Differential effect on mortality of the timing of initiation of renal replacement therapy according to the criteria used to diagnose acute kidney injury: an IDEAL-ICU substudy

Background This substudy of the randomized IDEAL-ICU trial assessed whether the timing of renal replacement therapy (RRT) initiation has a differential effect on 90-day mortality, according to the criteria used to diagnose acute kidney injury (AKI), in patients with early-stage septic shock. Methods Three groups were considered according to the criterion defining AKI: creatinine elevation only (group 1), reduced urinary output only (group 2), creatinine elevation plus reduced urinary output (group 3). Primary outcome was 90-day all-cause death. Secondary endpoints were RRT-free days, RRT dependence and renal function at discharge. We assessed the interaction between RRT strategy (early vs. delayed) and group, and the association between RRT strategy and mortality in each group by logistic regression. Results Of 488 patients enrolled, 205 (42%) patients were in group 1, 174 (35%) in group 2, and 100 (20%) in group 3. The effect of RRT initiation strategy on 90-day mortality across groups showed significant heterogeneity (adjusted interaction p = 0.021). Mortality was 58% vs. 42% for early vs. late RRT initiation, respectively, in group 1 ( p = 0.028); 57% vs. 67%, respectively, in group 2 ( p = 0.18); and 58% vs. 55%, respectively, in group 3 ( p = 0.79). There was no significant difference in secondary outcomes. Conclusion The timing of RRT initiation has a differential impact on outcome according to AKI diagnostic criteria. In patients with elevated creatinine only, early RRT initiation was associated with significantly increased mortality. In patients with reduced urine output only, late RRT initiation was associated with a nonsignificant, 10% absolute increase in mortality.