In critically ill adult ICU patients with stage 2–3 acute kidney injury, should renal replacement therapy be initiated early (within 12 hours of indication) or delayed (watchful waiting up to 48 hours)?

Early vs Delayed RRT in ICU

In critically ill adult ICU patients with stage 2-3 acute kidney injury without life-threatening complications, a delayed/standard strategy of RRT initiation (watchful waiting up to 48-72 hours) should be preferred over early initiation (within 12 hours), as it results in equivalent mortality while avoiding unnecessary RRT in approximately 40-50% of patients and reducing adverse events.

Evidence-Based Recommendation

The most recent and highest quality evidence comes from the STARRT-AKI trial 1, a multinational RCT of 3,019 critically ill patients published in 2020. This definitive study showed:

• No mortality benefit: 90-day mortality was identical between accelerated (within 12 hours) and standard strategies (43.9% vs 43.7%, RR 1.00)
• Harm from early initiation: Increased RRT dependence at 90 days (10.4% vs 6.0%, RR 1.74) and more adverse events (23.0% vs 16.5%, P<0.001)
• Avoidable RRT: 38% of delayed-strategy patients never required RRT

This is corroborated by the AKIKI trial 2 (2016,620 patients), which showed no mortality difference but found 49% of delayed-strategy patients avoided RRT entirely, with earlier recovery of diuresis (P<0.001).

When to Initiate RRT

Immediate Initiation (Within Hours)

Start RRT emergently when life-threatening complications exist 3:

• Severe hyperkalemia (refractory to medical management)
• Medically refractory volume overload with pulmonary edema
• Severe metabolic acidosis (pH typically <7.15)
• Uremic complications (pericarditis, encephalopathy, bleeding)

Standard/Delayed Strategy (Watchful Waiting)

For stage 2-3 AKI without the above emergent indications 3:

• Monitor closely for 48-72 hours
• Assess broader clinical context and trends rather than single BUN/creatinine thresholds
• Consider: hemodynamic status, fluid balance, urine output trends, response to diuretics
• Initiate RRT if emergent indications develop OR if AKI persists beyond 72 hours without recovery

Critical Nuances

The ELAIN trial 4 (2016,231 patients) showed mortality benefit with early RRT (39.3% vs 54.7%, P=0.03), but this contradicts larger, more recent trials. Key differences:

• Single-center study with smaller sample size
• Used NGAL biomarker >150 ng/mL as entry criterion (not standard practice)
• Earlier stage of AKI (KDIGO stage 2 vs stage 3 in other trials)
• Not externally validated

The 2022 Cochrane meta-analysis 5 of 12 RCTs (4,880 participants) confirms:

• Low certainty evidence: Early RRT may have little to no difference on 30-day mortality (RR 0.97)
• Moderate certainty evidence: No difference in mortality after 30 days (RR 0.99)
• High certainty evidence: Early RRT increases hypophosphatemia (RR 1.80), hypotension (RR 1.54), cardiac arrhythmias (RR 1.35), and infections (RR 1.33)

Practical Algorithm

1. Assess for emergent indications (hyperkalemia, acidosis, volume overload, uremic complications)

   • If present → Initiate RRT immediately

2. If no emergent indications but stage 2-3 AKI:

   • Optimize hemodynamics and fluid status
   • Trial diuretics if appropriate
   • Monitor for 48-72 hours with serial assessments:
     • Urine output (most robust predictor of recovery 6)
     • Serum creatinine trends
     • Fluid balance
     • Development of emergent indications

3. Initiate RRT if:

   • Emergent indication develops during observation
   • AKI persists >72 hours without improvement
   • Clinical deterioration despite supportive care

Common Pitfalls to Avoid

• Don't use single BUN/creatinine thresholds alone to trigger RRT 3
• Don't initiate RRT preemptively to "prevent" complications—this increases harm without benefit 1, 5
• Don't delay when emergent indications exist—these require immediate RRT 3
• Recognize that 40-50% of "delayed" patients recover without ever needing RRT 2, 1, 5

Resource Considerations

The Canadian Society of Nephrology 3 notes that aggressive early RRT initiation consumes more resources. In a healthcare system like Canada's, widespread adoption of early strategies should await stronger evidence, particularly given:

• Increased catheter-related infections 2
• Higher adverse event rates 1, 5
• Increased RRT dependence 1
• No mortality benefit 7, 2, 1, 5

For hemodynamically unstable patients, continuous RRT (CRRT) is preferred over intermittent RRT regardless of timing 3, 8, 3.