Objective: Autoimmune hemolytic anemia (AIHA) is the most prevalent form of acquired hemolytic anemia, representing a heterogeneous group of anemias. As the largest retrospective study of AIHA in Asia, this study aimed to characterize the clinical features, treatment patterns, response rates, and long-term outcomes, thereby providing benchmarks for risk stratification and the evaluation of novel therapies.

Methods: This retrospective cohort study analyzed 566 AIHA patients (449 warm AIHA [wAIHA] and 117 mixed AIHA [mAIHA]) hospitalized at the Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College from January 2014 to December 2023, with a median follow-up of 57 months (range: 0.03-420). We evaluated the baseline clinical and laboratory characteristics, treatment responses, and long-term outcomes, including overall survival (OS), event-free survival (EFS, composite endpoint of relapse or death), and complications (thrombosis and infections). Survival was analyzed by Kaplan-Meier and Cox regression models.

Results: Analysis of the 566 AIHA patients (median onset age 48 years; 70% female) revealed distinct differences between subtypes. Compared to wAIHA, mAIHA patients exhibited an earlier onset (median 43 vs. 48 years), poorer erythropoiesis (bone marrow responsiveness index [BMRI] 125.3 vs. 173.9; reticulocyte production index [RPI] 2.5 vs. 3.6), a bimodal age distribution (peaks at 20/60 years) and cold AIHA-like features (including reduced RBC counts, elevated MCV/MCH/MCHC, and higher lymphocytes percentages) despite comparable initial hemoglobin (Hb) levels and hemolytic markers. Notably, secondary forms were significantly more prevalent in mAIHA (39.3% vs 22.5%), with systemic lupus erythematosus and CD5-CD10- small B-cell lymphoproliferative disorders (particularly marginal zone lymphoma and lymphoplasmacytic lymphoma/Waldenström macroglobulinemia) predominating among autoimmune and lymphoproliferative causes, respectively. Our cohort exhibited 1-, 3-, and 5-year OS rates of 96%, 92%, and 89% respectively. While disease progression accounted for 25.8% of mortality, while infections represented the leading cause of non-AIHA deaths (31.8%). Multivariable Cox regression established five independent risk factors including age ≥60 (HR=3.04), glucocorticoid (GC) non-response (HR=2.45), opportunistic infections (HR=2.13), Hb <6 g/dL (HR=1.98), and BMRI <121 (HR=1.73). Serologically, IgM- (HR=3.18) and IgM+IgA-mediated (HR=6.56) subtypes had worse OS than IgG+C3 controls in wAIHA patients. The median EFS was 17 months (1-/3-/5-year EFS: 54.5%/37.5%/29.9%), with relapses occurring primarily during treatment tapering (71.8%) or with infections (27.1%, primarily respiratory). Thrombosis occurred in 10.2% (wAIHA) and 7.7% (mAIHA), mostly during active hemolysis (84.8%/66.7%), with 43.6% of thrombosis occurred within 1-month post-diagnosis. Among opportunistic infections, fungal infections predominated (80.6%; Aspergillus, Candida, Pneumocystis). First-line GCs were used in 98.9% of patients, with response rates of 92.3% in wAIHA and 97.3% in mAIHA, demonstrating non-inferior efficacy in mAIHA. Second-line therapies (64.5%) comprised rituximab (54.5%, with comparable efficacy between standard-dose and low-dose), cyclosporine (17.5%), androgens (13.2%), cyclophosphamide (4.7%), mycophenolate mofetil (3.0%), splenectomy (2.5%) and others. Refractory AIHA was observed in 3.2% (wAIHA) and 3.5% (mAIHA) cases, demonstrating resistance to ≥3 lines of therapies including splenectomy/immunosuppressants.

Conclusion: This study revealed that mAIHA presents with an earlier onset, more secondary forms, cold agglutination-like features and poorer bone marrow compensation compared to wAIHA. Independent risk factors for mortality were identified as age ≥60 (HR=3.04), GCs resistance (HR=2.45), opportunistic infections (HR=2.13), Hb <6 g/dL (HR=1.98), and impaired erythropoiesis (BMRI <121, HR=1.73). Serologically, IgM-mediated wAIHA or with concurrent IgA involvement (IgM+IgA-mediated) showed worse OS. Notably, cold agglutinin titers showed limited predictive value for OS or treatment response. A nomogram model was developed to facilitate risk stratification. Documentation of treatment patterns and responses illustrates the therapeutic landscape of AIHA in Asian populations and establish a crucial baseline for evaluating emerging therapies.

This content is only available as a PDF.
2025
Sign in via your Institution