MEB Science Day 2025 - PhD's and abstracts

Comparing animal developmental toxicity to human pregnancy outcomes for fifty pharmaceuticals: is the rat a better predictor than the rabbit?

Authors
Puck Roos¹, Sander Timmermans¹, Peter van Meer^1,2,3, Peter Theunissen^1,2

Affiliations
1.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
2.    Department of Pharmacology and Toxicology, Radboud University Medical Center, Nijmegen, The Netherlands
3.    Department of Pharmaceutics, Utrecht Institute of Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands 

Background: Embryofetal Development (EFD) studies are conducted to detect potential adverse effects of pharmaceuticals during pregnancy. For small molecules, EFD studies are generally performed in both rats and rabbits. However, multiple studies have found that EFD studies in a second species have limited added value compared to one species. The current study evaluated to what extent rat and rabbit EFD studies predict known adverse pregnancy outcomes in humans. 

Methods: A previous database of rat and rabbit EFD studies for pharmaceuticals was used for compound inclusion. Pharmaceuticals with human pregnancy outcomes reported in the Summary of Product Characteristics were included in the analysis. Comparison between animal EFD toxicity and human pregnancy outcomes was based on malformations and embryofetal lethality (MEFL). 

Results: Fifty pharmaceuticals were included. Thirty pharmaceuticals induced MEFL in humans. Five out of 30 pharmaceuticals induced MEFL only in humans, and not in animals. For most of these pharmaceuticals, effects in humans were equivocal and based on limited data. Twenty out of 30 pharmaceuticals induced MEFL in humans, rats and rabbits. Five out of thirty pharmaceuticals induced MEFL in humans and rats, but not in rabbits. None of the pharmaceuticals induced MEFL in humans and rabbits, but not in rats. 

Conclusions: We concluded that, based on a limited number of pharmaceuticals, rats seem to better predict adverse human pregnancy outcomes than rabbits. These data could contribute to reduction of animal studies, by prioritizing rat over rabbit in a tiered testing approach, or for qualification of in vitro assays for developmental toxicity testing. 

From guideline to practice: three  years of ICH S11 insights and  recommendations

Authors
Diana Tavares^1*, Hsiao-Tzu Chien^2,3*, Maria Elzbieta Sheean¹, Peter T. Theunissen^2,3, Peter van Meer^2,3,4**, Karen Van Malderen^5**

Affiliations
1.    Paediatric Medicines Office, European Medicines Agency, Amsterdam, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
3.    Department of Pharmacology and Toxicology, Radboud University Medical Center, Nijmegen, The Netherlands
4.    Department of Pharmaceutics, Utrecht Institute of Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands
5.    Federal Agency for Medicines and Health Products, Brussels, Belgium

*These authors have contributed equally to this work and share first authorship
**These authors have contributed equally to this work and share last authorship

Background: Juvenile Animal Studies (JAS) are conducted to ensure the children’s safety of paeditraic medicines. The ICH S11 guideline aims to guide the need for and design of JAS, and outlines a weight-of-evidence (WoE) approach. This study evaluates the impact of the ICH S11 guideline on the non-clinical strategy for paediatric medicines in the European Union.

Methods: We reviewed Paediatric Investigation Plans (PIPs) for 127 products approved between 2020 and 2023, analyzing associated regulatory assessments and final non-clinical strategies.

Results: Among the selected PIPs, 12 involved ongoing or completed JAS at the time of submission, while the remaining 115 were submitted before JAS initiation. In 75% of cases (86/115), WoE discussions led to an agreement on the proposed non-clinical strategy. However, in 25% of cases, disagreements emerged based on the WoE analysis, resulting in the addition (3%), modification (10%), or removal (11%) of JAS.

Conclusion: Our findings suggest that the ICH S11 guideline has facilitated science-driven discussions about the necessity and design of JAS, within the broader non-clinical strategy. By considering the developmental aspects of the pharmacological target, clinical relevance of toxicity findings, and clinical context, the guideline fosters effective dialogue and improves regulatory alignment. The WoE approach ensures the generation of relevant safety data to support paediatric drug development, while adhering to the 3Rs principles of non-clinical replacement, reduction, and refinement.

Oligonucleotide-based therapeutics: Hard to handle? - Evaluation of the non-clinical testing strategies addressing safety assessment from a regulatory perspective

Authors 
Clara Stock¹, Britt Duijndam¹, Kris Siezen¹, Anna M.G. Pasmooij^1,2

Affiliations 
1.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
2.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands

Background: The non-clinical safety assessment of oligonucleotide-based therapeutics (ONTs) lacks coherent regulatory guidance on a global level. Even though ONTs reflect characteristics similar to both small molecules and biopharmaceuticals, they also exhibit unique characteristics, such as specific off-target toxicities and different pharmacokinetic-pharmacodynamic relationships. To harmonize regulatory expectations, the ICH committed to drafting a new guideline addressing the non-clinical safety assessment of ONTs (ICH S13). We aim to identify critical aspects of the non-clinical testing strategy of ONTs, which require attention in the upcoming ICH S13 guideline.

Methods: The non-clinical safety study packages of all ONTs which applied for EU marketing authorization to date, supplemented by ONTs receiving CHMP Scientific Advice in 2020-2024, were collected. Scientific Advice letters were analyzed regarding questions and positions of the Applicants and subsequent opinion and suggestions from the CHMP.

Results: We identified 64 different ONTs, including 41 ASOs and 23 siRNAs. The non-clinical study packages were analyzed for their strategies to assess ONTs safety profiles, with focus on  secondary pharmacology, safety pharmacology, tissue distribution, genotoxicity, carcinogenicity, and reproductive and developmental toxicity. We will answer ONT specific safety questions, such as: How do Applicants address the relatively long tissue half-lives of ONTs when designing toxicology studies? How are species selected and surrogates used for safety assessment? Are there concerns for genotoxicity and how are they assessed? 

Conclusion: Ultimately, critical aspects of the non-clinical testing will be shared with the drafting group for ICH S13 and thereby support the development of a tailored guideline for ONTs.
 

European perspectives on ethical aspects around utilization of patient registries for medicines decision-making

Authors
Fabian Windfuhr¹, Sieta T. de Vries^1,2, Maria Melinder³, Diogo Almeida⁴, Bruno Sepodes⁴, Carla Torre⁴, Björn Wettermark³, Peter G.M. Mol^1,2

Affiliations
1.    Department of Clinical Pharmacy and Pharmacology, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
3.    Department of Pharmacy, Faculty of Pharmacy, Uppsala University, Uppsala, Sweden
4.    Faculdade de Farmácia, Universidade de Lisboa, Portugal

Objectives: To describe ethical considerations for using patient registry data to inform medicines decision-making.

Methods: European regulators, Health Technology Assessors (HTAs), payers, patients, academia, industry, and healthcare professionals (HCPs) were invited through email and social media to complete an online survey (Dec 2024 – Jan 2025), including participants’ demographics and perspectives regarding ethics-related aspects of patient registry usage (based on 5-point Likert scales). Data were analyzed descriptively.

Results: Included were 382 participants (159 regulators/HTA/payers, 41 industry, 42 patient representatives, 140 academics/HCPs/others). Participants represented 31 countries, had a mean age of 48 years (SD±11) and 56% were women. Most participants (65%) indicated that registries should collect identifiable data and process pseudonymized data, 51% preferred an ‘opt-out’ consent for this. For secondary use of registry data, 34% of respondents chose transfer of anonymized data as the preferred method, and ‘opt-out’ consent was preferred for this by 61%. Respondents perceived information on a registry’s use-and-access policy most important (mean 4.2; 95% CI: 4.1-4.3) and information on funding sources least important (mean 3.9; 95% CI: 3.8-4.0). Most participants (83%) preferred to find such information on a registry’s website, followed by the European Medicines Agency – Heads of Medicines Agencies Catalogues of real-world data sources (66%). Participants on average considered 51% (scale:0-100%; SD±32) funding of registries by industry acceptable.

Conclusions: Participants opted for higher granularity for data collection and processing by a registry, compared to data granularity for secondary use. Generally, ‘opt-out’ was preferred as a method for obtaining consent.

Authors
Geeske F. Grit^1,2, Maaike van Dartel³, Cornelis Boersma^5,6, Doranne Hilarius⁶, Peter G.M. Mol^1,2

Affiliations
1.    Department of Clinical Pharmacy and Pharmacology, University Medical Center Groningen, Groningen, The Netherlands
2.    Dutch Institute for Clinical Auditing, Leiden, The Netherlands
3.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
4.    Unit of Global Health, Department of Health Science, University Medical Centre Groningen, Groningen, The Netherlands
5.    Faculty of Management Sciences, Open University, Heerlen, The Netherlands
6.    Department of Pharmacy, Rode Kruis Ziekenhuis, Beverwijk, The Netherlands

Background: Registry based real-world data (RWD) can enhance the evaluation and appropriate use of new drugs for non-small cell lung cancer (NSCLC), as small treatment populations and single arm trials currently challenge the evaluation of drug effects for regulatory authorities, health technology assessment (HTA), payers, and clinicians. Therefore, we aimed to develop a minimal dataset enabling all relevant stakeholders to evaluate NSCLC drugs using registry-based RWD.

Methods: The minimal dataset was developed through a literature review of existing datasets, followed by two expert meetings to review the variables, and identify any missing important variables. First, experts - including pulmonologists, hospital pharmacists, and a patient representative – provided input on the minimal dataset (n=11). Subsequently, the Regulatory Science Network Netherlands organised an expert meeting including regulators, HTA, industry, academia, and a patient representative (n=15). 

Results: In literature, four existing suitable datasets were identified and used to define the initial minimal dataset. Expert input was then incorporated to further refine it. The final minimal dataset includes general patient characteristics (n=9), disease-specific factors (n=20), treatment details, including systemic therapy, surgery, and radiotherapy (n=28), outcomes such as death, progression, response (n=7), and patient reported outcome measures (n=18).

Conclusion: This minimal dataset contains essential variables required to address most stakeholder questions about drugs used in NSCLC. However, experts noted that not all variables can be defined for every specific regulatory and HTA decisional situation, as they depend on a specific question. Our dataset provides guidance for registry owners in building a relevant core NSCLC patient registry.

A Target Trial Emulation with an Extension to Subgroups: an Example for Relapsing-Remitting Multiple Sclerosis

Authors 
Stefan Verweij^1,2, Maarten J. Bijlsma³, Katrien Oude Rengerink², Jan Hillert⁴, Lars E. Forsberg⁴, Elena Flavia Mouresan⁴, Anna Glaser⁴  Robert J. Fox⁵, Eelco Hak¹, Peter G.M. Mol^2,6

Affiliations
1.    Unit of PharmacoTherapy, Epidemiology and Economics, Groningen Research Institute of Pharmacy, Groningen, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
3.    Max Planck Institute for Demographic Research, Rostock, Germany
4.    Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden
5.    Mellen Center for Multiple Sclerosis Treatment and Research, Neurological Institute, Cleveland Clinic, Cleveland, OH, USA
6.    Department of Clinical Pharmacy and Pharmacology, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands

Background: The target trial emulation (TTE) framework may help reduce bias in observational analyses and improve transparency of design and analytic decisions. However, conclusions are also limited to the restricted trial populations. Therefore, knowledge gaps for neglected subpopulations often remain. In this study, we emulated the CONFIRM trial using strict eligibility criteria, and compared the estimates to a pragmatic scenario which included these ineligible subgroups.

Methods: CONFIRM compared dimethyl fumarate (DMF) with glatiramer acetate (GA) in reducing relapses in 709 patients with relapsing-remitting multiple sclerosis (RRMS). Here, we designed a retrospective cohort study and followed two scenarios, where we 1) adhered to original trial criteria and 2) followed a pragmatic approach which included otherwise ineligible subgroups. Outcomes were the annualized relapse rate (ARR) and the probability of experiencing a relapse, adjusted for covariates in combination with overlap weights.

Results: Using strict eligibility criteria, 288 RRMS patients were included (111 DMF, 177 GA). In the pragmatic scenario, 1831 patients were included (1072 DMF, 759 GA). The ARR ratio for DMF versus GA was 0.76 (95%CI 0.56-1.02) in CONFIRM, 0.75 (95%CI: 0.44-1.29) in the strict scenario, and 0.58 (95% CI: 0.36-0.94) in the pragmatic scenario. Similarly, the HR was 0.93 (95%CI 0.64-1.34), 1.05 (95%CI: 0.83-1.33) and 0.56 (95%CI: 0.46-0.67) respectively. 

Conclusion: Comparable results between CONFIRM and the strict scenario were observed. In the second step, methodological advantages of TTEs were extended to excluded subgroups, thereby embracing the added value of observational data, that is, the wider variety in the patient population often ignored in experimental settings.
 

Evaluating model-based extrapolation for generic long-acting injectable (LAI) products: from single- to multiple-dose studies

Authors
D. Esther Lubberts^1,2, Jeroen V. Koomen^1,2, Douglas J. Eleveld¹,  Pieter J. Colin¹  

Affiliations
1.    Department of Anesthesiology, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands

Background: Long-acting injectable (LAI) products are formulations with prolonged drug release that offer several benefits including improved compliance and patient outcomes. Current EMA Guidelines require both single-dose (SD) and multiple-dose (MD) bioequivalence (BE) studies as evidence for marketing approval, which suggests a difference in absorption after SD versus MD administration. The development of generic LAIs is hampered by these lengthy and complex BE studies. This study investigates whether absorption is different after SD versus MD administration of currently approved LAIs. 

Methods: Population pharmacokinetic models evaluating a range of structural absorption models were developed for a sample of 5 LAI products based on SD data using non-linear mixed effect modeling. The best population pharmacokinetic model, as assessed by numerical and graphical evaluation of model fit, was used to predict pharmacokinetic profiles after MD administration. 

Results: The absorption of LAIs was best described with (parallel) first-order absorption structures (with and without lag-time). For the first LAI, 95% prediction intervals of Cmin, Cmax, AUC were 79.8%-105.8%, 80.6%-98.2%, and 93.0%-115%, respectively. Simulations after MD administration for the remaining products are still ongoing.     

Conclusion: Simplified first-order absorption models properly describe the SD pharmacokinetics for the included LAIs. The first results suggest that the SD models extrapolate well to MD data, which would indicate that absorption is not different between these scenarios for currently approved LAIs. 

Optimising methodologies for sample size determination in paediatric pharmacokinetic trials to support extrapolation based on exposure-matching.  

Authors
Loes C.H. Maton^1,2, Jeroen V. Koomen^1,2,  Roberto de Lisa³, Pieter J. Colin^2,3

Affiliations
1.    Medicines Evaluation Board (CBG-MEB), Utrecht, The Netherlands
2.    Department of Anaesthesiology, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
3.    European Medicines Agency (EMA), Amsterdam, The Netherlands

Background: Generation of evidence in the paediatric population is often hampered by difficulties in conducting clinical trials, e.g. due to limited sample size. Under defined circumstances, when there is sufficient confidence that disease progression and treatment response are similar, extrapolation of efficacy based on demonstrating similarity in exposure between the target and reference populations could provide an alternative or complement an efficacy trial with limited sample size. However, there is currently no established methodological guidance for concluding on whether similarity in exposure has been sufficiently demonstrated. This project aims to benchmark current strategies of demonstrating similarity in exposure and their impact on sample size estimations in paediatric clinical trials. 

Methods: Paediatric investigation plans (PIPs) with model-based evaluation of similarity in exposure were eligible for inclusion in this analysis if the Paediatric Committee (PDCO) of the European Medicines Agency referred them to the agency’s Modelling and Simulation Operational Expert Group (MSOEG) for review, particularly on the sample size selected for paediatric pharmacokinetic (PK) trials. Methodologies used to determine sample sizes according to the proposed analysis were extracted from these PIPs. These methodologies will be evaluated using an established population PK model. 

Expected results: Seventeen cases reviewed by MSOEG between September 2022 till September 2024 were selected. Three of these cases included a stand-alone analysis, whereas the others included a pooled analysis. Sample sizes were predominantly based on the acceptance criteria described by Wang et al¹. Only one case justified their sample size calculation using the Fisher Information Matrix. 

1.    Wang Y, Jadhav PR, Lala M, Gobburu JV. Clarification on precision criteria to derive sample size when designing pediatric pharmacokinetic studies. J Clin Pharmacol. 2012 Oct;52(10):1601-6. doi: 10.1177/0091270011422812. Epub 2011 Dec 12. PMID: 22162537.

PARP Inhibitors: Pool-Adjacent-Violators Algorithm (PAVA) Analysis of Anti-Tumor Activity and Toxicity

Authors
Njeri Kamau^1,2, Mark T.J. van Bussel^1,2, Steven Teerenstra^1,2, Kit C.B. Roes^1,2

Affiliation
1.    Department of Health Evidence, Radboud University Medical Center, Section Biostatistics, Nijmegen, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands

Background: The selected doses of many anticancer agents have been shown to be poorly tolerable as reflected by high discontinuation and dose reduction rates due to adverse events. To improve on dose selection for targeted agents, both anti-tumor activity as well as toxicity should be evaluated simultaneously during the dose escalation phase. The use of Pool-Adjacent-Violators Algorithm (PAVA) based isotonic regression to characterize dose-toxicity and dose-efficacy profiles of these agents could be used to improve dose selection. 

Methods: Dose-toxicity and dose-efficacy data were collected from phase 1 dose-escalation studies of  European Medicines Agency (EMA) approved poly(ADP-ribose) polymerase (PARP) inhibitors. Toxicity outcomes were based on predefined dose limiting toxicities and efficacy outcomes based on objective response rates. This data was analyzed using PAVA based isotonic regression to characterize the dose-response profiles. 

Results: The probability of the specified toxicity or efficacy outcomes at each dose level was determined using the PAVA model. Dose-response curves were plotted, providing an aid for the visualization of the efficacy-toxicity tradeoff. In all resulting dose-efficacy curves, after an initial increase, the probability of efficacy did not seem to increase much more beyond a certain dose (plateau effect). This effect was found to occur at a dose level lower than the final selected dose for all the agents assessed.

Conclusion: We demonstrate how using PAVA analysis to determine dose-response relationships can be a useful tool to assist in the dose selection of targeted agents in early phase oncology trials.

Evaluating sequencing and variant annotation platforms identifies discrepancies in risk interpretation 

Authors
Rafaella Buzatu^1,2, Ksenia Arkhipova¹, Christian Schröter¹, Tineke van den Hoorn², Marcel H.N. Hoefnagel², Micha Drukker¹

Affiliations
1.    Division of Cell Systems and Drug Safety, Leiden Academic Centre for Drug Research (LACDR), Leiden University, 2333 CC Leiden, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands

Background: Assessing the safety of cellular products requires identifying their genomic variations and quantifying their tumorigenic risk. However, there is no consensus on optimal sequencing techniques for detecting biologically significant variants or on integrating various databases to interpret pathogenicity.

Methods: We compared the effectiveness of RNA-sequencing and Whole Exome Sequencing (WES) for variant detection.

Results: Both platforms exhibited similar coding region coverage, with only 4.8% of WES exonic variants missed by transcriptomics due to lack of expression. RNA-sequencing additionally captured post-transcriptional events, accounting for 16.8% of its variants, while another 17.3% of its variants were identified in coding regions not covered by WES. Comparing RNA-sequencing variant-calling pipelines (e.g., GATK) showed limited overlap, while pathogenicity variant annotations from ClinVar, COSMIC, and AlphaMissense were frequently contradictory.

Conclusions: These findings emphasize the value of transcriptomics for variant detection and call for new approaches to integrate information on their functional consequences to develop more reliable risk assessment approaches.

Investigating the Definition, Utilization and Acceptability of Platform Technology: A Scoping Review

Authors
Stephanie M.P. Oskam^1,2, Lourens T. Bloem¹, Anna M.G. Pasmooij^1,2

Affiliations
1.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands

Background: Platform technology remains ambiguously defined. Multiple stakeholders, including the European Commission and the United States Food and Drug Administration (FDA), have defined the term differently. The European Commission has incorporated platform technology in the new proposal for the European Union general pharmaceutical legislation as a new type of application for a marketing authorisation. The FDA has a platform technology designation program, defining platform technology along the lines of medicine manufacturing. The various definitions of platform technology are creating confusion regarding its exact meaning. The objective of this scoping review is therefore to identify how platform technology is defined and utilized, and assess its regulatory and clinical acceptability in current literature. 

Methods: Publications in English discussing the terms ‘platform technology’ or ‘platform technologies’ in medicine development, regulation, or use from PubMed will be included. A PubMed search identified 2,572 publications. After removing two duplicates, 2,570 publications remained for full-text screening. The first reviewer will assess all publications for inclusion and extract relevant data from the included publications. A second reviewer will independently assess and extract 10%. The data extracted will be analysed using descriptive statistics.

Results: We aim to extract definitions of platform technology, its context and utilization (including substance class of medicines, disease area, and where in medicine development process), as well as benefits, challenges, regulatory considerations and opinions on clinical acceptability. Findings will be presented in tables and charts along with a narrative.

Conclusion: Further investigation into the definition, utilization and acceptability of platform technology may lead to advancements in medicine development, regulation, and use.
 

Exploring stakeholders’ experiences with the development and implementation of biomarker thresholds

Authors
Renske J. Grupstra^1,2, Audrey M.M. Hermans^1,2, Elisabeth Bakker^1,2, Viktoriia Starokozhko^1,2,3, Anna M.G. Pasmooij^1,4, Peter G.M. Mol^1,2,5

Affiliations
1.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
2.    Department of Clinical Pharmacy and Pharmacology, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
3.    Scientific Advice Working Party, European Medicines Agency, Amsterdam, The Netherlands
4.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute of Pharmaceutical Sciences, Faculty of Science, Utrecht University, Utrecht, The Netherlands
5.    Committee for Medicinal Products for Human Use, European Medicines Agency, Amsterdam, The Netherlands

Background: To increase the chance of individual patients’ treatment success, biomarkers can be used to select the right treatment for the right patient. When using biomarkers it is often necessary to determine a specific cutoff value, also known as a biomarker threshold. For example, for treatment purposes it may be useful to divide chronic kidney disease (CKD) patients into two groups based on kidney volume above or below a specific threshold. This study aims to gain insights into stakeholders’ experiences with the development and implementation of biomarker thresholds in different therapeutic areas. 

Methods: Semi-structured interviews were conducted with stakeholders with expertise in CKD, oncology and Alzheimer’s disease. Representatives from the following stakeholder groups were interviewed: pharmaceutical companies, diagnostic companies, regulatory agencies, Health Technology Assessment experts, healthcare professionals, and patient representatives. Interview transcripts will be analyzed inductively following principles of thematic analysis.  

Results: Between August 2024 and January 2025, 49 interviews were conducted with stakeholders with expertise in CKD (n = 13), oncology (n = 22), and Alzheimer’s disease (n = 14). Interview transcript analysis is expected to be completed in Q2 of 2025. 

Conclusion: We anticipate to draw lessons from issues encountered in the development and implementation of biomarker-based treatment trajectories and to collect case study examples of biomarker development where thresholds had to be established. Such insights may provide valuable lessons for future determination of biomarker thresholds and their acceptability. As part of the PRIME-CKD consortium, this work contributes to enhancing biomarker discovery and personalized medicine implementation in CKD.   

The origin of uncertainties and their mitigation strategies for oncology medicines

Authors
Anne C. Taams^1,2, K. Esther Broekman^2,4, Marloes T. Bazelier¹, Antoine C.G. Egberts^1,3, Lourens T. Bloem¹ 

Affiliations
1.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands 
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
3.    Department of Clinical Pharmacy, University Medical Centre Utrecht, Utrecht, The Netherlands 
4.    Department of Medical Oncology, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands

Background: Uncertainties exist throughout the lifecycle of every medicine. The European Medicines Agency (EMA) evaluates and documents uncertainties in the European Public Assessment Reports (EPARs). In a previous study, we showed that for oncology medicines at time of marketing authorisation, uncertainties mostly relate to effect size, efficacy in specific subpopulations, and long-term efficacy and safety.¹ This follow-up study explores the origins of these uncertainties and corresponding regulatory and clinical mitigation strategies.

Methods: We included all oncology medicines consisting of a new active substance and granted initial marketing authorisation by the European Commission between 2011-2022. Data were extracted from the text fragments about uncertainties and  mitigation strategies in the benefit-risk balance section of EPARs. Using inductive reasoning, we developed a classification system with mutually exclusive categories for origins of the uncertainties. Descriptive statistics were used to examine relationships between uncertainties, their origins, and related mitigation strategies.

Results: The origins of 790 uncertainties identified for 121 oncology medicines were classified into seven origin categories: no data, limited data, internal data validity, external data validity, inconsistent data, external knowledge, and other. Preliminary analysis revealed three predominant origin categories: limited data, internal data validity, and no data. The analysis of mitigation strategies is ongoing.

Conclusion: This study will provide new insights into the origins of uncertainties and their mitigation strategies, and how these are communicated via EPARs. Understanding these factors can help to improve regulatory decision-making and transparency.

1.    Uncertainties about the benefit-risk balance of oncology medicines assessed by the European Medicines Agency. Taams, A.C. et al. ESMO Open, Volume 9, Issue 12, 103991

Biomarkers for Neurodegenerative Diseases in Regulatory Decision-making by the European Medicines Agency

Authors
Audrey M.M. Hermans^1,2*, Elisabeth Bakker^2*, Viktoriia Starokozhko^1,2,3, Loes den Otter^1,4, André J.A. Elferink^1,3, Angela Bradshaw⁵, Lorenzo Guizzaro³, Peter G.M. Mol^1,2,3, Anna M.G. Pasmooij^1,6

Affiliations
1.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
2.    Department of Clinical Pharmacy and Pharmacology, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlands
3.    European Medicines Agency, Amsterdam, The Netherlands
4.    Mental Health and Neuroscience Research Institute, Maastricht University, Maastricht, The Netherlands
5.    Alzheimer Europe, Luxembourg, Luxembourg  
6.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University, Utrecht, The Netherlands

*These authors share first authorship.

Background: Biomarkers (BMs) are valuable tools to facilitate early diagnosis of (subtypes of) diseases, improve patient selection and stratification, and detect therapeutic effects or safety concerns. This study explores the extent to which BMs are utilised in the development of treatments for neurodegenerative diseases (NDDs), topics of discussion around BMs in regulatory advice- and decision-making processes, and sharing of BM-related data.

Methods: We screened the European Medicines Agency’s marketing authorisation application (MAA), qualification- (QA/QO) and scientific advice- (SA) procedures regarding NDDs and analysed those mentioning BMs.

Results: In total, 105 procedures were analysed; 57 SAs (Jan 2020–Dec 2022), 19 QAs/QOs (Jan 2008–Dec 2023), and 29 MAAs (Jan 1995–Dec 2023). The majority involved Alzheimer’s disease (AD; n=30), Parkinson’s disease (PD; n=9), and multiple sclerosis (MS; n=33). Imaging BMs were the most commonly discussed type of BM. Most BMs were used as pharmacodynamic/response measures. The acceptance and role of BMs differed between NDDs. In MAAs for AD, diagnostic BMs guiding patient selection were mostly discussed, whereas in MAAs for MS, imaging BMs (particularly lesions) were generally accepted as supportive/secondary endpoints.

Conclusion: Despite the established role of certain BMs, more precise and reliable BMs are necessary to improve diagnostic accuracy and treatment monitoring for NDDs. To implement novel BMs and facilitate development of new treatments, robust evidence bases showcasing biological plausibility and clear clinical benefits are essential. Collaboration and data-sharing among stakeholders is vital in generating this evidence and enhancing the understanding and management of NDDs.

Authors
Loes den Otter^1,2, Linda Reus¹, André J.A. Elferink¹, Willem M. Otte³, Kees P.J. Braun³, H.J. Marian Majoie^4,5

Affiliations
1.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
2.    Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience, Maastricht University, Maastricht, the Netherlands
3.    Department of Neurology and Neurosurgery, UMC Utrecht Brain Center, Utrecht, The Netherlands
4.    Department of Neurology, Academic Centre for Epileptology, Epilepsy Centre Kempenhaeghe 
5.    Maastricht University Medical Center+, Maastricht, the Netherlands

Background: A new anti-seizure medication (ASM) is typically evaluated in a comparative clinical trial in subjects with insufficiently controlled seizures where the ASM and placebo are given on top of standard of care. The magnitude of the placebo response considerably impacts the evaluation of efficacy of the new ASM and possibly subsequent regulatory drug approval. The current study investigated factors that could potentially influence the placebo response, in order to provide recommendations to increase the efficiency of future clinical trial designs in epilepsy. 

Methods: A systematic search was conducted in MEDLINE, EMBASE and CENTRAL to identify randomized, double-blind, placebo-controlled studies for ASMs in epilepsy or seizures published between 2003 and 2023. Studies that evaluated the 50% responder rate, i.e., proportion of subjects with a ≥ 50% reduction in seizure frequency, as outcome measure were included in the evaluation. Meta-analysis and meta-regression analyses were performed to evaluate if features of study design, patient population and epilepsy-specific features impact the 50% responder rate in the placebo group.

Results: A total of 5,177 studies were identified and 88 were included in the final analyses. Meta-analysis of the 50% responder rate in the placebo group indicated a large variation in response between the studies. Subgroup and meta-regression analyses identified several potential predictors of the placebo response, e.g., mean study duration. 

Conclusion: Further research is needed to investigate how to control for predictors of the placebo response to increase clinical trial efficiency. 

Authors
Sharon C.M. Essink^1,2, Inge M. Zomerdijk², Sabine M.J.M Straus³, Helga Gardarsdottir^1,4,5, Marie L. De Bruin¹

Affiliations
1.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands
2.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands
3.    European Medicines Agency, Amsterdam, The Netherlands
4.    Department of Clinical Pharmacy, University Medical Centre Utrecht, Utrecht, The Netherlands.
5.    Faculty of Pharmaceutical Sciences, University of Iceland, Reykjavik, Iceland

Background: The good pharmacovigilance practices Module XVI (Rev 3) outlines considerations for additional risk minimisation measures (aRMMs), however, lacks guidance on discontinuation. This study assessed aRMM discontinuations for centrally authorised medicinal products (CAPs) in the European Union (EU).

Methods: CAPs authorised with aRMMs in the EU between July 2012-December 2021 were included. aRMM discontinuations were identified using ‘EPAR—Product information’. The date of discontinuation, regulatory procedure, and supporting evidence were extracted from regulatory documents. Follow-up started at marketing authorisation (MA) until the first discontinuation of at least one aRMM, MA withdrawal, or 1 September 2024. Kaplan-Meier survival analyses estimated probabilities of discontinuing aRMMs for CAPs within five and ten years after MA. 

Results: Among 132 CAPs authorised with aRMMs, at least one aRMM was discontinued for 15 CAPs, with all aRMMs discontinued for eight. The probability of discontinuing at least one aRMM was 5.4% (95% CI 1.1–9.6; n at risk=84) within five years after MA and 27.0% (95% CI 12.1–39.3; n at risk=19) within ten years. Discontinuations were based on post-authorisation safety study (n=5), periodic safety update reports (n = 3), or other regulatory procedures (n=7). Supporting evidence included integration into clinical practice (n=5), coverage in product information (n=2), removal of risk due to application form change (n=1), or unclear reasons (n=7). 

Conclusions: Around 5% and 25% of CAPs authorised with aRMMs had at least one aRMM discontinued within five and ten years after MA, respectively. This highlights that evaluation is necessary to keep risk minimisation strategies aligned with medicinal product lifecycles. 

Authors
Doerine J. Postma^1,2, Peter A.G.M. de Smet³, Aukje K. Mantel-Teeuwisse¹, Hubert G.M. Leufkens¹, Kim Notenboom⁴ 

Affiliations
1.    Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University, Utrecht, The Netherlands
2.    Royal Dutch Pharmacists Association, The Hague, The Netherlands
3.    Radboud University Medical Centre, Radboud Institute for Health Sciences, Departments of IQ Healthcare and Clinical Pharmacy, Nijmegen, The Netherlands
4.    Dutch Medicines Evaluation Board, Utrecht, The Netherlands 

Background: Supply chain resilience is in the spotlight of global pharmaceutical policies. We assessed the upstream pharmaceutical supply chains to detect vulnerabilities that may increase the risk of medicine shortages.

Methods: The diversity of active pharmaceutical ingredient (API) and finished pharmaceutical product (FPP) manufacturers, their geographic locations and the interdependencies between these manufacturers and marketing authorisation holders (MAHs) was assessed for ten pharmaceutical substances with the largest number of patients in the Dutch outpatient setting in 2022. 

Results: For the 407 authorised medicinal products, 50 of the 90 API manufacturing sites were in Asia. For five pharmaceutical substances, most of the API sites were outside Europe. Of the 128 FPP manufacturing sites, 94 were in Europe. For all ten substances, at least 47% of FPP sites were in Europe.  

API manufacturing for 122 of the 407 products was entirely performed outside Europe, and FPP manufacturing for 66 of the 407 products. For four substances, more than half of the products depended on API manufacturing outside Europe.  

The number of distinct API and FPP manufacturing sites per substance was at least four. For some substances there was an overdependency on one API or FPP manufacturer. 

MAHs applied dual sourcing for API and FPP manufacturing for 61 of the authorised medicinal products. For three substances, none of the products listed at least two API and FPP manufacturing sites.

Conclusion: Our study indicates the need for granular assessment of the interdependencies between MAHs, API and FPP manufacturers to identify upstream supply chain vulnerabilities.