Therapeutic Drug Monitoring

Therapeutic drug monitoring measures the concentration of a drug in blood to keep it in the range where benefit is greatest and toxicity least. It applies to drugs with a narrow therapeutic window, large variation between patients in absorption and clearance, a concentration dependent effect, and a level that is easier to read than the clinical response. Drug levels are measured by immunoassay or chromatography with mass spectrometry, in serum, plasma or whole blood.

A level is useful only when interpreted against dose, the time of the last dose, time on treatment, renal and hepatic function, interacting drugs, adherence and clinical state. Levels are taken at steady state (after four to five half lives), usually as a trough just before the next dose, unless the drug needs a peak or timed level.

Drug toxicity accounts for a significant share of unplanned admissions in older adults, and digoxin, lithium, aminoglycosides, anticonvulsants and immunosuppressants are leading causes among monitored drugs.

Indications. Initiation or dose change of a drug with a narrow range, suspected toxicity, suspected non-adherence, treatment failure, change in renal or hepatic function, pregnancy, interacting drug started or stopped, change of formulation or brand, and unclear history after overdose.

Specimen. Serum or heparin plasma for most drugs. Whole blood in EDTA for tacrolimus, ciclosporin and sirolimus. Lithium is measured in plain serum, since lithium heparin tubes cause false high results. Gel separator tubes can absorb lipophilic drugs. Record the time of last dose, the time of sampling, and dose history.

Ranges are indicative and vary with the indication, assay and local practice. Conversion factors are given for common drugs.

Digoxin. Range: 0.5 to 0.9 ng/mL (0.6 to 1.2 nmol/L) in heart failure, and up to 2.0 ng/mL (2.6 nmol/L) in atrial fibrillation. Sample at least 6 to 12 hours after the last dose. Toxicity over 2.0 ng/mL, and also at lower levels with hypokalaemia, hypomagnesaemia, hypercalcaemia, hypothyroidism, kidney failure, hypoxia and drug interactions (amiodarone, verapamil, clarithromycin, quinidine, ciclosporin). Toxicity: nausea, visual disturbance, confusion, bradycardia, heart block, ventricular arrhythmia.

After antidigoxin Fab antibody fragments, total digoxin levels are falsely high, because bound drug is measured. Digoxin like immunoreactive substances in neonates, pregnancy and kidney or liver disease can give false positive results.

Lithium. Range: 0.6 to 0.8 mmol/L for maintenance (0.4 to 0.8 in older adults), and 0.8 to 1.0 mmol/L in acute mania. Sample 12 hours after the last dose. Toxicity: over 1.5 mmol/L (tremor, vomiting, diarrhoea, ataxia), over 2.5 mmol/L (confusion, seizures, coma). Chronic toxicity can occur at lower levels with severe symptoms.

Levels rise with dehydration, kidney disease, non-steroidal anti-inflammatory drugs, ACE inhibitors, angiotensin receptor blockers and thiazides. Dialysis is considered for a level over 4.0 mmol/L, or over 2.5 mmol/L with symptoms or kidney impairment, and levels may rebound after dialysis.

Phenytoin. Total range: 10 to 20 mg/L (40 to 80 micromol/L). Free range: 1 to 2 mg/L. Kinetics are saturable, so small dose rises can give large level rises. In low albumin, kidney failure, pregnancy and critical illness the total level underestimates the active free fraction, and a free level or a corrected value is required. Toxicity: nystagmus over 20 mg/L, ataxia over 30 mg/L, confusion and coma over 40 mg/L.

Valproate. Range: 50 to 100 mg/L (350 to 700 micromol/L), with a trough sample. The free fraction rises at high levels and in hypoalbuminaemia, so total levels correlate poorly with effect. Toxicity: tremor, sedation, hyperammonaemia, pancreatitis and hepatotoxicity.

Carbamazepine. Range: 4 to 12 mg/L (17 to 50 micromol/L). Autoinduction over 2 to 4 weeks lowers levels after initiation. Toxicity over 15 mg/L (diplopia, ataxia, hyponatraemia).

Vancomycin. Trough range: 10 to 20 mg/L, with area-under-the-curve to minimum inhibitory concentration ratio of 400 to 600 as the preferred measure of exposure. Persistently high troughs raise the risk of nephrotoxicity. Sample before the fourth dose, or earlier in kidney impairment.

Aminoglycosides (gentamicin, amikacin). Gentamicin trough under 1 mg/L, and under 2 mg/L with multiple daily dosing. Peak 5 to 10 mg/L with multiple daily dosing, and 15 to 30 mg/L with once daily dosing. Persistently raised trough levels predict ototoxicity and nephrotoxicity, and require dose interval extension or a change of agent.

Tacrolimus and ciclosporin. Trough in whole blood, with the range depending on the organ, time after transplant and regimen. Typical tacrolimus trough: 5 to 15 ng/mL (higher early after transplant, lower later). Typical ciclosporin trough: 100 to 300 ng/mL. Toxicity: nephrotoxicity, tremor, hypertension, hyperkalaemia and hypomagnesaemia. Subtherapeutic levels risk rejection. Strong interactions arise with azole antifungals, macrolides, diltiazem, grapefruit and enzyme inducers.

Theophylline. Range: 10 to 20 mg/L (55 to 110 micromol/L), and 5 to 15 mg/L in neonates and older adults. Toxicity over 20 mg/L (tachycardia, vomiting, tremor), arrhythmia and seizures over 40 mg/L. Smoking, ciprofloxacin, heart failure, liver disease and viral infection alter clearance.

Timing of the sample.

• Trough: just before the next dose, for most oral drugs.
• Peak: aminoglycosides, and some drugs with concentration dependent killing.
• Timed levels: lithium 12 hours after dose, digoxin at least 6 hours after dose, and after loading doses 24 hours.
• Steady state: four to five half lives after starting or changing dose. Digoxin takes 7 days, phenytoin and carbamazepine 1 to 2 weeks, lithium 4 to 5 days.

Reasons for unexpected levels.

• Low or undetectable: non-adherence, missed doses, malabsorption, enzyme induction, drug interaction, sampling soon after a dose for a trough, rapid metabolisers, or increased clearance in pregnancy and critical illness.
• High: overdose, renal or hepatic impairment, enzyme inhibition, dehydration, reduced albumin (for total levels of binding drugs), sampling before distribution is complete, and a recent dose.
• Normal level with toxicity: free fraction raised, active metabolite, or end organ sensitivity (digoxin in hypokalaemia).
• Normal level without response: resistance, wrong diagnosis, or poor adherence earlier.

Adherence assessment. Undetectable levels in a patient who reports adherence, or very variable levels, suggest missed doses.

Critical values and findings needing immediate action. Triggers vary by laboratory.

• Digoxin over 2.5 ng/mL, or any level with bradycardia, heart block or ventricular arrhythmia.
• Lithium over 1.5 mmol/L, or any level with neurological signs.
• Phenytoin over 30 mg/L, or valproate over 150 mg/L.
• Theophylline over 20 mg/L with arrhythmia, vomiting or seizures.
• Gentamicin trough over 2 mg/L with falling urine output.
• Tacrolimus or ciclosporin well over range with rising creatinine, or undetectable level in a transplant recipient.

Interferences and pitfalls.

• Sampling at the wrong time is the commonest error, and a level taken soon after a dose gives a falsely high trough.
• Lithium heparin tubes give falsely high lithium results.
• Gel separator tubes and some plastic tubes absorb lipophilic drugs and give falsely low results.
• Haemolysis, lipaemia and bilirubin interfere with some immunoassays.
• Immunoassays cross-react with metabolites (tacrolimus, ciclosporin, digoxin), and mass spectrometry is more specific.
• Different assays and different brands of the same drug give different results.
• Total levels of protein bound drugs mislead in low albumin, kidney failure and pregnancy, and free levels are needed.
• A level in range does not exclude toxicity, and a level out of range does not mandate a dose change without clinical correlation.

Special groups.

• Neonates and infants: immature clearance and different binding, with a need for age specific ranges.
• Children: higher weight based clearance for many drugs, so doses and ranges differ.
• Pregnancy: clearance of lamotrigine, phenytoin, carbamazepine and lithium changes, with a need for more frequent levels and postpartum readjustment.
• Older adults: reduced renal clearance and polypharmacy increase risk.
• Obesity, burns and critical illness: altered volume of distribution and clearance alter aminoglycoside and vancomycin dosing.
• Regions: variation in enzyme genotype (CYP2C19, CYP2D6, CYP3A5) influences exposure.

Follow up tests. Urea, creatinine and eGFR, liver function tests, albumin, electrolytes including potassium, magnesium and calcium, thyroid function tests, ECG, full blood count, free drug levels, drug metabolite levels, ammonia (valproate), sodium (carbamazepine), audiometry (aminoglycosides), pharmacogenetic testing, and repeat level after dose adjustment at steady state.