For years, Sara worked in a hospital, making care possible for others. Now, after being diagnosed with breast cancer, she was the one in need of care.
A major challenge in drug development is the translation of candidate molecules from the preclinical phase to the clinical setting.
The key elements to be considered are the pharmacology, toxicology and pharmacokinetic (PK) properties of the candidate molecule, and the quantitative relationships among dose, concentration, target engagement, efficacy, and/or toxicity.
When preparing a first in human (FIH) single-ascending dose (SAD) study design, clear and sound criteria for starting dose, dose escalation, and top dose selection need to be defined in order to safeguard trial subjects and mitigate the risks due to uncertainty (e.g. on PK, mode of action, target nature, relevance of animal models, etc) 1,2. All these aspects are covered in detail in the European Medicines Agency (EMA) FIH guideline 1, dated 2017, where it is also stressed how dose is only a “scaling factor”, and that the expected exposure in human is what really matters 1. The Food and Drug Administration (FDA) guideline 2, dated 2005, focuses only on starting dose selection, with an algorithm based on administered doses and observed toxicities in animals.
Depending on factors such as drug type, therapeutic area, level of translational uncertainty and of knowledge of the intended target, etc, the starting dose can be based on either:
In general, the starting dose should always correspond to an exposure lower than PAD 1 and should provide an exposure at least 10-fold lower than the one at NOAEL 2.
Abbreviations: MABEL = minimal anticipated biological effect level; PAD = pharmacologically active dose; PD = pharmacodynamic; NOAEL = no observed adverse event level; HED = human equivalent dose.
Note: The blue curve represents desired effect, while the red curve represents unwanted side effect.
Dose escalation should be performed with caution based on the observed human safety, tolerability and PK data as they become available during the FIH study 1. While the starting dose will be fixed, dose escalation will need to be flexible in order to accommodate for differences between observations and predictions and for any safety/tolerability signal observed at previous doses. Dose increments will depend on the steepness of the dose-toxicity or dose-effect relationship, and on potential PK nonlinearity 3.
The clinical top dose can be identified based on the estimated human pharmacodynamic (PD) range or based on exposure observed at NOAEL in the relevant animal species 1,2, depending on factors such as the type of participants (healthy volunteers or patients) and the benefit/risk ratio 1.
All available PK and PD data in animals, together with ex vivo / in vitro data need to be integrated to derive a prediction of human PK/PD, determine human doses corresponding to NOAEL, PAD, and/or MABEL, and thus support the planning of FIH trial 1,2. Well-established mathematical modelling methods exist to predict the relationship between dose and concentration in humans based on non-clinical data. Allometry, possibly coupled with population PK modelling, derives human PK by scaling animal PK parameters (such as clearances and volumes) correcting by weight (and possibly other variables that differ across species, such as the lifespan). For specific molecule types, and depending on the available information, in silico models representing the human physiology (Physiologically Based PK models, or PBPK) might be more appropriate for human PK prediction. Such models can be informed with in vivo data and in vitro pharmacology screens characterizing absorption, distribution, metabolism and excretion (ADME). If data permit, mathematical modelling methods to predict the human relationship between concentration and effect (target engagement, efficacy and/or toxicity) can also be applied, considering known relevant inter-species differences.
In summary, a clear and sound scientific rationale for dose selection in FIH trials helps in minimizing the risks and in taking into account sources of uncertainty when translating the preclinical data to the clinical stage of drug development. Mathematical modelling methods represent a valid tool to integrate and leverage all available data and make inference for unobserved settings.
If you would like to discuss your FIH study design needs with one of our experts within Parexel’s Clinical Pharmacology, Modeling and Simulation global team, please contact Joseph Kim (VP, US) at firstname.lastname@example.org or Neil Attkins (VP, EU) at email@example.com .
Generative AI and how we can harness its power in clinical development
Sep 26, 2023
RBQM Podcast Series | Episode 2: Cultivating Risk-based Behaviors
Mar 18, 2022
PBPK modeling solutions as a potential risk mitigation strategy for pH dependent DDIs
Mar 15, 2021
Putting AI to work in your safety program
May 21, 2021
RBQM is our future: Using holistic risk-based oversight, we can focus on a trial’s most critical factors
Jun 18, 2021
Tencent-Backed Wedoctor Is Going Public: What Are the Next Steps for Chinese Digital Health Companies?
Jun 30, 2021
Evidence Generation Strategy under Germany’s Digital Healthcare: Is More Better?
Aug 6, 2021
Highlights from ISMPP Asia Pacific Meeting 2021
Sep 20, 2021
Capitalizing on sensors in clinical trials
Sep 20, 2021
Actigraphy advances a patient-first approach: Reduce patient burden — and make the most of valuable data
Oct 15, 2021
CNS Summit Recap: The Future is Collaborative
Nov 22, 2021
How biotechs can strengthen their value story with advanced analytics
Feb 15, 2022