INTRODUCTION — Acquired pure red cell aplasia (PRCA) is a rare condition of profound anemia characterized by a very low reticulocyte count and the virtual absence of erythroid precursors in the bone marrow. All other cell lines are present and seem quantitatively and morphologically normal. Many cases of acquired PRCA are idiopathic. In others, underlying conditions, such as thymoma, myelodysplastic syndromes, lymphoma, leukemia, systemic autoimmune disorders, and viral infection (ie, parvovirus B19), or treatment with drugs, such as phenytoin or chloramphenicol, are identified [1]. (See "Acquired pure red cell aplasia in adults".)
PRCA has also been described in patients treated with epoetin and other erythropoiesis-stimulating agents (ESAs), resulting from the induction of neutralizing antibodies directed against the erythropoietin (EPO) molecule [2]. Most reported cases have been in patients receiving epoetin for chronic kidney disease (CKD)-related anemia.
PRCA due to anti-EPO antibodies in patients with CKD will be reviewed here. A discussion of other causes of acquired PRCA is presented separately. (See "Acquired pure red cell aplasia in adults".)
ETIOLOGY AND PATHOGENESIS — Most cases of non-erythropoietin (EPO)-related PRCA are mediated by immunoglobulin G (IgG) autoantibodies or cytotoxic T lymphocytes directed against erythroid precursor or progenitor cells [1,3]. Inhibition of erythropoiesis occurs primarily between the colony forming unit–erythroid (CFU-E) and proerythroblast stages. The proerythroblast is the first morphologically identifiable erythroid precursor in normal marrow and is virtually absent in the marrow of patients with PRCA. (See "Acquired pure red cell aplasia in adults".)
PRCA due to autoantibodies against endogenous EPO is rare in patients who have never been treated with erythropoiesis-stimulating agents (ESAs) [4-6]. The serum of patients with EPO-related PRCA inhibits the growth of erythroid progenitor cells in bone marrow cultures [2]. Neutralizing IgG antibodies to the protein component of exogenous ESAs crossreact with endogenous EPO.
There are several preparations of epoetins (ie, epoetin alfa, epoetin beta, epoetin omega, and others) produced by various manufacturers around the world. They differ from one another and from the native hormone in terms of glycosylation and sialic acid content [7]. The vast majority of cases of EPO-related PRCA have occurred in patients treated with a particular epoetin alfa product (Eprex, in single-use syringes), which was manufactured and distributed outside the United States. Altered antigenicity of this specific product has been suggested as the underlying cause of anti-EPO antibody development. However, EPO-related PRCA has been described with other preparations including darbepoetin alfa and methoxy polyethylene glycol-epoetin beta [8,9].
Virtually all reported cases of anti-EPO antibody-mediated PRCA have occurred in patients with chronic kidney disease (CKD) who have received the drug subcutaneously [2,10-12]. Although not specifically studied with epoetin, subcutaneous administration of other drugs has been associated with greater immunogenicity than with intravenous (IV) administration [13].
EPIDEMIOLOGY — Most cases of erythropoietin (EPO)-related PRCA occurred in patients with renal failure in Europe and the United Kingdom, but patients in Canada, Australia, and countries in Asia have also been affected [10,14].
The condition remains extremely rare, given the widespread use of epoetin and other erythropoiesis-stimulating agents (ESAs) [15-17]. The overall incidence of reported cases between 1989 and June 2004 was 1.6 per 10,000 patient-years of subcutaneous exposure [18]. There have been only a few cases reported in patients treated with intravenously (IV) administered epoetin, with an incidence of 0.02 per 10,000 patient-years of exposure [19].
Eprex — There have been over 200 reported cases of PRCA related to Eprex use [20], representing the vast majority of affected patients [10,14]. There was a substantial increase in the number of reported cases between 2001 and 2003, almost exclusively in patients with chronic kidney disease (CKD) who received Eprex subcutaneously [10,14,18].
This increase in the number of cases was later discovered to coincide with the release of a new formulation and packaging of this epoetin, as discussed above. The incidence was 3.43 per 10,000 patient-years of exposure for Eprex supplied in syringes with uncoated rubber stoppers compared with 0.23 and 0.17 per 10,000 patient-years for Eprex preparations in coated rubber-stopper syringes and stabilized with human serum albumin (HSA), respectively.
One report of the multinational Prospective Immunogenicity Surveillance Registry (PRIMS) described five confirmed episodes of EPO-related PRCA [21]. Three episodes occurred with a reformulated Eprex preparation, and two occurred with other ESAs among over 15,000 patients, with an incidence (per 100,000 patient-years) of 35.8 for Eprex and 14 for other ESAs (rate ratio 2.56, 95% CI 0.43-15.31), which was not statistically significantly different.
Non-Eprex cases — The overall incidence of reported cases between 1989 and 2004 of PRCA with subcutaneous use of non-Eprex preparations was much lower, ranging from 0.02 to 0.16 per 10,000 patient-years between 1989 and 2004, depending on the preparation [10,14,17,22-26]. Fortunately, EPO PRCA has become exceedingly rare since 2005 [26].
EPO antibody-mediated PRCA has only rarely been reported in children with kidney failure [14,27].
"Biosimilar" epoetin products are becoming increasingly available around the world; treatment with these agents has also been associated with development of anti-EPO both non-neutralizing and neutralizing antibodies, with a notable outbreak in Thailand related to subcutaneous use of a biosimilar produced outside the United States and the European Union [28-31].
SCREENING FOR ANTI-EPO ANTIBODIES — Among dialysis patients, routine screening for the presence of anti-erythropoietin (EPO) antibodies cannot be justified. This was shown in a study involving five centers in Canada, in which approximately 1500 hemodialysis, peritoneal dialysis, and predialysis patients were screened [32]. Using a radioimmunoprecipitation assay (RIPA), a low-positive test was noted in only one patient (who had previously been diagnosed with PRCA), and borderline positive assays were observed in only three individuals, none of whom had signs of PRCA. None of the four had neutralizing anti-EPO antibodies.
Similar results were obtained in 536 patients screened in 35 German dialysis units [33]. In this study, anti-EPO antibodies were detected in three EPO-hyporesponsive and three EPO-normoresponsive patients using an enzyme-linked immunosorbent assay (ELISA), with only one borderline result with a RIPA. There were no cases of PRCA.
However, we agree with the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) guidelines, which suggest that an evaluation for PRCA due to anti-EPO antibodies should occur in the patient exposed to at least eight weeks of erythropoiesis-stimulating agent (ESA) therapy who develops all of the following [34]:
●Decline in hemoglobin (Hb) level of >0.5 to 1 g/dL per week or transfusion requirement of at least one to two units per week to maintain adequate Hb
●Normal platelet and white blood cell (WBC) count
●Absolute reticulocyte count of <10,000/microL
CLINICAL MANIFESTATIONS — PRCA due to anti-erythropoietin (EPO) antibodies should be suspected if the hemoglobin (Hb) level declines by >2 g/dL per month or the reticulocyte count is <20,000/microL in an individual who has previously responded to EPO. PRCA generally does not occur, unless the patient has been on EPO for at least three to four weeks, and typically occurs after 6 to 18 months of exposure [10].
PRCA is specifically characterized by the following clinical features [10,35]:
●Decline in Hb level of >0.7 to 1 g/dL per week without transfusions or transfusion requirement of at least one unit per week to maintain adequate Hb, despite continued use of EPO at the same or increased doses
●Markedly reduced reticulocyte count (<10,000/microL)
●Drop in platelet counts, although generally not below the lower limit of normal
●Normal white blood cell (WBC) count
●Elevated serum transferrin saturation and serum ferritin, reflecting decreased utilization of iron secondary to diminished erythropoiesis
Allergic urticarial skin reactions at sites of earlier subcutaneous EPO injections have also been described [36].
EVALUATION AND DIAGNOSIS — To properly diagnose the disorder, a bone marrow aspirate and evaluation for the presence of anti-erythropoietin (EPO) antibodies must be performed.
Bone marrow aspirate — The bone marrow aspirate reveals severe erythroid hypoplasia, with <5 percent red blood cell precursors; evidence of a block in the maturation of erythroid precursors may be present. Platelet and white cell precursors are entirely normal.
Anti-EPO antibodies — Identification of antibodies to EPO is a critical component of the diagnosis. There are several available tests to detect antibodies that bind EPO, each with its own advantages and disadvantages [37]:
●Radioimmunoprecipitation assay (RIPA) appears to be the most accurate test for detecting anti-EPO antibodies. However, it is not standardized, may not detect low-affinity antibodies, requires radiolabeled antigen, and is time consuming and difficult to automate [10].
●Enzyme-linked immunosorbent assays (ELISAs) are more widely available but appear to have lower sensitivity and specificity than the RIPA [10,38].
●A biosensor assay, which is not as readily available as the previous two assays, may provide better characterization of detected antibodies (ie, binding affinity and isotype). However, it may detect very low-affinity antibodies in baseline sera that never cause PRCA.
In general, these assays can detect antibodies with a concentration of at least 500 ng/mL [37]. The neutralizing capability of the anti-EPO antibodies cannot be assessed with any of the above-mentioned assays. This requires a bioassay in which patient serum or immunoglobulin inhibits red blood cell precursor growth in bone marrow or cell line cultures [10,35]. In addition, non-neutralizing antibodies may be more common than neutralizing antibodies [31].
It is difficult to recommend use of any one of these assays over another given that they are not standardized, have advantages and disadvantages that are difficult to compare, and are not available commercially [39]. The World Health Organization (WHO) established an EPO-antibody reference panel that should be useful for assay standardization and improving diagnostic accuracy [40].
The sample in a suspected case may be sent to the manufacturer, where several different assays, including one for neutralizing capability, are performed [41]. Confirmation of a suspected case of anti-EPO antibody-mediated PRCA should show the presence of anti-EPO antibodies and evidence of neutralizing ability [14,35]. Bioassays generally detect antibodies with a concentration of at least 1 mcg/mL [37].
DIFFERENTIAL DIAGNOSIS — In addition to PRCA due to anti-erythropoietin (EPO) antibodies, there are a large number of other causes of anemia and hyporesponsiveness to EPO among patients with chronic kidney disease (CKD). They can be distinguished from PRCA largely because of the observation that PRCA due to anti-EPO antibodies is associated with a profound fall in hemoglobin (Hb) level and reticulocyte count [37]. By comparison, the other causes of EPO resistance are associated with less severe drops in counts. Other causes of EPO resistance are discussed separately. (See "Hyporesponse to erythropoiesis-stimulating agents (ESAs) in chronic kidney disease", section on 'Causes'.)
TREATMENT — There is limited experience with managing PRCA. Treatment consists of:
●Initial management including transfusions for symptomatic anemia and discontinuing all recombinant erythropoietin (EPO) products
●Immunosuppressive therapy to eradicate antibodies
Among patients who do not respond to immunosuppressive therapies and are eligible candidates, consideration should be given to kidney transplantation.
Initial management — The two most important initial steps in management in anti-EPO antibody-mediated PRCA are transfusions for symptomatic anemia and stopping all recombinant EPO products [10]. There are at least three observations underlying the recommendation to avoid any recombinant EPO preparation, including darbepoetin alfa, in patients with anti-EPO antibody-mediated PRCA:
●Anti-EPO antibodies crossreact not only with the endogenous hormone, but also with all recombinant EPO molecules, including darbepoetin alfa [2].
●Rechallenge with EPO preparations may cause an anamnestic antibody response, making it less possible for the antibody to either spontaneously disappear or return to clinically unimportant levels.
●Rechallenge with EPO may incite the formation of allergic skin and systemic reactions (eg, anaphylaxis), which might further complicate the patient's clinical picture [36].
However, there have been case reports of patients who have tolerated rechallenge [8,42-47]. Despite these case reports and other isolated instances in which patients appeared to tolerate resumption of EPO therapy or other EPO preparations, in some instances with concomitant immunosuppression [42,48], patients with EPO-associated PRCA should have EPO treatment discontinued and should not routinely be switched to an alternative EPO product or to darbepoetin alfa.
If, however, rechallenge is considered for clinical reasons or at the patient's request, the following practice appears reasonable [45]:
●Rechallenge only if anti-EPO antibody levels are below the lower limit of detection
●Monitor the absolute reticulocyte count and anti-EPO antibody levels and for systemic reactions
●Administer methoxy polyethylene glycol-epoetin beta, rather than EPO or darbepoetin
●Administer methoxy polyethylene glycol-epoetin beta intravenously, rather than subcutaneously, whenever feasible
Immunosuppressive therapy — Since PRCA in this setting is immune mediated, the condition should respond to immunosuppressive therapy, and, while spontaneous remissions after cessation of EPO therapy occur, they appear to be rare; immunosuppressive therapy should probably be provided in most cases [45].
There are no controlled data, but several different immunosuppressive regimens have been used. In one retrospective review of 47 cases in Europe [49], EPO therapy was stopped in all patients. Ten patients did not receive immunosuppressive therapy, one died within six weeks of diagnosis, and nine did not recover by a median follow-up of 12 months.
The remaining 37 patients were treated with a variety of immunosuppressive agents. The majority received steroids, alone or in combination with cyclophosphamide, intravenous immune globulin (IVIG), or plasmapheresis; six received cyclosporine alone; one received mycophenolate mofetil; and two received rituximab. Six patients received a kidney transplant. The following results were reported:
●Seventy-eight percent of all patients recovered, most within three months.
●Recovery occurred only after antibody levels became undetectable.
●No relapse of PRCA occurred following cessation of immunosuppression, but none of the patients were rechallenged with EPO or darbepoetin alfa following recovery.
●All six patients who received a kidney transplant recovered.
Favorable responses were seen with many of the regimens. This included corticosteroids with or without IVIG (10 of 18 recovered) and corticosteroids with cyclophosphamide (seven of eight recovered), as well as with cyclosporine alone (four of six recovered), which resulted in the fastest response [49,50]. In this study, none of the patients treated with rituximab, an antibody against B cells that has been effective for non-EPO-related PRCA, recovered [49], although response to rituximab has been reported elsewhere [51,52]. Only one of nine patients who were treated with IVIG alone recovered.
Tacrolimus has also been used successfully in a case report [53].
In one review, there was an analysis of EPO-related PRCA, as reported to the US Food and Drug Administration (FDA) and the manufacturers of several EPO preparations, as well as longer-term follow-up of patients previously reported and described above from the European PRCA Study Group [49]. It was concluded that the highest recovery rates were associated with combination cyclophosphamide and prednisone and that the highest rates of EPO responsiveness on rechallenge were in those patients whose anti-EPO antibody levels were undetectable [54].
The treatment recommendations that follow are made with the caveat that worldwide experience is very limited, and none of the data are from randomized, controlled trials. Given that the highest recovery rates appear to be associated with combination cyclophosphamide and prednisone, we prefer initial therapy consisting of prednisone (1 mg/kg per day) plus oral cyclophosphamide (50 to 100 mg per day). A reasonable alternative is cyclosporine alone at a dose of 200 mg per day (or 100 mg twice per day) [45,54]. Patients who do not respond to treatment with either oral cyclophosphamide plus prednisone or with cyclosporine alone may be treated with the other regimen; tacrolimus may also be considered as an alternative to cyclosporine. (See "General principles of the use of cyclophosphamide in rheumatic diseases".)
The cost, potential toxicities, and limited effectiveness [49] of IVIG and rituximab suggest that these treatments should perhaps be limited to those who fail to respond to other therapy. (See "Overview of intravenous immune globulin (IVIG) therapy".)
The optimal duration of immunosuppressive therapy and monitoring for anti-EPO antibody-mediated PRCA is unknown. However, the following recommendations appear reasonable:
●Continue treatment until antibody levels become undetectable; discontinue treatment three or four months after initiation among patients with no response.
●Monitor hemoglobin (Hb) weekly to assess for transfusion requirement and response to therapy.
●Monitor reticulocyte count and anti-EPO antibody levels every one to two weeks during treatment to assess response or lack thereof.
Since the majority of patients who respond to treatment do so within three to four months, immunosuppressive therapy should probably be discontinued if there is no response within this time. However, in the abovementioned study of 37 treated patients, 4 patients responded 5 to 18 months after treatment was initiated [49].
With the introduction of "biosimilar" epoetins and increasing use of unlicensed epoetin products in certain parts of the world, ongoing vigilance must be maintained for new occurrences of anti-EPO antibody-mediated PRCA [31,55].
Experimental therapies — Hypoxia-inducible factor prolyl hydroxylase inhibitors (HIF PHIs) are a novel class of oral drugs that stimulate the production of endogenous erythropoietin. The HIF-PHI roxadustat may be helpful for the management of anemia in patients with EPO-associated PRCA [56-60]. However, these data are limited to case reports and additional data are needed before these can be routinely recommended for management of EPO-related PRCA.
SUMMARY AND RECOMMENDATIONS
●The vast majority of cases of recombinant human erythropoietin (EPO)-related acquired pure red cell aplasia (PRCA) have occurred in patients treated with a particular epoetin alfa product: Eprex (in single-use syringes). The underlying cause may be organic compounds (leached by polysorbate from uncoated rubber stoppers in prefilled syringes) that are acting as adjuvants, resulting in anti-EPO antibody development. Virtually all reported cases of anti-EPO antibody-mediated PRCA have occurred in patients with chronic kidney disease who have received the drug subcutaneously. (See 'Etiology and pathogenesis' above.)
●EPO-related PRCA remains extremely rare, given the widespread use of erythropoiesis-stimulating agent (ESAs). There have been >200 reported cases of PRCA related to Eprex use, which represent the vast majority of affected patients. Following changes in storage and handling recommendations and discontinuation of subcutaneous administration of Eprex, the incidence of anti-EPO antibody-mediated PRCA in patients with kidney disease has declined. (See 'Epidemiology' above.)
●Screening for the presence of anti-EPO antibodies should not be routinely performed in dialysis patients. (See 'Screening for anti-EPO antibodies' above.)
●EPO-induced PRCA should be considered in the patient with significant anemia who has been treated with epoetin or another ESA for at least three to four weeks and has previously responded to treatment. The condition is characterized by a sudden decline in hemoglobin (Hb) level despite continued use of the ESA, markedly reduced reticulocyte count, and normal white blood cell (WBC) and platelet counts. (See 'Clinical manifestations' above.)
●To definitively diagnose EPO-induced PRCA, a bone marrow aspirate and evaluation for the presence of neutralizing anti-EPO antibodies should be performed. The bone marrow reveals severe erythroid hypoplasia, with <5 percent red blood cell precursors, and there may be evidence of a block in the maturation of erythroid precursors. Platelet and white cell precursors are entirely normal. Anti-EPO antibodies are detected by radioimmunoprecipitation assay (RIPA), enzyme-linked immunosorbent assays (ELISAs), or other assay, as available. The sample must be sent to the manufacturer for testing to document the presence of anti-EPO antibodies and evidence of their neutralizing ability. (See 'Evaluation and diagnosis' above.)
●We recommend the cessation of all recombinant ESAs in patients with PRCA (Grade 1A). We also recommend not switching to an alternative ESA (Grade 1A). Patients should be transfused for severe symptomatic anemia. Our recommendations concerning rechallenge are discussed below. (See 'Immunosuppressive therapy' above.)
●Given that spontaneous remissions after cessation of epoetin therapy are rare, we recommend the administration of immunosuppressive therapy in most patients (Grade 1B). We suggest initial therapy consisting of prednisone (1 mg/kg per day) plus oral cyclophosphamide (50 to 100 mg per day) for a maximum of three to four months (Grade 2C). A reasonable alternative for first-line therapy is cyclosporine alone at a dose of 200 mg per day (or 100 mg twice per day) for a maximum of three to four months. Patients who do not respond to initial treatment with either oral cyclophosphamide plus prednisone or with cyclosporine alone may be subsequently treated with the other regimen. Tacrolimus may also be considered as an alternative to cyclosporine. Consideration should be given to kidney transplantation in all eligible patients if there is no response to this strategy. (See 'Immunosuppressive therapy' above and "General principles of the use of cyclophosphamide in rheumatic diseases".)
●The optimal duration of immunosuppressive therapy and monitoring for anti-EPO antibody-mediated PRCA is unknown. However, the following recommendations appear reasonable (see 'Immunosuppressive therapy' above and "General principles of the use of cyclophosphamide in rheumatic diseases"):
•Continue treatment until antibody levels become undetectable; discontinue treatment three or four months after initiation among patients with no response.
•Monitor Hb weekly to assess for transfusion requirement and response to therapy.
•Monitor absolute reticulocyte count and anti-EPO antibody levels every one to two weeks during treatment to assess response or lack thereof.
●Rechallenge may be considered in patients in whom anti-EPO antibody levels are below or closer to the lower limit of detection. If performed, we recommend administration of the methoxy polyethylene glycol-epoetin beta intravenously in hemodialysis patients. The Hb, reticulocyte count, and anti-EPO antibody levels should be closely monitored. (See 'Initial management' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Stanley L Schrier, MD (deceased), who contributed to an earlier version of this topic review.
