# Renal Replacement Therapy in Critical Illness

## Indications for RRT in the ICU

### Conventional Indications (AEIOU)

The decision to initiate renal replacement therapy in the critically ill patient is guided by a set of conventional indications that can be organized using the mnemonic AEIOU. Acidosis refers to severe metabolic acidosis with a pH below 7.15 that is refractory to medical management including bicarbonate therapy. Electrolyte derangements, most critically refractory hyperkalemia with serum potassium exceeding 6.5 mEq/L accompanied by ECG changes despite maximal medical therapy including calcium, insulin with dextrose, sodium bicarbonate, and potassium-binding agents, constitute a compelling indication. Intoxication with dialyzable toxins such as methanol, ethylene glycol, lithium, salicylates, and metformin represents a unique indication in which the primary goal is toxin removal rather than renal support. Overload describes fluid accumulation exceeding 10 percent of body weight that is refractory to diuretic therapy, a threshold that has been independently associated with increased mortality in critically ill patients. Uremia with symptomatic manifestations including encephalopathy, pericarditis, neuropathy, and bleeding diathesis from platelet dysfunction represents the final conventional indication. These indications represent situations in which RRT is considered necessary regardless of the trajectory of renal function, and their presence should prompt initiation without delay.

### Timing of RRT Initiation

The question of when to initiate RRT in the absence of conventional indications has been one of the most intensely studied topics in critical care nephrology. The STARRT-AKI trial of 2020, the largest and most definitive study on this question, randomized critically ill patients with KDIGO Stage 2 or 3 AKI to an accelerated strategy with initiation within 12 hours versus a standard strategy in which RRT was initiated only when conventional indications developed or Stage 3 AKI persisted for more than 72 hours. The trial demonstrated no difference in 90-day mortality, which was 44 percent in both groups. Notably, the accelerated group experienced higher rates of catheter-related bloodstream infections and hypotension related to RRT initiation, underscoring the harms of unnecessary RRT. The IDEAL-ICU trial of 2018 addressed a similar question specifically in septic shock patients with AKI meeting the RIFLE Failure classification, comparing early initiation within 12 hours to delayed initiation at 48 hours. Again, no mortality difference was found, and importantly, 38 percent of patients in the delayed group never required RRT at all, having recovered renal function spontaneously. The AKIKI trial of 2016 compared early initiation at KDIGO Stage 3 with delayed initiation triggered only by conventional indications and similarly demonstrated no mortality difference, with 49 percent of patients in the delayed group avoiding RRT entirely. The AKIKI-2 trial of 2021 extended this investigation further by comparing a delayed strategy to an even more-delayed approach, finding that the more-delayed strategy was associated with a trend toward harm, suggesting that there is a point at which excessive delay becomes detrimental. The current consensus, synthesized from these landmark trials, is that routine early initiation of RRT offers no benefit and may cause harm, but clinicians should not unreasonably delay RRT once conventional indications are present. The optimal approach is to wait for conventional indications or persistent Stage 3 AKI that is not recovering, while maintaining vigilant monitoring for the development of urgent indications.

### Furosemide Stress Test for RRT Timing

The furosemide stress test provides a practical bedside tool for predicting which patients will progress to require RRT. After administration of a furosemide bolus of 1 mg/kg in diuretic-naive patients or 1.5 mg/kg in patients with prior diuretic exposure, a urine output below 200 mL over the subsequent 2 hours predicts the need for RRT with a sensitivity of 87 percent and a specificity of 84 percent. This test can help identify patients who are likely to progress and may benefit from early planning of vascular access and logistical preparation for RRT, even if initiation is deferred pending the development of conventional indications.

## Modalities

### Continuous Renal Replacement Therapy (CRRT)

Continuous renal replacement therapy is the preferred modality for hemodynamically unstable ICU patients. By providing treatment continuously over 24 hours, CRRT achieves slower rates of solute and fluid removal compared to intermittent modalities, resulting in superior hemodynamic tolerance and more stable intracranial pressure. The three principal CRRT modalities differ in their mechanism of solute clearance. Continuous venovenous hemofiltration, or CVVH, employs convective clearance exclusively, using a replacement fluid that is infused either before or after the hemofilter while the ultrafiltrate containing waste products is discarded. Continuous venovenous hemodialysis, or CVVHD, relies solely on diffusive clearance by running a dialysate solution through the hemofilter countercurrent to blood flow. Continuous venovenous hemodiafiltration, or CVVHDF, combines both convective and diffusive clearance and is the most commonly used modality in practice, offering the broadest spectrum of solute removal. Blood flow rates for CRRT are typically maintained at 150 to 250 mL/min, which is lower than those used in intermittent hemodialysis. The effluent rate, which represents the prescribed dose of CRRT, is targeted at 20 to 25 mL/kg/hr, though the actual delivered dose is typically 15 to 20 percent less than prescribed due to circuit interruptions, filter changes, and downtime.

### Intermittent Hemodialysis (IHD)

Intermittent hemodialysis delivers treatment in concentrated 3 to 4 hour sessions, typically performed every other day or daily depending on clinical needs. The higher blood flow rates of 300 to 400 mL/min enable rapid solute clearance, making IHD particularly efficient for emergent hyperkalemia and toxin removal where high dialysance is desirable. The principal disadvantage of IHD is the risk of hemodynamic instability resulting from the rapid removal of fluid and solutes, which produces intradialytic hypotension. For this reason, IHD is not recommended in hemodynamically unstable patients but becomes increasingly appropriate as patients improve and vasopressor support is weaned. IHD is generally better tolerated in the later phases of critical illness and offers practical advantages in terms of allowing patients time off the machine for mobilization, procedures, and other activities.

### Sustained Low-Efficiency Dialysis (SLED/Prolonged IHD)

Sustained low-efficiency dialysis, also known as prolonged intermittent hemodialysis, represents a hybrid approach that uses standard IHD equipment at reduced blood flow rates of approximately 200 mL/min and reduced dialysate flow rates of approximately 300 mL/min, delivered over extended treatment times of 6 to 12 hours. This approach provides better hemodynamic tolerance than standard IHD while achieving more efficient solute clearance than CRRT, owing to the shorter total treatment time that allows more downtime for procedures, diagnostic studies, and patient mobilization. A growing body of evidence suggests that SLED may be equivalent to CRRT even in hemodynamically unstable patients, and the modality offers significant practical advantages including the use of standard IHD machines without the need for specialized CRRT supplies, compatibility with dialysis nursing schedules, and lower overall cost.

| Feature | CRRT (CVVHDF) | IHD | SLED/Prolonged IHD |
|---|---|---|---|
| Treatment Duration | 24 hr/day (continuous) | 3–4 hr sessions | 6–12 hr sessions |
| Blood Flow Rate | 150–250 mL/min | 300–400 mL/min | ~200 mL/min |
| Hemodynamic Tolerance | Excellent | Poor (rapid fluid/solute shifts) | Good (intermediate) |
| Solute Clearance Rate | Slow and continuous | Rapid | Intermediate |
| Anticoagulation | Regional citrate preferred | Systemic heparin or none | Systemic heparin or citrate |
| Best For | Hemodynamically unstable, brain injury, precise fluid management | Stable patients, urgent K⁺/toxin removal | Transition from CRRT; hybrid approach |
| Mobility | Limited (continuous circuit) | Off-machine time allows mobilization | Some off-machine time |
| Cost | Higher (specialized supplies) | Lower (standard equipment) | Lower |
| Mortality Difference | No difference vs IHD in RCTs | No difference vs CRRT in RCTs | Equivalent in emerging evidence |

### CRRT vs. IHD

Multiple randomized controlled trials comparing CRRT to IHD in ICU patients with acute kidney injury have demonstrated no mortality difference between the modalities. CRRT offers advantages in hemodynamic stability, continuous fluid management, and stable intracranial pressure, making it the preferred choice in patients with brain injury, hemodynamic instability, or conditions requiring precise fluid balance. IHD offers advantages in faster solute removal, less cumulative anticoagulation exposure, lower cost, and greater nursing familiarity in many institutions. The practical choice between modalities is often driven by institutional protocols, available equipment and staffing, and the individual patient's hemodynamic status.

<image>Comparison diagram of three RRT modalities (CRRT, IHD, SLED) showing circuit configurations. For each modality: schematic of the extracorporeal circuit (access, blood pump, dialyzer/hemofilter, return) with flow rates labeled. Below each circuit: bar charts comparing key parameters — treatment duration (24h vs. 4h vs. 8h), blood flow rate, dialysate/replacement fluid rate, solute clearance efficiency, hemodynamic tolerance (rated as good/moderate/poor), and cost. Central comparison table listing specific indications for each modality: CRRT for hemodynamic instability/brain injury, IHD for stable patients/urgent K+ clearance/toxin removal, SLED as intermediate option. Include a decision node algorithm at the bottom for selecting modality based on hemodynamic status and clinical scenario.</image>

## CRRT Technical Considerations

### Vascular Access

Vascular access for CRRT requires a dual-lumen dialysis catheter, typically 13.5 French and non-tunneled. The right internal jugular vein is the preferred site, offering the best blood flow rates and the lowest recirculation. The femoral vein represents the easiest site for insertion and is a reasonable alternative, though it requires the patient to remain supine. The left internal jugular vein is used when other sites are unavailable but is associated with higher recirculation rates due to the tortuous path to the superior vena cava. The subclavian vein should be avoided whenever possible because of the risk of subclavian vein stenosis, which can compromise future arteriovenous fistula creation if the patient progresses to end-stage kidney disease. Catheter length should be matched to the insertion site, with 15 to 20 cm for internal jugular placement and 24 to 25 cm for femoral placement. Catheter malfunction is the most common cause of CRRT circuit interruption and typically results from fibrin sheath formation at the catheter tip or positional kinking.

### CRRT Prescription

The CRRT prescription encompasses several key parameters that must be individually adjusted. Blood flow rate is maintained at 150 to 250 mL/min, with higher flows improving solute clearance but increasing the risk of hemolysis. The effluent dose should target 20 to 25 mL/kg/hr as recommended by KDIGO guidelines. Two landmark trials established that higher doses do not provide additional benefit. The ATN trial of 2008 compared an intensive dose of 35 mL/kg/hr against a less intensive dose of 20 mL/kg/hr and found no mortality difference. The RENAL trial of 2009 similarly demonstrated no difference in 90-day mortality or renal recovery between 40 mL/kg/hr and 25 mL/kg/hr. Given that the actual delivered dose is typically 15 to 20 percent less than prescribed due to treatment interruptions, prescribing 25 to 30 mL/kg/hr to achieve a delivered dose of 20 to 25 mL/kg/hr is standard practice. The net ultrafiltration rate, which determines the patient-specific fluid removal target, is typically set between 0 and 200 mL/hr, with rates exceeding 1.5 to 2 mL/kg/hr risking hemodynamic instability. Replacement fluid can be administered as pre-dilution, where it is infused before the hemofilter, or post-dilution, where it is infused after the filter. Pre-dilution reduces the hematocrit within the filter, prolonging filter life by reducing clotting, but this dilutional effect decreases clearance efficiency by 15 to 20 percent, requiring compensatory dose increases. Post-dilution achieves more efficient solute clearance but at the cost of shorter filter life due to hemoconcentration within the filter.

### Anticoagulation for CRRT

| Anticoagulation | Mechanism | Target | Advantages | Disadvantages | Contraindication |
|---|---|---|---|---|---|
| Regional Citrate | Chelates ionized Ca²⁺ in circuit | Circuit iCa 0.25–0.35; systemic iCa 1.0–1.2 | Longer filter life, less bleeding (KDIGO preferred) | Citrate accumulation in liver failure (total:ionized Ca >2.5) | Severe hepatic failure |
| Systemic Heparin | Antithrombin-mediated anticoagulation | aPTT 45–55 sec | Familiar, reversible with protamine | Higher bleeding risk; HIT risk | Active bleeding, HIT |
| No Anticoagulation | Pre-dilution + saline flushes | N/A | No bleeding risk | Shorter filter life (12–18 hr) | N/A (used when others contraindicated) |

Regional citrate anticoagulation has become the preferred anticoagulation strategy for CRRT, as recommended by KDIGO guidelines. In this approach, citrate is infused into the circuit upstream of the filter, where it chelates ionized calcium and thereby prevents coagulation within the extracorporeal circuit. A separate calcium infusion is administered through a central venous catheter distal to the circuit to restore systemic ionized calcium levels. Circuit ionized calcium is monitored with a target of 0.25 to 0.35 mmol/L, while systemic ionized calcium is maintained at 1.0 to 1.2 mmol/L. Citrate accumulation, which occurs primarily in patients with hepatic failure or severe shock, is detected by a rising total-to-ionized calcium ratio exceeding 2.5, a phenomenon that occurs because citrate-calcium complexes contribute to total calcium measurements but not to ionized calcium. The CASH trial comparing citrate to heparin for CRRT anticoagulation demonstrated that citrate provides longer filter life and less bleeding with similar mortality. Systemic heparin anticoagulation, targeting an aPTT of 45 to 55 seconds, remains an alternative but carries a higher bleeding risk than citrate. For patients with coagulopathy defined by an INR above 2 or platelet count below 50,000, or for those with active bleeding, CRRT can be run without anticoagulation, employing pre-dilution and periodic saline flushes to extend filter life, though filter lifespan is reduced to 12 to 18 hours compared to 48 to 72 hours with citrate anticoagulation.

### Troubleshooting CRRT

Frequent filter clotting should prompt a systematic evaluation including increasing pre-dilution, verifying citrate dosing, excluding heparin-induced thrombocytopenia, and reviewing the filter change protocol. Hemodynamic instability during CRRT should be addressed by reducing the net ultrafiltration rate, slowing blood flow, and warming the replacement fluid. Electrolyte derangements are common during CRRT, with hypophosphatemia being the most frequent complication and requiring proactive phosphorus supplementation, as the continuous nature of CRRT efficiently removes phosphorus. Hypokalemia and hypomagnesemia also occur and require regular monitoring and supplementation. Circuit air detection alarms necessitate checking all connections and ensuring that lines are properly secured. Hypothermia is an underappreciated complication of CRRT, as the extracorporeal circuit causes heat loss of 1 to 2 degrees Celsius without warming, necessitating the use of an in-line warmer on the return line.

## Drug Dosing During CRRT

### Principles

Drug dosing during CRRT represents one of the most challenging aspects of critical care pharmacology. CRRT removes drugs based on several physicochemical properties including molecular weight, protein binding, volume of distribution, and sieving coefficient. Drugs with low molecular weight below 500 Daltons, low protein binding, and small volume of distribution are most efficiently removed by CRRT. The sieving coefficient, defined as the ratio of filtrate concentration to plasma concentration, provides a quantitative measure of drug removal, with values above 0.5 indicating significant clearance. Drug clearance by CRRT can be estimated by multiplying the sieving coefficient by the effluent flow rate. Hydrophilic drugs such as beta-lactam antibiotics, aminoglycosides, and vancomycin are significantly cleared by CRRT and require dose adjustment. Lipophilic drugs such as azole antifungals and macrolide antibiotics undergo minimal CRRT clearance and typically do not require adjustment.

### Common Drug Adjustments

Vancomycin dosing during CRRT requires a loading dose of 15 to 20 mg/kg followed by maintenance dosing of 10 to 15 mg/kg every 12 to 24 hours, guided by AUC-based therapeutic drug monitoring. Piperacillin-tazobactam should be dosed at 4.5 grams every 6 to 8 hours, with extended infusion preferred to optimize time above the minimum inhibitory concentration. Meropenem is typically dosed at 1 to 2 grams every 8 hours, with standard dosing usually adequate in CRRT patients. Cefepime requires adjustment to 1 to 2 grams every 8 to 12 hours. Fluconazole, being highly water-soluble with low protein binding, is significantly cleared by CRRT and requires a loading dose of 400 to 800 mg followed by a maintenance dose of 200 to 400 mg daily. Caspofungin, conversely, has high protein binding and a large volume of distribution, resulting in minimal CRRT clearance and requiring no dose adjustment. Given the complexity and variability of drug clearance during CRRT, pharmacy consultation should be obtained for all drug dosing decisions in patients on CRRT.

## Discontinuation of RRT

### Criteria for RRT Cessation

The decision to discontinue RRT requires assessment of multiple clinical parameters indicating renal recovery. A urine output exceeding 400 to 500 mL per day without diuretic support, or exceeding 2,000 mL per day with diuretics, suggests sufficient recovery to consider cessation. Spontaneous improvement in creatinine clearance, resolution of the indications that prompted RRT initiation including hyperkalemia, acidosis, and fluid overload, and hemodynamic stability off vasopressors all support discontinuation. A 6-hour measured creatinine clearance exceeding 15 to 20 mL/min provides a more objective measure of renal recovery and can guide the decision to stop RRT.

### Transition to IHD

As hemodynamic stability improves, transitioning from CRRT to SLED or IHD represents a logical step that allows for patient mobilization, reduced anticoagulation exposure, and improved nursing workflow. This transition should be accompanied by continued monitoring of renal function parameters, as many patients will continue to recover function over days to weeks following discontinuation of continuous modalities.

### Renal Recovery

Approximately 50 to 60 percent of ICU patients requiring RRT for AKI will ultimately recover renal function, though the timeline for recovery is variable and may extend over weeks to months. Risk factors for non-recovery include pre-existing chronic kidney disease, diabetes, older age, and the severity and duration of the acute kidney injury. All patients who required RRT during their ICU stay should receive nephrology referral and monitoring of creatinine and proteinuria at 3 months post-discharge, as the transition from AKI to chronic kidney disease represents a significant long-term risk that requires ongoing surveillance.

<image>CRRT circuit diagram showing a CVVHDF configuration. Starting from dual-lumen dialysis catheter in right internal jugular vein: blood drawn from "arterial" (red) lumen through blood pump at 200 mL/min. Pre-filter: citrate infusion (showing citrate bag with dose calculations) and pre-dilution replacement fluid. Blood enters hemofilter (hollow-fiber membrane shown in cross-section with blood compartment and dialysate compartment labeled). Dialysate flows counter-current at 1500 mL/hr. Effluent (spent dialysate + ultrafiltrate) collected in waste bag with volume measurement. Post-filter: calcium chloride infusion via separate CVC lumen (showing dose). Return through "venous" (blue) lumen. Air detector and pressure monitors at key points. Include monitoring panel showing: circuit ionized calcium target 0.25-0.35, systemic ionized calcium target 1.0-1.2, total/ionized Ca ratio for citrate accumulation assessment, access pressures, filter pressure, and transmembrane pressure.</image>

## Special Applications

### RRT for Toxin Removal

The use of RRT for toxin removal requires that the offending substance meet specific pharmacokinetic criteria: low molecular weight below 500 Daltons, low protein binding below 80 percent, small volume of distribution below 1 L/kg, and water solubility. When these criteria are met, intermittent hemodialysis is generally preferred over CRRT because of its higher clearance rates, which are critical for acute toxin removal. The EXTRIP workgroup has provided evidence-based recommendations for specific toxins. Hemodialysis is recommended for methanol and ethylene glycol when serum levels exceed 50 mg/dL, or in the presence of severe metabolic acidosis or end-organ damage. Lithium should be dialyzed when serum levels exceed 4 mEq/L in chronic toxicity or 5 mEq/L in acute toxicity with significant symptoms. Salicylate toxicity warrants hemodialysis when levels exceed 100 mg/dL, or in the presence of altered mental status, renal impairment, or severe acidosis. Metformin-associated lactic acidosis with pH below 7.0 and lactate above 20 mmol/L represents an indication for hemodialysis. Valproic acid overdose requires hemodialysis when serum levels exceed 1300 mg/L, or in the setting of cerebral edema or hemodynamic instability. Following IHD for toxin removal, CRRT may be continued to prevent rebound, which occurs when the toxin redistributes from tissue stores back into the bloodstream, a phenomenon particularly relevant for lithium and methanol.

### Therapeutic Plasma Exchange (TPE)

Therapeutic plasma exchange, while not strictly a form of renal replacement therapy, is often managed by critical care and nephrology teams. TPE removes high molecular weight substances including antibodies, immune complexes, and complement components that cannot be cleared by conventional RRT modalities. The principal indications include thrombotic thrombocytopenic purpura, anti-GBM disease, ANCA-associated vasculitis with diffuse alveolar hemorrhage, myasthenic crisis, and Guillain-Barre syndrome. Each session exchanges 1 to 1.5 plasma volumes, with replacement using fresh frozen plasma for TTP and albumin for most other indications.

## Key Clinical Pearls

- Routine early RRT initiation does not improve outcomes (STARRT-AKI, IDEAL-ICU, AKIKI) — wait for conventional indications unless urgent
- Nearly 50% of patients in delayed RRT strategies never require dialysis — patience prevents unnecessary RRT
- Prescribed CRRT dose of 20-25 mL/kg/hr is sufficient — higher doses (ATN trial, RENAL trial) provide no benefit
- Regional citrate anticoagulation is the preferred anticoagulation strategy for CRRT — longer filter life, less bleeding
- Monitor total/ionized calcium ratio during citrate-based CRRT: ratio >2.5 indicates citrate accumulation (especially in liver failure)
- Hypophosphatemia is the most common electrolyte derangement during CRRT — proactive phosphorus replacement is essential
- Drug dosing in CRRT requires pharmacy collaboration — hydrophilic, low-protein-bound drugs are significantly cleared
- Follow all ICU-AKI patients post-discharge for CKD development — the AKI-to-CKD transition is common and underappreciated

## References

1. STARRT-AKI Investigators. Timing of initiation of renal-replacement therapy in acute kidney injury. N Engl J Med. 2020;383(3):240-251.
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.
3. VA/NIH Acute Renal Failure Trial Network. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med. 2008;359(1):7-20.
4. RENAL Replacement Therapy Study Investigators. Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med. 2009;361(17):1627-1638.
5. Zarbock A, Küllmar M, Kindgen-Milles D, et al. Effect of regional citrate anticoagulation vs systemic heparin anticoagulation during continuous kidney replacement therapy on dialysis filter life span and mortality among critically ill patients with acute kidney injury. JAMA. 2020;324(16):1629-1639.