Inter-study variability in population pharmacokinetic meta-analysis: when and how to estimate it?

Area of Science:

• Pharmacokinetics
• Statistical modeling
• Drug development

Background:

• Population pharmacokinetic (PopPK) analysis increasingly pools data from multiple studies.
• Individual pharmacokinetic parameters can vary randomly between studies, introducing inter-study variability (ISV).
• Accurate estimation of variability is crucial for robust PopPK models.

Purpose of the Study:

• To investigate the impact of neglecting inter-study variability (ISV) in population pharmacokinetic analysis.
• To evaluate two methods for estimating random study effects (RSE) and their impact on inter-individual variability (IIV) and ISV.
• To determine the conditions under which ISV estimates are reliable.

Main Methods:

• Simulations were used to assess the bias and variability of parameter estimates under different scenarios.
• Two random study-effect (RSE) estimation methods were compared: a one-stage approach and a two-stage approach.
• A real-world pharmacokinetic dataset was analyzed to illustrate the application of an ISV model.

Main Results:

• Neglecting ISV did not bias fixed parameters or residual variability but could overestimate inter-individual variability (IIV).
• The one-stage RSE model demonstrated good performance for population parameter estimation.
• The two-stage model required rich sampling designs for accurate IIV estimation, and ISV estimates were reliable only when a large number of studies were pooled.

Conclusions:

• The one-stage RSE model is a suitable approach for population pharmacokinetic analysis when pooling data from multiple studies.
• Accurate estimation of inter-study variability (ISV) requires a substantial number of pooled studies.
• Homogeneity in study designs may lead to non-significant study effects, even when ISV is present.