BPS3082 Chap.6 Direct-Effect PK/PD Models
Direct-Effect PK/PD Models
Direct-effect PK/PD models connect concentration to response when there is no meaningful delay between plasma concentration and the observed effect. Model selection therefore begins with time: peak effect should occur close to peak concentration, and a systematic lag or hysteresis pattern points toward an effect-compartment or indirect-response structure.
A plausible static concentration-effect curve cannot by itself rescue the wrong temporal assumption. The simple Emax model separates two roles. Emax sets the modelled response plateau, while EC50 is the concentration that produces half of that maximum. Lower EC50 shifts the curve left and indicates greater potency within the model; it does not raise maximum efficacy.
The sigmoid Emax model adds the Hill coefficient, which controls steepness around EC50. A larger Hill value makes small concentration changes in that region produce larger response changes while leaving the plateau and half-maximal location conceptually distinct. Calculations at C equal to EC50 and well above it make the half-maximum property and diminishing returns visible.
Parameter interpretation must respect the observed concentration range. If every observation lies on the lower near-linear limb, Emax can be poorly identified. If the plateau is never approached, EC50 and Emax may trade off and extrapolation becomes uncertain.
PK differences also change the concentration input without necessarily changing PD parameters: faster clearance can shorten time in the responsive range even when Emax, EC50 and Hill are constant. This chapter uses curve comparison, units, time alignment and parameter uncertainty to turn a numerical response into a bounded pharmacological statement rather than an unsupported receptor claim.
What this chapter covers
- 01
Direct concentration-effect link
- 02
Emax parameter
- 03
EC50 parameter
- 04
Hill coefficient
- 05
Effect-time simulation
- 06
Peak timing check
- 07
Observed range and parameter identifiability
- 08
Potency efficacy and slope separation
- 09
PK inputs versus PD parameters
- 10
Hysteresis as a model warning
Worked example · free
Calculate effect at EC50
- +1Use E = Emax × C /(EC50 + C).
- +2E = 80 × 2 /(2 + 2) = 40%.
- +1At concentration equal to EC50, response is half of Emax.
Key terms
- Direct Effect
- A response assumed to track concentration without a meaningful temporal delay.
- Emax
- The fitted maximum response approached by the concentration-effect curve.
- EC50
- The concentration producing half of Emax.
- Hill Coefficient
- A parameter controlling the steepness of the sigmoid concentration-effect transition.
- Potency
- The concentration needed to produce a stated response fraction within the model.
Direct-Effect PK/PD Models FAQ
What is the difference between Emax and EC50?
Emax sets the response plateau, while EC50 locates the concentration giving half of that plateau. Lower EC50 shifts the curve left without necessarily changing the maximum. Their units also differ: Emax carries response units and EC50 carries concentration units. Estimate interpretation should be limited when the observed data do not cover both the rising region and plateau.
What does a larger Hill coefficient do?
It steepens the response transition around EC50. Small concentration changes in that region then produce larger response changes, while Emax and EC50 can remain unchanged. This can magnify the effect of exposure variability near the transition, but a fitted Hill value alone does not prove cooperative receptor binding.
How do I know a direct-effect model is appropriate?
Compare concentration and effect over time as well as on a concentration-effect plot. Peak effect should occur near peak plasma concentration without a systematic lag. A delayed peak or hysteresis pattern suggests an effect-compartment or indirect-response model. Confirm that sampling is dense enough to resolve timing before assigning the delay to biology.
Assessment move
Sketch three overlaid curve pairs in which only Emax, EC50 or Hill changes, and label both the changed feature and the feature that stays fixed. Work calculations at concentration equal to EC50, below it and far above it; after each result, explain half-maximum behaviour or diminishing returns in words and check response and concentration units.
Compare effect-time and concentration-time plots before accepting direct effect, marking the times of both peaks and looking for hysteresis. Build a second drill in which clearance changes but all PD parameters stay constant, then explain how the altered concentration input changes effect duration.
For each fitted model, ask whether the observed range reaches the plateau and surrounds EC50; if not, state which parameters or extrapolations are weak. Use precise language: efficacy for the plateau, potency for horizontal position and Hill for steepness.
Avoid interpreting empirical parameter estimates as receptor constants without independent biological evidence, and finish model answers with one temporal diagnostic and one limitation. Make a parameter-identifiability checklist for three experimental ranges: concentrations all below EC50, values spanning EC50, and values reaching the plateau.
Predict which parameter combinations become correlated in each design and which additional observations would be most informative. Next, simulate a modest clearance change and trace its effect through concentration, time above EC50 and predicted response without changing the PD curve. Use this to practise separating exposure explanations from pharmacodynamic explanations.
Finish by stating why a modelled potency difference alone does not establish safety, clinical superiority or receptor mechanism.
