Design Investigator-led, prospective, open-labelled marker-based strategy (hybrid) randomised trial. Background Allografts in 3% of kidney transplant patients fail annually. Development of antibodies against human leucocyte antigens is a validated predictive biomarker of allograft failure. Under immunosuppression is recognised to contribute, but whether increasing immunosuppression can prevent allograft failure in human leucocyte antigen Ab+ patients is unclear. Participants Renal transplant recipients > 1 year post-transplantation attending 13 United Kingdom transplant clinics, without specific exclusion criteria. Interventions Regular screening for human leucocyte antigen antibodies followed, in positive patients by interview and tailored optimisation of immunosuppression to tacrolimus, mycophenolate mofetil and prednisolone. Objective To determine if optimisation of immunosuppression in human leucocyte antigen Ab+ patients can cost-effectively prevent kidney allograft failure. Outcome Time to graft failure after 43 months follow-up in patients receiving the intervention, compared to controls, managed by standard of care. Costs and quality-adjusted life-years were used in the cost-effectiveness analysis. Randomisation and blinding Random allocation (1 : 1) to unblinded biomarker-led care or double-blinded standard of care stratified by human leucocyte antigen antibodies status (positive/negative) and in positives, presence of donor-specific antibodies (human leucocyte antigen antibodies against donor human leucocyte antigen) or not (human leucocyte antigen antibodies against non-donor human leucocyte antigen), baseline immunosuppression and transplant centre. Biomaker-led care human leucocyte antigen Ab+ patients received intervention. Human leucocyte antigen Ab-negative patients were screened every 8 months. Recruitment Began September 2013 and for 37 months. The primary endpoint, scheduled for June 2020, was moved to March 2020 because of COVID-19. Numbers randomised From 5519 screened, 2037 were randomised (1028 biomaker-led care, 1009 to standard of care) including 198 with human leucocyte antigen antibodies against donor human leucocyte antigen (106 biomaker-led care, 92 standard of care) and 818 with human leucocyte antigens antibodies against non-donor human leucocyte antigen (427 biomaker-led care, 391 standard of care). Numbers analysed Two patients were randomised in error so 2035 were included in the intention-to-treat analysis. Outcome The trial had 80% power to detect a hazard ratio of 0.49 in biomarker-led care DSA+ group, > 90% power to detect hazard ratio of 0.35 in biomarker-led care non-DSA+ group (with 5% type 1 error). Actual hazard ratios for graft failure in these biomarker-led care groups were 1.54 (95% CI: 0.72 to 3.30) and 0.97 (0.54 to 1.74), respectively. There was 90% power to demonstrate non-inferiority of overall biomarker-led care group with assumed hazard ratio of 1.4: This was not demonstrated as the upper confidence limit for graft failure exceeded 1.4: (1.02, 95% CI 0.72 to 1.44). The hazard ratio for biopsy-proven rejection in the overall biomarker-led care group was 0.5 [95% CI: 0.27 to 0.94: p = 0.03]. The screening approach was not cost-effective in terms of cost per quality-adjusted life-year. Harms No significant differences in other secondary endpoints or adverse events. Limitations Tailored interventions meant optimisation was not possible in some patients. We did not study pathology on protocol transplant biopsies in DSA+ patients. Conclusions No evidence that optimised immunosuppression in human leucocyte antigen Ab+ patients delays renal transplant failure. Informing patients of their human leucocyte antigen antibodies status appears to reduce graft rejection. Future work We need a better understanding of the pathophysiology of transplant failure to allow rational development of effective therapies. Trial registration This trial is registered as EudraCT (2012-004308-36) and ISRCTN (46157828). Funding This project was funded by the National Institute for Health and Care Research (NIHR) Efficacy and Mechanism Evaluation programme (11/100/34) and will be published in full in Efficacy and Mechanism Evaluation; Vol. 10, No. 5. See the NIHR Journals Library website for further project information.