BACKGROUND: Diffuse large B-cell lymphoma (DLBCL) is an aggressive cancer of the lymphatic system. About 30% to 40% of people with DLBCL experience relapse and 10% are refractory to first-line treatment usually consisting of R-CHOP chemotherapy. Of those eligible for second-line treatment, commonly consisting of salvage chemotherapy followed by autologous stem-cell transplantation (ASCT), around 50% experience relapse. With a median overall survival of less than six to 12 months, the prognosis of individuals who relapse or are refractory (r/r) to advanced lines of treatment or of those who are ineligible for ASCT, is very poor. With the introduction of chimeric antigen receptor (CAR) T-cell therapy, a novel treatment option for these people is available. OBJECTIVES: To assess the benefits and harms of chimeric antigen receptor (CAR) T-cell therapy for people with relapsed or refractory (r/r) DLBCL. SEARCH METHODS: An experienced information specialist performed a systematic database search for relevant articles on CENTRAL, MEDLINE and Embase until September 11th, 2020. We also searched trial registries and reference lists of identified studies up to this date. All search results were screened by two authors independently and a third author was involved in case of discrepancies. SELECTION CRITERIA: We included prospectively planned trials evaluating CAR T-cell therapy for people with r/r DLBCL. We had planned to include randomised controlled trials (RCTs) and we flexibly adapted eligibility criteria to the most reliable study designs available. We excluded studies involving fewer than 10 participants with r/r DLBCL and studies with a proportion of participants with r/r DLBCL below 70%, unless data were reported separately for this subgroup. DATA COLLECTION AND ANALYSIS: Two review authors extracted data and performed risk of bias ratings independently. A third author was involved in case of disagreements. As our search did not yield any completed RCTs, prospective controlled non-randomised studies of interventions (NRSIs) or prospective observational studies with a control group, we did not meta-analyse data and reported all results narratively. We adopted the GRADE approach to assess the certainty of the evidence for prioritised outcomes. MAIN RESULTS: We identified 13 eligible uncontrolled studies evaluating a single or multiple arms of CAR T-cell therapies. We also identified 38 ongoing studies, including three RCTs. Ten studies are awaiting classification due to completion with no retrievable results data or insufficient data to justify inclusion. The mean number of participants enrolled, treated with CAR T-cell therapy and evaluated in the included studies were 79 (range 12 to 344; data unavailable for two studies), 61 (range 12 to 294; data unavailable for one study) and 52 (range 11 to 256), respectively. Most studies included people with r/r DLBCL among people with other haematological B-cell malignancies. Participants had received at least a median of three prior treatment lines (data unavailable for four studies), 5% to 50% had undergone ASCT (data unavailable for five studies) and, except for two studies, 3% to 18% had undergone allogenic stem-cell transplantation (data unavailable for eight studies). The overall risk of bias was high for all studies, in particular, due to incomplete follow-up and the absence of blinding. None of the included studies had a control group so that no adequate comparative effect measures could be calculated. The duration of follow-up varied substantially between studies, in particular, for harms. Our certainty in the evidence is very low for all outcomes. Overall survival was reported by eight studies (567 participants). Four studies reported survival rates at 12 months which ranged between 48% and 59%, and one study reported an overall survival rate of 50.5% at 24 months. The evidence is very uncertain about the effect of CAR T-cell therapy on overall survival. Two studies including 294 participants at baseline and 59 participants at the longest follow-up (12 months or 18 months) described improvements of quality of life measured with the EuroQol 5-Dimension 5-Level visual analogue scale (EQ-5D-5L VAS) or Function Assessment of Cancer Therapy-Lymphoma (FACT-Lym). The evidence is very uncertain about the effect of CAR T-cell therapy on quality of life. None of the studies reported treatment-related mortality. Five studies (550 participants) reported the occurrence of adverse events among participants, ranging between 99% and 100% for any grade adverse events and 68% to 98% for adverse events grade ≥ 3. In three studies (253 participants), 56% to 68% of participants experienced serious adverse events, while in one study (28 participants), no serious adverse events occurred. CAR T-cell therapy may increase the risk of adverse events and serious adverse events but the evidence is very uncertain about the exact risk. The occurrence of cytokine release syndrome (CRS) was reported in 11 studies (675 participants) under use of various grading criteria. Five studies reported between 42% and 100% of participants experiencing CRS according to criteria described in Lee 2014. CAR T-cell therapy may increase the risk of CRS but the evidence is very uncertain about the exact risk. Nine studies (575 participants) reported results on progression-free survival, disease-free survival or relapse-free survival. Twelve-month progression-free survival rates were reported by four studies and ranged between 44% and 75%. In one study, relapse-free survival remained at a rate of 64% at both 12 and 18 months. The evidence is very uncertain about the effect of CAR T-cell therapy on progression-free survival. Thirteen studies (620 participants) provided data on complete response rates. At six months, three studies reported complete response rates between 40% and 45%. The evidence is very uncertain about the effect of CAR T-cell therapy on complete response rates. AUTHORS' CONCLUSIONS: The available evidence on the benefits and harms of CAR T-cell therapy for people with r/r DLBCL is limited, mainly because of the absence of comparative clinical trials. The results we present should be regarded in light of this limitation and conclusions should be drawn very carefully. Due to the uncertainty in the current evidence, a large number of ongoing investigations and a risk of substantial and potentially life-threatening complications requiring supplementary treatment, it is critical to continue evaluating the evidence on this new therapy.