Ibutinib-Associated Autoimmune Hemolytic Anemia in CLL

Abstract Autoimmune hemolytic anemia (AIHA) occurs in CLL at some time during the course of the disease in up to 7-10% of patients. The acute onset of AIHA may occur unrelated to therapy but has also been linked to treatment with chemo­therapeutic agents including chlor­ambucil, benda­mustine and particu­larly purine nucleosides such as fludarabine. Although the mechanism is still not well understood, chemo­therapy-induced changes in regulatory T-cells have been proposed as a trigger for autoimmunity and clinical hemolysis. In contrast to these cytotoxic therapies, ibrutinib, an inhibitor of Bruton’s tyrosine kinase recently approved for the treatment of CLL, appears to have a different mechanism of action and thus far has not been associated with AIHA in published reports. However, we report here a patient with CLL and a history of prior AIHA, who developed a recurrence of acute hemolysis after the initiation of ibrutinib. The patient is a 67-year-old man diagnosed with CLL in 2002 and treated for progressive disease with a single cycle of bendamustine in 2009. Although the lymphocytosis resolved rapidly, the hemoglobin also decreased from 14 g/dL to 5.2 g/dL by 3 weeks after the start of therapy. Due to the onset of Coombs-positive AIHA, chemotherapy was dis­con­tinued. Hemolysis resolved with prednisone therapy and did not recur after a slow taper. The CLL then remained asymptomatic until 2012 when night sweats developed at a white blood cell (wbc) count of 95,000/µL. Benda­mustine was re-started and despite a negative Coombs test prior to treatment, Coombs-positive AIHA developed again with the hemoglobin falling from normal to 7.0 g/dL within 4 weeks. After stabilization with transfusions and steroids, an additional cycle of bendamustine plus rituximab was administered without further complications and the patient’s symptoms and lympho­cytosis resolved. After the discontinuation of prednisone, hemolysis did not recur clinically although the Coombs test remained 1+ positive through early 2014. By May 2014 the wbc count had increased to 144,000/µL with the onset of a mild anemia (Hgb 12.3 g/dL) and symptomatic night sweats. Due to the history of repeated chem­otherapy-associated AIHA, alter­native therapy with ibrutinib, which had not been associated with AIHA, was instituted at 420 mg daily. However, within 2 weeks the hemoglobin decreased to 7.0 g/dL while the wbc count increased to 300,000/µL. A reticulocyte count was 16%, total bilirubin 3.2 mg/dL, haptoglobin <10 mg/dL, and the Coombs test was reported 3+ positive for IgG. Prednisone was started at 1 mg/Kg daily and ibrutinib was held. After 10 days of steroid therapy, the hemoglobin improved to 10 g/dL and ibrutinib was restarted. Over the next month, the hemoglobin continued to increase to 12.7 g/dL and the wbc count gradually decreased to 44,000/µL with 24% neutrophils, a marked increase in the absolute neutrophil count as compared to pre-treatment. The patient currently continues on ibrutinib with a slow prednisone taper and is now asymptomatic. In summary, this case suggests that some CLL patients may experience the acute onset of AIHA after starting ibrutinib with a clinical course similar to cases reported after fludarabine and other agents. In addition, a response to steroid therapy may allow the successful resumption of ibrutinib treatment. The actual likelihood of acute hemolysis after ibrutinib therapy in CLL is unclear, but may have been much greater in this case than in CLL patients without prior episodes of AIHA or a positive Coombs test. However, a potential mechanism for the precipitation or accel­eration of auto­immunity is suggested by the recent report by Byrd and colleagues (Blood 2013; 122:2539-49) showing that ibrutinib inhibits interleukin-2-inducible kinase (ITK). This inhibition skewed the CD4 T-cell popul­a­tions isolated from treated CLL patients toward a Th1 profile, a regulatory T-cell subset previously reported to drive the early auto­antibody response in an animal model of AIHA (Blood 2009; 113:389-95). Disclosures No relevant conflicts of interest to declare.