CRRT prescription at a glance
100 kg, septic, anuric, K 7, on inotropes. Start now, run continuous, prescribe a deliverable dose, keep the filter alive.
- Indication Refractory K: start now, treat in parallel.
- Modality: CVVHDF Continuous for BP, diffusion for potassium.
- Circuit life Citrate first; filtration fraction when clotting.
-
Stop 1 · Initiation
Indications and timing of initiation
Refractory K >6.0: start now. Acidosis alone: bicarbonate first.
1Potassium
Therapy-resistant K >6.0: start immediately (ICU Rec 1.1).
2Acidosis
pH <7.20 alone: IV bicarbonate first, unless another urgent indication.
3Volume or uraemia
Refractory overload or uraemic complications (KDIGO 2012).
Temporising therapy
Calcium, insulin-dextrose, salbutamol while access is placed.
Deferred initiation
Deferred over early start (KDIGO 2026 draft, 1C); start for refractory complications.
STARRT-AKIDeath 43.9% vs 43.7%; more RRT dependence if accelerated.
AKIKINo mortality difference; 49% of delayed never needed RRT.
IDEAL-ICUSeptic shock, 12 h vs 48 h: stopped for futility.
BICARICU-2Bicarbonate to pH 7.30: mortality unchanged; less KRT (secondary).
Memory hook“Refractory potassium starts the pump; acid alone gets the bicarbonate.” AKIKI 2: do not delay beyond standard triggers.
-
Stop 2 · Set-up
Vascular access and modality selection
Right IJ, continuous, diffusive high dialysate flow: CVVHDF.
- Access · KDIGO 2012 Right IJ first
- Non-tunnelled dual-lumen, ultrasound-guided; chest film before use.
- Continuous vs intermittent Unstable → CRRT
- Instability, high inotropes, raised ICP, liver failure (KDIGO 2026 draft: CRRT or acute PD, 2C).
- IRRT: stable, rapid removal. SLED / PIRRT: acceptable if unstable.
CRRT modalities
UF onlySCUF
Water only; no effective solute clearance.
ConvectionCVVH
Small and middle molecules; needs replacement fluid.
DiffusionCVVHD
Small solutes; uses dialysate.
BothCVVHDF
Diffusion + convection: typical ICU choice.
- Diffusion vs convection: similar outcomes; severe hyperkalaemia → high dialysate flow.
- Pre-dilution: less clearance per litre (dose higher), less clotting.
- Buffer: bicarbonate, not lactate, in shock or liver failure.
Exam trapIHD on maximal inotropes. Hemodiafe 2006: no difference; ICRAKI 2026: IHD not noninferior, higher mortality.
-
Stop 3 · Dose
Dose prescription
Prescribe 25–30 mL/kg/h to deliver 20–25 despite downtime.
Effluentthe dose= Dialysateflow+ Replacementflow+ Net UFfluid goalDose = 24 h effluent ÷ weight ÷ 24. Actual weight; BMI ≥30: ideal or adjusted (KDIGO 2026 draft).
- Prescribe 25–30 mL/kg/h
- About 2,500–3,000 mL/h at 100 kg.
- Deliver 20–25 mL/kg/h
- ICU guideline, KDIGO 2012 and 2026 draft. No HVHF in sepsis.
- Blood flow QB 150–200 mL/min
- Clinical Handbook range 100–250 mL/min.
- Net UF Modest, individualised
- 1.01–1.75 mL/kg/h appeared safe; >1.75 linked to mortality (observational).
Worked example · account for downtime
- 100 kg: 3,000 mL/h × 20 h = 60 L/day = 25 mL/kg/h delivered.
- 75 kg: deliver 1,500–1,875 mL/h; prescribe ~2,250 mL/h.
- Delivered dose low? Fix anticoagulation and circuit life first, then raise the rate.
Memory hook25–30prescribe 20–25deliver 150–200QB ~3,000mL/h · 100 kgSlightly above target, no higher.
Exam trap35–40 mL/kg/h as target. RENAL (40 vs 25): 90-day mortality 44.7% both. ATN: no benefit, more hypotension.
-
Stop 4 · Anticoagulation
Anticoagulation
Regional citrate if no contraindication: longer filter life, fewer bleeds.
Guideline recommendations
KDIGO AKI 2012Suggests citrate over heparin (2B).
ICU guideline 2026Citrate = heparin for outcomes; citrate if bleeding risk; UFH over LMWH.
KDIGO 2026 draftRecommends citrate over heparin (1B); UFH, epoprostenol, nafamostat alternatives.
- Citrate set-up · mind the units Fixed to blood flow
- ACD-A mL/h ≈ 1.5 × QB in mL/min: QB 150 → 225 mL/h.
- Never 150% of QB in mL/h. Change citrate in proportion to QB.
- Citrate monitoring Two calcium targets
- Post-filter iCa <0.35 mmol/L; separate calcium to normal systemic iCa.
- iCa 4-hourly, then 6-hourly; total Ca ≥12-hourly; acid–base, lactate.
Excess buffer
Ratio <2.5 + alkalosis: review fluid buffer and citrate rate.
Citrate accumulation
Ratio >2.5, anion-gap acidosis, rising Ca need: treat hypocalcaemia, reduce citrate.
Persists: stop citrate → bicarbonate CRRT, heparin or none.
- Avoid citrate: progressive lactic acidosis + shock + severe liver failure. Either alone: monitor.
- Other citrate issues: hypo/hypercalcaemia, hypernatraemia, hypomagnesaemia.
- Heparin: trials 2,000–5,000 IU, then 7–10 IU/kg/h; Handbook 500–1,000 IU, then ~500 IU/h.
- Heparin targets: aPTT q6h (RICH 45–60 s) or anti-Xa 0.3–0.6 IU/mL.
- Acute HIT: argatroban (bivalirudin, danaparoid); citrate in subacute/remote HIT.
- Bleeding: citrate; no anticoagulation only if unavoidable (pre-dilution, higher QB).
Memory hook“Over 2.5, citrate is piling up. Under 2.5 with alkalosis, too much buffer.”
Exam trapCalling accumulation an alkalosis. It is anion-gap acidosis, ratio >2.5. No shared ACT target.
-
Stop 5 · Troubleshooting
Filtration fraction and circuit troubleshooting
High FF: haemoconcentration, protein deposition, circuit clotting.
FFfraction= Total UFreplacement + net UF÷ Plasma flowQB × (1 − Hct)Dialysate not counted. Keep FF <20–25% (<20% if no anticoagulation).
Worked example
QB 200, Hct 0.30 → plasma flow 140 mL/min; UF 25 mL/min → FF ≈ 18%.
Limitation: ignores pre-filter replacement.
1 · Increase blood flow
Lowers FF; adjust citrate to the new QB.
2 · Pre-dilution
Shift replacement fluid pre-filter.
3 · Optimise anticoagulation
Post-filter iCa above target: citrate too low. Clotted: stop, change circuit, check coagulation.
Circuit pressures
- Rising pre-filter + TMP / pressure dropFilter clotting or fouling.
- Very negative access pressureCatheter position, kinking or poor flow.
- High return pressureClotted return lumen or kinked line.
- Blood-leak alarmMembrane rupture: change the circuit at once.
EmergencyK still high on CRRT (e.g. 7.4): ECG, repeat K, IV calcium for ECG changes, temporising therapy; check circuit and flows; raise dialysate flow. No sustained dose escalation.
Memory hook“Faster blood, dilute before, check the anticoagulant.”
-
Stop 6 · Monitoring
Drug dosing, electrolytes and nutrition
Antibiotics, phosphate, magnesium and amino acids leave in effluent.
- Antibiotics Full first dose
- Full, short infusion; maintenance by residual kidney + machine clearance.
- TDM: beta-lactams, vancomycin, aminoglycosides.
- Electrolytes K, PO4, Mg every 6 h
- Hypophosphataemia >50% (up to 65%): phosphate solution (~1 mmol/L) or supplement.
- Nutrition Amino acids leak
- 0.2 g/L effluent → protein 1.2–1.5 g/kg/day (1.7 if hypercatabolic).
- Daily checks The rest of the board
- ABG, Hct (FF), aPTT if heparin, fluid balance, delivered dose, cultures.
Red flagsSevere hypotension, bradycardia, bleeding: stop and reassess. Air embolism, hypothermia. Antibiotic under-dosing.
-
Stop 7 · Recovery
Discontinuation
Overload corrected, urine recovering, no new indication: trial off.
- Accepted criteria
- Urine >500 mL/day without diuretics, or ~2.4 L/day with.
- ICU guideline 2026
- Urine 300–600 mL/day without diuretics; K >5.5 or pH <7.3 predict failure.
- KDIGO 2026 draft
- >450 (no diuretics) or >2300 mL/24 h (diuretics) + 2-h CrCl ≥23 mL/min: predictors, not a rule.
Exam trapFalling creatinine on active CRRT is not recovery.
Memory hook“Dry first, urine back, no new reason: trial off.”
Rapid recall
What starts CRRT now in this patient?
Therapy-resistant K >6.0. Acidosis alone (pH <7.20) gets IV bicarbonate first.
Why CVVHDF, and why continuous?
Unstable → continuous; hyperkalaemia → diffusion, high dialysate flow.
Dose for a 100 kg patient?
Prescribe 25–30 mL/kg/h (~3,000 mL/h) to deliver 20–25; QB 150–200 mL/min.
What makes up the effluent?
Dialysate + replacement + net UF.
The FF equation and target?
FF = total UF ÷ plasma flow; plasma flow = QB × (1 − Hct). Keep <20–25%.
FF 35–40%, filter clotting?
Raise QB, shift replacement pre-filter, review anticoagulation. Clotted: hold, change circuit.
Citrate accumulation vs excess buffer?
Accumulation: ratio >2.5, anion-gap acidosis, rising Ca need. Excess buffer: alkalosis, ratio <2.5.
Dose and timing trials?
RENAL, ATN: more dose, no benefit. STARRT-AKI, AKIKI, IDEAL-ICU: earlier start, no benefit.
When do you trial off?
Overload corrected, urine >500 mL/day, no new indication.
Next tabs: detail, spoken answer, viva questions.
Sources: ICU RRT guideline 2026, KDIGO 2012 and 2026 draft, Clinical Handbook of Nephrology 2024.