Second-Line Therapy Sequencing in Primary Warm Autoimmune Hemolytic Anemia: Splenectomy-Rituximab Versus Rituximab-Splenectomy

Introduction: Primary warm autoimmune hemolytic anemia (wAIHA) is a hematologic condition that affects approximately 1 in 8000 people and occurs when IgG antibodies directed against self-erythrocytes cause hemolysis and exceed the bone marrow’s ability to compensate. First-line treatment with glucocorticoids yields sustained remission in less than 30% of individuals. For patients requiring second-line treatment, splenectomy was the preferred treatment for over 50 years, being recommended over rituximab by leading experts as recently as 2010. In the last decade this practice shifted due to concerns of tripartite risk of splenectomy with rates of infection, thrombosis, and perioperative mortality, juxtaposed against the increased experience with rituximab. These risks are outdated as they predate the modern-day vaccination, thromboprophylaxis, and laparoscopic splenectomy era, respectively. Summarily, the 2017 British guidelines dropped splenectomy from second- to third-line for patients with primary wAIHA, consistent with expert opinion in the United States in 2023. Over these same twenty years, the mortality and morbidity of splenectomy has greatly decreased, owing to updated preoperative vaccination standards, prevention of thrombosis, and widespread use of the laparoscopic (i.e., rather than open) technique. Despite these advancements, splenectomy is now used in fewer than 10% of patients with wAIHA. In the absence of a clinical trial randomizing patients to splenectomy versus rituximab, we conducted the first simulation of second-line therapeutic sequencing for adult patients with primary wAIHA with a focus on short-run and long-run effectiveness results.

Methods: We constructed a Markov simulation of adult patients with primary wAIHA and requiring second-line therapy to compare the effectiveness of strategy #1: splenectomy followed by rituximab, if necessary, versus strategy #2: rituximab followed by splenectomy, if necessary. Transition probabilities for initial response and time-variant relapse risk with rituximab and with splenectomy were drawn from retrospective primary-wAIHA-specific literature. For rituximab, we assumed that 1) patients maintained in remission for three years following treatment with a rituximab cycle would remain in remission and 2) all patients who relapse more than a year after receiving rituximab, would be retreated with rituximab. Costs were sourced from the Centers for Medicare & Medicaid Services and the Centers for Disease Control and Prevention. Utility values for patients with wAIHA were specific to the reported patient disease status, whether in relapse, partial remission (defined as hemoglobin >10g/dL) or complete remission (defined as hemoglobin >12g/dL). These were sourced from the Global Burden of Disease Study and weighted for the degree of anemia corresponding with each status. Effectiveness was calculated in quality-adjusted life-years (QALYs). The primary outcomes were lifetime discounted costs, QALYs, and net monetary benefits across accepted willingness-to-pay thresholds over short- and long-runs at 5 years and lifetime, respectively. Probabilistic sensitivity analysis was employed to vary all parameters across their uncertainty intervals simultaneously over 10,000 Monte Carlo iterations.

Results: In the short-run, strategies #1 and #2 cost $39,700 and $47,100 and accrued 5.21 and 5.26 QALYs (Table 1). In the long run, costs were $909,000 and $912,000 while accruing 22.0 and 22.1 QALYs. Net monetary benefit point estimates were similar across the accepted range of WTP thresholds in both the short- and long-runs.

Conclusion: The balance of costs and quality-adjusted life expectancy appears to be comparatively tightly balanced in the sequencing of second-line therapies in primary wAIHA with splenectomy-rituximab versus rituximab-splenectomy. The absence of a clear rituximab-early advantage on a population level may be driven by advancements in vaccination, thromboprophylaxis, and laparoscopic surgical techniques for splenectomy over the last 20 years. Value-of-information analysis would help identify parameters whose uncertainty reduction with future study in the real world would yield a treatment strategy with a definitive advantage. In the meantime, patients whose values and preferences align with splenectomy should not be dissuaded from pursuing it due to perceived risk.