The Decennial Index of the White Papers in Bioanalysis: ‘A Decade of Recommendations (2007–2016)’

BioanalysisVol. 9, No. 21 Special FeatureFree AccessThe Decennial Index of the White Papers in Bioanalysis: 'A Decade of Recommendations (2007–2016)'Wei Garofolo & Natasha SavoieWei Garofolo*Author for correspondence: Tel.: +1 514 461 0877; E-mail Address: weigarofolo@cfabs.org CFABS, Montreal, QC, Canada & Natasha Savoie CFABS, Montreal, QC, CanadaPublished Online:3 Nov 2017https://doi.org/10.4155/bio-2017-4979AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinkedInReddit First draft submitted: 24 July 2017; Accepted for publication: 7 August 2017; Published online: 3 November 2017Each year, the Workshop on Recent Issues in Bioanalysis (WRIB) gathers a wide range of industry opinion leaders and regulatory authorities working on bioanalysis, biomarkers and immunogenicity to discuss the most current topics of interest, and to provide potential solutions aiming to improve quality, increase regulatory compliance and achieve scientific excellence.These yearly 'hot' topics (covering both small and large molecules), are the starting point for fruitful exchanges of knowledge, and extensive sharing of ideas among presenters, regulators and attendees, and are distilled into a series of relevant recommendations. The resultant White Papers summarizing these conclusions and consensus points from each WRIB provide the global bioanalytical community with key information and practical solutions on topics and issues addressed each year. This series of White Papers has always been among the 'most read' articles in Bioanalysis over the past 10 years [17].WRIB White Papers are designed to be 'horizontal' documents providing consensus on multiple topics rather than 'vertical' (i.e., 'consensus documents on a single topic for standardizing industry practice'). It is a cumulative effort spanning more than a year, with tremendous work behind the scenes by each co-author, starting from exhaustive preparation/design of each issue long before the WRIB, then extensive working-dinner discussions to prepare open panel discussions at the WRIB, and finally the refinement and critical review of the draft recommendations before turning it into a consensus paper.In recent years, to reflect the new structure of three sequential, core workshop days at WRIB, each White Paper has been divided into three parts to reflect the focus of each core workshop day: Part 1 on LCMS, Part 2 on hybrid ligand-binding assays (LBA)/LCMS and Regulatory Inputs and Part 3 on LBA.Although the recommendations presented in this article demonstrate that some topics have been extensively discussed repeatedly in the last decade, the evolution of bioanalysis, biomarkers and immunogenicity, and the accompanying technologies and regulations continue to bring additional challenges and solutions to existing problems. Continued assembly of industry and regulators will ensure that best practices continue in order to market safe and effective pharmaceutical products. The future publication of the 2017 White Paper in Bioanalysis signals the global bioanalytical community's intent to continue providing a consensus document for such discussions.After a decade of discussions and consensus, this article acts as a general index, organized by topic in alphabetic order, to easily consult all these recommendations and their evolution over time.Alphabetic IndexNotes for the readers1) Select items from the Alphabetic Index in the drop-down menu to reach the paragraph of interest2) Click on the reference in the square brackets [X] to reach the White Paper of interestTopicsAnomalous resultsIn 2009, it was determined that analytically anomalous results must be investigated [2]. The recommendation was expanded the next year, stating that a standard operating procedure (SOP) must exist outlining the procedure for investigations. Investigations should not only be initiated in the case of out-of-specifications, but also if there are any anomalous results or unexpected trends in the data [3]. These recommendations were reconfirmed in 2015. Thorough and documented investigations of anomalous results, including an assessment of impact on the data, are expected. Samples with anomalous values selected and repeated by the laboratory should be defined a priori by SOPs and in compliance with current regulatory requirements. The extent of an investigation should be SOP-driven and aimed at identifying both systemic and isolated incidents [11].Antidrug antibodies: analysis in nonclinical studiesStriving for the highest sensitivity was not considered necessary for antidrug antibody (ADA) analysis in nonclinical studies. A screening cut point at the 99.9th percentile was preferred and consequently confirmatory assays and titration assays were not required. Detailed ADA analysis should be conducted if impaired PK/PD, toxicity or ADA related adverse events are observed. For high-risk molecules with an endogenous counterpart, it might also be helpful to assess neutralizing antibodies to evaluate the consequences of cross-neutralization of this endogenous molecule [16].Antidrug antibodies: duration & terminologyThe topic was extensively discussed in 2016, and several recommendations were reported. First, to illustrate and describe ADA duration, a graphical option can be considered when sample size is statistically significant. This option presented data as 'median duration of ADA' and used inter-quartile values (Q1, Q3) to describe the duration of ADA in half, 25% and 75% of the ADA-positive subjects. Moreover, to properly classify the ADA duration data, at least 1 year of immunogenicity data were recommended [16]. A 16- or 12-week duration were both deemed acceptable to characterize the patient as having an ADA as persistent response and an adequate sampling schedule should be selected. The duration of follow-up testing for ADA-positive patients should be data driven and should not default to continuous monitoring until subjects become baseline negative [16].Antidrug antibodies: effect on PK assaysIn 2013, it was recommended that when trying to determine the effect of ADAs on PK assays, a risk-based approach should be implemented as not all abnormal PK is inherently due to ADA. Knowledge of the expected drug therapeutic window and information regarding possible interference by ADA in the PK assay should be applied [7]. Routine assessment of ADA interference in PK assays was not recommended; however, it is important to understand the PK assay and to also consider data from other relevant assessments (e.g., PD and efficacy). Additionally, there is no need to routinely employ strategies to mitigate ADA interference, although there should be an understanding of how ADA may impact PK results [10].Antibody–drug conjugates: immunogenicity assaysAntibody–drug conjugates (ADCs) are only one of many classes of multiple functional domain biotherapeutics. As with other compounds with complex multidomain structures and potentially multistep modes of action, ADCs require a series of immunogenicity tests to fully understand the immunogenic profile. The thinking following discussions in 2014 was that specificity testing should be conducted for each domain with a functional endpoint, for example, the antibody and payload. Further testing may be merited based upon a customized risk-assessment [10].Antibody-drug conjugates: PK assaysIn 2013, it was recommended that during the initial discovery phase evaluation of ADCs, many analytes will need to be tested, requiring several assays to be developed and qualified. The type of analytes to track during later nonclinical and clinical phases will be determined by the exposure type that provides the signals that provide the best correlation with the clinically relevant patient outcomes. Assays for these analytes require full validation [7]. The main challenges for ADC bioanalysis discussed in 2014 included heterogeneity of the reference material containing molecules with different drug–antibody ratios, changing ADC composition in biological samples and availability of the adequately characterized reference materials for all the molecular species with different drug–antibody ratios. The complexity of ADCs requires multiple bioanalytical methods including mAb assays, unconjugated payload LCMS assays and ADC conjugate assays [9]. As the content of the conjugated payload in ADCs is typically vastly higher than the concentrations of the unconjugated payload in matrix, it could be challenging to prevent the cleavage of ADCs and artificial formation of unconjugated payload during in vivo sample collection, processing and analysis. Thus, it may be difficult or even not feasible to meet the small-molecule method validation acceptance criteria for the unconjugated payload assays. Acceptance criteria for payload assays may need to be widened based on the experimental scientific results. For quantification of ADCs, LCMS assays may be more sensitive and selective. They may be applicable to different ADCs with similar payloads, and they often require less time for method development and validation [9].Anticoagulant counterionsDiscussions regarding anticoagulant counterions began in 2008, when Health Canada began requesting data to determine the impact of different anticoagulant counterions on sample analysis results and in 2009, the discussions on this topic continued. It was stated at that time, however, that the anticoagulant used in the study protocol should be consistent with the validation, otherwise additional testing should be done [3]. Following 2 years of testing the impact of anticoagulant counterions by industry, it was concluded in 2011 that no significant impact had been observed. This could be verified during method development via benchtop stability testing with the different anticoagulant counterions of interest [5].Autosampler stability & reinjection reproducibilityEvaluations of autosampler stability and reinjection reproducibility have been a hot topic of discussion since 2008, and as recently as 2016. It was clarified in 2008 that autosampler stability and reinjection reproducibility are two different evaluations that test stability during different parts of samples analysis. Depending on how routine sample analysis is conducted in a bioanalytical laboratory, it may be necessary to perform one or both evaluations during method validation [2]. Best practice for extract stability dictates that stored low and high QC samples should be measured against a fresh calibration curve, and analyte/IS ratio should be used [14]. For small molecules, samples can be reinjected when a cause for a batch failure can be assigned, however, samples need to be reanalyzed when processed stability is surpassed or when no assignable cause is found. For large molecules, reanalysis applies when a batch fails [6].Biomarker assay validationDiscussions surrounding biomarker assays and validation began in 2012, when it was recommended in order to avoid confusion with the term GLP (Good Laboratory Practices) in biomarker analysis, either 'regulated' or 'non-regulated' bioanalysis should be used as part of nonclinical and clinical biomarker studies [6]. Biomarker assay validation (BAV) requirements were deliberated in 2015, when it was established that exploratory and confirmatory biomarker categories are not always distinct. The application of an assay may evolve during the drug development and the levels of assay validation required may vary depending on the intended use of the data [13]. Due to a variety of challenges, developing and validating the biomarker assay cannot always be performed to meet the standards of a PK assay as requested by the 2013 US FDA draft guidance. It was recommended to have at least one quality control (QC) in the same matrix as the study samples, included among low, medium and high QCs. Stability testing should be performed in both buffer and sample matrix under actual conditions used during sample analysis. Matrix stability should be tested with matrix QC samples where available. Precision and accuracy of a biomarker assay should be assessed. However, when the biomarker methods are relative quantitative or qualitative, assay validation efforts should be focused on assay precision or reproducibility [13]. Discussions regarding BAV continued in 2016 and recommendations were similar to those from the previous year; stability recommendations evolved and were expanded [16]. If reference standard material and isotopically labeled material are available, then standard bioanalytical method validation (BMV) PK assay criteria may be applied for small-molecule biomarker methods. An endogenous QC sample should be included during method validation in order to demonstrate the suitability of the biomarker measurement method. The acceptance criteria of the endogenous QC sample should be fit-for-purpose and based on the biology and variability of the biomarker. The approach should be data driven, and determined on acquired validation data. Whenever possible, matrix pools from both healthy and patient populations should always be included in the biomarker validation [13]. For biomarkers, it was further determined in 2016 that it is not necessary to monitor stability of both the spiked analytical QCs and endogenous sample QCs. Only the endogenous sample QCs need to be shown to be stable for protein biomarkers. Moreover, biomarker stability should be supported by using incurred samples so that biomarker variability induced by storage can be more precisely monitored over time. Individual samples should be used for assessing stability of biomarkers in the assay and the use of a 'mega-pool' to assess long-term stability is not applicable. BAV must be fit-for-purpose and endogenous samples should be used as QCs rather than buffer controls and surrogate QCs as much as possible to evaluate assay validation parameters [16].Biomarkers: small moleculesWhen assaying small-molecule biomarkers, it was recommended to use a stable-isotope labeled (SIL) analyte instead of a surrogate matrix when feasible. When selecting the labeled forms, it is preferable to use 15N or 13C analytes if the introduction of these isotopes is chemically possible. A consistent response ratio within the small-molecule biomarker assay between days was recommended as a minimum standard, while some proposed an equivalent one as a higher level of performance. When using surrogate matrices, pooled incurred samples as endogenous QCs provide a good understanding of data reproducibility between days, which suffices the incurred sample reanalysis (ISR) needs during the early stage of drug development. The formal ISR in the traditional design may be performed at a later stage. Regardless of which surrogate approach is used, matrix or analyte, a parallelism experiment for small-molecule biomarkers was recommended [14].Biomarkers in tissues: LCMSIt was noted in 2014 that LCMS is well suited for the bioanalysis of biomarkers in tissues as it is typically less dependent on reagents that could be potentially interfered with by the tissue matrix. One particular approach that may be suitable for the analysis of membrane proteins employs tissue disruption and harsh denaturing conditions for extraction followed by proteolytic digestion prior to LCMS analysis [9].Biosimilars: clinical immunogenicityBased on industry and regulatory experience reported in 2015, both one-assay and two-assay approaches for the demonstration of clinical immunogenic comparability for biosimilars were considered valid approaches and accepted by regulatory agencies [13]. However, it was recommended during the 2016 Workshop that one assay is preferred in which the biosimilar is used as both the capture and detector reagent [16].Biotransformation of biotherapeuticsDiscussions in 2015 were heavily focused on biotherapeutics and the impact of biotransformations. Ideally the quantified analyte reflects active drug. Measuring inactive drug may be regarded as supplementary information only and, as such, was not recommended in 2015 as the only quantitative assessment of drug levels [13]. An appropriate bioanalytical strategy should be developed to allow a differentiation between drug elimination and conversion. In order to come up with the most appropriate bioanalytical strategy, multiple technologies and tools, and various approaches including qualitative and quantitative methods may be needed to understand the impact of biotransformation of biotherapeutics on function. It was recommended that both LCMS and LBA techniques should be used in a complementary manner to understand biotherapeutic biotransformation: LCMS and hybrid LBA/LCMS for defining protein biotransformation and cell-based assays for understanding activity [13]. Discussions in 2016 concluded that hybrid LBA/LCMS methods are increasingly used to identify, characterize and quantify biotherapeutic biotransformations using an integrated approach. Understanding biotransformations will help scientists determine the impact on efficacy, safety, PD and PK, as well as help to correlate with in vitro monitoring. If biotransformations of biotherapeutics are present in quantities of more than 10% of the starting material, it was recommended to follow the MIST guidance regarding quantification recommendations.. For quantities below 10%, the minor modified biotherapeutic is likely to have less impact on the overall activity, depending on how important the biotransformation is to the biotherapeutic [15].Bispecific therapeutic proteinsBispecific therapeutic proteins may require additional mechanistic characterizations associated with immune responses in addition to ADA testing such as monitoring cytokine release and other immune system modulation events [13]. Screening and confirmatory assays should minimally be performed against the intact bispecific antibody per conventional testing strategies. Testing for specificity against the functional domain may be justified as the incidence and clinical impact merits. Specificity testing can be done through competitive binding experiments in the confirmatory tier of the testing strategy; alternatively, it can also be achieved by implementing fully independent, domain-based screening assays. The choice of these approaches should be case dependent [13].Calibration curve: regression modelWhen selecting a regression type, guidance documents state that the simplest model should be used [2]. In 2008, the acceptance of nonlinear calibration models and how much quadracity is acceptable were discussed [2]. A second look at the topic in 2015 resulted in the recommendation that the selection of the regression model should be step-by-step, beginning with a simple model (linear) and then moving to more complex models (quadratic), each testing the model improvement through the use of weighting, supported by a predefined statistical approach and documented as part of company SOPs [11].Calibration curve: slopeThere is no requirement for verifying the trend in slope values. However, the recommendation in 2011 states that although different calibration curve slopes may be obtained between different LCMS systems, a coefficient of variation of ≤15% in slope values within injection systems suggests that the robustness of the method is not affected [5]. This was expanded in 2015 to add that unexplained, extreme variations should be evaluated to assess impact on accuracy [11].Calibration curvesThe use of either bulk or fresh preparations of calibration standards can be used during sample analysis, with proper supporting stability data, if necessary [4]. However, when performing stability evaluations, a 'fresh' curve preparation means freshly fortified calibration standards only, while QC samples may not need to be freshly fortified since subject samples are generally frozen at least once and thawed [5].Capillary microsamplingCapillary microsampling, as a different collection technology, was addressed in 2014. It was recommended that microsampling should be applied based on scientific, ethical and regulatory considerations. The conduct of plasma capillary microsampling itself presents no major differences when compared with traditional plasma sampling. However, additional consideration may be required for ISR which will often be conducted on diluted samples, and additional testing may be required when using fixed volume capillaries. Appropriate training is crucial in its successful implementation [8].Carryover (LCMS)In 2010, carryover was discussed in detail and it was agreed that carryover must be investigated during method development, and minimized [4]. The most common procedure used by the industry is to inject a blank or zero sample after an upper limit of quantitation (ULQ) sample, or multiple ULQ samples. If carryover is inevitable, a nonrandomized sequence should be used.Chromatogram integrationChromatogram integration has been a topic of discussion since 2009. At that time, it was recommended that manually integrated chromatograms should occur infrequently since it is important to correctly set the integration method optimally for all samples in a run. However, good scientific judgment must be used in the cases where a change is required and changes must be documented with management approval. Manually modified chromatograms are not generally reported in the final report but they may be reported if desired [3]. In 2011, the recommendation was clarified for cases when manual modifications were judged necessary; they may be performed with scientific justification, before the results are regressed and documented in a controlled fashion [5]. Still, further clarifications were requested in 2016 in order to better define a manual modification. It was concluded that chromatographic and mass spectrometric parameters that impact the way data are acquired should be fixed in validation. Permissible changes are to those parameters that can be applied post acquisition, based on the general principal that integration must be correct and consistent. However, these changes should be infrequent and based on documented procedures [14].Collision-induced dissociation fragmentation in LCMSWhen collision-induced dissociation fragmentation inefficiency is observed, survivor scans can be successfully used with compounds that do not fragment well by MS such as cyclic peptides, oligonucleotides, steroids, opiates or similar [14].Co-administered compound stabilityFirst discussed in 2010, regulatory agencies began questioning the stability of samples containing co-administered compounds. Following extensive input from industry experts, no examples were provided where analyte stability was affected by co-administered compounds, thus, bringing into question the scientific rationale for such stability assessment requirements [4]. Discussions continued in 2011 and 2014, in an effort to determine the scientific needs of performing co-administered stability tests. In 2011, it was concluded that the decision to demonstrate analyte stability in the presence of such compounds should be based on a scientific rationale rather than using a systematic approach [5]. No cases of stability issues were reported as of the 2014 discussion either [8].Commercial kits: biomarkersCommercial kits are often used to assay biomarkers. The recommendation from the 2013 Workshop states that changes to an imperfect commercial kit can be made to enable validation. For example, only critical reagents from the kits may be used to build an assay. A strong consensus was reached on the fact that the data provided in the kit insert should not be used to support assay validation [7]. Attendees at the 2015 Workshop established that the best-practices for confirming the specificity of commercially available immunoassays used for clinical biomarker analyses include characterization of the commercial kit [13]. As determined in 2013, the presence of a signal and sensitivity specifications as determined by the vendor are not enough to confirm that the kit is detecting the biomarker of interest without additional testing; assessing assay specificity is the key. Reference standards from different sources should be tested. Parallelism testing should be performed to assess whether recombinant material is comparable to endogenous analyte. If only one kit lot is to be used for the entire study, interlot variability testing may not be required. However, if multiple lots are needed, it was recommended to establish a lot-bridging strategy and test as many lots as reasonably possible to assess risk. If interlot variability is observed during study sample analysis, the vendor(s) should be contacted to replace the critical reagents and if not possible, appropriate statistical assessments on the impact of variability and the use of correction factors can be explored to determine the best path forward [13].Critical reagentsOne key component of LBA methods is critical reagents. Adequate quality of critical reagents should include an early assessment of binding properties (e.g., cross-reactivity to matrix components), binding kinetics and any altered binding issues due to, for example, labeling or storage. Biophysical properties of critical reagents are required by regulators and include identity, concentration, storage conditions and stability information. Purity information is also important [13]. The retest date of critical reagents is first assigned based on previous history with the material, and stability is continually monitored in order to determine the final expiration date [13].Cross-validationWhen cross-validating methods, procedures and criteria may vary depending on the assay, the magnitude of the change and the sponsor's requirements [5]. For example, if a single study is done at two different sites, a cross-validation using spiked QC samples and incurred samples was recommended. If the same method SOP is used at the two sites, one cross-validation run using QC samples may be sufficient [8]. The recommendation from 2014 states that cross-validation with an existing technology is necessary only if the change of platform occurs within a program or a study; however, the presence of unique variables that could impact quality should be considered [8].Cross-validation (LBA)Discussions in 2014 concluded that changes in assay formats are inevitable when moving from early to later stage LBAs. Assays in support of earlier phases will typically be restricted by available critical reagents. It is likely that changes in PK profiles will occur as new, more specific critical reagents become available with the evolution of methods. It is not necessary to have the same result from different PK assays, but rather that there should be a scientific understanding of the species that the assays are measuring so that results can be interpreted appropriately [10]. Continued discussions in 2015 recommended that when there are differences in concentration results between two LBA platforms that had been previously cross-validated, an investigation would be needed to understand the reasons for the differences. The reasons for choosing one assay platform over another should be documented [13]. Additional discussions in 2016 concluded that when results demonstrate differences in concentrations when cross-validating methods, the use of correction factors is discouraged. If the cross-validation meets criteria but there are still significant differences between a subset of incurred sample data, the issue should be investigated as needed. Trends or correlations should be examined, as well as how the results relate to an identifiable cause [16].Cross-validating (LBA & LCMS)Prior to any cross-validation of LBA and LCMS techniques, it was recommended in 2014 to confirm that both technologies measure the same analyte [12]. It was not generally recommended to measure an analyte with both approaches; however, this may happen when one assay or technology runs into issues such as matrix interference, when early data are generated on one platform and then the platform is switched to the other; or when the molecular complexity necessitates additional characterization. When cross-validating between LCMS and LBA, QC samples and incurred samples should be used, and the general acceptance criteria that should be applied are two-thirds of sample results within 30% [9]. The following year, this recommendation was expanded to conclude that if cross-validation of LBA and LCMS data are required, then it is recommended to use the same capture reagent [12].Cut-pointsThe use of floating cut-points (CPs) by pooling the variability across all the assay runs is generally acceptable for screening CP evaluations even when the variances are significantly different between assay plates/runs, because the use of dynamic CPs is often impractical [13]. However, prior to directly implementing the floating CP approach, potential causes for the variance heterogeneity should be investigated and re-optimization of the assay should be considered when feasible [13]. For confirmation assays, the fixed CP approach is still the most viable strategy even when the means and/or variances are significantly different across assay plates/runs. Methods for floating CP strategy may be considered by spiking the negative control with excess study drug, but this requires further evaluation and validation. Titration CPs are necessary when the screening CPs are too low and fall on the lower plateau of the positive control (PC) dilution curve. The CPs determined during validation may sometimes not be suitable for testing the clinical study samples. If the false-positive rate of the clinical baseline samples based on the pre-study validation screening CP is under 2% or over 11%, the use of a study-specific CP should be considered [13].Dried blood spotsDried blood spot (DBS) analysis was highly anticipated as an