Nonlinear mixed effects modelling: structural models MCQs With Answer

Introduction: Nonlinear mixed effects modelling is a cornerstone of population pharmacokinetic/pharmacodynamic analysis. This blog focuses on structural models used in NLME: compartmental architectures, absorption and elimination kinetics, parameterizations (CL, V, Q, KA), and practical features such as lag-time, transit compartments and nonlinear clearance. For M.Pharm students preparing for clinical pharmacokinetics and therapeutic drug monitoring, understanding how structural choices influence parameter identifiability, model selection and interpretation of inter-individual variability is essential. The following MCQs deepen conceptual knowledge and model-building intuition, linking mathematical forms to biological processes and common software implementations used in population PK modelling.

Q1. Which structural model feature specifically accounts for a delay between drug administration and the start of observable absorption?

• First-order absorption rate constant (KA)
• Absorption lag time (Tlag)
• Peripheral compartment volume (V2)
• Inter-compartmental clearance (Q)

Correct Answer: Absorption lag time (Tlag)

Q2. In a two-compartment linear PK model after an IV bolus, which parameters define the distribution and elimination phases?

• KA and Tlag
• CL and V1 only
• Inter-compartmental clearance (Q), central V1, peripheral V2, and CL
• Bioavailability (F) and absorption rate

Correct Answer: Inter-compartmental clearance (Q), central V1, peripheral V2, and CL

Q3. Which absorption model can best describe a long, well-sustained input process such as an extended-release oral formulation?

• Zero-order absorption
• First-order absorption
• Michaelis-Menten elimination
• Two-compartment distribution without absorption

Correct Answer: Zero-order absorption

Q4. When would one prefer a transit compartment absorption model over a simple lag-time model?

• To model instantaneous absorption with no variability
• When absorption delay is variable and the absorption profile is dispersed over time
• When elimination is saturable
• When bioavailability is 100%

Correct Answer: When absorption delay is variable and the absorption profile is dispersed over time

Q5. Which structural model characteristic indicates nonlinear elimination consistent with capacity-limited metabolism?

• First-order clearance (CL independent of concentration)
• Michaelis-Menten kinetics with Vmax and Km
• Instantaneous distribution to a peripheral compartment
• Combined additive and proportional residual error

Correct Answer: Michaelis-Menten kinetics with Vmax and Km

Q6. In population PK, parameterizing a two-compartment oral model as CL, V1, Q, V2 and KA is preferred because:

• It reduces number of parameters to estimate to two
• It directly maps to physiological processes and aids interpretation of clearance and distribution
• It eliminates the need to model residual variability
• It forces bioavailability to be 1

Correct Answer: It directly maps to physiological processes and aids interpretation of clearance and distribution

Q7. Which residual unexplained variability model is most appropriate when assay error is proportional to concentration magnitude?

• Additive error model
• Proportional error model
• Logistic error model
• Inter-individual variability model

Correct Answer: Proportional error model

Q8. Adding random effects (η) on CL and V in an NLME structural model primarily captures what?

• Within-subject residual assay noise
• Inter-individual variability in pharmacokinetic parameters
• Time-varying changes in dose
• Covariate relationships like weight or renal function

Correct Answer: Inter-individual variability in pharmacokinetic parameters

Q9. Which structural assumption would be violated if estimated CL decreases as dose increases and may indicate saturable elimination?

• Linear (first-order) elimination
• One-compartment distribution
• Instantaneous absorption
• Fixed bioavailability

Correct Answer: Linear (first-order) elimination

Q10. Which parameterization change can improve identifiability when data are sparse and both central volume and peripheral volume are poorly informed?

• Fixing V2 or Q to literature values
• Adding a third peripheral compartment
• Replacing CL with bioavailability F
• Switching from first-order to zero-order absorption

Correct Answer: Fixing V2 or Q to literature values

Q11. In an oral two-compartment model, apparent clearance (CL/F) is estimated because:

• F (bioavailability) is always 1 for oral dosing
• Oral data cannot separately identify absolute clearance and bioavailability without IV data
• F cancels out when using transit compartments
• CL and F are orthogonal parameters and estimated independently

Correct Answer: Oral data cannot separately identify absolute clearance and bioavailability without IV data

Q12. When constructing a structural model, which diagnostic pattern suggests the need for an additional absorption lag or transit process?

• Systematic positive residuals in early post-dose observations
• Random scatter of residuals around zero at all times
• Underprediction at late concentrations only
• Perfect fit of observed versus predicted concentrations

Correct Answer: Systematic positive residuals in early post-dose observations

Q13. Allometric scaling of clearance with body weight commonly uses which exponent for adult populations?

• 0.75
• 1.0
• 0.33
• 2.0

Correct Answer: 0.75

Q14. In a model with Michaelis-Menten elimination, which parameter represents the concentration at which elimination rate is half-maximal?

• Vmax
• Km
• CL
• KA

Correct Answer: Km

Q15. Which structural change most directly addresses persistent heteroscedastic residuals increasing with concentration?

• Switching from proportional to additive residual error
• Using a combined (additive + proportional) residual error model
• Removing random effects on CL
• Adding a second peripheral compartment

Correct Answer: Using a combined (additive + proportional) residual error model

Q16. In an NLME structural model, correlation between ηCL and ηV suggests what?

• Artifacts in assay calibration only
• Subjects with higher clearance also tend to have larger volumes (or vice versa), indicating physiological linkage or identifiability coupling
• No effect on model predictions
• That residual error model is misspecified

Correct Answer: Subjects with higher clearance also tend to have larger volumes (or vice versa), indicating physiological linkage or identifiability coupling

Q17. When is a one-compartment model potentially insufficient and a multi-compartment structural model warranted?

• When concentration-time data show a single log-linear terminal phase only
• When early post-IV concentrations decline biexponentially indicating rapid distribution phase plus slower elimination
• When there is no absorption phase
• When doses are administered orally

Correct Answer: When early post-IV concentrations decline biexponentially indicating rapid distribution phase plus slower elimination

Q18. Which model component represents variability that cannot be explained by the structural model or covariates and differs between observations of the same individual?

• Inter-individual variability (IIV)
• Residual unexplained variability (RUV) or intra-individual variability
• Between-study variability
• Covariate effect

Correct Answer: Residual unexplained variability (RUV) or intra-individual variability

Q19. In population PK analysis, why might one parameterize elimination as CL rather than ke (elimination rate constant)?

• ke is dimensionally inconsistent
• CL is directly scalable with body size and relates to physiological elimination processes, facilitating covariate modeling
• ke cannot be estimated from IV data
• ke assumes nonlinear elimination

Correct Answer: CL is directly scalable with body size and relates to physiological elimination processes, facilitating covariate modeling

Q20. Which structural modelling approach helps distinguish between poor absorption versus high first-pass extraction when only oral data are available?

• Simultaneously fitting a parent-metabolite model with observed metabolite data or having IV reference data
• Using a one-compartment model with additive error
• Assuming bioavailability equals 1
• Fixing CL to zero

Correct Answer: Simultaneously fitting a parent-metabolite model with observed metabolite data or having IV reference data

Download