Safety of Multiple Doses of CAR T Cells

Chimeric antigen receptor (CAR) T cells have demonstrated promising results for cancers in pre-clinical models and early phase trials. However, some groups have observed serious adverse events (SAEs) and toxicities attributable to the administration of CAR T cells, the most clinically significant being cytokine release syndrome (CRS). In addition, a recent case report described a patient who developed an anaphylaxis reaction to CAR T cells, ultimately attributed to receipt of multiple doses of cells. As there are a number of ongoing clinical trials at our institution utilizing CAR T cells to treat various hematologic and solid malignancies and several patients have received multiple doses, we performed a retrospective review to assess whether early or late infusion toxicities were observed with subsequent infusions. We identified 47 of over 200 patients who received more than one dose of CAR T cells between January 2009 and December 2014. We assessed patient characteristics including type of malignancy, disease status at the time of subsequent infusions, cell product (autologous vs. allogeneic), presence or absence of prior lymphodepletion, number of infusions, dosing schedule, and the relationship to development of CRS. Each characteristic was examined for any correlation to developing a SAE. There were no early adverse events within 24 hours of infusion aside from one patient who developed fever, without other signs suggestive of CRS. The majority of non-hematologic grade 3-4 AEs were electrolyte disturbances and elevated liver function tests. Only two grade 3-4 AE's were deemed possibly attributed to CAR T cells (one of which was pain at a site of bony disease in a patient with neuroblastoma). 13/47 patients experienced SAEs, with 11 occurring after the 2^nd infusion and 2 after the 3^rd infusion. Cell doses ranged from 1 x 10⁶ - 2 x 10⁸ cells/m², with 8 patients receiving an identical dose as first infusion, 2 a lower dose, and 3 a higher dose. The median time of SAE occurrence in relation to the infusion was 5 weeks, with the earliest occurring 16 days following the second infusion. We further categorized the SAEs into the following categories: thromboembolic, pulmonary, fever/systemic inflammatory response, and electrolyte disturbance. A single SAE was deemed possibly related to a CAR T cell-induced hyperinflammatory response, occurring 16 days after the second infusion. Of note, we did not observe events consistent with anaphylaxis or suggestive of acquired immune response, and we found no evidence of human anti-mouse IgG antibody formation. We also found no relation between cell dose, cell type, lymphodepletion status, or dosing schedule. We can therefore conclude that repeated CAR T cell infusions are well tolerated, and that the majority of grade 3 and 4 adverse events are hematologic and electrolyte abnormalities that resolve without intervention. Furthermore, the majority of SAEs reported after multiple CAR T cell infusions were unrelated to the infusion itself. However, further evaluation of a larger cohort is necessary to determine whether an association between the timing of repeated infusions and CRS and other SAEs exists. These findings and our continued evaluation of patients receiving multiple infusions will help us to ensure the safety of administering multiple doses of CAR T cells in the future.