Add-on erythropoietin in autoimmune hemolytic anemia

In this issue of Blood Advances, Fattizzo et al¹ showed a significant increase in hemoglobin levels after subcutaneous injection of erythropoietin (EPO) in patients with autoimmune hemolytic anemia (AIHA).

Warm AIHA is usually due to immunoglobulin G (IgG) class antibodies, less frequently because of IgM and IgA warm antibodies.² Cold agglutinin disease (CAD) is a distinct clonal lymphoproliferative disorder, and typically complement fixing IgM autoantibodies bind to the I antigen preferably at low temperatures. However, a high thermal amplitude of the antibodies can lead to hemolysis at room temperature.

Symptomatic anemia is the principal indication for treatment, and compensated hemolysis (eg, >10 g/dL) may not necessarily require therapy. Initial therapy includes eventual administration of prednisolone and rituximab, based on 2 prospective randomized trials.³ Rituximab may be most efficient in second-line therapy; third-line treatment includes splenectomy or immunosuppression. Additionally, severe hemolysis requires salvage therapy including transfusions, glucocorticoids, IV immunoglobulin, or plasma exchange.³ 

In CAD, one may either target the clone by B-cell–directed therapy or the classical pathway of complement-mediated hemolysis by inhibiting upstream complement factors (C1s and C3). Rituximab monotherapy provides moderate response rates (∼50%) and has been considered as first-line therapy; if patients are fit enough, a combination with bendamustine provides a higher chance for long-term and complete responses.³ The onset of effect may show a considerable lag time of 1 to 3 months, and deepening of responses may be seen in the long term with the combination therapy. Because opsonized C3b-positive erythrocytes are mainly removed by the liver, splenectomy is usually ineffective and, therefore, contraindicated. Of note, treatment options commonly used in warm AIHA, including corticosteroids, do not usually work in CAD.³ In contrast, hemolysis is immediately inhibited by IV infusion of sutimlimab (TNT009 and BIVV009), a monoclonal IgG4 antibody directed against the complement factor C1s. Sutimlimab was authorized based on 3 trials as reviewed recently^4,5 and other evidence from a named patient use.^6,7 Blood transfusions require sample handling and transfusions at ∼37°C, and the target of cold agglutinins (I antigen) is expressed on virtually all adult erythrocytes. Therefore, prior inhibition of the classical complement pathway will be essential to protect newly transfused packed red blood cells from rapidly lysing again.

Inadequate reticulocytosis was observed in 20% to 40% of AIHA cases, correlating with a poor prognosis. Hence, over the last years, increased attention has been paid to bone marrow compensation.⁸ EPO levels may be inappropriately low in patients with AIHA. A negative correlation between bone marrow compensatory response and EPO was observed in a retrospective multicenter study of 51 patients with AIHA treated with recombinant human erythropoietin (rhEPO) after failure to respond to previous therapies. Patients showed an overall response rate of 55% after 15 days.⁹ 

Therefore, Fattizzo et al conducted a prospective trial to confirm the safety and efficacy of EPO in patients with AIHA.¹ They enrolled 47 patients in a single-arm trial. Important inclusion criteria were hemolysis, a hemoglobin value of <10 g/dL, an inadequate reticulocyte count reflected by a bone marrow responsiveness index of <121, and insufficient EPO levels.

Relevant exclusion criteria were deficient of nutrients, bleeding, moderate or severe renal insufficiency, unstable doses of steroids or immunosuppressants, and recent prior use of rituximab or IV Igs. However, an increase in steroid doses, rituximab, and Igs was allowed as concomitant therapy as early as 8 days after study enrollment.¹ 

Patients received flat doses of 40 000 international units of rhEPO subcutaneously once a week until Hb exceeded 11 g/dL. The dose was chosen based on the previous retrospective study and EPO use in myelodysplastic syndrome.¹¹ The results were compared with results of external controls (similar patients with AIHA who did not receive rhEPO). Hemoglobin values increased from baseline by a median of 1.4, 2.4, and 3.4 g/dL after 14 days, 1, and 3 months, respectively, and on average patients stopped EPO after 4 months. Simultaneously transfusion needs decreased from 30% to <10%. The overall response rates of 55% at 15 days were consistent with those observed in the retrospective study; they were 23 percentage points higher, and hemoglobin values were 0.8 g/dL higher than those in the control group at 15 days. Numerically higher response rates and higher Hb values persisted at 1 year. Despite the lack of randomization, the prospective study further supports a role for rhEPO as an add-on treatment to immunosuppressive therapy, which may reduce the risk for alloimmunization because of transfusions.

Where does rhEPO best fit into the current treatment landscape? Finding new treatments is particularly important for refractory patients with AIHA, who require (regular) blood transfusions. As proposed, rhEPO can be used to accelerate the response to immunosuppressive therapy, which is particularly important for rituximab therapy, as well as hemolytic crisis which may lead to intensive care unit admissions and an associated high mortality rate. Whether rhEPO use may decrease the burden of immunosuppressive therapy, and associated risk of infections, would require a large, randomized trial.

Not unexpectedly, patients with warm AIHA responded better than patients with CAD (73% vs 26% at day 15). However, we observed good responses when lower doses of rhEPO were injected 3 times weekly to 2 patients with CAD treated with sutimlimab, who also suffered from renal insufficiency.⁷ Although sutimlimab stopped hemolysis in these patients, their rise in Hb was slow and fastened by rhEPO. In the absence of proper dose finding, individual dose titration may be appropriate. However, based on the mechanism of action, and similar to sutimlimab, EPO cannot improve circulatory symptoms (acrocyanosis) which are due to agglutination of erythrocytes in CAD. In addition, EPO enhances platelet reactivity and platelet and endothelial activation in humans, indicating a thrombogenic effect of erythropoietin.¹⁰ Finally, EPO (in contrast to sutimlimab) cannot dampen inflammation induced by C3a or C5a, which may contribute to thrombosis. Together, this could at least theoretically enhance thrombosis risk, although thrombosis rates were similar between patients receiving EPO and controls. However, in view of the small sample sizes, the important benefits and potential risks of thromboprophylaxis should be discussed with patients with hemolysis.

In summary, add-on EPO can contribute to our current therapeutic armamentarium to treat a subgroup of patients most severely affected by AIHA.

Conflict-of-interest disclosure: Reimbursement for scientific advice and expenses related to presentations was provided to B.J. by Sanofi and Sobi. M.M.M. declares no competing financial interests.

Correspondence: Bernd Jilma, Department of Clinical Pharmacology, Medical University of Vienna, Waehringer Guertel 18-20, Vienna 1090, Austria; email: bernd.jilma@meduniwien.ac.at.