1. What Is Hemodynamics?
Hemodynamics describes the movement of blood through the cardiovascular system and the physical relationships among pressure, flow, vascular resistance, cardiac output, and blood volume.
In quantitative systems pharmacology (QSP), hemodynamics provides a mechanistic framework for connecting a drug's molecular or pharmacologic effects to changes in cardiovascular physiology. Rather than treating blood pressure as an isolated endpoint, a QSP model can represent blood pressure as an emergent consequence of interacting physiological processes.
A cardiovascular QSP model links drug effects to interacting hemodynamic processes rather than modeling a blood-pressure endpoint in isolation.
2. What Questions Can a Hemodynamic QSP Model Answer?
A hemodynamic QSP model can address questions that require integration across multiple physiological processes.
| Question | Model component | What it helps describe |
|---|---|---|
| What determines arterial pressure? | MAP, cardiac output, vascular resistance | The interaction between pressure generation and systemic vascular load |
| How does a drug lower blood pressure? | Pharmacodynamic mechanism | Changes in cardiac function, vascular tone, fluid balance, or combinations of these |
| Why does heart rate change after treatment? | Autonomic regulation | Baroreflex-mediated and direct drug effects on heart rate |
| How does vascular resistance affect flow? | Resistance-flow relationships | The dependence of pressure gradients and flow on vascular tone |
| Why can blood pressure recover after an acute perturbation? | Feedback control | Homeostatic responses such as baroreflex regulation |
| What happens under repeated dosing? | Dynamic drug-system model | The interaction between drug exposure, effect, and physiological adaptation |
The key distinction from a purely empirical model is that the QSP framework attempts to explain the observed response through interconnected physiological mechanisms.
3. The Core Hemodynamic Variables
Most mechanistic cardiovascular models are built around a relatively small set of fundamental variables. These variables are related mathematically, but each represents a different aspect of cardiovascular physiology.
| Variable | Meaning | Role in a QSP model |
|---|---|---|
| MAP | Mean arterial pressure | Represents the average arterial pressure driving systemic blood flow |
| CO | Cardiac output | Total blood flow generated by the heart per unit time |
| HR | Heart rate | Frequency of cardiac contractions |
| SV | Stroke volume | Blood ejected by the ventricle per beat |
| SVR | Systemic vascular resistance | Represents the resistance opposing systemic blood flow |
| BV | Blood volume | Provides a state variable connecting fluid balance with filling pressures and cardiac output |
A particularly important relationship is:
Cardiac output can therefore change because heart rate changes, stroke volume changes, or both.
Under a simplified systemic circulation representation, mean arterial pressure is related to cardiac output and systemic vascular resistance by:
More precisely, the pressure driving systemic flow is related to the difference between mean arterial pressure and venous pressure:
where \(P_{CV}\) represents an appropriate central venous or downstream pressure approximation. The simpler \(MAP\approx CO\times SVR\) relationship is often useful when downstream venous pressure is small relative to arterial pressure.
4. Pressure, Flow, and Resistance
At the level of a simplified vascular bed, blood flow depends on the pressure difference across the vascular system and the resistance to flow.
Equivalently:
This relationship is analogous to other transport systems in physiology and engineering. A pressure difference provides the driving force, while vascular resistance determines how much flow results from that pressure difference.
In a simplified vascular representation, flow is determined by the pressure gradient and resistance.
For QSP modeling, the importance of this relationship is that a drug can alter pressure indirectly. For example, a vasodilator may reduce vascular resistance, which can lower arterial pressure even when cardiac output changes in response.
5. Vascular Resistance and Vessel Radius
For idealized laminar flow through a cylindrical tube, Poiseuille's law illustrates the strong dependence of resistance on vessel radius:
Here, \(\eta\) is viscosity, \(L\) is vessel length, and \(r\) is vessel radius.
The fourth-power dependence on radius means that relatively small changes in vascular radius can produce large changes in theoretical hydraulic resistance. Real blood vessels are not rigid cylindrical tubes, and the circulation is substantially more complex, but the relationship provides useful mechanistic intuition for why vascular tone can have a large influence on systemic hemodynamics.
6. Cardiac Output, Stroke Volume, and Heart Rate
Cardiac output is the amount of blood pumped by the heart per unit time:
Stroke volume can be expressed as the difference between end-diastolic and end-systolic volume:
Therefore:
This decomposition is important for mechanistic modeling because drugs can influence cardiac output through different pathways.
| Mechanism | Potential model variable affected | Hemodynamic consequence |
|---|---|---|
| Chronotropy | Heart rate | Changes cardiac output through HR |
| Inotropy | Contractility / ESV | Changes stroke volume and potentially cardiac output |
| Preload | EDV / filling | Changes ventricular filling and stroke volume |
| Afterload | Vascular resistance / pressure | Can alter ventricular ejection and stroke volume |
A QSP model can explicitly represent these pathways when the scientific question requires them, rather than treating cardiac output as a single unexplained variable.
7. Blood Volume and Cardiovascular Filling
Blood volume is an important state variable in many cardiovascular systems models because changes in fluid volume can influence venous return, cardiac filling, cardiac output, and arterial pressure.
A simplified mass-balance representation can be written as:
The actual physiology can be expanded to include renal sodium and water handling, vascular permeability, interstitial fluid exchange, hormonal regulation, and other processes.
In a QSP model, this creates a connection between cardiovascular hemodynamics and other physiological systems. For example, a drug that changes renal sodium handling may alter extracellular fluid volume, which can subsequently affect cardiovascular filling and blood pressure.
8. The Baroreflex: A Key Homeostatic Mechanism
The cardiovascular system contains feedback mechanisms that oppose acute changes in arterial pressure. The baroreflex is one of the most important of these mechanisms.
In a simplified representation, arterial pressure is sensed by baroreceptors, and changes in the sensed pressure influence autonomic activity. Autonomic activity can then alter heart rate, contractility, and vascular tone.
A simplified baroreflex closes a feedback loop between arterial pressure and cardiovascular effectors.
A useful conceptual representation is:
The feedback loop means that the final effect of a drug may differ from its immediate direct pharmacologic effect. A direct decrease in vascular resistance, for example, may trigger compensatory changes in heart rate or contractility.
9. How Drug Effects Enter a Hemodynamic QSP Model
The pharmacologic component of a QSP model describes how drug exposure changes one or more physiological processes.
A simple direct-effect relationship might use an \(E_{\max}\) model:
Depending on the drug and mechanism, the effect could modify vascular resistance, heart rate, contractility, renal sodium handling, venous tone, or another physiological process.
For example, if drug concentration reduces vascular resistance, a simplified relationship might be:
The hemodynamic model then propagates that change through the cardiovascular system.
10. From Physiology to Differential Equations
QSP models represent physiological processes using differential equations, algebraic equations, or combinations of both.
Suppose a simplified model contains blood volume \(BV\), cardiac output \(CO\), and systemic vascular resistance \(SVR\). A basic algebraic pressure relationship might be:
If blood volume changes dynamically, a simple mass-balance equation could be:
Cardiac output can be represented as:
And stroke volume can depend on physiological state:
These equations form a connected system rather than independent formulas. A change in one state variable can therefore propagate through several equations and ultimately alter an observable such as blood pressure.
11. Cardiovascular Compartments in QSP Models
Unlike a simple PK compartment, a cardiovascular compartment can represent a physiologically meaningful blood or tissue space. The exact structure depends on the purpose of the model.
| Component | Potential representation | Typical role |
|---|---|---|
| Heart | Cardiac pump | Generates pressure and flow |
| Arterial compartment | Systemic arterial blood volume | Represents arterial pressure and compliance |
| Venous compartment | Systemic venous blood volume | Provides a reservoir and affects venous return |
| Peripheral vascular beds | Resistance elements | Represent regional vascular tone and flow |
| Pulmonary circulation | Pulmonary vascular compartment | Connects right and left cardiac output |
| Renal system | Fluid and electrolyte regulation | Links kidney function to volume and pressure |
More detailed models may subdivide the systemic circulation into multiple vascular beds, such as renal, cerebral, splanchnic, skeletal muscle, and cutaneous circulations. This can be useful when drug effects differ substantially among tissues.
12. Compliance and the Relationship Between Volume and Pressure
Blood vessels are compliant rather than perfectly rigid. Compliance describes how much volume changes for a given change in pressure:
For a simplified linear representation:
Arterial compliance is particularly important because arterial pressure depends not only on flow and resistance but also on how the vascular system stores and releases blood during the cardiac cycle.
A QSP model can therefore distinguish between:
- Resistance: opposition to blood flow.
- Compliance: the volume response to changes in pressure.
- Elastance: the inverse of compliance in appropriate formulations.
- Volume: the amount of blood contained within a vascular compartment.
Including these properties allows the model to represent dynamic pressure responses rather than only steady-state relationships.
13. Worked Example: How a Change in Vascular Resistance Changes Pressure
Consider a simplified cardiovascular system with:
- Cardiac output: 5 L/min
- Systemic vascular resistance: 18 mmHg·min/L
- Central venous pressure: 3 mmHg
Step 1: Calculate mean arterial pressure
Using the simplified pressure-flow relationship:
Step 2: Introduce a vasodilator effect
Suppose the drug reduces systemic vascular resistance by 25%, while cardiac output initially remains at 5 L/min.
Step 3: Recalculate pressure
Step 4: Interpret the result
Under these deliberately simplified assumptions, a 25% reduction in systemic vascular resistance lowers mean arterial pressure from approximately 93 to 70.5 mmHg.
A real QSP model would not necessarily stop here. The fall in arterial pressure could activate baroreflex responses, alter heart rate and contractility, change venous tone, and potentially modify cardiac output. The final predicted pressure would therefore emerge from the complete interconnected model.
14. Why the Direct Drug Effect Is Not Always the Final Effect
One of the central reasons to use QSP is that biological systems contain feedback.
Suppose a drug directly decreases vascular resistance. The immediate effect may be a fall in arterial pressure. The cardiovascular system can then respond through homeostatic mechanisms.
| Step | Mechanistic event |
|---|---|
| 1 | Drug concentration increases. |
| 2 | Drug reduces vascular resistance. |
| 3 | Mean arterial pressure decreases. |
| 4 | Baroreceptor signaling changes. |
| 5 | Autonomic activity changes heart rate and contractility. |
| 6 | Cardiac output and vascular tone change. |
| 7 | The final pressure response reflects the combined direct and compensatory effects. |
This is an example of why a mechanistic QSP model can provide information that would be difficult to obtain from a simple concentration-versus-blood-pressure regression.
15. Acute Responses Versus Steady-State Responses
Hemodynamic systems can respond on multiple time scales.
| Time scale | Examples | Potential QSP processes |
|---|---|---|
| Seconds to minutes | Heart rate, vascular tone | Autonomic responses and direct vascular effects |
| Minutes to hours | Cardiac filling, redistribution of blood flow | Changes in venous return and vascular compartments |
| Hours to days | Fluid balance | Renal sodium and water handling |
| Days to weeks | Physiological adaptation | Longer-term neurohormonal and tissue responses |
A model that only represents an acute response may be insufficient for questions about chronic treatment. Conversely, a highly detailed long-term model may be unnecessary when the scientific question concerns an immediate pharmacologic response.
Model scope should therefore be determined by the time scale of the scientific question.
16. Representing Interindividual Variability
Patients do not all have the same cardiovascular physiology. QSP models can represent variability in baseline parameters and physiological responses.
For example, systemic vascular resistance might vary among individuals:
where \(SVR_{pop}\) is a typical population value and \(\eta_i\) represents an individual-specific deviation.
Similar variability can be introduced for:
- Cardiac output.
- Heart rate.
- Vascular compliance.
- Drug sensitivity.
- Renal fluid handling.
- Baroreflex sensitivity.
Covariates can then be incorporated when scientifically justified. Age, body size, disease state, baseline cardiovascular function, and concomitant treatment may influence model parameters.
17. Extending Hemodynamic Models to Disease
A major advantage of QSP is the ability to represent disease as a perturbation of physiological mechanisms rather than simply as a categorical label.
For example, a cardiovascular disease model might represent changes in:
- Baseline vascular resistance.
- Arterial compliance.
- Cardiac contractility.
- Preload and venous return.
- Renal sodium and water handling.
- Autonomic regulation.
- Neurohormonal signaling.
The disease state can then be represented by altered parameter values, additional mechanisms, or both.
This creates a mechanistic bridge between normal physiology, disease physiology, and pharmacologic intervention:
Such a framework can be useful for understanding why the same pharmacologic mechanism may produce different responses in different disease states.
18. Modeling Multiple Drugs
Cardiovascular treatments often act through different mechanisms. A QSP model can represent these mechanisms separately and allow them to interact through the shared physiological system.
| Drug mechanism | Potential system target | Possible downstream effect |
|---|---|---|
| Vasodilation | Vascular resistance | Reduced pressure load |
| Positive inotropy | Contractility | Altered stroke volume and cardiac output |
| Chronotropic effect | Heart rate | Altered cardiac output |
| Diuresis / natriuresis | Fluid balance | Reduced circulating volume |
| Neurohormonal modulation | Regulatory pathways | Changes in vascular tone and fluid balance |
Because the drugs act on a shared physiological system, their combined response does not necessarily equal the sum of their individual effects. Feedback and nonlinear relationships can produce interactions that emerge only when the complete system is simulated.
19. What Can a Hemodynamic QSP Model Simulate?
Once a model has been calibrated and evaluated, simulations can be used to explore scenarios that may be difficult or impractical to test experimentally.
- Changes in blood pressure following a dose.
- Heart-rate responses to changes in arterial pressure.
- Effects of altered vascular resistance.
- Effects of changes in cardiac contractility.
- Responses to different dosing regimens.
- Combination-treatment scenarios.
- Differences between physiological and disease states.
- Potential effects of changes in renal fluid handling.
- Mechanistic explanations for interindividual variability.
For example, a virtual population can be simulated by sampling physiological and pharmacological parameters from distributions representing plausible patient variability.
The resulting simulations can then be summarized as predicted distributions of blood pressure, heart rate, cardiac output, or other endpoints rather than as a single deterministic trajectory.
20. How Is a Hemodynamic QSP Model Built?
Developing a useful QSP model requires translating biological knowledge into a mathematically coherent system.
- Define the scientific question. Determine what cardiovascular behavior the model needs to explain or predict.
- Define the physiological scope. Decide which cardiovascular, renal, neurohormonal, and pharmacologic processes are necessary.
- Define model states and parameters. Identify quantities that change dynamically and parameters that determine their behavior.
- Write mechanistic relationships. Convert physiological assumptions into differential and algebraic equations.
- Connect pharmacology to physiology. Specify how drug concentration or target engagement modifies physiological processes.
- Calibrate the model. Estimate uncertain parameters using appropriate experimental and clinical data.
- Evaluate model behavior. Compare predictions with observed data and assess whether the model reproduces important physiological features.
- Perform sensitivity analysis. Determine which parameters and mechanisms most strongly influence the outputs of interest.
- Simulate scenarios. Use the model to investigate doses, disease states, combinations, or physiological perturbations.
The objective is not to reproduce every detail of cardiovascular physiology. The objective is to construct a model that contains enough mechanistic structure to answer the intended scientific questions.
21. Sensitivity Analysis in Hemodynamic QSP
Sensitivity analysis asks how strongly model outputs depend on model parameters or mechanisms.
A local normalized sensitivity can be represented conceptually as:
where \(y\) is an output such as MAP and \(p\) is a model parameter such as vascular resistance, cardiac contractility, or drug potency.
High sensitivity indicates that uncertainty in the parameter can have a substantial influence on the predicted outcome.
Sensitivity analysis can therefore help identify:
- Which physiological mechanisms dominate a predicted response.
- Which parameters require more precise estimation.
- Which measurements would be most informative in future experiments.
- Where model uncertainty is likely to have the greatest effect on predictions.
22. Identifiability and the Limits of Hemodynamic Data
A mechanistically rich model can contain more parameters than the available data can reliably determine.
For example, if only blood pressure is measured, it may be difficult to determine whether a change arose primarily from cardiac output, systemic vascular resistance, blood volume, or compensatory feedback.
Additional measurements such as:
- Heart rate.
- Cardiac output.
- Stroke volume.
- Blood pressure.
- Renal biomarkers.
- Hormonal measurements.
- Drug concentrations.
can provide additional constraints on the model.
This is a fundamental QSP principle: model complexity should be supported by information in the available data.
23. What Hemodynamic QSP Models Do Not Tell Us Automatically
A mechanistic model can be biologically informed and mathematically sophisticated while still being subject to important limitations.
- Model structure is a simplification. Even a detailed cardiovascular QSP model represents only selected aspects of physiology.
- Parameter estimates depend on model assumptions. Changing the structure can change estimated parameter values.
- A good fit does not establish uniqueness. Multiple mechanisms may sometimes explain the same observed endpoint.
- Unmeasured mechanisms can remain uncertain. A model may contain processes that are weakly constrained by available data.
- Predictions are conditional. Extrapolation beyond the conditions used for calibration requires additional justification.
- Feedback can amplify or attenuate direct drug effects. The final physiological response may differ from the immediate pharmacologic action.
- Virtual populations are model-based. Simulated variability depends on the assumptions and distributions used to construct the population.
24. A Practical Hemodynamic QSP Workflow
- Start with the pharmacologic question. What mechanism or intervention needs to be understood?
- Map the physiological pathway. Identify how the drug could affect cardiovascular function.
- Define the minimum necessary system. Include the mechanisms needed to answer the question without unnecessary complexity.
- Write the governing equations. Translate assumptions about pressure, flow, resistance, volume, and feedback into mathematical relationships.
- Connect drug exposure to mechanism. Use an appropriate pharmacodynamic relationship.
- Calibrate against multiple data types. Where available, integrate drug concentration, blood pressure, heart rate, cardiac output, and other measurements.
- Evaluate model behavior. Check both quantitative fit and qualitative physiological behavior.
- Perform sensitivity and identifiability analyses. Determine which parameters and mechanisms are supported by the available information.
- Simulate clinically relevant scenarios. Explore doses, disease states, combinations, and patient variability.
- Use predictions to generate testable hypotheses. QSP is most useful when model predictions can inform subsequent experiments or clinical development decisions.
25. Key Takeaways
- Hemodynamics describes the relationships among blood flow, pressure, vascular resistance, cardiac function, and blood volume.
- QSP models represent these variables as an interconnected physiological system rather than isolated clinical endpoints.
- Cardiac output is determined by heart rate and stroke volume: \(CO=HR\times SV\).
- Mean arterial pressure is closely related to cardiac output and systemic vascular resistance, with downstream venous pressure included when appropriate.
- Vascular resistance provides a mechanistic link between vascular tone and blood flow.
- Blood volume connects cardiovascular physiology with renal and fluid-balance mechanisms.
- The baroreflex illustrates how physiological feedback can modify the direct effect of a drug.
- Drug effects can enter a QSP model by modifying vascular resistance, heart rate, contractility, fluid balance, neurohormonal pathways, or other mechanisms.
- QSP models can explain why the final physiological response may differ from the immediate pharmacologic effect.
- Mechanistic cardiovascular models can incorporate disease states, combination therapy, and interindividual variability.
- Sensitivity analysis helps identify mechanisms and parameters that most strongly influence predicted outcomes.
- Identifiability depends on the information contained in the available data; a biologically plausible mechanism is not necessarily identifiable from every dataset.
- The appropriate model is not necessarily the most detailed model. It is the model with sufficient mechanistic structure to address the scientific question and available data.
Where to Go Next
A natural progression from hemodynamic QSP is to study QSP models of cardiovascular disease, where blood pressure and flow are connected to disease mechanisms such as altered vascular resistance, cardiac dysfunction, fluid retention, and neurohormonal activation.
From there, increasingly integrated models can connect hemodynamics with renal physiology, glucose-insulin regulation, inflammation, pharmacokinetics, pharmacodynamics, and disease progression.
The next tutorial can build directly on the framework introduced here by examining how a cardiovascular QSP model represents cardiac function, vascular resistance, blood volume, and baroreflex regulation together.