Objectives: Among critically ill patients with severe acute kidney injury (AKI) who lack emergent indications for renal replacement therapy (RRT), a strategy of preemptive RRT initiation does not lead to improved outcomes. However, for patients with persistent AKI and without urgent indications for RRT, the safety of prolonged delays in RRT initiation is unclear. We sought to assess the association between progressively longer delays in RRT initiation and clinical outcomes.

Design: A post hoc secondary analysis.

Setting: The multinational STandard vs. Accelerated initiation of Renal Replacement Therapy in Acute Kidney Injury (STARRT-AKI) trial.

Patients: Participants allocated to the standard strategy of the STARRT-AKI trial.

Interventions: The exposure was time from randomization to RRT initiation, evaluated in quartiles and as a continuous variable.

Measurements and main results: The primary outcome was all-cause mortality at 90 days. Secondary outcomes were RRT dependence, RRT-free days, and hospital-free days, all at 90 days, as well length of ICU and hospital stay. Of the 1462 participants allocated to the standard strategy group, 903 (62%) received RRT. Median time (interquartile range) to RRT initiation was 12.1 hours (8.3-13.8 hr), 24.5 hours (21.8-26.5 hr), 46.8 hours (35.2-52.1 hr), and 96.1 hours (76.7-139.2 hr) in quartiles 1-4, respectively. Prolonged time to RRT initiation was associated with a lower risk of death at 90 days (quartile 4 vs. 1: adjusted odds ratio, 0.63 [95% CI, 0.42-0.94]); further analyses using cubic splines and inverse probability weighting to account for immortal time bias showed no association with the risk of death. There was no association between time to RRT initiation and RRT-free days, hospital-free days, or lengths of ICU or hospital stay. Longer delay to RRT initiation had a linear association with RRT dependence at 90 days.

Conclusions: Among patients with no urgent indications and who received RRT in the standard strategy of the STARRT-AKI trial, longer deferral of RRT initiation was not associated with a higher risk of mortality.