Therapeutic Drug Monitoring (TDM)
Using measured drug concentrations to individualise dosing — indications, sampling timing, target ranges, and model-informed precision dosing (MIPD).
What is Therapeutic Drug Monitoring?
Therapeutic drug monitoring (TDM) is the clinical practice of measuring drug concentrations in biological fluids (usually plasma or serum) to optimise dosing for individual patients. TDM is not applied to all drugs — only those where:
- The drug has a narrow therapeutic index (toxic concentration close to therapeutic)
- There is a well-established concentration–effect relationship
- Dose alone is insufficient to predict exposure due to high PK variability
- The clinical endpoint is delayed or difficult to measure (seizure control, organ rejection)
TDM integrates pharmacokinetic (PK) and pharmacodynamic (PD) principles. The goal is to achieve a target exposure (AUC, Cmax, trough) associated with efficacy while avoiding toxicity. See also: pharmacokinetics basics and drug clearance.
When is TDM Indicated?
Vancomycin, aminoglycosides, digoxin, lithium, phenytoin, carbamazepine, cyclosporine, tacrolimus, methotrexate.
Renal impairment (vancomycin, aminoglycosides), hepatic failure (phenytoin, cyclosporine), hypoalbuminaemia, ARC (augmented renal clearance), morbid obesity.
Digoxin toxicity, phenytoin ataxia/nystagmus, lithium toxicity (tremor, confusion, polyuria), aminoglycoside nephrotoxicity/ototoxicity.
CYP inhibitor/inducer added to a narrow-TI drug — e.g., adding rifampicin to cyclosporine, or adding fluconazole to phenytoin.
Neonates and infants (rapidly changing PK), critically ill patients (ARC, third-spacing, altered protein binding), pregnancy (increased Vd, GFR).
Antiepileptics (phenytoin, valproate), antiretrovirals, antimalarials — undetectable or sub-therapeutic levels may indicate non-adherence.
Sampling Timing
The clinical utility of TDM depends critically on when the sample is drawn. An incorrectly timed sample is often worse than no sample.
Drawn immediately before the next scheduled dose. Reflects minimum steady-state concentration. Used for: vancomycin (trough-only, though AUC now preferred), tacrolimus, cyclosporine, phenytoin, digoxin, aminoglycosides (conventional dosing). Must be at steady state (4–5 × t½).
Drawn ~1 h after end of IV infusion, or ~1–2 h post-oral dose. Used for: aminoglycosides (once-daily — peak 20–30 mg/L for gram-negative efficacy). Avoid drawing during distribution phase (false high).
For vancomycin AUC/MIC targets: draw 1–2 h post-dose start (mid-infusion Cp1) and 4–6 h later (Cp2). Use a linear-log trapezoidal method or software (InsightRX, DoseMeRx) to estimate AUC. ASHP/IDSA/SIDP 2020 vancomycin guidelines strongly recommend this approach.
Always drawn exactly 12 h after the last dose, regardless of dosing interval. This standardised timing is essential for meaningful comparison to published therapeutic ranges (0.6–1.2 mEq/L).
TDM Reference Table
| Drug | Target range | Sample timing | Indication | Guideline |
|---|---|---|---|---|
| Vancomycin | AUC/MIC 400–600 mg·h/L | 2 samples post-dose (AUC method) | MRSA; renal impairment | ASHP/IDSA/SIDP 2020 |
| Gentamicin / Tobramycin | Trough <1 mg/L (conventional); Peak 20–30 mg/L (once-daily) | Pre-dose trough; 1 h post-dose peak | Gram-negative sepsis; endocarditis synergy | IDSA guidelines |
| Phenytoin | 10–20 mg/L total; 1–2 mg/L free | Trough (pre-dose) | Epilepsy; hypoalbuminaemia — use free level | Correct for albumin |
| Digoxin | 0.5–2 ng/mL (heart failure: 0.5–0.9) | ≥6 h post-dose (distribution phase complete) | AF rate control; HFrEF | AHA/ACC |
| Lithium | 0.6–1.2 mEq/L (acute 0.8–1.2; maintenance 0.6–0.8) | 12 h post-last dose (standardised) | Bipolar disorder | NICE, BAP |
| Tacrolimus | 5–15 ng/mL (organ-specific, protocol-driven) | Trough (pre-dose) | Solid organ transplant | KDIGO; institutional protocol |
| Methotrexate (HD) | <0.1 µmol/L at 72 h post-infusion | 24 h, 48 h, 72 h post-infusion | High-dose MTX (ALL, osteosarcoma, CNS lymphoma) | Rescue protocol trigger |
| Ciclosporin (Cyclosporine) | C₀ or C₂ (2 h post-dose) per organ/protocol | Trough (C₀) or 2 h (C₂) | Solid organ; autoimmune | Institutional |
Ranges are general guidance. Always refer to current institutional protocols and drug-specific guidelines. References: ASHP/IDSA vancomycin 2020; IDSA aminoglycoside guidance; EMA product labelling.
Bayesian Model-Informed Precision Dosing (MIPD)
Traditional TDM uses a single measured concentration and a nomogram or linear equation to adjust dose. Bayesian MIPD is more powerful: it starts with a population PK model (prior) and updates it using the patient's own concentration–time data (likelihood) to produce individualised PK parameter estimates (posterior).
Key advantages over traditional approaches:
- Works with samples taken at any time point — not restricted to steady state or specific pre/post-dose windows
- Integrates all available clinical covariates (age, weight, SCr, CYP genotype) via the PopPK model
- Provides uncertainty estimates around predicted concentrations
- Particularly valuable in critical care, neonates, and patients with rapidly changing organ function
Commercially available MIPD platforms include InsightRX, DoseMeRx, MwPharm++, and JPKD (Japan). The 2020 vancomycin consensus guidelines explicitly endorse Bayesian AUC-guided dosing. See also: Therapeutic Drug Monitoring journal for current research.
FAQ
What is therapeutic drug monitoring?
TDM is the clinical practice of measuring drug concentrations in plasma or serum to guide individual dose adjustments. It applies to drugs with narrow therapeutic index, high PK variability, or delayed clinical endpoints where concentration is a better surrogate than dose.
Which drugs require TDM?
Common TDM drugs: vancomycin (AUC/MIC 400–600), aminoglycosides (trough/peak), phenytoin (10–20 mg/L), digoxin (0.5–2 ng/mL), lithium (0.6–1.2 mEq/L), tacrolimus (5–15 ng/mL), cyclosporine, methotrexate (HD), and antiepileptics (carbamazepine, valproate).
When should a TDM sample be collected?
Sample timing is drug-specific: trough (pre-dose at steady state) for most drugs; 12 h standardised for lithium; AUC two-point (1–2 h and 4–6 h post-dose) for vancomycin. Always document exact dose time and sample time.
What is Bayesian TDM?
Bayesian MIPD combines population PK priors with individual measured concentrations (at any time point) to estimate patient-specific CL and Vd. This allows precise dose adjustment even from a single random level — especially useful in ICU and paediatric patients.
Why has vancomycin TDM shifted from trough to AUC monitoring?
The 2020 ASHP/IDSA/SIDP vancomycin guidelines showed that targeting AUC/MIC (400–600 mg·h/L) rather than trough alone (15–20 mg/L) reduces nephrotoxicity without sacrificing efficacy. Trough-only monitoring was associated with higher toxicity at concentrations needed to achieve adequate AUC.