Therapeutic Drug Monitoring - Pharmacology

SECTION 1

Historical Context & Motivation

For much of medical history, clinicians relied on observable patient responses—symptom relief, adverse reactions, or frank toxicity—to guide drug dosing. This trial-and-error approach was fraught with danger, especially for drugs with narrow therapeutic indices, where the margin between a therapeutic and a toxic concentration is perilously slim. The concept of therapeutic drug monitoring (TDM) arose from the recognition that measuring actual drug concentrations in a patient's blood could replace guesswork with quantifiable, reproducible data. By linking plasma drug levels to pharmacological outcomes, TDM transformed medication management from a reactive practice into a proactive, individualized science.

1940s • Early Serum Assays

Pioneering spectrophotometric and microbiological assays enabled crude measurement of antibiotic levels in serum, establishing the principle that drug concentration correlates with clinical response.

1960s • Anticonvulsant Monitoring Begins

Landmark studies by Buchthal and Lennox demonstrated that plasma phenytoin concentrations predicted seizure control far more reliably than dose alone, catalyzing the modern TDM movement.

1970s • Immunoassay Revolution

The development of radioimmunoassay (RIA) and enzyme-multiplied immunoassay technique (EMIT) made rapid, routine quantification of drug levels feasible in hospital laboratories worldwide.

1980s–90s • Expansion to Immunosuppressants

The introduction of cyclosporine for organ transplantation demanded rigorous TDM, and population pharmacokinetic modeling using Bayesian methods emerged for individualized dosing.

2000s–Present • Pharmacogenomics Integration

Advances in LC-MS/MS technology and genetic testing now enable clinicians to combine drug level data with a patient's pharmacogenomic profile, ushering in an era of precision dosing.

The fundamental question TDM seeks to answer remains deceptively simple: Is this patient receiving enough drug to be effective, but not so much as to cause harm? Answering this question requires an understanding of pharmacokinetic variability, assay methodology, the concept of a therapeutic range, and the clinical judgment to interpret results in the context of a whole patient.

SECTION 2

Core Principles & Definitions

Therapeutic drug monitoring rests on several foundational pharmacological principles that connect a drug's plasma concentration to its clinical effect. Understanding these principles is essential before any meaningful interpretation of a drug level can occur. At the heart of TDM lies the assumption that a measurable relationship exists between the concentration of drug in the blood and the pharmacological effect at the target site. This relationship, while imperfect, is strong enough for certain drug classes that plasma concentration serves as a practical surrogate for the drug's activity at tissue receptor sites.

Therapeutic Range

The range of plasma drug concentrations associated with a high probability of therapeutic success and a low probability of dose-related toxicity. Also called the therapeutic window. Values below the range risk treatment failure; values above risk adverse effects.

Steady State

The condition achieved when the rate of drug input equals the rate of drug elimination. At steady state, plasma concentrations fluctuate predictably between peak (Cmax) and trough (Cmin) values with each dosing interval. Steady state is generally reached after 4–5 half-lives.

Narrow Therapeutic Index (NTI)

Drugs for which small changes in plasma concentration lead to significant changes in pharmacological response—either loss of efficacy or onset of toxicity. Classic examples include digoxin, lithium, warfarin, aminoglycosides, and phenytoin.

Trough Level Timing

Most TDM samples are drawn as trough levels—immediately before the next scheduled dose—because this represents the lowest drug concentration in the dosing interval and is the most reproducible sampling point for clinical comparison.

Interpatient Variability

Genetic polymorphisms in metabolizing enzymes (e.g., CYP450 isoforms), renal function, hepatic function, age, body composition, drug interactions, and disease states all contribute to pharmacokinetic variability, making standardized dosing unreliable for NTI drugs.

✦

KEY TAKEAWAY

Think of the therapeutic range like a target zone on a thermostat: set the temperature too low and the room is uncomfortably cold (subtherapeutic), set it too high and the system overheats (toxicity). TDM is the thermometer that tells you exactly where the system is so you can make precise adjustments, rather than simply turning the dial and hoping for the best.

SECTION 3

Visualizing the Therapeutic Window

The cyan curve shows plasma drug concentration rising to a peak (Cmax) after each dose and falling to a trough (Cmin) before the next. The green shaded region represents the therapeutic window bounded by the minimum effective concentration (MEC, amber dashed line) and the minimum toxic concentration (MTC, red dashed line). With repeated dosing at interval τ, the drug accumulates until steady state is reached (approximately 4–5 half-lives), at which point Cmax and Cmin stabilize.

The diagram above encapsulates the central rationale for TDM. When a patient receives repeated doses of a drug at a fixed interval (τ), the plasma concentration oscillates in a sawtooth pattern defined by the drug's absorption, distribution, and elimination kinetics. The clinician's goal is to keep these oscillations entirely within the therapeutic window. If the peak concentration exceeds the MTC, the patient may experience concentration-dependent toxicity—as seen with aminoglycoside nephrotoxicity. If the trough concentration falls below the MEC, the drug fails to exert its intended effect—as occurs when vancomycin troughs drop below the target, permitting bacterial regrowth. TDM provides the measured data points to verify that a patient's actual concentrations align with these theoretical targets.

SECTION 4

Pharmacokinetic Framework for TDM

The mathematical backbone of TDM is grounded in first-order pharmacokinetics. Most drugs monitored through TDM follow first-order elimination kinetics, meaning that the rate of drug elimination is proportional to the amount of drug in the body. A notable exception is phenytoin, which exhibits saturable (Michaelis-Menten) kinetics at therapeutic concentrations. Below are the key equations used in TDM calculations.

ELIMINATION HALF-LIFE

t₁/₂ = 0.693 / Ke

Where t₁/₂ = elimination half-life (hours), Ke = elimination rate constant (hr⁻¹), and 0.693 = ln(2). The half-life determines how long it takes for plasma concentration to decline by 50%.

STEADY-STATE TROUGH CONCENTRATION (IV INTERMITTENT)

Css,min = (S × F × Dose / Vd) × [e^(−Ke × τ) / (1 − e^(−Ke × τ))]

Css,min = steady-state trough concentration, S = salt factor, F = bioavailability, Vd = volume of distribution (L), τ = dosing interval (hours). This equation predicts the lowest concentration in a dosing interval at steady state.

CLEARANCE FROM MEASURED LEVELS

Cl = (Dose × F) / (AUC₀₋τ)

Cl = clearance (L/hr), AUC₀₋τ = area under the concentration-time curve over one dosing interval at steady state. Clearance reflects the body's overall ability to eliminate the drug and is a critical parameter for dose adjustment.

ADJUSTED PHENYTOIN LEVEL (WINTER-TOZER)

C_adjusted = C_measured / (0.2 × Albumin + 0.1)

Because phenytoin is highly protein-bound (≈90%), hypoalbuminemia increases the free (active) fraction. The Winter-Tozer equation adjusts the measured total concentration to an equivalent concentration if albumin were normal (4.4 g/dL). A lower albumin yields a higher adjusted concentration, indicating the patient may be closer to toxicity than the raw level suggests.

⚠

Clinical Note

Phenytoin is a frequent board-examination topic because its saturable metabolism means that small dose increases can produce disproportionately large rises in plasma concentration. Unlike first-order drugs, where doubling the dose doubles the steady-state level, phenytoin's Michaelis-Menten kinetics can cause levels to increase exponentially once the metabolic enzymes approach saturation.

SECTION 5

Commonly Monitored Drug Classes

Not every drug requires TDM—monitoring is reserved for medications where the benefit of measuring plasma levels clearly outweighs the cost and inconvenience of blood draws and laboratory analysis. The criteria for selecting drugs appropriate for TDM generally include a narrow therapeutic index, significant interpatient pharmacokinetic variability, a defined relationship between concentration and effect, and the availability of a validated assay. The following table summarizes the most clinically important drug classes subject to routine TDM.

Drug / Class	Therapeutic Range	Key Toxicities	Sample Timing
Vancomycin	AUC/MIC ≥ 400; trough 15–20 mcg/mL (serious infections)	Nephrotoxicity, ototoxicity, Red Man Syndrome (infusion-related)	Trough within 30 min before 4th dose; AUC-guided with 2 levels
Aminoglycosides (gentamicin, tobramycin)	Traditional: peak 5–10 mcg/mL, trough < 2 mcg/mL; Extended interval: random level < 1 mcg/mL	Nephrotoxicity (trough-related), ototoxicity (cumulative)	Peak 30 min post-infusion; trough 30 min pre-dose
Phenytoin	10–20 mcg/mL (total); 1–2 mcg/mL (free)	Nystagmus, ataxia, cardiac arrhythmias, gingival hyperplasia	Trough; adjust for albumin using Winter-Tozer equation
Digoxin	0.8–2.0 ng/mL (heart failure: 0.5–0.9 ng/mL)	Cardiac arrhythmias, nausea, visual disturbances (yellow halos)	At least 6–8 hours post-dose (distribution phase completion)
Lithium	0.6–1.2 mEq/L (acute mania up to 1.5 mEq/L)	Tremor, polyuria, hypothyroidism, renal impairment, seizures	Trough 12 hours post-dose
Tacrolimus / Cyclosporine	Tacrolimus: 5–15 ng/mL; Cyclosporine: 150–400 ng/mL (varies by organ and time post-transplant)	Nephrotoxicity, neurotoxicity, hyperglycemia, immunosuppression-related infections	Trough (C₀) 12 hours post-dose; C₂ monitoring for cyclosporine

Summary of commonly monitored drugs, therapeutic ranges, key toxicities, and recommended sampling times

This flowchart illustrates the clinical decision pathway for TDM. The process begins only after steady state is achieved (4–5 half-lives). If the measured trough falls outside the therapeutic range, the dose is recalculated using pharmacokinetic equations and the cycle repeats.

SECTION 6

Worked Example — Vancomycin Dose Adjustment

A 72-year-old male patient (80 kg, SCr 1.4 mg/dL) is receiving vancomycin 1,000 mg IV every 12 hours for a methicillin-resistant Staphylococcus aureus (MRSA) bacteremia. A trough level drawn just before the fourth dose returns at 8 mcg/mL. The target trough for serious MRSA infections is 15–20 mcg/mL. The clinical pharmacist is asked to recommend a new dose to achieve a target trough of 15 mcg/mL.

Vancomycin Trough-Based Dose Adjustment (Linear PK Method)

1
Step 1 — Confirm Steady State

Vancomycin's typical half-life in this patient population (elderly, mild renal impairment) is approximately 8–12 hours. With a dosing interval of 12 hours, four doses means roughly 48 hours of therapy. At a half-life of ≈10 hours, 4–5 half-lives = 40–50 hours. The patient is at or near steady state, so the measured trough is interpretable.

Steady state confirmed (≈ 48 hours ≥ 4 × t₁/₂)
2
Step 2 — Apply the Linear Proportionality Method

For drugs following first-order kinetics, steady-state concentrations are directly proportional to dose (assuming no change in clearance or interval). Therefore: New Dose / Old Dose = Target Trough / Measured Trough. Rearranging: New Dose = Old Dose × (Target Trough / Measured Trough) = 1,000 mg × (15 / 8).

New Dose = 1,000 × 1.875 = 1,875 mg
3
Step 3 — Round to a Practical Dose

Vancomycin is typically dosed in 250 mg increments. Rounding 1,875 mg to the nearest 250 mg gives 1,750 mg or 2,000 mg. Given that the patient's current trough is significantly below target in the setting of serious MRSA bacteremia, the pharmacist may recommend rounding up to 2,000 mg every 12 hours, with close monitoring for nephrotoxicity.

Recommended dose: 2,000 mg IV q12h
4
Step 4 — Predict the New Trough

Using the proportionality approach: Expected new trough = 8 mcg/mL × (2,000 / 1,000) = 16 mcg/mL. This falls within the target range of 15–20 mcg/mL.

Predicted trough = 16 mcg/mL (within target range of 15–20 mcg/mL)
5
Step 5 — Plan Follow-Up Monitoring

A repeat trough should be drawn once the patient reaches steady state on the new regimen—approximately 4–5 half-lives (40–50 hours) after the dose change. Additionally, renal function (SCr, BUN) should be monitored at least every 48–72 hours given the higher vancomycin dose and the patient's baseline renal impairment.

Recheck trough in ≈ 48 hours; monitor renal function q48–72h

💡

AUC-Guided Monitoring

Current 2020 ASHP/IDSA guidelines recommend transitioning from trough-only monitoring to AUC/MIC-guided dosing for vancomycin (target AUC₂₄/MIC of 400–600). This approach uses Bayesian software with two measured levels (peak and trough) to estimate AUC, reducing nephrotoxicity risk while maintaining efficacy. The linear proportionality method shown above remains a useful bedside estimation tool.

SECTION 7

Strengths & Limitations of TDM

Therapeutic drug monitoring is an indispensable tool in modern clinical practice, yet it is not without significant caveats. Recognizing both its strengths and limitations helps clinicians apply TDM judiciously rather than reflexively. A measured drug level is only as useful as the clinical context in which it is interpreted, and several factors can undermine the utility of even the most precise laboratory result.

Strengths	Limitations
Enables individualized dosing by accounting for patient-specific pharmacokinetic variability (renal/hepatic function, weight, age, drug interactions)	Requires accurate sample timing; an improperly drawn level (e.g., before distribution phase completion for digoxin) yields misleading data
Reduces risk of dose-related toxicity for narrow therapeutic index drugs, improving patient safety	Therapeutic ranges are population-based guidelines, not absolute thresholds; some patients experience toxicity within the 'therapeutic' range, and others tolerate supratherapeutic levels
Provides objective data to guide dose escalation when efficacy is uncertain (e.g., persistent seizures despite adequate phenytoin doses)	Assay interference (e.g., digoxin-like immunoreactive substances, DLIS) can produce falsely elevated or depressed results
Helps detect non-adherence, drug interactions, and changes in organ function that alter drug disposition	Measures total drug concentration (bound + unbound) in most assays; free drug levels are needed for highly protein-bound drugs in hypoalbuminemic patients
Cost-effective in preventing adverse drug events, hospitalizations, and treatment failures for high-risk medications	Cost and patient burden of repeated blood draws; turnaround time may delay clinical decision-making in urgent situations

Comparative strengths and limitations of TDM in clinical practice

✦

KEY TAKEAWAY

A drug level is a data point, not a diagnosis. Just as a GPS coordinate tells you where you are on the map but doesn't tell you where you should be going, a measured plasma concentration tells you the patient's current pharmacokinetic state but requires clinical context—symptoms, renal function trends, drug interactions, treatment goals—to guide the next step. TDM supports clinical judgment; it does not replace it.

SECTION 8

Connection to Precision Dosing & Pharmacogenomics

Traditional TDM—drawing a level, comparing it to a population-derived therapeutic range, and adjusting the dose—represents the first generation of individualized pharmacotherapy. The field is rapidly evolving toward model-informed precision dosing (MIPD), which integrates patient-specific covariates, measured drug concentrations, and population pharmacokinetic models using Bayesian estimation to predict the entire concentration-time profile and optimize doses prospectively. When combined with pharmacogenomic testing—identifying genetic variants in drug-metabolizing enzymes (e.g., CYP2C9 for phenytoin, CYP2D6 for codeine, UGT1A1 for irinotecan)—clinicians can anticipate pharmacokinetic variability before the first dose is even administered.

Feature	Traditional TDM	Precision Dosing (MIPD)
Data inputs	Single trough (or peak/trough pair) at steady state	Multiple levels at any time, plus covariates (weight, renal function, genotype)
Modeling approach	Linear proportionality or simple PK equations	Population PK models with Bayesian feedback; nonlinear mixed-effects modeling
Target metric	Trough concentration within published range	AUC, Cmax, AUC/MIC, or probability of target attainment (PTA)
Timing flexibility	Requires steady state and precise sample timing	Can use levels drawn at any time, even before steady state
Software requirement	Calculator or manual computation	Bayesian dosing software (e.g., InsightRx, DoseMeRx, MwPharm)

Traditional TDM versus Model-Informed Precision Dosing

As electronic health records increasingly embed pharmacokinetic decision-support tools and pharmacogenomic test results become routinely available, the distinction between TDM and precision dosing will continue to blur. Healthcare students today will practice in an environment where a patient's genotype informs the initial dose, Bayesian software refines it after the first measured level, and machine-learning algorithms may eventually automate dosing adjustments in near-real time. Understanding the traditional principles of TDM remains essential, however, because these fundamentals—steady state, half-life, therapeutic range, clearance—form the conceptual scaffolding on which precision dosing is built.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A clinician orders a vancomycin trough level on a patient who has received only two doses of vancomycin (1,000 mg IV q12h, with an estimated half-life of 8 hours). The result comes back at 9 mcg/mL. Should this result be used to make a dose adjustment? Explain your reasoning.

Show answer

This result should not be used for dose adjustment because the patient has not yet reached steady state. With a half-life of 8 hours, steady state requires approximately 4–5 half-lives, or 32–40 hours. After only two doses (24 hours), the drug is still accumulating and the measured trough does not yet reflect the ultimate steady-state trough. The concentration will continue to rise with subsequent doses. Drawing and acting on a pre-steady-state level risks overdosing the patient, since the eventual steady-state trough will be higher than the current value. The appropriate action is to wait until at least the fourth or fifth dose to obtain a reliable trough.

Question 2: BASIC CALCULATION

A patient has a measured phenytoin level of 8 mcg/mL and a serum albumin of 2.0 g/dL. Using the Winter-Tozer equation (C_adjusted = C_measured / (0.2 × Albumin + 0.1)), calculate the adjusted phenytoin concentration. Is this patient within the therapeutic range of 10–20 mcg/mL?

Show answer

Applying the Winter-Tozer equation: C_adjusted = 8 / (0.2 × 2.0 + 0.1) = 8 / (0.4 + 0.1) = 8 / 0.5 = 16 mcg/mL. The adjusted concentration of 16 mcg/mL falls within the therapeutic range of 10–20 mcg/mL. Although the measured total level of 8 mcg/mL appears subtherapeutic, this patient's hypoalbuminemia means a greater fraction of phenytoin is unbound and pharmacologically active. The adjusted value corrects for this and suggests the patient is actually adequately treated. No dose increase is necessary.

Question 3: INTERMEDIATE

A patient is receiving gentamicin 120 mg IV every 8 hours (traditional dosing). At steady state, the measured peak (drawn 30 minutes after the infusion ends) is 9 mcg/mL and the trough (drawn 30 minutes before the next dose) is 3.5 mcg/mL. The target peak is 5–10 mcg/mL and the target trough is < 2 mcg/mL. Which value is out of range, and what dosing strategy change would you recommend to correct the trough while maintaining the peak?

Show answer

The peak of 9 mcg/mL is within the target range (5–10 mcg/mL), but the trough of 3.5 mcg/mL exceeds the target of < 2 mcg/mL, placing the patient at increased risk for nephrotoxicity and ototoxicity. Because the peak is appropriate (indicating the dose is correct for efficacy), simply reducing the dose would lower both peak and trough, potentially compromising efficacy. The correct strategy is to extend the dosing interval—for example, changing from every 8 hours to every 12 hours—to allow more time for the drug to clear before the next dose, thereby lowering the trough while preserving the same peak concentration. A new trough should be checked after reaching steady state on the new interval.

Question 4: APPLIED

A 45-year-old renal transplant recipient on tacrolimus 5 mg PO twice daily has a steady-state trough of 18 ng/mL (target: 8–12 ng/mL for this post-transplant period). She was recently started on fluconazole for a fungal infection. Her previous tacrolimus trough on the same dose was 10 ng/mL before fluconazole was initiated. Explain the most likely mechanism for the elevated level and recommend a management plan.

Show answer

Fluconazole is a potent inhibitor of CYP3A4, the primary enzyme responsible for tacrolimus metabolism. By inhibiting CYP3A4, fluconazole decreases the clearance of tacrolimus, leading to drug accumulation and a supratherapeutic trough of 18 ng/mL—well above the target of 8–12 ng/mL. At this concentration, the patient is at elevated risk for tacrolimus toxicity including nephrotoxicity, neurotoxicity, and hyperglycemia. The management plan should include an empiric dose reduction (commonly 50% reduction when starting fluconazole) to approximately 2.5 mg PO twice daily, with a repeat trough level drawn after 3–5 days (approximately 4–5 tacrolimus half-lives). Additionally, renal function and electrolytes should be monitored closely, and the tacrolimus dose should be readjusted upward when fluconazole is discontinued to prevent subtherapeutic levels.

Question 5: CRITICAL THINKING

Two patients with epilepsy are both receiving phenytoin 300 mg daily and both have steady-state total phenytoin levels of 12 mcg/mL. Patient A has normal albumin (4.0 g/dL) and normal renal function. Patient B has an albumin of 1.8 g/dL and a GFR of 15 mL/min. Analyze why these identical total drug levels may represent very different clinical scenarios and discuss the pharmacokinetic basis for your analysis.

Show answer

Although both patients share an identical total phenytoin level of 12 mcg/mL, the pharmacologically active free (unbound) fraction differs dramatically. Phenytoin is approximately 90% protein-bound under normal conditions, leaving 10% free. For Patient A, with normal albumin, the free concentration is approximately 1.2 mcg/mL—well within the free therapeutic range of 1–2 mcg/mL. For Patient B, hypoalbuminemia reduces available binding sites, and uremia from renal failure further displaces phenytoin from albumin via competition with accumulated uremic toxins. Using the Winter-Tozer equation with renal impairment adjustment (C_adjusted = C_measured / (0.48 × 0.1 × Albumin + 0.1)), C_adjusted = 12 / (0.048 × 1.8 + 0.1) = 12 / 0.1864 ≈ 64 mcg/mL (with renal adjustment) or using the standard equation: 12 / (0.2 × 1.8 + 0.1) = 12 / 0.46 ≈ 26 mcg/mL. Either way, Patient B's adjusted level far exceeds the therapeutic range, indicating a dangerously elevated free fraction and a high risk of toxicity—despite the reassuring-looking total level. This case underscores why interpreting drug levels requires consideration of protein binding, renal function, and clinical context, and illustrates why a free phenytoin level should be ordered directly in patients with hypoalbuminemia or renal failure rather than relying solely on total levels with correction equations.

SUMMARY

Therapeutic Drug Monitoring — Summary

Summary

Therapeutic drug monitoring (TDM) is the practice of measuring plasma drug concentrations to individualize dosing for drugs with narrow therapeutic indices. The goal is to maintain concentrations within the therapeutic window—above the minimum effective concentration (MEC) and below the minimum toxic concentration (MTC). Commonly monitored drugs include vancomycin, aminoglycosides, phenytoin, digoxin, lithium, and immunosuppressants. Samples should be drawn at steady state (4–5 half-lives) and at the correct time relative to the dosing schedule—typically as trough levels drawn immediately before the next dose.

Key pharmacokinetic equations—including the half-life equation, steady-state trough prediction, and the Winter-Tozer correction for phenytoin—provide the mathematical foundation for dose adjustments. Clinicians must also account for interpatient variability driven by renal/hepatic function, protein binding status, drug-drug interactions, and pharmacogenomic polymorphisms. The field is advancing toward model-informed precision dosing using Bayesian estimation and population pharmacokinetic models, but the foundational principles of TDM remain essential for every healthcare professional.