Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences

jo urn a l h om ep ag e : w ww.jp harms ci.o rg

General Commentary

Quality risk management and data integrity in R&D laboratories

supporting CMC lifecycle of biological products
Brent S. Kendricka,*, John P. Gabrielsonb, Deanna Huntc, Merry Christied,
Steven Bowene, Christina Vesselyf, Richard S. Rogersg, Chad Clevelandc, Karl Malufc,
Shawn Roachh, Nadine Ritteri
a
First Principles Biopharma, LLC, Louisville, CO, USA
b
Similis Bio, JSR Life Sciences, LLC, Sunnyvale, CA, USA
c
KBI Biopharma, Inc., Louisville, CO, USA
d
Orchard Therapeutics, PLC., Boston, MA, USA
e
Eliquent Life Sciences, Washington, D.C., USA
f
Biologics Consulting, Alexandria, VA USA
g
Umoja Biopharma, Inc., Seattle, WA USA
h
Halloran Consulting Group, Inc. Boston, MA USA
i
Global Biotech Experts, LLC., Germantown, MD, USA

A R T I C L E I N F O A B S T R A C T

Article history: The development of pharmaceutical products is the critical bridge that moves a potential new medicine from aca-
Received 30 May 2024 demic discovery to applied treatment of patients. It translates an idea for a new drug to bench-level research on
Revised 12 September 2024 how it can be manufactured, formulated, characterized and controlled for use in non-clinical and early clinical trials.
Accepted 13 September 2024
From pre-clinical R&D discovery work through the commercial launch, substantial R&D CMC data is gener-
Available online 21 September 2024
ated to develop and optimize cGMP manufacturing and testing operations, while also supporting product
comparability, elucidating product / impurity structures, assessing critical quality attributes, developing the
Keywords:
drug delivery mode, and developing the product formulation for long-term stability. Significant R&D CMC
Biologics
Biological drug development
work continues post-approval to support continuous improvement and market expansion of the commercial
Quality systems product. These activities are crucial elements of Product Lifecycle Management, and taken together, they
Regulatory filings comprise Pharmaceutical Quality or Chemistry, Manufacturing and Controls (CMC). The objective of this
Biopharmaceutical characterization paper is to mitigate the regulatory ambiguity of R&D quality systems with practical, risk-based examples and
Biosimilars recommendations when conducting supportive CMC studies for biological products.
Comparability Making sound strategic CMC decisions under any circumstances assumes data from R&D studies are reliable,
traceable, and complete. While there are specific regulatory guidelines on phase-appropriate cGMP activities,
none exist for quality practices in R&D CMC laboratories conducting non-cGMP studies. Hindsight is not the
time to discover that R&D studies lack key elements that would otherwise have allowed the data to be
directly presented to regulators, if needed. There is a strong prospective business interest in protecting con-
siderable investments made for CMC R&D studies. Therefore, establishment of a robust and stage-appropriate
R&D laboratory quality system is essential for companies seeking to capitalize on prior knowledge, protect
investments, and be prepared for accelerated approval pathways.
© 2024 American Pharmacists Association. Published by Elsevier Inc. All rights are reserved, including those
for text and data mining, AI training, and similar technologies.

Abbreviations: cGMP, current good manufacturing practices; CMC, chemistry,

Introduction
manufacturing and controls; EC, established conditions; ICH, international conference
on harmonization; CQA, critical quality attribute; CGTs, cell and gene-based therapies;
ATMPs, advanced therapy medicinal products; R&D, research and development; Research and development (R&D) data are utilized extensively
CDMO, contract development and manufacturing organization; ALCOA, attributable, throughout chemistry, manufacturing, and controls (CMC) develop-
legible, contemporaneous, original, accurate; CFR, code of federal regulations; CTU, ment to inform process development and optimization, product com-
controlled temperature unit; IQ, installation qualification; OQ, operational qualifica-
tion; PQ, performance qualification; DV, data verification; QC, quality control; QA,
parability after process changes, elucidation of product structure,
quality assurance, DS, drug substance; DP, drug product; MS, mass spectrometry. characterization of product and process impurities, identification of
* Corresponding author at: Louisville 80027, CO, USA. critical quality attributes (CQAs), formulation optimization and
E-mail address: [email protected] (B.S. Kendrick).

https://doi.org/10.1016/j.xphs.2024.09.013
0022-3549/© 2024 American Pharmacists Association. Published by Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technolo-
gies.
3124 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

robustness, and long-term stability. Significant R&D CMC work con- Supportive information: R&D CMC data provided to health
tinues post-approval to support continuous improvement and mar- authorities that are not directly governed by cGMP regulations
ket expansion of the commercial product. There is increased
regulatory flexibility in leveraging prior knowledge and platform Using the ICHQ7 MCB/WCB example above, although not subject
experience gained from past products to justify CMC development to cGMP, it is expected that the R&D activities for establishing the
decisions with new products. Much of the supportive data may be MCB/WCB will be thoroughly documented, and the cell bank charac-
generated in non-cGMP R&D laboratories. For biosimilar products, terization and stability data will be reliable and accurate.
the comparative analytical assessment (CAA) studies against the ref- Regulatory guidance defines these non-cGMP R&D studies as
erence licensed product are almost entirely analytical data generated “Supportive Information” .3,4 Numerous international and regional
from non-cGMP methods in R&D. Also, numerous global Health guidances indicate which types of supportive CMC data are expected
Authorities now have accelerated approval pathways for new prod- for biological products for pre-clinical, clinical, and post-approval
ucts serving unmet medical needs. When CMC development is phases6,8,9 including emerging advanced therapies (gene and cell
squeezed by compressed clinical timelines, decisions often rely upon therapy products).10
R&D studies from early development. These activities are crucial ele- For purposes of R&D quality systems, in the sections below we
ments of Product Lifecycle Management, and taken together, they will further delineate the nature of R&D studies included in Support-
comprise Pharmaceutical Quality or CMC. ive Information sections of regulatory filings.
When R&D labs conduct CMC studies to support the development
and lifecycle management of a biological product, they are expected Potential regulatory risks
to generate accurate and reproducible data. This basic scientific tenet
is a simple principle that should be followed in all R&D labs. The reli- From the framework above, ‘Supportive Information”3 derived
ability of R&D data can be compromised by a host of hidden threats, from R&D CMC studies are presented in sections of Module 3 in a
including undetected inconsistent or unstable reagents or materials, product IND and BLA dossier (or equivalent clinical trial and product
uncalibrated or out-of-tolerance instrumentation (including pipet- registration applications outside the US). Not surprisingly, regulators
tors), malfunctioning equipment (e.g. refrigerators, freezers, incuba- expect that all R&D CMC supporting data provided for their review
tors), inappropriately developed or incorrectly applied testing will be accurate and reliable:
procedures, inadequate or inconsistent documentation practices,
“It is the manufacturer’s responsibility to submit complete and
incomplete or uncontrolled raw data files, undetected or undocu-
accurate information in marketing applications for evaluation by
mented analyst mistakes, and lack of appropriate checks and balan-
the FDA.”11
ces for inadvertent human errors.
In the absence of tailored R&D quality guidance from regulatory
Regulatory expectations for the quality of supportive CMC data
agencies, industry managers, laboratory and study directors and lead
apply whether the supportive CMC data are generated by the spon-
scientists’ may erroneously implement cGMP or GLP. Or they may do
sor’s own R&D laboratories or in R&D laboratories of contract devel-
nothing and simply hope for the best. Applying unnecessary cGMP or
opment, testing, and manufacturing organizations (CDMOs).12
GLP requirements creates an overly restricted environment which
The verification of authenticity and data integrity for supportive
may stifle innovation, extend R&D timelines, and add undue costs to
CMC data presented in Module 3 are among the specified objectives
drug development. But hoping that each R&D scientist will figure out
of FDA pre-approval inspections (PAIs) for new products.13 Although
best practices, or that mistakes will be effortlessly identified and cor-
Objective #1 of an FDA PAI is to confirm cGMP compliance for com-
rected, poses significant risk to a company’s investments in key CMC
mercial manufacturing and testing of drug substance and drug prod-
studies. R&D quality gaps can jeopardize IP claims, impact due dili-
uct, the other three PAI objectives are related to quality risk
gence audits by partners, misinform strategic decisions, trigger the
management and data integrity (underlines added):
need to recreate lost or unreliable R&D data sets, and undermine the
company’s reputation. Furthermore, R&D data is subject to regulatory “Objective 2: Conformance to Application (Covered on every
review and inspection, especially data that is utilized in cGMP PAI) . . .Coverage of this objective should also include inspection
method/process validation packages. It is therefore important to of laboratory methods, including auditing Research & Develop-
assure the reliability and traceability of the supportive data. ment notebooks. . . Inspections and audits under this objective
To outline the framework of laboratory quality practices, CMC and Objective 3 are to verify factual and contextual integrity of
activities1,2 can be categorized by the purpose of the activities being the information filed in the application. Information that has fac-
conducted: tual integrity is information that is original and corresponds
directly to that submitted to the agency (e.g., a chromatogram
Established conditions: production and testing activities governed showing a peak area that directly calculates to an assay value sub-
by statutory regulations for current good manufacturing practices mitted in a data summary sheet in the application). . .”
(cGMPs)
“Objective 3: Data Integrity Audit (Covered on every PAI) . . .Audit
the accuracy and completeness of data reported by the facility for
Regulatory guidance defines these operational commitments as
the product. Not every CMC data summary must be audited to
“Established Conditions,” (EC) .3,4 Several international and regional
accomplish this objective. The inspectional strategy may select key
regulatory guidances describe ‘sliding scale’ or phase-appropriate
data sets from drug development. . . During the inspection, compare
cGMP expectations for the production and testing operations for clin-
raw data—hardcopy or electronic—such as chromatograms, spectro-
ical trial materials.1,5−7
grams, laboratory analyst notebooks, and additional information
Some guidances also provide information on which point in the
from the laboratory with summary data filed in the CMC section.
Established Conditions the cGMP regulations start to apply. For
Raw data files should support a conclusion that the data/information
example, ICH Q7 explains that for recombinant biological drug sub-
reported by the site are complete and accurate. . .”
stance manufacturing processes, activities for the establishment of the
master and working cell banks (MCB/WCB) are not required to be “Objective 4: Commitment to Quality in Pharmaceutical Devel-
conducted under full cGMP.1 cGMP starts with the maintenance of opment (Covered on initial PAI and periodically on subsequent
working cell banks to be used for upstream production. PAIs) . . .Review the tools, procedures, or strategies put in place by
 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136 3125

the facility as part of its overarching pharmaceutical development For completeness, two other categories of CMC data are men-
program and determine whether the pharmaceutical develop- tioned here, but for reasons specified below are excluded from the
ment report for the application product aligns with the develop- scope of this paper:
ment program. . . Determine whether adequate risk assessment
activities are included as part of the pharmaceutical development  Tier 0 = R&D activities that are conducted for internal purposes
program. . . Confirm procedures are put in place to reduce or miti- only for the support of preliminary CMC decisions. We recom-
gate the risk. Assess the firm’s use of quality risk management mend that each company assess their own risks and implement
principles during development and verify that adequate steps are quality practices that best balance the level of risk against the cost
included as part of the development program. . .” of time and money. [Out of Scope]
 Phase-appropriate cGMP activities = we do not assign a Tier desig-
The impressions gathered by an FDA inspector − especially from nation to this category because they are already subjected to the
their assessment of quality practices in pharmaceutical development sliding-scale enforcement of statutory quality requirements. For
- can have long-reaching implications on a sponsor’s reputation for these, we defer to the numerous international and regional guid-
their attentive oversight of scientifically-sound and reliable product ance documents on phase-appropriate cGMP manufacturing and
development activities: QC testing practices given in the references. [Out of Scope]
“The information gathered from Objective 4 coverage during a PAI
is generally used for data analysis or internal trending by FDA and
may assist in identification of risk factors (e.g., risks related to pro- The practical R&D quality recommendations we provide for Tiers
cess, firm history, and product) for future PAI decisions. Coverage of 1 and 2 R&D data are adapted from relevant principles of quality
Objective 4 helps FDA’s decision-making related to the firm’s effec- compliance systems found in QC laboratories and US-based cGMP
tiveness in developing new products. . .[it] provides an opportunity regulations (CFRs)20−22 and ICH guidelines.3,23−25 Also, although we
for investigators to observe and document examples of mature qual- have experience in R&D quality practices from numerous laborato-
ity practices. . .” ries, our recommendations are not based on quality systems from
Regulatory scrutiny of data integrity systems became even more any specific company, product, or region. Rather, they are derived
urgent after Health Authority field inspections revealed widespread from the authors’ collective observations of successful − and disas-
gaps in data integrity in cGMP facilities.14 In response, four health trous − business and regulatory situations regarding the reliability
agencies generated guidance documents outlining the expectations (or not) of non-cGMP CMC data from R&D laboratories associated
for data integrity, and practices that would support data integrity with the development and commercial support of biological prod-
quality systems.15−18 The objective, as stated in the FDA guidance on ucts.
Data Integrity, is that: The benefits of implementing a comprehensive, consistent R&D
quality system are (1) to lower the business risks of losing significant
“(Data integrity) system design and controls should enable easy investments of time and money in R&D studies (that in hindsight are
detection of errors, omissions, and aberrant results throughout the too unreliable to trust), (2) to increase confidence in CMC decisions
data’s life cycle”. made by the company (backed by scientifically sound high-quality
data) without implementing costly full-on cGMP in R&D labs, and (3)
Because supportive CMC studies are not subject to the stringency to ensure that regulatory expectations are met for product IND and
of codified cGMP or GLP compliance, including their associated Part BLA dossiers, i.e., that they contain reliable, authentic, accurate, and
11 validation and control elements for computerized systems and traceable CMC data.
electronic data,19 the burden is on the sponsor to create an appropri-
ate quality system for their R&D CMC activities. R&D laboratory quality system framework

Risk-Based quality practices for CMC R&D laboratories In a broad sense, the integrity of all drug development data
throughout the pre-clinical, clinical and commercial development
The aim of this document is to provide an example quality system lifecycle can ultimately impact manufactured drug quality. The tiered
framework that is tailored for supportive CMC R&D studies (examples categories presented in Fig. 1 are intended to capture the relative risk
are provided in Table 1) that are conducted throughout the product to manufactured product quality of any given data point generated in
development lifecycle for any biopharmaceutical product modality. each area/tier. Tier 0 represents R&D screening / scouting studies that
Implementation of this framework could take many forms from an are not intended to be filed in a regulatory dossier and require the
organizational hierarchy perspective. Since R&D quality practices are least amount of quality system controls. cGMP Quality laboratory
neither codified nor consistently taught as a part of scientific training, data is intended to be on the front lines of quality control for each
R&D CMC departments should have a sound understanding of the manufactured batch, and there is a high risk of product quality issues
quality expectations applied to their studies. Without assuring each if strict cGMP controls are not followed.
person understands the rationale and justification behind the opera- Tiers 1 and 2 (detailed in Table 2, and the primary focus of this
tional practices, an R&D Quality System framework might simply paper) are proposed for non-cGMP CMC R&D studies designed to col-
exist on paper. While CMC development lab activities should not fall lect Supportive Information. R&D data from both Tier 1 and Tier 2
under formal Quality Assurance oversight, sources trained in formal may be submitted to regulatory agencies, but at different points in
QA practices can provide valuable advice in implementation and development, for different purposes, or to support specific change
maintenance of R&D CMC quality practices. requests. Many of the R&D CMC studies conducted in early develop-
In order to reflect the risk-based nature of R&D data generated for ment to support clinical trials are updated and confirmed in later
biological product development and commercialization, we have development, e.g. for PPQ studies. After commercial approval (BLA/
devised a tiered R&D quality risk management framework (Fig. 1). MAA), additional R&D work may be conducted for preliminary
The categories in Fig. 1 are based on when and how the data are used assessment of proposed post-approval changes (PAC), which (if going
in regulatory filings. This paper will focus exclusively on the R&D forward) would be confirmed under a PAC protocol.
studies defined in Tier’s 1 and 2, i.e. those identified as [In Scope] This paper provides a non-exhaustive list of examples to illustrate
below and in Fig. 1. how the framework of a Tier1/Tier 2 R&D Quality Risk Management
3126 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

Table 1
Examples of CMC R&D development activities.

CMC Support Elements R&D Study Types / Data Applications

Elucidation of Structure and Function, Identification and Identification and characterization of product variants (product-related substances and impurities)
Characterization of Impurities, Product-related Quality Identification and characterization of non-product-related impurities
Attribute Determination Characterization of product functional attributes
Product structure-function studies
Identification and characterization of product degradants
Identification of specified process residuals (e.g. HMW DNA, key HCPs)
Product Forced Degradation Studies Establishing product stability profile for potential degradation pathways
Used in validation of QC stability-indicating methods (e.g. specificity)
Photodegradation studies to support methods used photoconfirmatory studies
Used to compare pre and post change product degradant patterns
Used to compare method sensitivity and specificity in stability method bridging studies
Qualitative assessment of degradant patterns of reference product to biosimilar (CAA study)
Establishment, Stability, and Bridging of Product Reference Establishment and bridging of initial and interim product reference standards
Standards Stability monitoring of initial and interim product reference standards
Establishment of designated product or process impurity standards (e.g. process-specific HCPs)
Calibration and bridging of relative potency product reference standards
Establishment of primary and working product reference standards
Qualification of new working product reference standards
Bridging of designated product or process impurity standards (e.g. process-specific HCPs)
Analytical Method Development and Method Qualification Fit-for-purpose assessment of qualitative R&D analytical methods
QbD studies to establish QC method performance design space
DOE / Optimization studies to establish QC method procedural robustness
Qualification of R&D methods used in CMC study protocols (e.g. PPQ, comparability protocols)
Qualification of R&D methods used in biosimilar product Comparative Analytical Assessment (CAA) studies
Upstream and Downstream Process Development Characterization of production MCB/WCB or MVB/WVB
QbD / DOE Design Space Studies
Qualification of small-scale models (SSMs) of unit operations
Limit of In Vitro Cell Age (LIVCA) studies
Assessment of upstream culture conditions
Assessment of PARs and NORs for unit operations
Downstream process clearance and viral clearance studies
Column cleaning and lifetime studies
Process hold-point stability studies
Bulk DS freeze/thaw studies
Development of Product Formulation and Presentation Assessment of formulation excipients
Formulation QbD / DOE / Design space studies
Confirmation of formulation composition and concentration ranges
Assessment of container/closure systems
Drug Product Process Development Compatibility studies
In-use stability studies
Photostability confirmatory studies
Shipping stability studies
In-use microbiological challenge studies
Identify potential extractable compounds from container/closure system
Verify/validate R&D methods for detecting extractable compounds
Conduct of real-time leachable studies of product matrix in C/C system
Analysis of leachable study test samples for presence of extracted compounds
Comparability Studies Comparative assessment of pre and post change materials for:
- Changes in process scale
- Manufacturing site changes/additions
- Changes in process unit operations
- Product formulation changes
- Changes in product formulation/filling operation
- Changes in product container/closure system

Biosimilar Product RLD and CAA Studies Characterization of physiochemical and functional attributes of Reference Licensed Product(s) (RLDs)
Comparative analytical assessment (CAA) of RLD lots to Biosimilar Product lots
Supportive investigations for cGMP OOT/OOS events Generate hypothesis for potential assignable cause(s)
Conduct investigative experiments to confirm/refute hypothesis
Generate orthogonal corroborative data to support conclusions
Assess impact of corrective actions

program might be implemented. Specific operational points to con- ○ Attributable, Legible, Contemporaneous, Original or True Copy,
sider include: and Accurate
○ Record keeping that is designed to be Complete, Consistent,
- Control of raw materials/reagents used in R&D studies Enduring, and Available
- Chain of custody of R&D samples
- Management of equipment (including controlled temperature Control of raw materials and reagents
units, CTUs)
- Suitable performance of instrumentation at time of use Tiers 1 and 2: Raw materials (reagents, chromatography resins,
- Documentation/data integrity using the principles of ALCOA+ reference materials, etc.) need to have acceptable quality (usually
 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136 3127

Fig. 1. Risk-ranked categories of CMC data generated thorough out product development and commercialization.

compendial grade or better), or as the saying goes, “garbage in, gar- temperature. Confirmatory temperature checks should be done on at
bage out!” Use of compendial grade materials in early development least a weekly basis and recorded in a log. To capture temperature
will help mitigate quality questions as development proceeds. A risk- excursions that may have occurred between checks, it is recom-
based system should be in place to help assure the quality of raw mended to include a min/max digital thermometer in each CTU. At
materials and reagents. A risk-based system could include, for exam- each temperature check, document the observed temperature range
ple, reliance on supplier Certificates of Analyses, testing upon receipt, for the monitoring period and reset them for the next period. CTUs
and/or stability testing. used for specified applications (e.g. sample incubation steps, forced
All labs should label raw material/reagents with expiry dates if degradation studies) should also be monitored during the execution
not provided by the manufacturer. Raw materials that are expired or of the study (e.g. with min/max thermometers) to confirm how con-
no longer of suitable quality for their intended purpose should be dis- sistently the intended temperatures were met. Note that min/max
carded. If no expiry is provided by the manufacturer, or for in-house thermometers only provide the ranges; they do not capture the fre-
prepared reagents and critical reagents, expiration dates on materials quency or duration of excursions. The potential impact of tempera-
can be derived or extended either by scientific rationalization of the tures exceeding the expected range for the storage or study
expected stability of the material (i.e. using the best judgement of the conditions should be described and justified in the relevant Tier 1
scientist preparing the material) or through empirical studies. R&D data sets.
In addition to expiry dates on unopened containers, reagents Tier 2: CTUs used for storage and handling of test samples, reference
should be labelled with the date at which the bottle was opened. This standards, assay controls, and critical reagents or materials should also
is particularly important for unstable reagents (e.g. polysorbates). be monitored to assure they maintain the correct temperature continu-
Operators should record in their laboratory notebooks the lot num- ous control of conditions. This includes the storage and handling of ref-
bers and expiry for all raw materials utilized in the procedures, as erence licensed product for biosimilar CAA studies. However, the
well as lot numbers for critical consumables such as chromatography monitoring systems of Tier 2 CTUs should capture the range, frequency,
resins, columns, etc. If expired reagents or materials are used in a key and duration of temperature excursions, not just the overall min/max.
R&D study (e.g. for FMEA), the justification and any supporting stud- With scientific justification, some temperature-controlled Tier 2
ies should be documented. reagents and materials may be stored in Tier 1 level CTUs.
Additionally, shipments of highly critical materials and samples
(e.g., samples of the reference product for biosimilar CAA studies) for
Sample chain of custody
Tier 2 studies should include temperature monitoring devices that
are documented with each shipment. The potential impact of tem-
Tiers 1 and 2: Samples that are transferred between labs should be
peratures excursions (including range, frequency, and duration)
handled with standard chain of custody (CoC) procedures. Samples
should be described and justified in the relevant Tier 2 R&D data sets.
should have unique or distinguishing names, and receiving labs
should record the information necessary to maintain the chain of cus-
tody (e.g., the date of receipt and the condition and identity of sam-
R&D operator/analyst training
ples received) and location of storage.
Tier 1: Given the early phase of development and/or the R&D
Controlled temperature units nature of the Tier 1 work, it is acceptable for lab management to
determine if operators are adequately trained. Documented training
Tier 1: Controlled temperature units (CTUs) used for storing test records are unnecessary. For certain procedures, method controls or
samples, reference standards, assay controls, and critical reagents or system suitability measures may serve as evidence of suitable train-
materials should be monitored to assure they maintain the correct ing by confirming correct assay execution.
3128 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

Table 2
Examples of tier 1 and tier 2 R&D laboratory quality practices.

Category Tier 1 Tier 2

Raw materials, controlled materials Use USP/EP/ACS grade (where available). Follow Same as Tier 1
and reagents manufacturer’s storage and expiry. Discard
expired commercial materials unless justified
by the study (e.g. FMEA). Label containers
with date of opening. Use scientific judge-
ment and/or empirical studies for expiry of
in-house prepared reagents and controls.
Sample CoC Record material receipt, condition, and holding Same as Tier 1
location.
CTU control Monitor CTU temperatures of test samples and Monitor CTUs with critical test materials, stand-
high-risk reagents, controls, and materials. ards, and controls and high-risk materials and
Use devices that capture min/max temps in reagents with devices that capture range, fre-
monitoring periods quency and duration of temperature excur-
sions; use data loggers in shipments of highly
critical samples such as biosimilar reference
products and product reference standards
Operator/Analyst training Supervisor deems adequacy of technical and Same as Tier 1; evidence of suitable conduct of
quality training. System suitability measures study protocols and documentation practices
may be used to confirm correct assay execu- is expected. Consistent performance gaps
tion. should be addressed by lab management
Qualification and maintenance of In-house testing can be created to ensure equip- Tier 1 for many Tier 2 studies, but for high-risk
instruments and equipment ment is properly installed and satisfies end Tier 2 studies (e.g. CAA and PRS), the opera-
user(s) intentions. System suitability should tional status of R&D instruments should be
be conducted to verify equipment is operat- justified by a risk assessment of the purpose
ing as needed to produce reliable results. Sys- of the analysis (e.g. type of product attribute),
tems requiring calibration should be the design of the test (e.g. stand-alone vs side-
calibrated with the appropriate frequency by-side samples), and nature of the measure-
and documented. PMs should be performed ment (e.g. qualitative, relative quantitation, or
as needed to maintain system performance, empirically quantitative)
and repairs should be performed to remediate
performance issues.
ALCOA Attributable Lab notebook should contain name of analyst, Same as Tier 1, with peer review signatures and
equipment identifier(s), data collection and periodic review to verify satisfactory docu-
analysis software (and version). mentation by all analysts and reviewers
Audit trail desirable, but not mandatory Maintain an audit trail (electronic or paper),
with periodic review.to assure audit trail sys-
tems are functioning correctly
Admin-like system logins are acceptable for Same as Tier 1
shared instruments and software only if elec-
tronic data files generated with each use are
uniquely identifiable to the analyst executing
the work and locations of complete data sets
from the run are documented in individual
notebooks.
Legible Handwritten notes and written notebook Same as Tier 1
entries must be legible. Corrections must
retain legibility of original information.
Contemporaneous Lab notebook entries should be recorded as Same as Tier 1; peer reviews should be con-
close to the time of experimental execution as ducted in a similarly timely manner to allow
practical. for contemporaneous correction of errors or
omissions.
Original or true copy of data retained Data, metadata, and supplementary electronic Tier 1 plus: Minimum of monthly data backup
file(s) should be archived on a secure backup
server (deletions / overwrites of raw data are
not allowed). Minimum of quarterly data
backup. Paper copies (where no electronic
record is captured) should be printed and
retained.
Accuracy (match to Lab notebooks and reports Reviewed for scientific approach. If 2nd signa- Tier 1 plus: Data Verify (DV) results in note-
true value) ture is deemed necessary, sign as “read and books, reports as “read and verified” (repro-
understood”. Reports should be verified back cess data as needed for 100 % DV). the
to the laboratory notebook record using spot laboratory notebook (paper or electronic)
checking as appropriate, but exhaustive data record is verified to the source if different
verification is not required. from the laboratory notebook. All data within
reports are verified back to original laboratory
notebook records
Data collection software Utilize equipment vendor software, as appropri- Tier 1 plus: Engineering and/or procedural con-
ate. trols should be implemented in either the
Data analysis software Use scientific judgement to deem analysis soft- asset’s Operational SOP or the applicable
ware as fit-for-purpose. Ensure either auto- product-specific SOP (e.g., method procedure)
mated or manual data transformations are to ensure the equipment is operated correctly
appropriate. and data analysis has the appropriate

(continued)
 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136 3129

Table 2 (Continued)

Category Tier 1 Tier 2

accuracy; name and version of the software

application used to collect/process the data
should be documented (e.g., should future re-
processing be required). 21CFR11 is not
required, but relevant elements may be
adapted as deemed appropriate.
+ Complete, consistent, enduring, available Store 2° data files and/or processing parameters Same as Tier 1
(for dynamic records) needed to reproduce
results on a backup drive and/or as attach-
ments to ELN. Minimum of 30-day backup
frequency.
Ensure experiment design and data support the Same as Tier 1; confirm that all experiments
complete study. The lab notebook should defined in study protocol are present with
record the procedure in sufficient detail to complete data sets, that any protocol devia-
recreate the experiment. Printouts scanned to tions, additional experiments (i.e., repeats of
digital format. failures or invalids), or omitted data are docu-
mented and explained.

Tier 2: Operator training for Tier 2 related procedures may be for- quantitative assays for content/activity. Analysts should be trained
mally documented in a training record or may be based on docu- on factors that might impact the suitability of the pipette (e.g. drop-
mented experience (e.g., in a person’s curriculum vitae) to establish ping it!) and visual cues that may indicate if a device is out of calibra-
their technical qualifications to run the procedures. Supervisors tion.
should assess operator understanding of R&D quality systems by Tier 2: The key to effective protection against inaccurate measure-
their satisfactory conduct with study protocols and documentation ments is to recognize the potential impact of using poorly performing
practices. Reassessing an individual’s training qualifications would be equipment or out-of-tolerance instruments in critical R&D studies
at the discretion of the lab management. and adopt practical measures that can help the laboratory minimize
the risk of generating erroneous data upon which critical CMC deci-
Qualification/maintenance of instruments and equipment sions could be made. However, a higher burden of confidence in
instrument performance and measurement accuracy is generally
Tiers 1 and 2: Analytical instruments should be installed following expected in Tier 2 studies such as biosimilar CAA studies or for estab-
the manufacturer’s recommendations, and in-house or vendor-per- lishing Primary Reference Standards. For high-risk studies, the degree
formed testing using appropriate control samples may be utilized to of rigor used to confirm the operational status of R&D instruments
ensure they are properly installed and functioning as intended. For- should be justified by a risk assessment of the purpose of the analysis
mal installation and operational qualification (I/OQ) of R&D instru- (e.g. type of product attribute), the design of the test (e.g. stand-alone
mentation may be performed but are not necessarily required vs side-by-side samples), and nature of the measurement (e.g. quali-
especially when appropriate system suitability checks are conducted tative, relative quantitation, or empirically quantitative).
during routine analysis. For example, highly critical quantitative product attributes (i.e.,
System suitability controls (appropriate positive and negative calibration of potency, quantitation of sequence variants) should be
control samples) should be utilized, and corresponding system suit- tested with instruments and equipment with more rigorous qualifi-
ability results should be recorded, to demonstrate the instrumenta- cation/maintenance programs. On the other hand, for comparative
tion is operating appropriately at the time of use. For example, we qualitative measurements of samples tested side-by-side (e.g., Circu-
recommend workflow-based performance measures and the use of lar Dichroism spectroscopy, flow cytometry) or for quantitative prod-
known reference materials and buffer-blanks to demonstrate the uct attributes that are tested with multiple orthogonal methods (e.g.
suitable performance of a chromatographic or electrophoretic sepa- purity/degradants by chromatography and electrophoresis), the
ration method. assurance of accurate, reliable instrument performance may be more
Performance qualification (PQ) may be evidenced by the use of meaningfully demonstrated in each run with the use of sensitive and
appropriately sensitive and specific system suitability measures as specific suitability measures that are tailored to the product and the
established for each type of instrument and workflow. Preventative method, such as assay controls and reference standards.
maintenance (PM) activities should be done as needed to prevent or
remediate system performance problems. Attributable, legible, contemporaneous, original or true copy of data
Systems that require calibration to ensure accurate results should
have periodic calibration checks on a timely basis using appropriate Tier 1: There are several ALCOA+ principles that should be
controls, with documentation of their calibration status (e.g. in or out employed regardless of lab tier for raw data storage. Data should be
of tolerance). For example, mass spectrometers should be calibrated attributable to a user, where a lab notebook (paper or electronic) or
with appropriate mass standards at a suitable frequency based on similarly capable documentation system is used to identify the ana-
vendor recommendations or available data. pH meters should be cali- lyst who performed the work, unique identification of samples,
brated during the day in which they are used with a 3-point pH cali- unique identification of the equipment used to collect the data, and
bration set. Gravimetric systems should be calibrated with the data collection and analysis software names and versions.
appropriate frequency or at time of use with a NIST certified weight Handwritten notes and sketches must be legible and recorded in a
set when quantitative assay results depend on accurate weights. Vol- lab notebook as close to the time of experimental execution as practi-
umetric systems should have a calibration or verification program cal. Labs should consider employing an electronic lab notebook sys-
based on intended use. Pipettes used for volumetric transfers may tem, one that requires an individualized and unique login for each
have less stringent calibration requirements than those used for user, and ideally one that adheres to 21CFR part 11 requirements
3130 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

since these types of lab notebook systems are now widely available software may be sufficient, if they adequately assure data integrity.
and relatively affordable. This especially pertains to biosimilar devel- For example, some R&D instrument systems and data analysis soft-
opment that relies heavily on analytical similarity data stemming ware applications may be utilized by several analysts in different
from CMC development labs. Labs that utilize paper lab notebooks studies. For shared instruments, it may be acceptable to have a single
should consider that the paper record may be designated as “source lab-level login name and password for access to the system, analo-
data” and employ additional controls such as a witness signature to gous to having a single administrator login name and password. This
ensure data integrity. is especially efficient in cases where multiple users may need to
Tier 2: Includes all of Tier 1, plus electronic or paper audit trails for access the same instrument throughout the day and getting locked
data, metadata and supplementary electronic files can be imple- out by another user can impede workflows. Furthermore, vendor
mented so that results can be traced back to the source data, equip- software licenses (when approved for multiple users) may be
ment operator and date/time of use. designed towards single logins. Similarly, some equipment such as
pH meters, benchtop UV−Vis instruments, etc., may have attached
Data and documentation accuracy screen interfaces and software that do not accommodate user logins.
However, in cases like these, the identity of the analyst using the sys-
Tiers 1 and 2: Data integrity practices should also ensure that data tem for Tier 1 or Tier 2 studies must then be captured procedurally in
is complete, consistent, enduring, and available. Data and metadata the metadata associated with each use. Regardless of the administra-
generated by data collection and analysis software should be stored tive or procedural control strategy for shared R&D instruments and
in a way that allows re-creation of results. Lab notebooks should be software applications, identifying details of each login and use must
reviewed to ensure that the experimental design is sound, and that be captured in the documentation system (e.g. electronic or paper lab
the recorded data supports the design. Poor data accuracy, even dur- notebooks) that can tie the electronic data location, metadata and
ing Tier 1 activities, incurs significant risk to a company preserving results to the analyst who performed the work.
intellectual property. In Tier 2 work, poor data accuracy likely poses a After data are collected, the raw data files often need to be proc-
more serious risk to product quality and/or potential product approv- essed further to enable more accurate and precise results. For exam-
ability than during Tier 1. ple, chromatography and spectroscopy data are often truncated and/
While CMC development lab activities do not usually fall under or baseline corrected to remove unnecessary datapoints and system-
formal Quality Assurance oversight, a final confirmation of data accu- atic artifacts from the data. It is then further processed and/or trans-
racy may utilize data verification (DV) by a person independent from formed by suitable data analysis software, either provided by the
the individual who conducted the activity. DV involves checking equipment vendor, 3rd party analysis software, or in-house custom
reported quantitative results back to source reports, notebooks, or in code. Source (raw) data are often subjected to modification by inter-
some cases back to the source data by a second investigator recreat- polation and/or rounding when exported by the equipment software
ing the original data analysis on the original (or true copy) of the and/or imported by 3rd party analytical software.
dataset. When performing data transformations, it is very important to
We should note that even if lab-level DV has been conducted at avoid deleting or manipulating the R&D data in a way that creates
the time of Tier 1 and Tier 2 R&D studies, it may be months or years bias in interpretation of data, even unintentionally. All manipulations
before the studies are formally reported in regulatory submissions. In should be documented in a manner that allows objective, indepen-
order to assure confidence, especially of Tier 2 data, some companies dent assessment of the complete results.
may elect to review and confirm DV activities at the time of filing. Examples of generally acceptable R&D data transformations (par-
Such a look-back verification right before submitting critical R&D ticularly x, y datasets) include:
data may be particularly important where the historical data were
generated in multiple labs with disparate R&D quality practices, or at  Truncating flow-through and wash sections of chromatography
a time before a given lab had implemented them. data
Tier 1 Data Verification: Reviewed for scientific approach by a sec-  Interpolation, extrapolation, and smoothing that enhances data
ond scientist. A second scientist signs all notebooks, signifying that analysis without masking critical data
the experiment was “read and understood”. Reports are verified back  x- and y-axis shifting to align peaks and enable more direct data
to the laboratory notebook record using spot checking as appropriate, visualization and comparison
but an exhaustive data verification is not required.  x-, y-data normalization, i.e. subtracting an underlying baseline
Tier 2 Data Verification: A second person data verifies all results in slope or curvature with an appropriate algorithm to enable
notebooks and reports as “read and verified”. Some source data can improved visual / chemometric comparisons to a reference mate-
be reprocessed as needed to spot-check data analysis accuracy. The rial without masking relevant peaks.
laboratory notebook (paper or electronic) record is verified to the
source if different from the laboratory notebook. All data within Generally Forbidden R&D Data Transformations Include:
reports are verified back to the laboratory notebook record. Further,
the transcription from reports into submissions must be verified as  Fabricating, manipulating, falsifying, or deleting data in order to
well. There needs to be 100 % confidence that a third-party reviewer pass a specification or acceptance criterion
or inspector can go from submission to report to notebook with no  Intentionally misleading the interpretation of results
concern for data integrity. Peer review must be documented and veri-  Scaling data to mask potentially critical attributes or impurities
fied.  Adjusting system inputs or calibration constants to achieve a
desired result
Accuracy and integrity of electronic data collection and data analysis  Masking or eliminating the original results

Tiers 1 and 2: Accuracy of recorded raw data from equipment Data analysis software can either be equipment vendor provided,
hardware is typically dependent on data collection software provided 3rd party software (e.g., Excel, Sedfit, etc.), or in-house developed
by the equipment vendor. Although data collection and analysis soft- scripts (VBA, Python, Matlab, etc.).
ware does not require statutory 21CFR part 11 compliance for these Tier 1 Audit Trails, Data Archival and Operating Procedures: Audit
R&D activities, lab-defined risk-based procedural controls of R&D trails for data collection and analysis are unnecessary at this stage.
 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136 3131

Data archival is recommended at a quarterly frequency. Procedures times across different platforms or products, we suggest confirma-
for equipment operation are flexible, do not need to be documented tory data from studies such as these be generated under Tier 2.
prior to operation, but should be documented after execution in a lab
notebook sufficiently to recreate the procedure later. Elucidation of structure and function, identification and characterization
Tier 2 Audit Trails, Data Archival and Operating Procedures: Audit of impurities, product-related quality attribute determination
trails for data recorded in an electronic lab notebook are recom-
mended. Data, metadata, and supplementary electronic files can be Elucidation of a biological product’s structure and function is a
backed up with a greater frequency (minimum of monthly). If data part of CMC development as outlined in ICH Q6B23 and is submitted
transformations are applied, a description of the transformation to regulatory authorities in the Investigational New Drug (IND) dos-
should be documented, both the original and transformed data sier, IND amendments, marketing applications (BLA/MAA), and for
should be stored (and significant intermediate transformations), some post-approval changes. By late-stage development, biological
along with visual representations of pre- and post-transformed data product degradants that are observed in ICH shelf-life stability stud-
where applicable. ies of drug substance and drug product are also expected to be ana-
For example, when processing a gel image, an operator may run it lytically identified and characterized. Though they are sometimes
through appropriate software to enhance contrast, then again use the more molecularly complex than purified proteins, gene therapy and
same or a different software to compare band densities. In this case, cellular therapy products are expected to have an appropriate degree
the original gel image, the enhanced gel image, and the band density of characterization of structure and function. These characterization
quantitation should be stored. Engineering and/or procedural con- activities form the basis for identification of CQAs.
trols should be implemented in either the equipment’s operational Elucidating biochemical and biophysical structures of proteins,
SOP or the applicable product-specific SOP (e.g., method procedure) glycans, oligonucleotides, cell-based products and other biological
to ensure the equipment is operated correctly and data analysis has entities is technically challenging, requiring dozens of analytical tech-
the necessary accuracy. Any modifications to an operational proce- niques. Many characterization methods used for structural elucida-
dure should be documented in a lab notebook and applicable techni- tion and characterization of product degradants are cutting-edge
cal reports. R&D techniques such as advanced mass spectrometry and spectros-
As another example, peptide mapping-mass spectrometry data copy techniques that are run by highly specialized labs. Advanced
often exceeds 1 Gb for a single sample run on a high-resolution therapies often use electron microscopy or other image-based physi-
instrument. Data utilized to verify a particular peptide sequence may cal analysis systems to assess structural elements.
include primary (unfragmented) mass, fragmented masses (e.g. b- Methods used for characterization of functional attributes are also
and y-ions) and a variety of charge states (m/z values) for each. A sig- highly specialized, with advanced technologies such as surface plas-
nificant amount of subject matter expertise, and in some cases cus- mon resonance, biolayer interferometry, electrochemiluminescence.
tom software, is utilized in this type of data processing. Requiring a Even traditional methods of functional analysis such as cell-based
second analyst to independently verify every data point in this type bioassays are typically used in specialized R&D studies such as char-
of analysis isn’t feasible. However, key data should be documented in acterization of structure/function relationships.
a technical summary report, and for Tier 2, subsequently data veri- Also, R&D test samples might require specific types of processing
fied. for the analytical study. For example, fraction collection of peaks
from analytical chromatography columns may be performed to iden-
Assigning R&D development activities to a quality risk tier tify product related substances and impurities, or to conduct struc-
ture/function testing. While it usually produces a reliable indicator of
Realizing that the application of a tiered system to CMC R&D activ- structural identity, analytical fraction collection is fraught with the
ities may still be somewhat ambiguous, the following examples are possibility of creating artifacts due to sample manipulations. There-
provided. The information is not intended to be a prescriptive set of fore, it is important that R&D sample preparation steps are thor-
requirements, but instead should be used for guidance of how to oughly documented, including which samples and controls were
apply the tiered system to other activities as well. utilized and how they were prepared, collected, and stored.
The concept of phase-appropriate quality systems is an inherent Analytical sample preparation is often critical to R&D studies that
element of cGMP for investigative medical products, and similarly are conducted for the identification and characterization of product
there can be a ‘phase-appropriate’ consideration for some R&D CMC degradants. Characterization studies conducted on the morphological
studies. The recommended degrees of R&D quality system rigor we and compositional aspects of product particulates (e.g. subvisible or
present below are based on the level of risk that potentially inaccu- visible particles) may require careful handling of samples to avoid
rate supportive information could pose to the overall product devel- introducing artifacts by creating, or disrupting, agglomerated species.
opment decisions made at different stages of product development. On the other hand, forced-degradation treatments may be employed
When supporting some cGMP activities (for example comparability to purposefully generate potential degradants for identification and
studies, analytical method bridging studies, etc.), it may be necessary characterization. Either way, quality practices and sound documenta-
to perform studies on materials that are tied directly to the clinic. tion of the sample handling and preparation can become a rich source
This elevates the need for well-designed studies that fall within a of prior knowledge that could be leveraged in future studies such as
quality framework to reduce the possibility of generating data that process-change comparability, formulation changes, and bridging of
could negatively impact ongoing clinical trials. QC methods.
Often, multiple orthogonal methods are utilized to identify prod-
R&D CMC studies that are always Tier 2 uct structures and degradants. For example, primary sequence is
often determined by a combination of confirmation of the applicable
Some R&D CMC studies generate critical regulatory data regard- gene sequence in the biologic entity, intact mass measurements on
less of development phase, such as Comparative Analytical Assess- the drug substance, peptide mapping with UV detection at multiple
ments for biosimilar products, establishment of Product Reference wavelengths and with mass spectrometric (MS) detection for both
Standards, establishment of platform process unit operations, or the intact peptides (MS1) and peptides fragmented with a collision
qualification of platform analytical methods.26 To mitigate the risk of gas or other energy source (MS2). In some cases, datasets may be
inadequate data integrity, especially when leveraging data multiple only partially complete due to technical limitations (e.g., full
3132 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

sequence coverage with MS2 fragmentation may not always be possi- long-term use of a one-tier system may result in undesired drift of
ble or practical). A totality of evidence approach, where information the reference standard lots.
or partial information from multiple methods is assessed to support a To mitigate this risk with commercialized biological products, it is
reasonable conclusion as to the product’s structure and function, or expected that a two-tier system of primary and working product ref-
the identification and characterization of degradants and other impu- erence standards will be established. “Two-tier” means one product
rities, may be the objective of a series of coordinated orthogonal R&D batch is qualified as the primary/master reference standard, and a dif-
analytical studies. ferent product batch is qualified against the primary to serve as the
In most cases, studies for the elucidation of structure and function, secondary/working reference standard. Working reference standard
or characterization of impurities, are qualitative in nature. They are batches are directly utilized in QC testing, whereas the primary refer-
designed to answer the biomolecular questions of ‘What” species are ence standard batch is only used to qualify new lots of working refer-
structurally present, and for potency attributes, “How” do they func- ence standard.
tion. But quantitative questions such as “How Much” of these species With either system, new lots may be required if the current refer-
may be present in the process and product, and what (if any) control ence standard lot is seen to be degrading, if the inventory supply is
limits should be employed, are addressed and justified based on running low, or if there are substantial differences in physical or func-
cGMP manufacturing, release, and stability testing. tional product attributes after process changes. The qualification pro-
Given the qualitative nature of R&D CMC characterization activi- tocols for bridging one-tier reference standard lots, establishing the
ties, a Tier 1 data quality system is appropriate for most studies con- primary reference standard lot, and qualifying each new working ref-
ducted to provide supportive product characterization and erence standard lot are key elements of product development.
elucidation of structure data. However, there could be special cases Whether it is a one- or two-tiered system, the characterization, quali-
where highly critical R&D studies would benefit from risk-mitigation fication, bridging and stability of product reference standards require
elements described in Tier 2. numerous R&D physical and functional methods. The pedigree of
product reference standards, i.e. the analytical links (especially
Product forced degradation studies potency) between those reference standard batches used for QC of
clinical trial materials and those intended for use with QC of commer-
Subjecting drug substance or drug product to targeted stress con- cial product, form a key part of biological product regulatory submis-
ditions such as high temperatures, high/low pH, oxidation stress, sions.
light stress, agitation, etc. can reveal potentially relevant degradation Due to the critical nature of in-house product reference standards
pathways and product liabilities, enable qualitative comparison of in the control and consistency of product quality attributes such as
degradation patterns across formulations or manufacturing condi- identity and relative potency, it is recommended that these R&D
tions, and generate material for assessing stability-indicating proper- studies be conducted at the Tier 2 level.
ties of analytical methods. These studies may fall under Tier 1 or Tier
2 depending on the nature of the study. Analytical method development and method qualification
Tier 1: R&D laboratories should follow laboratory Tier 1 recom-
mendations for forced degradation studies utilized for manufactur- Analytical R&D laboratories provide data to demonstrate the capa-
ability assessments, establishing qualitative product stability profiles bilities of a wide variety of testing methods throughout a biopharma-
for potential degradation pathways, R&D photodegradation studies ceutical product’s development lifecycle. Test methods may come
and in preparation of source material used in validation of QC stabil- from compendial sources (either product monographs or general
ity-indicating methods and stability method bridging studies (e.g. methods) or may be developed de novo (non-compendial methods).
specificity). Pre-PPQ comparability studies to compare pre and post Methods that are used for QC product release and stability programs
change product degradant patterns. follow phase-appropriate method verification (if compendial) or vali-
Tier 2: Qualitative assessment of degradant patterns of reference dation (if non-compendial). All QC methods (compendial or non-
product to biosimilar (CAA study) and post-PPQ comparability stud- compendial) must be verified for suitable use in the cGMP laboratory
ies to compare degradant patterns should be conducted using Tier 2 in which they will be used.28
R&D data quality practices. While phase-appropriate method validation and verification are
outside of the scope of the R&D studies presented here, there are
Establishment, stability, and bridging of product reference standards emerging concepts in analytical method validation support and
method design space that leverage data generated in non-cGMP ana-
Numerous types of reference standards and materials are utilized lytical R&D laboratories. ICH Q1429 “Analytical Procedure Develop-
in manufacturing and testing operations. Those associated with com- ment” provides different strategies for the validation of cGMP
pendial methods or compendial product monographs, such as raw methods, stating: “Data gained during the development studies (e.g.,
materials and excipients, are typically obtained from commercial robustness data from a design of experiments (DoE) study) could be
sources who are responsible for their production, calibration, certifi- used as part of the validation data for the related analytical procedure
cation, and stability. Some existing commercial biological products performance characteristics, and studies do not necessarily need to
have certified reference standards for certain attributes, such as iden- be repeated.” In cases where a cGMP method validation package cites
tity or potency of plasma products, vaccines, cytokines, and hor- development studies for supportive data for parameters such as spec-
mones. But even when externally-certified biological product ificity (especially for validation of stability methods) or robustness,
reference standards exist, it is expected that each manufacturer will regulators expect to see the data presented in the method validation
use them to establish in-house (working) reference standards from report.7,28
their own production process.27 Another application of analytical method performance data is in
In early product development, product reference standard lots the qualification of test methods that are used in key R&D CMC study
may be managed in a one-tier system of interim reference standards. packages. Numerous regulatory guidance documents for biological
“One-tier” means the reference standard batch is directly utilized in products use the term R&D ‘method qualification’ or state that R&D
QC testing, and when a new one is required, it is qualified against the ‘methods should be qualified’ for use in designated CMC studies. Crit-
last reference standard batch. This practice is typically employed ical non-cGMP studies that should use suitably qualified methods
over a short-term with limited changes in reference lots because include reference standard characterization and bridging studies,
 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136 3133

enhanced analytics used in PPQ support, late-phase and post- essential datasets for support the proposed process control strategy.
approval analytical comparability studies, and biosimilar product Characterization of production MCB/WCB or MVB/WVB should utilize
comparative analytical assessments (CAAs). In fact, for biosimilar Tier 2 R&D compliance practices due to the direct impact these banks
products, FDA expects the R&D method qualification packages to be have on cGMP production.
submitted along with the method validation packages for cGMP QC
methods.
Development of product formulation and presentation
While there is no published regulatory definition of ‘method qual-
ification’, the practical distinction is that a method qualification study
Preliminary formulation development studies are often per-
establishes method performance capabilities for ICH Q2 parameters
formed in parallel with analytical and process development stud-
such as accuracy, precision, linearity, specificity, LOD/LOQ, etc.. . .
ies, and thus may pre-date the implementation of cGMP methods.
whereas a method validation study confirms these parameters within
Materials used in these studies may be supplied from different
pre-defined acceptance criteria.30 Method qualification data are often
processes, and methods used to analyze samples may change
utilized to help set method validation acceptance criteria by demon-
from one study to the next. To be able to make reliable conclu-
strating inherent performance range of ICH Q2 parameters. Method
sions from any early phase formulation development study, it is
qualification data are also valuable in establishing the minimum
necessary to document lot numbers of materials and reagents,
acceptance criteria ranges for quantitative attributes in comparability
SOP numbers, or in the absence of SOP’s, the details of the analyt-
studies and CAAs by showing the degree of precision from replicate
ical method should be captured within the formulation develop-
measurements of the same material.
ment protocol and/or report.
Tier 1: Analytical R&D laboratories should follow laboratory Tier 1
Early formulation development studies are performed in stages,
recommendations for method performance studies that support early
starting with pre-formulation studies which screen many conditions,
phase process and product characterization and comparability stud-
followed by optimization studies around potentially promising for-
ies.
mulations, and finally longer-term storage studies on the lead candi-
Tier 2: Analytical R&D laboratories should follow laboratory Tier 2
dates. The majority of methods employed will align with those
recommendations for protocol-driven method qualification studies
intended for product release and stability but may also include
that support process comparability studies, and CAA studies. As these
extended characterization methods for more complete screening and
will be protocol-driven studies, deviations should be documented in
evaluation.
the reports and failures to meet acceptance criteria should be investi-
Because formulation development studies are comparative
gated with any retests properly justified. Also, for studies that gener-
studies, often with the samples for a given timepoint tested
ate R&D method performance data intended to support cGMP
within the same analytical run, these studies are a little more for-
method validation packages. Use of Standard Operating Procedures
giving with respect to method and instrument qualification, and
for conducting methods may also be important for certain critical
the implementation of system suitability and other analytical
studies, such as CAA studies, which form an important bases for biosi-
controls. As such, the preliminary formulation development stud-
milar applications.
ies (pre-formulation and optimization) will likely fall into the cat-
egory of Tier 1, though the methods should still be confirmed as
Upstream/downstream process development
“fit for purpose”. As analytical methods are locked and phase
appropriately qualified, which hopefully will occur prior to the
Upstream and downstream process development comprises a
initiation of longer-term lead candidate studies, the activities will
wide range of activities throughout the product development life-
fall more into the realm of Tier 2.
cycle. Preclinical process development is typically minimal in scope,
Formulation development does not necessarily stop prior to Phase
with the goal of producing a relatively pure and safe drug substance
1 clinical trials. Clinical dosing and other factors may require changes
to test for biological effect(s) in-vitro and in-vivo animal models. As
to formulation concentration, route of administration, or the imple-
drug development proceeds towards toxicology material production,
mentation of a device to facilitate patient comfort and/or compliance.
processes are scaled up and more formally developed with the goal
The steps are largely the same in these types of studies, starting with
of enabling eventual cGMP production, often employing platform
screening, then optimization, then longer-term studies, and may be
processes where feasible. If pre-clinical studies demonstrate promis-
performed in a development laboratory instead of a full QC/cGMP
ing results, the manufacturing processes are further scaled up and
environment. However, the final confirmatory studies, which will
developed to support cGMP manufacturing of material for use in
likely be documented in a comparability assessment, would require
first-in-human clinical trials.
full cGMP to be applied for release and stability methods. Extended
As a drug proceeds through phase I − phase III clinical trials, pro-
characterization studies would fall into Tier 1 or Tier 2 depending on
cess development activities will typically begin focusing on a more
the phase at which the studies were performed, with later stage stud-
thorough assessment of manufacturing design space operating
ies leaning more toward Tier 2.
parameters and performance parameters following QbD principles to
ensure the DS and DP will meet acceptable limits for identity,
strength (potency), quality, and purity. Drug product process development
Tier 1: Process development laboratories should follow laboratory
Tier 1 recommendations for the development / optimization of These studies generally include drug product handling studies and
upstream and downstream processes, design space studies, process extractable / leachable studies. Drug product handling studies include
hold-point, DS freeze-thaw stability studies, and generally most characterization of drug product compatibility with containers, deliv-
early-stage development activities. ery devices, IV bags / tubing, in-use stability studies, photostability
Tier 2: Late-stage qualification of small-scale models (SSMs) of confirmatory studies, pumpability and shipping stability studies.
unit operations, Limit of In Vitro Cell Age (LIVCA) studies, assessment Extractable and leachable studies identify potential extractable com-
of PARs and NORs for unit operations, downstream process clearance pounds from container/closure system. These studies involve devel-
studies, column cleaning and lifetime studies and studies for Process opment / verification of R&D methods for detecting extractable
Validation (hold time studies, in use studies, extractables and leach- compounds, real-time leachable studies of product matrix in con-
able testing, etc.) all should follow Tier 2 practices as they are tainer closure system.
3134 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

Tier 1: pre-PPQ (including IND stage) drug product handling determinations of efficacy from clinical data) and should be collected
studies with an appropriate quality system
Tier 2: Post-PPQ drug product handling studies and in-use micro- The R&D methods used for comparative analytical assessments
biological challenge studies. are expected to be qualified to demonstrate their technical suitability
for the intended measurements of the different products. And as
with process change comparability studies, method qualification data
Comparability studies
may support part of the justification of comparative acceptance crite-
ria for the measured attribute. Due to the critical role of method per-
Analytical comparability studies are a major element of product
formance capabilities when comparing two different products, CAA
lifecycle management, whereby the impact of changes to
method qualification packages are subject to regulatory review and
manufacturing processes for drug substance or drug product are
inspection. Data integrity is highly scrutinized, including the pres-
assessed based on the measurable analytical differences in physical
ence or absence of raw data that should be there, and the fidelity of
and functional characteristics. If substantial analytical differences are
the conclusions to the complete data set of each experiment. Regula-
observed in test results, it could trigger the need for in vivo studies to
tors are particularly concerned about “cherry-picking” of selected
adjudicate the impact on product safety or efficacy data that were
R&D data from selected lots of biosimilar or reference products that
generated with drug product lots made in earlier versions of the pro-
are out-of-trend with the body of results.
cess.
In addition to the quality practices associated with the reagents,
Therefore, a key part of analytical comparability study protocols is
instruments, method qualification, study documentation and raw
to establish appropriate acceptance criteria for each product attribute
data sets, there are high regulatory expectations for the quality prac-
assessed. One of the first elements needed to justify quantitative
tices associated with the incurred batches of reference licensed prod-
ranges is the demonstrated degree of precision of each analytical
uct used in the studies. There must be a documented chain of custody
method. At a minimum, comparative acceptance criteria for each
for how all batches of reference product are shipped, received, stored,
method should consider the range of results obtained when testing
handled, prepared, and analyzed by R&D, which inventory manage-
the same material compared to itself.
ment and reconciliation of all batches. While some CAA studies are
Analytical comparability studies typically include historical and de
designed to examine expired or force-degraded preparations of refer-
novo data from QC lot release and stability testing. For QC methods,
ence product, studies without such an objective should assure the
the inherent degree of method intra- and inter-assay precision
results obtained from the reference product batches are a reflection
should be provided in the phase-appropriate method validation stud-
of the quality and stability of the commercial material without arti-
ies. But in addition to QC data from validated method, substantial
facts being subsequently introduced by chain-of-custody operations.
comparative data are generated using R&D characterization methods.
Due to the critical nature of the R&D studies and materials man-
A practical reason for conducting qualification of R&D methods used
agement for biosimilar product comparative analytical assessments,
in comparability studies is to provide data establishing the precision
a Tier 2 data quality system is strongly recommended. An adequate
and (if needed) intermediate precision of each method with the
quality system for CAA studies should also be in place for any ven-
intended sample types. The data can then be leveraged in the justifi-
dors/external contract testing organizations used to collect CAA data,
cation of comparative acceptance criteria from R&D studies.
with appropriate oversight by the Applicant to ensure adequate prac-
The level of cGMP associated with the manufacturing process and
tices are followed. Of note is the considerable regulatory scrutiny on
QC test methods would be commensurate with the phase of develop-
the ALCOA+ elements for CAA study to confirm the continuity and
ment. The scope expected for the design of comparability protocols
completeness of all raw data sets, electronic or paper. And the R&D
(e.g., the slate of analytical tests, number of pre- and post-change
laboratories executing CAA studies should be clearly informed that
batches to be assessed, replication plan, and type, range, and nature
they can be subject to pre-approval inspections as a part of the BLA/
of comparative acceptance criteria) may vary with the extent of the
MAA approval process for the purposes of assuring traceability of test
process change and the point in development at which the change is
materials and data integrity.13
made. For example, when a comparability study is executed during
early development of a biologic, extended characterization methods
may require only qualitative acceptance criteria or reporting of
Support of cGMP non-conformance investigations
results without pre-defined criteria. Whereas when the same
extended characterization methods are utilized for registration-
Lot disposition is normally informed by results generated by a QC
enabling product comparability, quantitative acceptance criteria may
laboratory, which is governed by applicable cGMP regulations. How-
be required.31 R&D data in comparability studies should be supported
ever, in some cases such as the occurrence of an out of specification
with equally sound data quality practices at any phase of product
(OOS) result in QC, or in other non-conformance (NC) related investi-
development. Therefore, we recommend a Tier 2 data quality system
gations, identification of the root cause of the NC may be facilitated
for comparability study data sets that will be provided to regulators,
by utilization of Characterization methods and process development
especially for PPQ stage and beyond.
activities performed in CMC development labs. For example, an
upstream (fermentation) raw material may have failed a specification
Comparative analytical assessments for diosimilars for a color test. As part of a formal investigation, characterization
tests revealed that the color is related to a non-toxic compound and
A plethora of orthogonal analytical data sets are also generated in process development scale-down model tests revealed that the com-
R&D for the structural and functional comparison of a proposed new pound is cleared by the drug substance purification process.
biosimilar product to the specified reference licensed product(s). In this example, it would be prudent to utilize a Tier 2 analytical
Regulators evaluate the comparative data sets to determine the lab data quality system for any quantitative testing for the colored
degree of analytical similarity and inform the nature of the preclinical compound, and a Tier 2 process development data quality system for
and clinical studies needed to adjudicate any observed in vitro differ- running a qualified scale-down purification process. The higher level
ences.32 Therefore, CAA data form an important bases for any biosi- of data quality compliance (over that of Tier 1) would help ensure
milar application, as they allow Applicants to leverage previous data integrity and confidence in the results when deciding on lot dis-
Health Authority determinations for the reference product (i.e., position.
 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136 3135

Table 3
Laboratory operations quality systems: cGMP, GLP, ISO 17,025, and R&D.

Laboratory Quality Practices GLP cGMP ISO R&D

Defined Laboratory Management Responsibilities X X X X

Adequate Facilities/Staff to Execute Workflow X X X X
Organizational Chart with Defined Roles X X X
Demonstrated Personnel Training X X X X
Master Schedule of Studies X − −
Study Director (Technical Lead) X − − X
Study Protocols and Study Reports X − − X
Independent QA Unit Study Audits X − X
Standard Operating Procedures X X X
Qualified Analytical Methods X
Validated Analytical Methods (phase appropriate) − X X
Validated Bioanalytical Methods X − X
Complete Chain of Custody of Test Samples X X X X
Control of Materials/Reagents Used in Workflow X X X X
Analysis of Test Articles in Dosing Solutions X − −
Animal Housing and Care Requirements X − −
Specimen or Sample Retention X X −
System Suitability Measures for Instrument PQ X X X X
Equipment/Instrument IO/Q and Preventative Maintenance X X X
Computer System Operational Traceability X X X X
Vendor and User Software Validation X X
Electronic System Part 11 Compliance X X
Data Integrity Requirements (Raw Data to Reported Results) X X X X
Documentation Change Control X X X
Data Archiving and Record Retention X X X X
Investigation of Deviations or Failures with CAPA X X X
Opinions and Interpretations of Study Results X − X X
Test Results Confirming Pass or Fail of Specifications − X −
Oversight of Vendor Activities (Materials and Instruments) − X X
Quality Oversight of Subcontracted Activities X X X

Existing lab operations quality statutes and voluntary standards implementation of all quality elements in ISO17025, and then under-
going a compliance inspection from a certifying agency, e.g. ANSI.38
International and regional current Good Manufacturing Practice Accreditation status verifies suitable implementation of quality
(cGMPs) statutes and associated product development regulations requirements for the testing competence, impartiality of reporting
establish specific requirements for the manufacturing and quality results, and consistent operation of laboratories. However, the lab
control operations for the production, testing, transport, and storage cannot be accredited if it is not sufficiently compliant with all of the
of clinical and commercial pharmaceutical products. However, many ISO 17,025 requirements, some of which may not be necessary for
critical supportive CMC process and product development activities R&D CMC studies. Compliance with selected relevant elements might
fall outside of the scope of cGMP regulations. And though the name facilitate R&D best practices, but it will not be sufficient for formal
may seem to suggest it, Good Laboratory Practice (GLP) statutes33 in ISO accreditation.
fact do not apply to any CMC activities conducted in R&D laboratories
for product development purposes. GLP applies to non-clinical phar-
macology/toxicology studies to assess product safety.
It is important to note that R&D laboratories claiming compliance Conclusion
to GLP must always satisfy all elements in the statute; there is no
‘sliding scale’ or ‘phase appropriate’ interpretation of GLP. For exam- At all phases of product development and commercialization, crit-
ple, GLP (bioanalytical) methods must be fully validated before they ical R&D-based CMC studies are needed to support product lifecycle
are used to generate any reportable data, and the method validation management decisions. It is vital to ensure that business and regula-
guidances34,35 are very different from those applicable to cGMP tory decisions based on these studies can trust the authenticity of the
methods.28,36 And GLP laboratories are subject to regulatory inspec- work and reliability of the data. With the absence of existing relevant
tion under different policies procedures than for cGMP laboratories. guidance on the nature of quality systems that should be used in
There is a voluntary consensus standard that defines quality prac- CMC R&D labs, it is up to each sponsor to establish its own approach.
tices for testing laboratories in technical industries. ISO/IEC “General Two opposing risks must be considered when setting up quality
requirements for the competence of testing and calibration laborato- systems in CMC development labs. On one hand, not defining suitable
ries”37 is an international reference for testing and calibration labora- quality practices to guide complete and consistent R&D activities are
tories wanting to demonstrate their capacity to deliver reliable a great risk to the ability to trust, or even reproduce, the data.
results. A major focus of ISO17025 is on measurement calibration and On the other, implementing overly restrictive quality systems can
metrology rather than CMC study protocols. Nonetheless there are impose operational burdens that lengthen CMC development time-
many valuable quality elements in ISO 17,025 which could be lines and make development costs prohibitively expensive, leading
adapted for CMC R&D laboratories. Several of them overlap with sim- to fewer products becoming available for patients.
ilar quality elements in cGMP and GLP statues; Table 3 provides a It is our hope that the concepts presented here may help sponsors
comparison of key laboratory quality requirements required by find the right balance of operational mechanisms to ensure the qual-
cGMP, GLP, ISO and R&D. ity of their R&D CMC studies throughout the product lifecycle. Practi-
Some CMC R&D facilities, particularly contract laboratories, have ces that assure consistency in R&D laboratory operations are vital in
elected to become ISO 17,025 accredited. Accreditation requires bringing safe and effective biologically-derived medicines to market.
3136 B.S. Kendrick et al. / Journal of Pharmaceutical Sciences 113 (2024) 3123−3136

Declaration of competing interest 17. WHO. Annex 4: guideline to data integrity; TRS 1033 (October 2021). Accessed
April 16, 2024. https://www.who.int/publications/m/item/annex-4-trs-1033.
18. Pharmaceutical Inspection Convention. Good practices for data management and
The authors declare that they have no known competing financial integrity in regulated GMP/GDP environments PI-041-1 (July 2021). Accessed April
interests or personal relationships that could have appeared to influ- 16, 2024. https://picscheme.org/docview/4234.
ence the work reported in this paper. 19. FDA guidance for industry. Part 11, Electronic records; electronic signatures —
scope and application (August 2003). Accessed April 16, 2024. https://www.fda.
gov/media/75414/download.
20. US code of federal regulations. 21 CFR Part 210, current good manufacturing prac-
References tice in manufacturing processing, packing, or holding of drugs (2024). Accessed
July 5, 2024. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-
1. International conference on harmonization. ICH Q7 good manufacturing practice 210.
for active pharmaceutical ingredients (2000). Accessed May 22, 2018. https://data 21. US code of federal regulations. 21 CFR part 211, current good manufacturing prac-
base.ich.org/sites/default/files/Q7%20Guideline.pdf. tice for finished pharmaceuticals (2024). Accessed July 5, 2024. https://www.ecfr.
2. International conference on harmonization. ICH Q10 pharmaceutical quality sys- gov/current/title-21/chapter-I/subchapter-C/part-211.
tem (2008). Accessed May 22, 2018. https://database.ich.org/sites/default/files/ 22. US code of federal regulations. 21 CFR Part 600, Biological Products; General
Q10%20Guideline.pdf. (2024). Accessed July 5, 2024. https://www.ecfr.gov/current/title-21/chapter-I/sub
3. International conference on harmonization. ICH Q12 technical and regulatory con- chapter-F/part-600.
siderations for pharmaceutical product lifecycle management (2019). Accessed 23. International conference on harmonization. ICH Q6B specifications test procedures
May 22, 2018. https://database.ich.org/sites/default/files/Q12_Guideline_ and acceptance criteria for biotechnological/biological products (2009). Accessed
Step4_2019_1119.pdf. May 22, 2018. https://www.fda.gov/media/71510/download.
4. FDA guidance for industry. Chemistry, manufacturing, and controls changes to an 24. International conference on harmonization. M4Q: the CTD - quality (2001).
approved application: certain biological products (June 2021). Accessed May 29, Accessed May 22, 2018. https://www.fda.gov/media/71581/download.
2024. https://www.fda.gov/media/109615/download. 25. International conference on harmonization. ICH Q8(R2) pharmaceutical develop-
5. FDA guidance for industry. CGMP for phase 1 investigational drugs (July 2008). ment (2009). Accessed May 22, 2018. https://www.fda.gov/media/71535/down
Accessed April 16, 2024. https://www.fda.gov/media/70975/download. load.
6. F.D.A. Guidance for Industry. INDs for phase 2 and phase 3 studies chemistry, 26. Gabrielson JP, Kendrick BS, Young JA. Universal qualification of analytical proce-
manufacturing, and controls information (May 2003). Accessed April 16, 2024. dures for characterization and control of biologics. J Pharm Sci. May 26, 2020.
https://www.fda.gov/media/70822/download. https://doi.org/10.1016/j.xphs.2020.05.012. Published online.
7. Pharmaceutical Inspection Convention. Pharmaceutical inspection co-operation 27. CASSS CMC Strategy Forum. Biological Product-Specific Reference Standard - Best
scheme (PIC/S) GMP particularities in the manufacture of medicinal products to be Practices - Part 1 and 2. August 2013.. August 2013. Accessed May 24, 2024;
used in clinical trials on human subjects, PI-021-2 (September 2007). Accessed https://www.casss.org/papers-and-presentations/resource/cmc-summary-paper-
April 16, 2024. https://picscheme.org/docview/3456. biological-product-specific-reference-standard-best-practices-part-1-and-2.
8. FDA guidance for industry. content and format of investigational new drug appli- 28. FDA guidance for industry. Analytical procedures and methods validation for drugs
cations (INDs) for phase 1 studies of drugs, including well characterized, therapeu- and biologics (July 2015). Accessed April 16, 2024. https://www.fda.gov/media/
tic, biotechnology-derived products (November 1995). Accessed April 16, 2024. 87801/download.
https://www.fda.gov/media/71203/download. 29. International conference on harmonization. ICH Q14 analytical procedure develop-
9. E.M.A. Guideline. Guideline on the requirements for quality documentation con- ment (August 2022). Accessed November 2, 2023. https://www.fda.gov/media/
cerning biological investigational medicinal products in clinical trials (January 161202/download.
2022). Accessed April 16, 2024. https://www.ema.europa.eu/en/documents/scien 30. Ritter N., Advant S.J., Hennessey J., et al. What is test method qualification? BioPro-
tific-guideline/guideline-requirements-quality-documentation-concerning-biologi cess Int. September 2004:2−11.
cal-investigational-medicinal-products-clinical-trials-revision-2_en.pdf. 31. Weiss WF, Gabrielson JP, Al-Azzam W, et al. Technical decision making with higher
10. EMA. Guideline on quality, non-clinical and clinical requirements for investigational order structure data: perspectives on higher order structure characterization from
advanced therapy medicinal products in clinical trials (Draft) (March 2024). Accessed the biopharmaceutical industry. J Pharm Sci. 2016;105(12):3465–3470. https://doi.
April 16, 2024. https://www.ema.europa.eu/en/documents/scientific-guideline/draft- org/10.1016/j.xphs.2016.09.003.
guideline-quality-non-clinical-clinical-requirements-investigational-advanced-ther 32. Center for Drug Evaluation and Research. Development of Therapeutic Protein Biosi-
apy-medicinal-products-clinical-trials-second-version_en.pdf. milars: Comparative Analytical Assessment and Other Quality-Related Considerations
11. Marks, Peter. Statement on data accuracy issues with recently approved gene ther- Guidance For Industry. May 2019.. Accessed May 24, 2024; https://www.fda.gov/
apy (April 2024). Accessed April 15, 2024. https://www.fda.gov/news-events/press- media/159261/download.
announcements/statement-data-accuracy-issues-recently-approved-gene-therapy. 33. Good Laboratory Practices for Nonclinical Studies. 21 CFR Part 58. Accessed April
12. FDA Guidance for Industry. Contract Manufacturing Arrangements For Drugs: Quality 16, 2024. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-A/part-58.
Agreements. November 2016. Accessed April 10, 2024; https://www.fda.gov/ 34. FDA Guidance for Industry. Bioanalytical method validation (May 2018). Accessed
media/86193/download. April 16, 2024. https://www.fda.gov/files/drugs/published/Bioanalytical-Method-
13. FDA. Compliance program 7346.832−preapproval inspections (October 2022). Validation-Guidance-for-Industry.pdf.
Accessed April 12, 2024. https://www.fda.gov/media/121512/download. 35. I.nternational Conference on Harmonization. ICH M10 bioanalytical method vali-
14. Unger B. FDA Data Integrity Enforcement Trends and Practical Mitigation Measures. dation and study sample analysis (May 2022). Accessed April 16, 2024. https://
Food and Drug Law Institute (FDLI); April 2018... Accessed April 16, 2024; https:// www.fda.gov/media/162903/download.
www.fdli.org/2018/04/update-fda-data-integrity-enforcement-trends-practical- 36. International conference on harmonization. ICH Q2(R2) validation of analytical
mitigation-measures/. procedures (November 2023). Accessed April 16, 2024. https://www.fda.gov/
15. FDA guidance for industry. Data integrity and compliance with drug CGMP ques- media/161201/download.
tions and answers (December 2018). Accessed April 16, 2024. https://www.fda. 37. ISO/IEC. 17025:2017 General requirements for the competence of testing and cali-
gov/media/119267/download. bration laboratories (2018). ISO. Accessed April 16, 2024. https://www.iso.org/pub
16. M.H.R.A. GxP Data integrity guidance and definitions (March 2018). Accessed April 16, lication/PUB100424.html.
2024. https://assets.publishing.service.gov.uk/media/5aa2b9ede5274a3e391e37f3/ 38. ANAB/ANSI accreditation program ISO/IEC 17025. Accessed April 16, 2024. https://
MHRA_GxP_data_integrity_guide_March_edited_Final.pdf. anab.ansi.org/standard/iso-iec-17025/.