Chimeric antigen receptor T cell (CAR-T) therapy is an effective treatment for B cell malignancies. A certain fraction of patients, however, experience post-CAR-T relapse, and due to the difficulty of precise relapse prediction, biomarkers that can predict the strength and duration of CAR-T efficacy are needed before CAR-T infusion. Therefore, we performed a single-center cohort study including 91 diffuse large B cell lymphoma (DLBCL) patients treated with CAR-T in order to identify such a new prognostic biomarker. After confirming that each of the already reported prognostic parameters (disease status at leukapheresis, primary refractoriness, number of treatment lines, CD3+ cell counts at leukapheresis) has only limited predictive performance, we established a new composite parameter by integrating these four variables, and found that it predicts progression-free survival (PFS) after CAR-T infusion with statistical significance. Moreover, after comprehensive correlation analyses of this new composite parameter with all individual laboratory variables, we determined that the standard deviation of red blood cell distribution width (RDW-SD) at leukapheresis shows significant correlation with the composite parameter and may be a prognostic biomarker (R2 = 0.76, p = 0.02). Validation analysis indicated that a higher RDW-SD is significantly associated with poorer PFS after CAR-T cell therapy (HR, 3.46, P = 0.03). Thus, this study suggests that a single parameter, RDW-SD at leukapheresis, is a novel, useful biomarker that can be obtained early to predict therapeutic effects of CAR-T cell therapy. Post-CAR-T maintenance or re-induction therapies should be adopted for higher risk patients, who may relapse after CAR-T therapy.