Feedback from The European Bioanalysis Forum Workshop: Taking Tiered Approach to The Next Level

Keywords::

• early GLP
• first into human
• metabolites
• regulatory validated
• scientific validation
• stage of development
• tiered approach
• tissue homogenates
• urine

In this article, we give feedback on the progress in industry on their efforts to provide practical and tangible solutions for a harmonized implementation of the principles of tiered approach. By describing tiered approach as different levels of scientific validation applied as an alternative to apply established regulatory validation principles [1–4] for studies where the guidance was not the intended scope, we hope to provide an acceptable handle for its adoption for an array of study types in industry. The principles of tiered approach became gradually known in regulated bioanalysis about a decade ago [5,6], received positive comments by Health Authority (HA) representatives [7], and were further propagated in regulated bioanalysis with the support of industry consortia [8,9]. Going forward, the bioanalytical community identified more areas in scope of application for tiered approach in a recently published special focus issue of Bioanalysis [10]. The practice of adopting less rigorous (or exhaustive) validation requirements in earlier stages of development, as highlighted in recent (draft) guidance [11] and confirmed during discussions at the recent Crystal City V meeting (Baltimore, MD, USA, 3–5 December 2013) stimulated the European Bioanalysis Forum (EBF) to organize a special workshop on Tiered Approach in June 2014 (Brussels, Belgium) to seek further alignment and provide practical solutions to bring tiered approach to the next level; in other words, in day-to-day practice. The aim of the workshop, which started from a simple paradigm that a validated assay does not necessarily equate to valid data, was fourfold:

• Promote an ‘evolution’ instead of a ‘revolution’;

• Manage expectations of believers and nonbelievers of tiered approach and get alignment with all stakeholders of the validity of applying tiered approach as an alternative for regulatory validation in some areas of drug development;

• Balance ‘change management’ versus the ‘risk management’;

• Acknowledge the contributions from many in industry to date and, where appropriate, use these existing building blocks to propose a more detailed process to apply tiered approach.

Capitalizing on a paragraph in the recent draft of US FDA guidance [11], stating that “For pivotal studies that require regulatory action for approval or labeling, such as bioequivalence (BE) or pharmacokinetic (PK) studies, the bioanalytical methods should be fully validated. For exploratory methods used for the sponsor’s internal decision-making, less validation may be sufficient,” the workshop tried to reflect on how to provide a strategy to allow BA scientists to define a consistent and defendable answer on three questions:

• Is the study ‘pivotal’ with respect to regulatory action for approval or labeling?

• Is the study in support of an internal decision?

• Will the study be included to support regulatory actions for labeling?

In addition, the workshop aimed to provide a toolkit to give BA scientists better insight on the types of studies supported and to help with the internal decision process on how to define assay specifications in support of a valid decisions. It was not the goal of the workshop to stimulate progress toward a tick-box or decision tree.

The organizing committee of the workshop (i.e., the authors of this manuscript) was mindful of Pharma and CRO expectations. In the absence of an agreed and globally endorsed alternative, performing work in the regulated validated arena is a quality statement and strong selling proposition for any CRO. As a consequence, there may currently be not a financial model to support approaches other than regulated validated assays.

Similarly, we identified the importance to focus on the expectations of the stakeholders (pharmacokinetics, toxicokinetics and pharmacologists and their management or quality assurance specialists), who may not be comfortable if the bioanalytical lab steps out of the regulated validated arena. We also aimed to familiarize all stakeholders with the scientific validity of data generated using the principles of tiered approach, and ensure all that it is not the intention of the bioanalytical community to promote ‘less valid data’ but rather to promote justified ‘scientific validation’. In support of this challenge, the workshop proposes a refinement of semantics for scientific and regulatory validated assay, which should facilitate discussions with our stakeholders. The difference is summarized in the following definitions:

• Regulatory validation:

  • Assay validations to provide concentration data that are scientifically accurate, reproducible and reconstructable to allow valid decision-making for the intended purpose of the study and comply with regulated BA standards as specified by Health Authorities (HA) guidance documents.

• Scientific validation:

  • Assay validations to provide concentration data that are scientifically accurate, reproducible and reconstructable to allow valid decision-making for the intended purpose of the study and can withstand independent review, including scientific review from regulators if so required – although not applying all elements specified by HA guidance documents.

Preworkshop survey

Prior to the workshop, the organizing committee organized a survey within the EBF community with 12 questions related to their current practice and the desire to further stimulate the use of tiered approach. It is not the intention of this manuscript to give detailed feedback of this survey, but a few highlights are important to underpin this manuscript and the workshop. Twenty-seven companies responded to the survey.

When asked to provide a good definition of tiered approach, different answers were given, but most of them included the words ‘fit-for-purpose’, ‘science’, ‘validation’, ‘stage of development’ and/or ‘documented’. From these answers, the team distilled a comprehensive definition for tiered approach which tried to capture the spirit of the survey feedback:

“A science driven fit-for-purpose strategy, applying a predefined appropriate level of bioanalytical quality, for the analysis of biological samples originating from preclinical and clinical studies. Level of pre- or in-study validation, acceptance criteria and documentation depends on the type of the study, intended use of the concentration data and/or considering the stage of drug development in which the study data are generated.”

All 27 respondents agreed that full validation as per regulatory guidance is not required for all the studies currently being analyzed. Also, all respondents acknowledged the benefit of changing our industry paradigm from ‘one size fits all’ validation into a mixture of ‘regulatory validation’ and ‘scientific validation’.

Taking a deeper dive and probing the companies on their current practices and preferred future practices on a broad array of studies in early and late development, the survey responses identified the areas of focus for the Brussels workshop. Four areas were defined:

• Agreement to continue using regulatory validated assays. Examples of studies suggested are:

  • Pivotal late development preclinical studies and clinical studies.

• Agreement to continue using scientific validated assay in line with the principles of tiered approach. Examples of studies suggested are:

  • Early non-GLP TK/PK studies, inclusive of non-GLP MABEL study;

  • In vitro studies;

  • Early development safety pharmacology studies.

• Opportunity to intensify the use of scientific validated assay in line with the principles of tiered approach. Examples of studies suggested are:

  • Early development (pre)clinical metabolite analysis in all matrices;

  • Tissue analysis (predominantly preclinical) and urine analysis (clinical);

  • Late development safety pharmacology studies;

  • Microdosing studies (cold or accelerator MS);

  • Animal PK and PK/PD studies;

  • Biomarker analysis in preclinical.

• Opportunity to start using scientific validation instead of regulatory validation. Examples of studies suggested are:

  • Early development GLP studies: for example, BA support for 28-day toxicology in rodents and nonrodents;

  • Early development clinical studies (single (SAD) and multiple dose (MAD) studies in human volunteers, early PK-PD (e.g., proof of concept, enabling in healthy volunteers, early PK studies other than SAD/MAD (e.g., food effect);

  • Reproductive toxicology studies;

  • Late development safety pharmacology;

  • LBA: Late development preclinical studies in special matrix.

There was no challenge from the survey data or at the workshop that the industry needs to comply with regulated bioanalytical standards for clinical or preclinical studies in late development (area 1) or to continue the use of other approaches in a discovery setting or for early PK studies (area 2). However, many of the respondents recognized the added value of a more intensive use of scientific validation (area 3) and already suggested some study types qualifying for this: tissue homogenate analysis; metabolites analysis; in vitro studies or urine analysis. In fact, for some of these study types, the EBF has already published recommendation papers [8,12–14]. Last, a significant part of the respondents challenged the use of regulatory validated assays for early clinical studies (area 4), not only for SAD and MAD studies but also for a whole array of other early development clinical studies. Surprisingly, the same challenge was given also for early GLP studies, more specifically for 28-day toxicology studies.

Design of the workshop

Inspired by many prior discussions during EBF internal meetings on tiered approach since 2008, aforementioned publications and guidance, the workshop also built on a recent discussion paper [15]. In addition to bundle prior discussion on tiered approach, that paper identified some reasons for the staggering application of tiered approach in industry and provided practical suggestion to stimulate its use.

On day 1, the morning session of the workshop started with an introductory presentation from the EBF in which an overview was presented of the current status of tiered approach. In this introduction, an interesting historic reflection was given to illustrate the challenge of the intended scope of the guidance versus the increasing number of studies requiring bioanalytical support today versus 15 (at the time of Crystal City II defining the current 2001-FDA guidance) or 25 years ago (at the time of Crystal City I, defining the acceptance criteria of, in essence, the bioanalysis clinical studies in healthy volunteers). At the time of the conception of the bioanalytical quality paradigm in the early 1990s, the majority of the studies requiring bioanalytical support fitted the scope of the discussions and subsequent guidance. This is less true today, where a variety of different study types with different questions in addition to these clinical is on the bioanalysts workbench today. From a filing perspective, the clinical Phase-I type studies will still be the vast majority, but in laboratory workload the balance with other studies is very different. The session continued with multi-faceted industry feedback. Steve White (GlaxoSmithKline), Stuart McDougall (Covance) and Michelle Manton (UCB Pharma/RQA) presented an industry perspective on tiered approach from an innovator, a CRO and a QA perspective. The session continued with practical examples giving current practice and hurdles for a small molecule portfolio (Tom Verhaeghe, Janssen R&D), for biomarkers (Marianne Scheel Fjording, Novo Nordisk). Mark Seymour (Xceleron) presented the journey of the accelerator MS community moving from ‘no validation’ into trying to fully comply with ‘regulatory validation’ and landing on acknowledging the advantages of ‘scientific validation’ [16,17]. Magnus Knutsson (Ferring, on behalf of EBF) concluded the morning session by summarizing the input from all presenters and preparing the delegates for the hands-on sessions of the workshop.

These hands-on sessions started from the input and outcome from the survey and the suggestions made in [15], the latter identifying the lack of harmonized process with sufficient practical detail for scientific validation and differences in interpretation of tiered approach as a reason of its poor adoption. Hence, a number of study types were identified to involve the workshop delegates to propose practical acceptance criteria. To each of these study types, a hands-on session was dedicated, sometimes preceded by one or two industry presentations to paint the landscape.

The breakout with focus on LC–MS-based assay included hands-on sessions on tissue analysis, metabolite analysis, urine analysis, nonpivotal clinical studies and nonpivotal preclinical studies. The breakout with focus on LBA-based assays included hands-on on general principles of tiered approach for LBA and biomarkers. From a drug development perspective, these hands-on sessions could also be classified as focusing on assay appropriate scientific validation (i.e., irrespective of the stage of development in which the study is supported, the proposed quality criteria for scientific validation support that valid and documented decisions can be made from the reported concentrations) or stage appropriate scientific validation (i.e., depending on the stage of development in which the study is supported, the proposed quality criteria for scientific validation may be different. In all stages, the reported concentrations support valid and documented decisions in the study).

LC–MS-based assays workshops

Prior to breaking out into hands-on session defining scientific validation for urine, tissues and metabolites (all three classified as assay-appropriate scientific validation), Morten Kall (Lundbeck) and Martijn Hilhorst (PRA Healtsciences) gave their company perspective on tiered approach for metabolite analysis and support of Phase I studies, respectively.

On day 2, the hands on session for LC–MS-based assays continued with two scenarios qualifying for stage of development as driver for scientific versus regulatory validation: early development nonpivotal clinical studies and preclinical (GLP) studies.

In practice, during both breakouts, the delegates were asked to provide their proposed criteria on two sets of parameters: acceptance criteria for scientific validation and acceptance criteria for study acceptance using a scientific validated assay.

The parameters discussed for validation were:

• Requirement of a certificate of analysis;

• Matrix of quality control (QC); and calibration samples;

• Intra-assay and inter-assay variability;

• Preparation process, number of assay runs and acceptance criteria for QCs and calibration samples;

• Selectivity;

• Extraction recovery;

• Requirements for internal standard;

• Carryover;

• Matrix effect;

• Dilution integrity;

• Definition of Lower Limit of Quantification (LLOQ);

• Selectivity toward comedication (drug–drug interaction studies);

• Selectivity toward over the counter medication;

• Costability in fixed-dose combination programs;

• Stability: processed sample, short and long term, F/T, stock solution stability and other matrices as appropriate (e.g., whole blood);

• Other parameters as appropriate for the assay (e.g., hemolytic, hyperlipedimic);

• Sampling conditions.

The parameters discussed for study sample analysis were:

• Accuracy and precision;

• Incurred Sample Reanalysis (ISR);

• Carry over;

• Dilution QC;

• IS variability;

• Anomalous results;

• Preparation of Cal and QC;

• Number of QC levels;

• Number of CAL points and QC replicates;

• Single or duplicate cal lines;

• Additional in study validation parameters in case no prestudy validation is performed.

LBA-based assays workshops

Prior to breaking out into hands-on sessions to discuss how tiered approach translates from the small molecule community into LBA assays, John Smeraglia (UCB Pharma) and Esther Biemans (QPS the Netherlands) engaged the delegates on the practical implications of risk-based approach and a case study on using duplicates versus singlets in LBA, respectively.

The LBA breakout on day 1 illustrated the difference of adoption of tiered approach in the LBA community versus the LC–MS community. Where in the latter these principles are being discussed and/or adopted for some time, it is only recently that the LBA community is joining this discussion at a more intense level. The sharing of thoughts during the breakout provided an important foundation for the discussions of day 2.

Building on the ‘prepared mind’ from the day 1 breakouts, the hands-on session for LBA-based assays on day 2 continued with proposing the assay criteria for stage of development of appropriate scientific validation for both PK and biomarker assays.

In practice, the delegates were asked to provide their proposed criteria for validation parameters: acceptance criteria for the validation and for the analysis of study samples using a scientific validated assay for both PK and biomarker assays.

The parameters discussed for validation were:

• Requirement of a certificate of analysis;

• Matrix of QC and calibration samples;

• Intra-assay and inter-assay variability;

• Preparation process, number of assay runs and acceptance criteria for QCs and calibration samples;

• Definition of LLOQ and upper Limit of quantification (ULOQ);

• Selectivity in individuals;

• Specificity of assay reagents;

• Interference of structurally related compounds;

• Dilution linearity;

• Stability: short and long term, F/T, whole blood;

• Other parameters as appropriate for the assay (e.g., hemolytic, hyperlipidemic);

• Sample collection procedures.

The parameters discussed for study samples analysis were:

• Accuracy and precision;

• ISR;

• Number of QC levels;

• Number of CAL points and QC replicates.

Documentation

On day 2, a plenary session focused on ensuring appropriate documentation and reporting practices for scientific validation discussion, the need and the content of a validation plan/protocol and/or validation report for the different assay types discussed.

Highlights of the discussions at the workshop

At the workshop, thanks to the broad and diverse expertise of the delegates (representing 37 companies, with equal spread of CRO vs innovator companies and LC–MS vs LBA experts), the discussion allowed a very complementary view on the tiered approach paradigm. This facilitated an exchange of opinions, experiences and practices, and resulted in a comprehensive draft proposal from the workshop on each of the types of assays discussed.

Certainly on the areas qualifying for assay-specific scientific validation (urine, tissue, metabolites), there was a consensus that there is a quick win for our industry if we can propose the minimal assay criteria for scientific validation. Of course, we would stimulate more discussion in and input from a broader group including the regulators prior to applying them in a transparent and harmonized way in industry. The overall gain would be very tangible; in other words, anybody involved in setting up an assay for urine or tissue would have a clear guidance on what the minimal requirements and acceptance criteria for a scientific validated assay would look like and in documenting and reporting back, we would speak the same language.

For LBA, there was an agreement that the complexity of the macro-molecules (proteins and therapeutic antibodies) and the specificity of the assay reagents are important parameters to take into account for scientific validated assays.

Similarly, the discussions around stage appropriate scientific validation for early development clinical studies yielded a starting position on which to build a consensus. Although the current use of the regulatory assay has created a comfort zone, there is a growing understanding that the regulatory framework is jeopardizing the need for more scientific freedom in early studies where neither dose, concentration, metabolites, matrix variations or the presence of interference may be better served allowing more scientific freedom by using scientific validation principles.

For early GLP studies (e.g., 28-day rodent and nonrodent studies), however, and in contrast to the clear signal from the survey which challenged the need for a regulatory validated assay for those studies, the delegates at the workshop were taken out of their comfort zone when discussing this suggestion. As iterated during the workshop and re-emphasized in this manuscript, the current thinking of the organizers of the workshop is that, although from a scientific perspective applying scientific validation instead of regulatory validation is a defendable approach, early GLP studies would not be the first area to propose to step away from regulatory validation practices. The inclusion of these studies in a later stage, when scientific validation has become an established practice, should not jeopardize today’s implementation of scientific validation in the areas where we are comfortable to start.

Our goal remains to promote human health through the rapid and safe development of new medicines. The drug exposure data generated by bioanalytical techniques is only a small (but essential) part of understanding the safety and efficacy of a drug. Without increasing risk to patients, a rigorous and scientifically justified application of a tiered approach to bioanalysis can ultimately increase drug development speed and reduce drug development costs.

Next steps

At the workshop, the delegates provided input for each area qualifying for scientific validation. The organizing committee has consolidated this input and shared with the delegates and with the broader EBF community for input and additional comments. The detailed results will be presented at international meetings in 2014 and early 2015, with a session already scheduled at the Seventh EBF Open Symposium (19–21 November 2014, Barcelona, Spain) for further input by the broader community and from there onward published as an EBF Recommendation.

Disclaimer

The views expressed in this discussion paper are the ones of the European Bioanalysis Forum and do not necessarily represent the views of its individual member companies.

Acknowledgement

The authors thank Martijn Hilhorst (PRA Health Sciences) Michelle Manton (UCB Pharma/RQA), Tom Verhaeghe (Janssen R&D), Mark Seymour (Xeleron) and Esther Biemans (QPS the Netherlands) for their contribution to the Workshop.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending or royalties.

No writing assistance was utilized in the production of this manuscript.