Clinical pharmacokinetics and therapeutic drug monitoring

In short

Clinical pharmacokinetics applies the mathematics of absorption, distribution, metabolism and excretion to one individual patient so that a dose, an interval and a route can be chosen rationally rather than by habit. Therapeutic drug monitoring (TDM) is the clinical use of measured drug concentrations - interpreted with the patient's organ function, timing of the sample and clinical state - to individualise therapy for drugs with a narrow therapeutic index. Together they are the quantitative core of ward-level clinical pharmacy: digoxin, phenytoin, vancomycin, amikacin, lithium, carbamazepine, tacrolimus, cyclosporine, valproate and theophylline all need them. For Pharm.D students this is a heavily weighted unit of Clinical Pharmacy / Clinical Pharmacokinetics and Pharmacotherapeutics (PCI curriculum), and it is the single most asked area in interviews for clinical pharmacist, TDM-cell, medical-affairs, pharmacovigilance-PK and clinical-research posts at Indian hospitals (AIIMS, PGIMER, CMC Vellore, Apollo, Manipal, Amrita, NIMS) and CROs.

Explained

Core parameters and what each one actually controls

• Bioavailability (F): fraction of an administered dose reaching systemic circulation unchanged. F = 1 for IV; oral F is reduced by incomplete absorption and first-pass metabolism (propranolol, morphine, lignocaine have high first-pass).
• Volume of distribution (Vd): the apparent volume relating amount of drug in body to plasma concentration. It is a proportionality constant, not an anatomical space - digoxin's Vd is about 7 L/kg because it is bound in muscle, while gentamicin's is about 0.25 L/kg (extracellular water).
• Vd determines the LOADING dose. High Vd drugs need large loading doses; Vd rises in oedema/ascites (hydrophilic drugs) and falls with dehydration.
• Clearance (CL): volume of plasma cleared of drug per unit time (L/h or mL/min). CL = CL(hepatic) + CL(renal) + others. CL determines the MAINTENANCE dose.
• Elimination rate constant (k or kel): fraction eliminated per unit time, h⁻¹. k = CL/Vd.
• Half-life (t½): time for concentration to halve. t½ = 0.693/k = 0.693 Vd/CL. It is a *derived* parameter - it changes only if Vd or CL changes. t½ determines the dosing interval and the time to steady state.
• Protein binding: only the unbound (free) fraction is pharmacologically active and available for clearance. Phenytoin (about 90% bound to albumin), valproate, warfarin, ceftriaxone are highly bound; hypoalbuminaemia, uraemia and pregnancy raise the free fraction.
• Extraction ratio and first-pass: hepatic CL depends on liver blood flow for high-extraction drugs and on enzyme capacity for low-extraction drugs - this is why liver disease and enzyme inducers affect drugs differently.

Kinetic orders and compartment models

• First-order (linear) kinetics: rate of elimination is proportional to concentration; a constant *fraction* is removed per unit time; plasma concentration is proportional to dose; log C versus t is a straight line. Most drugs at therapeutic doses.
• Zero-order (non-linear/saturation) kinetics: a constant *amount* is removed per unit time because enzymes are saturated; small dose increases cause disproportionate rises. Classic examples: ethanol, high-dose aspirin, phenytoin, theophylline at the top of its range.
• Michaelis-Menten (mixed-order) kinetics: rate = Vmax·C/(Km + C). When C ≪ Km it behaves first-order; when C ≫ Km it is zero-order. Phenytoin is the exam standard - Km in adults is roughly 4 mg/L and Vmax about 7 mg/kg/day, so a rise from 300 to 400 mg/day can push a level from 10 to 25 mg/L.
• One-compartment model: body behaves as a single well-stirred space; single exponential decline after IV bolus.
• Two-compartment model: rapid distribution (α) phase followed by slower elimination (β) phase. Clinically vital - a digoxin level drawn during the 6-8 h distribution phase is falsely high, which is why samples are taken at least 6 h (practically 8-12 h) post-dose.
• Flip-flop kinetics: when absorption is slower than elimination (some sustained-release products), the terminal slope reflects absorption, not elimination.
• Steady state (Css): reached when rate in = rate out; takes about 4-5 half-lives (about 94-97% of Css). A loading dose reaches the target *concentration* immediately but does NOT shorten the time to steady state.
• Accumulation and fluctuation: shortening the interval relative to t½ raises the average level and flattens peak-trough swing; the same daily dose given as one large dose raises the peak and lowers the trough.

Therapeutic drug monitoring - when, how and why

• Definition: measurement of drug (or metabolite) concentration in a biological fluid at a defined time, interpreted together with clinical and laboratory findings, to optimise the dose.
• Criteria for a TDM candidate drug (examined almost every year): narrow therapeutic index; wide inter-individual PK variability; no easily measured clinical end-point; a defined concentration-effect relationship; saturable or non-linear kinetics; suspected toxicity, non-compliance, interaction or treatment failure; altered physiology (renal/hepatic failure, pregnancy, neonates, burns, obesity, critical illness); long-term therapy.
• Drugs NOT suited to TDM: those with a clear measurable end-point (antihypertensives - measure BP; oral anticoagulants - measure INR; oral hypoglycaemics - measure blood glucose; diuretics - weight and urine output) or a wide therapeutic index (most penicillins).
• Sampling discipline - the commonest source of error:
  • Trough: immediately before the next dose (within 30 min) - the standard for most anticonvulsants, vancomycin, lithium, tacrolimus, cyclosporine.
  • Peak: 30 min after the end of a 30-min IV infusion for aminoglycosides; about 1 h after an IM dose.
  • Digoxin: at least 6 h, usually 8-12 h, after the dose.
  • Lithium: exactly 12 h after the evening dose (the 12-h standardised serum lithium level).
  • Wait for steady state (4-5 t½) unless checking for toxicity or using a Bayesian method.
• Assay methods: immunoassays (FPIA, CLIA, EMIT, ELISA) in most Indian hospital labs - cheap and fast but liable to cross-reactivity (for example digoxin-like immunoreactive substances in renal failure, neonates and pregnancy); chromatographic methods (HPLC, LC-MS/MS) are the reference for immunosuppressants, antiretrovirals and newer antiepileptics; dried blood spot sampling is used in some paediatric and research settings.
• Reporting: always record drug, dose, route, time of last dose, time of sampling, duration of therapy, serum creatinine, albumin, co-medication and indication - a level without these is uninterpretable.

Dose individualisation and special populations

• Renal impairment: estimate renal function first. Cockcroft-Gault creatinine clearance is still the standard for drug dosing in Indian practice; CKD-EPI/MDRD eGFR is used for staging. Reduce dose or extend interval for renally cleared drugs (aminoglycosides, vancomycin, digoxin, lithium, methotrexate, most beta-lactams).
• Dialysis: drugs with low Vd, low protein binding and low molecular weight are removed by haemodialysis (aminoglycosides, lithium, vancomycin partially with high-flux membranes); supplementary post-dialysis doses are needed.
• Hepatic impairment: reduces CL of drugs with high hepatic extraction; Child-Pugh grading guides reduction; also reduces albumin, raising free fraction of phenytoin and valproate.
• Paediatrics and neonates: immature enzymes and renal function at birth, then higher clearance per kg in children; dosing by mg/kg or body-surface area, never by adult-dose fractions alone.
• Geriatrics: falling renal function, lower albumin, lower muscle mass (so serum creatinine under-reads the true loss of function), lower total body water - start low, go slow.
• Obesity: use adjusted body weight for aminoglycosides (ABW = IBW + 0.4 × (TBW − IBW)) and total body weight for the vancomycin loading dose; lipophilic drugs distribute into fat.
• Pregnancy and critical illness: expanded volume, raised GFR, hypoalbuminaemia and capillary leak all change exposure; septic patients often need higher initial antibiotic doses.
• Correcting phenytoin for low albumin (Sheiner-Tozer): essential whenever albumin is low or the patient is uraemic - otherwise a "subtherapeutic" total level hides an adequate free level.
• Interaction-driven adjustment: enzyme inducers (rifampicin, carbamazepine, phenytoin, phenobarbitone) lower levels; inhibitors (erythromycin/clarithromycin, azole antifungals, ritonavir, valproate on lamotrigine) raise them. Detailed interaction mechanisms belong to the drug-interactions note.

Drug-wise therapeutic ranges and practice points

• Ranges differ slightly between laboratories and guidelines; always use the reporting laboratory's quoted range. Indicative adult values commonly taught:
• Digoxin: 0.5-2 ng/mL historically; for heart failure 0.5-0.9 ng/mL is now preferred. Toxicity is worsened by hypokalaemia, hypomagnesaemia, hypercalcaemia, hypothyroidism and renal failure. Features - nausea, vomiting, visual yellow-green halos, bradycardia, arrhythmias.
• Phenytoin: total 10-20 mg/L, free 1-2 mg/L. Nystagmus at about 20, ataxia at about 30, lethargy/coma above about 40 mg/L. Zero-order at the top of the range; increase by 25-50 mg steps once above 300 mg/day.
• Carbamazepine: 4-12 mg/L; auto-induces its own metabolism over 2-4 weeks, so levels fall on a fixed dose.
• Valproate: 50-100 mg/L; highly protein-bound and saturable binding, so free fraction rises with dose.
• Phenobarbitone: 15-40 mg/L; very long t½ (about 4-5 days), so steady state takes 2-3 weeks.
• Lithium: 0.6-1.2 mEq/L maintenance (0.8-1.2 in acute mania); above 1.5 toxic, above 2.5 a medical emergency. Thiazides, ACE inhibitors, NSAIDs and dehydration raise levels.
• Vancomycin: modern guidance is AUC₂₄/MIC 400-600 (MIC 1 mg/L) for MRSA, with trough 10-15 mg/L (15-20 in deep-seated infection) still used where AUC software is unavailable, as in many Indian hospitals.
• Aminoglycosides: conventional gentamicin/tobramycin peak 5-10 mg/L and trough below 2 mg/L; amikacin peak 20-30 and trough below 5-8 mg/L. Extended-interval (once-daily) dosing exploits concentration-dependent killing and reduces nephrotoxicity; it is monitored by a single mid-interval level with the Hartford nomogram.
• Theophylline: 10-20 mg/L (lower in neonatal apnoea); affected by smoking (induction), heart failure and ciprofloxacin.
• Immunosuppressants: cyclosporine trough about 100-400 ng/mL and tacrolimus about 5-20 ng/mL, both varying widely by organ, protocol and time post-transplant - follow the transplant unit's protocol.
• Others monitored: methotrexate (for leucovorin rescue), amiodarone, procainamide, lamotrigine, linezolid (toxicity), antiretrovirals in research settings.

Indian regulatory and practice context

• Pharmacy Practice Regulations, 2015 (Pharmacy Council of India, under the Pharmacy Act, 1948) formally list patient counselling, drug therapy review and monitoring as duties of a registered pharmacist - the legal basis for a clinical pharmacist's TDM role.
• PCI Pharm.D regulations make Clinical Pharmacokinetics and Pharmacotherapeutics, Clinical Pharmacy and the clerkship/residency rotations compulsory, so TDM is both taught and practised during the internship year.
• NABH hospital accreditation standards on medication management expect documented monitoring of high-risk/narrow-therapeutic-index drugs; this is what creates hospital TDM cells and pharmacist posts.
• Drugs and Cosmetics Act, 1940 and Rules, 1945 (with the New Drugs and Clinical Trials Rules, 2019) govern bioavailability/bioequivalence studies, where the same PK parameters - Cmax, Tmax, AUC - are the regulatory end-points; CDSCO is the regulator.
• Laboratory quality for TDM assays falls under NABL (ISO 15189) medical-laboratory accreditation.
• Advanced tools: Bayesian dose-optimisation software and population PK models are standard abroad and are spreading in large Indian transplant and ICU units, but trough-based protocols remain the norm in most hospitals.

Key points

• Clearance - decides the maintenance dose; volume of distribution - decides the loading dose; half-life - decides the interval and the time to steady state.
• k = CL/Vd - the elimination rate constant is derived, not independent.
• 4-5 half-lives - time to reach steady state, and also the time for a changed dose to show its full effect.
• Loading dose - achieves the target concentration at once but does not change time to steady state.
• First-order - constant fraction eliminated; zero-order - constant amount eliminated, so levels rise disproportionately.
• Michaelis-Menten - phenytoin's kinetics; a 33% dose rise can more than double the level.
• Narrow therapeutic index - the single strongest indication for TDM; a wide index drug rarely needs it.
• Sample timing - trough just before the next dose; aminoglycoside peak 30 min after infusion; digoxin at least 6-8 h post-dose; lithium exactly 12 h.
• Two-compartment distribution phase - the reason an early digoxin sample reads falsely high.
• Protein binding - only free drug acts; correct phenytoin for hypoalbuminaemia with Sheiner-Tozer.
• Cockcroft-Gault - the creatinine clearance equation used for drug dose adjustment in renal impairment.
• Dialysable drugs - low Vd, low binding, small molecule (lithium, aminoglycosides); need post-dialysis supplements.
• Carbamazepine auto-induction - levels fall on an unchanged dose over 2-4 weeks.
• Vancomycin - AUC₂₄/MIC 400-600 is current guidance; trough 10-20 mg/L used where software is unavailable.
• Extended-interval aminoglycosides - high peak for concentration-dependent killing, long trough-free period to limit nephrotoxicity.
• Immunoassay cross-reactivity - digoxin-like immunoreactive substances cause falsely high digoxin levels in renal failure, pregnancy and neonates.
• A level without its timing and clinical data is uninterpretable - the report must carry dose, last dose time, sampling time, creatinine and albumin.
• Pharmacy Practice Regulations, 2015 - the Indian legal mandate for pharmacist-led drug therapy monitoring.

Formulas and examples

Example 1 - loading and maintenance dose of theophylline. A 60 kg adult with acute asthma; target Css 15 mg/L, Vd 0.5 L/kg, CL 2.4 L/h, F = 1 for IV aminophylline (take the salt factor as 0.8 for aminophylline).

• Vd = 0.5 × 60 = 30 L.
• LD (as theophylline base) = 30 × 15 = 450 mg; as aminophylline = 450/0.8 = 562.5 mg, round to about 560 mg IV over 30 min.
• Maintenance rate = CL × Css = 2.4 × 15 = 36 mg/h of theophylline = 36/0.8 = 45 mg/h of aminophylline.
• t½ = 0.693 × 30 / 2.4 = 8.66 h, so steady state in about 4 × 8.66 ≈ 35 h - check a level then.

Example 2 - gentamicin in renal impairment. Male, 70 years, 65 kg, serum creatinine 1.8 mg/dL.

• CrCl = [(140 − 70) × 65] / (72 × 1.8) = 4,550 / 129.6 = 35.1 mL/min.
• Normal gentamicin CL ≈ CrCl, so CL ≈ 35.1 mL/min = 2.11 L/h. With Vd = 0.25 × 65 = 16.25 L, k = 2.11/16.25 = 0.130 h⁻¹ and t½ = 0.693/0.130 = 5.3 h (normal is about 2-3 h).
• Because t½ has roughly doubled, extend the interval: an extended-interval dose of 5 mg/kg = 325 mg given every 36-48 h rather than every 24 h, with a level to confirm; trough must fall below 1 mg/L before the next dose.

Example 3 - phenytoin dose increase (Michaelis-Menten). A patient on 300 mg/day has a steady-state level of 8 mg/L. Take Km = 4 mg/L.

• From R = Vmax·Css/(Km + Css): 300 = Vmax × 8/(4 + 8) → Vmax = 300 × 12/8 = 450 mg/day.
• Dose needed for a target of 15 mg/L: R = 450 × 15/(4 + 15) = 6,750/19 = 355 mg/day - so only about 50-55 mg/day more.
• Dose needed for 20 mg/L: R = 450 × 20/24 = 375 mg/day. A further 20 mg/day moves the level from 15 to 20 mg/L - this is why increments above 300 mg/day are 25-50 mg and never 100 mg.

Example 4 - Sheiner-Tozer correction. Reported total phenytoin 6.5 mg/L; albumin 2.2 g/dL; normal renal function.

• C corrected = 6.5 / (0.2 × 2.2 + 0.1) = 6.5 / (0.44 + 0.1) = 6.5 / 0.54 = 12.0 mg/L - within the 10-20 mg/L range, so the dose should NOT be increased.
• If the same patient had CrCl 15 mL/min: C corrected = 6.5 / (0.1 × 2.2 + 0.1) = 6.5/0.32 = 20.3 mg/L - at the top of the range; increasing the dose could cause toxicity.

Example 5 - digoxin loading dose and a mistimed level. 70 kg patient with atrial fibrillation, normal renal function; Vd 7 L/kg, target 1.2 ng/mL (= 1.2 µg/L), oral F = 0.7.

• Vd = 7 × 70 = 490 L. LD = (490 × 1.2)/0.7 = 588/0.7 = 840 µg, given as about 0.25 mg six-hourly in divided doses to about 0.75-1 mg total over 24 h with ECG review.
• A level drawn 2 h after the dose reads 3.1 ng/mL. This is the distribution phase, not toxicity - repeat at 8-12 h post-dose before changing anything. A truly toxic level is interpreted with potassium, magnesium, renal function and the ECG, not alone.

Where it is used

• Hospital TDM cells and clinical pharmacy units - AIIMS, PGIMER, CMC Vellore, NIMHANS (lithium and antiepileptics), SCTIMST, Apollo, Fortis, Manipal, Amrita: pharmacists interpret levels, recommend doses and document in the case sheet during ward rounds.
• ICU and transplant services - vancomycin and amikacin dosing in sepsis, tacrolimus/cyclosporine after renal and liver transplant, often with extended-interval or Bayesian protocols.
• Oncology and nephrology - high-dose methotrexate level-guided leucovorin rescue; dose adjustment for dialysis and CKD patients.
• Clinical research and CRO work - Cmax, Tmax, AUC, Vd and CL are the end-points of bioavailability/bioequivalence studies run under CDSCO and the New Drugs and Clinical Trials Rules, 2019; roles in PK analysis, NCA/WinNonlin work and protocol writing.
• Pharmaceutical medical affairs and pharmacovigilance - PK-based dose recommendations for renal/hepatic impairment in package inserts, and causality assessment of concentration-related adverse reactions.
• Paediatrics and neonatology - weight- and age-based dosing of gentamicin, vancomycin, phenobarbitone and caffeine citrate, where TDM prevents both failure and toxicity.
• Teaching and nutrition/TPN services, antimicrobial stewardship programmes - pharmacist-led dose optimisation is now a measured NABH/stewardship indicator.

Interview questions

• Q: Which pharmacokinetic parameter decides the loading dose and which decides the maintenance dose? A: Volume of distribution decides the loading dose, because it relates the amount in the body to the target concentration; clearance decides the maintenance dose, because at steady state input rate must equal CL × Css.
• Q: Why does a loading dose not shorten the time to steady state? A: Time to steady state depends only on half-life (4-5 t½); a loading dose merely starts the patient at the target concentration instead of climbing to it.
• Q: Name the criteria that make a drug suitable for TDM. A: Narrow therapeutic index, wide inter-patient PK variability, no easily measurable clinical end-point, an established concentration-effect relationship, and situations such as non-linear kinetics, suspected toxicity, non-compliance, interactions or organ failure.
• Q: Why is digoxin not sampled soon after the dose? A: Digoxin follows two-compartment kinetics with a 6-8 h distribution phase into tissue; an early sample is falsely high and could prompt a wrong dose reduction. Sample 8-12 h post-dose.
• Q: A patient on 300 mg/day phenytoin has a level of 8 mg/L. Would you add 100 mg? A: No - phenytoin shows saturable Michaelis-Menten kinetics; 100 mg could push the level well above 20 mg/L. Increase by 25-50 mg/day and recheck after about 1-2 weeks.
• Q: Phenytoin total level is 6 mg/L in a patient with albumin 2 g/dL. What do you do? A: Apply the Sheiner-Tozer correction - 6 / (0.2 × 2 + 0.1) = 12 mg/L, which is therapeutic, so the dose stays unchanged; measure free phenytoin if available.
• Q: Why is once-daily aminoglycoside dosing preferred? A: Aminoglycosides kill in a concentration-dependent way with a post-antibiotic effect, so a high peak improves efficacy, while the long low-concentration period reduces renal tubular and cochlear uptake - that is, less nephro- and ototoxicity with equal or better efficacy.
• Q: Which drugs are removed by haemodialysis and why? A: Small, water-soluble, poorly protein-bound drugs with a low volume of distribution - lithium, aminoglycosides, methanol, salicylates; they need a supplementary dose after dialysis.
• Q: Why does carbamazepine's level fall even though the dose is unchanged? A: It induces its own metabolising enzymes (auto-induction) over 2-4 weeks, raising its clearance and shortening its half-life.
• Q: What has changed in vancomycin monitoring? A: Current consensus guidance targets AUC₂₄/MIC of 400-600 rather than trough alone for serious MRSA infection; trough-based monitoring (10-20 mg/L) is still used where AUC software or two-level sampling is not available.
• Q: Why can a hypoalbuminaemic patient have toxicity at a "therapeutic" total phenytoin level? A: Lower albumin means a higher free fraction, so the pharmacologically active free concentration is high even though the total measured level looks normal.
• Q: What information must accompany a TDM request for it to be interpretable? A: Drug, dose and route, time of the last dose, exact sampling time, duration of therapy, serum creatinine and albumin, co-medication and the clinical reason for the request.

Common mistakes

• Treating half-life as an independent parameter - it changes only through Vd or CL, so it cannot be "adjusted" on its own.
• Using linear proportionality (new dose = old dose × target/measured) for phenytoin, where saturable kinetics make it dangerously wrong.
• Sampling before steady state, or at the wrong time, and then changing the dose on that number.
• Ignoring protein binding - not correcting phenytoin or valproate levels in hypoalbuminaemia or renal failure.
• Using serum creatinine alone instead of estimated creatinine clearance in elderly or low-muscle-mass patients, who have a normal creatinine despite poor renal function.
• Confusing salt factor with bioavailability - aminophylline/theophylline and phenytoin sodium/phenytoin acid doses are not interchangeable.
• Treating a number instead of the patient - acting on a digoxin level without checking potassium, magnesium, renal function and the ECG.
• Mixing units in calculations: ng/mL with mg/L, L/h with mL/min, mg with µg.

Related topics

• Clinical pharmacy practice: ward rounds, medication review and drug information · Drug interactions and medication errors · Clinical toxicology: poisoning and its management · Pharmacovigilance · Biopharmaceutics and bioequivalence · Pharmacotherapeutics of renal and hepatic disease.