8.3 Indications & Timing for RRT in Critical Illness: AKIKI & STARRT-AKI

Urgent Indications for Renal Replacement Therapy: The "AEIOU" Framework

In critically ill patients with acute kidney injury, the decision to initiate renal replacement therapy (RRT) begins with identifying absolute, life-threatening clinical emergencies. When present, these conditions mandate emergent dialytic intervention regardless of KDIGO stage or urine output trajectory. The classical "AEIOU" mnemonic structures this urgent evaluation:

Loading diagram...

1. Acidosis (Metabolic)

Severe, uncompensated metabolic acidosis with an arterial $\text{pH} < 7.15\text{ to }7.20$ and serum bicarbonate $< 10\text{ to }12\text{ mEq/L}$ that is refractory to conservative medical therapy (or where exogenous sodium bicarbonate infusion is contraindicated due to severe hypernatremia, hypocalcemia, or volume overload). Acidemia below 7.15 severely impairs myocardial contractility, induces peripheral vasodilation, desensitizes vascular smooth muscle to exogenous catecholamines/vasopressors, and triggers malignant ventricular arrhythmias.

2. Electrolytes (Hyperkalemia)

Severe, refractory hyperkalemia (serum potassium $> 6.5\text{ mEq/L}$) accompanied by electrocardiographic signs of membrane instability (peaked T waves, PR prolongation, loss of P waves, QRS widening, or sine-wave patterns) that fails to stabilize following aggressive pharmacotherapy:

• Intravenous calcium gluconate or calcium chloride (myocardial membrane stabilization);
• Intravenous regular insulin plus $50\%$ dextrose (intracellular potassium shift);
• Inhaled beta-2 agonists (albuterol $10\text{ to }20\text{ mg}$ nebulized);
• Gastrointestinal potassium binders (sodium zirconium cyclosilicate [Lokelma] or patiromer). Severe refractory hypermagnesemia ($> 8\text{ to }10\text{ mg/dL}$) with flaccid paralysis or respiratory depression, or severe refractory hypercalcemic crisis in the setting of oliguric renal failure also represent dialytic indications.

3. Ingestions (Dialyzable Toxins: "SLIME")

Emergent extracorporeal removal is indicated for specific low-molecular-weight, water-soluble, low-volume-of-distribution toxins summarized by the SLIME mnemonic:

• S — Salicylates: Serum level $> 90\text{ to }100\text{ mg/dL}$ ($> 80\text{ mg/dL}$ in renal failure) or presence of severe neurological toxicity, pulmonary edema, or refractory acid-base decompensation;
• L — Lithium (EXTRIP 2015): Recommended if kidney function is impaired and $[\text{Li}^+] > 4.0\text{ mEq/L}$, or with decreased consciousness, seizures, or life-threatening dysrhythmias at any level; suggested if $[\text{Li}^+] > 5.0\text{ mEq/L}$, with significant confusion, or if the expected time to reach $< 1.0\text{ mEq/L}$ exceeds 36 hours;
• I — Isopropanol: When accompanied by severe hemorrhagic gastritis or refractory hypotension (though conservative support is often attempted first);
• M — Methanol (EXTRIP 2015): Coma, seizures, new visual deficits, pH $\le 7.15$, anion gap $> 24$, or persistent acidosis despite treatment; or a methanol level $> 70\text{ mg/dL}$ on fomepizole ($> 60$ on ethanol, $> 50$ without an alcohol dehydrogenase blocker);
• E — Ethylene Glycol (EXTRIP 2023): Anion gap $> 27\text{ mmol/L}$ (suggested at 23-27), coma, seizures, or AKI; on fomepizole, ECTR is suggested for an EG level $> 50\text{ mmol/L}$ ($> 310\text{ mg/dL}$) or osmolal gap $> 50$; a glycolate concentration above $12\text{ mmol/L}$ (suggested at 8-12) is an alternative trigger where it can be measured;
• Additional dialyzable toxins: Theophylline, toxic overdoses of metformin presenting with profound metformin-associated lactic acidosis (MALA), carbamazepine, and valproic acid.

4. Overload (Fluid)

Diuretic-refractory hypervolemia manifest as worsening hypoxemic respiratory failure (flash pulmonary edema) or severe systemic venous congestion in an oliguric or anuric patient. In patients unresponsive to maximal loop diuresis ($> 160\text{ to }240\text{ mg}$ IV furosemide equivalent combined with a thiazide such as metolazone or IV chlorothiazide), extracorporeal ultrafiltration is life-saving.

5. Uremia (End-Organ Manifestations)

Overt uremic organ dysfunction:

• Uremic Pericarditis: Manifest as retrosternal pleuritic chest pain and a scratchy, pathognomonic triphasic friction rub, frequently with a serosanguinous pericardial effusion.

• Uremic Encephalopathy: Cognitive slowing, asterixis (flapping tremor), multifocal myoclonus, seizures, and progressive coma.
• Uremic Bleeding Diathesis: Impaired primary hemostasis resulting from uremic toxins (such as guanidinosuccinic acid) interfering with platelet glycoprotein IIb/IIIa receptors and von Willebrand factor binding. Clinically reversed by desmopressin (DDAVP $0.3\ \mu\text{g/kg}$ IV), cryoprecipitate, and emergent dialytic clearance.

---

The Timing Controversy: Accelerated vs. Watchful Waiting Strategies

While urgent "AEIOU" indications demand immediate intervention, the management of patients with severe AKI (KDIGO Stage 2 or 3) who lack immediate life-threatening complications historically generated fierce debate.

• The Pro-Early Hypothesis: Proponents argued that preemptive, accelerated initiation of RRT would prevent the lethal systemic sequelae of uremia, eliminate toxic inflammatory mediators, maintain tight fluid balance, and avert emergency dialysis under unstable conditions.
• The Pro-Delayed (Watchful Waiting) Hypothesis: Opponents highlighted that renal replacement therapy is inherently invasive and hazardous. Catheter insertion risks pneumothorax, arterial puncture, and catheter-related bloodstream infections (CRBSI). Dialytic circuits induce hypotension, activate systemic leukocytes, strip away essential trace elements and antibiotic levels, and cause fluctuations in renal perfusion that retard endogenous tubular regeneration. Crucially, many patients with Stage 2 or 3 AKI spontaneously recover renal function if given adequate time and supportive medical management.

---

Landmark Randomized Controlled Trials

Four landmark multicenter randomized trials definitively resolved this debate over the past decade:

Loading diagram...

1. The AKIKI Trial (NEJM 2016)

• Design: Multicenter randomized trial in 31 French ICUs enrolling 620 patients with KDIGO Stage 3 AKI receiving invasive mechanical ventilation, vasopressor therapy, or both.
• Strategies:
  • Early Strategy: RRT initiated immediately within 6 hours of documenting KDIGO Stage 3 AKI.
  • Delayed Strategy: RRT withheld until urgent indications developed (hyperkalemia $> 6.0\text{ mEq/L}$, pH $< 7.15$, acute pulmonary edema, or BUN $> 112\text{ mg/dL}$) or persistent oliguria/anuria exceeded 72 hours.
• Results:
  • 60-day all-cause mortality was virtually identical: $48.5\%$ in the early group vs. $49.7\%$ in the delayed group ($p = 0.79$).
  • $49\%$ of patients in the delayed group spontaneously recovered renal function and never received RRT.
  • The early initiation group suffered significantly higher rates of catheter-related bloodstream infections ($10\%$ vs. $5\%$, $p = 0.03$).

2. The IDEAL-ICU Trial (NEJM 2018)

• Design: Multicenter French trial enrolling 488 patients with septic shock and severe AKI in the RIFLE "Failure" stage.
• Strategies: Early initiation (within 12 hours of Failure stage) vs. Delayed initiation (deferred for 48 hours unless AEIOU indications arose).
• Results: The trial was terminated early for futility at the planned interim analysis. 90-day mortality showed no significant difference ($58\%$ early vs. $54\%$ delayed, $p = 0.38$). In the delayed group, $38\%$ of patients never required RRT.

3. The STARRT-AKI Trial (NEJM 2020)

• Design: The definitive international, multicenter trial across 168 centers in 15 countries, randomizing 3,019 critically ill patients (2,927 in the modified intention-to-treat analysis) with KDIGO Stage 2 or 3 AKI.
• Strategies:
  • Accelerated Strategy: RRT initiated within 12 hours of meeting eligibility (median time to initiation: 6.1 hours).
  • Standard (Watchful Waiting) Strategy: RRT withheld until urgent indications developed (severe hyperkalemia, severe acidosis, volume overload) or persistent renal failure lasted $\ge 72\text{ hours}$ (median time to initiation in those who received RRT: 31.1 hours).
• Results:
  • 90-Day All-Cause Mortality: Identical between groups: $43.9\%$ (643/1,465) in the accelerated group vs. $43.7\%$ (639/1,462) in the standard group (relative risk 1.00; $95\%$ CI 0.93 to 1.09; $p = 0.92$).
  • RRT Avoidance: In the standard group, $38\%$ of patients never received RRT due to spontaneous renal recovery or death before intervention.
  • Adverse Events: The accelerated group experienced significantly higher rates of adverse events ($23.0\%$ vs. $16.5\%$, $p < 0.001$), driven by severe hypotension and hypophosphatemia.
  • Persistent Dialysis Dependence: Among 90-day survivors, the rate of chronic dialysis dependence was significantly higher in the accelerated group ($10.4\%$ vs. $6.0\%$; RR 1.74, 95% CI 1.24 to 2.43). Premature catheter insertion, circuit hypotension, and dialytic removal of physiological regulatory molecules impaired long-term parenchymal renal recovery.

4. The AKIKI-2 Trial (Lancet 2021)

• Design: Tested whether postponing RRT initiation even further than the standard delayed strategy provided incremental benefit or caused harm. Compared a "delayed" strategy (dialysis when oliguria $> 72\text{ hours}$ or BUN reaches $112\text{ mg/dL}$) with an "ultra-delayed" strategy (withholding until BUN reached $> 140\text{ mg/dL}$ or urgent AEIOU criteria).
• Results: The primary outcome, RRT-free days at day 28, did not differ (median 12 vs. 10 days). The more-delayed strategy offered no benefit and had a signal of harm: 60-day mortality was numerically higher (about 55% vs. 44%), reaching significance only in an adjusted analysis. The practical lesson is that once oliguria exceeds 72 hours or BUN exceeds 112 mg/dL, postponing RRT further gains nothing.

Comparison Table of Landmark RRT Timing Trials

---

Clinical Consensus and Pragmatic Management

Synthesizing this robust trial evidence establishes a standard ICU clinical pathway:

1. Screen for Immediate Emergent Indications: Immediately initiate RRT if refractory "AEIOU" emergencies are present.
2. If No Emergent Criteria (Stage 2/3 AKI): Adopt a watchful waiting strategy with aggressive medical support:
   • Optimize systemic hemodynamics and renal perfusion pressure (target MAP $\ge 65\text{ mmHg}$, or MAP 75 to 85 mmHg in chronic hypertensive patients);
   • Administer a furosemide stress test or trial of loop diuretics to assess tubular responsiveness;
   • Restrict unnecessary maintenance fluids to prevent progressive fluid accumulation;
   • Administer potassium binders and sodium bicarbonate for mild-to-moderate acid-base/electrolyte abnormalities;
   • Perform serial clinical examinations, blood gas analyses, and electrolyte panels every 6 to 12 hours.
3. Initiate RRT When: Refractory AEIOU criteria emerge, progressive hypervolemia compromises respiratory mechanics, or significant oliguria persists past 48 to 72 hours with escalating azotemia (BUN approaching 100 to 112 mg/dL).

---

Effluent Dose Target: Evidence from ATN and RENAL Trials

Once the decision to initiate CRRT is made, the clinician must prescribe the dialytic intensity, quantified as the effluent flow rate per unit body weight ($Q_{\text{effluent}}$, in $\text{mL/kg/h}$).

Landmark Intensity Trials

Two landmark multicenter randomized trials evaluated whether "intensive" or high-dose CRRT improved patient outcomes compared to standard dosing:

1. The VA/NIH Acute Renal Failure Trial Network (ATN) Study (NEJM 2008):
   • Randomized 1,124 critically ill patients with AKI and multiorgan failure to intensive therapy (CVVHDF at $35\text{ mL/kg/h}$ or daily intermittent hemodialysis) versus standard therapy (CVVHDF at $20\text{ mL/kg/h}$ or thrice-weekly intermittent hemodialysis).
   • Results: 60-day mortality was identical: $53.6\%$ intensive vs. $51.5\%$ standard ($p = 0.47$). No differences were seen in renal recovery or ICU length of stay. Intensive therapy significantly increased hypotensive episodes and hypophosphatemia.
2. The ANZICS RENAL Trial (NEJM 2009):
   • Randomized 1,508 critically ill patients in Australia and New Zealand to high-dose CVVHDF ($40\text{ mL/kg/h}$) versus standard-dose CVVHDF ($25\text{ mL/kg/h}$).
   • Results: 90-day mortality was identical: $44.7\%$ in the high-dose group vs. $44.7\%$ in the standard-dose group ($p = 0.99$). High-dose CRRT caused significantly higher rates of severe hypophosphatemia ($65\%$ vs. $54\%$) and increased the clearance of essential trace minerals and therapeutic antibiotics.

Clinical Standard and the "Prescribed vs. Delivered" Gap

Based on the ATN and RENAL trials, the KDIGO guidelines recommend a delivered effluent dose of $20\text{ to }25\text{ mL/kg/h}$ for CRRT in acute kidney injury.