Severe acute kidney injury not treated with renal replacement therapy: characteristics and outcome

BACKGROUND: Only a proportion of critically ill patients with severe [RIFLE (Risk, Injury, Failure, Loss, End-stage renal disease) criteria, class-F] acute kidney injury (AKI) appear to receive renal replacement therapy (RRT). The aim of this study was to study the characteristics and outcome of patients with severe (RIFLE-F) AKI who did not receive RRT. METHODS: We identified all consecutive patients admitted to our institution that developed RIFLE-F AKI by creatinine criteria over a 3-year period and did not receive RRT, and compared their characteristics and outcomes with those of RIFLE-F RRT-treated patients. RESULTS: Within the study period, 20,126 patients were admitted to our institution for >24 h. Among them, 2949 were admitted to the intensive care unit (ICU) and 195 developed RIFLE-F AKI. Of these, 90 received RRT (RRT patients) and 105 did not (no-RRT patients). Compared with RRT patients, no-RRT patients were similar in terms of age, gender and ward of origin. However, they had a shorter median ICU stay (2.7 versus 7.9 days; P < 0.001), required less mechanical ventilation (56.2 versus 70%; P < 0.05) and had a lower mean Acute Physiology and Chronic Health Evaluation III score (82.7 versus 86.7; P < 0.05). The two main reasons these patients did not receive RRT were limitations of medical therapy (LOMT) orders in 41 (39%) cases and expected renal functional improvement in 59 (56.2%). Mortality in no-RRT patients was 58.1% compared with 55.5% in the RRT group (P = 0.72). After exclusion of LOMT patients, the mortality of the no-RRT group, although lower than that of the RRT group, remained high (30.5 versus 55%; P < 0.001). Most of these deaths occurred after ICU discharge and appeared secondary to underlying chronic diseases or recurrence of the initial insult. CONCLUSIONS: After exclusion of LOMT patients, about a third of critically ill patients with severe (RIFLE-F) AKI did not receive RRT. A third of these patients died in hospital. The timing of the deaths and their underlying causes do not suggest that a broader application of RRT would have changed patient outcomes.