Take Home Message
For acute kidney injury (AKI) where there is no indication for immediate renal-replacement therapy (RRT), this study provides reasonable evidence that there is no difference in mortality when initiation of RRT is delayed. Early initiation of RRT may in addition be associated with greater long-term dependence on RRT.
- AKI is common in intensive care and associated with a high mortality.
- There is largely a consensus on when to start RRT for AKI when metabolic or fluid-balance criteria are met, but not in the case of AKI alone.
- Early initiation of RRT could prevent exposure to adverse physiological conditions but may be inappropriate for patients whose renal function recovers before meeting standard criteria for RRT and is itself associated with risks.
- Conduct a prospective study comparing early versus delayed initiation of RRT for AKI (in the absence of indications for immediate RRT).
- 168 hospitals in 15 countries participated.
- October 2015-September 2019. 3019 patients randomly assigned, final number in modified intention to treat analysis 2927.
- Inclusion criteria: 18 years, admitted to ITU with renal dysfunction (Cr> 100 for women, Cr >130 for men) and stage 2-3 severe acute kidney injury as per KDIGO classification. Exclusion criteria: emergency indications for RRT, advanced CKD, previous RRT, uncommon causes for AKI.
- Physicians asked to confirm absence of indication for immediate RRT or likelihood of immediate renal recovery.
- 1:1 randomisation.
- Accelerated strategy – RRT started as soon as possible (withing 12 hours).
- Standard strategy – RRT delayed until K 6, pH 7.20, HCO3 12, P:F ratio200, clinical volume overload or 72 hours post randomisation.
- Primary outcome – all cause mortality at 90 days
- Secondary outcomes – dependence on RRT at 90 days, sustained reduction in renal function, death in ITU at 28 days, number of days off RRT at 90 days, ventilator/vasopressor free days at 28 days, length of hospital stay and health related quality of life at 90 days.
- All reported adverse events were reviewed.
- RRT initiated at a median of 6.1 hours in 1418/1465 patients vs 31.1 hours in 903/1462 patients (accelerated strategy vs standard strategy group).
- 90-day mortality 43.9% for accelerated strategy vs 43.7% for standard strategy (RR 1.00, 95% CI 0.93 to 1.09).
- Dependence on RRT at 90 days: 85/814 patients (10.4%) for accelerated strategy vs 49/815 patients (6.0%) for standard strategy (RR 1.74, 95% CI 1.24 to 2.43).
- More adverse events in accelerated strategy group vs standard strategy group (23% vs 16.5%, RR 1.4 95% CI 1.21 to 1.62), but no difference in serious adverse events.
- Large multi-centre, multi-national study: adequately powered and likely representative of diverse patient populations but possibility for significant differences in clinical practice between countries.
- Inclusion criteria (clinical equipoise) fairly subjective as determined by individual physicians.
- Reporting of adverse events likely to be incomplete.
- 90-day end point for dependence on RRT may not be meaningful – longer follow-up would be interesting.
- This is generally a well-designed study and lends weight to the assertion that the timing of RRT in AKI does not affect mortality in the absence of clinical indications for immediate RRT. Of note is the very high mortality in both groups.
- The study also suggests the accelerated strategy leads to greater long-term dependence on RRT. The authors propose that exposure to RRT may inhibit spontaneous renal recovery, and (though this is only correlation) a significantly smaller proportion of patients were started on RRT in the standard group.
- Overall, there appears reasonable evidence to delay starting RRT where there is clinical equipoise in AKI.
Summary by Dr Nicloa Ingram