Volume 38 Number 6, 313-372

Volume 38 Number 6 June 2020

www.chromatographyonline.com
LCGC NORTH AMERICA

LC PRACTICAL APPLICATIONS DATA HANDLING

Stability Studies MULTIDIMENSIONAL Controlling Peak
and Testing of Integration to Ensure
Pharmaceuticals SEPARATIONS FOR Data Integrity
BIOTHERAPEUTIC
GC CHARACTERIZATION THEORY &
Troubleshooting FUNDAMENTALS
Reduced Peak Starting Up Your Laboratory
June 2020

Size in GC After a Pandemic Shutdown

 MACHEREY-NAGEL
Columns and supplies
Chromatography

Your solutions at www.mn-net.com

www.mn-net.com
www.mn-net.com
www.mn-net.com
 What does it take to solve
one of the biggest challenges
in chromatography?

Be one of the first to see

what PREMIER can do for you.

waters.com/PREMIER
316  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

MANUSCRIPTS: For manuscript preparation guidelines, see chromatographyonline. 485F US Highway One South, Digital Production Manager
com/lcgc-author-guidelines, or call The Editor, (732) 596-0276. LCGC welcomes unso- Suite 210 Sabina Advani
licited articles, manuscripts, photographs, illustrations, and other materials but can- Iselin, NJ 08830 SAdvani@mjhlifesciences.com
not be held responsible for their safekeeping or return. Every precaution is taken to (732) 596-0276
Fax: (732) 647-1235 Managing Editor, Special Projects
ensure accuracy, but LCGC cannot accept responsibility for the accuracy of informa- Kaylynn Chiarello-Ebner
tion supplied herein or for any opinion expressed. KEbner@mjhlifesciences.com
PUBLISHING/SALES
SUBSCRIPTIONS: For subscription and circulation information: LCGC, P.O. Box
Senior Vice President,
457, Cranbury, NJ 08512-0457, or email mmhinfo@mmhgroup.com. Delivery of MARKETING/OPERATIONS
Industry Sciences
LCGC outside the United States is 14 days after printing. (LCGC Europe and Michael J. Tessalone Marketing Manager
LCGC Asia Pacific are available free of charge to users and specifiers of chro- MTessalone@mjhlifesciences.com Brianne Pangaro
matographic equipment in Western Europe and Asia and Australia, respectively.) BPangaro@mjhlifesciences.com
Associate Publisher
Edward Fantuzzi C.A.S.T. Data and List Information
CHANGE OF ADDRESS: Send change of address to LCGC, P.O. Box 457, EFantuzzi@mjhlifesciences.com Melissa Stillwell
Cranbury, NJ 08512-0457; alternately, send change via e-mail to MStillwell@mmhgroup.com
fulfill@mjhassoc.com or go to the following URL: Sales Manager
http://mmhpubs.mmhgroup.com/Welcome.aspx?pubid=LCGC Brianne Molnar Reprints
BMolnar@mjhlifesciences.com Alexandra Rockenstein
Allow four to six weeks for change. PUBLICATIONS MAIL AGREEMENT
ARockenstein@mjhlifesciences.com
No. 40612608. Return all undeliverable Canadian addresses Senior Director, Digital Media
to: IMEX Global Solutions, P.O. Box 25542, London, ON, N6C Michael Kushner Audience Development Manager
6B2, CANADA. Canadian GST number: R-124213133RT001. MKushner@mjhlifesciences.com Jessica Stariha
JStariha@mmhgroup.com
C.A.S.T. DATA AND LIST INFORMATION: Contact Melissa Stillwell,
EDITORIAL
tel. (218) 740-6831, e-mail MStillwell@mmhgroup.com.
CORPORATE
REPRINTS: Contact Michael J. Tessalone, Editorial Director
e-mail: MTessalone@mjhlifesciences.com Laura Bush Chairman & Founder
INTERNATIONAL LICENSING: Contact Kim Scaffidi, LBush@mjhlifesciences.com Mike Hennessy Sr
e-mail: kscaffidi@mjhassoc.com
Managing Editor Vice Chairman
John Chasse Jack Lepping
CUSTOMER INQUIRIES: Customer inquiries can be forwarded directly to MJH JChasse@mjhlifesciences.com
Life Sciences, Attn: Subscriptions, 2 Clarke Drive, Suite 100, Cranbury, NJ President & CEO
08512; e-mail: mmhinfo@mmhgroup.com Senior Technical Editor Mike Hennessy Jr
Jerome Workman
JWorkman@mjhlifesciences.com Chief Financial Officer
Neil Glasser, CPA/CFE
ycled P
Rec a Associate Editor
%
Cindy Delonas Executive Vice President,
pe
50

r
10-20%

Global Medical Affairs &

aste

CDelonas@mjhlifesciences.com
Corporate Development
rW

Po
s t C s u me
on Creative Director, Publishing Joe Petroziello
Melissa Feinen
mfeinen@mdmag.com Executive Vice President,
© 2020 MultiMedia Pharma Sciences, LLC. All rights reserved. No part of this publi- Operations
Senior Art Director Tom Tolvé
cation may be reproduced or transmitted in any form or by any means, electronic or
Gwendolyn Salas
mechanical including by photocopy, recording, or information storage and retrieval
gsalas@mjhlifesciences.com Senior Vice President, Content
without permission in writing from the publisher. Authorization to photocopy items
Silas Inman
for internal/educational or personal use, or the internal/educational or personal Graphic Designer
use of specific clients is granted by MultiMedia Pharma Sciences, LLC. for libraries Courtney Soden Senior Vice President, I.T.
and other users registered with the Copyright Clearance Center, 222 Rosewood Dr., csoden@mjhlifesciences.com & Enterprise Systems
Danvers, MA 01923, (978) 750-8400, fax (978) 646-8700, or visit http://www.copyright. John Moricone
com online. For uses beyond those listed above, please direct your written request
to Permission Dept. email: ARockenstein@mjhlifesciences.com Senior Vice President, Audience
CONTENT MARKETING
Generation & Product Fulfillment
MultiMedia Pharma Sciences, LLC. provides certain customer contact data (such as Joy Puzzo
Custom Content Writer
customer’s name, addresses, phone numbers, and e-mail addresses) to third parties
Allissa Marrapodi
who wish to promote relevant products, services, and other opportunities that may Vice President, Human
AMarrapodi@mjhlifesciences.com
be of interest to you. If you do not want MultiMedia Pharma Sciences, LLC. to make Resources & Administration
your contact information available to third parties for marketing purposes, simply Webcast Operations Manager Shari Lundenberg
email mmhinfo@mmhgroup.com and a customer service representative will assist Kristen Moore
you in removing your name from MultiMedia Pharma Sciences, LLC. lists.
Vice President, Business
KMoore@mjhlifesciences.com
Intelligence
LCGC North America does not verify any claims or other information a ­ ppearing in Project Manager Chris Hennessy
any of the advertisements contained in the publication, and cannot take responsi- Vania Oliveira
bility for any losses or other damages incurred by readers in reliance of such content. Executive Creative Director,
VOliveira@mmhgroup.com
Creative Services
To subscribe, email mmhinfo@mmhgroup.com. Jeff Brown
Quantification of Adeno-
Associated Virus Aggregation
The Field-Flow Fractionation Platform

Adeno-associated viruses (AAVs) are increasingly used for gene therapy due
to their versatility and safety. One of the biggest concerns for manufacturing
a uniform AAV suspension is the presence of viral aggregates, which
can create problems with transduction efficiency, biodistribution, and
immunogenicity. These large AAV aggregates are challenging to separate
and characterize by traditional column-based chromatography techniques
such as size exclusion chromatography (SEC).

Asymmetrical Flow Field-Flow Fractionation

with Multi Angle Light Scattering (AF4-
MALS) can separate and size large AAV
aggregates, and discern a difference in
aggregate concentration due to the
stressing protocol. Some or all of the large
aggregates would be filtered out by SEC,
resulting in incorrect determination of the
aggregate content or the false conclusion
that no aggregates are present.

For more details visit www.postnova.com and search for ‘virus’.

NovaFFF Software
MALS DLS UV RI ICP-MS

Contact us for more information: www.postnova.com

318  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

CONTENTS
COLUMNS

320 LC TROUBLESHOOTING
Recovering from a COVID-19 Shutdown:
Tips and Tricks for Starting Up, Part I

Volume 38 Number 6, 313-372

Dwight R. Stoll and Tony Taylor
When you restart liquid chromatography (LC) instrumentation that Volume 38 Number 6 June 2020
www.chromatographyonline.com

was idle during the COVID-19 shutdown, you need to follow a

systematic approach. Otherwise, problems may appear in days or
weeks following startup.

LCGC NORTH AMERICA

325 PERSPECTIVES IN MODERN HPLC
Stability Studies and Testing of Pharmaceuticals:
An Overview
LC PRACTICAL APPLICATIONS DATA HANDLING

Kim Huynh-Ba and Michael W. Dong Stability Studies

and Testing of
Pharmaceuticals
MULTIDIMENSIONAL
SEPARATIONS FOR
Controlling Peak
Integration to Ensure
Data Integrity
BIOTHERAPEUTIC

Determining product shelf life is a regulatory requirement for

GC CHARACTERIZATION THEORY &
Troubleshooting FUNDAMENTALS
Reduced Peak Starting Up Your Laboratory

June 2020
Size in GC After a Pandemic Shutdown

pharmaceuticals and many other regulated consumer products. In this

comprehensive overview of stability studies and testing, we summarize COVER IMAGERY BY
current regulatory requirements, share industry practices for forced KENTOH -
degradation, and explain approaches for reduced testing and data stock.adobe.com
evaluation to expedite stability study timelines.

338 FOCUS ON BIOPHARMACEUTICAL ANALYSIS

Multidimensional Separation Techniques for
Characterization of Biotherapeutics
Anurag S. Rathore, Ramesh Kumar, and Ira S. Krull DEPARTMENTS
Multidimensional separations, in which two or more separation 355 The Application
methods are coupled, are a valuable analytical tool for higher Notebook
peak capacity and improved selectivity for the analysis of complex
samples like biotherapeutics.

371 FUNDAMENTALS
Troubleshooting Gas Chromatography:
Reduced Peak Size (Loss of Sensitivity)
Tony Taylor
There are many potential causes of reduced peak size in gas
chromatography (GC), and an inexperienced GC user may not
know where to begin the troubleshooting process. Here, we
review potential causes for reduced peak size in GC systems.

FEATURE ARTICLE

346 Are You Controlling Peak Integration to Ensure Data Integrity?

R.D. McDowall
Chromatography data systems (CDS) have been at the center of
multiple FDA 483 citations and warning letters, with an emphasis
on peak integration and interpretation of chromatograms. Here,
we review the issues associated with ensuring compliance when
performing peak integration.

LCGC North America (ISSN 1527-5949 print) (ISSN 1939-1889 digital) is published monthly by MultiMedia Healthcare, LLC., 2 Clarke Drive, Suite 100, Cranbury, NJ 08512, and is
distributed free of charge to users and specifiers of chromatographic equipment in the United States and Canada. LCGC is available on a paid subscription basis to nonquali-
fied readers in the United States and its possessions at the rate of: 1 year (13 issues), $83.95; 2 years (26 issues), $151.20; in Canada and Mexico: 1 year (13 issues), $107.10; 2 years
(26 issues), $168.53; in all other countries: 1 year (13 issues), $157.50; 2 years (26 issues), $281.40. Periodicals postage paid at Trenton, NJ 08650 and additional mailing offices.
POSTMASTER: Please send address changes to LCGC, P.O. Box 457, Cranbury NJ 08512-0457. PUBLICATIONS MAIL AGREEMENT NO. 40612608, Return Undeliverable Canadian
Addresses to: IMEX Global Solutions, P. O. Box 25542, London, ON N6C 6B2, CANADA Canadian GST number: R-124213133RT001. Printed in the USA.
WWW.CHROMATOGRAPHYONLINE.COM
Editorial Advisory Board
• Kevin D. Altria – GlaxoSmithKline, Ware, United Kingdom
• Jared L. Anderson – Iowa State University, Ames, Iowa
• Daniel W. Armstrong – University of Texas, Arlington, Texas
• David S. Bell – Restek, Bellefonte, Pennsylvania
• Dennis D. Blevins – Agilent Technologies, Wilmington, Delaware
• Zachary S. Breitbach – AbbVie Inc., North Chicago, Illinois
• Ken Broeckhoven – Department of Chemical Engineering,
Vrije Universiteit Brussel, Brussels, Belgium
• Deirdre Cabooter – Department of Pharmaceutical and Pharmacological
Sciences, KU Leuven (University of Leuven), Belgium
• Peter Carr – Department of Chemistry,
University of Minnesota, Minneapolis, Minnesota
• Jean-Pierre Chervet – Antec Scientific, Zoeterwoude, The Netherlands
• André de Villiers – Stellenbosch University, Stellenbosch, South Africa
• John W. Dolan – LC Resources, McMinnville, Oregon
• Michael W. Dong – MWD Consulting, Norwalk, Connecticut
• Anthony F. Fell – School of Pharmacy,
University of Bradford, Bradford, United Kingdom
• Francesco Gasparrini – Dipartimento di Studi di Chimica e Tecnologia delle
Sostanze Biologicamente Attive, Università “La Sapienza,” Rome, Italy
• Joseph L. Glajch – JLG AP Consulting, Cambridge, Massachusetts
• Davy Guillarme – University of Geneva, University of Lausanne, Geneva, Switzerland
• Richard Hartwick – PharmAssist Analytical Laboratory, Inc.,
South New Berlin, New York
• Milton T.W. Hearn – Center for Bioprocess Technology,
Monash University, Clayton, Victoria, Australia
• Emily Hilder – University of South Australia, Adelaide, Australia
• John V. Hinshaw – Serveron Corporation, Beaverton, Oregon
• Kiyokatsu Jinno – School of Materials Science,
Toyohashi University of Technology, Toyohashi, Japan
• Ira S. Krull – Professor Emeritus, Department of Chemistry and
Chemical Biology, Northeastern University, Boston, Massachusetts
WinGPC UniChrom
GPC/SEC, IPC and 2D of
(Bio)Polymers and Proteins
Macromolecular Chromatography
Data System
Ensure Compliance and
Data Integrity
Instrument control for all major
LCs and specialty detectors MultiWorkstation
Data Evaluation, Client/Server
or MultiWorkstation
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  319
• Ronald E. Majors – Analytical consultant, West Chester, Pennsylvania
• Debby Mangelings – Department of Analytical Chemistry and
Pharmaceutical Technology, Vrije Universiteit Brussel, Brussels, Belgium
• R.D. McDowall – McDowall Consulting, Bromley, United Kingdom
• Michael D. McGinley – Phenomenex, Inc., Torrance, California
• Victoria A. McGuffin – Department of Chemistry,
Michigan State University, East Lansing, Michigan
• Mary Ellen McNally – FMC Agricultural Solutions, Newark, Delaware
• Imre Molnár – Molnar Research Institute, Berlin, Germany
• Glenn I. Ouchi – Brego Research, San Jose, California
• Colin Poole – Department of Chemistry,
Wayne State University, Detroit, Michigan
• Douglas E. Raynie – Department of Chemistry and Biochemistry,
South Dakota State University, Brookings, South Dakota
• Fred E. Regnier – Department of Chemistry, Purdue University,
West Lafayette, Indiana
• Koen Sandra – Research Institute for Chromatography, Kortrijk, Belgium
• Pat Sandra – Research Institute for Chromatography, Kortrijk, Belgium
• Peter Schoenmakers – Department of Chemical Engineering,
University of Amsterdam, Amsterdam, The Netherlands
• Kevin Schug – University of Texas, Arlington, Texas
• Dwight Stoll – Gustavus Adolphus College, St. Peter, Minnesota
• Michael E. Swartz – Stealth Biotherapeutics, Newton, Massachusetts
• Caroline West – University of Orléans, France
• Thomas Wheat – Chromatographic Consulting, LLC, Hopedale,
Massachusetts
CONSULTING EDITORS:
Jason Anspach – Phenomenex, Inc.; David Henderson – Trinity College;
Tom Jupille – LC Resources; Sam Margolis – The National Institute of
Standards and Technology; Joy R. Miksic – Bioanalytical Solutions LLC
Data Evaluation PERFECT
SEPARATION SOLUTIONS
Request a Live Demo at your Desk:
info@pss-polymer.com
Client/Server www.pss-polymer.com
320  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
LC TROUBLE SHOOTING
Recovering from a COVID-19 Shutdown:
Tips and Tricks for Starting Up, Part I
COVID-19-related laboratory shutdowns are sure to cause a myriad of problems with liquid chromatography (LC) instrumentation
across the globe. Taking a systematic approach to restarting these systems will save time and money in the long run, by
preventing problems that may otherwise appear in days or weeks following startup.
Dwight R. Stoll and Tony Taylor
I n March of this year, many organizations took
unprecedented steps to halt the spread of
COVID-19, including severely restricting work
in laboratories, or even shutting down entire
laboratories, buildings, and work sites for
weeks at a time. While some of these shut-
downs were planned days in advance and
executed well, I have heard many stories from
scientists indicating that the shutdowns were
sudden, and did not allow time to properly
prepare their analytical instrumentation to
be idle for weeks or months at a time. Unfor-
tunately, this means that these scientists are
going to encounter many challenges when
they return to the laboratory, which will neces-
sarily include a lot of troubleshooting to figure
out why their systems are not working prop-
erly before they can return to their normal
experience of producing high quality data.
For this month’s installment of “LC Trouble-
shooting,” I’ve decided to commit the entire
column to providing some tips and tricks that
should be useful as scientists go about the
process of restarting their laboratories. I’ve
invited Tony Taylor to join me in sharing some
of the advice and procedures he has devel-
oped over the years. Although we’ve never
worked through pandemic-related laboratory
shutdowns and startups, some of our expe-
riences with troubleshooting liquid chroma-
tography (LC) problems during normal times
should prove very useful under these strange
circumstances as well.
Dwight Stoll
As we discussed the topics to address
in this installment, we decided to focus
on the challenges that are likely to
appear most often. When the pandemic
has passed, and we all have the chance
to discuss problems we encountered
and strange observations we made,
we’re sure that the list will be very long,
and highly diverse. In the following sec-
tions we’ve tackled specific aspects of
LC systems and their operation, and
tried to anticipate the most common
shutdown-related problems and their
solutions. In next month’s installment,
we’ll address other important topics,
including column cleaning and health
assessment, overall LC system health,
and making the decision to start collect-
ing data again.
We strongly encourage you to take
care and use a systematic approach to
bringing your LC system back on-line.
A little effort at this point with preven-
tative measures can save a lot of time
and money later on when you are ready
to start collecting useful data again.
Remember that the column is the heart
of any LC system, and arguably the most
sensitive to potential shutdown-related
problems. Be sure to disconnect the
column from the flow path before doing
anything else with the system, and only
reconnect it to the flow path after you
are sure that the rest of the system (that
is, solvent bottles, lines, pump, and flow
path upstream from the column) is free
of contamination and blockages.
A final general cautionary note here
is that great care should be taken to
avoid precipitation of salts inside the
LC system. Aside from evaporation of
water from buffer solutions, precipita-
tion of buffer salts can also easily occur
when a buffer solution comes into con-
tact with a pure organic solvent. This
is because most commonly used buf-
fer salts (although inorganic salts, such
as potassium phosphate, are generally
worse than organic salts such as ammo-
nium acetate) are highly insoluble in
organic solvents, such as methanol and
acetonitrile (1). So, one should avoid—
at all costs—a situation where a LC sys-
tem component containing an aqueous
buffer is flushed immediately with a
pure organic solvent. Such a change-
over should involve flushing the buf-
Icon Image: Joe Zugcic / Zugcic Photographers, Inc.
fer first with water, and then the
organic solvent.
Solvents and Buffers
Figure 1 shows a picture of a buffer bot-
tle I observed in an LC laboratory about
a year ago. The solution was dilute (1
mM) phosphoric acid in water. Even
though this is fairly acidic (pH ~2.7),
there was a very healthy microbial com-
munity growing inside (affectionately
Bio Fermented Alcohol Detection
Utilizing Isocratic Cation Exchange
HPLC with Hamilton’s HC-75 Column

Concern over detrimental greenhouse gases produced from the
combustion of fossil fuels has been growing for the past few decades.
Identification of new and more diverse methods of energy use has been
investigated around the world. One such avenue is combustible alcohols
produced from biomasses like sugar beet or sugar cane. To reduce
dependence on fossil fuels, a concerted effort has been made to develop
processes where biomass can be fermented biologically by bacteria or
yeast to produce combustible alcohols. Combustible alcohols are desired
over other technologies due to the convenience of switching from petrol
based fuels. Ethanol is already added to most gasolines to increase
octane efficiency. With anywhere from 5–25% ethanol being added to
petrol products in most of the world and up to 100% ethanol fuel in Brazil.1
Additionally, the combustion of these alcohols tend to be more efficient
and lead to less carbon monoxide side products.2 Production of ethanol
from biological sources produce side products, specifically propanol
and methanol amongst other ketones, fatty acids, and esters. Due to
azeotropic formation of ethanol with water it is difficult to distill pure
200 proof ethanol products. However, HPLC analysis provides a good
platform for ethanol determination after biomass fermentation.3 To aid
the detection of ethanol purity, Hamilton has developed a method utilizing
the HC-75 (H+ Form) HPLC column, which provides good separation
of the desired analytes. With ~8% crosslinking of the PS-DVB backbone
the resulting porous gel stationary phase matrix allows for enhanced
interactions between stationary phase and analyte. Detection was
accomplished with refractive index, but other methods of detection
can be applied (i.e. UV, or mass spectrometry).

Column Information
Packing Material HC-75 (H + Form), 9 µm 3

P/N 79544

Chromatographic Conditions
2
Gradient Isocratic

Temperature 80°C
1
Injection Volume 10 µL

Detection Refractive Index

Dimensions 305 x 7.8 mm 0 5 10 15 20 25

Time (minutes)
Eluent A 0.55 mM Succinic Acid
Compounds: 1: 2-Propanol 2: Ethanol 3: Methanol
Flow Rate 0.6 mL/min

1) E. Gnansounou, A. Dauriat, Ethanol fuel from biomass: A review. J. Sci. & Indust. Res.
(64) 2005, 809–21.
2) S. Onuki, J. A. Koziel, W. S. Jenks, L. Cai, D. Grewell, J.H. van Leeuwen, Taking ethanol
quality beyond fuel grade: A review. J. inst. Brew. (122) 2016, 588–98.
3) S.Y. Lee, R. Sankaran, K.W. Chew, et al. Waste to bioenergy: a review on the recent
conversion technologies. BMC Energy. 1:4, 2019.
©2020 Hamilton Company. All rights reserved.
All other trademarks are owned and/or registered by Hamilton Company in the U.S. and/or other countries.
Author: Adam L. Moore, PhD, Hamilton Company Lit. No. L80105 — 05/2020

Web: www.hamiltoncompany.com Hamilton Americas & Pacific Rim Hamilton Europe, Asia & Africa
Hamilton Company Inc. Hamilton Central Europe S.R.L.
USA: 800-648-5950 4970 Energy Way
Reno, Nevada 89502 USA
str. Hamilton no. 2-4
307210 Giarmata, Romania
Europe: +40-356-635-055 Tel: +1-775-858-3000 Tel: +40-356-635-055
Fax: +1-775-856-7259 Fax: +40-356-635-060
To find a representative in your area, please visit hamiltoncompany.com/contacts. sales@hamiltoncompany.com contact.lab.ro@hamilton-ce.com
322  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
FIGURE 1: Image of an LC solvent bottle
with visible (cloudy) microbial growth.
known as “floaties”). The good news
in this particular case is that the scien-
tists were able to recover their pumping
system without any long-term effects;
however, the column they were using
at the time suffered a performance loss,
and was never quite the same after-
ward. We expect that this experience
is going to be repeated in hundreds of
laboratories as a result of the COVID-
related laboratory shutdowns. Many of
the aqueous buffer solutions used in
LC (and even high-performance liquid
chromatography [HPLC] grade water,
if given enough time and exposure to
laboratory dust) are environments quite
favorable to microbes, particularly those
in the middle of the pH range. Simply
stated, many scientists are going to
return to laboratories where they will be
able to see visible microbial growth in
the solvent bottles of their LCs.
So…what to do?
Occasionally we have problems in our
own laboratories with solvent bottles
that become contaminated somehow,
whether resulting from a bad lot of sol-
vent from a vendor, or an instrument
operator pouring something into the
bottle that does not belong in there.
Whenever this happens, our approach
depends on whether or not the instru-
ment is using mass spectrometric (MS)
detection. If it does, our preference is
to simply not use that solvent bottle
again for that instrument, because if
thoroughly cleaning the bottle requires
detergent or other cleaning agents, the
potential for problems associated with
those cleaning agents is greater than
the benefit of having a clean bottle to
work with. For example, some deter-
gents can be difficult to remove from
glassware, to the point where they are
not detectable by MS, and thus what
seems like a simple bottle cleaning step
can turn into a background contaminant
peak in the MS that is detectable for
months. When using MS detection, we
feel the best solution to a badly con-
taminated solvent bottle is a brand new
solvent bottle. In other words, just get
rid of the contaminated bottle without
bothering to try to clean it.
On the other hand, if the contami-
nated bottle is on an instrument using
a less sensitive detector (for example,
ultraviolet-visible [UV-vis] spectros-
copy or an evaporative light scatter-
ing detector [ELSD]), then the bottle
usually can be cleaned without major
impact on the rest of the system. After
discarding the contaminated solvent,
wash thoroughly with a detergent
designed for use with glassware, rinse
four times with HPLC-grade water, and
finally with HPLC- grade organic sol-
vent. We find that working through a
series of organic solvents of increasing
dipolarity—such as cyclohexane, aceto-
nitrile, and methanol (in that order)—is
effective for removing contaminants of
varying degrees of dipolarity. Before
use again on the instrument, be sure to
rinse with whatever solvent you plan to
put in the bottle.
Now, if you come back to your labo-
ratory and find microbes in your solvent
bottles, then the solvent line feeding the
pump is also contaminated, in addition
to the solvent bottle itself. Before put-
ting this solvent line back into the solvent
bottle you have just cleaned, consider
what exactly you will do with the solvent
line. Again, if this is an LC–MS instrument,
you should seriously consider replacing
the entire solvent line and solvent inlet
filter that goes in the solvent bottle. In
any case, if the solvent inlet filter is
porous—whether sintered glass or stain-
less steel—just throw it away, and replace
it with a new one. These porous filters are
perfect environments to support micro-
bial growth, and getting them truly clean
after they have been contaminated is
extremely difficult. Inlet filters that are
based on screens or mesh are more eas-
ily cleaned. Remove the filter from the
solvent line, remove the visible debris
with a brush, and then soak in detergent,
followed by several rinses with water,
and finally organic solvent. Sonicating
sintered glass filters should be avoided,
but sonicating plastic or metal ones is
okay. As with the bottle, rinse the filter a
few times with the solvent you intend to
use it with before putting it in the clean
bottle. For the solvent line itself, it is best
to remove the line from the pump and
attach a syringe (10 mL or so will do) so
that you can pull large quantities of rins-
ing solvent through the line. For older
systems that have degasser modules
separate from the pump itself, discon-
nect the solvent line at the pump so that
you can pull cleaning solvents through
the degasser also. Pulling detergent
solution through the degasser should
be avoided, because this may persist in
the tubing for a long time. We find that
using warm water (up to 60 °C) is quite
helpful for flushing the internal tubing
of the degasser. For newer pumps with
integrated degassers, it should be suf-
ficient to disconnect the solvent line at
the inlet to the degasser on the pump
unit. Using the same steps as with the
bottle and filter, pull each liquid through
the solvent line as a means of rinsing the
inside of the line. For older systems with
degassers that have large internal vol-
umes, something on the order of 50 mL
of each liquid will be enough to give the
solvent line and degasser a good flush.
With newer systems where you only need
to flush the line external to the pump; 10
mL or so should be good enough.
Dirty and Dried-Out Pumps
Before attempting to move any liquid
through the pump itself, it is best to dis-
connect the high pressure outlet capil-
 WWW.CHROMATOGRAPHYONLINE.COM
lary from the pump and connect a waste
line running to a waste container to col-
lect whatever goes through the pump
until you are confident that the pump is
delivering clean, debris-free solvent. If it
is obvious that one or more solvent lines
has been contaminated with microbes,
it will be best to flush the pump thor-
oughly with clean solvent before con-
necting the pump to the rest of the
system. Our approach to this step is to
put all of the solvent lines associated
with the pump (typically two for a binary
pump or four for a quaternary pump)
into a single bottle of HPLC-grade
water, and run 30 mL of water through
each channel of the pump. Next, put all
of the solvent lines into a single bottle
of isopropyl alcohol (IPA), and run 30 mL
of it through each channel. Finally, move
each solvent line back to its respective
solvent bottle (for example, one in water,
one in buffer, and one in acetonitrile),
and flush each channel with 30 mL of the
solvent intended for use in the method.
It is important to recognize that there
are some filters in modern LC systems
that might be easy to overlook. It is
always helpful to consult the technical
manual for your LC system to see where
these might be. For example, some
pumps have porous frits or screens on
the upstream side of the inlet check
valve. Other pumps have porous frits on
the downstream side of the outlet check
Unbeatable accuracy
again and again
The precise and time-pressured nature of your work
mean you need products you can rely on, every time.
That’s why every product across the comprehensive
Supelco® portfolio is meticulously tested against the
tightest specifications to ensure the highest industry
standards and compliance with applicable regulations.
So for unbeatable accuracy every time, choose
Supelco® analytical products.
For more information, please visit:
SigmaAldrich.com/SuccessReplicated
MilliporeSigma, the vibrant M and Supelco are trademarks of Merck KGaA, Darmstadt,
Germany or its affiliates. All other trademarks are the property of their respective owners.
Detailed information on trademarks is available via publicly accessible resources.
© 2020 Merck KGaA, Darmstadt, Germany and/or its affiliates. All Rights Reserved.
The Life Science Business of Merck KGaA, Darmstadt, Germany operates as
MilliporeSigma in the US and Canada.
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  323
valve. These can also suffer from partial
or total obstruction, leading to inconsis-
tent flow (if on the inlet side), or higher
than expected pressure (if on the outlet
side). If you know you are dealing with
a situation where it is likely these filters
or screens have been fouled due to pro-
longed inactivity of the pump, microbial
growth, or other debris, it will be a good
idea to simply replace these filters or
screens at the beginning of your startup
process. If you want to try cleaning them
before replacing, remove and sonicate
in warm water for 10 min, followed by
IPA for 10 minutes. If problems persist
that are attributed to these parts, just
replace them entirely.
If the pump has dried out entirely
because it was left on with an empty
solvent bottle, it can be difficult to
remove air bubbles from pump heads
during startup. Air bubbles can also
accumulate in pump heads if the sys-
tem is not used for an extended period
of time. In our experience, the best way
to resolvate a completely dry pump or
remove air bubbles from the pump
heads is to purge thoroughly with IPA,
because it wets most materials used in
LC pumps well.
Once the pump has been thoroughly
purged with clean solvent and it is deliv-
ering the expected flow from each chan-
nel at low pressure, it can be connected
to the rest of the system. Again, take
care to be sure that, if the rest of the sys-
tem contains a buffer solution, the first
solvent pumped from the clean pump
to the rest of the system is water, so that
precipitation of salts can be avoided.
Component Blockages Due
to Precipitation and Debris
Some users will return to their labora-
tories to find that there is a significant
obstruction somewhere in their LC sys-
tem that is preventing flow at reason-
able pressures. This obstruction could
result from debris associated with con-
taminated solvents, pump or valve seal
material if salt precipitation occurs in
those components, or precipitation of
salts inside of some other components
such as a connecting capillary. The first
step in resolving such obstructions is
to clean the pump as discussed above,
and make sure it will pump water with
a waste tube connected directly to
the high pressure outlet of the pump.
Then, systematically add one connec-
tion or component at a time, working
your way from the pump outlet toward
the detector. In most systems, the next
component after the pump will be the
autosampler, with a fairly long connect-
ing capillary between them. Disconnect
this capillary from the sampler, con-
nect it to the pump, and see if you can
pump water through this capillary at
reasonable pressure; what pressure is
324  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
reasonable depends on the dimen-
THEIRS IS sions of the capillary. A benchmark to
FAMILIAR. remember is that the pressure drop
OURS, across a 300 mm length of 120-µm
i.d. (0.005 in.) capillary when pumping
SUPERIOR. water at room temperature and 1 mL/
min is about 10 bar. If you double the
length of this capillary, the expected
pressure will double. If you double
the diameter, the expected pressure
drops by a factor of four. In our experi-
ence, it is best to simply replace the
capillary if the pressure is more than
two times higher than the expected
pressure after 5 min of flushing with
water. Once you have verified that this
capillary is okay, either by flushing or
replacing it, then the next likely com-
ponent is the autosampler itself. Here
there are multiple sources of poten-
tial obstruction, including the sampler
needle and needle seat capillary in the
case of a flow-through needle design,
or a sample loop capillary if using a
fixed volume injection design. If the
pressure increases dramatically when
the sampler is reconnected, then it is
likely that one of these components is
obstructed. Occasionally, this can be
resolved by removing the capillary and
sonicating in warm water, but the easy
fix is to simply replace a capillary that
you know is obstructed based on the
pressure it adds to the system when
it is in the flow path. You should keep
adding one component at a time to
the flow path until you discover the
one that is obstructing the flow.
Finding the component in a sys-
tem that is leading to higher-than-
expected pressures can be tricky
sometimes, but most of the time the
ION CHROMATOGRAPHY BY
systematic approach described here
will identify the culprit.
One other component that deserves
special mention here is the flow cell of
optical detectors. Most UV and fluo-
Get the support your lab deserves. rescence detectors have quartz com-
Learn more at LeaveTheFamiliar.com
ponents in contact with the mobile
phase that can be damaged if left in
contact with highly alkaline solutions
for extended periods, or contami-
© 2019 Metrohm USA, Inc. Metrohm and design® is a registered trademark of Metrohm Ltd.
nated by microbial growth or other
particles that will reduce the trans-
mission of light through the flow cell.
Most manufacturers recommend flush-
ing the flow cell with warm water (up
to 60 °C), followed by IPA. If excessive
noise or poor sensitivity are observed
after cleaning, the flow cell windows
may have to be replaced.
Summary
Given the variety of ways different labo-
ratories were shut down early on in the
global COVID-19 outbreak, it is likely
that LC users will encounter a wide vari-
ety of problems with their instruments
when they return to their laboratories
and resume work with their instruments.
In this installment of “LC Troubleshoot-
ing,” we have highlighted some of the
problems we expect users to encounter
most frequently, and made suggestions
for resolving these problems in a system-
atic and efficient way. Taking the time to
thoroughly clean and examine your LC
system when you return to the labora-
tory will pay dividends later by avoiding
costly interruptions to normal work.
References
(1) A.P. Schellinger and P.W. Carr, LCGC
North Amer. 22(6), 544–548 (2004).
ABOUT THE COLUMN EDITOR
Dwight R. Stoll is the edi-
tor of “LC Troubleshooting.”
Stoll is a professor and the
co-chair of chemistry at Gus-
tavus Adolphus College in St.
Peter, Minnesota. His primary research
focus is on the development of 2D-LC for
both targeted and untargeted analyses.
He has authored or coauthored more
than 60 peer-reviewed publications and
four book chapters in separation science
and more than 100 conference presenta-
tions. He is also a member of LCGC’s edi-
torial advisory board. Direct correspond-
ence to: LCGCedit@mmhgroup.com
ABOUT THE CO-AUTHOR
Tony Taylor is the Chief Sci-
ence Officer of Arch Sciences
Group and the Technical
Director of CHROMacademy.
Direct correspondence to:
LCGCedit@mmhgroup.com.
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  325

PER SPEC TI V E S
IN MO DERN HPLC

Stability Studies and Testing of

Pharmaceuticals: An Overview
This installment is the first of a series of three white papers on stability studies and testing of pharmaceuticals, as well as the
development and validation of stability-indicating high performance liquid chromatography (HPLC) methods. The series is
co-authored by Kim Huynh-Ba, a subject-matter expert on stability testing and regulatory compliance, and Michael Dong,
the columnist on “Perspectives in Modern HPLC.” This first installment provides a comprehensive and updated overview
of stability studies and testing of small molecule drugs, current regulatory requirements, and industry practices for forced
degradation, as well as possible approaches for reduced testing and data evaluation to expedite stability study timelines.
Kim Huynh-Ba and Michael W. Dong

D etermining product shelf life is a reg-

ulatory requirement for pharmaceu-
ticals and many other regulated consumer
products. The shelf life of medicines is set
following stringent regulations; therefore,
efficient application of stability science is
critical. The shelf life (expiration dating or
expiry) is displayed on labels of pharma-
ceutical products to ensure the integrity,
quality, and potency of the product when
used within that time period. Shelf life is
established using data that are generated
to verify the label claim, and approved
by the regulatory agencies. An expiration
date is required by regional laws to ensure
the safety, efficacy, and quality of the drug
products, and that these criteria are main-
tained throughout the labeled shelf life of
the pharmaceutical product.
Most companies have established stan-
best practices, and regulatory expecta-
tions of stability programs.
The stability profile, a critical quality attri-
bute (CQA) of a pharmaceutical entity, is
based primarily on the physicochemical
properties of the drug substance (DS) and
drug product (DP). The DS is the active
pharmaceutical ingredient (API), with small
amounts of impurities and degradation
products. A DP typically contains a single DS;
when the DP formulation has more than one
DS, it is referred to as a combination product.
For example, many over-the-counter (OTC)
products contain multiple active ingredients.
For small-molecule orally delivered DPs,
such as tablets or capsules, a typical expira-
tion period is two to five years under room
temperature storage. Table I summarizes the
characteristics of stability studies and testing.
Each is discussed further in this paper.
Stability Studies in
New Drug Development
During drug discovery, the medicinal
chemist focuses on synthesizing new
chemical entities (NCEs) that interact selec-
tively with the molecular targets (such as
receptors or enzymes), leading to potential
disease mitigation (1). The structural motifs
of a leading NCE candidate are then opti-
mized for biological activities and safety by
the synthesis of a series of analogs. The
optimization tools include in vitro and in
vivo animal studies, (such as biochemical
target binding, biomarkers, or animal mod-
els), bioavailability (pharmacokinetics), and
toxicological evaluations (1,2). The desired
NCEs need to be reasonably stable to be
viable drug candidates, which are scaled
up for clinical trials using simple formula-
tions. Eventually, the final DS is formulated
Icon Image: Joe Zugcic / Zugcic Photographers, Inc.

dard operating procedures (SOPs) to sup-
plement regulatory guidelines to provide
more specific details, to ensure that stabil-
ity studies are appropriate for their specific
product types. Nevertheless, stability pro-
grams and their practices can vary widely,
particularly between large and small
pharmaceutical companies, often due to
the depth of knowledge and available
resources. The primary aim of this paper is
to increase understanding of the science,
In this paper, we use solid oral dos-
age forms, such as tablets or capsules, as
examples. Four types of testing for orally
available products, listed in Table II, are
discussed. Biological products are not
covered here.
Other dosage forms such as parenterals
also have similar types of tests, including
identity, chemical testing (assay and impu-
rities), physical testing (pH, clarity, particu-
lates), and sterility tests as appropriate.
into a commercial drug product, and sub-
mitted for regulatory approval.
In most pharmaceutical companies, the
nomination of an NCE to the status of a
drug development candidate triggers the
formation of a multidisciplinary technical
development team. This team is responsi-
ble for taking the drug candidate into clini-
cal trials in humans, and eventually to drug
approval, production, and commercializa-
tion. The team is responsible for activities
326  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

referred to in regulatory documents under

TABLE I: Summary of characteristics of stability studies and testing
the heading of “Chemistry, Manufacturing,
Stability Study and Control” (CMC), and thus the team is
Stability Testing Characteristics
Characteristics
often called the CMC team.
Stability studies must be • Drug regulations require expiration dates to be posted on Stability studies for clinical trial materi-
conducted to meet the labels for all pharmaceutical products.
regulatory expectations • Stability data of DS and DP are submitted to regulatory agencies. als (CTMs) are conducted to monitor their
CQAs and help identify which formulation
• Formal stability studies are used to justify storage conditions
and the expiration dating of a packaged product. will result in a successful candidate for reg-
Stability studies fall • Other studies are secondary or those used to support the ulatory submission. In the United States,
into different types primary stability data set.
an expiration date is not required on the
• Stress studies include photostability studies and studies that sup-
port temperature excursions throughout the supply chain. label of a CTM, but this is required in many
• Stability studies are conducted at several environmental condi- countries, such as those in Europe. Forced
tions based on the desired labeled storage conditions. degradation studies (also referred to as
• For products to be stored at room temperature, studies are stress studies) are used to challenge the
Studies are conducted conducted at room temperature and accelerated
under multiple storage conditions. Intermediate condition can be studied if samples stability-indicating power of the analytical
conditions at accelerated method. The accelerated studies, which
conditions may not meet room temperature specifications. require storage under higher tempera-
• Refrigerated or freezing conditions are used for
products that are not chemically stable at room tempera- ture and humidity compared to normal
ture (such as biologics). room-temperature conditions, allow the
• Attributes that change over time are required to be samples to degrade at a faster rate. Data
assessed in a stability study. from accelerated studies may be extrapo-
• The chemical, physical, microbiological, and lated to project the results of a longer-term,
Stability-indicating
performance characteristics of DP must meet
analytical procedures
acceptance criteria throughout the shelf life.
controlled-room-temperature study. Most
must be followed pharmaceutical products exhibit a linear
• Forced degradation studies are performed to challenge the
stability-indicating power of the method and to degradation trend. Based on the stability
evaluate the major degradative pathways of the API. data, a retest period is assigned to the DS,
• Stability testing consumes a significant part of internal and an expiration date to the DP.
Stability studies are
analytical testing and outsourcing resources of
resource intensive The CMC team typically consists of scien-
pharmaceutical companies.
tists from various functions, such as process
chemistry (synthetic organic chemistry), for-
mulation (pharmaceutics), analytical chem-
TABLE II: Four categories of stability tests for tablet or capsule drug products istry (analytical development and quality
Category Typical Tests control [QC]), pharmacokinetics and drug
metabolism (PKDM), outsourcing, regu-
The primary stability-indicating assay is typically an HPLC-UV
Chemical testing method for potency and chemical impurities. Others are tests latory affairs, supply chain, and project
for chiral impurities (HPLC-UV) and moisture (Karl Fisher) management. The degree of involvement
Tests for appearance, color, solid-state characterization (crystallinity, depends on the development phase. For
Physical testing
polymorph by X-ray power diffraction [XRPD]), thermal methods) example, the analytical development chem-
Performance testing
Dissolution (or disintegration) tests to measure the ist is more involved in the discovery and
rate of release of the solid dosage form. early-phases of development to support DS
Microbial limit testing Tests indicated in USP General Chapters <61> and <62> and DP processes. In contrast, quality con-
trol chemists and regulatory staff are more
active in the later clinical phases for CTM
TABLE III: Storage conditions established by the ICH Q1A (R2) guideline for zone I/II. manufacturing and regulatory filings (3,4).
Some of the team activities include:
Label
Storage Conditions Storage Conditions to be Studied • conducting the necessary synthetic
Conditions
process scale-up of the DS (process
Long-term storage 25 ± 2 °C and 60 ± 5% RH
Controlled chemistry staff)
room Intermediate condition 30 ± 2 °C and 65 ± 5% RH
• performing detailed characterization of
temperature
Accelerated condition 40 ± 2 °C and 75 ± 5% RH the API, developing and validating ana-
Refrigerated Refrigerated condition 5 ± 3 °C lytical procedures for DS and DP, and
Freezer Freezing condition –20 ± 5 °C identification of critical impurities and
degradation products (analytical devel-
opment staff)
WWW.CHROMATOGRAPHYONLINE.COM
• finding the optimum solid-state form
(salt, crystallinity) or polymorph of the
API (analytical staff)
• conducting stability studies to support
clinical development and registration
(analytical and QC staff)
• designing and developing processes for
formulations of CTMs and the final com-
mercial DP (pharmaceutics staff)
• setting acceptance criteria for CQAs
(specifications) to monitor clinical quality
and establishing commercial specifica-
tions (QA and regulatory staff)
• manufacturing CTMs (CMC team and out-
sourcing and project management staff)
• assembling the CMC submission pack-
age for regulatory filings (CMC team
and regulatory staff).
In this series of white papers on stability
testing, we focus on the tasks performed
by the analytical development and QC
scientists within the CMC team. The first
task for the analytical chemist assigned to
the CMC team is to develop a reasonable
stability-indicating method for monitoring
Changing the ART of analytical chromatography with μPACTM capLC columns:
Unrivalled separation power Micromachined reproducibility Flow rate flexibility Plug-and-play connectivity
Designed for increased robustness and throughput
without losing sensitivity.
Discover our products on
www.pharmafluidics.com
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  327
the quality of CTMs. This method is typi-
cally a “composite” analytical procedure
that determines both the potency of the
API and chemical impurities of the DS. Ide-
ally, it can also be used for the assessment
of DP samples. Once a feasible analytical
procedure is established, forced degra-
dation studies are performed to evaluate
the specificity of the analytical procedure.
This method is used in the release testing
of the early clinical batches and testing of
samples from initial stability studies to sup-
port clinical trials (3–5).
The stability-indicating assay and
impurities method is typically a gradient
reversed-phase HPLC method with UV
detection that can separate the API and
all the known impurities and degradation
products (5). This method is validated to
ensure adequate analytical performance
(specificity, linearity at 100% target concen-
tration and expected impurities concentra-
tions, accuracy, precision, sensitivity, and
solution stability) for “formal stability stud-
ies” (defined as studies conducted under
1 - 15 µl/min
Good Manufacturing Practices [GMP]) sup-
porting the establishment of expiration
dates for regulatory filings (5,6). Details on
the development and validation of the sta-
bility-indicating methods will be covered
in the second and third installments of the
series. Forced degradation and stability
studies are discussed in later sections here.
A better understanding of stability
studies and their regulatory requirements
is essential for the analytical chemist for
several reasons (6,7). First, these stud-
ies require many different tests (shown
in Table II) on multiple batches stored at
different storage conditions and pulled at
numerous time intervals. Second, the pro-
cess development or formulation scientists
may request many experimental batches
of DS or formulations to be placed on sta-
bility, leading to redundant testing, bring-
ing little added value if not managed with
a science- and risk-based approach. There-
fore, a better understanding of the design,
intent, and best practices of stability stud-
ies and a proper interpretation of associ-
328  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
TABLE IV: An example of stress conditions for drug substance
Factor Being
Conditions
Studied
a) Open dish
Light b) In the immediate container
c) In the secondary container
Heat (solution) a) 60 °C
Heat (solid) a) 50 °C, 60 °C, 70 °C, 80 °C
Accelerated a) Open dish 40 °C/75% RH or greater
stability b) In the immediate container
a) 3% peroxide at RT*
Peroxide
b) 3% peroxide at <40 °C
a) 0.1–1 N HCl (at RT)
Acid (solution)
b) 0.1–1 N HCl (at 60 oC)
a) 0.1–1 N NaOH (at RT)
Base (solution)
b) 0.1–1 N NaOH (at 60 °C)
* RT = room temperature
ated regulations can lead to more efficient
implementation of stability programs.
Objectives of this Article
In this installment, we discuss:
• regulations and regulatory expectations
of stability programs
• the goals of forced degradation stud-
ies and examples of forced-degradation
protocols
• a strategy for conducting stability stud-
ies to expedite regulatory filings
• modern predictive stability software
tools and statistical approaches to expe-
dite stability studies in a science- and
risk-based approach.
Regulations and Regulatory
Expectations of Stability Programs
Stability is a Critical Quality
Attribute for Pharmaceuticals
In the Quality by Design (QbD) approach,
CQAs describe the properties or attri-
butes pertinent to product quality (8–11).
For pharmaceutical products, CQAs are
characteristics that impact the safety and
efficacy of drug products. Acceptable
limits or product specifications must be
established and verified by release test-
ing (11). Stability studies of DS and DP are
conducted throughout the drug develop-
ment process, from the preclinical stage to
final product approval, with the study size
dependent on the phase of development.
The initial analytical development activities
include the development of analytical pro-
Time
Q1B exposure levels
(2x to 10x)
Up to 24 h
Up to 24 h
Up to 7 d
a) 30 min to 2 h
b) 30 min to 2 d
a) 30 min to 2 h
b) 30 min to 2 d
a) 30 min to 2 h
b) 30 min to 2 d
cedures, establishment of acceptance cri-
teria, forced degradation studies, DS and
DP method validation, and stress tests (3,4,
12). In early phases, stability studies may
be conducted under limited accelerated
conditions, and often without a fully vali-
dated suite of analytical methods. CTMs
are released against a set of acceptance
criteria (such as specifications) for clinical
use by the Quality Assurance (QA) group.
Once a product proceeds to Phase III
development and preparation for product
registration, GMP regulations must be fol-
lowed for finished pharmaceutical products
starting from the registration batches (13).
The registration batches and commer-
cial products destined to be distributed
for patients’ use must undergo stringent
release testing by QC to meet the estab-
lished commercial specifications for identity,
purity, potency, and quality of the finished
product. While QC performs release test-
ing of the registration batches and issues
a Certificate of Analysis (CoA), it is officially
released and signed-off by the QA group
(or by a “Qualified Person” [QP] in Euro-
pean Union pharmaceutical regulation) for
market distribution after receiving approval.
A portion of the registration batches is gen-
erally used for “formal (registration) stabil-
ity studies,” and upon commercialization,
stability samples are evaluated annually in
marketed product stability programs.
Stability testing of CTMs typically
includes tests for appearance, identifica-
tion and quantification (assay), impurities,
dissolution (for solid dosage forms), mois-
ture, and additional physical character-
izations such as X-ray powder diffraction
(XRPD) tests for crystallinity and polymorph
form. This set of tests (customized for the
specific DS or DP, or compendial tests) is
documented in the product specifications.
They must be used to assess the identity,
physical form (color, size, and crystallin-
ity), purity (chemical and chiral purity) (3,4),
water content, performance (dissolution),
and microbial activity (4,14,15), to ensure
that drug quality is maintained throughout
the stability studies.
In a stability study, representative
batches of pharmaceuticals are stored in
controlled stability chambers at different
temperatures and relative humidity (RH)
levels in proposed packaging (containers,
closure systems) (6,7). These storage con-
ditions represent long-term, accelerated,
or stress storage conditions. Samples are
periodically removed from these chambers
according to a pre-approved schedule,
and tested to verify that the pharmaceu-
tical products meet specifications within
the established expiration date under the
recommended storage and packaging
conditions. Formal stability testing to sup-
port regulatory filings must be conducted
with validated methods according to GMP
regulations (13,16). These stability data
are also used to support the manufactur-
ing, registration, shipping, and transporta-
tion of commercial products. Furthermore,
stability data are collected through clinical
formulation phases, process development,
proposed packaging selection to establish
the drug product’s shelf life, storage condi-
tions, and commercial specifications.
Post-marketing studies are conducted
to verify that any changes to raw material,
manufacturing steps, or container or closure
systems have not affected the DP through
its expiration date based on product com-
mitment. Any planned change of raw mate-
rial, manufacturing, or packaging must be
evaluated to determine if additional stability
studies or regulatory filings are needed.
The United States Food and Drug Admin-
istration (FDA) has determined that the sta-
bility profile is a critical attribute of a pharma-
ceutical product. Therefore, stability studies
are necessary to support any submission
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  329

such as an Investigational New Drug (IND),
New Drug Application (NDA), or Abbrevi-
ated New Drug Application (ANDA) filing (1).
The stability data packages within the CMC
section are typically the most extensive non-
clinical sections in these submissions. They
include tabulated data, graphs, narratives to
summarize the stability profile of the pack-
aged product to justify specifications, and
proposed expiry for the product.
Harmonization of Regulations: ICH
Quality Guidelines for Stability Studies
ment process, according to ICH Q1A (R2)
and ICH Q2 (R1) (8,16). Although regula-
tions mandate that these studies must be
performed, the guidance does not provide
directions on procedures or conditions to
use. The actual protocols employed appear
to vary widely in different organizations. An
overview of the chemistry fundamentals of
drug forced degradation studies and best
practices can be found in books, articles,
and other resources (17–20).
Forced degradation studies are per-
formed to investigate the major degrada-
tive pathways of the DS and DP and support
the initial development of the DS stability-
indicating methods (3,5,12). High-stress
conditions are employed to subject DS and
DP to conditions more severe than acceler-
ated conditions to serve several purposes:
• to provide insight into the degradation
pathways of DS and DP and their mech-
anisms, facilitating the development

YOUR LAB WILL FETCH MORE CASH

Given that stability testing is expensive and
labor-intensive, it is imperative to understand
the regulatory requirements and their intent
to avoid unnecessary studies. In addition,
regulations vary significantly from country to
country, which impacts the cost of new drug
development and the registration timeline
for a global product launch. Therefore, in
the early 1990s, the International Council for
Harmonisation of Technical Requirements
for Pharmaceuticals for Human Use (ICH),
consisting of representatives from regulatory
agencies and scientists from pharmaceutical
industries, was formed. The council’s role is
to harmonize the quality requirements for
pharmaceuticals to minimize redundant test-
ing, and reduce the time and cost of new
product development. The ICH stability
guideline, ICH Q1, was the first quality guide-
line established by the council, indicating the
significant need for harmonization of global
stability requirements.
The ICH initially harmonized three geo-
graphic regions: the United States, the
European Union, and Japan, which encom-
passed Zone 1 and Zone 2 of the World
WITH A PROTON ONSITE GENERATOR
Health Organization (WHO) climatic zones.
• H2, N2 and Zero Air • Proven Safe
However, many countries outside Zone 1
and 2 joined the ICH, or voluntarily adopted • Consistent Purity • Cost Effective
these guidelines later with various modifica- • Consistent Pressure • Eliminates Cylinder
tions. The current stability guideline (Q1A Storage and Delivery
[R2]) harmonized the storage conditions, Issues
frequency of required testing, and the mini-
mum amount of data needed for registra-
tion (6–8). Table III lists the stability storage
conditions established by ICH Q1A (R2) (8).

Forced Degradation Studies

Forced degradation studies, or stress test-
ing, are required by regulations and scientific
necessity during the initial method develop-
+1.203.949.8697
www.ProtonOnSite.com
330  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

TABLE V: An elaborate forced degradation protocol used in an automated laboratory Thermal Degradation
The effect of temperature is studied by expos-
Stressing Condition Temperature (°C) Stressing Time (days)
ing the drug substance to high temperatures
Water 70 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 5, 7 in increments of 10 oC from 50 to 80 oC. Ther-
0.1 N HCl 70 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 5, 7 mal degradation is the primary degradation
0.1 N NaOH 70 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 5, 7 pathway of the DS in the solid state.
0.3% H2O2 Ambient 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 5, 7, 10, 14
Solution State Acid/Base Degradation
Cool White Fluores-
Ambient 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 5, 7, 10, 14 Solution-state degradation mechanisms may
cent Light (solution)
Cool White Fluores- be different, and should also be explored.
Ambient 0, 1, 2, 3, 5, 7, 9, 10, 11, 12, 13, 14
cent Light (solid) If the drug substance is not soluble under
Suntest—UV and aqueous conditions, then a small amount
Ambient 8 h (1xICH), 16 h (2xICH)
Visible Light (solid) (up to 10%) of organic solvent (acetonitrile
PH 2 Buffer 70 0, 1, 3, 5, 7 or methanol) can be used to solubilize the
PH 4 Buffer 70 0, 1, 3, 5, 7 DS, and the acid or base can then be added
PH 6 Buffer 70 0, 1, 3, 5, 7 to stress the samples. Attention should
be given to the functional groups pres-
PH 8 Buffer 70 0, 1, 3, 5, 7
ent in the drug molecule when selecting
PH 10 Buffer 70 0, 1, 3, 5, 7
a co-solvent (20).
Week 1: 60 It is not advisable to use a high concen-
In solid form at
Week 2: 70 0, 1, 3, 5, 7, 8, 10, 12, 14, 15, 17, 19, 21
ambient humidity
Week 3: 80
tration of acid or base, because doing so is
neither practical nor necessary. A few papers
Week 1: 60
In solid form at 75%RH Week 2: 70 0, 1, 3, 5, 7, 8, 10, 12, 14, 15, 17, 19, 21 have suggested a series of extensive steps
Week 3: 80 of pH changing from pH 2 to pH 10 buffer. It
may not be necessary as a routine practice,
unless for a specific DS or as part of a stress
Design of Experiments (DoE) to establish a
TABLE VI: Typical forced degradation studies for drug product of a solid-dosage form
stability model (21).
Forced Degradation Study Suggested Conditions
Heat Expose drug products at 50 ºC for up to 1 month Photostability
Expose drug products to 40 ºC and Photostability studies are essential, and
Humidity
75% RH and 25 ºC and 90% RH must be performed to generate primary
Photostability
Expose drug product in a petri dish without light-induced degradation products by
cover at twice or three times the ICH Q1B exposure level subjecting one representative batch of the
DS to the light exposure listed in ICH Q1B.
The standard light exposure is 1.2 million
of stable formulations and selection of in both solid and solution states. Table IV lux hours of visible light and 200 lux hours
suitable packaging lists an example set of stress conditions of near UV, with storage conditions con-
• to obtain samples to verify method for DS. An important goal of forced deg- trolled at room temperature.
specificity and structure elucidation of radation studies is to generate a potential These studies can be repeated when
significant degradation products level of degradation products that may there is a change of DS samples, such as a
• to allow the differentiation of impurities form during manufacturing and on stabil- new synthetic route, with different crystal-
and degradation products from the DS, ity storage. Typically, forced degradation linity form, different supplier, or a change
excipients, and other interference studies are conducted until a 5–20% loss of DP formulation.
• to facilitate the rapid establishment of of API is observed (18,19). This range is set During the method development phase,
CQAs for selection or elimination of to produce a reasonable amount of deg- forced degradation samples are typically
selected testing. radation products to facilitate method evaluated using a preliminary mass spec-
• to rapidly identify any excipient incom- development and to avoid secondary trometry (MS) compatible HPLC method
patibility issues with the DS degradation products from an unstable with a photodiode array detector (PDA)
• to generate potential data to support degradant (which are not observed under to collect information on peak area under
the justification of specifications. actual stability conditions). It should be the curve. The MS and PDA data are used
Forced degradation studies are con- noted that the conditions should be for peak tracking, peak purity assessment,
ducted under high temperature, humidity, selected based on the physicochemical identification of degradation products,
acid/base, oxidative, and light conditions properties of the individual DS and DP. impurities, and interferences (3,5,19).
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  331

Table V shows an example of a more extensive, elaborate forced these chambers be controlled within ±2 oC and ±5% RH for
degradation protocol used by a pharmaceutical laboratory to col- chambers operated at 25 °C/60% RH, 30 °C/65% RH, and 40
lect data for most of their NCEs using a four-temperature thermal °C/75% RH. The refrigerated condition is controlled at 5 oC ± 3
stressing condition matrix (ambient, 60 oC, 70 oC, 80 oC) for acid, oC (2 to 8 oC) with ambient humidity (monitored but not con-
base, oxidative, light, humidity, buffered solution, and prolonged trolled). These chambers are equipped with sensors for the
stressed duration (up to 21 days). This sample-intensive protocol continuous monitoring of temperature and relative humidity.
was made feasible by this centralized and automated laboratory These data are captured electronically or by chart recorders
specializing in supporting forced degradation studies for the with backup power. Excursions will be noted with an alarm
development site. The laboratory utilized automated sample stor- system. When the chambers are out of tolerance (see above),
age and retrieval system called “Powderium” (a powder dispens- investigations must be performed to identify the cause of the
ing workstation), robotics, programmable liquid handling, and issue and determine the impact of the stored samples (7).
refrigerated storage of quenched samples (18,19). The stability chambers are fully qualified and maintained, and
Table VI provides a summary of forced degradation studies per- their access is limited to authorized personnel to ensure that sam-
formed on solid dosage forms. Note that acid, base, and oxidative ple removal is tracked and controlled. Maintenance and service
exposures are not used, because these conditions are not repre- records must be available for inspection during audits. Downtime
sentative of realistic storage or exposure situations. The length of must be limited, and a backup plan must be developed to avoid
exposure should be evaluated carefully, depending on the dosage issues that can affect sample storage and data integrity. An annual
forms or the formulation. For example, capsule formulations or chamber inventory program is also necessary to track the sample
film-coated tablets may not be able to withstand a temperature storage and disposition at any time.
higher than 60 oC for an extended length of time.
Similar to the DS, a goal of 5% to 20% loss of active ingredient Stability Study Strategies
is recommended. If DS or DP is stable and the desired degrada- to Expedite “First-in-Human” Clinical Trials
tion level cannot be reached under high-stress conditions, stud- Minimizing “Time to Market” is an essential strategy for most
ies should be terminated. Overstressing a sample may lead to pharmaceutical companies (1). Given that by regulatory agencies
secondary degradation products that are not present in formal expect to see three months of stability data in IND filings to enable
stability studies. In contrast, understressing may lead to insufficient
degradation products for method development or identification
(12). For each condition of the forced degradation study, the chro-
matogram is compared with that of a control sample to document
the mass balance of the API (% loss of API) and % degradation
(normalized peak area %). For all conditions, the peak shapes
and peak purity (by PDA and MS) of all components should be
assessed for coelution or presence of interferences (3–5).
Accelerated and stress studies are performed to induce quicker
degradation than what would be observed in recommended stor-
age conditions. Results from these studies are used to estimate the
stability profile of the DS and DP at long term storage conditions,
establish the actual degradation pathways, and demonstrate the
intrinsic stability of the DS. It is recommended that these studies
be implemented under GMP conditions because the results of
these studies may be used as supporting information in regulatory
document packages. Full shelf-life studies are still needed to verify
the retest date of the DS and expiration date of the DP.
While discussions on forced degradation studies for NCEs in
NDA applications are generally plentiful, those for ANDA are often
insufficient and may result in regulatory deficiencies. For instance,
the impurity profiles may not be completely derived or sufficiently
discussed (20). In addition, the labeling for generic drug products
should be concordant with that of the reference listed drug (RLD)
and with a compendial monograph, as applicable (20).

Stability Storage
Stability samples are kept in a controlled temperature in humid-
ity stability chambers. The ICH Q1A (R2) guideline requires that
332  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
TABLE VII: Typical stability studies to support early-phase clinical trials
Intended Storage Conditions to Be Studied
25 ºC and 60% RH
Room temperature or 30 ºC and 65% RH
40 ºC and 75% RH
2–8 ºC
Refrigerated
25 ºC and 60% RH
TABLE VIII: Example of a simple bracketing design
Strength 50 mg
Batch A B
50 mL T T
Container
100 mL - -
size
250 mL T T
TABLE IX: Examples of factors to consider in bracketing
1. Container size
2. Container wall thickness
3. The surface area of the volume
4. The volume of the product
5. The weight of the product
6. The permeability rate of water per dosage unit
7. The strength of a similar drug/excipient drug ratio
Phase I clinical trials, it is customary to per-
form preliminary accelerated stability stud-
ies on the Good Laboratory Practices (GLP)
toxicological DS and prototype DP batches
to provide supporting data for quicker IND
submissions. Early-phase stability studies are
typically short, as shown in Table VII, and the
study conditions depend on where the clini-
cal trials will take place. During Phase I, the
analytical procedures are usually not fully vali-
dated, though specificity, accuracy, linearity,
and precision studies should be completed.
These time-saving strategies allow the phar-
maceutical companies to start Phase I clinical
trials as soon as possible after the nomina-
tion of the drug candidates.
Characterization studies of the API and DP
continue during formulation development. It
is recommended that the forced degrada-
tion studies be repeated before the release
testing and the initiation of the formal stabil-
ity studies under GMP regulations and ICH
Q1A (R2) guideline (8). Examples of standard
Most tools are science- and risk-based,
and have a robust statistical underpin-
Time Points (in Months) ning that conforms to the principles of
Time point zero, 1, 2, 3, 6, 9, QbD and DoE, which are endorsed by
and 12, or end of the study regulatory agencies for pharmaceutical
1, 2, 3, 6 development (9).
Time point zero, 1, 2, 3, 6, 9,
and 12, or end of study Modern Predictive
1, 2, 3, 6 Methods and Approaches
One widely used approach is the Acceler-
ated Stability Assessment Program (ASAP),
which uses a modified Arrhenius approxi-
mation. This approach uses degradation
100 mg 250 mg 500 mg
kinetics at a few stressed conditions in 1–2
C A B C A B C A B C
weeks, and extrapolates to lower tempera-
T - - - - - - T T T tures and humidity for longer timeframes.
- - - - - - - - - - The advantages of this strategy are that
T - - - - - - T T T the length of these experiments is short
and the experiments use fewer resources.
The accuracy of this model assumes a first-
order reaction, which is often the case for
hydrolysis-based degradation (6,22). The
disadvantage is that the presumed kinetic
model limits the degradation to less than
a percent and requires extrapolation of
three factors simultaneously: time, temper-
ature, and humidity. These disadvantages
are minimized when the DS and DP have a
simple kinetic pathway, and the primary sta-
bility drivers are temperature and humidity
(23). Another drawback could occur when
stability protocols that are listed for different the kinetic route is complex or nonlinear
dosage forms at different drug develop- or involves physical changes, which may
ment phases are discussed elsewhere (6,7). result in the linear model not being accu-
Results of the formal stability studies used to rate. Recently, this approach was also used
determine the drug products’ expiration are to select appropriate packaging based on
included in the stability section of the NDA its thermal and moisture-barrier character-
or ANDA. istics, and to optimize the drug:excipient
ratio of the formulation based on the
Predictive Tools to moisture content (22). This approach has
Establish Product Expiry gained popularity, mainly to set expiry for
Stability data are used to establish the clinical materials, and has been accepted
expiry of finished products according to by multiple regulatory agencies (24).
US FDA Good Manufacturing Practices Another successful tool is the Predic-
regulation 21 CFR 211.137 (21). The data tive Characterization Study (PCS), a rapid
from the accelerated condition are used to development of robust stability models
determine a tentative expiration date, but using a semi-empirical design space. It is
testing at full shelf life is necessary to con- a systematic study that evaluates the prod-
firm the approved expiry. uct data of samples exposed to a series
In this section, we discuss examples of storage conditions to build predictive
of useful software tools and statistical stability models to establish the design
approaches to evaluate stability data and space and control strategy based on the
to expedite stability studies for establish- QbD concept, such as setting specifica-
ing product expiry. tions and selecting proposed container–
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  333

closure systems. This approach is also found acceptable by many
regulatory agencies (23–25).
Improved modeling tools have enabled stability predictions
with accelerated timeframes compared to those of the traditional
extrapolation approach. Besides providing a more accurate way
to justify an expiry assignment, these models are used to increase
the understanding of the critical factors that influence the quality
of the DS and DP, as demonstrated in empirical science and risk-
based approaches detailed in ICH Q8–Q11. These approaches
are often utilized by large pharmaceutical companies with bet-
ter modeling resources and expertise. A regulatory template
has been shared to standardize on critical elements to use risk-
based predictive stability data for setting up shelf life to support
clinical development. (26)
Trending Analysis
The purpose of the stability program is to establish a retest period
for DS or shelf life for DP that will apply to all future commercial
batches. Stability data of CQAs of individual batches should
remain within specification throughout the assigned retest period
of the DS or shelf life of DP. Therefore, the trending of all stability
data is very important. Most degradation trends of API are linear;
however, the degradation pathway could be a linear, quadratic, or
cubic function. The typical ICH Q1E approach for the determina-
tion of shelf life through the analysis of data based on a quantita-
tive attribute that is expected to change over time is to determine
the period of time at which the 95% one-sided (or two-sided) con-
fidence limit for the mean curve intersects the acceptance crite-
rion (27,28). The annual commitment lots should also be evaluated
against the product trend to assure the consistency of the stability
profile of the DP and verify if there is any unintended change that
may impact the DP.
size) encompass and represent the stability behaviors of products of
the intermediate levels, eliminating the need for testing at the inter-
mediate levels. A bracketing schedule can be applied to multiple
strengths of identical or closely related formulations.
Table VIII shows an example in which the 100 mL container
size is bracketed by the 50 mL and 250 mL container sizes; thus,
stability samples stored in the 100 mL containers are not tested.
Similarly, the 100 mg and 250 mg tablet strengths are bracketed
by the 50 mg and 500 mg strengths. Assuming that the stability
profiles of these different strength formulations are the same
and the compositions of these tablets are similar, regardless of
strength, then the testing of 100 mg and 250 mg tablets may
not be needed. ICH Q1A (R2) requires that three batches (A, B,
C) be made available for submission. According to the guide-
lines, a pharmaceutical company could reduce its testing from
potentially 36 configurations to only 12 configurations, resulting
in significant savings of resources and time (6,7).
Table IX lists some factors to which a bracketing approach can
potentially apply, assuming that the container–closure system
and the headspace of the containers do not have any impact.
The disadvantage of the bracketing concept is that when one of
the results is out-of-specification; all bracketed configurations
data could then be at risk. Full testing would need to be acti-
vated mid-study, complicating the data set, and negating the
benefit of any testing reduction. In addition, all bracketed fac-

Reduced Testing with

Bracketing and Matrixing
Conducting stability programs is expensive and time consuming.
To reduce costs and increase the efficiency of the stability pro-
gram while maintaining regulatory compliance, many companies
employ bracketing or matrixing options that reduce the amount
of testing and resources required (6,7). ICH Q1A (R2) and ICH
Q1D are guidelines to refer to when reduced testing is applied.
Although the bracketing and matrixing concepts are included
in multiple global stability guidelines, limited numbers of pub-
lished applications can be found. It is even more challenging to
provide examples for the matrixing for an NCE. Here, we explain
the similarities and differences between these approaches to
illustrate this application.

Bracketing
Bracketing refers to a study design in which only the extreme vari-
ables, such as extreme strengths, container sizes, or container fills,
are tested. The plan assumes that the stability behaviors of prod-
ucts manufactured or packaged at these extreme levels (such as, for
example, highest vs. lowest strength, or largest vs. smallest package
334  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

TABLE X: An example of a two-thirds factorial matrixing design

annual product monitoring stability stud-
ies rather than for DPs from NCEs.
Timepoint Months on stability
Batch 0 3 6 9 12 18 24 36 Matrixing
A T T - T T - T T Matrixing is a statistical design of the sta-
Strength 1 B T T T - T T - T bility schedule in which a selected group
C T - T T T T T T of samples of the total number of pos-
sible samples is tested at a specific time
A T - T T T T T T
point. At the next time point, a different
Strength 2 B T T - T T - T T
group should be tested. At each time
C T T T - T T - T
point, the stability data of that group of
T = for stability testing scheduled for that timepoint samples will represent the stability of the
whole set of studies.
TABLE XI: Examples of factors for which matrixing can be considered Similar to the bracketing approach, this
design also assumes that all the batches
have similar stability profiles; thus, there
1. Different batches is no need to generate all the results. The
2. Different concentrations
3. Different filling long-term trends of all configurations are
4. Different container size approximately linear across the whole set
5. Different closure composition of all studies, and the comparative sta-
6. Different closure system
bility of each presentation can be evalu-
ated with the reduced schedule. Many
options are available with the matrixing
tors (for example strengths/formulations if only a subset will be commercialized. concept, which could reduce the testing
and package configurations) will need to Due to the risk, the bracketing concept schedule by a third, a half, or two thirds,
be registered in all end markets, even is used more often for post-approval or depending on the aggressiveness of the

TABLE XII: Example of a stability data record (6)

STABILITY ANALYTICAL RECORD

Sample Name: Manufacturing Date:
Storage condition:
Lot#: Manufacturing Site:
Sample Orienta-
Study #: Expiration Date:
tion (if applicable):
Protocol #: Testing Site:
Packaging Information:
Study Start Date: Packaging Site:
Packaging Date:
Study Purpose:
Acceptance Time
Test Name Method 1 Mo 2 Mo 3 Mo 6 Mo 9 Mo 12 Mo
Criteria Zero
Pull Date
Test Date
Appearance
Assay
Impurities
Individual
Total
Dissolution
Average
% RSD
Range
Moisture
Completed By: Date:
Reviewed By: Date:
Approved By: Date:
WWW.CHROMATOGRAPHYONLINE.COM
statistical extrapolations and risk assess-
ment for the design (6,7).
Table X shows an example of a two-
thirds factorial matrixing design. In this
schedule, there are two strengths and
three batches made of each strength.
All presentations are tested at time zero,
12 months, and the end of the study (36
months) for the highest possible precision
as these time points are critical to the DP
stability profile. For all other time points (3,
6, 9, 18, and 24 months), only two-thirds
of the presentations get tested at a rotat-
ing schedule. Therefore, all configurations
are tested at different time points for the
same amount of times in total.
The matrixing approach requires that
all samples be placed in stability cham-
bers, including at the timepoints that
are reduced. Therefore, when there is an
instability issue, the testing schedule can
be changed to test all configurations. If
some data points are not available, there
will be a minimum impact on the entire
study of multiple presentations.
Table XI lists some examples of factors
where matrixing can be used to reduce
testing. Matrixing also provides more
flexibility because different reduced
testing options can be performed for
various tests depending on the test-
ing confidence. Realistically, matrixing
should be used only for label storage
conditions. The accelerated stability data
may provide some insight into the sta-
bility profile of the product at the label
storage conditions. Because matrixing is
a statistical concept, evaluation of data
can be more involved, and assessment
of all the presentations as a complete
set is more complicated, which explains
why this approach is not used widely.
These reduced testing approaches are
discussed in detail in many global guide-
lines; however, a lack of knowledge of
matrixing among regulatory reviewers
and internal quality and regulatory pro-
fessionals may hinder the use of this
application to reduced stability testing.
Stability Reports and
Regulatory Submissions
Stability reports are required in all phases
of regulatory submissions regardless of
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  335
TABLE XIII: Questions about the conclusions of a stability report
► What is the stability profile of the product?
► Are all lots within the stability trend of the product?
► Can differences be attributed to a particular factor such as packaging or strength?
► Do all tests meet specifications?
► Is there any test closely monitored? Are there any tests to be performed
more often?
► Is there any package closely monitored? Is there any package to be tested
more often?
► Do all tests meet specifications?
► Do the data support the expiration date?
► Are there any stability trends?
► Discuss any out-of-specification investigation.
► Discuss any statistical evaluation of the data set.
►  ould the stability database be satisfactory to the proposed or approved expiration
W
dating?
lifecycle state (such as NDA, ANDA, or according to ICH Q1E or statements
supplement NDA). The stability data- that no overall trends and little variabil-
base comprises laboratory data gen- ity are observed in the dataset. Table
erated and stored electronically or as XIII lists typical questions that should
printed reports. These stability data are be addressed in the stability conclusion.
entered into stability reports, which are “Poolability” of different batches should
essential to communicate information be evaluated to determine the consis-
internally for audit purposes, product tency of the representative stability data
investigations, justification of specifica- and trends with respect to the manufac-
tions, to support product development, turing process and analytical method,
or simply as a communication tool. All by comparing the intercepts and the
detailed information in the report must slopes of all batches (6,7,24).
be verified for accuracy. For an NDA or
ANDA submission, the stability report is Summary and Conclusions
typically the most extensive non-clinical A robust and science-based stability
section to summarize the stability data program is required for the registration
of the product during its shelf life (6,28). and commercialization of any pharma-
Stability data can be presented in a tab- ceutical product. Regulations in this
ular format, in a graphical representation, area are well established; however, the
a narrative, or a combination of formats. application and interpretation of the
Table XII is an example of a stability data regulations vary between companies
record. The data tables can be generated and regions. ICH harmonizes the main
manually in a document or printed out regions to help companies use a stan-
from a laboratory information manage- dardized approach.
ment system (LIMS) or stability manage- This article provides a high-level
ment software. All information and data summary of regulations and regula-
must be reviewed and verified. The con- tory expectations of stability programs.
clusion of the stability report discusses It reviews the goals and practices of
whether the data indicate that the prod- forced degradation studies to generate
uct continues to meet the quality specifi- sample data to confirm method speci-
cations or acceptance criteria established ficity. A strategy for conducting accel-
for the study and whether the data sup- erated stability studies on preliminary
port the proposed or approved expiration batches to expedite initial regulatory
dating period (6,7). filing is described. Finally, predictive
Stability reports should include a sta- stability software-based and statistical
tistical evaluation of the currently avail- approaches such as trending, bracket-
able data and a discussion of the results ing, and matrixing are explained. 
336  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
Acknowledgments
The authors express their gratitude to
the following colleagues for their time
and efforts to provide timely reviews of
the manuscript to improve content accu-
racy and clarity: Jane Weitzel of Zinata,
Jing Capucao of J&J, Mike Shifflet of
Johnson and Johnson Consumer Health
Care, Adrijana Torbovska of Farmahem,
Alice Krumenaker of TW Metals, LLC;
Mark Shapiro of MCS Pharma Consult-
ing, Madhavi Mahavadi of BioMarin
Pharmaceuticals, and Joshua Ayers of
ASQ Solutions.
References
(1) R.G. Hill and H.P. Rang, Eds., Drug Dis-
covery and Development: Technology in
Transition (Churchill Livingston, Elsevier,
Edinburgh, Scotland, 2nd Ed., 2012)
(2) M.W. Dong, Drug Development Process,
Short Course presented at Pittcon, Phila-
delphia, Pennsylvania, March 2019.
(3) M.W. Dong, LCGC North Am. 33(11),
764–775 (2015).
(4) D. Kou, L. Wigman, P. Yehl, and M.W.
Dong, LCGC North Am. 33(12), 900–909,
(2015).
(5) M.W. Dong, HPLC and UHPLC for Practic-
ing Scientists (John Wiley & Sons, Hobo-
ken, New Jersey, 2nd Ed., 2019), Chap-
ters 9–11.
(6) K. Huynh-Ba, Ed., Handbook of Stabil-
ity Testing of Pharmaceutical Products:
Regulations, Methodologies and Best
Practices (Springer, New York, 2009).
(7) K. Huynh-Ba, Ed., Pharmaceutical Stabil-
ity Testing to Support Global Markets
(Springer, New York, 2010).
(8) International Council for Harmonisation
of Technical Requirements for Pharma-
ceuticals for Human Use (ICH) Q1A (R2),
Stability Testing of New Pharmaceutical
Products (Geneva, Switzerland, 2003).
(9) International Council for Harmonisation
of Technical Requirements for Pharma-
ceuticals for Human Use (ICH) Q8(R2),
Pharmaceutical Development (Geneva,
Switzerland, 2009).
(10) International Council for Harmonisation
of Technical Requirements for Pharma-
ceuticals for Human Use (ICH), Q3A(R2),
Impurities in New Drug Substance,
Q3B(R2), Impurities in New Drug Prod-
ucts (Geneva, Switzerland, 2006).
(11) International Council for Harmonisa-
tion of Technical Requirements for
Pharmaceuticals for Human Use (ICH)
Q6A, Specifications: Test Procedures
and Acceptance Criteria for New Drug
Substances and New Drug Products,
(Geneva, Switzerland, 1999).
(12) H. Bruemmer and N. Belikova, Pharm.
Technol. 41(9), 46–52 (2017).
(13) 21 Code of Federal Regulations (CFR),
Part 211.194(a), Current Good Manufac-
turing Practice for Finished Pharmaceu-
tical Products (US Government Printing
Office, Washington, DC, 2019).
(14) United States Pharmacopeia Conven-
tion, USP 37/NF 32, General Chapter
USP <1225>, Validation of Compendial
Methods, (United States Pharmacopeial
Convention, Rockville, Maryland, 2013).
(15) United States Pharmacopeia Conven-
tion, USP 37/NF 32, General Chapter USP
<1226>, Verification of Compendial Pro-
cedures (United States Pharmacopeial
Convention, Rockville, Maryland, 2013,
revised 2018).
(16) International Council for Harmonisation
of Technical Requirements for Pharma-
ceuticals for Human Use (ICH) Q2 (R1),
Validation of Analytical Procedures:
Methodology (Geneva, Switzerland,
November 1996, updated 2015).
(17) S.W. Baertschi, K.M. Alsante and R.A.
Reed, Eds., Pharmaceutical Stress
Testing: Predicting Drug Degradation
(CRC Press, Boca Raton, Florida, 2nd
Ed., 2011).
(18) H.T. Rasmussen, W. Li, D. Redlich, M.I.
Jimidar, HPLC Method Development, In
Handbook of HPLC in Pharmaceutical
Analysis, S. Ahuja and M. W. Dong, Eds.,
(Elsevier, Amsterdam, 2005), Chapter 6.
(19) M.W. Dong and H.T. Rasmussen, HPLC
Method Development Short Course,
Eastern Analytical Symposium, Somer-
set, New Jersey, 2004.
(20) R. Maheswaran, Pharm. Technol. 36(5),
1–6 (2012).
(21) 21 Code of Federal Regulations (CFR),
Part 211.137, Current Good Manufactur-
ing Practice for Finished Pharmaceuti-
cal Products (US Government Printing
Office, Washington, DC, 2019).
(22) K. Waterman, Understanding and Pre-
dicting Pharmaceutical product Shelf life,
In Handbook of Stability Testing of Phar-
maceutical Products: Regulations, Meth-
odologies and Best Practices (Springer,
New York, 2009), Chapter 6.
(23) K. Huynh-Ba et al., Meeting Report,
Analytical Approaches to Ensure Prod-
uct Quality – AAPS Joint Face-to-Face
Meeting of the Stability, the Pharmaceu-
tical Impurities, and the CMC Statistics
Focus Group, AAPS Open 3 (1), 2017.
https://aapsopen.springeropen.com/
articles/10.1186/s41120-017-0011-z
(24) H. Williams et al., Pharm. Technol. 41(3),
52–57 (2017).
(25) D. Lavrich, Rapid Development of
Robust Stability Models Using Semi-
Empirical Design Space, AAPS Webi-
nar, March 24, 2016. https://aapsopen.
s pring eropen.com /ar ticles/10.118 6/
s41120-017-0018-5#article-info
(26) H. Williams, Pharm Technol. 42(8),
42– 47 (2018).
(27) WHO Expert Committee on Specifica-
tions for Pharmaceutical Preparations
Fifty-second report, annex 10, Stabil-
ity testing of ac tive pharmaceutical
ingredients and finished pharmaceuti-
cal products (World Health Organiza-
tion, 2018).
(28) International Council for Harmonisa-
tion of Technical Requirements for
Pharmaceuticals for Human Use (ICH)
Q1E, Evaluation of Stabilit y Data
(Geneva, Switzerland, 2003).
ABOUT THE CO-AUTHOR
Kim Huynh-Ba
is the managing director
of Pharmalytik LLC. (www.
pharmaly tik.com), which
provides consulting ser-
vices in Stability Sciences, Quality
Management Systems, and Analytical
Development. She is an Adjunct Pro-
fessor at Temple University’s School
of Pharmacy and Illinois Institute of
Technology (IIT). Kim is a member of
the US Pharmacopeia’s Council of
Experts, and the chair of the Chemical
Medicines Monograph IV Expert Com-
mittee, USP Good Documentation
Practices Expert Panel, USP Organic
Impurities of Drug Substance and
Drug Products Expert Panel. She is on
the editorial board of AAPS Open, the
Journal of GXP Compliance, and the
Journal of Validation Technology. Kim
has authored ~30 technical publica-
tions and is the editor of two books on
stability testing to support registration
in global markets.
ABOUT THE AUTHOR
Michael W. Dong
is a principal of MWD
Consulting, which provides
training and consulting
services in HPLC and UHPLC,
method improvements, pharmaceutical
analysis, and drug quality. He was formerly
a Senior Scientist at Genentech, Research
Fellow at Purdue Pharma, and Senior
Staff Scientist at Applied Biosystems/
PerkinElmer. He holds a PhD in Analytical
Chemistry from City University of New York.
He has more than 100 publications and
a best-selling book in chromatography.
He is an editorial advisory board
member of LCGC North America and
the Chinese American Chromatography
Association. Direct correspondence to:
LCGCedit@MMHGroup.com
338  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

FOCUS
ON
BIOPHAR M ACEU TIC AL
ANALYSIS

Multidimensional Separation Techniques

for Characterization of Biotherapeutics
Multidimensional separations aim to combine two or more separation methods, and these are increasingly being employed to
overcome the limitations of one-dimensional separations. The outcome typically includes higher peak capacity and improved
selectivity. Such an approach can be invaluable when analyzing complex mixtures. In this article, we discuss some commonly used
two-dimensional systems, their pros and cons, and their applications in the biopharmaceutical industry.

Anurag S. Rathore, Ramesh Kumar, and Ira S. Krull

T he quality of a biotherapeutic is
defined by its various biological,
immunological, structural, and physi-
cochemical attributes. Some of these
are post-translational modifications
(PTMs), other product-related het-
erogeneities, process related impu-
rities, and host cell related impuri-
ties. An array of different techniques
contributes to the analytical toolbox.
These are utilized for characterization
of the biotherapeutic product, and
especially the myriad of other species
that are also present in the final drug
product (1–3). One of the most com-
monly used analytical techniques for
routine process evaluations and qual-
ity control in the industry is reversed-
phase liquid chromatography-mass
of large proteins may generate more
than 100 peptides. Given the random
distribution of peaks, even a system
with a peak capacity of 10,000 would
be able to resolve only 98% of the mix-
ture of 100 tryptic peptides (6). Thus,
in the case of biopharmaceuticals,
which are mostly large molecules, use
of one-dimensional ( 1D) chromatogra-
phy usually does not yield complete,
resolved product coverage (6,7).
Multidimensional (MD) separation
systems, in which two or more sepa-
ration methods are coupled, can be
used in such cases to increase the
peak capacity of the system. This will
then allow the complete resolution, it
is hoped, of all constituent analytes
(8). These systems improve sample
Based on how the eluate is trans-
ferred from one dimension to another,
MD separations can be classified as
on-line and off-line. The MD system is
said to be on-line when the eluate of
one dimension is directly transferred
to the next dimension, and separation
occurs in real time in both the dimen-
sions. On the other hand, in the case
of off-line systems, eluates from the
first dimension are collected as frac-
tions, and then later subjected to sep-
aration in the next dimension (9,12).
Both approaches have some advan-
tages and disadvantages (Table I).
Additionally, MD approaches can
also be classified as heart-cutting or
comprehensive ( 1 D× 2 D), depending
on the part of the 1 D elution con-
Icon Image: (Columns) Joe Zugcic / Zugcic Photographers, Inc.

spectrometry (LC–MS/MS). Due to its
robustness, reproducibility, and easy
coupling with the MS, LC–MS has
become the state-of-the-art technol-
ogy for routine characterization of
biotherapeutics in the industry (3,4).
However, LC-based separations suf-
fer from a significant limitation aris-
resolution manifold by first resolving
the components in the first dimension.
The eluate is then transferred and fur-
ther resolved in the second or subse-
quent dimensions (9). Such a separa-
tion was first reported with the use of
two-dimensional (2D) paper chroma-
tography in 1944 (8). Since then, differ-
tent that is transferred in the second
dimension. In the heart-cutting MD
approach, only a part of the 1 D elu-
ate is subjected to separation in the
next dimension. However, in a com-
prehensive approach, the whole elu-
ate of the 1 D step is subjected to a
2D separation (9).

ing from incomplete separation and
co-elution of target analytes (5). One
reason for lack of baseline resolution
is that LC can generate peak capaci-
ties up to 1000, and proteolytic digest
ent separation techniques have been
coupled in a two-dimensional or mul-
tidimensional manner, and have been
proven useful for the analysis of com-
plex biological samples (9–11).
Whether on-line or off-line, an MD
separation system is useful if it fulfills
two fundamental requirements, as
suggested by Giddings (8). First, the
separation techniques in the MD set
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  339

up must be orthogonal. This means TABLE I: Advantages and disadvantages of on-line and off-line coupling
that the resolving principles of the
Parameter On-line Off-line
separation techniques must be based
on different chemical or physical prop- Speed Fast (12) Slow (12)
erties of the analytes. Thus ideally, for Automation Feasible (10) Difficult (10,13)
a given analyte, the peak distribution Mobile phase compositions Limited options (12) More options (12)
in the two dimensions must be uncor- Optimization Difficult (12) Easy (12)
related. The second principle states Sample manipulation
that the resolution of the first dimen- Not allowed (10) Possible (10)
between dimensions
sion must not be lost during separa- Resolution Less (9) More (9)
tion in subsequent dimensions (8,12).

Multidimensional LC–LC Coupling
One of the most successful MD sys-
tems, widely accepted by both aca-
demia and industry alike, is that of
2D–LC-based separations (14). These
separations involve separation of com-
pounds at two different conditions,
using either different modes of LC or
using a single-mode operated at two
different conditions. Many LC modes,
such as ion exchange chromatography
(IEC), size exclusion chromatography
(SEC) and reversed-phase LC, have
been combined. Such couplings have
been successfully implemented for the
analysis of biopharmaceuticals (11,15).
However, not all LC modes can be
combined with each other, due to sol-
vent immiscibility and other compat-
ibility issues (16). These 2D-LC systems
have been widely accepted by the bio-
pharmaceutical industry. Some of the
applications of these techniques in the
industry are summarized in Table II.
Some of the most commonly used,
2D-LC methods, which have been
applied for the characterization of bio-
therapeutics, are discussed below. In
all these methods, reversed-phase LC
has been used for the second-dimen-
sion separation, likely because of the
high-quality reversed-phase LC col-
umns available and their compatibility
with MS (34).
Reversed-phase LC x
Reversed-phase LC Coupling
This separation system involves resolv-
ing the analytes via reversed-phase

ADVANTAGE
See What It Can Do for You and Your Lab

Sign up today to access Restek’s

years of chromatography knowledge at
www.restek.com/advantage

Pure Chromatography
340  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

TABLE II: Biopharmaceutical applications for multidimensional separation techniques

1st Dimension 2nd Dimension Detectors Applications References

HIC RPLC MS Product quality of antibody–drug conjugates (17)
SEC WCX DAD Aggregate analysis and charge variant profiling (18)
(AEX and
RPLC CAD and MS Polysorbate stability in mAb formulations (19)
RPLC)/CEX
IEC/SEC RPLC MS Charge and size variants of mAb (20)
HIC Native SEC MS Drug to antibody ratio of ADCs (21)
HIC RPLC MS Structural isoforms and PTM analysis of ADCs (21)
SEC RPLC MS mAb degradation products (22)
SEC RPLC DAD/MS Impurity and stability analysis of ADCs (23)
CEX RPLC MS mAb isoform and subunit analysis (24)
SCX RPLC DAD and MS Peptide mapping of mAb (6)
HILIC RPLC DAD and MS Peptide mapping of mAb (6)
RPLC RPLC DAD and MS Peptide mapping of mAb (6)
RPLC RPLC MS Host cell protein analysis (25–27)
SEC iCIEF CCD Charge heterogeneity of mAb heavy and light chains (28)
RPLC/IEX CGE PDA Disulfide isomer heterogeneity in mAb (29)
RPLC CZE UV and LIF Peptide mapping of an immunoconjugate (30)
RPLC CZE MS Peptide mapping of mAb (31)
CIEF CZE MS Charge variant analysis of mAb (32)
CZE CZE MS Charge variant analysis of mAb (33)
Abbreviations: ADC: antibody drug conjugate; AEX: anion exchange chromatography; CAD: charged aerosol detector; CCD: charge coupled device; CEX:
cation exchange chromatography; CGE: capillary gel electrophoresis; CIEF: capillary isoelectric focusing; CZE: capillary zone electrophoresis; DAD: diode ar-
ray detector; HIC: hydrophobic interaction chromatography; HILIC: hydrophilic interaction liquid chromatography; iCIEF: imaged capillary isoelectric focusing;
IEC: ion exchange chromatography; LIF: laser induced fluorescence; mAb: monoclonal antibody; MS: mass spectrometer; PDA: photodiode array detector;
PTM: post translational modification; RPLC: reversed-phase liquid chromatography; SEC: size exclusion chromatography; SCX: strong cation exchange chro-
matography; UV: ultraviolet; WCX: weak cation exchange chromatography.

The impact of the use of different

TABLE III: Advantages and limitations of commonly used 2D separation techniques
cell lines and purification strategies
Techniques Advantages Limitations on the composition of HCPs in bio-
SCX-RPLC
Most commonly Low peak capacity, poor recovery pharmaceuticals was analyzed via an
used method (36) of hydrophobic peptides (36)
on-line, comprehensive 2D-LC–MS/
RPLC-RPLC High peak capacity (37) Only partially orthogonal (45) MS analysis. Six PTG1 mAb samples
HILIC-RPLC High orthogonality (36) Peptide precipitation (35) were trypsinized and characterized
SEC-RPLC High orthogonality (37) Low peak capacity (37) using MS with high pH reversed-phase
IEC-RPLC High orthogonality (37) Low peak capacity (37) LC (pH 10) in the first dimension, and
then with low pH reversed-phase LC
High orthogonality, peak Eluate and injection volume
LC-CE (pH 2.5) in the second dimension.
capacity, and resolution; fast (8) disparity, difficult coupling (9)
Abbreviations: SCX: strong cation exchange RPLC: reversed phase liquid chromatography HILIC: hydrophilic in- The 2D-LC system was quite stable,
teraction liquid chromatography SEC: size-exclusion chromatography IEC: ion exchange chromatography LC: liquid with excellent reproducibility of frac-
chromatography CE: capillary electrophoresis.
tionation in the first dimension, and
overall retention times in the second
LC in both dimensions, but at two dif- This difference in peptide retention dimension. A reproducibility of 0.1%
ferent pH conditions. The principle is mechanisms is utilized to constitute in retention time was obtained over
that at high pH (such as pH 10), pep- the two dimensions of this separation 48 h of separation, and the reproduc-
tide retention on the reversed-phase system (35). For instance, researchers ibility of the chromatographic setup
LC column is based on solvophobic have reported an on-line 2D-LC–MS/ was independent of the number of
interactions, while at low pH, both MS approach for a comprehensive fractionations in the first dimension. A
solvophobic and electrostatic inter- analysis of host cell proteins (HCPs) total of 33 HCPs were identified over 5
actions determine peptide retention. in monoclonal antibodies (mAbs) (27). orders of magnitude, with the lowest
WWW.CHROMATOGRAPHYONLINE.COM
concentration at 16 ppm. To decrease
analysis time and improve peak shape,
the second dimension was operated
at a high flow rate and high tempera-
ture, which then led to a 15% loss in
peak capacity as compared to that
obtained with standard LC conditions
(27). Limited orthogonality is, how-
ever, a major issue of this approach
(36). Column stability and possible
erosion of fused silica tubes at high
pH are some of the other factors that
must be considered before setting up
this separation system (35).
HILIC x Reversed-Phase LC Coupling
One of the best orthogonalities dem-
onstrated is exhibited by the hydro-
philic interaction chromatography
(HILIC) x reversed-phase LC combina-
tion (36). In the first dimension, HILIC
separates analytes mainly based on
their separation between the mobile
phase and water-enriched layer
adsorbed onto the stationary phase.
In the second dimension, reversed-
phase LC separates them on the basis
of their hydrophobicity. The use of this
setup has been recently reported in a
study in which a commercially avail-
able 2D-LC system was used for the
bottom-up analysis of trastuzumab.
A novel dual loop interface was used
for transferring the samples between
the two dimensions. The use of high
elution strength and a high volume of
injection solvent, however, led to com-
plete or partial elution of polar pep-
tides in the void volume. Several pre-
cautions, such as high flow rate and low
temperature of the second dimension,
were taken to minimize peptide break-
through in reversed-phase LC, but it
could not be completely eliminated.
Unfortunately, the HILIC x reversed-
phase LC combination was challeng-
ing, and offered limited coverage of
the separation space as well. The study
compared HILIC x reversed-phase LC
with strong cation exchange (SCX)
x reversed-phase LC and reversed-
phase LC x reversed-phase LC, and it
found that the other two techniques
are superior for the separation of tryp-
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  341
tic peptides of trastuzumab. Although
the system offers high orthogonal-
ity, concerns persist due to the pres-
ence of a high organic content in the
mobile phases (35).
SEC x Reversed-Phase LC Coupling
Size-exclusion chromatography (SEC) x
reversed-phase LC separates analytes
on the basis of size and hydrophobic-
ity in the first and second dimensions,
respectively. The system is thus highly
orthogonal, and it is useful for peptide
separations. Synthetic polymers and
biopolymers are some of the major
application areas of this system (37).
A 2D-LC–MS based, semi-automated
platform was developed for charac-
terization of impurities in monoclonal
antibodies, with a week-long data
collection time. Breakdown products
were identified in stress induced mAb
samples, using SEC–reversed-phase
LC–MS/MS analysis. The low molecu-
lar weight breakdown products were
enriched by fractionation via SEC in
the first dimension, and the fractions
were resolved via C8 chromatography
in the second dimension. Both on-line
and off-line mass spectrometry analy-
ses were performed, using the C8 elu-
ate. A part of the C8 eluate was thus
directed to the MS while a major part
was collected as fractions. While on-
line analysis was used for intact mass
analysis, the off-line set up included
analysis under both reducing, as well
as non-reducing conditions, so as to
elucidate impurity structure. Various
anticipated products, as well as unpre-
dicted modifications, such as frag-
ment antigen-binding (Fab) fragments
and hydrolyzed, free light chain, were
identified. The approach was also pro-
posed to be useful for charge variant
and aggregation analysis of mAbs.
IEC x Reversed-Phase LC Coupling
In this mode of duality HPLC, ion
exchange chromatography (IEC) is
used in the first dimension, which
separates molecules based on their
charge. Analytes are transferred from
IEC in the first dimension to reversed-
phase LC in the second dimension,
as they are eluted with the change in
the pH or ionic concentration of the
mobile phase. The advantage of using
this system is its ease of use, as well
as the minimal sample manipulation
requirements. Both cation-exchange
and anion-exchange chromatography
have been coupled to reversed-phase
LC to separate protein mixtures and
tryptic digests (3,37). For example,
rapid analysis of charge variants of
mAbs without much sample manipu-
lation was performed, with a generic
2D-LC–MS method. Two independent
gradients were used on a single chro-
matographic system, consisting of
two independent column compart-
ments, with built-in switching valves.
Reversed-phase trap cartridges were
used in the second dimension for
desalting first dimension fractions
before MS analysis. These cartridges
enabled fast conditioning and re-
equilibration of columns, with minimal
backpressure, during the switching of
columns. The poorly resolved and low
abundance peaks of IEC in the first
dimension, were preconcentrated on
the reversed-phase traps using mul-
tiple injections. Using an in-process
sample of a mAb, fractions of interest
were collected from IEC, and then sub-
sequently analyzed via reversed-phase
LC, combined with an on-line elec-
trospray ionization (ESI)-time-of-flight
(TOF-MS) based detection. Chain spe-
cific information about mAbs was also
deduced by a disulfide reduction step.
This was incorporated as an optional
step into the workflow in an on-line
manner. Contamination, sample loss,
and degradation were minimized with
the use of an on-line setup. The whole
system was fully automated to allow
for high throughput analysis, minimum
sample handling, and rapid analysis.
The system was shown to be useful
for charge, sequence, and chemically
unstable mAb variant analysis. Minor
glycoforms were also identified at the
intact protein level, due to the sepa-
ration of charged isoforms in the first
dimension by IEC (20).
342  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

in order to hasten the separation

process, often deteriorate the over-
(a) all, analytical performance of the
500
2D setup (27).
mAU (214 nm)

400

300
Multidimensional LC-CE Coupling
Due to the limited orthogonality and
200
other disadvantages associated with
100
LC modes in the second dimension,
0 alternative separation techniques,
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
Time (min) having different separation principles,
(b) x10 7 have been attempted to improve the
1
3
2.8 separation further. Capillary electro-
2.6
2.4
phoresis (CE) is an attractive alterna-
Counts (x107)

2.2
2
1.8 tive in this regard, because it separates
1.6
1.4 analytes based on their charge to size
1.2
1
0.8
ratios, complementary to all modes
0.6
0.4 of LC (10,39). Electrophoresis with
0.2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
chromatography has been described
Acquisition Time (min) as a powerful 2D-separation method
by Giddings (9). The use of CE as the
second dimension in 2D systems can
FIGURE 1: Off-line LC-CE-MS of a protein digest. A separation of a protein digest was performed thus provide high efficiency and rela-
using reversed-phase LC in the first dimension and capillary electrophoresis (CE) in the second tively fast separations as compared to
dimension. (a) Separation profile of the digest with reversed-phase LC in the first dimension. Sol- LC-based methods. However, special
vent A and solvent B were 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respec-
tively. Overlapping peaks at 38–39 min are circled in green. (b) CE–MS profile of the fraction,
coupling techniques are required for
containing the peaks circled in green in panel a, in the second dimension. CE separation was CE based setups, due to the use of
performed in a 56 cm long polyvinyl alcohol coated capillary at 30 kV. Background electrolyte was small volumes and high electric fields
2% acetic acid and detection was performed at 214 nm. (13). The history of coupling chroma-
tography and electrophoresis dates
The coupling of SCX in the first or less correlated selectivity, and a back to 1948, and since then the cou-
dimension and reversed-phase LC in lower than maximum peak capacity pling has been shown to improve the
the second is the most widely used is usually achieved (10,36). Another resolution of various biological and
2D-LC system, especially in bottom- major limitation of the 2D–LC system environmental samples (9).
up proteomics-based studies. Ana- is that the separation processes have Various methods based on LC–
lytes are separated on the basis of longer analysis times. Various meth- CE, CE–LC, and SEC–CE have been
their charge in the first dimension, ods to speed up the LC-based separa- reported in the literature, primarily for
and hydrophobicity in the second tions via altering column length, tem- proteome and peptide level analysis.
dimension (35). This method, however, perature, and particle size have been An in-depth review of the applica-
suffers from a significant limitation reported. However, each of these tions of different modes of CE in MD
in terms of the peak capacity in the methods has its own limitations (37). In separations has been recently pub-
first dimension. The SCX chromato- 2D-LC, for example, to achieve faster lished (13). The ability of CE to resolve
gram is sparsely populated with posi- separation, the second dimension LC LC peaks has been discussed in the
tively charged tryptic peptides, most column is run at high flow rates, which literature (40), and is demonstrated
of which (+2 and +3 charged pep- compromises resolution. The second in Figure 1. In this work, the tryptic
tides) tend to be eluted close to each dimension can also be made faster digest of a protein was resolved using
other. Another major limitation of this by operating the column at high tem- reversed-phase LC in the first dimen-
method is the limited orthogonality of perature, which reduces the viscosity sion, then fractionated, with each
the system (about 50%). Furthermore, of the mobile phase, and allows the fraction subjected to CE–MS analy-
hydrophobic peptides may be eluted user to increase the eluent flow rate, sis in the second dimension. Several
with a poor peak shape, or may be but this may affect the stability of overlapping peaks were observed in
incompletely recovered with SCX (36). the stationary phase and the analyte LC that were efficiently resolved by
The major limitation with these sys- (38). Thus, the use of high tempera- CE in the second dimension (Figure
tems, however, is that they show more ture and other operating parameters, 1). For instance, the peaks from 38
WWW.CHROMATOGRAPHYONLINE.COM
to 39 min (circled in green in Figure
1a), remained unresolved in reversed-
phase LC, as observed in the corre-
sponding UV chromatogram. CE–MS
of the corresponding fraction showed
baseline resolution of the same peaks
(Figure 1b).
Although these reports clearly
demonstrate the utility of LC–CE in
biotherapeutics, LC–CE suffers from
significant limitations in terms of injec-
tion and elution volume disparities,
between the LC and CE dimensions.
There is also the need to isolate CE
from the other parts of the system, due
to its high electric field. Due to these
reasons, LC–CE has still not gained as
much attention as 2D-LC (9). A com-
parative analysis of the advantages
and disadvantages of the commonly
used MD methods is given in Table III.
Nevertheless, the coupling of LC with
CE has been described for analysis
of natural compounds, peptide frag-
ments, tryptic peptides of a protein
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  343
mixture, immunoconjugates of a mAb,
and disulfide isomer heterogeneity of
IgG2-mAbs (9,10,40–43). For example,
a size exclusion chromatography-
imaged capillary isoelectric focusing
(SEC-iCIEF) method was proposed for
the quantitative charge heterogeneity
analysis of heavy and light chains of a
mAb. Over 10 different mAbs, includ-
ing both IgG1 and IgG4 subtypes,
were characterized using this platform.
The platform was found to be less
time-consuming, less labor-intensive,
more reproducible, and allowed fully
quantitative analysis of charge variants
compared to the traditional 2D poly-
acrylamide gel electrophoresis (PAGE)
analysis. Results were obtained in a sin-
gle day, but only with automated SEC
and iCIEF platforms. The method was
also used for characterization of mAbs
under accelerated stability conditions,
and it was proposed to be useful for
generating simplified electrophero-
grams for conjugated antibodies. It
was also useful for obtaining a deeper
understanding about the conjugation
process. Although this method was
well suited for analyzing reduced and
denatured mAbs, and offered several
advantages over 2D PAGE, it required
large amounts of protein sample
for analysis (28).
Four-Dimensional (4D) HPLC-MS
MD setups are not limited to just two-
dimensional analysis. Use of more
than two dimensions is also emerg-
ing for the analysis of biotherapeu-
tics. For instance, recently, an on-line
4D HPLC-MS approach was reported
for characterization of mAb variants
(44). The mAb sample was resolved
and fractionated using IEC in the first
dimension. These fractions were then
transferred to the 2D reversed-phase
column, where they were reduced
and washed. The reduced mAb chains
were then eluted onto the third col-
umn, where they were digested.
344  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
The digested peptides were then detected using LC–MS in
the fourth dimension. The whole system was fully automated
from fractionation to peptide mapping. Prefractionation by
IEC allowed reduction and digestion of separated charge
variants, resulting in higher signal intensity and sensitivity.
The coupling of MS/MS allowed for localizing the modifica-
tions at the amino acid level. The system was highly sensi-
tive, and it was able to detect deamidation present, even at
0.5% of the main peak. Five out of six oxidation sites were
identified in the study. The system, however, showed lim-
ited retention of small and hydrophilic peptides. Although
the authors presented IEC in the first dimension, the 4D
set-up was designed so that SEC could be coupled in place
of the IEC for the characterization of size variants.
Conclusions
Biopharmaceuticals are complex molecules, and they
demand complex, analytical procedures with maximum
resolving power. To meet this need, one-dimensional analyti-
cal methods are not sufficient for achieving complete analy-
sis. Multidimensional systems offer an alternative, promis-
ing approach in this regard. 2D-LC is commonly used as a
multidimensional separation system for biopharmaceutical
analysis, and it has now been accepted by many biophar-
maceutical companies. Simplified instrumentation, method
Over 25 Years of
Quality HPLC and
SFC Columns
Offering a wide range
of phases and dimensions for
both analytical and prep
www.pci-hplc.com | 609.860.1803
development, and data analysis strategies in the future are
expected to make it a routine characterization technique for
quality control processes in the industry. Various modes are
available for 2D-LC systems, but these have limited orthogo-
nality and peak capacity. Recent years have witnessed the
growth of CE as a method of choice for MD separation sys-
tems. Although these are not yet as widely used as 2D LC-
based methods, interest in their deployment is growing rap-
idly. They represent one of the major developments in the
industry which should become more obvious in the future.
References
(1) A.J. Reason, A. Weiskopf, and A.S. Rathore, BioPharm Intl. 27(7),
1–8 (2014).
(2) A.S. Rathore, I.S. Krull, and S. Joshi, LCGC North Am. 36(6),
376–384 (2018).
(3) A.S. Rathore, I.S. Krull, and S. Joshi, LCGC North Am. 36(11),
814–822 (2018).
(4) M. Pioch, S.C. Bunz, and C. Neusüß, Electrophoresis 33(11),
1517–1530 (2012).
(5) S.K. Grebe and R.J. Singh, Clin. Biochem. Rev. 32, 5–31 (2011).
(6) G. Vanhoenacker, I. Vandenheede, F. David, P. Sandra, and K.
Sandra, Anal. Bioanal. Chem. 407(1), 355–366 (2015).
(7) J.R. Lill, Analytical Characterization of Biotherapeutics (John
Wiley & Sons, Hoboken, New Jersey, 1st Ed., 2017) pp. 1-14.
(8) W. Grochocki, M.J. Markuszewski, and J.P. Quirino, Electropho-
resis 36(1), 135–143 (2015).
(9) L. Ranjbar, J.P. Foley, and M.C. Breadmore, Anal. Chim. Acta
950, 7–31 (2016).
(10) P. Cesla, P. Dinisová, and J. Fischer, Sens. Electroanal. 6, 357–
368 (2011).
(11) B.W.J. Pirok, D.R. Stoll, and P.J. Schoenmakers, Anal. Chem.
91(1), 240–263 (2019).
(12) C.R. Evans and J.W. Jorgenson, Anal. Bioanal. Chem. 378, 1952–
1961 (2004).
(13) F.J. Kohl, L. Sánchez-Hernández, and C. Neusüß, Electrophore-
sis 36(1), 144–158 (2015).
(14) X. Wang, S. Buckenmaier, and D. Stoll, J. Appl. Bioanal. 3(5),
120–126 (2017).
(15) H. Wang and S. Hanash, J. Chromatogr. B 787, 11–18 (2003).
(16) P.W. Carr and D.R. Stoll, Two-dimensional Liquid Chromatogra-
phy (Agilent Technologies, 2015).
(17) R.E. Birdsall, H. Shion, F.W. Kotch, A. Xu, T.J. Porter, and W.
Chen, MAbs 7(6), 1036–1044 (2015).
(18) S. Schneider, Online 2D-LC Characterization of Monoclonal
Antibodies with Size Exclusion and Weak Cation Exchange
Chromatography (Agilent Technologies, 2016).
(19) Y. Li, D. Hewitt, Y. K. Lentz, J. A. Ji, T. Y. Zhang, and K. Zhang,
Anal. Chem. 86(10), 5150–5157 (2014).
(20) M. Alvarez, G. Tremintin, J. Wang, M. Eng, Y. Kao, J. Jeong, V.T.
Ling, and O. V Borisov, Anal. Biochem. 419(1), 17–25 (2011).
(21) J.J. Gilroy and C.M. Eakin, J. Chromatogr. B 1060, 182–189 (2017).
(22) M.T. Mazur, R.S. Seipert, D. Mahon, Q. Zhou, and T. Liu, AAPS J.
14(3), 530–541 (2012).
(23) Y. Li, C. Gu, J. Gruenhagen, K. Zhang, P. Yehl, N.P. Chetwyn, and
C.D. Medley, J. Chromatogr. A 1393, 81–88 (2015).
(24) D.R. Stoll, D.C. Harmes, J. Danforth, D. Guillarme, S. Fekete, and
A. Beck, Anal. Chem. 87(16), 8307–8315 (2015).
WWW.CHROMATOGRAPHYONLINE.COM
(25) K. Sandra, A. Ortiz, and P. Sandra, LCGC
Europe 28(10s), 45–48 (2015).
(26) J.H. Thompson, W.K. Chung, M. Zhu, L.
Tie, Y. Lu, N. Aboulaich, R. Strouse, and
W.D. Mo, Rapid Commun. Mass Spec-
trom. 28(8), 855–860 (2014).
(27) C.E. Doneanu, A. Xenopoulos, K. Fad-
gen, J. Murphy, S.J. Skilton, H. Prentice,
M. Stapels, and W. Chen, MAbs 4(1),
24–44 (2012).
(28) R.P. Vanam, M.A. Schneider, and M.S.
Marlow, MAbs 7(6), 1118–1127 (2015).
(29) R.K. Rober t s, H.J. Holovic s, M.R.
Thompson, and C.W. Demarest, Elec-
trophoresis 31(3), 448–458 (2010).
(30) J. Liu, H. Zhao, K.J. Volk, S.E. Klohr, E.H.
Kerns, and M.S. Lee, J. Chromatogr. A
735, 357–366 (1996).
(31) R. Kumar, R.L. Shah, and A.S. Rathore,
J. Chromatogr. A (2020). doi:10.1016/j.
chroma.2020.460954
(32) J. Hühner, K. Jooß, and C. Neusüß, Elec-
trophoresis 38(6), 914–921 (2017).
(33) K. Jooß, J. Hühner, S. Kiessig, B. Moritz,
and C. Neusüß, Anal. Bioanal. Chem.
409(26), 6057–6067 (2017).
(34) K. Sandra and P. Sandra, Bioanalysis
7(22), 2843–2847 (2015).
HPLC Columns for
Organic Acid Analysis
bensonpolymeric.com
775.356.5755
LCGC0719 Benson HPH Ad pr2f.indd 1
40419080306_BENSON
magenta
cyan
yellow
black POLYMERIC_pgxx_HPH_1-1.pgs 07.23.2019 20:56
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  345
(35) F. Xie, R.D. Smith, and Y. Shen, J. Chro- ABOUT THE CO-AUTHOR
matogr. A 1261, 78–90 (2012).
(36) M. Gilar, P. Olivova, A.E. Daly, and J.C.
Ramesh Kumar is a grad-
Gebler, Anal. Chem. 77(19), 6426–6434 uate student at the Depart-
(2005). ment of Chemical Engineer-
ing at the Indian Institute of
(37) D.R. Stoll, X. Li, X. Wang, P.W. Carr,
Technology in Delhi, India.
S.E.G. Porter, and S.C. Rutanb, J. Chro-
matogr. A 1168(1–2), 3–43 (2007).
(38) M. Kivilompolo, J. Pó, and T. Hyö-
tyläinen, LCGC Europe 24(5), 232–243 ABOUT THE COLUMN EDITOR
(2011).
(39) D. Chen, X. Shen, and L. Sun, Analyst Ira S. Krull is a Professor
142(12), 2118–2127 (2017). Emeritus with the Depart-
(40) R. Aebersold and H. D. Morrison, J. ment of Chemistry and
Chromatogr. 516, 79–88 (1990). Chemical Biology at North-
eastern University in Bos-
(41) S.E. Rudnick, V.J. Hilser, and G.D.
Worosila, J. Chromatogr. A 672(1–2),
ton, Massachusetts, and a member of
219–229 (1994). LCGC’s editorial advisory board.
(42) H.J. Issaq, K.C. Chan, C. Liu, and Q. Li,
Electrophoresis 22(6), 1133–1135 (2001).
(43) E. Tamizi and A. Jouyban, Electrophore-
ABOUT THE COLUMN EDITOR
sis 36, 831–858 (2015).
(44) C. Gstöttner, D. Klemm, M. Haberger, A. Anurag S. Rathore is a
Bathke, H. Wegele, C. Bell, and R. Kopf, professor in the Depart-
Anal. Chem. 90(3), 2119–2125 (2018). ment of Chemical Engineer-
(45) C. Song, M. Ye, G. Han, X. Jiang, F. ing at the Indian Institute of
Wang, Z. Yu, R. Chen, and H. Zou, Anal. Technology in Delhi, India.
Chem. 82(1), 53–6 (2010).
Over 40 years of experience
providing high quality polymeric HPLC
columns for the analysis of samples
containing carbohydrates and
organic acids.
Benson polymeric TM
2019-07-12 3:46 PM
ADVANSTAR_PDF/X-1a
346  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
Are You Controlling Peak Integration to Ensure Data Integrity?
Peak integration and interpretation of chromatograms is one of the current areas when inspectors are looking to ensure
the integrity of chromatographic data and results. Are you in control of integration, or are you integrating into compliance?
R D McDowall
C hromatography Data Systems (CDS)
have been at the center of multiple
FDA 483 citations and warning letters since
the Able Laboratories fraud case in 2005
(1). Between 2005 and 2015, the inspection
emphasis was mainly on unofficial testing,
deletion of data files, time traveling, and no
segregation of duties for users. As compa-
nies have changed their working practices to
stop these practices, the regulatory focus has
moved to peak integration and invalidated
out of specification (OOS) results. This “Data
Integrity Focus” column will review compliant
approaches to peak integration, and the next
article will discuss OOS investigations.
Compliant peak integration has been of
interest since 2015, enough so that I have
published three previous articles in LCGC
publications in that time. The first discussed
manual intervention and manual integration
in 2015 (2), and in the first series of “Data
Integrity Focus,” Mark Newton and I dis-
cussed peak integration and the sequence
of integrating data files in a run (3). Most
recently, Heather Longden and I revisited
that first 2015 article in 2019, and revised the
overall approach (4).
We’ll start our peak integration journey by
looking at the applicable GMP regulations
and guidance documents.
A Regulatory Refresher
It is important to understand the regulatory
requirements for laboratory controls and
records as these provide a major input to any
discussion about integration. The US GMP
regulations for laboratory controls include
this requirement:
21 CFR 211.160(b): “Laboratory controls shall
include the establishment of scientifically
sound and appropriate specifications, stan-
dards, sampling plans, and test procedures
designed to assure that components, drug
product containers, closures, in-process
materials, labeling, and drug products con-
form to appropriate standards of identity,
strength, quality, and purity (5).”
This is a relatively simple regulation to
understand; everything done in the labo-
ratory including peak integration must be
scientifically sound. A similar requirement
for scientific soundness in the laboratory is
found in Section 11.12 of EU GMP Part 2
(also known as ICH Q7) for active pharma-
ceutical ingredients (6,7).
As chromatography is a comparative
analytical technique, all injections should be
integrated in the same way as much as pos-
sible. However, this generalization is tested
near the limits of quantification and where
complex mixtures are separated.
Next, we have:
21 CFR 211.194(a): “Laboratory records shall
include complete data derived from all tests
necessary to assure compliance with estab-
lished specifications and standards, includ-
ing examinations and assays,….. (5).”
The requirement for laboratory records
is also simple to understand, but appar-
ently difficult to follow. Complete data
means “everything,” including poor injec-
tions and poor integration and was dis-
cussed in “Data Integrity Focus” last year
(8) and other articles (9–11).
The EU GMP regulations are not as sim-
ple to interpret, as Chapter 4 refers to raw
data which is not defined (12); a discussion
of the meaning of raw data has been pre-
sented earlier, and is equivalent to com-
plete data in the US GMP regulations (8).
If you are subject to FDA’s Pre-Approval
Inspections (PAI), Compliance Program
Guide (CPG) 7346.832 for inspectors was
updated in September 2019 (13). Under
Objective 3, the data integrity audit, the
following areas are specifically mentioned
in the new version for the inspector to
focus on:
• Manipulation of a poorly defined analyti-
cal procedure and associated data analy-
sis to obtain passing results
• Missing data or unreliable data
• Data or information submitted to the
application that were potentially unreli-
able or misleading and the relevance of
these data or information
• Unexplained or inappropriate gaps in a
chromatographic or analytical sequence
• A pattern of inappropriately disregarding
test results
• Inadequate or lack of justification for not
reporting data/information.
More detail on this revised version of CPG
7346.832 can be found in a recent “Focus on
Quality” column (14).
Although there have been many regula-
tory guidance documents from the MHRA,
WHO, FDA, and PIC/S (15–18), none of these
mention peak integration. The best guid-
ance for peak integration is PDA Technical
Report 80, which has a large section on chro-
matographic integration, and the guidance
document is good in that it illustrates both
acceptable and unacceptable peak integra-
tion practices (19).
ICH M10 section 3.3.6 outlines the cur-
rent thinking for integration of chromato-
grams in bioanalysis (20).
Chromatogram integration and reintegra-
tion should be described in a study plan, pro-
tocol or standard operating procedure (SOP).
Any deviation from the procedures
described a priori should be discussed in the
Bioanalytical Report.
The list of chromatograms that required
reintegration, including any manual inte-
grations, and the reasons for reintegration,
should be included in the Bioanalytical
Report. Original and reintegrated chromato-
grams, and initial and repeat integration
results, should be kept for future reference,
and submitted in the Bioanalytical Report for
comparative BA/BE (bioavailability or bio-
equivalence) studies.
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  347

Gone is the burdensome FDA require-
ment for a manager to preapprove any
manual integration (21), but this is replaced
by control via a plan or procedure, together
with a listing of chromatograms requiring
manual integration (moving baselines) and
the before and after chromatograms. For a
study involving the analysis of >1000 samples,
this could be quite a large burden especially
if the elimination phase, is relatively long and
the drug is slowly metabolized or excreted.
These requirements emphasize the need to
have robust and reliable analytical proce-
dures, as discussed in an earlier “Data Integ-
rity Focus” article (22).
A Parcel of Rogues
A small selection of FDA 483 observations
and warning letter citations of chromato-
graphic peak integration failures are pre-
sented in Table I, with the key problems
highlighted in bold. One piece of free advice
I would offer is not to say to an inspector that
chromatographers “play with parameters,”
even if it is to get “proper” integration. Sum-
marizing these citations, we have the follow-
ing areas of non-compliance:
• Integrating into compliance
• Failure to retain integration parameters
(for example, a lack of complete data)
• Lack of a procedure for manual integration
• Inappropriate use of integrate inhibit
• Unscientific integration
• No quality oversight
In my view, the FDA are wrong in their
requirement for a procedure on manual inte-
gration. To avoid these issues, we must have
a procedure that is structured and compliant,
and a scientific approach to peak integra-
tion as a whole. What is required is a holistic
approach with an SOP for all chromatographic
integration, of which manual integration is an
important subset. Further information about
an integration SOP, including banned prac-
tices such as peak skimming or enhancing,
can be found in the articles by McDowall,
Newton and McDowall and Longden and
McDowall (2–4). The purpose of the SOP is
not to describe the principles of peak integra-
tion, as this can be found in the excellent book
by Normal Dyson (23), or in tutorials and man-
uals from CDS suppliers. A suggested peak
integration workflow is shown in Figure 1 (2,
24), and we will discuss various aspects of this
workflow throughout this article.
Why Manually Integrate Peaks?
We need to consider why we need to inte-
grate peaks manually. The reason is that
there are situations where a CDS cannot inte-
grate peaks correctly, and some examples of
this are:
• Split peaks
• Shoulder peaks
• Tailing peaks
• Baseline noise
• Negative peaks
• Co-eluting peaks
• Rising or falling baselines
• Slowly eluting peaks (where the CDS
has difficulty identifying the peak start
and end).

Robust Methods, Fewer Injections

Software to streamline method development using in silico optimization

ACD/Method Selection Suite

Generate Robust Minimize Experiments

Methods in silico Build optimized methods
Optimize pH, gradient, with minimal injections
temperature, & more
through simulation Predict the Impact of Track Peaks
4 5 35º 40º Method Changes Automatically
6 7 15º 25º
Visualize the effect of Consistently identify
Co pH Te
varying methods without peaks of interest based
lum m
n p added experiments on spectral similarity

Learn more www.acdlabs.com/methdev

A trademark of Advanced Chemistry Development, Inc. (ACD/Labs)
348  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM

TABLE I: A Selection of FDA 483 observations and warning letter citations on peak integration

Company Peak Integration Citation

Unimark Remedies Ltd (25)
Divi’s Laboratories (26)
Leiner Health Products (27)
Kashiv BioSciences (28)
Megafine Pharma Ltd (29)
During the inspection, the investigator found no procedures for manual integration or review of electronic
and printed analytical data for (b)(4) stability samples. Electronic integration parameters were not saved
or recorded manually. When the next samples were analyzed, the previous parameters were overwritten
during the subsequent analyses.
(T)he software you use to conduct HPLC analyses of API for unknown impurities is configured to permit
extensive use of the “inhibit integration” function without scientific justification. For example,.....your
software was set to inhibit peak integration at four different time periods throughout the analysis. Your
firm reintegrated multiple chromatograms to determine <redacted> levels; however, the parameters for
the reintegration were not retained.
In addition, our investigators documented many iinstances with extensive manipulation of data with no
explanation regarding why the manipulation was conducted. This manipulation would include changing
integration parameters or relabeling peaks such that previously resolved peaks would not be integrated and
included in the calculation for impurities…There was no Standard Operation Procedures (SOP) to describe the
policy, standard practice, and circumstances under which manual integration would be allowed…..
Integration of chromatograms for method STM-0076 <redacted> has been performed inconsistently..... The
recommendation is often, but not always followed and results in this area being integrated. There is a lack of
scientific justification to support if the tailing portion should be integrated. The change.....was implemented
without fully evaluating the impact on previously processed data.
Your test methods were not capable of demonstrating the purity of your drugs .....(A)nalysts reprocessed
data up to 12 times, and only included the final result in the report for review by Quality Assurance. Your
Deputy Manager, Quality Control stated that it is common practice to “play with parameters” to get the
proper integration (29).

You failed to establish adequate test procedures. For example, your analyst manually integrated a HPLC
Hubei Danjiangkou Danao
test for <redacted> API despite the fact that the chromatogram lacked peak resolution.....You lacked an
Pharmaceutical Co Ltd (30)
approved protocol for manual integration or quality oversight of the practice.

Your firm failed to properly integrate co-eluting peaks during impurity testing of phentermine HCL
capsules, which resulted in your analysis failing to detect out-of-specification (OOS) results for at least
KVK-Tech, Inc (31)
one lot of drug product.....Multiple examples of your firm’s failure to properly integrate closely
co-eluting peaks were observed during our inspection.

The reasons for the inability of the integra-
tion method may be due to:
• Poor method development and valida-
tion where the analytical procedure or
the integration method is not optimized
or robust.
• Complex sample matrices resulting in
interfering peaks that may still be present
after sample preparation, such as biologi-
cal samples, contrast media, and fermen-
tation media.
• Complex mixtures of similar analytes
These situations may result in a heavy
manual integration workload, as the
CDS method is not able to integrate all
peaks automatically.
A Suggested Peak
Integration Workflow: 1
The first part of the integration workflow in
Figure 1 is that all peak integration in the first
instance must always be conducted using
automatic integration. There are no excep-
tions to this statement. None, repeat none.
If the peak integration for the sequence
is acceptable, then the reportable results
can be calculated and reviewed; high fives
all around! However, if the peak integra-
tion is not acceptable we come to the
first decision point: Is manual integration
allowed for this analysis? If not, we trigger
a laboratory investigation.
The big problem with the citations about
the lack of a manual integration procedure
in Table I is that there is no definition of the
term “manual integration.”
Defining Manual Integration
In a “Questions of Quality” column on
integration, it was noted that there was
no definition of manual integration, and it
was implicitly defined as manual placement
of the baseline by a chromatographer (2).
Mark Newton and I reiterated this defini-
tion in our “Data Integrity Focus” article in
2018 (3). The PDA’s Technical Report 80, also
published in 2018, formally defined manual
integration as:
“(The) (p)rocess used by a person to modify
the integration of a peak area by modifying
the baseline, splitting peaks or dropping a
baseline as assigned by the chromatogra-
phy software to overrule the pre-established
integration parameters within the chromato-
graphic software (19).”
Is this definition acceptable?
• It is wordy, repetitious, and could be
better phrased.
• The use of the word “overrule” is conten-
tious. Chromatography is a dynamic pro-
cess, and a CDS application can struggle
to separate overlapping peaks which are
obvious to a trained eye.
• The biggest issue with this definition
is that there is no mention of scien-
tific soundless, as defined in the FDA
GMP regulations (5).
WWW.CHROMATOGRAPHYONLINE.COM
Laboratory
No
Investigation
1. Rename
peaks or adjust
windows 2. Adjustment of
Manual
Intervention
FIGURE 1: A suggested peak integration workflow (2).
Therefore, a simpler, more concise, and
better definition of manual integration
should be “manual repositioning of peak
baselines with scientific justification for
their positioning.”
Implicit within this definition is the use of
CDS software; otherwise, you’d be draw-
ing baselines on paper. However, this also
requires that the chromatographer is trained,
and, ideally, software technical controls
should prevent manual repositioning of
baselines where this is not justified by the
type of the analysis.
Should Manual
Integration Be Banned?
From the citations in Table I, would it be rea-
sonable to ban manual integration in regu-
lated laboratories? Let’s think this through.
Experienced analysts know that chromato-
graphic analysis can be affected by tempera-
ture, humidity, and column history, as well
as mobile phase preparation, such that one
day’s analysis often varies slightly from the
previous day’s run. To achieve consistent out-
put and measurement, it is critical to adapt
or optimize factors such as peak detec-
tion threshold or retention time windows
to ensure consistent, correct, and accurate
peak identification and measurement (4). But
how can reviewers, approvers, quality, or out-
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  349
Chromatographic Run
Automatic
integration
Manual Complete
Integration No Acceptable? Yes Analysis
OK?
Yes
Manual
Integration
Options
integration
parameters 3. Manual
reposition of
baselines
Manual
Reintegration
side auditors recognize the legitimate versus
egregious use of integration?
Banning the use of manual integration is a
common response to avoid questions about
data integrity. However, there are three out-
comes to this crude (and stupid) action:
• Laboratories will have to accept poor and
inconsistent integration
• Analysts will find a workaround that per-
mits them to integrate each run with a
different set of integration parameters
(typically involves performing quanti-
fication in a LIMS, or worse, a spread-
sheet, without traceability back to the
integration methods)
• Analysts will be forced to spend hours
of their day developing complex and
manipulative methods to address varia-
tions between chromatograms with a
single processing method. Typically, this
will require many “integration events”
that could even include placing peak
starts and ends at specific time points;
in effect, performing manual integration
under the guise of an automated method
to deceive a reviewer (4).
Consider, also, the nature of an analysis.
My experience is based on analysis of small
heterocyclic molecules; however, I acknowl-
edge that there are biologicals, macro-
molecules, and chiral separations that will
350  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
TABLE II: A suggested approach when manual integration is and is not allowed
Manual Integration is Allowed When
• Peak split by CDS software
• Co-eluting peak
• Well-defined peaks on the
shoulders of other peaks
• For impurity peaks close to the limits
of quantification or detection
• Matrix interference
• Negative spikes in the baseline
• Failure to separate peaks
1If each type of manual integration is allowed by the applicable SOP or analytical procedure.
2
Manual integration is dependent on the nature of the analyte (simple heterocyclic, biological, or
macromolecule; and sample matrix such as liquid, emulsion, or biological, cell culture medium).
have broader peaks and there will be the
impact of the sample matrix that could
impact resolution of compounds. Where is
the laboratory in the research, development,
and manufacturing spectrum? The closer to
product registration and manufacturing, the
analytical procedure should be validated
and its performance understood, and the
need for manual integration should be
restricted to analysis of impurities near the
limits of quantification or detection, albeit
with the caveats about the analyte and
matrix above. The key question is, “What
can be justified scientifically?”
In the wrong hands, with the wrong
intent, and without a robust training and
review process, altering chromatographic
peak processing parameters has been mis-
used by analysts to falsify results. How can
this be managed?
A Suggested Peak
Integration Workflow: 2
Returning to the workflow depicted in Figure
1, we move to a decision point that deter-
mines if manual integration (whatever that
term may cover) is permitted for an indi-
vidual analytical procedure. If not, the next
stage is a laboratory investigation.
What methods could we consider for
inclusion for no manual integration?
• Measurement of active pharmaceutical
ingredients (APIs)
• Registered methods for finished product
for the active ingredient
• Stability indicating methods (main peak)
1,2 Manual Integration is Not Allowed For
• Improper peak identification
• Peak shaving or clipping
• Peak juicing or enhancement
• Raising the baseline above the
detector signal
• Lowering the baseline below the
detector signal
• Selectively adjusting integration
events
• Insufficient sensitivity
• Using integration parameters
inappropriately (peak inhibition)
• Resolving poor chromatography
due to instrument failure
• One position could be if these peaks can-
not be integrated correctly are you out of
control? We will consider if this is tenable
when we have finished discussing the
flow chart.
In my view, manual integration must
be specifically prohibited in the following
circumstances:
• Symmetrical peaks that have acceptable
baseline to baseline fitting following
automatic integration.
• Enhancing or shaving peak areas to meet
SST acceptance criteria or allowing a run
to meet the test specification.
Now we come to what constitutes “man-
ual integration. Figure 1 presents three
options that I have selected for discussion;
your laboratory SOP may have more areas
depending on the work performed. The
ideal outcome you require is consistent and
appropriate manual integration that is scien-
tifically defensible. Thus, you need to avoid
situations where you have inconsistent or
inappropriate integration.
Manual Intervention vs.
Manual Integration
Let me introduce two terms: manual inte-
gration and manual intervention (2). The
former term you know and love; it has been
defined above and is manual placement
of baselines by a chromatographer. But
manual intervention? Is this just playing with
words to confuse you?
Manual intervention is defined as chang-
ing integration parameters without manual
baseline placement. My intention is to iden-
tify an area for data interpretation that rep-
resents the middle ground between auto-
matic integration and manual integration.
As you can see from Figure 1, there are two
types of manual intervention. Let me eluci-
date my reasoning.
• Option 1: Peaks have slipped out of a
retention window, and they are not cor-
rectly identified. Otherwise, automatic
integration is acceptable, and all that is
required is to change the peak windows
in the integration method and repro-
cess. Peak areas are not changed by this
approach. Providing that all changes to
integration parameters are recorded
by the CDS application (either in a new
version of the integration method or
the audit trail), this is acceptable, and
can be justified on a rational basis, as
there are no other changes to the run
data. Of course, if manual integration
was banned, you would be initiating a
laboratory investigation already. This is
not the smartest move.
• Option 2: Parameters in the integration or
processing method need to be adjusted
and then applied to all injections in the
run. An example could be change of
the peak threshold or minimum area to
reduce the impact of baseline noise. Peak
areas may or may not be changed under
his option, but, importantly, there is no
manual placement of the baselines by an
analyst. Again, all changes should be cap-
tured by the software to ensure complete
data. Don’t even think of an investigation.
• Option 3: To complete the discussion, if
permitted, manual placement of baselines
by the chromatographer is required due to,
say, a late running peak or noise if under-
taking an impurity analysis. Peaks areas will
be changed by the reintegration. Obviously,
all changes will be monitored by the CDS
and recorded, especially the number of
attempts at integration.
Note that the three options in Figure 1
are shown on different levels. This is my way
of illustrating the descent of a chromatog-
rapher into the manual integration nether-
world. In the first option you may think of the
first option as purgatory, and the second as
limbo, while the third option is... well, you get
the idea.
WWW.CHROMATOGRAPHYONLINE.COM JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  351

Options 1 and 2 are manual intervention, No Manual

Integration
but the baseline placement is performed Allowed

by the CDS and is not altered by a chroma-

tographer. These are the preferred options, Automatic
Integration
Manual
Integration
System
Suitability Test
Standards
Other
Checks
Samples

and easier to justify scientifically. Manual

intervention is also acceptable, according to
the PDA’s Technical Report 80 (19). Option 3 • Check integration
• Manual
• Only where
permitted
• Must meet
predefined
• Must meet
predefined
• Drift checks
and QC samples • Peak shape
intervention: • Standards & acceptance acceptance • Must meet acceptable
is where a chromatographer goes through adjust parameters samples
the same
treated criteria criteria predefined • Peak areas
generated
acceptance
individual chromatograms and repositions criteria

the baselines where appropriate. This latter

Order of
point is important; it is an exercise of scien- Evaluation

tific judgment that needs to be backed up FIGURE 2: Scope and order of chromatographic integration (3).
by the procedures within the integration
SOP and the changes retained in the audit This places responsibility on any labora- • Standards injections
trail of the CDS. Table II shows suggestions tory to develop robust analytical chromato- • Other checks, such as quality control sam-
for when manual integration is and is not graphic procedures with reliable separations ples and blanks
allowed, depending on the nature of the that are fit for use as discussed in a previous • Samples.
analysis undertaken by a laboratory. “Data Integrity Focus” article (22). A key com- If any of these items fail acceptance crite-
ponent of this approach is that the resultant ria, STOP! Do not review anything involving
Why Use Automatic Integration? peak integration must be consistent. This is a samples until assay criteria are met. Looking
This is a very simple question to answer from subtle, but vital, difference that is not always at sample results prior to accepting the run
a business perspective: sequences where appreciated. could be considered “integrating into com-
integration is automatic are very much pliance,” as you wanted to invalidate testing
quicker to review. In contrast, manual inte- Sequence of Data File Integration you didn’t like (or accept one you did like).
gration is slow, especially for sequences with In 2018, Mark Newton and I discussed the Fresenius Kabi Oncology received a warning
a large number of injections with complex order of integration and processing of any letter from the FDA for such behavior (32).
separations where each injection must be injection sequence (3), and, due to the
integrated manually. Often the review can importance of correct integration, it is worth Is the FDA Inhibiting Integration?
take longer than the actual analysis (3). Con- repeating here. Figure 2 shows the order of The use of inhibit integration function is cur-
sequently, a faster review process means that integration evaluation: rently a hot regulatory topic, as can be seen
analysis is completed quicker, a batch can • Automatic integration and manual from in the FDA regulatory citation of Divi’s
be released sooner, and the organization intervention Laboratories (shown in Table I), where this
cash flow benefits. There is also the bonus of • Manual integration (if permitted) function was used four times in the middle of
lower regulatory scrutiny. • System suitability test injections a chromatogram.

CAMAG® HPTLC PRO

Fully automated sample analysis and
evaluation system for routine quality control

Explore the next dimension of High-Performance

Thin-Layer Chromatography at camag.com
352  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
This citation is based on 21 CFR 211.160(b) (5) that was discussed
earlier in this article, and the key questions to ask are if, where
and when can integrate inhibit be used. It has been said in some
audits and inspections that this function cannot be used. This is
an untenable situation, as there is no explicit or implicit statement
in the GxP regulations for this attitude. However, it comes back to
scientific soundness, and a laboratory must be able to justify the
use of the function.
Let us consider the following scenarios:
• There is baseline perturbation with a large negative peak after
an injection.
• A peak of interest elutes shortly after the negative peak. The use of
integrate inhibit is fully justified from the start of the injection, void
volume, solvent front, and until the baseline has returned to nor-
mal and before elution of the peak of interest. Otherwise, there is
a large probability that baseline placement of the analyte could be
adversely influenced by the negative peak. This would also result
in more manual integration.
• Similar scenarios occur when extraneous peaks are washed from
the column, or baseline perturbations from mobile phase change
during a gradient elution, or wash at the end of a chromatogram.
What is more problematic is the use of integrate inhibit in
the middle of a run, as cited above. If system, blank, or other
non-sample peaks occur in the middle of a chromatogram, tra-
ditionally those peaks were not integrated. Because of suspi-
SERVING ROYALTY. EXCEEDING EXPECTATIONS. EVERY MOMENT.
• Provider of top brand HPLC instrumentation products
• Equivalent to corresponding OEM products
• Serving customers for over 35 years
• Reduce product repair expenses by up to 30%
• Lifetime Warranty on manufacturing defects
www.sciencix.com 800.682.6480 sales@sciencix.com
cions that the excluded peaks might be real impurities, exclud-
ing these system peaks needs to be carefully documented and
justified in the method development and validation reports;
otherwise, they should be integrated and marked clearly as
system peaks.
System Evaluation Injections
Trial injections using actual samples feature in many warning let-
ters (33) and Question 13 of the FDA Data Integrity Guidance (16),
which states:
“FDA prohibits sampling and testing with the goal of achieving a spe-
cific result or to overcome an unacceptable result (e.g., testing different
samples until the desired passing result is obtained).”
This is correct and should never be acceptable in a GxP
laboratory SOP.
However, consider the following situation: You are analyzing low
volume samples from a non-clinical study. There is a total volume of
20 µL plasma sample that is extracted, and there is only enough for
a single injection from each sample. Ask yourself the question: Are
you going to commit an analytical run of samples without checking
that the system is ready? The cost of repeating the study is a high
six figure sum if a run does not work. Therefore, from a practical
perspective, we need a way of checking that a system is ready for
analysis, but does not involve testing into compliance with samples.
Ah, somebody says to use SST injections. The problem is that you
may need several replicates to determine if the system is ready, and
SST should never be started until you are confident that the system
is equilibrated.
Consider the following approach for system evaluation or readi-
ness injections or equilibration checks (4):
• The ability to use system evaluation injections must be docu-
mented in an applicable SOP or analytical procedure.
• The minimum column equilibration time needs to be documented in
the method to avoid excessive system readiness injections.
• Only system evaluation injections prepared from a suitable refer-
ence standard can be used to evaluate if the chromatographic sys-
tem ready. Records of the solution preparation must be available.
Ideally, a test mixture which mimics the separation characteristics,
but is easily distinguishable from real samples could be used.
• Should the maximum number of system evaluation injections
that can be made be documented in the procedure before a
problem with the chromatographic system needs to be investi-
gated? If the cause is thought to be an equilibration issue, waiting
and injecting again should be sufficient. If the system continues
to not behave, then an investigation is needed, the cause should
be found, remediated, and documented in the instrument log
book before checking the system evaluation again. If the prob-
lem requires maintenance to resolve it (for example, a pump
seal replacement), then requalification of the pump should be
conducted and documented before beginning the analysis.
• Using sample preparations must be prohibited in this procedure,
accompanied with staff training.
WWW.CHROMATOGRAPHYONLINE.COM
• System evaluation injections are part of
complete data for the analytical run, and
must be included in the instrument log
book entries, along with any investiga-
tion and remediation work on the instru-
ment. A common practice is to store
the data from these tests in a separate
folder or location to the real analyses.
This practice needs careful management
and documentation, as it becomes dif-
ficult to connect those injections to the
official laboratory work. Ideally, all work,
including system evaluation injections,
should be stored in the same location.
Five Rules of Integration
An integration SOP was discussed earlier. To
help understand what should be in it and
the associated training, there are five rules of
integration to consider (4):
Rule 1: The main function of a CDS is not
to correct your poor chromatography. This
places greater emphasis on the development
of robust chromatographic procedures so
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  353
that the factors involved in the separation are
known and controlled adequately (22). Chro-
matographic methods should be developed
with automatic integration and the norm
and not the exception. Management need
to understand that adequate time must be
given to method development and validation.
This is especially true for pharmacopoeial
methods that never work as written.
Rule 2: Never use default integration
parameters, always configure specific
integration for each method. Without
exception, peak integration and result pro-
cessing must be defined and validated for
each method so that all peak windows and
names are defined, and, if necessary, any
system peaks are identified. Using a default
or generic method results in excessive need
for manual integration to name and calculate
peaks, and so forth.
Rule 3: Always use automatic integration
as a first option and control manual inte-
gration practices. Remember that the use
of manual integration is a regulatory concern,
and use needs to be scientifically sound. But
also be aware that, as discussed earlier, man-
ual integration slows down a process, so see
Rule 1 to get the right method depending
on the sample matrix and peaks of interest.
Rule 4: Understand how the CDS works
and how the numbers are generated. This
requires basic training in the principles of peak
integration and how a CDS works. The prob-
lem is that with mergers, acquisitions, and
encouraging experiences analysts to retire
and employ younger workers, skills are being
eroded, and a CDS can be looked at as a
black box that always gives the right answers.
Rule 5: Use your brain – Think! This rule
is sometimes difficult to follow, but it follows
on from Rule 4. You can have what appears
to be a perfect separation and peak integra-
tion, but look at peak start and end place-
ment—do they look right? Use the zoom
and overlay functions of the CDS to see of
standards and samples have the right peak
354  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  JUNE 2020WWW.CHROMATOGRAPHYONLINE.COM
shape, etc. We discussed in the article on
second person review the need to have
an adequate sized monitor; this is a case
where one is essential (34). The analyst has
the responsibility to execute applicable
procedures correctly, which includes cor-
rect peak integration. The reviewer, how-
ever, also has a role to ensure that all inte-
gration (whether automated, optimized,
or manually placed) follows the method
guidance for placing baselines as the SOP
describes, especially when the representa-
tive area for unresolved peaks are being
estimated. Significant peak area manipula-
tion should be easily noticed by an experi-
enced reviewer.
Quo Vadis Peak Integration?
If you think that peak integration is a regula-
tory issue now, what will it be like in the future?
The May 2019 supplement to LCGC Europe
gives an interesting glimpse of the future
in which Wahab and associates (35) discuss
advanced signal processing techniques that
could be used in chromatographic integra-
tion. The techniques listed are:
• Deconvolution of extra column effects
by Fourier transformation for removing
band broadening.
• Peak area extraction by iterative curve
fitting for partial overlapping peaks in
a chromatogram.
• Model free approaches for peak infor-
mation extraction, another approach for
extracting peak areas from overlapping
peaks in complex matrices.
• Direct resolution by Power Law increases
resolution by reducing peak width
and trailing.
• Direct resolution enhancement by
even derivative peak sharpening
also increases resolution by reducing
peak width.
It is beyond the scope of this article
to present and discuss what is already in
Wahab’s paper, but if any of these tech-
niques are integrated into a chromatogra-
phy data system, then their use needs to
be justified scientifically. This means from
development through validation to use of
a method.
If regulators are worried by peak inte-
gration now, they could be paranoid in
the future!
Summary
Control of peak integration is imperative in a
regulated laboratory. Good peak integration
requires good chromatography. Good chro-
matography requires a well developed and
robust analytical procedure. The bottom line
is: Are you in control of the analytical proce-
dure and peak integration?
References
(1) Able Laboratories Form 483 Observations.
2005 23 Dec 2019]; Available from: https://
www.fda.gov/media/70711/download.
(2) R.D. McDowall, LCGC Europe 28(6), 336–342
(2015).
(3) M.E. Newton and R.D. McDowall, LCGC
North Am. 36(5), 330–335 (2018).
(4) H. Longden and R.D. McDowall, LCGC
Europe 21(12), 641–651 (2019).
(5) 21 CFR 211 Current Good Manufacturing
Practice for Finished Pharmaceutical Prod-
ucts. (Food and Drug Administration: Sliver
Spring, MD, 2008).
(6) ICH Q7(R1) - Good Manufacturing Prac-
tice for Active Pharmaceutical Ingredients.
(International Conference on Harmonisation:
Geneva, 2016).
(7) EudraLex - Volume 4 Good Manufactur-
ing Practice (GMP) guidelines, Part 2 - Basic
Requirements for Active Substances used as
Starting Materials (European Commission,
Brussels, Belgium, 2014).
(8) R.D. McDowall, LCGC North. Am. 37(4), 265–
268 (2019).
(9) R.D. McDowall, Spectroscopy 28(4), 18–25
(2013).
(10) R.D. McDowall, Spectroscopy 31(11), 18-21
(2016).
(11) R.D. McDowall, Spectroscopy 33(12), 8–11
(2018).
(12) EudraLex - Volume 4 Good Manufacturing
Practice (GMP) Guidelines, Chapter 4 Docu-
mentation (European Commission, Brussels,
Belgium, 2011).
(13) FDA Compliance Program Guide CPG
7346.832 Pre-Approval Inspections (Food
and Drug Administration: Silver Spring,
Maryland, 2019).
(14) R.D. McDowall, Spectroscopy 34(12), 14–19
(2019).
(15) MHRA GxP Data Integrity Guidance and Defini-
tions (Medicines and Healthcare Products Regu-
latory Agency, London, United Kingdom, 2018).
(16) FDA Guidance for Industry- Data Integrity and
Compliance With Drug CGMP Questions and
Answers (Food and Drug Administration: Silver
Spring, Maryland, 2018).
(17) WHO Technical Report Series No.996 Annex
5 Guidance on Good Data and Records Man-
agement Practices (World Health Organisation:
Geneva, Switzerland, 2016).
(18) PIC/S PI-041-3 Good Practices for Data Man-
agement and Integrity in Regulated GMP / GDP
Environments Draft. (Pharmaceutical Inspec-
tion Convention/Pharmaceutical Inspection
Cooperation Scheme, Geneva, 2018).
(19) Technical Report 80: Data Integrity Manage-
ment System for Pharmaceutical Laboratories
(Parenteral Drug Association [PDA], Bethesda,
Maryland, 2018).
(20) ICH M10 Bioanalytical Method Validation
Stage 2 (International Council on Harmaoni-
sation: Geneva Switzerland, 2019 [draft]).
(21) FDA Guidance for Industry: Bioanalytical
Methods Validation (Food and Drug Admin-
istration: Silver Spring, Maryland, 2018).
(22) R.D. McDowall, LCGC North. Am. 38(4), 233–
240 (2020).
(23) N. Dyson, Chromatographic Integration
Methods (Cambridge: Royal Society of
Chemistry, Cambridge, United Kindom, 2nd
Ed., 1998).
(24) R.D McDowall, Validation of Chromatography
Data Systems: Ensuring Data Integrity, Meet-
ing Business and Regulatory Requirements
(Cambridge: Royal Society of Chemistry,
Cambridge, United Kindom, 2nd Ed., 2017).
(25) FDA Warning Letter Unimark Remedies,
India, (Food and Drug Administration: Silver
Spring, Maryland, 2014).
(26) FDA Warning Letter: Divi’s Laboratories Ltd.
(Unit II) ( Warning Letter 320-17-34). (Food
and Drug Administration: Silver Spring,
Maryland, 2017).
(27) FDA Warning Letter Leiner Health laborato-
ries (Food and Drug Administration: Rockville,
Maryland, 2006).
(28) FDA 483 Observations: Kashiv BioSciences.
(Food and Drug Administration: Silver Spring,
Maryland, 2018).
(29) FDA Warning Letter: Magafine Pharma Ltd.,
US FDA:, Silver Spring, Maryland, 2017)
(30) FDA Warning Letter: Hubei Danjiangkou
Danao Pharmaceutical Co., Ltd. (Food and
Drug Administration, Silver Spring, Mary-
land, 2017).
(31) FDA Warning Letter: KVK-Tech, Inc. (Food
and Drug Administration, Silver Spring, Mary-
land, 2020).
(32) FDA Warning Letter Fresenius Kabi Oncol-
ogy (Food and Drug Administration: Silver
Spring, Maryland, 2017) Available from:
https://www.fda.gov/ICECI/EnforcementAc-
tions/WarningLetters/2017/ucm589941.htm.
(33) R.D.McDowall, LCGC Europe 27(9), 486–492
(2014).
(34) R.D. McDowall, LCGC North. Am. 38(3), 163–
172, 193 (2020).
(35) M.F. Wahab, G. Hellinghausen, and D.W.
Armstrong, LCGC Europe 32(s5), 22–28 (2019).
R.D. McDowall
is the director of R.D. McDowall Lim-
ited in the UK. Direct correspondence
to: rdmcdowall@btconnect.com
 SUPPLEMENT TO

THE
APPLICATION
NOTEBOOK

June 2020
www.chromatographyonline.com
TABLE OF CONTENTS

THE APPLICATION
NOTEBOOK
Environmental
357 The Extraction of PFAS Molecules from Spiked Soil
Alicia D. Stell, CEM Corporation

Medical/Biological
359 Tailored Analysis of Phospholipid Classes Using
iHILIC-Fusion(+) as First Dimension for Online
Two-Dimensional HILIC–Reversed-Phase LC–MS
Patrick O. Helmer*, Carina M. Wienken*, Wen Jiang†, and Heiko Hayen*,
*Institute of Inorganic and Analytical Chemistry, University of Münster, Münster,

Germany, †Hilicon AB

361 Online SPE and LC–MS Analysis of Thyroid Hormones

in Human Serum
MilliporeSigma

362 Analysis of Fentanyl and Its Analogues in Human

Urine by LC–MS/MS
Shun-Hsin Liang and Frances Carroll,Restek Corporation

364 Chondroitin Sulfate Analyzed by SEC–MALS

Wyatt Technology

Food and Beverage

365 HILIC Columns for Phosphorylated Sugar Analysis
Showa Denko America, Inc.

Pharmaceutical/Drug Discovery
366 Amino Acid Analysis by European Pharmacopeia
to Support Coronavirus Research
Pickering Laboratories, Inc.
367 Size-Exclusion Chromatography for the Impurity
Analysis of Adeno-Associated Virus Serotypes
Stephan M. Koza and Weibin Chen,Waters Corporation
369 An Anion-Exchange Chromatography Method
for Monitoring Empty Capsid Content in
Adeno-Associated Virus Serotype AAV8
Hua Yang, Stephan M. Koza, and Weibin Chen,Waters Corporation

356 THE APPLICATION NOTEBOOK – JUNE 2020

 ENVIRONMENTAL

The Extraction of PFAS Molecules from Spiked Soil

Alicia D. Stell, CEM Corporation

Per- and polyfluoroalkyl substances (PFAS) are a group of manufactured Sample Preparation and EDGE Method
chemicals that are used in a wide variety of industries because of their Five grams of clean sandy loam was weighed directly into a Q-Cup®
resistance to stains, grease, and high temperatures. They possess a chain containing the S1 Q-Disc® stack (C9-G1-C9 sandwich). Each sample
of linked carbon atoms with fluorine atoms branching off the main chain. was spiked with 2 ng or 200 ng of standard in HPLC-grade methanol,
The presence of the strong carbon-fluorine bond contributes to the stability resulting in low spike and high spike samples, respectively. Each set of
of these compounds, earning them the nickname “forever chemicals.” spikes was done in triplicate. The sample was extracted with the EDGE
PFAS are used in products such as nonstick cookware, firefighting foam, using 80:20 methanol:water with 0.3% ammonium hydroxide using the
and stain-resistant carpets. method provided. Each set of spikes was extracted on the EDGE, and then
Because of their persistent nature and their widespread use, this group a blank extraction of the system was done with a Q-Cup containing the S1
of substances has leached into the environment with limited methods Q-Disc stack (C9-G1-C9 sandwich) to assess the level of carryover. Each
of remediation. Furthermore, these compounds have been found to extraction was collected in a polypropylene conical tube using an EDGE
bioaccumulate in animals and humans, and exposure in humans has rack. The sample was brought up to a final volume of 20 mL using 80:20
been shown to cause adverse health outcomes, including cancer, infertility, methanol:water with 0.3% ammonium hydroxide, and 20 µL of formic
and endocrine disruption. Thus, the assessment of the levels of PFAS in acid was added to each sample to neutralize the sample. The samples
the environment is important to the health and safety of humans. were then analyzed by Pace Analytical.
The EDGE, an automated solvent extraction system, was used to extract
a subset of PFAS molecules from spiked soil samples. The EDGE was EDGE Method
able to extract the soil samples in less than 10 min. The extraction yielded Q-Disc: S1 stack (C9-G1-C9 sandwich)
excellent recoveries and standard deviations. Furthermore, there was no Extraction Solvent: 80:20 methanol:water with 0.3% ammonium
carryover found between samples. The EDGE is an excellent choice for hydroxide
laboratories seeking to automate their PFAS extraction. Cycle 1
Top Add: 10 mL
Method Bottom Add: 0 mL
Rinse: 0 mL
Reagents Temperature: 65 °C
Clean sandy loam was purchased from MilliporeSigma. A PFAS standard Hold Time: 3 min
containing 24 different compounds (Part number 99207) was purchased Cycle 2
from Absolute Standards, Inc. HPLC-grade methanol, HPLC-grade water, Top Add: 10 mL
HPLC-grade formic acid, and ammonium hydroxide were purchased from Bottom Add: 0 mL
Fisher Scientific. Rinse: 0 mL

THE APPLICATION NOTEBOOK – JUNE 2020 357

ENVIRONMENTAL

Table I: Average recovery results from the low spike and high spike soil samples

Compound Low Spike RSD (n = 3) High Spike RSD (n = 3)

1H, 1H, 2H, 2H-perfluorodecane sulfonic acid (8:2 FTS) 83.00% 7.81% 87.00% 7.00%
1H, 1H, 2H, 2H-perfluorooctane sulfonic acid (6:2 FTS) 80.00% 3.00% 87.67% 7.09%
1H,1H,2H,2H-perfluorohexane sulfonic acid (4:2 FTS) 86.67% 10.02% 87.67% 5.13%
N-ethylperfluoro-1-octanesulfonamidoacetic acid (EtFOSAA) 85.67% 6.66% 86.00% 5.57%
N-methylperfluoro-1-octanesulfonamidoacetic acid (MeFOSAA) 78.67% 1.15% 81.33% 4.16%
Perfluoro-1-butanesulfonic acid (PFBS) 78.00% 3.00% 86.00% 1.73%
Perfluoro-1-decanesulfonic acid (PFDS) 77.00% 6.56% 84.33% 0.58%
Perfluoro-1-heptanesulfonic acid (PFHpS) 81.00% 2.65% 87.00% 2.65%
Perfluoro-1-nonanesulfonic acid (PFNS) 76.00% 6.56% 84.67% 2.89%
Perfluoro-1-octanesulfonamide (PFOSA) 81.00% 8.00% 86.67% 5.03%
Perfluoro-1-pentanesulfonic acid (PFPeS) 78.00% 4.36% 86.33% 0.58%
Perfluorohexanesulfonic acid (PFHxS) 98.33% 2.89% 89.67% 3.79%
Perfluoro-n-butanoic acid (PFBA) 85.65% 4.01% 88.17% 2.20%
Perfluoro-n-decanoic acid (PFDA) 101.33% 7.51% 93.33% 2.89%
Perfluoro-n-dodecanoic acid (PFDoA) 79.67% 7.37% 79.33% 2.08%
Perfluoro-n-heptanoic acid (PFHpA) 90.67% 9.02% 81.67% 3.21%
Perfluoro-n-hexanoic acid (PFHxA) 100.67% 9.02% 91.08% 6.64%
Perfluoro-n-nonanoic acid (PFNA) 96.33% 3.21% 92.67% 2.52%
Perfluoro-n-octanoic acid (PFOA) 82.77% 3.87% 85.53% 2.25%
Perfluoro-n-pentanoic acid (PFPeA) 79.50% 0.71% 85.67% 4.04%
Perfluoro-n-tetradecanoic acid (PFTeDA) 86.67% 4.73% 88.67% 1.15%
Perfluoro-n-tridecanoic acid (PFTrDA) 63.00% 3.46% 68.33% 3.51%
Perfluoro-n-u0ecanoic acid (PFUdA) 76.33% 1.53% 79.67% 3.51%
Perfluorooctanesulfonic acid (PFOS) 84.00% 4.00% 79.60% 0.53%

Temperature: 65 °C Conclusion
Hold Time: 4 min The analytical assessment of PFAS compounds is critical because of their
Wash 1 widespread nature, high stability, and adverse health effects. The EDGE
Wash Solvent: Methanol was able to rapidly and efficiently extract spiked soil samples with excel-
Wash Volume: 10 mL lent recoveries and RSD values. The EDGE also saw no carryover after
Temperature: 50 °C extractions into the subsequent extraction. The EDGE is an excellent ex-
Hold: 3 s traction tool for laboratories seeking to automate their PFAS extractions
Wash 2 with great efficiency.
Wash Solvent: 80:20 methanol:water with 0.3% ammonium hydroxide
Wash Volume: 10 mL
Temperature: None
Hold: None

Results
The recovery data in Table I, from the low and high spikes, indicated that
the samples were extracted with high efficiency, with recoveries ranging
from 63% to 101%. The resulting RSD values were also low, indicating
the recovery data were reproducible. The extraction data from the empty CEM Corporation
Q-Cup indicated that there was no carryover of the spiked compounds 3100 Smith Farm Road, Matthews, NC 28104
within the system, indicating the wash was aggressive enough to remove tel. (800) 726-3331
any residual PFAS compounds. Website: www.cem.com

358 THE APPLICATION NOTEBOOK – JUNE 2020

 MEDICAL/BIOLOGICAL

Tailored Analysis of Phospholipid Classes

Using iHILIC-Fusion(+) as First Dimension for Online
Two-Dimensional HILIC–Reversed-Phase LC–MS
Patrick O. Helmer*, Carina M. Wienken*, Wen Jiang†, and Heiko Hayen*,
*Institute
of Inorganic and Analytical Chemistry, University of Münster, Münster, Germany, †Hilicon AB

Phospholipids (PLs) are a large lipid subgroup that is involved 1D HILIC Separation:
in important cell functions in all organisms. They are the Column: 20 × 2.1 mm, 5-µm, 100 Å, iHILIC®-Fusion(+) (P/N
main components of biological membranes and are essential 100.022.0510, HILICON)
for many biologic processes within and between cells, for Eluents: A) ammonium formate solution (35 mM, pH 3.5) and
example, signal transduction, cell growth, and various 95:5 (v/v) acetonitrile; B) acetonitrile
transport processes. The basic building blocks for PLs are fatty Gradient Elution: 0–0.2 min, 97% B; 0.2–0.5 min, from 97%
acids linked to a glycerol backbone and polar head groups to 93% B; 0.5–2.75 min, 93% B; 2.75–7.5 min, from 93% to
(Figure 1). In addition to the main membrane PLs, such as 60% B; 7.5–11 min, 60% B; 11–11.5 min, from 60% to 97%
phosphatidylcholine (PC) and phosphatidylethanolamine B; re-equilibration is parallel to reversed-phase LC separation
(PE), other PL classes are of great importance as well. Some at 97% B.
PLs have significantly lower concentrations. For example, Flow Rate: 0.3 mL/min; 0.05 mL/min (during parallel
in eukaryotic organisms, cardiolipin (CL) is exclusively reversed‑phase LC separation)
located in the inner mitochondrial membrane (1). Thus, Column Temperature: 40 °C
the concentration of CL in total lipid extracts is very low in Injection Volume: 2–20 µL
comparison to the major membrane lipids, such as PLs and Heart-Cut Setup for PL Transfer: An online heart-cut 2D
cholesterol, but also triacylglycerols (TGs) (2,3). Furthermore, HILIC–reversed-phase LC–MS setup was developed for PL
ion suppression effects complicate the mass spectrometric
(MS) analysis, and a tailored analysis of low abundant PL
classes is often required.
In our previous study, we demonstrated the advantage
of performing sample preparation by solid-phase extraction
(SPE) in hydrophilic interaction liquid chromatography
PC
(HILIC) mode (4). The nonpolar fraction of lipids was
first separated from polar PLs using an offline HILIC SPE
PE
method, and thereafter separation by reversed-phase liquid
chromatography (LC) was performed. Offline methods often
require solvent evaporation and reconstituting analytes in
a suitable solvent for the second dimension, which is a CL
time‑consuming approach and risks losing labile PLs. In this
application, a novel online heart-cut two-dimensional (2D)
HILIC–reversed-phase LC–MS method is presented for the
separation of PL classes and nonpolar lipids, where HILIC was
polar
utilized in the first dimension (1D) and reversed-phase LC in
second dimension (2D). TG cholesterol

Experimental
Lipid Standards: Triacylglycerol (TG 48:0), cholesterol (Chol),
phosphatidylcholine (PC 32:0), phosphatidylethanolamine
(PE 32:0), cardiolipin (CL 64:4), and yeast total lipid extract
(Saccharomyces cerevisiae).
LC–MS/MS Setup: Thermo Scientific Ultimate 3000 system nonpolar
with dual gradient pump hyphenated to a Q Exactive™ Plus
Hybrid Quadrupole-Orbitrap™ mass spectrometer. Ionization
Figure 1: Selected structures of nonpolar lipids and polar PLs. HILIC
was performed utilizing electrospray ionization in negative separation of PLs takes place according to their polar head groups
ionization mode as described by Helmer et al. (3). (shown in green).

THE APPLICATION NOTEBOOK – JUNE 2020 359

MEDICAL/BIOLOGICAL

transfer from 1D (HILIC) to 2D (reversed-phase LC). As described

earlier, a six-port valve equipped with a sample loop (300 µL,
PEEK) was utilized as the transfer (Figure 2) (3). At first, the HILIC
eluate was directed to the waste. By switching the valve from A (a) TG PE CL PC
(2_1) to B (1_6) position, the selected PL fraction was transferred lipid standards
into the loop for the separation in the second dimension. Since we
focused on CL in this study, the transfer time window was set at its 1D HILIC (b)
retention time of 6.35–6.95 min.
CL 66:4 S. cerevisiae
CL 68:4
Results and Conclusion
CL 70:4
A tailored analysis of CL species in yeast (S. cerevisiae) was
achieved utilizing two-dimensional HILIC–reversed-phase (c)
LC–MS. In the first dimension, the separation of PL classes was
carried out using a HILIC method. In addition to that, nonpolar 1D-RPLC
2D RPLC
lipids did not show retention and were eluted at the void volume. 0 15 30
As shown in Figure 3(a), the lipid classes (TG, PE, CL, and PC) yeast CL
were well separated according to their different head groups
12 20
with this new method. The application of this method to a yeast Time (min)
total lipid extract (Figure 3[b]) shows the intended coelution
of the CL species in yeast, where one typical CL species is CL
Figure 3: (a) 1D HILIC separation of representative lipid standards;
66:4 with 66 carbon atoms and four double bonds in the four (b) coelution of dominant CL species in yeast (S. cerevisiae) within the
linked fatty acids. transfer window; (c) comparison of reversed-phase LC TICs in 1D (red)
In the developed 2D HILIC–reversed-phase LC separation and 2D (black).
method, a transfer window with all CL species from 1D
HILIC was first determined for the transfer via a heart-cut second dimension separation of low abundant PL classes.
setup to the 2D reversed-phase LC system. Compared to the Interestingly, low abundant CL species can be sensitively
conventional one-dimensional (1D) reversed-phase LC (red), detected and observed, even in the TIC. In summary, the
2D reversed‑phase LC (black) shows a significant decrease automated online 2D method via heart-cut transfer offers a fast
of background signals (mainly TGs), and low abundant CL and gentle sample preparation and clean-up procedure that is
species can even be observed in the total ion current (TIC) very useful for analyzing low abundant PL classes as well as
chromatogram (Figure 3[c]). Furthermore, by the reduction of labile PLs, such as PL oxidation products (3).
the background as observed in 1D reversed-phase LC, higher
CL signal intensities have been observed (3). References
The selective HILIC pre-separation of lipid classes in the (1) J.M. Berg et al., Stryer Biochemistry (Springer), (2018).
first dimension significantly reduced the matrix effects in the (2) M. Lange et al., Chromatographia 82, 77–100 (2018).
(3) P.O. Helmer et al., J. Chromatogr. A. 460918, in press https://doi.
org/10.1016/j.chroma.2020.460918
(4) P.O. Helmer et al., Rapid Commun. Mass Spectrom. 34, e8566 (2020).
https://doi.org/10.1002/rcm.8566

Pump 1 HILIC Waste

2 1
Position A: 2_1
3 Valve 6
Position B: 1_6
4 5

Pump 2 RP-MS

HILICON AB
Tvistevägen 48, SE-90736 Umeå, Sweden
Tel.: +46 (90) 193469
Figure 2: Valve setup for the transfer of PL classes from 1D HILIC E-mail: info@hilicon.com
to 2D reversed-phase (RP) LC via heart-cut approach. Website: www.hilicon.com

360 THE APPLICATION NOTEBOOK – JUNE 2020

 MEDICAL/BIOLOGICAL

Online SPE and LC–MS Analysis of Thyroid

Hormones in Human Serum
MilliporeSigma

Thyroid hormones play critical roles in the regulation of biological

processes, including growth, metabolism, protein synthesis, and 4750
1:651.80>605.50(+) CE: -24.0 1:777.70>731.80(+) CE: -28.0

brain development. Specifically, both 3,3’,5,5’-tetraiodo-L-thyronine 4500

4250
4000

(thyroxine or T4) and 3,3’,5-triiodo-L-thyronine (T3) are essential for 3750

3500
3250

development and maintenance of normal physiological functions. For 3000

2750
2500
a clinical laboratory, measurements of total T4 and total T3, along with 2250
2000
1750
estimates of free T4 (FT4) and free T3 (FT3), are important for the 1500
1250

diagnosis and monitoring of thyroid diseases. Most clinical laboratories

1000
750
500

measure thyroid hormones using immunoassays, which offer a relatively 250

0
-250

rapid, high patient sample throughput that lends itself to automation, but 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Time (min)
5.5 6.0 6.5 7.0 7.5

are compromised by problems with assay interference and by changes
in protein levels that alter free hormone availability (1).
Liquid chromatography–mass spectrometry (LC–MS) can offer superior
specificity and speed over immunoassays for determination of thyroid
hormones in biological matrices such as serum and tissues. The present
work demonstrates successful online solid-phase extraction (SPE) with LC–
MS for rapid determination of T4, T3, and 3,3’,5’-triiodo-L- thyronine (rT3)
from biological matrices.
Experimental
Materials: Supel™ Genie RP-Amide (RPA) and C8 (results not shown)
online SPE cartridges (2 cm × 4.0 mm i.d.); human serum; and protein
precipitation solvent: methanol with 1% (w/v) ammonium formate.
Sample Processing: Human serum spiked with analytes was protein
precipitated by vortex mixing with the precipitation solvent at a 1:3
ratio. The mixture was then centrifuged at 10,000 ×g for 3 min and
the supernatant was collected and directly injected for LC–MS analysis.
Online SPE–LC–MS Setup: Six-port switching valve and two pumps; one
for sample loading and washing, the other for sample elution. To minimize
potential peak broadening from the cartridges, the flow of sample loading–
washing and the subsequent elution are in reversed directions.
Figure 1: Representative LC–MS chromatogram of thyroid hormones
in human serum with RPA online cartridge.
time. We used RPA online cartridges with LC–MS analysis for the
detection of thyroid hormones from human serum. The samples were
protein precipitated with methanol containing ammonium formate
and after centrifugation then directly injected for online SPE and
LC–MS analysis. Sample loading–washing was performed entirely by
the instrument, eliminating the time-consuming solvent evaporation
and reconstitution steps.
The RPA cartridges captured a trace amount (100 ng/mL × 2 µL in
this case) of thyroid hormones from human serum (Figure 1). All three
analytes were well resolved with a peak width at half height <6 s and
tailing factor from 1.5–1.8. The total run time was within 6 min.
An online SPE–LC–MS method can be used for rapid detection of
thyroid hormones in human serum with minimal hands-on effort and
time-consuming steps.
Reference
(1) N. Kahric-Janicic, S.J. Soldin, O.P. Soldin, T. West, J. Gu, and J. Jonklaas,
Thyroid 17(4), 303–11 (2007).

Results and Discussion

Conventional SPE typically involves multiple labor-intensive and time-
consuming steps: conditioning, sample loading, washing, elution, MilliporeSigma
evaporation, and reconstitution. Supel Genie online cartridges were 400 Summit Drive, Burlington, MA 01803
developed to automate the sample preparation process, minimize tel. (800) 645-5476
hands-on time and human error, and reduce overall sample processing Website: www.sigmaaldrich.com

Table I: Ruggedness of online SPE–LC–MS with RPA cartridges from 120 consecutive injections of human serum samples.
Peak Area
Retention Time (Min) Retention Time Reproducibility Peak Area
Analyte MRM Quantifier Reproducibility
(Avg. n = 120) (%RSD, n = 120) (Avg. n = 120)
(%RSD, n = 120)
3,3’,5-triiodo-L-thyronine (T3) 651.8 / 605.5 4.03 0.2 27046 5.1
3,3’,5-triiodo-L-thyronine (rT3) 651.8 / 605.5 4.43 0.2 33723 6.2
3,3’,5,5’-tetraiodo-L-thyronine (T4) 777.7 / 731.8 4.79 0.2 23766 7.7

THE APPLICATION NOTEBOOK – JUNE 2020 361

MEDICAL/BIOLOGICAL

Analysis of Fentanyl and Its

Analogues in Human Urine
by LC–MS/MS
Shun-Hsin Liang and Frances Carroll,Restek Corporation

Abuse of synthetic opioid prescription painkillers such as

fentanyl, along with a rapidly growing list of illicit analogues,
is a significant public health problem. In this study, we
developed a simple dilute-and-shoot method that provides
a fast 3.5-min analysis of fentanyl and related compounds
(norfentanyl, acetyl fentanyl, alfentanil, butyryl fentanyl,
carfentanil, remifentanil, and sufentanil) in human urine by
liquid chromatography–tandem mass spectrometry (LC–MS/
MS) using a Raptor Biphenyl column.
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00
Time (min)

In recent years, the illicit use of synthetic opioids has skyrocketed,

and communities worldwide are now dealing with an ongoing Figure 1: The Raptor Biphenyl column effectively separated all target
epidemic. Of the thousands of synthetic opioid overdose deaths compounds in urine with no observed matrix interferences. Peak
per year, most are related to fentanyl and its analogues. With elution order: norfentanyl-D5, norfentanyl, remifentanil, acetyl fentanyl-
13C , acetyl fentanyl, alfentanil, fentanyl-D , fentanyl, carfentanil-D ,
their very high analgesic properties, synthetic opioid drugs 6 5 5
carfentanil, butyryl fentanyl, sufentanil-D5, sufentanil.
such as fentanyl, alfentanil, remifentanil, and sufentanil are
potent painkillers that have valid medical applications; however, sample preparation procedure was coupled with a fast (3.5 min)
they are also extremely addictive and are targets for abuse. In chromatographic analysis using a Raptor Biphenyl column. This
addition to abuse of these prescription drugs, the current opioid method provides accurate, precise identification, and quantitation
crisis is fueled by a growing number of illicit analogues, such as of fentanyl and related compounds, making it suitable for a variety
acetyl fentanyl and butyryl fentanyl, which have been designed of testing applications, including clinical toxicology, forensic
specifically to evade prosecution by drug enforcement agencies. analysis, workplace drug testing, and pharmaceutical research.
As the number of opioid drugs and deaths increases, so does
the need for a fast, accurate method for the simultaneous analysis Experimental Conditions
of fentanyl and its analogues. Therefore, we developed this LC–
MS/MS method for measuring fentanyl, six analogues, and one Sample Preparation
metabolite (norfentanyl) in human urine. A simple dilute-and-shoot The analytes were fortified into pooled human urine. An 80 µL
urine aliquot was mixed with 320 µL of 70:30 water–
Table I: Analyte transitions methanol solution (fivefold dilution) and 10 µL of in-
ternal standard (40 ng/mL in methanol) in a Thomson
Precursor Product Ion Product Ion
Analyte Internal Standard SINGLE StEP filter vial (Restek cat. #25895). After fil-
Ion Quantifier Qualifier
Norfentanyl 233.27 84.15 56.06 Norfentanyl-D5 tering through the 0.2 µm PVDF membrane, 5 µL was
13 injected into the LC–MS/MS.
Acetyl fentanyl 323.37 188.25 105.15 Acetyl fentanyl- C6
Fentanyl 337.37 188.26 105.08 Fentanyl-D5
Calibration Standards
Butyryl fentanyl 351.43 188.20 105.15 Carfentanil-D5
and Quality Control Samples
Remifentanil 377.37 113.15 317.30 Norfentanyl-D5 The calibration standards were prepared in pooled human
Sufentanil 387.40 238.19 111.06 Sufentanil-D5 urine at 0.05, 0.10, 0.25, 0.50, 1.00, 2.50, 5.00, 10.0,
Carfentanil 395.40 113.14 335.35 Carfentanil-D5 25.0, and 50.0 ng/mL. Three levels of QC samples (0.75,
Alfentanil 417.47 268.31 197.23 Acetyl fentanyl-13C6 4.0, and 20 ng/mL) were prepared in urine for testing ac-
Norfentanyl-D5 238.30 84.15 — — curacy and precision with established calibration standard
Acetyl fentanyl-13C6 329.37 188.25 — — curves. Recovery analyses were performed on three differ-
ent days. All standards and QC samples were subjected to
Fentanyl-D5 342.47 188.27 — —
the sample preparation procedure described.
Sufentanil-D5 392.40 238.25 — —
LC–MS/MS analysis of fentanyl and its analogues was
Carfentanil-D5 400.40 340.41 — —
performed on an ACQUITY UPLC instrument coupled

362 THE APPLICATION NOTEBOOK – JUNE 2020

 MEDICAL/BIOLOGICAL

with a Waters Xevo TQ-S mass spectrometer. Instrument conditions to 50 ng/mL for fentanyl, alfentanil, acetyl fentanyl, butyryl fen-
were as follows, and analyte transitions are provided in Table I. tanyl, and sufentanil; from 0.10 to 50 ng/mL for remifentanil; and
from 0.25 to 50 ng/mL for norfentanyl and carfentanil. All analytes
Analytical column: Raptor Biphenyl (5 µm, showed acceptable linearity with r2 values of 0.996 or greater and
50 mm × 2.1 mm; cat. #9309552) deviations of <12% (<20% for the lowest concentrated standard).
Guard column: Raptor Biphenyl EXP guard column
cartridge, (5 µm, 5 mm × 2.1 mm; Accuracy and Precision
cat. #930950252) Based on three independent experiments conducted on multiple
Mobile phase A: 0.1% Formic acid in water days, method accuracy for the analysis of fentanyl and its ana-
Mobile phase B: 0.1% Formic acid in methanol logues was demonstrated by the %recovery values, which were
Gradient Time (min) %B within 10% of the nominal concentration for all compounds at all
0.00 30 QC levels. The %RSD range was 0.5–8.3% and 3.4–8.4% for in-
2.50 70 traday and interday comparisons, respectively, indicating accept-
2.51 30 able method precision (Table II).
3.50 30
Flow rate: 0.4 mL/min Conclusions
Injection volume: 5 µL A simple dilute-and-shoot method was developed for the quantita-
Column temp.: 40 °C tive analysis of fentanyl and its analogues in human urine. The
Ion mode: Positive ESI analytical method was demonstrated to be fast, rugged, and sen-
sitive with acceptable accuracy and precision for urine sample
Results analysis. The Raptor Biphenyl column is well suited for the analy-
Chromatographic Performance sis of these synthetic opioid compounds, and this method can
All eight analytes were well separated within a 2.5-min gradient be applied to clinical toxicology, forensic analysis, workplace drug
elution (3.5-min total analysis time) on a Raptor Biphenyl column testing, and pharmaceutical research.
(Figure 1). No significant matrix interference was observed to neg-
atively affect quantification of the fivefold diluted urine samples.
The 5-µm particle Raptor Biphenyl column used here is a superfi-
cially porous particle (SPP) column. It was selected for this meth-
od in part because it provides similar performance to a smaller
particle size fully porous particle (FPP) column, but it generates
less system back pressure.

Linearity
Linear responses were obtained for all compounds and the calibra- Restek Corporation
tion ranges encompassed typical concentration levels monitored 110 Benner Circle, Bellefonte, PA 16823
for both research and abuse. Using 1/x weighted linear regression tel. 1 (814) 353-1300
(1/x2 for butyryl fentanyl), calibration linearity ranged from 0.05 Website: www.restek.com

Table II: Accuracy and precision results for fentanyl and related compounds in urine QC samples

QC Level 1 (0.750 ng/mL) QC Level 2 (4.00 ng/mL) QC Level 3 (20.0 ng/mL)

Average Conc. Average % Average Conc. Average % Average Conc. Average %

Analyte %RSD %RSD %RSD
(ng/mL) Accuracy (ng/mL) Accuracy (ng/mL) Accuracy
Acetyl fentanyl 0.761 102 1.54 3.99 99.7 2.08 19.9 99.3 0.856
Alfentanil 0.733 97.6 3.34 3.96 98.9 8.38 20.9 104 6.73
Butyryl fentanyl 0.741 98.9 6.29 3.77 94.3 6.01 20.8 104 4.95
Carfentanil 0.757 101 7.34 3.76 94.0 4.64 20.6 103 4.24
Fentanyl 0.761 102 1.98 3.96 99.1 2.31 19.9 99.6 1.04
Norfentanyl 0.768 103 6.50 4.04 101 1.84 20.1 101 2.55
Remifentanil 0.765 102 3.42 3.97 99.2 3.68 20.8 104 4.14
Sufentanil 0.752 100 1.67 3.93 98.3 1.28 20.1 100 0.943

THE APPLICATION NOTEBOOK – JUNE 2020 363

MEDICAL/BIOLOGICAL

Chondroitin Sulfate Analyzed by SEC–MALS

Wyatt Technology

Chondroitin sulfate is used to treat arthritis and other conditions

involving cartilage. Different preparations of the product have
different molecular properties, which are related in turn to therapeutic DŽůĂƌDĂƐƐǀƐ͘dŝŵĞ භ CSLM
^>D
ϭ͘ϬdžϭϬ
1.0x106ϲ භ CS
^
effectiveness. Such properties can be studied—and quantified—by MALS Signal
D>^^ŝŐŶĂů dRI Signal
ĚZ/^ŝŐŶĂů
light scattering.
Two chondroitin sulfate samples were compared by means of 1.0x105ϱ
ϭ͘ϬdžϭϬ
>ŽǁͲDwǁ Chondroitin
Low-M ŚŽŶĚƌŽŝƚŝŶ
a DAWN multi-angle light scattering (MALS) instrument (Wyatt Sulfate
^ƵůĨĂƚĞ

Mass (g/mol)
DŽůĂƌDĂƐƐ;ŐͬŵŽůͿ
Technology) coupled to size-exclusion chromatography (SEC). A 1.0x104ϰ
ϭ͘ϬdžϭϬ
differential refractive index (dRI) detector served to measure
dRI Signal
ĚZ/^ŝŐŶĂů
concentration. MALS and dRI measurements are combined in

Molecular
ϭ͘ϬdžϭϬ
1.0x103ϯ
ASTRA software to calculate molar mass distributions, moments
such as the weight-average molar mass Mw, and the polydispersity
index Mw/Mn. 1.0x102Ϯ
ϭ͘ϬdžϭϬ
Analysis by MALS relies on first principles and therefore
overcomes the uncertainties associated with column calibration, 1.0x101ϭ
ϭ͘ϬdžϭϬ
which arise from differences in elution properties between the 6.0
ϲ͘Ϭ 8.0
ϴ͘Ϭ 10.0
ϭϬ͘Ϭ 12.0
ϭϮ͘Ϭ 14.0
ϭϰ͘Ϭ
dŝŵĞ;ŵŝŶͿ
Time (min)
sample and reference molecules related to conformation, density,
and column interactions. SEC–MALS also eliminates the need to
perform frequent recalibration. Figure 1: Molar mass versus time for two chondroitin sulfate samples
measured by SEC–MALS. The dRI signals for both samples and the LS signal
for one sample are superimposed.
Results
Figure 1 shows chromatograms and molar masses obtained from
MALS measurements for two samples, CS and CSLM. The dRI trace Conclusions
alone is shown for CSLM in red, while both dRI and MALS signals The SEC–MALS method provides complete information about the mo-
are shown for CS, in blue. Not all of the differences between the two lar mass of chondroitin sulfate, determining distributions and a full
samples appear in standard SEC analysis, but are readily revealed set of moments. The analysis indicated similarities and discrepancies
by SEC–MALS. between the two samples. The presence of a high-molecular-weight
MALS indicates that CS has a Mw of 14.8 ± 0.2 kDa and the peak in CS, not observed by dRI, was detected unequivocally by MALS,
polydispersity index (Mw /Mn) is 1.21 ± 0.03. The low molar mass thanks to its enhanced sensitivity to high-molecular-weight species.
chondroitin sulfate LMCS sample has a Mw of 6.9 ± 0.2 kDa and a The fact that similar molar masses and sizes of each sample were
polydispersity of 1.4 ± 0.1. MALS analysis in fact determines a full eluted at the same time is a good indication of similar conformation
set of moments including Mn, Mw, and Mz, as well as cumulative and and chemical properties. Absolute macromolecular characterization
differential molar mass distributions (not shown). by SEC–MALS provides rich, detailed information without the biases
inherent in standard analytical SEC, for confidence in essential
• Two additional peaks were observed in both samples at late physiochemical properties.
elution times, though at different ratios to the main peak. The
molar masses of these species are calculated by MALS as just
a few hundred g/mol.
• In the CS sample, an early peak at 7 min is observed in the
MALS trace but is barely present in the dRI trace. This high ratio
of MALS to dRI is typical of high-molar-mass species; further
analysis shows that this peak contains a fraction with molar
mass ranging from 40–500 kDa. Wyatt Technology Corporation
6330 Hollister Ave., Santa Barbara, CA 93117
tel. (805) 681-9009
Website: www.wyatt.com

364 THE APPLICATION NOTEBOOK – JUNE 2020

 FOOD/BEVERAGE

HILIC Columns for Phosphorylated Sugar Analysis

Showa Denko America, Inc.

Various types of phosphorylated saccharides are involved with

glycolysis, glyconeogenesis, glycometabolism, and other naturally
occurring biological phenomena. One example of a phosphorylated
1. Glucose-6-phosphate
saccharide includes fructose-2,6-bisphosphate, the most powerful 3 (G6P)
activator of phosphofructokinase. Recent research has shown 2. Fructose-6-phosphate
(F6P)
the phosphorylated saccharide intermediates created during 1 3. Glucose-1-phosphate
2
glyconeogenesis and glycometabolism can be used in other ways (G1P)
4. Fructose-1-phosphate
than naturally intended for a number of applications, drastically (F1P)
broadening the interest in phosphorylated saccharide. 4
Phosphorylated saccharides occupy a central position in
carbohydrate metabolism. Numerous sugar interconversions occur
when the sugars are linked to phosphates and di-phosphates.
When this occurs, the phosphorylated saccharides can function
as glycosyl donors in many transglycosylation reactions, providing 0 2 4 6 8 10 12 14 min
a variety of oligosaccharides, glycosides, and polysaccharides.
Phosphorylated sucrose and fructose have been synthesized by
the transglycosylation process in plants, glycogen in animal tis- Figure 1: The analysis of phosphorylated saccharides using the Shodex
sues, and chitin. The isolated phosphorylated sucrose has further VT-50 2D column.
shown to potentially be effective in experimental drugs. How the Column: HILICpak VT-50 2D, Column temperature: 60 ºC, Injection
new drugs interact within the body is directly affected by which volume: 5 µL, Eluent: 25 mM HCOONH4 aq./CH3CN : 80/20, Flow
carbon contains the phosphate group. Phosphorylated fructose rate: 0.3 mL/min. Detector: ESI–MS (SIM Negative: m/z 259).
has been recently suggested to be a part of the creation of ATP. Sample: 1-and 6- phosphorylated saccharides.
Phosphorylated mannose and glucose were found to be potent
inhibitors of lysosomal hydrolase endocytose by fibroblasts, plants, HPLC system was coupled with an ESI–MS (SIM negative: m/z
and bacteria. It was recently proposed that phosphorylated 259) detector.
mannose is an essential component of the recognition marker on
acid hydrolases. This was direct evidence from the phosphorylated Results
mannose’s ability to uptake forms of lysosomal enzymes. The aqueous sample containing the four phosphorylated sugars
With the knowledge of these few applications, the field has was analyzed successfully using HILIC and MS detection with the
been exponentially growing. Shodex HILICpak VT-50 2D column (Figure 1). From this single
Four variations of phosphorylate saccharides were analyzed by a run, each phosphorylated saccharide was prominently separat-
Shodex HILICpak VT-50 2D column under LC–MS conditions. The ed. The closest peaks were glucose-6-phosphate and fructose-
sample contained two phosphorylated glucose and two phosphorylated 6-phosphate, both eluted between 6 and 7 min.
fructose, each phosphorylated at a different carbon, providing different
biological functions. This analytical condition can also be used with other
detectors including reflective index (RI), evaporative light scattering
detector (ELSD), and corona charged aerosol detector (CAD).

Experimental Conditions
The analysis of glucose-6-phosphate, fructose-6-phosphate, glucose- TM

1-phosphate, and fructose-1-phosphate was accomplished using

the Shodex HILICpak VT-50 2D (2.0 mm ID × 150 mm ID, 5 µ), a
HILIC column suitable for LC–MS. The column temperature was Shodex™/Showa Denko America, Inc.
60 ºC and the flow rate was 0.3 mL/min. The eluent conditions 420 Lexington Avenue Suite 2335A, New York, NY, 10170
were 25 mM HCOONH4 aq./CH3CN = 80/20. An injection volume tel. (212) 370-0033 x109
of 5 µL of 1 µM of each sugar was used for the experiment. The Website: www.shodexhplc.com

THE APPLICATION NOTEBOOK – JUNE 2020 365

PHARMA/DRUG

Amino Acid Analysis by European Pharmacopeia

to Support Coronavirus Research
Pickering Laboratories, Inc.

Pharmaceutical companies worldwide are currently searching

for vaccines, therapeutic agents, and prophylactics to combat

Ammonia
COVID-19. The focus on biologic drugs during the current
pandemic highlights the crucial role amino acids analysis
plays in research and production of pharmaceuticals.

Lysine
Threonine
The European Pharmacopoeia (Ph. Eur.) defines requirements for

Methionine

Phenylalanine
Isoleucine
Valine
the qualitative and quantitative composition of medicines, as well

Leucine
as the tests to be carried out on medicines and on substances and
materials used in their production.

Proline
Pickering Laboratories, Inc. offers a complete solution for amino
acids analysis according to Ph. Eur.. This includes the Onyx PCX 0 40 min
10 20 30
post-column derivatization instrument, analytical columns and
GARDs, buffers, and Trione® ninhydrin reagent. The methods pre- Figure 2: Sodium chromatogram of alternative amino acids ana-
lyzed using Ph. Eur. 9.0 methods (3 ug/mL each, 50 uL injection).
sented in this application note were optimized to comply with system
suitability requirements of Ph. Eur. 9.0 methods.

Ammonia
Equipment
• Quaternary HPLC pump, autosampler, UV-vis detector
• Onyx PCX post-column derivatization system

Analytical columns and eluants

Isoleucine
Leucine
Cystine

For Sodium-based methods: High-efficiency sodium cation-exchange

column, 4.6 x 110 mm, catalog number 1154110T. Eluants: Na315,
Cysteine
Proline

Na425, Na640, RG011

For Lithium-based methods: High-efficiency lithium cation-exchange 0 10 20 30 40 50 min

column, 4.6 x 75 mm, catalog number 0354675T. Eluants: 1700–

Figure 3: Lithium chromatogram of amino acids used as reference
1125, Li365, Li375, RG003
solutions for cysteine analysis (3 ug/mL each, 50 uL injection).

Post-column reagent To make it easier to start using Pickering Laboratories methods, we offer
Trione® ninhydrin reagent chemistry kits that include an analytical column, GARD buffers, and
reagents for amino acids analysis. All parts of the kit could be ordered
individually if needed. Please contact Pickering Laboratories if you have
Ammonia

any questions regarding this application.

Serine

Alanine
Threonine

Lysine
Isoleucine
Leucine
Valine

Arginine
Cystine
Proline

Pickering Laboratories, Inc.

min
0 10 20 30 40
1280 Speca Park Way, Mountain View, CA 04043
Figure 1: Sodium chromatogram of amino acids analyzed using Ph. Eur. tel. (800) 654-3330, (650) 694-6700
9.0 methods (3 ug/mL each, 50 uL injection). Website: www.pickeringlabs.com

366 THE APPLICATION NOTEBOOK – JUNE 2020

 PHARMA/DRUG

Size-Exclusion Chromatography for the Impurity

Analysis of Adeno-Associated Virus Serotypes
Stephan M. Koza and Weibin Chen,Waters Corporation

Monitoring the size heterogeneity of AAV-based gene
therapy therapeutics is potentially important to ensure
consistent product quality and efficacy. We demonstrate
that the levels of both high-molecular weight (HMW) and
low-molecular weight (LMW) impurities in AAV capsid
preparations can be separated on a Waters™ 450 Å Protein
BEH SEC column for a series of AAV serotypes.
As the development of gene therapy products accelerates, the
need to develop sound and efficient analytical strategies to
help guide the development of manufacturing processes and
evaluate the quality of clinical adeno-associated virus-(AAV)–
based gene therapy materials has become more important.
Among other critical quality attributes, the levels of potential
AAV high molecular weight (HMW) species and AAV fragments
or low molecular weight (LMW) species may also require
monitoring (1). Here we present optimized non-denaturing
size-exclusion chromatography (SEC) methods that can
separate soluble AAV for several AAV-CMV-GFP serotypes
including AAV1, AAV2, AAV5, AAV6, AAV8, and AAV9.
Experimental Conditions
AAV sample: AAV serotypes ranging from 1.6 × 1012 to
6.7 × 1013 GC/mL
LC system: ACQUITY™ UPLC™ H-Class Bio
Sample temp.: 6 ˚C
Column temp.: 25 ˚C
Injection volume: 3 µL
Column: XBridge™ Protein BEH SEC, 450 Å,
3.5 µm, 7.8 × 300 mm
Fluorescence
detector: Excitation: 280 nm; Emission: 350 nm
Mobile phase: 10 mM NaH2PO4, 10 mM Na2HPO4,
200–400 m KCl, pH 6.6 (HCl)
Flow rate: 0.6 mL/min
Results and Discussion
The upper analyte size limitation for SEC separations of
proteins is approximately 100–200 nm (subvisible) depending
on SEC particle size. Above these limits, the analyte may be
altered due to shear forces or trapped by the frits or packed
bed of the column. Therefore, these subvisible entities are
analyzed using dynamic light scattering and nanoparticle
tracking analysis (NTA) methods, among others (2). With these
considerations, the SEC separation of soluble AAV monomer,
and HMW, and LMW was evaluated on a 450 Å pore-size
BEH diol-bonded SEC column. This column was previously
demonstrated to be effective in the separation of IgM pentamer

ACQUITY UPLC Protein BEH SEC, 450 Å, 2.5 µm, 4.6 x 300 mm Column
0.05
Flow rate: 0.35 mL/min 5
0.045 Injection volume: 5 µL
0.04 4
0.035
UV (280 nm)

0.03

0.025 3
0.02
2
0.015 1a
0.01 1b
0.005

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9.5 10 10.5 11 11.5 12 12.5
Minutes

Figure 1: Shown is the SEC separation of several protein structures. Compounds are: 1a. IgM dipentamer (1.8 kDa), 1b. IgM pentamer (900 kDa), 2.
thyroglobulin (667 kDa), 3. apoferritin (443 kDa), 4. β-amylase (200 kDa), 5. IgG (150 kDa). UV absorbance (280 nm) and the identities of the peaks observed
for the IgM sample were confirmed by SEC–MALS analysis.

THE APPLICATION NOTEBOOK – JUNE 2020 367

PHARMA/DRUG

and dipentamer in Waters laboratories (Figure 1) and has a

reported diameter of 35 nm (3). AAV has a total protein and 1.0x107
7

ssDNA molecular weight of 5500 kDa; however, the compact RI

RI
MALS
structure has a diameter of only 25 nm (4). Therefore, it 8
8.0x108
Molar Mass
Molar Mass
was predicted that these high-efficiency SEC particles could

(g/mol)
provide the accessible pore volume needed. A 3.5 µm particle

Mass (g/mol)
6.0x1066
6.9 E6
6.9 E6 (± 0.3 E6)
E6)
size was selected over 2.5 µm particles to reduce potential

MolarMass
sample sieving and shearing effects. Intrinsic protein

(V)
voltage (V)
6
0.15
0.15

detectorvoltage
4.0x106

Molar
fluorescence detection was also employed to provide maximum 0.10
0.10
3.8
3.8 E6
E6 (±
(± 0.03 E6)
optical sensitivity.

detector
0.05
0.05

6
2.0x106
The SEC separation of a null control sample (AAV without
0.00
0.00
3.0 4.0 5.0 6.0 7.0
7.0 8.0 9.0
Minutes
Minutes

ssDNA) was evaluated using a Wyatt microDAWN Multi-Angle

Light Scattering (MALS) detector and a Waters ACQUITY 6.0 6.5 7.0
Minutes
Minutes
Refractive Index (RI) detector (Figure 2). The SEC–MALS data
confirmed that adequate separation between the dimeric and
Figure 2: SEC-MALS of AAV8-null sample using refractive index (RI) for
monomeric AAV forms was observed. Putative multimeric forms concentration measurement are shown. The MALS (red) and RI (blue)
preceding dimer were also observed by fluorescence but could signals are normalized and the average and distribution of determined
not be assigned molecular weights due to their low abundance. molar masses (green) were determined using Wyatt Astra (v. 7.3.1.9) based
on a dn/dc of 0.185 and using a “sphere” model for the icosahedral AAV.
A mobile-phase consisting of 20 mM sodium phosphate,
pH 6.6, with varying amounts of KCl (200–400 mM) was
found to maximize the recovery of dimer and multimer for 0.60
AAV1-GFP AAV6-GFP
30.00

serotypes AAV1, AAV2, AAV5, AAV6, AAV8, and AAV9 (Figure (200 mM KCl) 20.00
0.60
(350 mM KCl)
30.00

20.00

3). We observe that the peak shapes for both the dimer 0.40 10.00

D=1.6%
10.00
0.40
0.00 0.00

HMW species and monomeric AAV forms are symmetrical

0.00 5.00 10.00 15.00 20.00 25.00
0.00 5.00 10.00 15.00 20.00 25.00

0.20 0.20
M=0.5%
F=0.2%
and return to baseline appropriately. In addition, detectable D=0.2%
0.00
0.00

levels of multimeric AAV forms were observed for several of 1.00

AAV2-GFP AAV8-GFP
the serotypes. Significant amounts of LMW forms were only
40.00

(350 mM KCl)
100.00
0.80 2.00
(400 mM KCl)D=1.5%
D=1.1%
20.00
50.00
0.60

observed in AAV9 and AAV6.

FLR (relative)

0.00 M=1.5% 0.00

0.40 0.00 5.00 10.00 15.00 20.00 25.00

1.00
0.00 5.00 10.00 15.00 20.00 25.00

0.20

Conclusions 0.00
0.00

-0.20

A BEH SEC column with an average pore size of 450 Å

AAV5-GFP 20.00

AAV9-GFP 100.00

and a particle diameter of 3.5 µm was demonstrated to be 0.40

(250 mM KCl) 2.00
(250 mM KCl) D= 1.5%
15.00 80.00

10.00 60.00

D=1.2%
40.00
5.00

effective in the separation of AAV monomers from their HMW 0.20

0.00

0.00 5.00 10.00 15.00 20.00 25.00

1.00
20.00

0.00

M=0.9%
0.00 5.00 10.00 15.00 20.00 25.00

F=0.8%
dimers, lower valency multimers, and LMW fragments. The M=0.3%
F=0.3%

minimal amount of ionic strength (KCl) required for optimal 0.00 0.00

peak shape and recovery varied by serotype, and separations 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00

Minutes Minutes
are presented for serotypes AAV1, AAV2, AAV5, AAV6, AAV8,
and AAV9. Figure 3: Shown are the SEC separations of a series of AAV serotype
control samples containing ssDNA coding for green fluorescent protein
(GFP). The peak percentages for dimer (D), multimer (M), and fragments (F)
References
are provided. The chromatogram baselines are zoomed approximately 50x
(1) J. F. Wright, Gene Therapy 15(11), 840–8 (2008). versus the full-scale chromatogram shown in the inset.
(2) (a) J. F. Carpenter, T. W. Randolph, W. Jiskoot, D. J. Crommelin, C. R.
Middaugh, G. Winter, Y. X. Fan, S. Kirshner, D. Verthelyi, S. Kozlowski, K.
A. Clouse, P. G. Swann, A. Rosenberg, and B. Cherney, J Pharm Sci 98(4),
1201–5 (2009); (b) B. Slutter, and W. Jiskoot, Expert Opin. Drug Delivery
13(2), 167–70 (2016).
(3) R. Saber, S. Sarkar, P. Gill, B. Nazari, and F. Faridani, Sci. Iran. 18(6),
1643–1646 (2011).
(4) S. Vernon, J. Stasny, A. Neurath, and B. Rubin, J. Gen. Virol. 10(3), 267–
272 (1971). Waters Corporation
34 Maple Street, Milford, MA 01757
Waters, The Science of What’s Possible, ACQUITY, UPLC, and XBridge are trademarks tel. 508 478 2000
of Waters Corporation. All other trademarks are the property of their respective owners. Website: www.waters.com

368 THE APPLICATION NOTEBOOK – JUNE 2020


An Anion-Exchange Chromatography Method
for Monitoring Empty Capsid Content in
Adeno-Associated Virus Serotype AAV8
Hua Yang, Stephan M. Koza, and Weibin Chen,Waters Corporation
The single-stranded DNA (ssDNA) content of AAV capsids in
AAV-based gene therapy preparations impacts the efficacy
of this treatment modality. We demonstrate that AAV8
capsids without and with full length ssDNA can be separated
and their relative abundances determined on a Waters™
Protein-Pak™ Hi Res Q strong anion-exchange column.
During the recombinant adeno-associated virus (AAV)
biomanufacturing process, capsids that do not contain
ssDNA (empty capsids) are also produced. Empty capsids are
unable to deliver ssDNA to the targeted cells and as a result
are important to monitor during process development and
product manufacturing. An anion-exchange chromatography
(AEX) separation driven by differences in surface charge can
monitor changes in empty capsid levels (1,2). AEX using
fluorescent detection consumes small sample amounts and
can provide reliable and reproducible results.
Experimental Conditions
Samples: AAV8-CMV-GFP (full capsid) and AAV8-null (empty capsid)
2 × 1012 capsids/mL; other AAV-CMV-GFP serotypes: ranging from
1.3 × 1013 to 6.7 × 1013 GC/mL = genome copies (physical titer)/mL
LC system: ACQUITY™ UPLC™ H-Class Bio
Sample temp.: 10 ˚C
Column temp.: 30 ˚C
Injection volume: 0.2–6 µL
Column: Waters Protein-Pak Hi Res Q, 4.6 × 100 mm
20.00
EU
10.00
0.00
20.00
EU
10.00
0.00
20.00
EU
10.00
0.00
8.00 10.00 12.00 14.00
Minutes
Figure 1: AAV8 ssDNA (CMV-GFP) empty and full capsids are separated on a Protein-Pak Hi Res Q Column using an optimized AEX method (see
experimental conditions for method details).
THE APPLICATION NOTEBOOK – JUNE 2020
PHARMA/DRUG
Fluorescence
detector: Excitation: 280 nm; Emission: 350 nm
Gradient: 70 mM bis-tris propane, pH 9.0, 100–300 mM
tetramethylammonium chloride in 20 min
at 0.4 mL/min
Results and Discussion
The AEX separation of empty and full AAV8 capsid was optimized
for several parameters including pH and salt types (3). In addition,
intrinsic protein fluorescence was monitored to provide greater
sensitivity and selectivity, and less baseline drift (2). The separation
achieved is shown in Figure 1. We observe that the ssDNA “empty”
capsids are partially coelute with the ssDNA “full” capsids while
showing no significant change in the UV absorbance ratio (280
nm/260 nm, data not shown), indicating the unlikelihood of
monitoring AAV8 with partial ssDNA by AEX.
A series of chromatograms for AAV8 samples with differing
percentages of empty capsids is presented in Figure 2. The
mixtures were generated from >99% pure “empty” and “full”
samples. The sample concentrations were normalized based
on FLR peak areas from a size-exclusion chromatography
(SEC) separation with fluorescence detection and assumed a
“full” to “empty” fluorescence response factor (RF F/E) of 1.3 (2).
For AEX, however, 1.9 was calculated for RF F/E (Area 100% Full/
Area 100% Empty). The discrepancy in RF F/E values between SEC
and AEX may be the result of changes in quantum efficiency
under different mobile-phase conditions. RF F/E is then divided
into the full capsid peak area to calculate the percentage of
Empty/full capsid mix
Empty capsid
Full capsid
16.00 18.00 20.00
369
PHARMA/DRUG

24.00 100
0% Empty
22.00 90 100 x A Empty
12.5% Empty % Empty
80 RF F/E

Measured % Empty
20.00 25% Empty (A Empty +
18.00 50% Empty 70 A Full
16.00
75% Empty 60
100% Empty 50
14.00
40
EU

12.00
30
10.00
20
8.00 y = 0.9264x + 3.2146
10 R = 0.9997
6.00
0
4.00 0 20 40 60 80 100
2.00 Predicted % Empty
0.00

8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00

Minutes

Figure 2: Shown on the left is an overlay of AEX chromatograms for AAV8 samples with differing levels of empty capsid. On the right is the
correlation between predicted and measured empty capsid content. Calculation of the percentage of empty capsid was based on the equation
shown, which corrects for the difference in fluorescence response between empty and full capsids. AEmpty and AFull are the AEX peak areas of the
empty and full capsids, and RFF/E is the fluorescence response factor.

empty capsid in a sample (Figure 2). The accuracy of this

approach depends on the veracity of RF F/E, which requires
4.00
reasonably pure “full” and “empty” samples, and accurate AAV1
EU

2.00

capsid concentrations. While RF F/E was not rigorously defined 0.00

4.00

in this work, the estimate of RF F/E used here effectively

EU

AAV2
2.00

demonstrates the approach to interpretation of the AEX data. 0.00

Determining RF F/E is not essential, if only relative comparisons 2.00

EU

AAV5
of empty capsid levels between samples are required. 0.00
4.00

Samples of several AAV-CMV-GFP serotypes were also

AAV6
EU

2.00

evaluated (Figure 3). All tested serotypes were retained on 0.00

the column, but their retention varied significantly. 10.00

EU

AAV8
0.00

Conclusions
AAV9*
EU

5.00
AEX will not likely provide a measurement of AAV8 capsid
0.00
carrying partial ssDNA, therefore complementary methods F
E
10.00
EU

such as analytical ultracentrifugation (AUC) would be needed AAV E/F mix

0.00
for that analysis. However, due to the technical challenges 2.00 4.00 6.00 8.00 10.00 12.00 14.00
Minutes
16.00 18.00 20.00 22.00 24.00 26.00

and costs of these methods, AEX may still provide utility as

an initial screening tool in support of manufacturing process
development and as a more precise assessment of changes in
empty capsid content in product quality testing.
Due to differences in capsid protein surfaces, method
optimization may be needed when separating empty and
full capsids of other AAV serotypes, and the AEX method
described here can serve as a useful starting point.
Figure 3: Chromatograms of AAV1, AAV2, AAV5, AAV6, AAV8,
and AAV9 serotypes (full capsids, CMV-GFP ssDNA). For comparison,
the bottom chromatogram shows the optimized separation of AAV8
empty and full capsid. Separation conditions are as described in the
text for all serotypes except for AAV9 (*) where the salt gradient was
0–200 mM KCl.

References
(1) X. Fu, et al, Hum. Gene. Ther. Methods 30(4), 144 (2019).
(2) C. Wang, et al, Mol. Ther. Methods Clin. 19, 257 (2019).
(3) H. Yang, et al, Waters Application Note 720006825EN (2020).
Waters Corporation
Waters, The Science of What’s Possible, Protein-Pak, ACQUITY, and UPLC are 34 Maple Street, Milford, MA 01757
trademarks of Waters Corporation. All other trademarks are the property of their tel. 508 478 2000
respective owners. Website: www.waters.com

370 THE APPLICATION NOTEBOOK – JUNE 2020

WWW.CHROMATOGRAPHYONLINE.COM
FUN DA MENTAL S
Troubleshooting Gas Chromatography:
Reduced Peak Size (Loss of Sensitivity)
Tony Taylor
W e are frequently asked about issues
with reduced peak size in gas chro-
matography (GC). There are so many poten-
tial causes for reduced peak size that an inex-
perienced GC user may not know where to
begin in the troubleshooting process. Sen-
sitivity problems can be grouped by way of
their presentation within the chromatogram:
All peaks sizes (heights and or areas)
decrease–retention times do not change.
As always, it’s best to start with the most
obvious possible causes..
• If operating in split mode, check the inlet
split ratio in the acquisition method and
correct, if required.
• If splitless mode with pressure pulse,
check the inlet pulse pressure and dura-
tion in the acquisition method.
• Check that the inlet and detector temper-
atures have been set correctly within the
acquisition method.
• Check that the sample vial contains
sufficient liquid and use separate vials
containing the same sample liquid for
repeated injections to rule out the pos-
sibility of sample loss through a compro-
mised septum (especially when analytes
are highly volatile), and confirm the prob-
lem is reproducible.
• Check that the autosampler syringe
plunger is free within the barrel and does
not leak. Observe an injection cycle and
confirm that sample is being aspirated
from the vial and that it the correct volume.
• Check and replace the inlet septum as
required, and check that the correct liner
has been installed.
• Check that the sample preparation or dilu-
tions are correct as per the sample prepa-
ration method.
• With flame based ionizing detectors,
JUNE 2020  LCGC NORTH AMERICA  VOLUME 38 NUMBER 6  371
check that the fuel gas ratios are appro-
priate and that all flow rates are as they
should be (using a flow meter, checking
one gas flow at a time), and check all
applied voltages according to the manu-
facturers specification.
• With MS detectors in selected ion mode,
check the masses and dwell times of each
ion within each SIM group to ensure that
the cumulative ion count is matched to
the acquisition method.
• With MS detectors in scan mode, check
the MS tune and verify that the repeller
or accelerator voltage has not increased
dramatically (indicating a dirty ion source),
that the electron multiplier or MCP volt-
ages have not increased dramatically (indi-
cating a worn-out detector), and that the
ionization energy is correctly set (typically
70 eV for electron ionization).
• Check the detector attenuation range
and set appropriately, as a secondary
check. When the detector attenuation is
incorrectly set, the signal-to-noise ratio
for a given peak will not reduce, but the
chromatogram will appear ”smaller.”
All peaks sizes (heights and areas) decrease.
Retention times shift, no evidence of loss
of efficiency (peak broadening)
• Check that the correct column dimensions
have been entered into the data system
and that the correct column is installed,
paying attention to the stationary-phase
film thickness.
• Check that the carrier gas is flowing at the
correct volumetric flow rate using a cali-
brated flow meter.
• Check that the carrier gas flow program-
ming method is correct by selecting either
constant pressure or constant flow oper-
ating modes. In the latter, the carrier gas
pressure will be ramped to achieve con-
stant linear velocity of carrier through the
column as the oven temperature increases
and the carrier gas viscosity reduces.
• Change the inlet septum to overcome any
leaks that may be occurring during the
injection phase.
All peaks sizes (heights and or areas)
decrease–peaks are broadened, retention
time shift may occur.
• Check that the correct column has been
installed and column dimensions and car-
rier gas flow rate are appropriate.
• Check the column logs; if the column is
old or has been used with dirty sample
matrices, suspect that the column effi-
ciency might have been reduced.
• Run a column test mix and compare with
the result obtained originally.
• If loss of column efficiency is suspected,
trim 0.5–1 meters of column at the
inlet end.
• Verify that the column has been installed
to the correct distance within the GC inlet
and detector. For flame-based and ioniz-
ing detectors, check that the gas flows are
correct paying particular attention to any
make-up gas flow rates.
• When all peak heights or areas reduce
and the peaks broaden, the most obvious
cause is a loss of efficiency within the chro-
matographic system. This phenomenon
Icon image: Monster Ztudio/Adobe Stock
should also lead a reduction in signal-to-
noise ratio for analyte peaks.
Tony Taylor is the Chief Sci-
ence Officer of Arch Sciences
Group and the Technical
Director of CHROMacademy.
Direct correspondence to:
LCGCedit@mmhgroup.com.