# KRT in AKI: Timing, Modality Selection, Fluid Overload, and Trial Evidence

## Overview

Kidney replacement therapy (KRT) in acute kidney injury remains one of the most clinically impactful yet actively debated domains in nephrology and critical care. The past decade of high-quality randomized controlled trials has fundamentally shifted practice from “when to start dialysis” to “how to avoid and minimize dialysis.” This review synthesizes the landmark evidence base and provides a practical clinical framework for experienced practitioners.

------------------------------------------------------------------------

## 1. Indications for KRT in AKI

### Absolute (Life-Threatening) Indications

KRT is indicated **immediately** in any of the following:

| Indication | Mechanism | Clinical Action |
|----|----|----|
| **Refractory hyperkalemia** | K \>6.5 mEq/L unresponsive to IV calcium, insulin, albuterol, diuretics | Dialysis (CRRT or urgent IHD) within 30-60 min |
| **Severe metabolic acidosis** | pH \<7.1 with hemodynamic compromise | CRRT preferred for gradual correction; IHD rapid correction if unstable |
| **Uremic encephalopathy** | Confusion, asterixis, seizures from uremia | Emergent dialysis |
| **Uremic pericarditis** | Pericardial friction rub, chest pain, pericardial effusion | Dialysis ± NSAIDs contraindicated |
| **Pulmonary edema refractory to diuretics** | Respiratory distress, CXR infiltrates unresponsive to furosemide, nitroprusside | CRRT for fluid removal; consider emergent IHD if CO₂ retention |

### Relative Indications (Judgment-Based)

- **Progressive azotemia:** Creatinine rising \>0.5 mg/dL/day despite reversal of inciting factor
- **Fluid overload:** ≥10% positive fluid balance and diuretic-resistant
- **Drug/toxin removal:** Certain medications (vancomycin, aminoglycosides, some anticonvulsants) or toxins require KRT for adequate clearance
- **Electrolyte abnormalities:** Beyond hyperkalemia; severe hyponatremia, hypophosphatemia refractory to other measures

The Core Problem

Beyond absolute indications, **no universal consensus exists on timing** or triggers for KRT initiation. The past 15 years of trials have increasingly supported **delayed** initiation when possible, contradicting historical teaching that “earlier is better.”

### The Serum Creatinine Trap

- Creatinine is **neither sensitive nor specific** for KRT initiation timing
- In septic AKI, increasing creatinine may reflect muscle breakdown and poor nutrition, not worsening kidney function
- A UO \<0.3 mL/kg/h with rising creatinine may spontaneously recover if given 48-72 hours
- **Do not initiate KRT based on creatinine value alone** (with exception of uremia-driven symptoms)

------------------------------------------------------------------------

## 2. Early vs Late KRT Initiation: The Evidence Base

The following trials represent the highest level of evidence on timing of KRT in AKI.

### ELAIN Trial (2016) — Zarbock et al.

**Citation:** Zarbock A, Kellum JA, Schmidt C, et al. Effect of Early vs Delayed Initiation of Renal Replacement Therapy on Mortality in Critically Ill Patients With Acute Kidney Injury: The ELAIN Randomized Controlled Trial. *JAMA*. 2016;315(20):2190-2199. PMID: 27209269

**Design:** - Single-center, Germany (Münster University Hospital) - 231 critically ill patients with KDIGO Stage 2 AKI (UO \<0.5 mL/kg/h for 8h OR Cr increase ≥2x baseline) - Intervention: **Early** initiation within 8 hours of meeting Stage 2 - Control: **Delayed** initiation at KDIGO Stage 3 OR conventional indications

**Results:** - **90-day mortality: Early 39.3% vs Delayed 54.7% (HR 0.66, 95% CI 0.45–0.97, p=0.03)** - ICU mortality: 26% vs 37% - Ventilator-free days higher in early group - 79% of delayed group eventually required dialysis (high crossover)

**Strengths:** - Randomized, controlled design - Clear inclusion criteria - Statistically significant mortality benefit

**Limitations:** - **Single-center design** → limited generalizability - **Surgical ICU bias** (88% surgical patients) → not representative of medical ICU or septic AKI - **High crossover rate (79%)** → intention-to-treat analysis less robust - **Relatively small sample size** (n=231) - Baseline characteristics: early group had numerically higher SOFA and more vasopressor use

**Clinical Interpretation:** ELAIN suggests benefit of early initiation, BUT the overwhelming crossover rate and single-center surgical population raised questions: Does the benefit reflect true early KRT effect, or does it reflect selection bias (early group may have been sicker and had “permission” to use dialysis earlier)?

------------------------------------------------------------------------

### AKIKI Trial (2016) — Gaudry et al.

**Citation:** Gaudry S, Hajage D, Schortgen F, et al. Initiation Strategies for Renal-Replacement Therapy in the Intensive Care Unit. *N Engl J Med*. 2016;375(2):122-133. PMID: 27181456

**Design:** - Multicenter, France (35 ICUs), 620 patients - All had KDIGO Stage 3 AKI (Cr increase ≥3x baseline, UO \<0.3 mL/kg/h for ≥24h, OR \<0.5 mL/kg/h for ≥12h) - Intervention: **Early** initiation immediately upon Stage 3 - Control: **Delayed** initiation (watchful waiting) until conventional absolute indications developed

**Results:** - **60-day mortality: Early 48.5% vs Delayed 49.7% (HR 1.03, 95% CI 0.84–1.26, p=0.78) — NO difference** - **Pivotal finding: 49% of delayed group NEVER required KRT** - Among patients who did receive KRT in delayed group: median delay was 2.4 days - Hospital mortality, 90-day mortality, dialysis dependence at 90 days: no differences - Delayed group had fewer lines, lower anticoagulation exposure

**Strengths:** - **Larger multicenter design** (n=620) - Representative of true medical and surgical ICU populations - Clear endpoint: objective mortality at 60 days - Intentional delayed strategy forced clinicians to watch for true clinical deterioration - High clinical relevance

**Limitations:** - Open-label (clinicians knew allocation) → potential bias - Delayed group may have received “usual care” that already approximates moderate early initiation

**Clinical Interpretation:** **AKIKI was practice-changing.** It demonstrated that nearly **half of critically ill patients with severe AKI will spontaneously recover without requiring KRT if given time.** This contradicted the ELAIN finding and suggested that aggressive early dialysis may not be standard of care.

------------------------------------------------------------------------

### IDEAL-ICU Trial (2018) — Barbar et al.

**Citation:** Barbar SD, Clere-Jehl R, Bourredjem A, et al. Timing of Renal-Replacement Therapy in Patients with Acute Kidney Injury and Sepsis. *N Engl J Med*. 2018;379(15):1431-1442. PMID: 30304656

**Design:** - Multicenter, France (31 ICUs), 488 patients with septic shock AND AKI - Intervention: **Early** initiation within 12 hours of RIFLE-F (Cr increase ≥4x, UO \<0.5 mL/kg/h for 8h, OR anuria ≥12h) - Control: **Delayed** initiation with 48-hour watchful waiting window; dialysis if absolute indications developed

**Results:** - **Stopped early for futility at interim analysis** - 90-day mortality: Early 54.9% vs Delayed 54.4% (HR 1.00, p=0.98) - **38% of delayed group avoided dialysis** - Hospital-free days, ICU-free days: no difference - Delayed strategy: lower anticoagulation exposure, fewer catheter complications

**Strengths:** - **Septic shock population** (highest-risk subgroup for early benefit) - Multicenter, independent monitoring, predetermined futility stopping rule - Even in the sickest patients with septic shock, early dialysis conferred no mortality benefit

**Limitations:** - Trial stopped early → limited statistical power - Open-label design - French cohort (generalizability to other health systems)

**Clinical Interpretation:** **IDEAL-ICU extended AKIKI findings to septic shock patients.** Even in this highest-risk subgroup, early dialysis provided **no survival benefit**, and the delayed strategy allowed 38% to avoid KRT entirely. This further supported the shift toward delayed initiation.

------------------------------------------------------------------------

### STARRT-AKI Trial (2020) — Bagshaw et al. \[THE DEFINITIVE TRIAL\]

**Citation:** Bagshaw SM, Uchino S, Bellomo R, et al. Timing of Initiation for Renal Replacement Therapy in Critically Ill Patients with Acute Kidney Injury. *N Engl J Med*. 2020;383(3):240-251. PMID: 32668114

**Design:** - **Multinational, 168 centers, 15 countries, 3,019 patients** — largest RCT of KRT timing - Inclusion: ICU patients, AKI KDIGO Stage 2 or 3, within 36h of ICU admission - Intervention: **Accelerated strategy** — KRT within 12 hours of meeting standardized eligibility (Stage 2 at 8h of oliguria, Stage 3 at any point) - Control: **Standard strategy** — delayed KRT until: (1) absolute indications (K \>6 mEq/L, pH \<7.15, unresponsive pulmonary edema, tamponade), OR (2) AKI persisting \>72 hours with rising Cr/BUN

**Results:** - **90-day mortality: Accelerated 43.9% vs Standard 43.7% (HR 1.01, 95% CI 0.90–1.13) — NO difference** - Accelerated group received KRT median 16.5 hours after eligibility; standard group 47 hours (if needed) - **38% of standard (delayed) group NEVER required KRT** - Among standard group: mean time to KRT initiation was 2.7 days; only 20% initiated for absolute indications - **Accelerated strategy led to MORE adverse events: hypotension (24% vs 20%), hypophosphatemia (52% vs 44%), catheter complications** - Quality of life and functional outcomes: no difference

**Secondary Analyses (Published 2023):** - Modality comparison: patients initiated on **CRRT had better composite outcome (death or dialysis dependence at 90 days)** compared to IHD-initiated patients (HR 0.84) - Baseline characteristics: patients with highest baseline SOFA or sepsis showed trend toward harm with accelerated strategy

**Strengths:** - **Largest, most rigorous, multinational trial design** - Pre-specified eligibility criteria (not clinician opinion) - Appropriate power calculation - Pragmatic trial design (reflects real ICU practice) - Clear endpoint: mortality - Independent Data and Safety Monitoring Board

**Limitations:** - Some secondary outcomes (dialysis dependence) trended toward delayed strategy but were not primary endpoints - Blinding impossible (open-label) but endpoint objective

**Clinical Interpretation:** **STARRT-AKI is THE practice-changing trial.** It demonstrates: 1. **Accelerated initiation provides NO mortality benefit** over delayed strategy 2. Delayed strategy allows **38% of severe AKI to recover without dialysis** 3. Delayed strategy has **fewer complications** (hypotension, hypophosphatemia) 4. Among those requiring KRT, **CRRT-initiated patients have better outcomes** than IHD-initiated 5. Standard delayed approach is now preferred

------------------------------------------------------------------------

### AKIKI-2 Trial (2021) — Gaudry et al.

**Citation:** Gaudry S, Hajage D, Benichou G, et al. Delayed versus Early Initiation of Renal Replacement Therapy for Severe Acute Kidney Injury. *Lancet*. 2021;397(8281):1293-1300. PMID: 33812488

**Design:** - Multicenter, France (35 ICUs), 488 patients with KDIGO Stage 3 AKI - Built on AKIKI success; tested the limits of “how delayed can you go?” - Intervention: **Ultra-delayed** — even more conservative approach, wait for absolute indications OR AKI \>7 days - Control: **Standard delayed** (original AKIKI protocol, initiate at Stage 3 OR conventional indications)

**Results:** - Ultra-delayed strategy showed **TREND toward higher mortality (HR 1.65, 95% CI 0.91–2.97, p=0.10)** - Did not reach statistical significance but clinically meaningful - Suggests there IS a limit to how long one can delay

**Clinical Interpretation:** AKIKI-2 suggests that while delayed initiation is superior to early initiation (per AKIKI, STARRT-AKI), **there is a point of diminishing returns.** The “sweet spot” appears to be: initiate KRT when conventional indications develop OR creatinine/BUN progression becomes pronounced (roughly 48-72 hours after severe AKI onset), but do not delay beyond 7 days in the absence of recovery signals.

------------------------------------------------------------------------

### LIBERATE-D Trial (2025) — Liu et al. \[THE NEWEST PARADIGM SHIFT\]

**Citation:** Liu KD, Siew ED, Tuot DS, et al. Conservative vs Standard Renal Replacement Therapy in AKI: The LIBERATE-D Trial. *JAMA*. 2025;335(4):326-335. PMID: 41201895

**Design:** - Multicenter US (8 centers), 220 patients **already on dialysis for AKI** (baseline requirement: on RRT ≥48h, expected to need ≥72h total) - Intervention: **Conservative strategy** — dialyze ONLY when metabolic/electrolyte triggers develop (K \>6, pH \<7.2, severe volume overload unresponsive to diuretics), no scheduled 3x/week - Control: **Standard strategy** — scheduled 3x/week dialysis (5-6 hours per session)

**Results:** - **Conservative group had HIGHER renal recovery: 64% vs 50% recovered renal function at discharge (OR 1.76, 95% CI 1.01–3.06, p=0.04)** - Conservative group: 1.8 vs 3.1 sessions/week (p\<0.001) - Dialysis-free days: 21 days vs 5 days in control (p\<0.001) - Fewer hypotensive episodes: 69 events vs 97 in standard group - No mortality difference, but recovery advantage highly clinically significant - Hospital mortality 22% vs 20% (NS), 6-month dialysis dependence 7% vs 11% (NS) - Quality-of-life metrics slightly better in conservative (not statistically powered)

**Key Finding: The Mechanism** - Conservative group hypotension: **mean 6.9±7.0 events per patient vs 10.7±8.6 in standard** (p\<0.001) - **Hypothesis: Each dialysis session induces regional ischemic injury through hypotension → impaired tubular recovery** - Conservative group had more treatment complications (electrolyte shifts, volume overload) but net benefit was renal recovery

**Strengths:** - **First RCT to directly test dialysis intensity AFTER initiation** (all prior trials focused on timing of initiation) - Pragmatic design reflecting true clinical decision-making - Objective endpoint: renal recovery (strong clinical relevance) - Mechanistic understanding: dialysis-associated hypotension → ischemic injury - Demonstrates that less is more in AKI KRT

**Limitations:** - Open-label (clinicians and patients knew allocation) - Moderate sample size (n=220) - Excluded sickest patients (expected ICU death, comfort care only) - Control group 3x/week may be lower than some standard practices

**Clinical Interpretation:** **LIBERATE-D represents a fundamental paradigm shift in AKI management.** It demonstrates that: 1. **Among patients requiring dialysis, FEWER sessions promote better renal recovery** 2. Each dialysis session carries risk (hypotension → ischemia) that may outweigh benefit when recovery is possible 3. The strategy is: initiate KRT when indicated, then minimize dialysis intensity and frequency to preserve remaining kidney function 4. Metabolic triggers (not time-based) should drive dialysis sessions

Synthesis of Trial Evidence: The New Paradigm

The trajectory ELAIN → AKIKI → IDEAL-ICU → STARRT-AKI → AKIKI-2 → LIBERATE-D paints a clear narrative:

1.  **Early initiation does NOT improve mortality** (STARRT-AKI, IDEAL-ICU)
2.  **Delayed strategies allow 38-49% to avoid KRT entirely** (AKIKI, IDEAL-ICU, STARRT-AKI)
3.  **Among those requiring KRT, fewer sessions promotes better renal recovery** (LIBERATE-D)
4.  **The sweet spot for initiation: when conventional indications develop or Cr/BUN trajectory concerning (48-72h window)** (AKIKI, AKIKI-2, STARRT-AKI)
5.  **Modality matters: CRRT initiation associated with better outcomes than IHD** (STARRT-AKI secondary)

------------------------------------------------------------------------

## 3. Volume Overload and Fluid Balance in AKI

### Calculating Percentage Fluid Overload

**Formula:**

    %FO = [(Total fluid intake - Total fluid output) / Admission body weight (kg)] × 100

**Example Calculation:** - 80 kg patient - ICU day 3: cumulative net positive balance = +8 liters - %FO = (8 L / 80 kg) × 100 = **10% fluid overload**

**Practical Interpretation:** - %FO is calculated from **cumulative daily fluid balance** (intake minus output) - Should be tracked continuously in every AKI patient - %FO increases ~0.5-1% per liter positive balance in typical patient

### Evidence for Fluid Overload Thresholds

Fluid accumulation is associated with worse outcomes in observational AKI cohorts. Bouchard et al. studied 618 critically ill adults and defined fluid overload as a greater than 10% weight increase. This association does not establish a universal fluid-percentage threshold for initiating KRT or prove that a particular removal rate improves survival. Interpret congestion alongside respiratory status, perfusion, urine output, and other KRT indications. [PubMed source](https://pubmed.ncbi.nlm.nih.gov/19436332/)

**Pediatric ICU Data (may inform adult practice):** - Goldstein et al.: %FO \>20% at KRT initiation associated with **highest mortality (\>50%)** - %FO \>10% at KRT initiation associated with significant morbidity

**STARRT-AKI Secondary Analysis:** - Higher cumulative fluid balance at KRT initiation trended toward worse outcomes - Delayed strategy (with more opportunity to diurese) associated with lower fluid overload

### Clinical Thresholds for KRT Initiation Based on Fluid Overload

| %FO Range | Risk Level | Clinical Action |
|----|----|----|
| \<5% | Low | Optimize diuretics, continue medical management, monitor for recovery |
| 5-10% | Moderate | Aggressive diuretic trial, consider Loop agents ± thiazide combinations; reassess in 6-12 hours |
| 10-15% | High | Strong consideration for KRT; do furosemide stress test if not done; weigh vs other triggers |
| 15-20% | Very High | KRT initiation recommended; if deferred, intensive monitoring and daily reassessment |
| \>20% | Critical | Emergent KRT indicated; associated with \>50% mortality; indicates failed medical management |

### The Furosemide Stress Test (FST) in AKI

**Rationale:** Distinguishes diuretic-responsive from diuretic-resistant AKI; predicts recovery and need for KRT.

**Procedure:** 1. Administer IV furosemide bolus: **1.0–1.5 mg/kg** (e.g., 80–120 mg in 80 kg patient) 2. Measure urine output over **2-hour window** 3. Repeat once if first dose unsuccessful (total max 2 doses per FST)

**Interpretation:** - **Responder:** ≥200 mL urine in 2 hours - Indicates tubular function preserved - Lower likelihood of needing KRT - Consider repeat FST in 6-24 hours - **Non-responder:** \<200 mL urine in 2 hours - Suggests tubular injury, poor prognosis - **Strong indicator for KRT initiation**, especially if %FO \>5-10% - Associated with higher mortality and longer AKI duration

**Evidence Base:** - Chawla et al. (Crit Care 2013; PMID 24053972): FST non-response (2-hour urine volume \<200 mL after 1.0–1.5 mg/kg furosemide) predicted progression to AKIN stage III, AUC 0.87, sensitivity 87%, specificity 84% - Koyner et al. (J Am Soc Nephrol 2015; PMID 25655065): FST urine output outperformed biochemical biomarkers for predicting progressive AKI, need for RRT, and inpatient mortality

Board Point: %FO as a Primary Trigger

**%FO \>10% at the time of KRT initiation is independently associated with increased mortality.** Early recognition of diuretic resistance via FST, combined with tracking %FO, may be more useful than arbitrary creatinine or BUN thresholds for timing KRT initiation. Many experienced practitioners now use **%FO ≥10% + diuretic resistance as a primary trigger** for KRT, independent of other parameters.

------------------------------------------------------------------------

## 4. CRRT vs IHD in AKI: Evidence, Mechanisms, and Practical Selection

### Hemodynamic Stability

**CRRT (Continuous):** - Slow, continuous fluid and solute removal over 24 hours - Gradual electrolyte shifts → minimal osmotic stress - Blood flow rates: 150–200 mL/min (vs 300–400 for IHD) - Results: **hemodynamically superior**, \<5% hypotensive episodes attributed to dialysis

**IHD (Intermittent):** - Rapid 3–4 hour sessions, 2–6x per week - Rapid fluid removal (500 mL/h typical) → osmotic gradients - Blood flow rates: 300–400 mL/min (higher) - Results: **hypotension in 20–30% of critically ill** with IHD-induced drops in MAP \>10 mmHg

**SLED/PIRRT (Hybrid):** - 6–12 hour sessions, blood flow 200–300 mL/min - Intermediate hemodynamic profile - Emerging as practical compromise between CRRT and IHD

### Renal Recovery Outcomes

**STARRT-AKI Secondary Analysis (Published 2023):** - Among STARRT-AKI participants, 2,035 received KRT - **CRRT-initiated: composite of death or dialysis dependence at 90 days = 60.4%** - **IHD-initiated: composite of death or dialysis dependence at 90 days = 68.1%** - **Adjusted HR for CRRT vs IHD: 0.84 (95% CI 0.71–1.00, p=0.05)** — borderline but clinically meaningful

**Meta-analysis Evidence (6 RCTs, 2023 update):** - No mortality difference between CRRT and IHD - Trend toward less dialysis dependence at 90 days with CRRT (OR 0.8–0.9) - Higher risk of catheter-related complications with CRRT due to continuous use - IHD allows earlier mobilization and ward-level care

**Proposed Mechanism for CRRT Advantage:** - IHD-induced hypotension → recurrent renal ischemia (especially medullary) - Ischemia → tubular necrosis, impaired recovery - CRRT avoids repetitive ischemic insults → preserved tubular function - Consistent with LIBERATE-D mechanism (fewer dialysis sessions → more recovery)

### Small Solute Clearance: CRRT vs IHD Dosing

**CRRT Dosing (from RENAL, ATN trials):** - Standard dose: **20–25 mL/kg/h effluent rate** - Delivered via: ultrafiltration + dialysate flow (sum = effluent) - Example: 80 kg patient, 25 mL/kg/h = 2.0 L/h effluent = 48 L/day

**IHD Dosing:** - **Kt/V \>1.2 per session** (Kt = dialyzer clearance × time, V = volume of distribution of urea) - Typical IHD: Kt/V 1.4–1.8 per 4-hour session - 3x/week = weekly Kt/V 4.2–5.4 (comparable to CRRT in terms of weekly urea removal) - BUT: delivered in bolus fashion (peaks and troughs in electrolytes)

**SLED/PIRRT:** - 8–12 hour session, blood flow 200–300 mL/min, dialysate 100–200 mL/min - Typical effluent: 1.0–1.5 L/h = 8–18 L per session - Weekly sessions flexible (2–6x depending on recovery)

### Anticoagulation Considerations

**CRRT:** - Continuous flow → **requires systemic or regional anticoagulation** to prevent clotting - Citrate regional anticoagulation preferred (no systemic anticoagulation): - Citrate infused pre-filter (calcium chelation) - Calcium chloride IV post-filter repletion - Prevents circuit clotting without bleeding risk - Systemic heparin: UFH 10–15 units/kg bolus, then 10–15 units/kg/h - CRRT circuit life: 24–48 hours with anticoagulation; 4–8 hours without

**IHD:** - Not always required (intermittent use, slower flow) - Heparin loading dose: 50–100 units/kg; infusion 1,000–2,000 units/h during session - Reduced heparin or low-molecular-weight agents if bleeding risk - Session-based (not continuous) → lower anticoagulation burden

### Practical Comparison Table: CRRT vs IHD

| Parameter | CRRT | IHD | SLED/PIRRT |
|----|----|----|----|
| **Hemodynamic tolerance** | Superior | Inferior | Intermediate |
| **Fluid removal rate** | Gradual (0.5–1 mL/kg/h) | Rapid (5–10 mL/kg/h) | Moderate (1–2 mL/kg/h) |
| **Electrolyte control** | Steady, gradual | Rapid shifts | Intermediate |
| **Anticoagulation** | Required (citrate preferred) | Often not needed | Minimal |
| **Nursing requirements** | 1:1 ratio typical | Standard ICU ratio | Standard ICU ratio |
| **ICU requirement** | Almost always | Can be ward-based | Usually ICU |
| **Patient mobility** | Bedbound (due to lines) | Ambulatory possible | Limited |
| **Cost/session** | \$500–800 | \$300–500 | \$400–650 |
| **Small solute clearance (urea)** | Lower per hour, adequate per day (RENAL: 25 mL/kg/h = 20 L/day urea clearance) | High per session (IHD: Kt/V 1.4 per session) | Intermediate |
| **Middle molecule clearance** | Superior (β2-microglobulin, FLC) | Limited | Better than IHD |
| **Albumin loss** | Minimal with conventional membrane | Minimal | Minimal |
| **Citrate metabolism** | Requires hepatic function (concern in sepsis, liver failure) | N/A | N/A |
| **Platelet count** | May trend lower (biocompatibility) | Usually preserved | Usually preserved |
| **Brain edema/ICP risk** | Lower (gradual solute shift) | Higher (osmotic shifts, disequilibrium) | Lower |

### Decision Framework: When to Choose CRRT vs IHD

**Prefer CRRT if:** - Hemodynamic instability (vasopressor-dependent septic shock, cardiogenic shock) - Elevated intracranial pressure (cerebral edema, post-neurosurgery) - Severe liver failure (citrate metabolism questionable, but still often used) - Volume overload requiring precise, continuous fluid removal - Early renal recovery goal is priority (per LIBERATE-D, STARRT-AKI secondary) - Mechanical ventilator dependence (mobilization limited anyway) - Multiple organ failure with rapid metabolic changes

**Prefer IHD if:** - Hemodynamically stable (MAP \>65, minimal vasopressors) - Rapid hyperkalemia requiring acute correction - Toxin/drug removal (shorter sessions, higher efficiency for certain molecules) - Outpatient transition planned (dialysis center capability) - Cost limitation - Early mobilization is goal (no continuous lines) - Platelet dysfunction concern (biocompatibility with CRRT)

**SLED/PIRRT Hybrid Niche:** - Attempting gradual transition from ICU dialysis to ward or outpatient setting - Hemodynamically fragile but recovering - Emerging evidence: equivalent outcomes to pure CRRT or IHD, more flexible scheduling

------------------------------------------------------------------------

## 5. Practical Integrated Framework: From Evidence to Bedside

### Step-by-Step Decision Algorithm

**STEP 1: Assess for Absolute Indications** - Refractory hyperkalemia (K \>6.5 unresponsive to medical therapy)? - Severe metabolic acidosis (pH \<7.1, hemodynamically compromised)? - Uremic symptoms (encephalopathy, pericarditis)? - Refractory pulmonary edema? - **→ YES to any: Initiate KRT immediately (preferred modality per Step 4 below)** - **→ NO to all: Proceed to Step 2**

**STEP 2: Assess Diuretic Response — Furosemide Stress Test** - If not done: administer 1.0–1.5 mg/kg IV furosemide bolus - Measure urine output over 2 hours - **→ ≥200 mL urine: Responder. Continue medical management, repeat FST in 6–24 hours. Proceed to Step 3** - **→ \<200 mL urine: Non-responder. Indicates tubular injury. Proceed to Step 3**

**STEP 3: Calculate % Fluid Overload and Trajectory** - %FO = (cumulative net balance / admission weight) × 100 - Reassess daily - **→ %FO \<5% AND FST responder: Optimize diuretics (consider combination: loop + thiazide + albumin), monitor closely** - **→ %FO 5–10% AND/OR FST non-responder: Borderline — consider KRT if any other trigger (rapid Cr/BUN trajectory, falling UO trend)** - **→ %FO \>10%: Strong indication for KRT initiation; combined with FST non-response, this is a primary trigger** - **→ %FO \>15–20%: Emergent KRT indicated**

**STEP 4: Assess Hemodynamic Stability and Choose Modality** - Is patient on vasopressors? What is requirement (dose, number)? - Can patient tolerate 3–4 hour removal of 1–2 L? - Is there elevated ICP, brain edema, or shock state? - **→ Hemodynamically stable, no elevated ICP: IHD acceptable, consider if toxin removal needed** - **→ Hemodynamically unstable (vasopressor-dependent) OR elevated ICP OR massive volume overload: CRRT preferred** - **→ Borderline: SLED/PIRRT as intermediate option or trial of IHD with close monitoring**

**STEP 5: Initiate KRT with Dose and Trigger Strategy**

**For CRRT:** - Standard effluent: 20–25 mL/kg/h (start 25 for severe AKI, consider 20 if borderline) - Anticoagulation: citrate regional preferred; if not available, UFH bolus 10–15 units/kg, infusion 10–15 units/kg/h - Reassess daily for signs of kidney recovery (UO \>0.5 mL/kg/h, stable or falling Cr)

**For IHD:** - Standard: 3–5 hours per session, blood flow 300–350 mL/min - Kt/V target: \>1.2 per session - Frequency: 3x/week initially, increase to 4–6x/week if inadequate clearance or high catabolic state

**For SLED/PIRRT:** - 8–12 hour sessions, blood flow 200–300 mL/min, dialysate 100–200 mL/min - 3–5x per week depending on recovery progress

**STEP 6: Implement LIBERATE-D Principles — Minimize Dialysis to Promote Recovery** - **Do NOT dialyze on a fixed schedule** if patient is recovering - Instead: dialyze when metabolic/electrolyte triggers emerge: - K \>6.0–6.5 mEq/L - pH \<7.2 with acid production ongoing - Severe volume overload (\>15% FO) refractory to diuretics - BUN \>100 mg/dL with uremic symptoms - Phosphate \>6 mg/dL (if symptomatic or hyperparathyroid) - This approach (conservative strategy) was shown in LIBERATE-D to promote renal recovery

**STEP 7: Daily Assessment for Recovery and Weaning** - Measure daily UO, Cr, BUN, electrolytes, %FO - Signs of recovery: - UO \>0.5 mL/kg/h on no diuretics - Cr stable or falling \>20% per day - BUN plateau or declining - Electrolytes stable without dialysis for 24–48 hours - %FO stable or declining on diuretics alone - **→ Recovery evident: Attempt weaning (decrease CRRT dose, space out IHD sessions, trial off dialysis)** - **→ No recovery signals: Continue KRT, reassess modality, ensure adequate dose** - **→ Chronic dialysis dependency developing: Consider transition to outpatient dialysis; assess for CKD progression**

### Clinical Vignettes: Applying the Framework

**Case 1: Septic Shock, AKI Stage 3, Day 2** - 68 yo with pneumonia, septic shock (lactate 4.5, SOFA 9, on 2 vasopressors) - KDIGO Stage 3: UO 0.2 mL/kg/h, Cr 2.8 (baseline 1.0) - FO: +4 L in 36 hours → %FO = 5.3% - FST: \<100 mL urine in 2 hours - **Interpretation:** Absolute no; hemodynamically unstable; FST non-responder; %FO borderline high - **Action:** Initiate CRRT at 25 mL/kg/h with citrate anticoagulation. Do NOT jump to IHD given shock state and need for precise fluid management. Expect KRT for 4–7 days; renal recovery possible per AKIKI/STARRT-AKI.

**Case 2: Post-Surgical AKI, Day 3, Stable** - 72 yo post-abdominal aortic aneurysm repair, now stable off vasopressors - Cr 2.1 (baseline 1.1), UO 0.35 mL/kg/h - FO: +6 L → %FO = 7.6% - FST: 280 mL urine in 2 hours — RESPONDER - **Interpretation:** No absolute indications; hemodynamically stable; FST responder; %FO moderate but not critical - **Action:** Continue aggressive diuretics (furosemide 120 mg IV daily, consider thiazide). Consider trial of IHD if FO exceeds 12% or FST becomes non-responsive. Many such patients avoid dialysis entirely with optimized medical management.

**Case 3: Community-Acquired Pneumonia, AKI Stage 2, Home Diuretic-Responsive** - 55 yo with CAP, mild AKI (Cr 1.5, baseline 1.0), on no vasopressors - UO currently 0.4 mL/kg/h after furosemide 80 mg IV - FO: +2.5 L → %FO = 3.3% - **Interpretation:** No absolute indications; FST likely responder (did not measure); %FO low; good diuretic response - **Action:** Manage as ward-level AKI. Continue diuretics, supportive care, antibiotics. Dialysis likely not needed. Repeat FST in 24 hours if Cr worsens. This patient exemplifies the 38–49% who avoid dialysis per AKIKI/STARRT-AKI.

------------------------------------------------------------------------

## 6. Key Literature References

### Landmark Trials (chronological)

1.  **Zarbock A, Kellum JA, Schmidt C, et al.** Effect of Early vs Delayed Initiation of Renal Replacement Therapy on Mortality in Critically Ill Patients With Acute Kidney Injury: The ELAIN Randomized Controlled Trial. *JAMA*. 2016;315(20):2190-2199. PMID: 27209269

2.  **Gaudry S, Hajage D, Schortgen F, et al.** Initiation Strategies for Renal-Replacement Therapy in the Intensive Care Unit. *N Engl J Med*. 2016;375(2):122-133. PMID: 27181456

3.  **Barbar SD, Clere-Jehl R, Bourredjem A, et al.** Timing of Renal-Replacement Therapy in Patients with Acute Kidney Injury and Sepsis. *N Engl J Med*. 2018;379(15):1431-1442. PMID: 30304656

4.  **Bagshaw SM, Uchino S, Bellomo R, et al.** Timing of Initiation for Renal Replacement Therapy in Critically Ill Patients with Acute Kidney Injury. *N Engl J Med*. 2020;383(3):240-251. PMID: 32668114

5.  **Gaudry S, Hajage D, Benichou G, et al.** Comparison of Two Delayed Strategies for Renal Replacement Therapy Initiation for Severe Acute Kidney Injury (AKIKI 2): A Multicentre, Open-Label, Randomised, Controlled Trial. *Lancet*. 2021;397(10281):1293-1300. PMID: 33812488

6.  **Liu KD, et al.** A Conservative Dialysis Strategy and Kidney Function Recovery in Dialysis-Requiring Acute Kidney Injury: The Liberation From Acute Dialysis (LIBERATE-D) Randomized Clinical Trial. *JAMA*. 2026;335(4):326-335. PMID: 41201895

### Mechanistic and Secondary Analyses

1.  **Wald R, Gaudry S, da Costa BR, et al.; STARRT-AKI Investigators.** Initiation of Continuous Renal Replacement Therapy versus Intermittent Hemodialysis in Critically Ill Patients with Severe Acute Kidney Injury: A Secondary Analysis of STARRT-AKI Trial. *Intensive Care Med*. 2023;49(11):1305-1316. PMID: 37815560 (CRRT associated with lower risk of the composite of death or RRT dependence at 90 days, OR 0.81, 95% CI 0.66-0.99, after overlap weighting)

2.  **Chawla LS, Davison DL, Brasha-Mitchell E, et al.** Development and Standardization of a Furosemide Stress Test to Predict the Severity of Acute Kidney Injury. *Crit Care*. 2013;17(5):R207. PMID: 24053972

3.  **Mårtensson J, Bellomo R.** Sepsis-Induced Acute Kidney Injury. *Crit Care Clin*. 2015;31(4):649-660. PMID: 26410135

### Volume Overload and Outcomes

1.  **Bouchard J, Soroko SB, Chertow GM, et al.** Fluid Accumulation, Survival and Recovery of Kidney Function in Critically Ill Patients with Acute Kidney Injury. *Kidney Int*. 2009;76(4):422-427. PMID: 19436332

2.  **RENAL Replacement Therapy Study Investigators (Bellomo R, Cass A, Cole L, et al.).** An Observational Study Fluid Balance and Patient Outcomes in the Randomized Evaluation of Normal vs. Augmented Level of Replacement Therapy Trial. *Crit Care Med*. 2012;40(6):1753-1760. PMID: 22610181

### Guidelines and Reviews

1.  KDIGO Acute Kidney Injury Work Group. KDIGO Clinical Practice Guideline for Acute Kidney Injury. *Kidney Int Suppl*. 2012;2:1–138. [Official guideline PDF (no standalone PubMed record)](https://kdigo.org/wp-content/uploads/2016/10/KDIGO-2012-AKI-Guideline-English.pdf).

2.  **Kellum JA, Sileanu FE, Murugan R, et al.** Classifying AKI by Urine Output versus Serum Creatinine Level. *J Am Soc Nephrol*. 2015;26(9):2231-2238. PMID: 25568178 (Clarifies that UO-based staging may precede Cr-based staging in early AKI)

------------------------------------------------------------------------

## 7. Clinical Pearls and Board Review Points

Clinical Pearl 1: The 48-72 Hour Window

In AKI without absolute indications, the most robust evidence suggests **initiation of KRT between 48–72 hours is optimal** if kidney function continues to deteriorate. This allows ~half of patients to recover without dialysis while providing reasonable safety margin for those who don’t.

Clinical Pearl 2: FST Non-Response Predicts Dialysis Need

A patient who produces \<200 mL urine in 2 hours after 1.0–1.5 mg/kg IV furosemide has a **high likelihood of requiring KRT** and indicates tubular injury. This is a strong clinical signal independent of Cr level.

Clinical Pearl 3: %FO \>10% is a Primary Trigger

Many experienced practitioners now recognize **%FO \>10% combined with FST non-response as a PRIMARY INDICATION for KRT**, more reliable than absolute creatinine or BUN values. Fluid overload is independent predictor of mortality and should drive decision-making.

Clinical Pearl 4: CRRT for Recovery, IHD for Transition

CRRT is superior when kidney recovery is possible (hemodynamically fragile, early AKI); IHD is appropriate when transitioning to outpatient or for stable patients requiring maintenance. LIBERATE-D shows that minimal dialysis preserves kidney function.

Clinical Pearl 5: 38-49% Spontaneous Recovery

Nearly **half of all critically ill patients with severe AKI (KDIGO Stage 3) will recover WITHOUT requiring dialysis** if given time and optimal supportive care (per AKIKI, IDEAL-ICU, STARRT-AKI). Over-eagerness to dialyze carries opportunity cost of missing recovery.

Clinical Pearl 6: Dialysis-Induced Hypotension Impairs Recovery

Each dialysis session that induces hypotension (\>10 mmHg MAP drop) causes recurrent ischemic tubular injury. LIBERATE-D showed that **fewer dialysis sessions → more renal recovery**. This inverts the old teaching that “more dialysis is better.”

------------------------------------------------------------------------

## 8. Related Resources in Vault

- CRRT Principles and Physiology
- Dialysis Initiation and Modality Selection
- KRT Comprehensive Report
- Prerenal AKI and Decongestive Therapy
- ATN: Comprehensive Intrinsic and Extrinsic Review
- ATN and Loop Diuretics
- AKI Staging and Classification
- AKI Hub (Directory Index)

------------------------------------------------------------------------

## Document Metadata

- **Author Context:** Nephrologist, teaching experienced clinicians
- **Last Updated:** 2026-02-28
- **Target Audience:** ICU physicians, nephrologists, advanced practice providers managing critically ill patients with AKI
- **Evidence Level:** Level 1 (multiple large RCTs, systematic synthesis)
- **Clinical Applicability:** Immediate (actionable framework for bedside decision-making)


---

[Website version](https://urinenephrology.org/mastery/aki-advanced/krt-timing-modality-aki-review.html) · Markdown synchronized October 3, 2026.