Renal replacement therapy

Contents

Indications

No single biochemical threshold mandates RRT; initiation is based on clinical context and trajectory. The AEIOU mnemonic covers the principal indications:

Indication Details
A — Acidosis Metabolic acidosis refractory to medical management (pH <7.1 despite bicarbonate and treatment of cause)
E — Electrolytes Hyperkalaemia refractory to medical treatment (K >6.5 mmol/L or lower with ECG changes/haemodynamic compromise)
I — Intoxication Dialysable toxins (see below)
O — Overload Volume overload causing pulmonary oedema refractory to diuretics
U — Uraemia Symptomatic uraemia: encephalopathy, pericarditis (friction rub), uraemic bleeding (platelet dysfunction)

Dialysable Toxins

RRT removes water-soluble, low-protein-bound, low-molecular-weight toxins:

  • Removed: methanol, ethylene glycol, lithium, salicylates, metformin, aminoglycosides (overdose)
  • Not effectively removed: paracetamol, TCAs, digoxin (high protein binding), most benzodiazepines

Modalities

Intermittent Haemodialysis (IHD)

  • Sessions 3–5h, 3–5×/week (or daily if needed)
  • Solute clearance by diffusion across high-permeability membrane; blood and dialysate flow counter-current
  • Most efficient per-hour clearance; suited to haemodynamically stable patients
  • Rapid fluid and solute shifts can cause haemodynamic instability → avoid in cardiovascularly unstable patients
  • Dialysis disequilibrium syndrome: rare; rapid urea clearance → osmotic gradient → cerebral oedema; prevent with short initial sessions

Continuous Renal Replacement Therapy (CRRT)

Preferred in haemodynamically unstable patients: gradual fluid and solute removal over 24h.

Modality Mechanism Principle
CVVH (continuous venovenous haemofiltration) Convection Ultrafiltrate removed by transmembrane pressure; replacement fluid added
CVVHD (continuous venovenous haemodialysis) Diffusion Dialysate run countercurrent to blood; no replacement fluid
CVVHDF (continuous venovenous haemodiafiltration) Convection + diffusion Combines both; highest clearance

Convection removes larger molecules (middle molecules, inflammatory cytokines) more effectively than diffusion. Clinical benefit of cytokine clearance in sepsis is biologically plausible but not proven in RCTs.

SLED / SLEDD

  • Sustained low-efficiency (daily) dialysis: 8–12h sessions using modified IHD equipment at lower blood and dialysate flow rates
  • Haemodynamic tolerability intermediate between IHD and CRRT
  • Useful for patients who are borderline haemodynamically stable or when CRRT circuit life is poor

Peritoneal Dialysis (PD)

  • Rarely used in adult ICU in high-income settings; commonly used in paediatric ICU and resource-limited settings
  • Slower clearance; infection risk (peritonitis); limited by abdominal surgery, respiratory compromise

Dosing

Dose is expressed as effluent volume per hour per kilogram body weight (mL/kg/h):

Effluent dose (mL/kg/h) = (Ultrafiltrate rate + Dialysate rate + Net fluid removal) / Patient weight

Evidence

RENAL Trial (Bellomo, NEJM 2009): 1508 critically ill patients randomised to 25 mL/kg/h vs 40 mL/kg/h effluent dose CRRT. No difference in 90-day mortality (44.7% vs 44.7%). Higher dose not beneficial and associated with more electrolyte disturbances and hypophosphataemia.

ATN Trial (VA/NIH, NEJM 2008): similar finding; higher-intensity RRT (more frequent IHD + higher CRRT dose) vs lower intensity → no mortality benefit at 60 days.

Current recommendation: target 20–25 mL/kg/h delivered dose. To achieve this, prescribe 25–30 mL/kg/h to account for circuit downtime (filter changes, procedures, clotting: CRRT is typically running only 70–80% of prescribed time).

Timing of Initiation

When to start RRT in AKI without an absolute indication (AEIOU) is debated.

Key Trials

IDEAL-ICU Trial (Barbar, NEJM 2018): 488 patients with stage 3 AKI from septic shock; early initiation (within 12h) vs delayed (wait for absolute indication or spontaneous recovery). No 90-day mortality difference. Importantly, ~49% of the delayed group never required RRT — suggesting many patients with severe AKI recover without it if given time.

STARRT-AKI Trial (2020, NEJM): 3019 critically ill patients with AKI; accelerated strategy (initiation at stage 2 AKI) vs standard strategy (absolute indications only). No difference in 90-day survival. Accelerated group had no benefit in RRT dependence at 90 days, and had more adverse events.

Implication: there is no evidence to support routine early initiation of RRT in AKI. Allow time for spontaneous recovery unless absolute indications (AEIOU) develop or AKI is clearly progressive and causing clinical deterioration.

Anticoagulation for CRRT

Circuit anticoagulation prevents filter clotting and extends circuit life. Options:

Regional Citrate Anticoagulation (RCA)

  • KDIGO 2012 and ESICM guidelines: preferred method for CRRT anticoagulation
  • Mechanism: citrate infused into circuit before the filter → chelates ionised calcium (Ca²⁺) within the circuit → prevents coagulation (calcium is an essential cofactor for multiple clotting factors)
  • Calcium infused into patient (post-filter or via separate line) to restore systemic ionised calcium
  • Result: anticoagulation confined to the circuit; patient's systemic coagulation is unaffected
  • Benefits: longer circuit life than UFH; no systemic bleeding risk; preferred in patients with HIT, coagulopathy, or bleeding risk

Citrate toxicity / accumulation:

  • Normally, citrate is rapidly metabolised in the liver (→ bicarbonate) and in peripheral tissues
  • If hepatic function is severely impaired: citrate accumulates systemically → chelates systemic Ca²⁺
  • Signs: ↑ total serum calcium (citrate-calcium complexes) + ↓ ionised calcium + metabolic alkalosis + elevated anion gap
  • Diagnostic: total:ionised calcium ratio >2.5 = citrate accumulation
  • Management: reduce citrate infusion rate; may need to switch to UFH or heparin-free if severe liver failure

Contraindications to RCA: severe liver failure (citrate not metabolised); severe metabolic alkalosis.

Unfractionated Heparin (UFH)

  • Systemic anticoagulation; standard alternative to citrate
  • Shorter circuit life than RCA; bleeding risk in patients with coagulopathy
  • Monitoring: APTT target 45–60 seconds (pre-filter sample)

HIT (Heparin-Induced Thrombocytopenia)

  • Stop all heparin immediately, including flushes and LMWH
  • Alternative anticoagulants for CRRT:
    • Argatroban (direct thrombin inhibitor; hepatically metabolised — preferred if renal failure; dose 0.5–2 mcg/kg/min; monitor APTT)
    • Danaparoid (heparinoid; some cross-reactivity with HIT antibodies — avoid if strongly positive 4Ts)
    • Bivalirudin (direct thrombin inhibitor; partially renally eliminated; used in some centres)
    • Fondaparinux: factor Xa inhibitor; renally eliminated — avoid in severe AKI needing CRRT
    • Citrate (regional): no systemic anticoagulant → effectively avoids HIT problem; preferred if no contraindication

Practical Considerations

Access

  • Vascular access catheter (permcath / temporary): typically femoral, jugular, or subclavian
  • Femoral catheters: easier to insert, higher infection risk; suitable for short term (<3–5 days)
  • Internal jugular: preferred for longer-term access; right side preferred (straighter to SVC)
  • Subclavian: avoid in AKI (risk of subclavian stenosis compromising future AV fistula maturation)

Electrolyte monitoring

  • Phosphate: commonly removed by CRRT → hypophosphataemia (add phosphate to replacement fluid)
  • Potassium: add to replacement fluid/dialysate based on levels
  • Bicarbonate: citrate-based replacement fluid provides bicarbonate equivalent; monitor acid-base

Drug dosing in CRRT

  • Many drugs are removed by CRRT: antibiotics (particularly β-lactams, vancomycin), antifungals, antivirals
  • Underdosing in sepsis is a real risk — check microbiological guidance; vancomycin levels must be monitored; pip-tazo, meropenem doses are increased in CRRT
  • Conversely, haemofiltration removes toxins: important in poisoning management

Key Trials

Trial Year Question Finding
ATN (VA/NIH, NEJM) 2008 High-intensity vs low-intensity RRT in AKI No mortality benefit from higher intensity (35 mL/kg/h vs 20 mL/kg/h effluent)
RENAL (Bellomo, NEJM) 2009 25 vs 40 mL/kg/h CRRT dose No mortality difference; higher dose not beneficial
IDEAL-ICU (Barbar, NEJM) 2018 Early (12h) vs delayed initiation in septic shock AKI No mortality difference; ~49% of delayed group recovered without RRT
STARRT-AKI (NEJM) 2020 Accelerated vs standard RRT initiation in AKI No 90-day mortality or RRT dependence benefit from early initiation

Viva Questions

1. What are the indications for starting RRT in the ICU and how do you decide when to begin?

Indications fall into absolute (AEIOU criteria) and clinical judgement categories. Absolute: refractory acidosis (pH <7.1), hyperkalaemia refractory to medical treatment, intoxication with dialysable toxin, volume overload refractory to diuretics, symptomatic uraemia (encephalopathy, pericarditis, uraemic bleeding). Clinical judgement: AKI that is clearly progressive with a deteriorating clinical trajectory, and in which spontaneous recovery is unlikely within a timeframe that would allow clinical stabilisation. The timing trials (IDEAL-ICU, STARRT-AKI) both show no benefit from early initiation — approximately half of delayed patients recovered spontaneously without ever needing RRT. This argues against early initiation based solely on AKI staging. I would consider starting when AEIOU criteria develop, when the trajectory makes spontaneous recovery very unlikely, or when the patient is clinically deteriorating due to AKI despite maximal medical management.

2. Why is regional citrate anticoagulation preferred for CRRT and what are the risks?

Regional citrate anticoagulation (RCA) is preferred by KDIGO and ESICM guidelines because it achieves effective anticoagulation within the CRRT circuit without systemic anticoagulation effects — reducing bleeding risk and allowing CRRT use in patients with coagulopathy or recent surgery. Mechanistically, citrate chelates ionised calcium within the circuit, impairing the calcium-dependent coagulation cascade. Systemically, citrate is metabolised to bicarbonate in the liver and peripheral tissues, and supplemental calcium is infused to restore ionised calcium. The principal risk is citrate accumulation, which occurs when hepatic and peripheral citrate metabolism is severely impaired (e.g. fulminant hepatic failure). Accumulated citrate forms calcium-citrate complexes, causing low ionised calcium despite normal total calcium. The diagnostic marker is a total:ionised calcium ratio >2.5. Treatment involves reducing the citrate infusion rate; if severe liver failure prevents metabolism, switch to systemic UFH or use heparin-free CRRT with increased flush protocols.

3. A patient on CRRT develops a total calcium of 2.8 mmol/L but an ionised calcium of 0.82 mmol/L. What does this mean and what do you do?

This discordance — high total calcium but low ionised calcium — is the hallmark of citrate accumulation. Total calcium is elevated because calcium is bound to accumulated citrate (forming calcium-citrate complexes measured in the total calcium assay), while ionised (biologically active) calcium is low because it is chelated. The total:ionised calcium ratio is 2.8 / 0.82 = 3.4 (normal <2.5), confirming accumulation. This indicates that the liver (and peripheral tissues) cannot metabolise the citrate infusion rate fast enough — most likely due to hepatic dysfunction (sepsis-related, drug-induced, or intrinsic liver disease). Management: reduce the citrate infusion rate or temporarily switch to citrate-free CRRT; infuse calcium to correct the ionised calcium until the situation resolves; reassess hepatic function; consider switching to systemic UFH anticoagulation or heparin-free running if liver failure is severe and persistent. Monitor acid-base carefully — citrate accumulation is also associated with metabolic alkalosis.

4. What dose of CRRT should you prescribe and why does delivered dose differ from prescribed dose?

Based on the RENAL and ATN trials, there is no benefit from doses above 20–25 mL/kg/h delivered effluent. To reliably achieve 20–25 mL/kg/h delivered dose, I would prescribe 25–30 mL/kg/h, because CRRT circuits are not running 100% of the time. Filter clotting requiring circuit changes, alarms, patient transport (procedures, imaging, theatre), nursing repositioning, and routine nursing care collectively result in approximately 20–30% downtime. If I prescribe 25 mL/kg/h but the circuit is only running 70% of the time, the patient actually receives ~17.5 mL/kg/h — below the target therapeutic dose. Prescribing 25–30 mL/kg/h accounts for this and ensures adequate delivered dose. Actual effluent volumes should be reviewed daily to confirm the target is being achieved, and adjusted if downtime is higher than expected (e.g. frequently clotting filters suggest anticoagulation needs review; frequent alarms may indicate access problems).