,as other abnormalities(such as MDS in elderly CKD patients)may be present and do not necessarily exclude PRCA.Without a BM examination before the develop-ment of suspected epo-PRCA,it is not possible to know whether the abnormalities were pre-existing or due to the disease.BM abnormalities may be quite common in the elderly,even in the absence of overt disease.Despite these caveats,the presence of other abnormalities in addition to the lack of red cell precursors should raise the suspicion of other associated diagnoses.BM criteria for distinguishing epo-PRCA and MDSClassic MDS is easily distinguished on BM exam-ination from classic PRCA(without myelodyspla-sia),but it should always be remembered that the two conditions can co-exist in CKD patients.If there is any reason other than anaemia and a low reticulocyte count to suspect MDS in a CKD patient about to start epoetin therapy,a BM examination should be performed before commen-cing therapy,as it is known that renal doses of epoetin are ineffective in most patients with MDS. The situation may be confusing in about5%of MDS patients who have<5%erythroblasts(17), but in these patients other features(BM and non-BM)will support the diagnosis.BM morphology reviewBone marrow slides were available from16of20UK patients and were independently reviewed by two haematologists from the working group(J.M.and J.P.-W.).Bone marrow aspirates were available in 14of16cases and trephine biopsies in13of16.Bone marrowfindings were consistent with a diagnosis of PRCA in all cases.In10cases,there was a complete absence of BM erythroid cells.In the remaining six, there was erythroid hypoplasia with only an occa-sional late erythroblast seen.Infive of these six cases,the remaining erythroblasts were<5%of all nucleated cells and in one case they were6%, consistent with PRCA as outlined by these recom-mendations.Giant pronormoblasts,characteristic of parvovirus-induced PRCA,were not observed in any of the cases.Bone marrow eosinophilia was noted in all except one case(median8%,range 5–17%;normal range0.3–4%)(18).Bone marrow cellularity was evaluated in13patients in whom trephine biopsies were available.In view of the normal reduction in cellularity reported with increasing age(19,20),age-adjusted cellularity was normal in nine and increased in four patients.There was no evidence of abnormal BM infiltration or other underlying pathology in any of the16patients. Bone marrow analyses were performed after recov-ery of PRCA in two patients;in both cases,the BM values had reverted to normal with full recovery of erythropoiesis and disappearance of eosinophilia. Distinguishing epo-PRCA from other forms of PRCAAs the cardinal diagnostic feature of epo-PRCA is the presence of anti-erythropoietin antibodies in the serum,it is not possible to distinguish epo-PRCA from other forms of PRCA on the basis of BM examination alone.As discussed above,however, the possibility of multi-causality should always beCasadevall et al. 392considered,and the BM examination can also exclude other causes of anaemia,including poor compliance with epoetin treatment.Clinical experi-ence from before the epoetin era suggests that BM findings in patients with CKD and severe chronic anaemia could hardly ever be confused with PRCA, as the remaining level of erythropoietin is sufficient to inhibit apoptosis and the morphological picture of PRCA is not seen.When epoetin is administered intravenously,there is little opportunity for non-compliance,but patients who self-administer sub-cutaneously(now contra-indicated in Europe for EPREXÒ)may miss doses(accidentally or inten-tionally).Other haematological recommendationsConsensus statements•A full assessment of peripheral blood[full blood count/complete blood count(FBC/CBC)]is mandatory,with bloodfilm examination and absolute reticulocyte count.•A reticulocyte count£30·109/L or a50% decrease(without a change in epoetin dose) should be the trigger for further investigation.•Testing for anti-erythropoietin antibodies is mandatory.•Exclusion of other causes of PRCA is manda-tory in antibody-negative patients.DiscussionReticulocyte countingAlthough the reticulocyte count is the best labor-atory marker of erythropoiesis available,peripheral blood reticulocytes have not been extensively stud-ied in renal disease(21),and many nephrologists have only recently become aware of their import-ance.The reticulocyte count provides information about BM activity and the effectiveness of daily red cell production(22,23)and is thus crucial in the diagnosis of patients with anaemia.For example, reticulocytopenia occurs where there is failure of red cell production.In contrast,reticulocytosis reflects increased red cell production by the BM when there is peripheral destruction of red cells as in haemolysis or loss of red cells as in acute haemorrhage.The manual counting of reticulocytes has been the standard routine laboratory method since it wasfirst developed,although it is now increasingly being replaced by semi-automated or fully automated methods(22).Although it requires only a light microscope,a trained technician is needed to perform a manual count and can be a source of variability.In addition,the values obtained are percentage or proportional values, and not an absolute count.The relative reticulocyte count is misleading when the red blood cell count is abnormal,as in severe anaemia,and must be corrected accordingly(22).However,the correction is based on the assumption of a linear relationship between packed cell volume and reticulocyte mat-uration time,which is appropriate only in patients with an intact haematopoietic system,in whom erythropoietin production is normal and the BM responds appropriately(23).The best way of correcting for the effects of anaemia is to calculate the absolute number of reticulocytes,which is normally between50and80·109/L for a normal haemoglobin level.Automated reticulocyte counts are now increasingly being used and provide superior precision to manual counts.The modified criteria of Eckardt and Casadevall (3)specify that in epo-PRCA,the absolute reticu-locyte count is below10·109/L.This is based on clinical observation and gives a clear separation of CKD patients without PRCA,almost all of whom have a reticulocyte count above30·109/L(median about70·109/L),when treated with epoetin. While this is an acceptable diagnostic criterion,it may be too stringent if the reticulocyte count is used as a trigger for further investigation and risks missing an early diagnosis of epo-PRCA.A repeat-edly obtained value of<30·109/L or a50%fall in reticulocyte count is a more useful early warning marker.In order to avoid missing the diagnosis until epo-PRCA is fully developed,the absolute (not percentage)automated reticulocyte count should be determined at monthly intervals(at the same time as the FBC/CBC)for all CKD patients starting epoetin therapy,and repeated within1wk if it is abnormal.As most cases of epo-PRCA have developed within1–2yr of starting epoetin treat-ment,the frequency of determining reticulocyte count could be decreased to every2–3months after this time.For many nephrologists,however,it may be more practical to link the reticulocyte count/ CBC to haemoglobin measurement and obtain both monthly.It should be noted that the reticu-locyte count in MDS may sometimes be below 10·109/L,and a definitive distinction is dependent upon other parameters such as the presence or absence of antibodies against epoetin or relevant cytogenetic data.If there are dysplastic changes on careful examination of the bloodfilm,then a BM examination should be performed,including cyto-genetic analysis.Antibody testingThe overriding feature of epo-PRCA is the presence of anti-erythropoietin antibodies,and antibodyDiagnosis of epoetin-induced PRCA393testing is therefore mandatory.The various tests available to detect the presence of anti-erythropoie-tin antibodies include radioimmunoprecipitation assays(RIPA)(24),enzyme-linked immunosorbent assays(ELISA)(25–27)and surface plasmon res-onance methods(BIACORE biosensor immunoas-says)(28,29).RIPA and ELISA are the most widely used,but may give misleading results if incorrectly formatted.All of these assays vary in sensitivity and specificity,and it should be remem-bered that a negative result from antibody testing, particularly with ELISA,might reflect an inability of the assay used to detect such antibodies in a given patient.Thus a second method of detection should be used if there is suspicion of epo-PRCA.If the BM examination and the clinical picture strongly suggest PRCA but the antibody test is negative,other causes or co-existent BM disorders must be excluded.The definitive assay for biologically active anti-erythropoietin antibodies is a bioassay,which determines the capacity of serum from a patient with suspected epo-PRCA to neutralise the action of epoetin in stimulating the proliferation of colonies of cells derived from normal BM or from an epoetin-dependent cell line(12).As virtually all anti-erythropoietin antibodies from patients with epo-PRCA have proved to be neutralising,and because the assays are not widely used,therapeutic decisions may be made in the absence of bioassay results.However,if possible,this assay should be included in the investigation.Haemoglobin levelsAs discussed previously,it is often impossible to document the rate of fall of haemoglobin accu-rately.In addition,different rates of fall are seen in different patients with suspected epo-PRCA.The need to document a rapid decrease in haemoglobin level is circumvented by mandatory measurement of the reticulocyte count,which will then act as a more sensitive trigger for further investigation. However,an unexplained,sudden,rapid and sus-tained decrease in haemoglobin level in a CKD patient who has previously responded to epoetin therapy should act as a trigger for careful evalua-tion.In view of these considerations,it was not thought helpful to specify a rate of fall of haemo-globin concentration as a diagnostic recommenda-tion.The loss of effect because of causes other than PRCA also manifests as a decrease in haemoglobin level,but this is usually clearly distinguishable from the sudden decrease in PRCA.If there is any reason to suspect another cause of loss of effect,this should be investigated.Other testsIn patients with a BM examination consistent with PRCA but in whom the presence of anti-erythro-poietin antibodies has not been detected by two different tests(RIPA/ELISA/BIACOR),it is neces-sary to exclude other causes of PRCA.Recent parvovirus B19infection can be detected by IgM serology.Alternatively,the more sensitive polym-erase chain reaction assay can be performed, although this is not widely used.A careful history may provide clues as to the possibility of parvo-virus B19infection.Immunophenotyping of lymphocytes and their subsets will be helpful to investigate the possibility of lymphoproliferative or T-large granular lymphocyte disorders(30),while computed tomography and magnetic resonance imaging of the thorax may detect a thymoma associated with PRCA.Paroxysmal nocturnal haemoglobinuria should be tested for by determin-ing CD55and CD59cell markers on red cells and neutrophils.Serum erythropoietin concentration is generally low in CKD patients and is usually undetectable in epo-PRCA;the assay used to measure this is not routinely used.Reticulocyte maturity can now be measured byfluorescence intensity in an automated reticulocyte counting instrument(31).However,the method is not yet widely used or well validated. Assessment of cultures of erythroid burst-forming units and colony-forming units(which measure the erythroid proliferation potential of the BM)are useful in that inhibition of erythroid growth by patient’s serum is strongly indicative of epo-PRCA, but unfortunately these assays are performed only in specialised laboratories.Although other causes of PRCA may be unlikely if anti-erythropoietin antibodies have been demonstrated,and only limited investigations are appropriate in such patients,physicians should keep an open mind about the possibility of dual pathology.Recommendations for monitoring and investigatingepo-PRCAIn the light of their discussions,the members of the ad hoc working group formulated a set of recom-mendations for haematological approaches to monitoring and investigating suspected epo-PRCA (Table4).It must again be emphasised that these are not evidence-based guidelines,but are recom-mendations based on what the participating physi-cians considered to be best practice.A sequence of investigations was not thought to be appropriate,and on suspicion of PRCA,all the investigations should be performed.Although itCasadevall et al. 394。