Drug clearance and dosing in liver and renal impairment

At a glance

• No blood test measures drug clearance directly. You infer it from creatinine for the kidney and from synthetic and excretory function (albumin and INR; bilirubin) plus encephalopathy and ascites for the liver.
• Kidney: decide whether function is stable, which estimate the drug's dosing used (Cockcroft-Gault or eGFR), and whether creatinine is misleading.
• Liver: transaminases tell you about injury, not capacity. Dosing tables use Child-Pugh class.
• Loading doses depend on volume of distribution and rarely change. Maintenance doses follow clearance.
• The pharmacokinetic principles (Dettli fraction, extraction ratio, dialysability) are in kidney impairment pharmacokinetics and liver disease and critical illness. This note is about reading the tests.

Renal function: what each number tells you

Measure	What it estimates	Use for dosing when	Main errors
Serum creatinine	Filtration, scaled by muscle mass	Never alone	Low in frailty, amputation, cirrhosis; raised by trimethoprim, cimetidine, dolutegravir (blocked secretion)
eGFR (CKD-EPI)	GFR for a 1.73 m² adult, mL/min/1.73 m²	Most newer drugs; metformin and SGLT2 inhibitors in Australian guidance	Assumes steady state and average muscle mass
Cockcroft-Gault CrCl	Creatinine clearance, mL/min	DOACs, aminoglycosides, enoxaparin, many older PIs	Weight choice in obesity; overestimates GFR because creatinine is secreted
Cystatin C (eGFRcr-cys)	GFR, independent of muscle	Low muscle mass, extremes of size, high-stakes narrow-margin drugs	Raised by steroids, thyroid disease, inflammation
24-hour urine CrCl	Creatinine clearance	Rarely; collection errors are common	Overestimates GFR by 10% to 20%
Measured GFR (nuclear medicine)	True GFR	Carboplatin, some chemotherapy, donor work-up	Cost and availability

Australian laboratories report CKD-EPI eGFR (KHA 2024), historically the 2009 equation; check which version your laboratory now uses. KDIGO 2024 recommends a race-free equation such as CKD-EPI 2021, which gives higher values in older Australians (Bongetti, MJA 2025). Detail is in measuring renal function.

Reading it for a prescription, step by step

1. Is function stable? Compare with previous results. A creatinine that has risen in the last 48 hours is not at steady state, and every equation overestimates GFR.
2. What did the product information use? Most DOAC, low molecular weight heparin and aminoglycoside thresholds are in Cockcroft-Gault CrCl.
3. Is the patient an average size? For eGFR, de-index: eGFR × BSA / 1.73. For Cockcroft-Gault in obesity, use the weight that TG or the AMH table specifies (usually lean or adjusted weight), not actual weight.
4. Is creatinine believable? Think about muscle mass and secretion blockers. Add cystatin C when the answer changes the dose.
5. Is there a drug level? Where TDM exists (gentamicin, vancomycin, lithium, digoxin), the level overrides any formula.

Worked example 1: the frail older man

An 88-year-old man weighs 52 kg and is 165 cm tall. Creatinine is 70 µmol/L and the laboratory reports eGFR 80 mL/min/1.73 m². He has new AF and is to start rivaroxaban.

• Cockcroft-Gault: (140 − 88) × 52 / (0.814 × 70) = 2,704 / 57 = 47 mL/min.
• De-indexed eGFR: 80 × 1.54 / 1.73 = about 71 mL/min.
• The Australian rivaroxaban PI gives 15 mg daily for CrCl 15 to 49 mL/min and 20 mg above that. The eGFR suggests 20 mg; Cockcroft-Gault correctly gives 15 mg.
• His low muscle mass means even 47 mL/min may be generous. A cystatin C-based eGFR is reasonable before starting.

Pitfall: a "normal" creatinine of 70 µmol/L in a 52 kg 88-year-old is not normal kidney function.

Worked example 2: AKI on treatment-dose enoxaparin

A 64-year-old woman on enoxaparin 1 mg/kg twice daily for pulmonary embolism becomes septic and oliguric. Creatinine rises from 80 to 150 µmol/L in 24 hours; eGFR is reported as 38.

• Creatinine is still climbing. Her real GFR is probably under 15 mL/min, so the eGFR of 38 is meaningless.
• Enoxaparin accumulates below CrCl 30 mL/min. Switch to an IV unfractionated heparin infusion, which is not renally cleared and can be stopped quickly.
• Hold metformin and nephrotoxins; review gabapentinoids, opioids and antibiotics daily. Keep full antibiotic loading doses. See AKI.

Worked example 3: obesity and actual body weight

A 46-year-old man weighs 130 kg and is 180 cm tall. Creatinine is 90 µmol/L; eGFR is reported as over 90.

• Cockcroft-Gault with actual weight: (140 − 46) × 130 / (0.814 × 90) = 167 mL/min.
• With ideal body weight (50 + 0.9 × (180 − 152) = 75 kg): 96 mL/min.
• Using 167 mL/min for an aminoglycoside or vancomycin maintenance dose overshoots. Start from TG's weight-adjusted dosing and let levels guide the rest.

Who is most often misjudged

• Older people: falling muscle mass hides falling GFR, and polypharmacy multiplies the harm (deprescribing).
• Aboriginal and Torres Strait Islander people have higher rates of CKD, often from a younger age and in remote communities where repeat testing is harder. Check kidney function before starting renally cleared drugs and build monitoring into the plan with the local Aboriginal Community Controlled Health Service.
• Pregnancy: GFR rises by about 50%, so creatinine falls and renally cleared drugs (enoxaparin, lithium) are cleared faster.

Liver function: what the tests can and can't say

Test	Reflects	Relevance to clearance
ALT, AST	Hepatocyte injury	Poor. ALT of 1,000 in acute hepatitis may leave clearance near normal; cirrhosis can have a normal ALT
ALP, GGT	Cholestasis, enzyme induction	Biliary-excreted drugs (some chemotherapy) need adjustment in obstruction
Bilirubin	Uptake, conjugation, excretion	Used for chemotherapy dose bands (eviQ)
Albumin	Synthesis (also low in inflammation and nephrotic syndrome)	Free fraction of phenytoin and valproate rises
INR	Synthesis of clotting factors	Also raised by vitamin K deficiency and warfarin

Child-Pugh score

Item	1 point	2 points	3 points
Bilirubin (µmol/L)	Under 34	34 to 50	Over 50
Albumin (g/L)	Over 35	28 to 35	Under 28
INR	Under 1.7	1.7 to 2.3	Over 2.3
Ascites	None	Mild or diuretic-controlled	Moderate to severe or refractory
Encephalopathy	None	Grade 1 to 2	Grade 3 to 4

Class A 5 to 6, B 7 to 9, C 10 to 15. Product information hepatic impairment studies are classed this way. MELD 3.0 predicts transplant-list mortality and is not a dosing tool.

Worked example 4: AF in cirrhosis

A 58-year-old man with alcohol-related cirrhosis has bilirubin 45 µmol/L (2), albumin 30 g/L (2), INR 1.6 (1), mild ascites on spironolactone (2) and no encephalopathy (1). ALT is 140 U/L. Score 8: Child-Pugh B.

• Rivaroxaban is avoided. The Australian PI contraindicates it in hepatic disease with coagulopathy and clinically relevant bleeding risk, including Child-Pugh B and C. It allows Child-Pugh B without coagulopathy, but exposure rises markedly in B and the EHRA guide advises against it there.
• The trial exclusions don't separate the others. RE-LY (dabigatran), ARISTOTLE (apixaban) and ENGAGE AF (edoxaban) all excluded ALT over 2 times the upper limit, and the apixaban and edoxaban trials also excluded bilirubin of 1.5 times or more. He fails both, so whichever DOAC he gets is used off the trial evidence.
• Apixaban with caution is the usual choice in Child-Pugh B. It has hepatic pharmacokinetic and observational data in cirrhosis and is the least renally cleared DOAC (about 27%), which matters when hepatorenal physiology or diuretics can drop the GFR quickly.
• Dabigatran is about 80% renally cleared and its label says use is "not recommended" with liver enzymes over 2 times the upper limit (his ALT is 3.5 times). Edoxaban is used with caution in B, like apixaban. No DOAC in Child-Pugh C.
• The apixaban, edoxaban and rivaroxaban labels all contraindicate hepatic disease with coagulopathy and clinically relevant bleeding risk. Before starting, treat varices (non-selective beta-blocker or banding) and agree with hepatology that bleeding risk is acceptable.
• Warfarin is hard to monitor when the baseline INR is already raised by liver disease.

Other Child-Pugh B choices follow the same logic:

Need	Choice	Avoid or reduce
Analgesia	Paracetamol, often capped at 2 to 3 g daily	NSAIDs (renal failure, bleeding); codeine (unreliable activation)
Opioid	Low dose, longer interval; hydromorphone or fentanyl	Usual oral morphine or oxycodone doses (shunting raises exposure)
Benzodiazepine	Oxazepam or lorazepam	Diazepam (half-life more than doubles)
Steroid	Prednisolone	Prednisone (needs hepatic activation)
Diabetes	Insulin	Metformin in decompensated disease

Clinical care is in cirrhosis.

Pitfall: reading normal transaminases as "liver fine" in a patient with low albumin, a raised INR and ascites.

When the result changes management

• Dose or interval of renally cleared drugs: DOACs, enoxaparin, gabapentinoids, aminoglycosides, vancomycin, acyclovir, baclofen, metformin, sotalol, lithium.
• Drug choice: changing from morphine to fentanyl in severe CKD, from diazepam to oxazepam in Child-Pugh B or C, from rivaroxaban to apixaban in Child-Pugh B.
• Monitoring: start TDM, or switch to a drug that can be monitored or reversed.
• Chemotherapy: carboplatin dosing by the Calvert formula (dose = target AUC × (GFR + 25)), using measured GFR or de-indexed eGFR; bilirubin bands for anthracyclines and taxanes.

Pitfalls

• Dosing from eGFR when the drug label uses Cockcroft-Gault, especially in small older women and frail men.
• Using actual body weight in Cockcroft-Gault for a 130 kg patient, which overestimates clearance.
• Trusting any equation in evolving AKI.
• Reducing an antibiotic loading dose because creatinine is high.
• Reading albumin as purely hepatic: sepsis and nephrotic syndrome lower it too.
• Using MELD, or an isolated ALT, to decide a dose.
• Forgetting that a hepatically metabolised drug can have a renally cleared active metabolite (morphine, oxypurinol), covered in kidney impairment pharmacokinetics.