Vincristine plus dexamethasone pulses are generally used throughout maintenance treatment for childhood acute lymphoblastic leukaemia. However, previous studies remain inconclusive about the benefit of this maintenance therapy and the absence of randomised, controlled trials in patients with low-risk or high-risk acute lymphoblastic leukaemia provides uncertainty. We therefore aimed to determine if this therapy could be safely omitted beyond 1 year of treatment without leading to an inferior outcome in any risk subgroup of childhood acute lymphoblastic leukaemia.This open-label, multicentre, randomised, phase 3, non-inferiority trial involved 20 major medical centres across China. We enrolled patients who were aged 0-18 years with newly diagnosed acute lymphoblastic leukaemia that was subsequently in continuous remission for 1 year after initial treatment. Patients with secondary malignancy or primary immunodeficiency were excluded. Eligible patients were classified as having low-risk, intermediate-risk, or high-risk acute lymphoblastic leukaemia based on minimal residual disease and immunophenotypic and genetic features of leukaemic cells. Randomisation and analyses were done separately for the low-risk and intermediate-to-high-risk cohorts. Randomisation was generated by the study biostatistician with a block size of six. Stratification factors included participating centre, sex, and age at diagnosis; the low-risk cohort was additionally stratified for ETV6-RUNX1 status, and the intermediate-to-high-risk cohort for cell lineage. Patients in each risk cohort were randomly assigned (1:1) to either receive (ie, the control group) or not receive (ie, the experimental group) seven pulses of intravenous vincristine (1·5 mg/m2) plus oral dexamethasone (6 mg/m2 per day for 7 days) during the second year of treatment. The primary endpoint was difference in 5-year event-free survival between the experimental group and the control group for both the low-risk and intermediate-to-high-risk cohorts, with a non-inferiority margin of 0·05 (5%). The analysis was by intention to treat. This trial is registered with the Chinese Clinical Trial Registry, ChiCTR-IPR-14005706.Between Jan 1, 2015, and Feb 20, 2020, 6141 paediatric patients with newly diagnosed acute lymphoblastic leukaemia were registered to this study. Approximately 1 year after diagnosis and treatment, 5054 patients in continuous remission were randomly assigned, including 2923 (1442 in the control group and 1481 in the experimental group) with low-risk acute lymphoblastic leukaemia and 2131 (1071 control, 1060 experimental) with intermediate-to-high risk acute lymphoblastic leukaemia. Median follow-up for patients who were alive at the time of analysis was 3·7 years (IQR 2·8-4·7). Among patients with low-risk acute lymphoblastic leukaemia, no difference was observed in 5-year event-free survival between the control group and the experimental group (90·3% [95% CI 88·4-92·2] vs 90·2% [88·2-92·2]; p=0·90). The one-sided 95% upper confidence bound for the difference in 5-year event-free survival probability was 0·024, establishing non-inferiority. Among patients with intermediate-to-high-risk acute lymphoblastic leukaemia, no difference was observed in 5-year event-free survival between the control group and the experimental group (82·8% [95% CI 80·0-85·7] vs 80·8% [77·7-84·0]; p=0·90), but the one-sided 95% upper confidence bound for the difference in 5-year event-free survival probability was 0·055, giving a borderline inferior result for those in the experimental group. In the low-risk cohort, we found no differences in the rates of infections, symptomatic osteonecrosis, or other complications during the second year of maintenance treatment between patients in the control and experimental groups. Patients with intermediate-to-high-risk acute lymphoblastic leukaemia in the control group were more likely to develop grade 3-4 pneumonia (26 [2·4%] of 1071 vs ten [0·9%] of 1060) and vincristine-related peripheral neuropathy (17 [1·6%] vs six [0·6%]) compared with the experimental group. Incidence of grade 5 fatal infection was similar between the control group and the experimental group in both the low-risk cohort (two [0·1%] of 1442 vs five [0·3%] of 1481) and intermediate-to-high risk cohort (six [0·6%] of 1071 vs five [0·5%] of 1060).Vincristine plus dexamethasone pulses might be omitted beyond 1 year of treatment for children with low-risk acute lymphoblastic leukaemia. Additional studies are needed for intermediate-to-high-risk acute lymphoblastic leukaemia.VIVA China Children's Cancer Foundation, the National Natural Science Foundation of China, the China fourth round of Three-Year Public Health Action Plan (2015-2017), Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences, US National Cancer Institute, St Baldrick's Foundation, and the American Lebanese Syrian Associated Charities.For the Chinese translation of the abstract see Supplementary Materials section.Copyright © 2021 Elsevier Ltd. All rights reserved.