Risk of Hyponatremia and Associated Clinical Characteristics in Patients with Acute Lymphoblastic Leukemia after CD19 Targeted Chimeric Antigen Receptor (CAR) T-Cells

Introduction: CD19-specific chimeric antigen receptor (CAR) T-cells have demonstrated high response rates in patients with acute lymphoblastic leukemia (ALL) in multiple clinical trials. The most prevalent toxicities of CD19 CAR T-cells are cytokine release syndrome (CRS), associated with elevation of proinflammatory cytokines such as IL-6, and neurotoxicity. There is increasing evidence that suggests IL-6 is involved in regulation of vasopressin secretion. The combination of IL-6 elevation, vasopressin secretion, and increased water intake may result in hyponatremia, creating a challenge in fluid management, particularly in the setting of neurotoxicity. However, the prevalence and severity of hyponatremia in patients receiving CD19 CAR T-cells remain unknown.

Methods: We performed a retrospective analysis of adult patients with relapsed ALL who received CD19 CAR T-cells at Memorial Sloan Kettering Cancer Center. Hyponatremia was graded using the CTCAE v.4.0 definitions. Association of characteristics with post-infusion hyponatremia was assessed using Fisher's exact test or Wilcoxon rank sum test. Variables reviewed were grade and duration of hyponatremia, fever, CRS, neurotoxicity, C-reactive protein (CRP), IL-6, T-cell dose, and conditioning chemotherapy (fludarabine vs. cyclophosphamide).

Results: Fifty-one patients were included in our analysis, with a median age of 44 years (range, 23-74). Following treatment, ≥ grade 1 hyponatremia occurred in 21 (41%) of patients with median sodium of 133 (range, 124-140): 13 patients (25%) had grade 1 and 8 patients (16%) had grade 3. When assessed by a change in serum sodium from baseline, 29 patients (57%) experienced >7 mEq change in sodium post-infusion. No patients developed grade 4 or grade 5 hyponatremia. Hyponatremia occurred at a median of 5 days after treatment (range, 1-32) with a median duration of 3 days (range, 0-25). Hyponatremia significantly correlated with development of grade 2-4 fever (p=0.006), severe CRS (p=0.011), any grade neurotoxicity (p=0.005), peak CRP (p<0.0001), but not with peak IL-6 (0.173), conditioning chemotherapy (p=0.75), and T-cell dose (p=0.374).

Conclusion: In our cohort of adult patients with ALL treated with CD19 CAR T-cells, hyponatremia was common, mostly grades 0-1, and occurred early within the first week of treatment, correlating with the development of CAR-associated toxicities including fever, CRS, and neurotoxicity. Notably, while CRP, considered to be a reliable surrogate marker of IL-6, significantly correlated with hyponatremia, absolute magnitude of IL-6 elevation did not. Further analysis to determine the etiology and mechanism of hyponatremia will be presented, which will provide more insights to a cause-and-effect relationship between hyponatremia and CAR-associated toxicities. Because hyponatremia occurs commonly during the times of CRS, it is important to carefully monitor electrolytes, volume status, and use of IV fluids to prevent development of worsening or severe hyponatremia which can further complicate neurotoxicity in patients receiving CD19 CAR T-cells.