EMA finalises revised PBPK guideline: expanded requirements for DDI predictions and special population extrapolation

The European Medicines Agency (EMA) published the final revision of its Physiologically Based Pharmacokinetic Modelling guideline (EMA/CPMP/EWP/7716/2016 Rev.2) on 10 November 2026, effective for all marketing authorisation applications (MAA) submitted from 1 April 2027. The guideline covers: (1) Model qualification requirements — the minimum evidence package needed to demonstrate that a PBPK platform (Simcyp, GastroPlus, PK-Sim) is qualified for the specific intended use; (2) Drug-drug interaction predictions — updated standards for CYP inhibition/induction PBPK DDI predictions used to waive or inform clinical DDI studies; (3) Hepatic and renal impairment extrapolation — criteria for using PBPK to replace dedicated impairment studies; (4) Paediatric dose extrapolation — requirements for PBPK-based paediatric dose finding under Paediatric Investigation Plans (PIPs). The revision aligns with FDA's 2018 PBPK guidance and the finalised ICH M15 Population PK guideline.

Physiologically based pharmacokinetic (PBPK) models are mechanistic mathematical frameworks that describe drug disposition using real physiological parameters — tissue volumes, blood flows, enzyme abundances, transporter expression levels — rather than empirical curve-fitting. They allow a priori prediction of how a drug will behave in special populations (elderly, paediatric, hepatic/renal impairment, pregnancy) or in the presence of perpetrators of drug-drug interactions, before clinical studies are conducted. Regulatory acceptance of PBPK models can waive the requirement for dedicated clinical pharmacology studies — a significant time and cost saving. The new EMA guideline tightens the qualification standard: sensitivity analyses must be reported, verification datasets must include diverse populations, and model limitations must be explicitly stated alongside predictions. See our PBPK model glossary entry and pharmacokinetics basics guide.

The most practically impactful section of the revised guideline covers CYP inhibition and induction DDI predictions. Under the previous guideline, PBPK-predicted DDIs required a confirmatory clinical DDI study in most cases. Under Rev.2, PBPK predictions meeting the new qualification criteria can fully waive clinical studies for moderate inhibitors/inducers in specific CYP substrates — provided the model has been qualified against ≥3 clinical DDI datasets for that perpetrator-substrate pair. This is expected to reduce Phase I DDI programme burden significantly, especially for narrow-therapeutic-index drug interactions where clinical DDI study design is ethically constrained. The change also impacts labelling of drug interactions — waivers require explicit statement of the PBPK prediction and its uncertainty in the SmPC/product information.

The revised guideline formalises PBPK-based paediatric dose extrapolation as an acceptable alternative to clinical PK studies in children when: (1) the adult PBPK model is qualified with ≥2 validation datasets; (2) ontogeny functions for the relevant CYP enzymes are incorporated (Simcyp paediatric or GastroPlus paediatric maturation parameters); (3) sensitivity analyses on uncertain paediatric parameters (GFR maturation, renal transporter expression) are presented. Paediatric PBPK predictions must be confirmed by sparse PK sampling in the first paediatric cohort, with a pre-specified criterion for dose adjustment. This reduces the number of full PK studies required in children, addressing both ethical concerns (venepuncture burden) and practical recruitment challenges. See our paediatric pharmacology guide for developmental PK context and allometric scaling for the complementary size-scaling approach.