Impact of renal replacement therapy on survival in patients with KDIGO stage 3 acute kidney injury: A propensity score matched analysis

AIM: To investigate the impact of renal replacement therapy (RRT) on 90-day mortality in critically ill patients suffering from KDIGO stage 3 acute kidney injury (AKI) with or without life-threatening complications using propensity score matching analysis. METHODS: We conducted a retrospective analysis of critically ill adult patients with KDIGO Stage 3 AKI with or without RRT during ICU stay between 1/1/2011-31/12/2013. Cox regression analysis and propensity score matching methods were used to determine predictors for 90-day mortality. RESULTS: Among 661 patients, 50.5% received RRT. The unadjusted 90-day mortality rate was 42.5% and 54.1% in patients who had or had not received RRT, respectively. After adjustment with propensity score based on the probability of receiving RRT, the cox regression analysis showed that RRT was associated with a lower 90-day mortality (p<0.001). Among 322 propensity-matched pairs, RRT was associated with lower ICU (23.6% vs. 39.8%, p=0.002), hospital (33.5% vs. 55.9%, p<0.001) and 90-day mortality (34.2% vs. 58.4%, p<0.001), and a higher 90-day renal recovery rate (57.8% vs. 45.3% full recovery, p=0.026) compared with no RRT. When an alternate propensity model was used, the benefits associated with RRT were very similar, except 90-day renal recovery became insignificant. CONCLUSION: Our observational study found that in critically ill patients with KDIGO Stage 3 AKI, RRT may be associated with lower 90-day mortality. The benefit of RRT on renal recovery was less prominent. Medical futility and practice variations may complicate study interpretation. To avoid these limitations, large-scale multicenter, non-observational study is recommended.