Clinical characteristics and outcomes of acquired hemophilia A before and after emicizumab approval in Japan

Acquired hemophilia A (AHA) is a rare disorder, with an annual incidence rate of 1.34 to 1.48 per million, caused by autoantibodies to coagulation factor VIII (FVIII). AHA may cause potentially severe bleeding.^1,2 The disease is more common in older adults, with a median age of onset in the 70s,^2-8 and nearly 90% of patients are aged >50 years.³ Older patients with AHA often have comorbidities that can prevent sufficient treatment and make them vulnerable to treatment-related side effects such as infection.⁵ Because autoantibodies to FVIII are the leading cause of AHA, immunosuppressive therapy (IST) is recommended as the primary treatment.^9,10 Prednisolone monotherapy or combined with cytotoxic agents are considered as first-line therapy,^9,10 and ∼60% to 70% of patients with AHA achieve a complete response (CR).^2,4-6 However, AHA has a high mortality rate, with hemorrhage accounting for 3% to 9% of deaths and IST-related deaths such as infection accounting for 14%.^2,4-6 

Bypassing agents (BPAs) are used to control active bleeding in ∼45% to 70% of patients with AHA.^2-7 Recombinant activated FVII and activated prothrombin complex concentrate are mainly used. The efficacy rate of BPAs exceeds 90%^7,11; however, some disadvantages exist, such as high costs and frequent administration until hemostasis is achieved. Patients must be hospitalized to receive BPAs because of their short half-lives and the need for frequent doses. Moreover, patients often need bed rest until hemostasis is achieved,³ prolonging their hospitalization and impairing their activities of daily living (ADLs).³ Therefore, preventing rebleeding that requires BPAs is essential.

Emicizumab is a recombinant, humanized, bispecific monoclonal antibody with FVIII-mimetic activity.¹² Currently, emicizumab is approved for the prevention of bleeding in patients with congenital hemophilia A, irrespective of FVIII inhibitor status. Recently, several case series have demonstrated the efficacy of emicizumab in preventing bleeding in patients with AHA.^13,14 In a prospective clinical study of emicizumab in patients with AHA conducted in Japan from 2020 to 2021 (AGEHA study), no major bleeding occurred in 11 patients who completed emicizumab treatment during the treatment.¹⁵ Furthermore, in a phase 2 study conducted in Germany and Austria from 2021 to 2022 (GTH-AHA-EMI study), with ∼4 times more patients than the AGEHA study, showed excellent bleeding prophylaxis with emicizumab, with a mean bleeding rate of 0.04 bleeds per patient-week.¹⁶ In Japan, emicizumab was approved for preventing bleeding in patients with AHA in 2022 based on the results of the AGEHA study.¹⁵ In addition to its bleeding prevention effect, the AGEHA study suggests that emicizumab may reduce the need for BPAs and transfusions, improve the Eastern Cooperative Oncology Group performance status, and enable early hospital discharge.¹⁵ In a systematic comparison of results from the GTH-AH 01/2010 study and the GTH-AHA-EMI study, it was suggested that patients treated with emicizumab experienced better bleeding protection and fewer fatal infections; furthermore, they showed improved overall survival.¹⁷ 

Owing to the rarity of the disease, real-world data on emicizumab use in patients with AHA are scarce. Here, we report a series of 19 patients with AHA from a single center, including 7 treated with emicizumab for hemostatic therapy. We treated the patients using an almost unified frontline treatment protocol, except for emicizumab, and investigated the course of treatment with and without emicizumab.

This was a single-center retrospective study using data from Kobe City Medical Center General Hospital (KCMGH) from 1 April 2010 to 31 March 2024. We reviewed 19 patients newly diagnosed with AHA who were treated at KCMGH. Clinical data were retrospectively obtained from electronic medical records. The institutional review board of KCMGH (approval number 24032) approved this study, which was performed according to the Declaration of Helsinki.

AHA was diagnosed based on decreased FVIII coagulation activity (FVIII:C) and the presence of FVIII inhibitor. FVIII:C and FVIII inhibitor titers were monitored during immunosuppressive and hemostatic therapy at least every 1 to 2 weeks until FVIII:C >50 IU/dL was achieved. In patients who were administered emicizumab, the activated partial thromboplastin time (APTT) was shortened because FVIII does not need to be activated by other coagulation factors owing to the FVIII-mimetic activity of emicizumab. FVIII:C and FVIII inhibitor titers were measured using the 1-stage coagulation-based and Bethesda assays, respectively. Because both assays are based on the APTT principle, FVIII:C and FVIII inhibitor titers were not accurately measured when emicizumab was administered. In these patients, FVIII:C and FVIII inhibitor titers were measured by central testing for correct assessment; FVIII:C was measured using a 1-stage coagulation-based assay with emicizumab in plasma samples neutralized by adding 2 anti-emicizumab idiotype monoclonal antibodies ex vivo; similarly, FVIII inhibitor titers were measured using a Bethesda assay with ex vivo emicizumab neutralization.¹⁸ 

We administered prednisolone as first-line therapy to eradicate autoantibodies to FVIII. The initial dose of prednisolone was 1 mg/kg per day for 2 to 4 weeks, and the dose was gradually tapered based on clinical findings, such as bleeding symptoms, coagulation activity, and inhibitor titer. When prednisolone was insufficient, other immunosuppressive agents such as cyclophosphamide (50-100 mg/d), rituximab (375 mg/m² per week for 4 consecutive weeks), or cyclosporine (200-250 mg/d) were added as second-line therapy following treatment guidelines for AHA.^9,10 The selection of second-line agents was based on the judgment of the attending physicians.

Severe bleeding was defined based on the European Acquired Hemophilia Registry criteria as life, limb, or organ threatening; central nervous system bleeding; hemoglobin <8 g/dL or a drop of >2 g/dL; and red blood cell (RBC) transfusion requirement of >2 units in 24 hours.⁴ Other bleeding events were defined as nonsevere bleeding. Almost all patients with severe bleeding were treated with BPAs, mainly recombinant human-activated FVII (rFVIIa; Novoseven; Novo-Nordisk) and activated prothrombin complex concentrate (factor VIII inhibitor bypassing activity [FEIBA]; Takeda Pharmaceuticals), repeatedly administered until hemostasis was achieved based on treatment guidelines for AHA. Regarding patients treated with emicizumab, we only used rFVIIa because it is not considered to promote coagulation even when combined with emicizumab.¹⁹ Furthermore, some patients were treated with invasive strategies such as endoscopic hemostasis, interventional radiology, and surgical operations.

The dose of emicizumab was fixed at 6 mg/kg on day 1 and 3 mg/kg on day 2 (loading dose), followed by 1.5 mg/kg weekly from day 8 onward (maintenance dose).¹⁵ We discontinued emicizumab based on the attending physician’s decision, referring to the criteria described in the AGEHA study: (1) FVIII activity measured in the absence of interference of emicizumab and coagulation factor products exceeded 50 IU/dL; and (2) >72 hours had passed since the last use of coagulation factor products for the last bleed requiring treatment to ensure hemostasis.¹⁵ 

We defined the hospital stay duration as the number of days from the date of admission until death, discharge, or transfer to another hospital. The day of the first IST administration was used as the reference for day 1, because IST and emicizumab were initiated on the same day in almost all patients in the emicizumab group (emi group). We used the response criteria from the GTH-AH 01/2010 study, which include partial remission (PR), defined as FVIII:C restored to >50 IU/dL and no active bleeding after stopping any hemostatic drug for >24 hours, and CR, defined as PR plus a negative inhibitor test, prednisolone tapered to <15 mg/d, and stopping any other IST.⁶ 

We described the patients’ characteristics using medians and ranges for continuous variables and numbers and percentages for categorical variables. All statistical analyses were performed using R version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria). Fisher's exact test was used to compare categorical variables between groups, and the Wilcoxon rank-sum test (Mann-Whitney U test) was used for continuous variables. P values <.05 were considered statistically significant. The Kaplan-Meier method was used to analyze the time to PR and CR and overall survival, with survival curves generated accordingly. The log-rank test was used to compare differences between groups. Missing data were handled using listwise deletion. All tests were 2-sided, and 95% confidence intervals (CIs) were calculated.

Table 1 displays the patient characteristics. Seven patients were diagnosed with AHA after emicizumab became available in Japan, and all 7 patients were treated with emicizumab. The median age was 81 years (range, 58-89) in the non-emicizumab group (non-emi group; patients 1-12) and 76 years (range, 60-87) in the emi group (patients 13-19). The non-emi and emi groups comprised 6 (50%) and 3 men (43%), respectively.

All patients had comorbidities, with lifestyle-related diseases such as hypertension, diabetes mellitus, and dyslipidemia being common. Ten (56%), 7 (37%), and 6 (32%) patients had hypertension, diabetes mellitus, and dyslipidemia, respectively. The onset of AHA led to the discovery of active cancer in 2 patients (11%) and recurrent autoimmune disease in 1 (5%).

All patients had prolonged APTT, with medians of 82.3 seconds (range, 51.1-128) and 84.8 seconds (range, 44.6-96.1) in the non-emi group and emi group, respectively. These patients had anemia, with median hemoglobin levels of 7.45 g/dL (range, 6.1-9.4) and 8.4 g/dL (range, 5.1-16.9) in the non-emi group and emi group, respectively. The median FVIII:C levels were 1.0% (range, <1%-7%) and 1.0% (range, <1%-10%), and the median FVIII inhibitor titers were 43.75 Bethesda units per mL (BU/mL) (range, 2.1-11 500) and 622 BU/mL (range, 1.0-4085) in the non-emi group and emi group, respectively. The AHA parameters of patients 3 and 5 were measured at other institutions before referral to the KCMGH, and FVIII:C of patient 3 was not measured at diagnosis. Supplemental Table 1 provides detailed patient characteristics.

Table 2 shows bleeding events and hemostatic therapy. At diagnosis, 11 (92%) and 3 patients (43%) in the non-emi group and emi group had severe bleeding, respectively. Intramuscular and subcutaneous hemorrhages were the most common bleeding symptoms. The median times from the first bleeding episode to diagnosis were 13 days (range, 2-162) in the non-emi group and 43 days (range, 7-185) in the emi group. Among the 11 patients with severe bleeding in the non-emi group, 7 (58%) were administered BPAs, and 2 (17%) required invasive procedures. Among the 3 patients with severe bleeding in the emi group, 2 (29%) received BPAs, and 1 (14%) required an invasive procedure.

After initiating IST, 8 patients (67%) in the non-emi group had bleeding symptoms, with a total bleeding of 27 times. Seven patients (58%) experienced severe bleeding, and the ratio of severe bleeding to all bleeding episodes was 93% (25 of 27 bleeding episodes). All 8 patients received BPAs. During IST, 3 patients (25%) died owing to uncontrolled bleeding. Six patients (50%) were transferred to other facilities because of the side effects of IST, which decreased their performance status. Only 3 patients (25%) were discharged.

In the emi group, all 7 patients received fixed-dose emicizumab treatment. Four patients (57%) had bleeding symptoms after IST initiation and emicizumab administration, with a total bleeding time of 6. One patient (14%) had severe bleeding, and the ratio of severe bleeding to all bleeding episodes was 17% (1 of 6 bleeding episodes). Severe bleeding occurred 9 days after emicizumab administration. Patients with severe bleeding received BPAs, whereas those with nonsevere bleeding did not receive any hemostatic treatment. The median time from admission to the start of rehabilitation was 8 days (range, 2-19). All patients were discharged, and ADLs were maintained.

The total dosage of BPAs was 521 times (43.4 doses per patient) and 49 times (7 doses per patient) in the non-emi group and emi group, respectively. The median doses of BPAs were 38 (range, 1-242) and 24.5 (range, 19-30) in the non-emi group and emi group, respectively. The median times from diagnosis to BPA initiation were 9 days (range, 0-35) in the non-emi group and 0 days (range, 0-0) in the emi group. Invasive procedures to stop active bleeding were required for 5 (42%) and 1 patients (14%) in the non-emi group and emi group, respectively. The total number of invasive procedures was 18 in the non-emi group and 2 in the emi group. Median hospital stays were 73.5 days (range, 29-202) and 44 days (range, 15-83) in the non-emi and emi groups, respectively. Eleven patients (92%) in the non-emi group and 2 in the emi group (29% of the emi group) required RBC transfusion. The total units of RBC transfusion were 320 units (26.7 units per patient) and 28 units (4 units per patient), and the median units of RBC transfusion were 22 units (range, 2-82) and 14 units (range, 10-18) in the non-emi group and emi group, respectively. Figure 1 and supplemental Figure 1 show events and responses to treatment in the first year and the total follow-up period, respectively.

Table 2 presents the IST and its outcomes. After diagnosis, all patients, except patient 3, were treated with prednisolone 1 mg/kg as a first-line induction IST. Patient 3 was initially treated with methylprednisolone pulse therapy (methylprednisolone 100 mg/day for 3 days), followed by prednisolone (1 mg/kg) as prescribed by a previous physician. As second-line therapy, 1 patient (patient 1) received cyclosporine, 2 patients (patients 5 and 8) received cyclophosphamide, and 1 (patient 14) received cyclosporine and rituximab because of refractoriness to steroid monotherapy or worsening bleeding symptoms during steroid tapering. The median time from diagnosis to IST initiation was 1.5 days (range, 0-22) in the non-emi group and 0 days (range, 0-14) in the emi group.

In the non-emi group, 8 patients achieved PR (67% of the non-emi group). Six patients (50%) achieved CR. The median times to PR and CR from IST initiation were 57 days (range, 38-90) and 98.5 days (range, 69-156), respectively. Three patients (patients 1, 5, and 7) were readmitted because of a relapse of bleeding symptoms. In the emi group, 5 patients achieved PR (71% of the emi group), and 4 (57%) achieved CR. The median times to PR and CR from IST initiation were 43 days (range, 8-85) and 118 days (range, 51-172), respectively. The 1-year cumulative incidence of PR was 67% (95% CI, 41-90) in the non-emi group and 71% (95% CI, 39-95.9) in the emi group (Figure 2A). The 1-year cumulative incidence of CR was 50% (95% CI, 26-79) in the non-emi group and 57% (95% CI, 27-90.2) in the emi group (Figure 2B).

Three patients were lost to follow-up. Of the remaining 16 patients, 8 died (42%). The 1-year overall survival was 47% (95% CI, 18%-72%) in the non-emi group and 100% (95% CI, 100%-100%) in the emi group (Figure 3). Three patients (16%) died in KCMGH owing to bleeding (patients 2 and 12 from low gastrointestinal bleeding; patient 4 from acute subdural hematoma). Two patients died after transfer to other hospitals (patient 1 from ventricular fibrillation; patient 9 from sepsis), and 3 died at home (patient 3 from acute myocardial infarction; patients 6 and 10, unknown cause of death). Seven patients (58%) in the non-emi group developed infections during IST, whereas no infections occurred in the emi group. Supplemental Table 2 shows detailed information about treatments and outcome.

This report summarizes the clinical characteristics and outcomes of 19 patients with AHA, including 7 treated with emicizumab for hemostatic therapy at KCMGH. The findings highlight the bleeding-preventive effect of emicizumab in patients with AHA. After IST initiation, bleeding occurred in 67% and 57% of patients in the non-emi group and emi group, respectively. However, severe bleeding episodes occurred in 58% and 14% of the non-emi group and emi group, respectively. The ratio of severe bleeding to all bleeding episodes was 93% in the non-emi group and 17% in the emi group. Despite differences in patient backgrounds, these results suggested that emicizumab is useful in the real world. This information provides important insights into future treatment strategies for AHA.

Emicizumab prevents bleeding in patients with congenital hemophilia A with or without FVIII inhibitor.²⁰ In addition to congenital hemophilia A,⁹ several studies have demonstrated similar efficacy for patients with AHA.^13,14 In the AGEHA study conducted in Japan, none of the 12 patients who received emicizumab had severe bleeding. Five bleeding episodes that required treatment occurred in 2 of the 12 patients. Most of them occurred around the initiation of emicizumab maintenance dose; in other words, before emicizumab concentration stabilized.¹⁵ In the GTH-AHA-EMI study, which involved >4 times as many patients as the AGEHA study (AGEHA, n = 11; GTH-AHA-EMI, n = 47), the patients received emicizumab with no IST during first 12 weeks.¹⁶ Over the course of 12 weeks of prophylaxis with emicizumab, the mean bleeding rate was 0.04 bleeds per patient-week, and the GTH-AHA-EMI study clearly demonstrated preventive effects of emicizumab on bleeding.¹⁶ The GTH-AHA-EMI study also demonstrated that plasma concentrations of emicizumab reached steady-state levels within the first week using a 1-week loading regimen,¹⁶ and a 1-week loading regimen is recommended in the guidelines.²¹ 

In our study, we administered emicizumab with a 1-week loading regimen to all 7 patients in the emi group. Bleeding occurred in 67% of patients in the non-emi group after IST initiation, comparable with the results of a previous study.²² In contrast, severe bleeding after IST and emicizumab initiation occurred in 16.7% of patients in the emi group, comparable with the rate of treated bleedings in the AGEHA study.¹⁵ Additionally, 3 patients died from bleeding in the non-emi group, whereas no bleeding-related death was recorded in the emi group. One patient experienced a single severe bleeding event 9 days after the first emicizumab administration, which could be because the plasma emicizumab concentration had not yet reached a steady state.

The AGEHA study also suggested that emicizumab reduces the need for BPAs and transfusion.¹⁵ However, data comparing the treatment outcomes with and without emicizumab are lacking. In our study, BPAs were needed in 67% and 29% of the patients in the non-emi group and emi group, respectively. The number of doses administered per patient was 43.4 and 7, respectively. RBC transfusions were required in 92% and 29% of patients in the non-emi group and emi group, respectively, with 26.7 units and 4 units of RBCs per patient. The results from our real-world data support the efficacy of emicizumab, other than hemostasis, in patients with AHA, as predicted in the AGEHA study.¹⁵ Our study suggests that emicizumab prophylaxis for patients with AHA reduces severe bleeding from multiple perspectives, including the number of BPAs and RBC transfusions. These results are consistent with the findings of previous single-arm studies and their suggestions.

All patients were treated with prednisolone (1 mg/kg) as the first-line therapy, except for patient 3, who received methylprednisolone pulse therapy at the referring hospital. Eight patients (67%) in the non-emi group and 5 patients (71%) in the emi group achieved PR, with median times to PR of 57 days (range, 38-90) and 43 days (range, 8-85), respectively. These results are comparable with those of previous reports,⁶ and the similar effect of IST in both groups suggests that the reduction in bleeding was mainly due to emicizumab. In this study, the median FVIII inhibitor titer of the emi group was as high as 622 BU/mL, which is higher than the values reported in previous reports. For example, the median inhibitor titers were 12.8 BU/mL in the European registry and 19 BU/mL in the registries of Austria and Germany.^6,23 Furthermore, 2 of the 7 patients in the emi group did not achieve PR. An FVIII inhibitor titer >20 BU/mL is considered a poor prognostic factor for IST,⁶ and the emi group in this study was at considerable risk of bleeding. Despite this, fewer bleeding episodes were observed in the emi group than in the non-emi group.

Seven of all 19 patients (37%) had infectious complications during IST, and at least 1 patient (patient 9) died from infection. Infections are a more common cause of death in patients with AHA than bleeding, and infections are considered mainly to be due to IST.⁶ In a large propensity score-matched comparison of patients with AHA treated with a standardized IST protocol and patients treated with emicizumab only for bleeding prophylaxis, patients with emicizumab prophylaxis had less bleeding (12-week bleed-free survival, 67% vs 33%), fewer fatal infections (number of patients, 0% vs 11%), and better overall survival (24-week survival rate, 90% vs 76%).¹⁷ It is suggested that the use of emicizumab for patients with AHA may reduce the intensity of IST and therefore weaken the side effects of IST. Currently, IST is recommended as a first-line treatment to eradicate FVIII inhibitors, and prednisolone at a dose of 1 mg/kg per day is suggested for corticosteroid therapy in recommendations and treatment guidelines in Japan,^9,10 and we followed these recommendations herein. However, these guidelines were published before emicizumab approval. New IST strategies combined with emicizumab need to be discussed in the future.

All patients in the emi group were discharged with maintained ADLs, whereas only 33% of those in the non-emi group were discharged. In the non-emi group, 6 patients (50%) were transferred to another institution. This result might be due to the shortening of bed rest due to reduced bleeding, earlier rehabilitation, and shorter hospital stays resulting from the hemorrhage control effect of emicizumab. The median lengths of hospital stay were 44 days and 73.5 days in the emi group and non-emi group, respectively. These comparisons support the efficacy of emicizumab, other than hemostasis, in patients with AHA, as predicted in the AGEHA study.¹⁵ 

The median doses of BPAs to stop active bleeding were 38 doses (range, 1-242) in the non-emi group and 24.5 doses (range, 19-30) in the emi group. Treatment with emicizumab can result in a reduced use of BPAs and shortened hospital stays, ultimately leading to cost savings.

In the AGEHA study, 3 treatment-related diseases were recorded, including thrombocytopenia, increased prothrombin fragment 1 + 2, and deep vein thrombosis.¹⁵ In our study, emicizumab was well tolerated without side effects, including thromboembolism, even when used with rFVIIa. After leaving the hospital, all patients in the emi group continued to receive emicizumab in an outpatient setting without adverse events.

Our study has some limitations. This retrospective study, conducted in a single center with a limited number of patients to compare the treatment course between the non-emi and emi groups, has potential biases in patient backgrounds. Severe bleeding at diagnosis was higher in the non-emi group (92%) than the emi group (43%), possibly overestimating the bleeding-preventive effect of emicizumab in the emi group. Moreover, this difference could affect the overall survival. The short follow-up period for the emi group necessitates long-term follow-up to evaluate the long-term effects and disadvantages of emicizumab. However, the results of this study are important for assessing the benefit-risk profile of emicizumab, because it would not be feasible to conduct a comparative study on the efficacy of emicizumab owing to the rarity and severity of AHA.

Our report investigates the treatment course for patients with AHA before and after the introduction of emicizumab, demonstrating its effectiveness in preventing bleeding. Furthermore, emicizumab may lead to shorter hospital stays, maintained ADLs, reduced costs, and improved prognosis in patients with AHA. However, real-world data from more patients and long-term follow-up are needed to confirm these findings.

The authors thank the medical and nursing staff at Kobe City Medical Center General Hospital.

Contribution: D.K. contributed to the design of the study, data analysis, interpretation of the results, and writing the manuscript; M.N., Y.S., T.I., and T.K. critically reviewed the manuscript; and all authors approved the final version of manuscript.

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

Correspondence: Yoshimitsu Shimomura, Department of Hematology, Kobe City Hospital Organization Kobe City Medical Center General Hospital, Minamimachi 2-1-1, Minatojima, Chuo-ku, Kobe 650-0047, Japan; email: shimomura_0119@yahoo.co.jp.