Chapter 7 — Pharmacokinetics in CRRT

📘 Scope note: This chapter is the pure academic framework. No calculators, code, or interactive tooling — by design. It gives the reasoning structure required to interpret any dosing recommendation you encounter, rather than a lookup table to apply without understanding.

CRRT dosing errors run in both directions, and both are dangerous: under-dosing antimicrobials in the septic patient with the highest bacterial burden, and accumulating drug in the patient with no renal reserve. The framework below is what lets you reason about a drug you have never dosed on CRRT before.

7.1 The Central Equation

Everything reduces to one relationship:

CL_total = CL_renal + CL_non-renal + CL_CRRT

In severe AKI, CL_renal approaches zero. CL_CRRT can therefore become the dominant elimination pathway — which is why a drug's "renal failure dose" from a formulary is frequently wrong for a patient on CRRT. That dose assumes no extracorporeal clearance.

💡 The most common clinical error in this chapter's territory: applying an anuric-patient dose reduction to a patient receiving 25 mL/kg/h of effluent. For a small, unbound, low-Vd drug, CRRT may be providing clearance comparable to a meaningful fraction of native renal function. The dose reduction can produce sub-therapeutic concentrations in exactly the patients who can least afford them.

7.2 The Four Drug Properties That Determine CRRT Clearance

1. Protein binding — the dominant variable

Only the unbound fraction can cross the membrane. This is the single most important determinant, and follows directly from the sieving coefficient (Ch. 1):

Sc ≈ unbound fraction (fu) for small molecules whose size does not itself restrict passage.

⚠️ Protein binding is not a fixed drug constant in critical illness. Hypoalbuminaemia, hyperbilirubinaemia, uraemia and acidosis all displace drug from albumin, raising the unbound fraction and therefore raising CRRT clearance — sometimes substantially, and unpredictably. Published sieving coefficients derived from healthy volunteers or stable patients may materially underestimate clearance in the hypoalbuminaemic septic patient. The magnitude of this effect in critical illness is not well quantified for most drugs.

2. Volume of distribution (Vd)

Vd determines how much of the drug is even available to the circuit at any moment.

• Low Vd (roughly <0.7 L/kg) — drug largely confined to plasma and extracellular fluid. A high proportion of the body burden passes through the filter. CRRT removal is efficient. Examples: aminoglycosides, β-lactams, vancomycin.
• High Vd (>1–2 L/kg) — drug sequestered in tissue. Plasma holds only a small fraction of the total body burden, so even excellent plasma clearance removes little. CRRT removal is inefficient. Examples: amiodarone, digoxin, many antidepressants.

💡 This is why CRRT is a poor rescue for many overdoses even when the drug is small and unbound — if it is not in the plasma, the filter cannot reach it. Vd is also the reason a drug can be efficiently cleared during the session yet rebound afterwards as tissue stores redistribute.

Critical illness alters Vd. Sepsis with capillary leak and aggressive fluid resuscitation expands the Vd of hydrophilic drugs, lowering peak concentrations. For concentration-dependent agents this matters most at the loading dose.

💡 Loading doses are generally not reduced for renal failure or for CRRT. The loading dose is governed by Vd, not by clearance. In the oedematous septic patient the loading dose may need to be larger, not smaller. Reduce the maintenance dose or extend the interval — not the load.

3. Molecular weight

Most antimicrobials and ICU drugs are well under 1,500 Da and pass conventional CRRT membranes freely. Molecular weight is therefore rarely the limiting factor for small-molecule drugs — protein binding and Vd dominate.

MW becomes relevant at the extremes: very large molecules (monoclonal antibodies, most peptide hormones) are effectively not cleared by conventional membranes.

4. Hydrophilicity vs lipophilicity

A useful shorthand that bundles several of the above:

💡 A rapid triage rule: hydrophilic, renally cleared, low protein binding, low Vd → assume clinically significant CRRT clearance and verify the dose. Lipophilic, hepatically cleared, highly bound, high Vd → CRRT clearance is likely negligible; dose for hepatic function instead.

7.3 How the Modality Changes Clearance

Drug clearance mirrors the solute-clearance physics of Chapter 1 exactly.

Three modality-related principles:

1. Effluent rate is the best single predictor of drug clearance for small, unbound drugs — which means a dose change (Ch. 4) is also a drug-dosing change. Increasing effluent from 20 to 35 mL/kg/h increases antimicrobial clearance proportionally.
2. Pre-dilution reduces drug clearance by the same correction factor that reduces solute clearance (Ch. 4). Heavy pre-dilution means less drug removed than the effluent number implies.
3. Adsorption can remove drug invisibly. Aminoglycosides and certain other agents adsorb to some membrane polymers. This clearance appears in no effluent-based calculation — it is front-loaded and decays as sites saturate (Ch. 1).

7.4 Applying PK/PD Targets

The dosing strategy depends on which pharmacodynamic index drives efficacy:

💡 The dose-versus-interval decision is a pharmacodynamic question, not a renal one. Two drugs cleared identically by the filter can require opposite adjustments depending on whether their killing is time- or concentration-dependent.

7.5 Therapeutic Drug Monitoring

⚠️ Declared gap — the honest headline of this chapter.

Published CRRT dosing recommendations carry substantial uncertainty. They are derived predominantly from small pharmacokinetic studies, often with heterogeneous modalities, effluent doses, dilution modes, membranes and patient populations, and frequently from patients less critically ill than those in front of you. There is no large randomised trial establishing that any particular CRRT dosing regimen improves clinical outcomes for any antimicrobial.

Consequently: where therapeutic drug monitoring is available, use it. It resolves in one measurement what the framework above can only estimate. Where TDM is unavailable, reason explicitly through protein binding, Vd, effluent rate, dilution mode and PD index — and document the reasoning, because it will need revisiting when the prescription changes.

Practical monitoring notes:

• Any change to the CRRT prescription is a change to drug clearance. Increasing effluent dose, switching modality, or altering the pre/post-dilution split all alter drug exposure. Reassess dosing whenever the circuit prescription changes.
• Circuit downtime alters exposure too. A patient off the circuit for eight hours a day is not receiving the clearance the prescription implies.
• Beware sampling errors — levels drawn from the circuit rather than the patient, or drawn without a defined relationship to the dose, are uninterpretable.

7.6 A Structured Reasoning Sequence

When facing an unfamiliar drug in a patient on CRRT, work through in this order:

1. What is the protein binding? Highly bound → CRRT clearance likely small. Poorly bound → proceed.
2. What is the Vd? High → CRRT removes little of the body burden regardless. Low → CRRT clearance is likely significant.
3. What fraction of total clearance was renal in health? Predominantly renal → CRRT now matters greatly. Predominantly hepatic → dose for hepatic function.
4. What is the effluent rate, modality and dilution mode? This quantifies the extracorporeal contribution.
5. What is the PD index? Determines how to adjust — dose size versus interval versus infusion strategy.
6. Is the loading dose separately considered? It usually should not be reduced, and in capillary leak may need increasing.
7. Is TDM available? If yes, this supersedes the estimate.
8. Has anything about the circuit changed? If yes, return to step 4.

7.7 Chapter Summary

1. CL_CRRT can be the dominant elimination pathway — formulary "renal failure" doses that assume anuria will under-dose these patients.
2. Only the unbound fraction sieves, and critical illness raises the unbound fraction unpredictably.
3. Vd determines availability to the circuit; high-Vd drugs resist removal no matter how good the clearance.
4. Loading doses are governed by Vd, not clearance — do not reduce them, and consider increasing them in capillary leak.
5. Effluent rate, modality and pre-dilution fraction all change drug clearance — a CRRT prescription change is a drug prescription change.
6. The evidence base for CRRT drug dosing is thin. Use TDM where available; reason explicitly and document where it is not.