CD19-targeted chimeric antigen receptor-engineered (CD19 CAR) T-cell therapy has shown significant efficacy for relapsed or refractory (R/R) B-cell malignancies. Yet, CD19 CAR T cells fail to induce durable responses in most patients. Second infusions of CD19 CAR T cells (CART2) have been considered as a possible approach to improve outcomes. We analyzed data from 44 patients with R/R B-cell malignancies (acute lymphoblastic leukemia [ALL], n = 14; chronic lymphocytic leukemia [CLL], n = 9; non-Hodgkin lymphoma [NHL], n = 21) who received CART2 on a phase 1/2 trial (NCT01865617) at our institution. Despite a CART2 dose increase in 82% of patients, we observed a low incidence of severe toxicity after CART2 (grade ≥3 cytokine release syndrome, 9%; grade ≥3 neurotoxicity, 11%). After CART2, complete response (CR) was achieved in 22% of CLL, 19% of NHL, and 21% of ALL patients. The median durations of response after CART2 in CLL, NHL, and ALL patients were 33, 6, and 4 months, respectively. Addition of fludarabine to cyclophosphamide-based lymphodepletion before the first CAR T-cell infusion (CART1) and an increase in the CART2 dose compared with CART1 were independently associated with higher overall response rates and longer progression-free survival after CART2. We observed durable CAR T-cell persistence after CART2 in patients who received cyclophosphamide and fludarabine (Cy-Flu) lymphodepletion before CART1 and a higher CART2 compared with CART1 cell dose. The identification of 2 modifiable pretreatment factors independently associated with better outcomes after CART2 suggests strategies to improve in vivo CAR T-cell kinetics and responses after repeat CAR T-cell infusions, and has implications for the design of trials of novel CAR T-cell products after failure of prior CAR T-cell immunotherapies.