Acute kidney injury (AKI) is very common amongst critically ill patients and mortality rates are high. Current evidence is unclear on the most appropriate timing of initiation of renal replacement therapy (RRT) in patients who are critically ill with AKI.

The STARRT-AKI trial examined the impact of the timing of RRT in patients with AKI on mortality and quality of life at 90 days and mortality at one year. NIHR funded an expanded UK arm to the STARRT-AKI trial to examine the cost-effectiveness of accelerated versus standard timing of initiation of RRT in critically ill patients with AKI. As part of this evaluation, the study will utilise clinical and administrative data in the UK to quantify longer term impacts of the timing of RRT on patient outcomes.

The trial recruited 193 patients across England and Scotland. This was well below the target recruitment of 560. The NIHR agreed to fund an observational study which collects the same data on patients, but for whom the timing of RRT is decided by the treating clinician. Trial and observational data will be combined to address the question of cost-effectiveness. We hope to meet the target recruitment of 560 patients through expansion of the observational cohort. Patients are recruited from critical care units with consent provided by the patient, their legal personal representative or a professional consultee. Patients are retrospectively consented if and when they recover consciousness and have capacity to consent. We will collect data on quality of life and primary health and social care resource use at 3, 6 and 12 months. The data will be used to inform a lifetime simulation model of patient costs and outcomes after severe acute kidney injury. The model will simulate patient health trajectories in the form of discrete health states which will include long term dialysis.

Data from the UKRR of patients enrolled in the randomized controlled trial and observational study will be used to estimate the likelihood of patients requiring long term dialysis or a kidney transplant in the simulation model.

The model will estimate the long-term costs and outcomes of early vs standard initiation of RRT in patients with AKI. We will report the cost-effectiveness of early versus standard initiation of RRT. RRT is expensive and patients usually require critical care admission. The information about long term kidney function is essential to inform commissioners of care in the UK and clinical guidelines on the timing of RRT.