PBPK vs. Population PK Models

1. Two Ways to Build a PK Model

Physiologically based pharmacokinetic (PBPK) models and population pharmacokinetic (PopPK) models are both mathematical frameworks for describing drug concentration over time, but they approach the problem from different directions.

A PBPK model starts with a representation of physiology: organs and tissues, blood flows, tissue composition, enzyme and transporter activity, and drug-specific properties. A PopPK model generally starts with observed concentration-time data and uses a statistical model to characterize typical PK behavior, between-subject variability, residual variability, and relationships with covariates.

2. What Is a PBPK Model?

A physiologically based pharmacokinetic model represents the body using compartments that correspond more closely to physiological organs or tissues than the abstract compartments used in conventional compartmental PK.

A typical PBPK model may include blood, liver, kidney, lung, muscle, adipose tissue, brain, and other tissues. Each tissue can have its own volume, blood flow, tissue-to-plasma partition behavior, and mechanisms of drug uptake or elimination.

The model combines these physiological characteristics with drug-specific properties such as lipophilicity, molecular size, plasma protein binding, intrinsic clearance, permeability, and enzyme or transporter interactions.

The result is a system of differential equations describing drug amounts or concentrations throughout the modeled physiological system.

3. What Is a Population PK Model?

Population pharmacokinetics describes how PK parameters vary across individuals in a population while simultaneously accounting for unexplained residual variability in observed concentrations.

A PopPK model commonly contains three related components:

1. A structural model describing the typical concentration-time behavior, such as a one- or two-compartment model with absorption and elimination.
2. A statistical model for between-subject variability (BSV) describing how individual parameters vary around typical population values.
3. An observation or residual error model describing differences between measured concentrations and model-predicted concentrations.

Covariates such as body weight, renal function, age, sex, disease status, genotype, or concomitant medications can be incorporated when supported by the data and scientific rationale.

The individual clearance of subject \(i\), for example, might be represented as:

Here, \(CL_{\text{pop}}\) is a typical population clearance, body weight is a covariate, and \(\eta_i\) represents unexplained individual deviation from the typical value after accounting for the covariate.

4. PBPK vs. Population PK: The Central Differences

These distinctions are tendencies rather than absolute rules. Modern pharmacometric workflows frequently combine mechanistic and statistical components rather than treating PBPK and PopPK as mutually exclusive alternatives.

5. How the Mathematical Structures Differ

Population PK equations

A conventional two-compartment PopPK model may describe drug movement using central and peripheral amounts:

The parameters may subsequently be expressed in terms of clearance and volumes, such as \(CL\), \(V_1\), and \(V_2\).

PBPK equations

A PBPK model instead describes drug amounts in physiological tissues. A simplified tissue balance might look like:

where \(A_T\) is drug amount in a tissue, \(Q_T\) is tissue blood flow, and \(C_A\) and \(C_V\) represent arterial and venous concentrations under a simplified flow-limited representation.

More mechanistic tissue models can incorporate permeability-limited distribution, active transport, binding, metabolism, intracellular compartments, or other processes.

6. What Do the Parameters Mean?

This difference has an important consequence: a parameter can have similar terminology but a different role in the two frameworks. For example, “clearance” in a PopPK model is often an estimated statistical parameter, whereas a PBPK model may construct systemic clearance from mechanistic organ processes.

7. What Data Does Each Approach Need?

The two approaches can draw on different information sources.

Typical PBPK inputs

• Physicochemical properties such as molecular weight and lipophilicity.
• Plasma or blood protein binding.
• In vitro metabolic data.
• Enzyme and transporter information.
• Intrinsic clearance measurements.
• Tissue distribution information.
• Organ volumes and blood flows.
• Physiological characteristics of the target population.
• Clinical PK data for model development and verification when available.

Typical PopPK inputs

• Observed concentration-time measurements.
• Exact dosing histories.
• Sampling times.
• Subject demographics.
• Measures of renal or hepatic function.
• Concomitant medications.
• Disease characteristics and other potential covariates.

A major practical difference is that PBPK can use substantial nonclinical and mechanistic information to inform predictions before extensive clinical PK data exist. PopPK is fundamentally dependent on observed clinical concentration data for estimation of the population model.

8. How Do the Models Represent Patient Variability?

Variability is a central concern in clinical pharmacology. Patients can differ substantially in clearance, distribution, absorption, enzyme activity, transporter activity, and other determinants of exposure.

Population PK

PopPK commonly represents individual parameters using random effects:

The variance of \(\eta_{CL}\) describes the magnitude of between-subject variability under the selected model.

Covariates can then explain some of this variability. For example:

PBPK

PBPK can represent variability by changing physiological inputs. For example, individuals may differ in body weight, organ volumes, blood flows, renal function, enzyme abundance, or transporter activity.

Instead of estimating a single random effect directly on clearance, a PBPK simulation may generate a virtual population in which several physiological quantities vary simultaneously.

9. Why PBPK Can Be Useful for Extrapolation

One of the major attractions of PBPK is the possibility of making predictions in situations where direct clinical observations are limited.

For example, a PBPK framework may be used to explore how PK could change with:

• Organ impairment.
• Age or developmental stage.
• Drug-drug interactions.
• Different ethnic or physiological populations.
• Pregnancy.
• Changes in enzyme or transporter activity.
• Different formulations or routes of administration.
• Unstudied dosing conditions.

The mechanistic structure provides a pathway for translating changes in physiology or drug properties into predicted changes in exposure.

However, mechanistic structure does not eliminate uncertainty. Predictions are only as credible as the assumptions, input parameters, and validation evidence supporting the model.

10. Where Population PK Is Especially Useful

Population PK is particularly useful when the scientific question concerns observed clinical variability.

• Estimating typical clearance and volume in patients.
• Quantifying between-subject variability.
• Identifying clinically relevant covariates.
• Supporting individualized dosing.
• Analyzing sparse sampling designs.
• Combining data from multiple studies.
• Characterizing PK in patient populations.
• Estimating individual exposure using Bayesian methods.
• Supporting therapeutic drug monitoring.
• Connecting clinical PK to PK/PD or exposure-response models.

Because PopPK models are built directly from clinical observations, they are especially valuable when the central question is how drug exposure behaves in the population that will actually receive the drug.

11. Where PBPK Is Especially Useful

PBPK is especially useful when mechanistic extrapolation is important and sufficient physiological and drug-specific information is available.

• First-in-human and early clinical development.
• Prediction of drug-drug interactions.
• Special populations.
• Organ impairment scenarios.
• Changes in enzyme or transporter activity.
• Formulation and absorption questions.
• Translation between species or development stages.
• Exploring untested physiological conditions.
• Mechanistically evaluating drug interaction pathways.
• Integrating in vitro and in vivo information.

The ability to connect drug properties and physiological processes to predicted concentrations is the defining feature of the PBPK approach.

12. Worked Example: The Same Drug Through Two Modeling Lenses

Consider a hypothetical drug administered intravenously to adults. Clinical data suggest that the drug has an apparent clearance of 5 L/h and a central volume of 25 L.

Step 1: Population PK interpretation

A simple one-compartment PopPK model could estimate:

The corresponding elimination rate constant would be:

and the approximate half-life would be:

The PopPK analysis could then investigate whether body weight, renal function, or another covariate explains variation in clearance among patients.

Step 2: PBPK interpretation

A PBPK model would not ordinarily begin by treating 5 L/h as the complete mechanistic description of elimination. Instead, it might represent hepatic and renal processes separately.

For example, a simplified systemic clearance relationship could be written as:

The hepatic component could itself depend on hepatic blood flow, unbound fraction, intrinsic metabolic clearance, and the assumptions of the selected hepatic model.

Thus, the two approaches can arrive at a similar observed concentration-time profile while giving different interpretations of what produces that profile.

13. How Are PBPK and PopPK Models Developed?

Population PK workflow

1. Define the scientific question.
2. Assemble dosing and concentration data.
3. Explore concentration-time observations.
4. Select a structural PK model.
5. Estimate typical population parameters.
6. Estimate between-subject and residual variability.
7. Evaluate covariates.
8. Perform model diagnostics.
9. Validate the model using appropriate internal or external information.
10. Use the model for prediction, simulation, or exposure-response analysis.

PBPK workflow

1. Define the physiological system and intended use.
2. Assemble physiological parameters.
3. Characterize drug-specific properties.
4. Represent absorption, distribution, metabolism, and excretion mechanisms.
5. Estimate or obtain uncertain drug-specific parameters.
6. Calibrate the model where appropriate.
7. Verify predictions against observed data.
8. Perform sensitivity and uncertainty analyses.
9. Simulate the target population or scenario.
10. Use the model within its qualified domain of applicability.

14. Validation Means Different Things in the Two Frameworks

Model evaluation is essential for both PBPK and PopPK, but the relevant questions differ.

A model should therefore be evaluated in relation to its intended use. A model developed to describe clinical concentration data may require different evidence than a model intended to predict a drug-drug interaction or a special population that was not directly studied.

15. Uncertainty in PBPK and PopPK Models

Both approaches contain uncertainty, but the sources may differ.

Population PK uncertainty

• Limited concentration data.
• Sampling imbalance.
• Model misspecification.
• Uncertainty in random-effect estimates.
• Uncertainty in covariate relationships.
• Uncertainty when extrapolating beyond the studied population.

PBPK uncertainty

• Uncertainty in physiological parameters.
• Uncertainty in drug-specific properties.
• Uncertainty in enzyme or transporter activity.
• Uncertainty in tissue partitioning.
• Uncertainty in intrinsic clearance.
• Structural uncertainty in mechanistic assumptions.

PBPK models can contain many more mechanistic inputs than a conventional PopPK model. This makes systematic sensitivity and uncertainty analysis especially important.

16. Can PBPK and Population PK Be Used Together?

Yes. In modern pharmacometrics, PBPK and PopPK can be complementary rather than competing approaches.

A mechanistic PBPK model can provide information about plausible physiological relationships, while a PopPK analysis can quantify observed clinical variability and identify covariates in patient data.

17. Which Modeling Approach Fits the Question?

The appropriate framework depends primarily on the scientific question, available information, and intended use of the model.

The table is a guide to the type of question each framework naturally addresses, not a rule that only one modeling approach can be used.

18. Common Misconceptions About PBPK and PopPK

Misconception 1: PBPK is always more mechanistic, so it is automatically more accurate.

PBPK contains more explicit physiological structure, but greater complexity does not guarantee more accurate predictions. Accuracy depends on the quality of the model structure, input data, assumptions, and verification evidence.

Misconception 2: PopPK is purely empirical.

PopPK models can incorporate substantial mechanistic knowledge. Covariates can be selected based on physiological relationships, and structural models can be informed by prior PK knowledge.

Misconception 3: PBPK does not need clinical data.

PBPK can make predictions before extensive clinical data are available, but clinical data are highly valuable for model verification, refinement, and assessment of predictive performance.

Misconception 4: A PBPK model cannot contain statistical variability.

PBPK simulations can incorporate distributions of physiological and biological parameters and can therefore represent substantial interindividual variability.

Misconception 5: PopPK cannot extrapolate.

PopPK models can be used for simulation and prediction, particularly when the target scenario is sufficiently represented by the development data. Extrapolation beyond the evidence supporting the model requires additional justification.

19. PBPK and PopPK Across Drug Development

20. A Practical Modeling Workflow

1. Start with the scientific question. Define exactly what needs to be described, explained, predicted, or extrapolated.
2. Identify the available evidence. Separate clinical concentration data from physiological, in vitro, and preclinical information.
3. Determine the intended use. Describing observed clinical PK may require a different framework from predicting an unstudied drug-drug interaction.
4. Choose an appropriate structural framework. This may be PopPK, PBPK, or a combination of approaches.
5. Specify assumptions explicitly. Important assumptions should be identifiable and scientifically defensible.
6. Estimate or obtain uncertain parameters. Distinguish measured, estimated, literature-derived, and assumed inputs.
7. Evaluate the model. Examine goodness of fit, predictive performance, biological plausibility, sensitivity, and uncertainty as appropriate.
8. Validate for the intended use. Verification should be relevant to the predictions the model is expected to make.
9. Simulate scenarios carefully. Clearly distinguish observed evidence from model-based predictions.
10. Update the model as evidence accumulates. New clinical or mechanistic data can change the appropriate parameter estimates or model structure.

Where to Go Next

A natural progression after comparing PBPK and population PK is to study each framework in greater depth.

For PBPK, useful next topics include physiological organ models, tissue partition coefficients, hepatic clearance models, renal clearance, enzyme and transporter modeling, drug-drug interactions, and virtual populations.

For PopPK, the next steps include nonlinear mixed-effects modeling, interindividual variability, residual error models, covariate modeling, model diagnostics, Bayesian individual estimation, and simulation.

Once both frameworks are understood separately, an especially useful topic is integrating PBPK and population PK approaches in model-based drug development.

References and Further Reading

1. Rowland M, Tozer TN. Clinical Pharmacokinetics and Pharmacodynamics: Concepts and Applications. Wolters Kluwer.
2. Ette EI, Williams PJ. Population pharmacokinetics I: background, concepts, and models. Annals of Pharmacotherapy.
3. Jones HM, Rowland-Yeo K. Basic concepts in physiologically based pharmacokinetic modeling in drug discovery and development. CPT: Pharmacometrics & Systems Pharmacology.
4. Jamei M. Recent advances in development and application of physiologically based pharmacokinetic models: a transition from academic curiosity to regulatory acceptance. Current Pharmacology Reports.
5. FDA. Physiologically Based Pharmacokinetic Analyses — Format and Content: Guidance for Industry.
6. FDA. Population Pharmacokinetics — Guidance for Industry.
7. EMA. Guideline on the reporting of physiologically based pharmacokinetic modelling and simulation.