Outpatient subcutaneous alemtuzumab is feasible and safe for aplastic anemia and is associated with high response rates

Aplastic anemia (AA) is a life-threatening disease characterized by cytopenias and a hypocellular bone marrow (BM).^1,2 Many hypotheses have been described to explain the complex pathophysiology of AA, which is not yet fully understood. A dysregulated immune system leads to a hematopoietic stem cell attack by autoreactive T cells, and adequate responses to immunosuppressive therapy are observed.^2,3 

The standard treatment for patients who lack an HLA-identical sibling donor is a combination of horse antithymocyte globulin (ATG; h-ATG) and cyclosporine (CsA).^4,5 This regimen has been shown to achieve an overall hematologic response rate of 60% to70%.^6,7 To improve the outcomes of h-ATG and CsA, several clinical trials have been conducted, but they have failed to demonstrate the superiority of replacing h-ATG with rabbit ATG (r-ATG) or alemtuzumab (ALZ) or adding sirolimus, mycophenolate mofetil, or hematopoietic growth factors.^4,8-12 In a phase 3 multicenter prospective trial, adding eltrombopag to h-ATG and CsA improved hematologic response, with faster and higher quality of responses with no significant additional adverse events, although relapses did not differ significantly.^3,13-17 

Unfortunately, h-ATG was discontinued in most South American and Asian countries in 2007 and in several European countries for some years.^10,18 Moreover, the access to effective therapies for AA is severely limited in several countries, which can be attributed to various factors such as availability, regulatory approvals, and reimbursement challenges.

ALZ, a humanized anti-CD52 antibody, exerts a lymphocytotoxic effect by targeting CD52-bearing cells through both antibody-dependent cellular cytotoxicity and complement-mediated lysis mechanisms.^19-24 ALZ can be administered by subcutaneous (SC) injections with comparable pharmacokinetics with IV administration, allowing for outpatient use.¹¹ In AA, the use of ALZ in monotherapy in the second-line setting has demonstrated an overall response rate (ORR) similar to that of r-ATG. However, its effectiveness appears to be more limited in treatment-naïve AA cases, and its use in such context remains controversial.^17,25,26 

Here, we describe a multicenter retrospective analysis of SC ALZ in association with CsA in patients with AA.

This is a collaborative retrospective study including a total of 61 patients from 2 centers in Brazil (n = 49) and 2 in the United Kingdom (n = 12). Retrospective data were collected in compliance with local institutional review board guidelines. Eligible patients were aged ≥18 years, had a diagnosis of acquired AA, and were not eligible for hematopoietic stem cell transplantation.^18,19 All patients who were offered immunosuppression did not have a suitable donor and/or were unfit for allogeneic transplant based on age and/or comorbidities.²⁷ 

AA was categorized as per Camitta criteria.²⁸ Severe AA (SAA) was defined as BM cellularity <25% (or 25% to 50% if <30% of residual cells are hematopoietic) and severe pancytopenia with at least 2 of the following criteria: absolute neutrophil count (ANC) ≤0.5 × 10⁹/L, absolute reticulocyte count ≤60 × 10⁹/L, or platelet count ≤20 × 10⁹/L.¹⁹ Very SAA (VSAA) was defined by the same criteria for SAA but with ANC ≤0.2×10⁹/L. Nonsevere AA (NSAA) was defined by BM cellularity <25% (or 25% to 50% if <30% of residual cells are hematopoietic) and peripheral blood cytopenia not fulfilling the criteria for SAA or VSAA. All patients were tested for paroxysmal nocturnal hemoglobinuria by flow cytometric assay.

ALZ administration was based on local protocols. In Brazilian centers, ALZ was used subcutaneously at doses of 3 mg, 10 mg, 30 mg, 30 mg, and 30 mg on 5 consecutive days (total dose, 103 mg)¹⁹ For patients aged >60 years, day 5 was omitted (total dose, 73 mg). In Brazilian centers, CsA was started the day after the last ALZ injection, and the dose was adjusted to be kept between 200 and 400 ng/mL during the time of treatment. In UK centers, ALZ dose varies between 100 and 150 mg, and CsA was not used in all patients, depending on previous exposure and institutional protocols. CsA was used for at least 12 months in all centers.

Premedication included antihistamines and oral paracetamol. A strict anti-infectious policy was followed for all patients, which included fluconazole, cotrimoxazole, and acyclovir. All centers performed molecular monitoring for cytomegalovirus (CMV) by polymerase chain reaction testing, and the threshold used to initiate treatment varied across centers, following institutional guidelines: in Brazilian Centers, the threshold was ≥10⁴ UI/L and in the UK centers ≥10³ UI/L. Epstein-Barr virus (EBV) monitoring (also by polymerase chain reaction testing) was only performed in the UK centers.

The response to immunosuppression was assessed at 3, 6, and 12 months after the ALZ infusion. Complete response (CR) was defined as a hemoglobin level ≥10 g/dL, an ANC ≥1 × 10⁹/L, and a platelet count ≥100 × 10⁹/L.²⁰ The criteria for a partial response (PR) were transfusion independence (both red cells and platelets), with a blood lineage recovery insufficient for CR, no longer meeting the criteria for SAA.

We also evaluated ORR, time to first response, best response, time to best response, overall survival (OS), relapse, the incidence of clinically relevant clonal evolution, and evolution to hemolytic paroxysmal nocturnal hemoglobinuria.

Median and range were used for the descriptive analysis of continuous variables, whereas categorical variables were presented as percentages. Cumulative incidence (CI) rates were calculated for response, nonrelapse mortality, and relapse. Ninety-five percent confidence intervals were estimated using the Greenwood formula. The probability of OS was calculated using the Kaplan-Meier estimation method and compared using log-rank tests. Adjusted probabilities for outcomes were estimated using the Cox proportional hazards method (OS) and Fine-Gray risk regression model (response, nonrelapse mortality, and relapse), expressed in hazard ratio (HR) and 95% confidence interval. The CI of response and relapse were calculated, considering death as a competing event. All P values were 2-sided. The analyses were performed using R 3.1.2 (Vienna, Austria) and IBM SPSS statistics version 29 (SPSS Inc, Chicago, IL).

We analyzed 64 treatments with ALZ in 61 patients. Clinical features of the patients are shown in Table 1. The median age was 52 years (range, 18-81), and 34 patients (56%) were male. Most patients (76%) had SAA (61%) or VSAA (15%), and in 34 patients (56%), a paroxysmal nocturnal hemoglobinuria (PNH) clone was detected. Among patients with a PNH clone, the median granulocyte clone size before ALZ treatment was 3.6% (range, 0.11%-99.4%), and the median monocyte clone size was 3.5% (range, 0.06%-99.2%).

At diagnosis, the median hemoglobin was 7.5 g/dL (range, 4.1-11.3), and the median reticulocyte count was 27 × 10⁹/L (range, 1.3 × 10⁹ to 100 × 10⁹/L), neutrophils 0.65 ×10⁹/L (range, 0.05 × 10⁹ to 6.78 × 10⁹/L), lymphocytes 1.6 × 10⁹/L (range, 0.23 × 10⁹ to 5.54 × 10⁹/L), and platelets 11 × 10⁹/L (range, 2 × 10⁹ to 576 × 10⁹/L). Before ALZ, the median hemoglobin was 7.3 g/dL (range, 4.3-11.3), and the median reticulocyte count was 33 × 10⁹/L (range, 1.2 × 10⁹ to 104 × 10⁹/L), neutrophils 0.53 × 10⁹/L (range, 0.05 × 10⁹ to 3.40 × 10⁹/L), lymphocytes 1.52 × 10⁹/L (range, 0.59 × 10⁹ to 3.30 × 10⁹/L), and platelets 11 × 10⁹/L (range, 1 × 10⁹ to 326 × 10⁹/L).

The most commonly used ALZ dose was 103 mg (n = 43 [67%]). Characteristics of ALZ treatments are presented in Table 2. Most treatments (n = 37 [58%]) were in the first line, in 17 cases (26%) in the second line, and in 10 cases (16%) after at least 2 lines of therapy. The median time between diagnosis and ALZ treatment was 8 months (range, 0.4-294).

We observed a hematologic response in 38 of 64 treatments (59.4%). The ORR was 41% at 3 months, 56% at 6 months, and 59.4% at 12 months (CR, 21.9%; PR, 37.5%). The median time to ALZ response was 3 months, ranging from 1 to 11 months. Only 3 patients presented responses later than 6 months after therapy (7, 10, and 11 months after ALZ).

Table 3 describes the ORR observed according to several possible risk factors. ORR was significantly higher in patients aged <30 years (77%) and in patients aged between 30 and 65 years (63%) and was lower in patients aged >65 years (31%; P = .02). There was also higher ORR in patients with hemoglobin levels before treatment >7.5 g/dL (84% vs 46% for those with lower hemoglobin levels; P = .01), and in patients with platelet counts before treatment >20 × 10⁹/L (84% vs 48% for those with lower platelet counts; P = .03). ORR was also higher in those treated with the total dose of 103 mg (70%) than those treated with different doses (38%; P = .03). However, only a few patients aged >65 years received doses >73 mg, and there could have been some confounding effect of these 2 risk factors. ORR was similar between patients with SAA (61%) and NSAA (53%), in patients with (61%) or without (57%) PNH clones, and in first line (60%) or second line (65%) vs further lines (50%). ORR was also not influenced by reticulocyte, neutrophil, or lymphocyte counts before treatment or by the duration of CsA treatment.

CI of response was 40.6% at 3 months, 54.7% at 6 months, and 59.4% at 12 months. CI of response at 12 months was significantly higher in patients aged <65 years (67%) than in older patients (31%; P = .03; Figure 1), patients with platelet counts >20 ×10⁹/L before treatment (84% vs 48% for those with lower platelet counts; P = .006), and those treated with the total dose of 103 mg (70% vs 38% for those treated with different doses; P = .02). After multivariate analyses, only platelet counts >20 ×10⁹/L remained significantly associated with a higher CI of response (HR, 0.33; 95% confidence interval, 0.15-0.71; P = .005).

At the data cutoff, among the 38 responses, 29 had maintained the response until the conclusion of this analysis. There have been 9 relapses (23%). Most patients were already off CsA therapy at the time of relapse. Three patients were re-treated with ALZ (only 1 responded), 2 received CsA monotherapy (both responded), 1 responded to danazol monotherapy, 2 patients were referred to stem cell transplantation, and 1 died from infectious complications. The median time to relapse was 25 months (range, 2-66). The median duration of response was 38 months, ranging from 2 to 157 months (13 years). Among those without relapse, the median duration of response was 41 months (range, 8-157). There are 12 patients with responses lasting >5 years.

Allogeneic hematopoietic stem cell transplantation was performed in 11 patients, 9 of whom had not responded to ALZ and 2 relapsed after a PR. Transplant was performed after a median time of 11 months (range, 6-70) after ALZ. Among patients who underwent transplant, 7 patients are still alive with sustained response and no chronic graft-verus-host disease, and 4 died from non–relapse-related causes.

Four patients had a clonal expansion to hemolytic paroxysmal nocturnal hemoglobinuria after ALZ and have been treated with complement inhibitor. One patient who had had a PR to ALZ had a clonal expansion to acute myelogenous leukemia 4 years after treatment, was refractory to treatment, and died from disease-related complications.

The median follow-up time after ALZ was 50 months (range, 8-157). OS was 86% at 1 year, 78% at 2 years, and 70% at 4 years. The causes of death were infectious complications related to persistent cytopenias (n = 13), transplant-related causes (n = 5), and clonal evolution to acute myeloid leukemia (n = 1). OS at 4 years was significantly higher in patients who responded to ALZ (90% vs 44% for those who did not respond; P < .0001; Figure 2). OS was also higher in patients with hemoglobin levels >7.5 g/dL before treatment (86% vs 41% for those with lower hemoglobin; P = .004) and in patients with platelet counts >20 ×10⁹/L before treatment (84% vs 43% for those with lower platelet counts; P = .005). After a multivariate analysis, the only risk factor that maintained a statistically significant association with poorer OS was lack of response to ALZ (HR, 9.88; 95% confidence interval, 2.12-46.0; P = .004).

All patients received ALZ as outpatients. Most patients completed the treatment without any relevant adverse event. Infusion reactions occurred in 10 treatments (16%), with self-limited fever in all cases. Nine patients (14%) presented febrile neutropenia after ALZ. CMV reactivation was observed in only 13 treatments (20%). In most cases, viral loads were relatively low and decreased spontaneously, and only in 5 cases, CMV treatment was performed and was successful. No cases of CMV disease were reported. There were 2 cases of EBV reactivation and no EBV-related diseases. Three patients with previous exposure to hepatitis B virus received prophylactic lamivudine or entecavir. Noninfectious adverse events were rare and included mild increase of liver enzymes in 1 patient. No late infectious complications possibly associated with ALZ-induced prolonged immunosuppression have been reported yet.

In this multicenter study, we retrospectively analyzed 61 patients who received ALZ alone or in association with CsA for the treatment of AA and observed an ORR of 54.7% at 6 months and 59.4% (CR, 21.9%; PR, 37.5%) at 12 months. Responses were higher in patients aged <65 years (67%) and in those treated with the total dose of 103 mg (70%). ORR was also significantly higher in those with higher hemoglobin levels and platelets counts, although it was similar significant between patients with SAA and NSAA. Moreover, ORR was also similar in the first line and further lines and in patients with or without PNH clones. OS of 4 years was significantly higher in patients who responded to ALZ (90% vs 44% for those who did not respond). Long-term responses were seen, and some patients improved their responses over time. There were no major adverse events, and infections were infrequent.

To the best of our knowledge, this was the largest series of AA cases treated with ALZ. The ORR reported here is consistent with previous experiences of ALZ.^{19,25,26,29-31} ALZ has been shown to be an effective treatment for AA in multiple case series as an alternative to treat autoimmune cytopenias, including SAA.³² In a previous analysis comparing 2 different ALZ doses, the ORR in AA was 35%, with better response rate in an escalating dose (10 mg on day 1; 20 mg on day 2; and 30 mg on day 3) with minimal toxicities.³¹ In another previous experience with 100 mg of SC ALZ plus CsA, the ORR was 57%.³⁰ 

The total dose of 103 mg (ALZ), divided over 5 consecutive days, was based on the phase 2 collaborative study of the European Group for Blood and Marrow Transplantation, which reported an ORR of 58%, very similar to the ORR reported in this study, and infectious complications were also not a major concern.^19,29,33 

In a randomized study with ALZ monotherapy, the ORR was comparable with that of r-ATG (37% and 33%, respectively), when applied as salvage therapy (after h-ATG failure).²⁵ In the most recent large randomized trial, the ORR of h-ATG plus CsA was only 41%, lower than those previously reported with the combination of h-ATG and CsA.³ 

In our series, 29 patients (48%) are maintaining their response after a median duration of response of 41 months (range, 8-157). Our results differ from a recent update from the phase 2 trial of ALZ for 42 relapsed patients with SAA, in which only 7 patients (17%) had durable long-term response at 5 years after ALZ. Furthermore, in our series, we observed only 1 case of clonal evolution, whereas in their trial, 9 cases of clonal evolution were seen, with high-risk evolution occurring in 6.²⁶ 

Our study has several limitations that need to be acknowledged. One limitation is the use of diverse treatments, with variations in doses and heterogeneity in concurrent CsA therapy. However, no statistically significant differences in terms of response or relapse rates were observed between patients who received CsA and those who did not.

Therefore, ALZ emerges as an effective and safe alternative to h-ATG for patients with AA who require immunosuppressive treatment in centers where h-ATG is not accessible. We also acknowledge that our results are particularly applicable for centers and countries with limited resources, where the current standard of care (h-ATG, CsA, and eltrombopag) is not available. Based on our results, we believe that in countries where h-ATG is not available, the combination of outpatient SC ALZ, at a full dose of 103 mg, in association with CsA, for at least 12 months, is a feasible option. More recently, it would be an adequate option to add eltrombopag to the combination, because it has become available to patients with AA in most countries and may improve hematologic response and quality of responses, as observed with h-ATG and CsA.

Besides, it should be noted that late responses (between 6 and 12 months) were observed in a few patients, and this observation may have important clinical relevance considering next-line therapies. Based on these observations, for patients without persistent severe or very severe neutropenia and/or patients not receiving heavy transfusion, we have been waiting at least 12 months before indicating further lines of therapy to allow time for late responses.

It should also be emphasized that all centers in our series are also transplant centers that perform large number of allogeneic hematopoietic cell transplantation for AA, including transplants from unrelated donors and also haploidentical donors in first line, depending on the severity of the disease. All patients who were offered immunosuppression did not have a suitable donor and/or were unfit based on age and/or comorbidities. All remaining patients were referred to an allogeneic transplant. Indeed, every effort should be made to find a donor to offer a curative approach to suitable candidates. With encouraging results using haploidentical transplants,³⁴ this option should also be considered in centers having experience with haploidentical transplants if a matched donor is not available. Haploidentical transplant remains especially important in countries with restricted resources where immunosuppressive therapies and supportive care are suboptimal.³⁵ 

Additionally, the SC administration of ALZ in outpatient setting offers notable advantages, including simplified logistics and potentially lower risks and cost savings than the conventional approach of administering r-ATG in inpatient setting over a period of at least 5 to 7 days.^19,33 Besides, r-ATG seems to be associated with lower responses than the ones described here with ALZ, although no head-to-head comparisons have been performed yet.

In conclusion, SC ALZ appears to be a feasible, effective, and safe alternative to r-ATG in patients diagnosed with AA requiring immunosuppressive treatment. This is particularly relevant in health care facilities and regions with restricted availability of h-ATG. However, it would be imperative to validate our findings through a randomized prospective trial, which would also allow for the assessment of potential advantages associated with incorporating eltrombopag into ALZ-based treatment regimens.

The authors thank all physicians, nurses and data managers that treated patients and collected the data for this study.

A.R.d.F. was supported by the Conselho Nacional de Desenvolvimento Cientfico e Tecnológico (CNPq). The study was partially funded by the Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior (CAPES).

Contribution: A.R.d.F., C.C.J., and C.A.-R. designed the research, analyzed data, and wrote the manuscript; V.C.d.M. analyzed data and wrote the manuscript; and P.E.Y., I.B.R., L.A., R.K., M.G., T.M., A.H., E.V., T.D., A.M.R., P.S., and A.K. contributed with patient data and wrote and reviewed the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Ana Rita da Fonseca, Department of Clinical and Experimental Hematology. Universidade Federal de São Paulo, R. Dr. Diogo de Faria, 824 - Vila Clementino, São Paulo, SP 04037-002, Brazil; email: anaritafonsecabm@gmail.com.