Impact of Cytokine Release Syndrome on Outcomes after T-Cell Replete Peripheral Blood Haploidentical Donor Transplantation

Background: Haploidentical hematopoietic cell transplantation (haplo-HCT) is an advancing and potentially curative therapy for patients with hematologic malignancies when the availability of HLA-matched donors are limited. The common source for haplo-HCT is donor bone marrow, but peripheral blood (PB) haplo-HCT is an emerging and less invasive option with similar survival outcomes. However, multiple series showed an increased incidence of culture negative fever early after haplo-HCT, which has been attributed to cytokine release syndrome (CRS) and more commonly observed with PB haplo-HCT. In this study, we investigated the incidence and impact of CRS on clinical outcomes and financial toxicity in PB haplo-HCT patients treated at a single center.

Methods: A total of 40 patients who underwent PB haplo-HCT in the Ohio State University between Jan 2015 - Dec 2017 were included. CRS was defined and graded based on the criteria proposed by Lee, et al (Blood, 2014). Symptoms occurring before day +14 were included, patients with a documented infection (e.g. positive culture data) and clinical suspicion of sepsis were excluded. Overall survival (OS) was defined as the time from transplantation until death. Transplant-related mortality (TRM) was defined as death before day +28 from any cause and patients who died after day +28 with no evidence of disease.

Results:. The median age at HCT was 53 years (range, 19 to 74), 63% were male. The most common diagnosis was acute myeloid leukemia (47.5%), followed by lymphoma (22.5%), acute lymphoblastic leukemia (7.5%) and myelodysplastic syndrome (7.5%). Two patients had undergone previous autologous HCT. At the time of haplo-HCT, 62% of patients were in complete remission, 38% had relapsed/refractory disease. Donor was the child in 58%, sibling in 24% and parent in 18% with an antigen match ratio of 5/10 in 50%, 6/10 in 30%, 7/10 in 12.5%, 8/10 in 7.5%. Mismatch for sex and ABO blood group were seen in 50% and 45%, respectively. Twenty percent of patients received myeloablative conditioning with fludarabine/busulfan/thiotepa, while 80% received reduced-intensity regimen with fludarabine/cyclophosphamide/total body irradiation. The incidence of CRS was 85%, including grade 1-2 CRS in 77.5% and grade 3-4 CRS in 7.5%. Time to the onset of CRS was <4 days in all cases, with 82% showing evidence of CRS within the first 48 hours. The incidence and severity of CRS did not differ by patient age, sex, remission status at the time of HCT, sex mismatch, ABO mismatch, degree of HLA matching, type of conditioning, or total nucleated cell dose. In 34 patients with CRS, most common symptoms were fever (88.2%), hypotension (15%), and worsening oxygen requirement (12%). Two patients (6%) experienced grade 2 creatinine elevation and another 2 (6%) experienced grade 1-2 transaminitis. Among 3 patients with severe CRS, 1 received tocilizumab and 1 received methylprednisolone after the onset of CRS. Patients with mild CRS (n= 31) had significantly better OS than the patients with severe CRS and no CRS (n=9) (median OS not reached, 11 months, 10.5 months, respectively, p= 0.005). Similarly, relapse-free survival (RFS) was longer in the mild CRS group than the patients with severe CRS and no CRS (p= 0.02). TRM was 20% in patients with mild CRS, when compared to 66% (2/3) in severe CRS and 33% (2/6) in patients without CRS. The duration of hospitalization was significantly longer in patients who experienced severe CRS as compared to patients with mild CRS or without CRS (40 vs 27.6 vs 27.4 days, p= 0.01). Concomitantly, inpatient cost of admission for HCT was significantly higher in patients with severe CRS (mean, $849,571) as compared to patients with mild CRS ($543,074) and without CRS ($574,632) (p= 0.04).

Conclusions: CRS is commonly observed after PB haplo-HCT and usually manifests with fever within 48 hours of cell infusion. We observed favorable survival outcomes in patients who experienced mild CRS, which may suggest an association between mild CRS and graft-versus-leukemia effect. A minority of patients developed severe CRS requiring intensive care unit support, which was associated with poor OS, high TRM, prolonged hospital stay and increased financial toxicity. Further studies investigating biomarkers that can predict the development of severe CRS might enable the incorporation of CRS prophylaxis (e.g. with tocilizumab) into standard conditioning regimens for PB haplo-HCT.