AKI Renal Replacement Timing

The optimal timing of renal replacement therapy (RRT) in critically ill patients with AKI was one of the most debated questions in nephrology and critical care. Practice varied widely, with some centers initiating RRT early based on AKI severity criteria and others delaying until conventional indications (refractory hyperkalemia, acidosis, volume overload) developed. The absence of randomized evidence left clinicians relying on physiological reasoning and observational data that were heavily confounded by indication.

The AKIKI trial randomized 620 critically ill patients with KDIGO stage 3 AKI requiring mechanical ventilation or vasopressors to early (within 6 hours) or delayed RRT initiation. At 60 days, mortality was nearly identical: 48.5% in the early group versus 49.7% in the delayed group (p=0.79). Crucially, 49% of patients in the delayed group never required RRT at all, and the delayed group had fewer catheter-related bloodstream infections (5% vs 10%, p=0.03). This demonstrated that a watchful waiting approach was safe and avoided unnecessary RRT in a substantial proportion of patients.

The STARRT-AKI trial confirmed and extended these findings in the largest RRT timing trial. Among 2,927 critically ill patients across 168 hospitals in 15 countries, accelerated RRT (within 12 hours of eligibility) showed no mortality benefit at 90 days: 43.9% versus 43.7% for the standard strategy (RR 1.00; 95% CI 0.93-1.09; p=0.92). Furthermore, among survivors, the accelerated group had higher rates of continued RRT dependence at 90 days (10.4% vs 6.0%; RR 1.74; 95% CI 1.24-2.43). Only 61.8% of the standard group ultimately required RRT.

KDIGO and critical care guidelines shifted to recommend against routine early RRT initiation based solely on AKI stage or biochemical criteria. The consensus is to initiate RRT when life-threatening complications develop (refractory hyperkalemia, severe metabolic acidosis, refractory volume overload) or clinical deterioration occurs, allowing time for potential kidney recovery. A watchful waiting approach with close monitoring is preferred.

The AKIKI and STARRT-AKI trials established that early RRT initiation does not improve survival and may increase RRT dependence among survivors. A standard or delayed approach, initiating for conventional indications, is now the preferred strategy. This avoids unnecessary RRT in approximately 40-50% of patients who would recover kidney function. Remaining challenges include defining the optimal monitoring strategy to identify patients who will not recover without RRT and developing biomarkers to predict AKI trajectory.