BACKGROUND: Vasovagal reactions are the most common acute systemic complications of whole blood donations and are associated with significant morbidity and medicolegal costs. Blood services worldwide have implemented various strategies to prevent vasovagal reactions, often without robust evidence. We compared standard practice in England with four vasovagal reaction prevention interventions used in other services or previously tested in small-scale studies. METHODS: This cluster-randomised, stepped-wedge, crossover trial involved 73 blood donation sites (clusters) across England, UK. Each site was randomly allocated to one of 64 possible sequences for implementing one or more of four interventions over four 9-month periods, yielding a total of 292 site-periods for a 36-month overall trial duration. Participants within each period were not unique, as repeat donations were included. Cluster randomisation was chosen for operational feasibility. Potential carry-over effects were minimised by use of stepped-wedge, crossover, and factorial design elements. The interventions assessed were: 500 mL pre-donation isotonic drink (vs standard 500 mL plain water), extended 3-minute post-donation rest on the donation chair (vs standard 2-minute rest), modified applied muscle tension (AMT) exercise (vs current AMT practice); and psychosocial intervention using preparatory materials (vs no materials). The primary outcome was in-session vasovagal reaction with loss of consciousness. Analysis was by intention-to-treat, including data from all donors attending the donation sites during the trial, unless the donor opted-out. This trial registered with isrctn.com (ISRCTN 10412338) and is complete. FINDINGS: Between Nov 4, 2019, and Nov 3, 2022, we recruited 1 379 095 blood donors who made 4 134 712 blood donations. The median age of donors at time of baseline trial visit was 41 years (IQR 30-54), 785 271 (56·9%) of 1 379 095 donors were female and 593 824 (43·1%) were male. 461 954 (33·5%) were first-time donors. We recorded 4388 vasovagal reactions with loss of consciousness and 60 517 total in-session vasovagal reactions (ie, with and without loss of consciousness). Compared with standard practices, none of the interventions clearly reduced the primary outcome (joint p=0·21) with an odds ratio of 0·98 (95% CI 0·92-1·04) for isotonic drink, 0·99 (0·92-1·06) for extended chair rest, 1·12 (1·00-1·26) for modified AMT, and 1·03 (0·93-1·14) for psychosocial intervention. Absolute event rates per 10 000 donations were 10·7 without intervention and 10·5 with intervention. Findings were similar for total in-session vasovagal reactions with higher absolute event rates. INTERPRETATION: Four interventions used to prevent donation-related vasovagal reactions showed no clear benefits compared with standard practices in England, suggesting potential policy implications for blood services worldwide to streamline donation practices and save resources. FUNDING: NHS Blood and Transplant and the NIHR Blood and Transplant Research Unit in Donor Health and Behaviour (formerly Donor Health and Genomics).