Skip to content
evident
Original paper
2020NEJMNephrology

STARRT-AKI

Timing of Initiation of Renal-Replacement Therapy in Acute Kidney Injury

RCTAcute kidney injury

On this page

Overview

In critically ill severe AKI, accelerated RRT did not lower 90-day death; among survivors, RRT dependence was 10.4% vs 6.0% (accelerated vs standard).

Clinical takeaway

Do not start RRT immediately just because severe AKI criteria are met. Accelerated RRT did not lower 90-day death (43.9% vs 43.7%; RR 1.00; 95% CI 0.93-1.09; P=0.92). Among 90-day survivors, continued RRT dependence was 10.4% vs 6.0% (RR 1.74; 95% CI 1.24-2.43).

Key result

In the modified intention-to-treat population, 90-day death was 43.9% with accelerated RRT versus 43.7% with a standard strategy (RR 1.00; 95% CI 0.93-1.09; P=0.92).

Practice impact

Delay RRT until a conventional indication or AKI past 72 hours; earlier start did not lower 90-day death and had more RRT dependence among survivors.

Evidence

Study design

Multinational, randomized, controlled trial comparing an accelerated renal-replacement strategy with a standard strategy in critically ill patients with severe acute kidney injury.

Enrollment

3,019

Follow-up

Death from any cause at 90 days.

Geography

Multinational

Clinical question

Does starting renal-replacement therapy within 12 hours after eligibility criteria are met lower 90-day death compared with a standard strategy in critically ill patients with severe AKI?

Population

  • Critically ill patients with severe acute kidney injury.

Intervention

Accelerated strategy: renal-replacement therapy initiated within 12 hours after the patient met eligibility criteria.

Comparator

Standard strategy: renal-replacement therapy discouraged unless conventional indications developed or acute kidney injury persisted for more than 72 hours.

Primary outcome

Death from any cause at 90 days.

An accelerated renal-replacement strategy was not associated with a lower risk of death at 90 days than a standard strategy (RR 1.00; 95% CI 0.93-1.09; P=0.92) in the modified intention-to-treat population.

Relative risk / 1 / 95% CI 0.93-1.09 / p=0.92

Key results

  • Of 3019 patients who underwent randomization, 2927 (97.0%) were included in the modified intention-to-treat analysis (1465 accelerated-strategy, 1462 standard-strategy).
  • Renal-replacement therapy was performed in 1418 of 1465 (96.8%) in the accelerated-strategy group and in 903 of 1462 (61.8%) in the standard-strategy group.
  • At 90 days, death had occurred in 643 of 1465 (43.9%) with the accelerated strategy versus 639 of 1462 (43.7%) with the standard strategy; relative risk 1.00 (95% CI 0.93-1.09; P=0.92).
  • Among 90-day survivors, continued dependence on renal-replacement therapy was confirmed in 85 of 814 (10.4%) in the accelerated-strategy group versus 49 of 815 (6.0%) in the standard-strategy group; relative risk 1.74 (95% CI 1.24-2.43).

Harms

  • Adverse events occurred in 346 of 1503 patients (23.0%) in the accelerated-strategy group and in 245 of 1489 patients (16.5%) in the standard-strategy group (P<0.001); these denominators differ from the modified intention-to-treat population.
  • Among 90-day survivors, continued RRT dependence was more common after the accelerated strategy (10.4% vs 6.0%; RR 1.74; 95% CI 1.24-2.43).

Clinical Use

Practice impact

  • Do not initiate RRT solely because severe AKI eligibility criteria are met if there is no conventional indication and AKI has not persisted beyond 72 hours.
  • A standard strategy led to RRT in 61.8% of patients and did not increase 90-day death compared with nearly universal early RRT.

Applicability

  • Most applicable to critically ill patients with severe acute kidney injury in whom the timing of RRT is still a choice.

Limitations

  • The primary analysis used a modified intention-to-treat population (2927 of 3019 randomized).
  • The standard strategy still led to RRT in 61.8% of patients, so this is not a comparison with never starting dialysis.
  • Adverse-event rates used a different denominator (1503 vs 1489) than the modified intention-to-treat death analysis.

Common misinterpretations

  • This is not a trial of whether RRT works; it tests when to start it.
  • A null 90-day mortality result is not evidence that immediate RRT is required, and it is not evidence that delaying RRT is harmless in patients who already have a conventional indication.

Citation

STARRT-AKI Investigators, Canadian Critical Care Trials Group, Australian and New Zealand Intensive Care Society Clinical Trials Group, et al. Timing of Initiation of Renal-Replacement Therapy in Acute Kidney Injury. N Engl J Med. 2020;383(3):240-251. doi:10.1056/NEJMoa2000741

Among critically ill patients with acute kidney injury, an accelerated renal-replacement strategy was not associated with a lower risk of death at 90 days than a standard strategy.