When a patient takes a medication, the drug does not instantly and uniformly distribute throughout their entire body. Instead, it travels through the bloodstream, permeates vascular walls, diffuses into tissues, binds to cellular targets, and undergoes metabolic clearance. Traditional pharmacology often simplifies this journey by treating the body as a single, well-mixed bucket (a one-compartment model). While computationally straightforward, this abstraction fails when a drug takes time to transition between blood and peripheral tissues, or when local tissue architecture dictates its therapeutic effect.
To capture this reality, researchers use multi-compartment and physiologically based pharmacokinetic (PBPK) modeling. By representing the body as an interconnected network of distinct physiological zones governed by systems of differential equations, these models allow us to track not just how much drug is in the body, but where it is over time.
From One Bucket to Multi-Compartment Networks
A one-compartment model assumes instantaneous distribution. As Zhu et al. and Jiang et al. demonstrate, this simplification can successfully characterize population-level clearance rates for certain drugs where entry and exit rates dominate. However, for drugs that undergo extensive or slow tissue distribution, a multi-compartment framework is necessary.
A typical two-compartment model splits the body into a central compartment (highly perfused tissues like blood and organs) and a peripheral compartment (poorly perfused tissues like fat and muscle). As Mao et al. explain, accounting for this distribution phase significantly reduces prediction errors immediately following an intravenous infusion. Without it, a model might misinterpret a rapid drop in blood concentration as drug elimination, when the drug is actually just moving into the surrounding tissues.
This distinction is vital in critical clinical settings. For instance, Wu utilizes a two-compartment model to map the complex disposition of immunosuppressants in pediatric liver transplant patients, where body weight and changing physiology alter tissue distribution. Similarly, Xie et al. apply systems of differential equations to capture the non-instantaneous distribution phase of colistin sulfate in critically ill patients, allowing for personalized dosing based on renal and liver function. By tracking the exchange rates between these compartments, systems can even automate drug delivery; Shanechi et al. leverage a multi-compartment linear model within a brain-machine interface to account for metabolic delays and precisely maintain medically induced comas.
Grounding Math in Anatomy: PBPK Modeling
While abstract compartments rely on statistical curve-fitting of blood samples, Physiologically Based Pharmacokinetic (PBPK) modeling grounds the mathematical architecture in actual anatomy. Instead of arbitrary "central" or "peripheral" boxes, a PBPK model maps out specific organs (e.g., liver, kidneys, lungs) connected by the actual blood flow rates of the species being studied.
As Zhang et al. demonstrate with a 14-compartment model, this mechanistic approach integrates real-world physiological data (like organ blood flow) with the drug's chemical properties (like solubility and metabolic stability). This allows researchers to predict human drug exposure within narrow error margins before extensive clinical trials even begin. Yuan et al. emphasize that by incorporating these explicit physiological and biochemical parameters, PBPK models become essential for optimizing multi-component therapies and predicting complex drug-drug interactions.
Furthermore, PBPK models can account for physical barriers within organs. Ai et al. show how combining flow-limited modules with membrane-rate-limiting modules improves the accuracy of predicting drug residues in complex biological systems where physical membranes restrict distribution.
Bridging the Gap: In Vitro to In Vivo Translation
One of the most powerful applications of PBPK modeling is its ability to serve as a predictive bridge between laboratory benchwork and human biology. Hermans et al. highlight how in silico physiological absorption modeling can simulate how variations in a drug's laboratory dissolution rate impact its actual systemic absorption in humans. This reduces the need to run expensive clinical trials for every minor formulation change.
This predictive power is also transforming regulatory approvals. Zhang et al. demonstrate that by integrating laboratory-measured particle size distributions with regional lung deposition patterns, PBPK models can verify the bioequivalence of generic inhaled aerosols against brand-name products, satisfying regulatory standards without exhaustive human testing.
Spatial Continuums and Localized Drug Delivery
When drug delivery is highly localized—such as from a drug-eluting stent or directly into a tumor—even organ-level compartments become too broad. Here, researchers turn to spatial continuum models that discretize target tissues into microscopic volume elements.
Hwang et al. argue that to optimize drug-eluting stents, models must move beyond average concentrations to account for local diffusion, convection, and tissue binding. Using a 2-dimensional transient finite volume model, Balakrishnan et al. show that optimal arterial drug retention occurs when the drug's release kinetics from the stent coating precisely match the tissue's physical absorption rate.
Similarly, in oncology, Gasselhuber demonstrates through a multi-compartment model that tracking the movement of smart drug carriers between vascular, extracellular, and intracellular spaces allows for a quantitative comparison of therapies, showing how temperature-sensitive liposomes can maximize tumor cell uptake while minimizing systemic toxicity.
Integrating Pharmacokinetics with Pharmacodynamics (PK/PD)
Ultimately, knowing where a drug goes (pharmacokinetics) is only useful if we can predict what it does (pharmacodynamics). Modern frameworks integrate these two layers. As Sprunk et al. outline, a dual-layer approach tracks the distribution across anatomical compartments and then applies mathematical functions (like the Hill equation) to determine the resulting physiological impact.
This integration enables model-informed clinical decisions. Wu et al. emphasize that combining blood concentration monitoring with clinical response metrics provides a comprehensive understanding of therapeutic effects. By scaling these integrated models to virtual patient populations, researchers like Byun et al. can predict minimally effective doses for clinical trials, while Huang et al. can quantify how unexpected covariates—such as co-administered drugs altering intestinal permeability—impact the systemic clearance of advanced therapeutics.
Key Foundational Work
Correlation of transarterial transport of various dextrans with their physicochemical properties (2000) — contributed 1 concept elements to this direction.
Impact of transport and drug properties on the local pharmacology of drug-eluting stents (2003) — contributed 1 concept elements to this direction.