Autoimmune hemolytic anemia (AIHA) is a complex disorder and autoantibodies optimally fight against red blood cells at 37°C in warm AIHA (wAIHA). Rituximab is the preferred second-line treatment after failure of glucocorticoids. The treatment of relapsed and refractory (RR) wAIHA is challenging, especially in emergencies with severe anemia and crossmatching incompatibility.

Bortezomib induces the apoptosis of plasma cells that produce autoantibodies and have multiple immunoregulatory functions. The beneficial effect of bortezomib in refractory wAIHA has been described in several case reports, with scant adverse effects [1]. A combination of rituximab and bortezomib might simultaneously clear autoreactive B lymphocytes and plasma cells to achieve a higher and faster response. We previously reported retrospectively data of seven patients with RR wAIHA who were administered rituximab combined with bortezomib (RB regimen) in 2020 [2]. Those patients quickly responded to the treatment, with mild adverse effects. Therefore, this investigator-initiated prospective single-arm trial investigated the effectiveness and safety of combination of rituximab and bortezomib in patients with RR wAIHA. The study was approved by the Ethics Committee of the Peking Union Medical College Hospital (PUMCH) (project number: HS-2053, approved August 2019). The trial is registered at ClinicalTrials.gov (NCT04083014).

We defined wAIHA as anemia with a positive direct antiglobulin test (DAT) for IgG, with or without complement, and signs of hemolysis. The patients were enrolled from PUMCH based on the following eligibility criteria: age ≥ 18 years; diagnosed with primary or secondary wAIHA without treatment indication of systems other than the hematologic system; relapse after glucocorticoid therapy was reduced or discontinued, or glucocorticoid dependence (prednisone maintenance dose > 10 mg/day), or no response after 4 weeks of glucocorticoid therapy, hemoglobin (HGB) ≤ 100 g/L with signs of hemolysis including elevated bilirubin, reticulocyte count, and lactic dehydrogenase (LDH), and normal cardiac, liver, and kidney functions. Patients with wAIHA with uncontrollable systemic infection or other serious diseases; active hepatitis B virus (HBV) were excluded. Patients with AIHA showing only C3d positive on DAT and those with cold agglutinin disease (CAD) who had positive cold agglutinin test were excluded.

Treatment for patients included two-course RB regimen and maintenance. The RB regimen comprised a single dose of 500 mg rituximab infusion (Hanlikang; Shanghai Henlius Biotech Inc., Shanghai, China) on Day 0 and subcutaneous injections of 1.3 mg/m² bortezomib (Jiangsu Hansoh Pharmaceutical Group Co. Ltd., Jiangsu, China) twice weekly for 2 weeks on Days 1, 4, 8, and 11. The same course was repeated after 3 months. Six cycles of rituximab (a single dose of 500 mg) were administered to responders every 3 months as maintenance (Figure S1). If patients were already medicated with glucocorticoids, the dose was rapidly reduced to the minimum maintenance dose (dose reduced by 50% every 3 days to a maintenance dose of 5–15 mg daily). Immunosuppressants other than glucocorticoids were not administered. Transfusion and intravenous immunoglobulin (IVIg) could be allowed as salvage treatment for patients whose HGB < 60 g/L.

The primary endpoint was the ORR after 6 months. The secondary endpoints comprised the interval (days) required to achieve a response (HGB increase of > 20 g/L and transfusion independence) and safety. The response last for more than 3 months was considered to be valid. A total of 31 patients were enrolled between September 2, 2019 and March 1, 2023. One patient was excluded because of withdrawal of informed consent before completing one course of RB regimen, and thus 30 patients were included in the analysis. Patients' characteristics are shown in Table 1. Among the 30 patients, 26 (86.7%) completed the planed two-course RB regimen and four (13.3%) received only one. 14/26 (53.8%) patients received regular rituximab maintenance therapy every 3 months (range 2–6 courses). Figure S1 shows a flowchart of the patients.

After the first course of RB treatment, response (increase in HGB of > 20 g/L and transfusion independence) was achieved in 26 patients, with a median of 15 (IQR 8.0–21) days, and 24 patients showed normalized HGB > 120 and > 110 g/L in males and females at a median of 41 days (IQR 21.0–64). The ORR at 6 months was 80.0% (24/30), with a CR rate of 56.7% (17/30) and PR rate of 23.3% (7/30), respectively (Figure 1). For the whole cohort, HGB level significantly increased and indicators of hemolysis, namely RET%, ARC, serum total bilirubin (Tbil), serum direct bilirubin (Dbil), and LDH decreased during the treatment (Figure S2).

Twenty-six (86.7%) patients were enrolled with concomitant corticosteroids, and the median prednisone-equivalent dose was 50 mg/day (range 4–100). At 6 months, seven patients (23.3%) were not combined with corticosteroids, and 23 patients (76.7%) were combined with corticosteroids, with a median dose of 5 mg/day (range 3–30 mg/day). Thirteen (43.3%) patients require transfusions in the 4 weeks prior to entry. Within 4 weeks of treatment, only 2/30 (6.7%) patients required blood transfusion support.

With a median follow-up of 22 months (range 5–56), the OR rate was 80.0% with a CR rate of 63.3%; four (13.3%) patients were lost to follow-up, and three (10.0%) patients died of severe pneumonia after 5, 6, and 10 months of treatment, respectively. The median OS was not reached. The predictive 24-month OS rate was 89.9% (Figure 1). Five (20.8%) of 24 responders relapsed at a median of 10 months (range 7–24). The median RFS was not reached. The predictive 24-month RFS rate was 76.0% (Figure 1). Fourteen patients switched to the next treatment. The time to next treatment (TTNT) was 7 months (range 1–25). The salvage therapies are listed in Table S1.

Factors that may influence the ORR and relapse were analyzed, including age, sex, interval between diagnosis and RB treatment, primary/secondary AIHA, disease status at enrollment (refractory vs. relapsed), lines of previous treatment (1 vs. ≥ 2), previous exposure to rituximab. Multivariate analysis verified that male patients were more likely to relapse than female patients (OR 54.102, 95% CI 2.214–1321.996, p = 0.014) (Table S2). Log-rank analysis revealed that female patients had a significantly better RFS than male patients (p < 0.001, Figure 1).

AEs were reported in 18 (60.0%) patients. The most prevalent AEs were lung infection (33.3%), fever (10.0%), thrombocytopenia (10.0%), nausea (6.7%), and herpes zoster (6.7%). Seven (23.3%) patients experienced grade ≥ 3 AEs (Table S3).

The ratios of B lymphocytes (p = 0.013), NK cells (p = 0.009) were significantly lower, whereas the ratios of CD3⁺CD8⁺ T cells (p = 0.029), CD8⁺CD38⁺ T cells (p < 0.001), and CD8⁺HLA-DR⁺ T cells (p < 0.001) were significantly higher in RR wAIHA patients before the RB regimen than in 799 healthy adults (median age, 40 [37–48] years). After 6 months of treatment, compared with those of the baseline, the ratios and counts of B lymphocytes significantly decreased, but the ratios of NK cells, T cells, and CD3⁺CD8⁺ significantly increased (p < 0.05). The median count of Treg (CD25⁺FOXP3⁺) cells was 8.72 cells/μL (IQR 3.06–27.5, reference range 28–142) before RB treatment and did not increase significantly 3 months after treatment (p > 0.1). The level of serum immunoglobulin G (p = 0.048) and IgA (p = 0.037) decreased significantly, whereas level of serum IgM (p = 0.377) did not present descending trend (Table S4).

This prospective phase 2 study showed that the combination of rituximab and bortezomib (RB regimen) was effective, with low recurrence rate in wAIHA patients. The 6-month ORR (80.0%) and CRR (56.7%) were comparable with those of the current recommended second-line rituximab monotherapy (ORR 70%–80% and CRR 25%–75%) [3-5]. An initial response (HGB increase of > 20 g/L and transfusion independence) was achieved in 26 (86.7%) patients, with a median of 15 (IQR 8–21) days after the first dose of rituximab in the RB treatment. The response is faster than rituximab monotherapy which takes 3–6 weeks to effect.

Yao et al. [6] reported that in seven patients with wAIHA treated with four cycles of LowR-BD regimen (rituximab 100 mg on Day 1 combined with bortezomib 1.3 mg/m² on Day 2), an ORR of 85.71% and CR of 28.57% were obtained. However, concomitant glucocorticoids, including dexamethasone 20 mg on Days 2 and 3 in the regimen, 5 mg dexamethasone + 40 mg methylprednisolone on Day 1 to prevent allergic reactions of rituximab, and 20–40 mg/day prednisone after LowR-BD regimen, interfered with response evaluation.

Factors predicting response to rituximab were analyzed in several studies. In the GIMEMA study [3], an increased response rate was associated with young age and a shorter interval between diagnosis and rituximab treatment in patients administered low-dose rituximab. In a meta-analysis of 21 studies of AIHA [4], younger age and wAIHA were confirmed as predictive factors for response following rituximab therapy. In another meta-analysis [5], including 37 investigations encompassing 1057 AIHA patients treated by rituximab, ORR and CRR were only significantly associated with age. In our study, relapsed patients had higher ORR than refractory patients, and male patients were more likely to relapse than female patients. No difference in response and relapse could be identified with respect to age, primary/secondary AIHA.

Assessment of immune function showed that the percentages of B cells were decreased significantly in RR wAIHA patients compared with those in healthy adults, and B cells were undetectable in all patients after treatment, with no difference between responders and non-responders. Though B cells play a key role in the production of autoantibodies, B cells at baseline had no correlation with response to RB treatment, and patients without detectable B cells could also relapse. CD8⁺ T cells were activated and Treg cells decreased in RR wAIHA patients and showed no significant change after rituximab and bortezomib treatment. Anti-B-cell therapeutic strategies for RR AIHA may be switched to anti-T-cell strategies when treatment failure occurs.

In conclusion, for RR wAHIA patients, combination treatment with rituximab and bortezomib demonstrates a quicker response and similar ORR, CRR, and safety to rituximab monotherapy, and reduces their need for corticosteroids. Rituximab maintenance therapy may improve long-term response.