Background: CD19-targeted chimeric antigen receptor-engineered (CD19 CAR) T cells achieve high response rates in patients (pts) with relapsed or refractory (R/R) aggressive B-cell non-Hodgkin lymphoma (NHL), but are limited by cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). Pivotal trial data suggested distinct toxicity risks across CD19 CAR T-cell products, but differences in pt and disease characteristics may have confounded these observations. Thus, we assessed the independent impact of 3 CD19 CAR T-cell products (axicabtagene ciloleucel[axicel], tisagenlecleucel [tisacel], and JCAR014) on CRS and ICANS severity in 136 pts with R/R aggressive NHL. Methods: We retrospectively analyzed aggressive NHL pts treated at our institutions with cyclophosphamide and fludarabine lymphodepletion (LD) followed by CD19 CAR T-cell therapy. Axicel and tisacel pts were treated off trial using commercial products. JCAR014 (defined-composition 4-1BB-costimulated CD19 CAR T cells) was administered in all pts at the dose of 2x10⁶/kg on a phase I/II clinical trial (NCT01865617). CRS and ICANS were graded according to the ASTCT criteria and CTCAE 4.03, respectively. We used multivariable proportional odds logistic regression to model CRS and ICANS grade. Results: The CAR T-cell product was axicel, tisacel, or JCAR014 in 50%, 28%, and 22% of pts, respectively. Compared to axicel pts, we observed higher preLD LDH levels in tisacel and JCAR014 pts, and lower preLD albumin with tisacel (p < 0.001) with comparable age and hematopoietic cell transplantation comorbidity (HCT-CI) indexes across CAR T-cell products. Higher day-28 overall response rate by Lugano criteria was observed after axicel (71%) compared to tisacel (56%) and JCAR014 (53%). Adjusting for age, HCT-CI, preLD LDH, preLD albumin, CAR T-cell product type was associated with CRS severity (tisacel versus [vs] axicel, OR = 0.45, p = 0.05; JCAR014 vs axicel, OR = 0.29, p = 0.005;). Age had limited or no impact on CRS severity (OR 95%CI, 0.97-1.02), while the effect of HCT-CI was undetermined (OR 95%CI, 0.85-1.27). In a multivariable model including the same covariates as above, CAR T-cell product type (tisacel vs axicel, OR =.14, p <.001; JCAR014 vs axicel, OR = 0.31, p = 0.009), preLD LDH (OR, 3.96 per log10 increase; p = 0.04) and age (OR per 10-year increase, 1.32; p =.06) were associated with ICANS severity. Interaction effect testing suggested effect modification of age by the CAR T-cell product type (tisacel/JCAR014 versus axicel, p = 0.06); using a multivariable model including this interaction term, the predicted probabilities of grade ≥3 ICANS in a 70 year-old after axicel, tisacel, and JCAR014 were 40%, 6%, and 8%, respectively. Conclusions: CAR T-cell product type independently impacts CRS and ICANS severity in NHL pts. Our findings provide key insights to guide patient and CAR T-cell product selection.