Tutorials › Pharmacometrics › Indirect Response Stimulation Models
Pharmacokinetics · PK/PD Modeling

Indirect Response Stimulation Models

Learn how indirect response stimulation models describe drug effects that increase the production or appearance of a response, why the response can persist after drug concentrations fall, and how the model parameters connect drug exposure to response dynamics.

IntermediatePK/PD ModelingIndirect ResponsePharmacometrics
01 · The big picture

1. What Is an Indirect Response Stimulation Model?

An indirect response stimulation model is a pharmacodynamic model in which a drug increases the rate at which a response variable is produced, generated, or appears. Instead of assuming that drug concentration directly determines the response instantaneously, the model describes the response as the result of a dynamic process.

This distinction is important because many drug effects are not immediate reflections of plasma concentration. A drug may stimulate the production of a biomarker, increase the synthesis of a physiological mediator, or enhance the appearance of an endogenous response. The response can therefore continue to change even after the concentration has begun to decline.

Drug C(t) Stimulated production kin · stimulation − kout · R Response R(t) Drug changes the response indirectly by changing its production rate.

In an indirect response stimulation model, drug concentration modifies the rate of response production rather than determining the response instantaneously.

Core idea: the drug acts on the rate of change of the response. This introduces a dynamic response process that can produce delays, hysteresis, and persistence of effect even when plasma concentrations change rapidly.
02 · Why indirect response?

2. Why Do We Need an Indirect Response Model?

A direct-effect model assumes that response follows concentration according to a static relationship such as an Emax model:

$$E(C)=E_0+\frac{E_{\max}C}{EC_{50}+C}$$

Such a model can be appropriate when the effect is closely linked to the measured concentration. However, it becomes inadequate when the biological response requires an intermediate process.

SituationWhy a direct model may be insufficient
Drug stimulates biomarker synthesisThe biomarker must be produced over time before the response changes substantially.
Drug increases an endogenous mediatorThe mediator has its own turnover and persistence.
Drug stimulates cell productionNew cells or cellular products cannot appear instantaneously.
Response shows delayed onsetThe observed delay may arise from turnover rather than distribution alone.
Response persists after drug declineThe response compartment can retain memory of prior drug exposure.

In these situations, the response is governed by a differential equation rather than a static concentration-effect equation.

03 · The basic model

3. The Fundamental Indirect Response Equation

Let R(t) denote the response at time t. A general turnover model can be written as:

$$\frac{dR}{dt}=R_{\mathrm{in}}-R_{\mathrm{out}}$$

where:

  • Rin is the rate of response production or input.
  • Rout is the rate of response loss or output.

For a first-order loss process:

$$R_{\mathrm{out}}=k_{\mathrm{out}}R$$

If the drug stimulates production, the input term becomes concentration-dependent:

$$\frac{dR}{dt}=R_{\mathrm{in}}\,f(C)-k_{\mathrm{out}}R$$

Here, f(C) describes how drug concentration changes the production rate.

Important distinction: the drug does not directly set R. It changes the rate at which R is produced, while the response simultaneously turns over according to its own elimination or loss process.
04 · Baseline response

4. The Baseline Response and Turnover

Before drug administration, assume the system is at steady state. With no drug effect, the production rate equals the loss rate:

$$R_{\mathrm{in}}=k_{\mathrm{out}}R_0$$

Therefore:

$$R_{\mathrm{in}}=k_{\mathrm{out}}R_0$$

where R0 is the baseline response.

This relationship is central to indirect response models because it means that baseline response and turnover determine the underlying production rate. If the baseline response and the turnover rate are known or estimated, the production rate can be derived.

Response half-life

The response turnover rate can also be expressed through a response half-life:

$$t_{1/2,R}=\frac{\ln(2)}{k_{\mathrm{out}}}$$

A small kout corresponds to slow turnover and a long response half-life. A large kout corresponds to rapid turnover and a short response half-life.

05 · Stimulation function

5. How Does Drug Concentration Stimulate Production?

A common choice is an Emax-type stimulation function:

$$f(C)=1+\frac{S_{\max}C}{SC_{50}+C}$$

The indirect response model then becomes:

$$\frac{dR}{dt}=R_{\mathrm{in}}\left(1+\frac{S_{\max}C}{SC_{50}+C}\right)-k_{\mathrm{out}}R$$
ParameterInterpretation
SmaxMaximum fractional stimulation above baseline production.
SC50Concentration producing half of the maximum stimulatory increase.
RinBaseline production rate.
koutFirst-order response loss or turnover rate.

At zero concentration:

$$f(0)=1$$

so the system returns to its baseline production rate. At very high concentration:

$$\lim_{C\rightarrow\infty}f(C)=1+S_{\max}$$

Thus, Smax represents the maximum fractional increase in production relative to the baseline production rate.

06 · The classic framework

6. Four Indirect Response Model Types

Indirect response models are commonly organized according to whether the drug stimulates or inhibits production or loss of the response.

ModelDrug actionGeneral form
Type IStimulates response productionDrug increases Rin
Type IIInhibits response lossDrug decreases kout
Type IIIInhibits response productionDrug decreases Rin
Type IVStimulates response lossDrug increases kout

This tutorial focuses on Type I stimulation, where drug concentration increases the production or input rate of the response.

07 · Dynamic behavior

7. Why Does the Response Lag Behind Concentration?

The key source of delay is the response turnover process. Even if drug concentration changes rapidly, the response cannot necessarily change at the same rate because it is governed by its own production and loss.

Suppose concentration suddenly increases. The stimulation term increases immediately, but the response itself must accumulate over time.

Concentration Indirect response Time Magnitude

The response can rise and fall later than concentration because response production and turnover introduce a dynamic time scale.

This delay is sometimes described as hysteresis when response is plotted directly against concentration. The same concentration can correspond to different response values depending on whether concentration is rising or falling.

08 · Exposure-response loops

8. Indirect Response and Hysteresis

With a direct Emax model, response is a single-valued function of concentration. If concentration rises and then falls, the response follows the same concentration-effect relationship in both directions.

For an indirect response model, the response contains information about the history of prior drug exposure. As a result, a concentration-response plot can form a counterclockwise or clockwise loop.

Interpretation: hysteresis does not automatically imply that drug concentration has been measured incorrectly. It can be a mathematical consequence of an effect compartment, an indirect response process, or another delayed mechanism.

Indirect response models are therefore useful when the observed response cannot be adequately explained by an instantaneous concentration-effect relationship.

09 · Mathematical structure

9. Deriving the Stimulation Model

Start with the general turnover equation:

$$\frac{dR}{dt}=R_{\mathrm{in}}-k_{\mathrm{out}}R$$

At baseline, the system is at steady state:

$$R_{\mathrm{in}}=k_{\mathrm{out}}R_0$$

Now introduce drug stimulation through an Emax function:

$$S(C)=1+\frac{S_{\max}C}{SC_{50}+C}$$

The stimulated model becomes:

$$\frac{dR}{dt}=R_{\mathrm{in}}S(C)-k_{\mathrm{out}}R$$

Substituting the baseline relationship gives:

$$\frac{dR}{dt}=k_{\mathrm{out}}R_0\left(1+\frac{S_{\max}C}{SC_{50}+C}\right)-k_{\mathrm{out}}R$$

This form makes the biological interpretation particularly clear: baseline turnover establishes the response, while the drug modifies the production term.

10 · A useful limiting case

10. What Happens at Constant Drug Concentration?

Suppose concentration remains constant at C. The stimulation factor is then constant:

$$S=1+\frac{S_{\max}C}{SC_{50}+C}$$

The model becomes:

$$\frac{dR}{dt}=R_{\mathrm{in}}S-k_{\mathrm{out}}R$$

The steady-state response is obtained by setting the derivative to zero:

$$R_{ss}=\frac{R_{\mathrm{in}}S}{k_{\mathrm{out}}}$$

Because Rin = koutR0:

$$R_{ss}=R_0S$$

Therefore:

$$R_{ss}=R_0\left(1+\frac{S_{\max}C}{SC_{50}+C}\right)$$

This is useful because it shows that the long-run response at constant concentration follows an Emax-type relationship even though the transient response is governed by a differential equation.

11 · Response time course

11. The Response Approaches Its New Steady State Gradually

For a constant concentration, the solution to the first-order turnover model is:

$$R(t)=R_{ss}+\left[R(0)-R_{ss}\right]e^{-k_{\mathrm{out}}t}$$

If the drug is suddenly introduced and the initial response equals baseline, then:

$$R(0)=R_0$$

and therefore:

$$R(t)=R_{ss}+(R_0-R_{ss})e^{-k_{\mathrm{out}}t}$$

The response moves exponentially toward its new steady state with a time scale controlled by kout.

Response half-lifeApproximate fraction of the steady-state change completed
1 half-life50%
2 half-lives75%
3 half-lives87.5%
4 half-lives93.75%
5 half-lives96.875%

Thus, even if the drug reaches its target concentration rapidly, the pharmacodynamic response may take several response half-lives to approach its new equilibrium.

12 · Parameters

12. The Main Parameters

ParameterMeaningPrimary influence
R0Baseline responseSets the starting level of the response.
RinBaseline production rateDetermines the amount entering the response system per unit time.
koutResponse turnover rateControls the speed of response turnover.
t1/2,RResponse half-lifeProvides an interpretable time scale for response turnover.
SmaxMaximum stimulatory effectControls the maximum fractional increase in production.
SC50Concentration for half-maximal stimulationControls the concentration scale of the stimulation relationship.

These parameters have different roles. Smax and SC50 primarily determine the magnitude and concentration dependence of stimulation, while kout determines how rapidly the response system turns over.

14 · Worked example

14. Worked Example: A Drug That Stimulates Response Production

Consider a hypothetical drug with the following properties:

  • Baseline response: R0 = 100 units
  • Response turnover rate: kout = 0.20 h−1
  • Maximum stimulation: Smax = 2.0
  • Half-maximal stimulatory concentration: SC50 = 10 mg/L
  • Constant drug concentration: C = 10 mg/L

Step 1: Calculate baseline production

$$R_{\mathrm{in}}=k_{\mathrm{out}}R_0=(0.20)(100)=20\text{ units/h}$$

Step 2: Calculate the stimulation factor

$$S(C)=1+\frac{(2.0)(10)}{10+10}=1+1=2$$

Thus, at 10 mg/L, the drug doubles the baseline production rate.

Step 3: Calculate the stimulated production rate

$$R_{\mathrm{in,drug}}=20(2)=40\text{ units/h}$$

Step 4: Calculate the new steady-state response

$$R_{ss}=\frac{40}{0.20}=200\text{ units}$$

The new steady-state response is therefore 200 units, compared with the baseline of 100 units.

Step 5: Calculate the response after 5 hours

Assuming the drug concentration is established immediately and the response begins at baseline:

$$R(5)=200+(100-200)e^{-0.20(5)}$$
$$R(5)=200-100e^{-1}\approx163.2\text{ units}$$

After five hours, the response is approximately 163 units, even though the eventual steady-state response is 200 units. The response therefore continues to rise because the turnover process takes time to reach equilibrium.

What this example illustrates: the drug immediately changes the production rate, but the response changes gradually. The magnitude of the effect is governed by the stimulation function, while the speed of the response is governed by kout.
15 · Interpretation

15. How Should the Parameters Be Interpreted?

Smax: maximum stimulation

Smax determines the maximum fractional increase in response production. If Smax = 2, the maximum stimulated production rate is three times baseline because the stimulation factor approaches 1 + 2 = 3.

SC50: concentration scale

SC50 is the concentration producing half of the maximum additional stimulation. It is analogous to EC50 in a standard Emax relationship, but it is interpreted within the production-rate component of the indirect response model.

kout: response turnover

kout controls how rapidly the response system turns over. It determines the response half-life:

$$t_{1/2,R}=\frac{0.693}{k_{\mathrm{out}}}$$

Importantly, kout does not determine the maximum pharmacodynamic stimulation by itself. It primarily determines how quickly the response approaches changes in its equilibrium.

16 · Estimation

16. Estimating Indirect Response Parameters

Indirect response models are typically fitted simultaneously to concentration and response observations, or to response data using a previously established PK model.

  1. Specify the PK model. Describe the concentration-time profile driving the response.
  2. Specify the response turnover model. Define the baseline production and loss processes.
  3. Choose the stimulation function. An Emax-type function is one common option.
  4. Estimate the parameters. Estimate quantities such as kout, Smax, and SC50.
  5. Evaluate diagnostics. Compare predicted and observed response profiles and examine residual behavior.
  6. Assess identifiability. Determine whether the available concentration and response data contain enough information to estimate the parameters separately.
Identifiability matters: a delayed response can arise from several mechanisms. Without informative sampling, it may be difficult to distinguish slow response turnover from distributional delay, an effect compartment, or another mechanistic process.
17 · Study design

17. Sampling Requirements for Indirect Response Models

The sampling design should capture both the driving concentration profile and the delayed response.

Sampling objectiveWhy it matters
Early concentration samplesCharacterize the exposure driving the response.
Samples during concentration declineHelp determine whether response persists after concentration falls.
Response measurements throughout the effect periodProvide information about response turnover.
Late observationsHelp characterize return toward baseline.
Multiple concentration levelsImprove information about the stimulation function and SC50.

If response samples are collected only near the maximum effect, there may be insufficient information to estimate the turnover rate accurately. Likewise, without sufficient concentration information, the stimulation parameters may be difficult to separate from response turnover.

18 · Common mistakes

18. Common Interpretation Mistakes

  • Treating the response as an instantaneous function of concentration. The whole point of an indirect response model is that response dynamics matter.
  • Confusing Smax with the absolute response. Smax describes stimulation of the production process, not necessarily the maximum measured response itself.
  • Ignoring baseline turnover. R0 and kout define the underlying response system.
  • Interpreting kout as a PK elimination rate. It describes turnover of the response, not necessarily elimination of the drug.
  • Assuming delayed response proves an indirect mechanism. Distributional delays and other mechanisms can also produce delayed pharmacodynamic effects.
  • Ignoring the PK model. Incorrect concentration predictions can propagate directly into the estimated PD parameters.
  • Overinterpreting parameter estimates without diagnostics. A numerical fit alone does not establish that the mechanism is uniquely identified.
19 · Model comparison

19. Direct Emax Versus Indirect Response Models

FeatureDirect EmaxIndirect response stimulation
Response determined by concentrationYes, directlyNo; concentration changes response production
Dynamic response turnoverNot explicitly representedExplicitly represented
Delayed response possibleNot without an additional mechanismYes
Hysteresis possibleNot for a static relationshipYes
Response half-life parameterNot intrinsic to the basic modelkout defines response turnover
Mechanistic interpretationConcentration-effect relationshipDrug-modified production and turnover

The choice between these approaches should be driven by the scientific question and the observed temporal behavior rather than by a preference for one model class.

20 · Extensions

20. Extensions of the Basic Model

The simple Type I model can be extended in several ways when the biology or data require additional structure.

  • Time-varying PK input: use a full concentration-time profile rather than a constant concentration.
  • Nonlinear stimulation: use Hill-type or other concentration-effect functions.
  • Multiple response compartments: represent sequential biological processes.
  • Covariate models: allow response turnover or stimulation parameters to vary with patient characteristics.
  • Population PK/PD: estimate typical parameters and between-subject variability simultaneously.
  • Effect compartments: introduce a separate effect-site concentration when distributional delay is the more appropriate explanation for the observed hysteresis.

The model should become more complex only when the additional structure is supported by the data and needed for the scientific purpose.

21 · Practical workflow

21. A Practical Workflow for Indirect Response Modeling

  1. Define the response. Determine exactly what biological or clinical quantity is being modeled.
  2. Characterize baseline turnover. Estimate or establish R0 and the response turnover process.
  3. Build the PK component. Obtain an adequate description of C(t).
  4. Inspect temporal relationships. Determine whether the response is delayed relative to concentration.
  5. Specify the stimulation function. Start with an interpretable form such as an Emax relationship when appropriate.
  6. Fit the integrated PK/PD model. Estimate parameters using an appropriate nonlinear modeling framework.
  7. Check diagnostics. Evaluate observed-versus-predicted plots, residuals, parameter precision, and plausibility.
  8. Test alternative mechanisms where justified. Compare indirect response, effect-compartment, or other mechanistic models when the data can support the distinction.
  9. Use the model for simulation. Explore concentration, response, dosing, and recovery scenarios within the model's applicable range.

22. Key Takeaways

  • An indirect response stimulation model describes a drug that increases the production or input rate of a response.
  • The basic turnover equation is dR/dt = Rinf(C) − koutR.
  • At baseline, production and loss balance: Rin = koutR0.
  • A common stimulation function is an Emax-type relationship involving Smax and SC50.
  • Smax controls the maximum fractional stimulation of production, while SC50 controls the concentration scale.
  • kout controls response turnover and determines the response half-life through t1/2,R = 0.693/kout.
  • The response can lag behind concentration because production and turnover take time.
  • This dynamic behavior can produce concentration-response hysteresis without requiring an instantaneous concentration-effect relationship.
  • Indirect response models can be connected directly to a PK model so that concentration drives response production.
  • Delayed response does not by itself prove an indirect mechanism; effect compartments and other mechanisms can produce similar observations.
  • Good sampling of both concentration and response over time is important for identifying the model parameters.
  • The appropriate model is the one that adequately represents the observed dynamics and answers the scientific question without adding unsupported complexity.
Next step

Where to Go Next

A natural next step is to study the other indirect response model types: inhibition of response loss, inhibition of response production, and stimulation of response loss. These models use the same turnover framework but place the drug effect on different components of the response system.

From there, the framework can be extended to population PK/PD modeling, effect-compartment models, time-dependent pharmacodynamics, and mechanistic biomarker models.

← Back to Pharmacokinetics Tutorials