Long-term outcome and risk factors for relapse in pediatric warm autoimmune hemolytic anemia.

Abstract

Abstract Warm-antibody autoimmune hemolytic anemia (wAIHA) in children is a rare condition with limited data to predict outcomes or optimize therapy. Only two cohorts with more than 100 pediatric patients have been published. We took advantage of the national OBS'CEREVANCE cohort to clarify the long-term outcomes of wAIHA in children and determine whether initial clinical and biological characteristics are associated with distinct outcome patterns. We analyzed 241 children with wAIHA (excluding Evans syndrome) from the French multicenter OBS'CEREVANCE cohort (1990–2021) with a median follow-up of 4.7 years. Initial corticosteroid therapy was administered to 89% of patients. Among the 241 included patients, 42.2% (n=109) received subsequent-line therapy (SLT), with most (n=82) initiating treatment beyond 30 days (median time: 5.5 months, range 0–200 months). A total of 16.2% (n=39) received more than two lines of treatment. Rituximab was the most common second-line therapy (n=55), administered in 52.7% of cases within the first three months after diagnosis. One year post-diagnosis, 64.8% achieved complete remission without relapse (R) or SLT, including 19.9% still on corticosteroids. Meanwhile, 22.8% had SLT before corticosteroid discontinuation, and 12.4% relapsed after stopping corticosteroids. In the present study, we observed that SLTs were sometimes combined prematurely, before their optimal response window (6–8 weeks for rituximab, 3 months for immunosuppressants). While aggressive escalation may be justified in cases of severe, refractory hemolysis, each additional treatment increases infection risk, necessitates a careful risk-benefit assessment and appropriate anti-infective prophylaxis. During follow-up, 35.2% developed immunopathological manifestations (IM), including 10.3% with another autoimmune disease. A total of 15 patients (6.2%) were diagnosed with a primary immunodeficiency, with 5 diagnosed at the time of AIHA diagnosis and the remainder identified later. Identified PID cases during follow-up included ALPS, CVID, hyper IgM syndrome, and PID related to mutations in the following genes: RAG2, NFKB1, and ITK. Infectious complications were reported in 17.4% of patients, and two patients died. Among patients who received early rituximab within the first month, 17.6% relapsed or required subsequent-line therapy (R/SLT) after the expected 8-week response window. In patients treated with corticosteroids alone during the first month (77.5%), the cumulative incidence of relapse (CIR) or initiation of subsequent-line therapy (CIR-SLT) was 42.7% at 2 years and 46.2% at 5 years. Those with R/SLT were less likely to have inadequate reticulocytosis (33.7% vs 48.5%, P=0.03). Primary AIHA was more common in patients without R/SLT (72.3% vs 47.7%, P<0.001). An optimal corticosteroid duration to reduce CIR-SLT after discontinuation could not be determined. Interestingly, no patients had undergone splenectomy since 2010, reflecting a shift away from this approach in pediatric AIHA. These findings underscore the need for long-term follow-up and further study on corticosteroid duration and early SLT use. This study, the largest cohort analysis of pediatric wAIHA to date, provides valuable insights into treatment responses and long-term outcomes. Identifying underlying immune disorders, such as PID, at diagnosis, as well as associated IM, is crucial given their incidence in this population and their implications for tailored therapeutic management. While corticosteroids remain the cornerstone of treatment, nearly half of the patients required additional therapy. Future studies should focus on determining the optimal corticosteroid duration and evaluating the potential benefits of early rituximab use in childhood wAIHA, particularly in relation to associated IM and the prognostic value of BMRI at diagnosis.