How To Approach Identification of Unknown Peaks During pharmaceutical sample Analyses.

1. LIFE SCIENCE I TECHNICAL BULLETIN ISSUE N°49 / JULY 2011
HOW TO APPROACH THE IDENTIFICATION OF ANOMALOUS PEAKS
DURING PHARMACEUTICAL SAMPLE ANALYSES
AUTHOR: NICK TOLTL, PH.D., MANAGER, R&D, SGS LIFE SCIENCE SERVICES, CANADA
The sudden appearance of an unknown peak during an HPLC analysis of a pharmaceutical product can be a critical
finding causing delays and requiring considerable resource (both time and money) to resolve. The immediate need
to investigate an unknown peak is a common problem and is generally accompanied with aggressive timelines
and stressful conditions and the issue can be caused by many things which can range from simple sample prepa-
ration contamination, all the way to unexpected degradation/stability properties of the product. In any case, it is
essential that a rapid and effective investigation take place to quickly determine the identity of the anomalous peak
and determine the root cause of its presence in the product.
The source of an unknown peak can be peak in order to ascertain the impact this BEFORE ATTEMPTING PEAK
attributed to: peak may have on the affected product IDENTIFICATION
and ultimately the patient. However,
• Laboratory sample contamination determining the identity of the peak can The isolation and identification of an
• Instrument related peaks be challenging because: unknown peak is not a trivial exercise
• Method resolution (e.g. a vendor and if done incorrectly, can cause more
modification to HPLC column pack- 1. The level of the peak is usually small problems. Prior to making the decision
ing altering the column performance (0.05 – 0.2% area percent of the to begin the isolation and identification
revealing a previously undetected main peak) process, some in-lab investigations
entity) 2. There is limited knowledge on the and standard OOS practices should be
• Raw material impurities or contami- source of the peak performed.
nation 3. Without a structure, spectral proper-
• Product or aliquot degradation (a ties such as relative response factor 1. Re-perform the sample preparation
material stability issue or reaction are unknown in scrupulously pre-rinsed glassware
with another excipient) 4. The peak may not be well resolved to rule out laboratory contamination
• Leachables/extractables (from a from other peaks in the analysis 2. Re-analyze the sample on a different
manufacturing step or a container instrument
closure system) While each investigation is situation spe- 3. Analyze the raw materials used in
• Manufacturing process cross-con- cific, these are some general guidelines the manufacturing process individu-
tamination and strategies that can be utilized to ally using the drug product impurity
• Some other unknown source achieve success in these endeavours. methodology in an attempt to iden-
In a typical study, the strategy for tify the source of the peak
Regardless, once an impurity analysis on obtaining the necessary data to identify 4. Review historical data to determine
a drug product reveals that an anomalous the unknown peak is not “set in stone” . if the unknown entity had been
peak causing a failure of a specification, While there is certain flexibility to the present (maybe at lower levels) in
it jeopardizes the release of material into plan, the basic tools necessary to make a prior batches
production, or even has the potential to conclusive identification include (but are 5. If cross contamination is suspected,
initiate a product recall. not limited to) diode array, LC-MS/MS, review batch and equipment records
As a first step, one needs to assess if accurate mass MS, and NMR. to assess potential candidate com-
the peak is real and not an artifact. Once ponents
confirmed, it is critical to determine the
identity and the source of the unknown

2. LIFE SCIENCE I TECHNICAL BULLETIN 2
STAGE 1 – DIODE ARRAY STAGE 3 – LC-MS/MS AND but this exercise is not trivial. A substan-
ANALYSIS HIGH RESOLUTION MS ANALY- tial peak collection effort is required to
SES collect enough of the unknown peak
Once the decision is made to initiate the to perform an effective NMR analysis.
unknown peak investigation, the first Mass spectrometry is a powerful tool On a case by case basis, various peak
data that should be generated in an un- during the investigation of unknown fraction collection strategies can be
known peak investigation is a UV spec- peaks but caution must also be em- used to obtain enough sample for NMR.
trum using a diode array detector. This ployed when interpreting the data Unknown peaks (~0.1%) that are well
serves three purposes: 1) the UV spec- generated using this technique. Points resolved from interferences and have
trum can give some critical structural to consider when analyzing the MS data maximized the on-column load of the
information about the unknown peak, 2) include: material may still require anywhere from
it may also give an indication if the peak 20-40 injections/collections to obtain
is related to the parent molecule being 1. The unknown peak being analyzed the necessary material for 1D and 2D
analyzed and 3) it enables peak track- may not be homogenous NMR experimentation. However, in an
ing capability during situations where 2. The response of the peak in the MS ideal situation, it is beneficial to collect
method modifications are required. detector may not correlate with the enough material so that a weight can be
UV data measured. Once weighed, the extinction
STAGE 2 - USE OF VOLATILE 3. In cases where the peak is not coefficient of the unknown peak at the
BUFFERS TO ACHIEVE AN LC- homogenous, the MS will afford detection wavelength can be determined
MS COMPATIBLE METHOD data consistent with the moiety which enables the appropriate quantita-
in the peak that is the most stable tion of the unknown peak in the HPLC
In the event that the method being used ion, and not necessarily the largest impurity analysis.
utilizes volatile buffers that are compat- component
ible for LC-MS analyses, then method Upon collection, the mobile phase can
development is not required and one That being said, MS and MS/MS data be removed using nitrogen flow. The
can proceed directly to performing combined with structural libraries can use of heat is not recommended as the
mass spectral analyses. However, many be an effective way of accomplishing chemical stability of the unknown peak
methods in the pharmaceutical industry a structural elucidation of an unknown is not established. Once dry, the sample
use phosphate buffer systems in HPLC peak. In addition, the determination of can be submitted for NMR analyses.
mobile phase. Phosphate buffers are not a high resolution MS data may aid in Consultation with an NMR scientist will
compatible with LC-MS analyses so the the investigation. The use of a 10 ppm establish the appropriate NMR strategy
first step in the unknown peak investiga- tolerance for suggested molecular to pursue. The utilization of 1D and 2D
tion is to substitute a volatile buffer into formula matches may result in multiple techniques combined with spectral
the mobile phase. For low pH situations, hits. Therefore, information such as the prediction software should afford the
formic acid or ammonium acetate are presence of heavy atoms is important structure of the unknown peak.
suggested as mobile phase modifiers. to help narrow down the search for the
Also, attempt to adjust the pH to match molecular formula of the unknown peak. DISCUSSION
to the original method pH to minimize Generally, at this point, a holistic analysis
shifts in retention times. Lastly, avoid considering the UV profile, molecular The point to remember is that each
using sodium salts, if possible, in either weight of the unknown peak and how situation is unique and will require an
the buffer or pH adjustment process that compares to the main active individualized case-by-case approach to
as counter-ion adducts in MS analyses pharmaceutical ingredient or excipients, identifying the unknown peak. However,
make data interpretations difficult. Once fragmentation patterns, and any other the general strategy to identifying an
the mobile phase has been adjusted, available information or historical data unknown peak is summarized in the
the analysis needs to be re-run (using should be performed. flowchart (figure 1).
diode array detection) to confirm that the
chromatographic profile remains rela- In some cases, GC-MS may also prove It is also very advantageous to collect
tively unchanged from the original mobile to be a valuable tool but should also be as much available information before
phase composition and the unknown used with caution due to thermal degra- initiating the investigation. Information
peak remains well resolved from interfer- dation or lack of volatility. such as structure of the parent molecule,
ences. The diode array data also ensures excipient formulation, known degradation
that the unknown peak continues to be STAGE 4 – PEAK ISOLATION pathways of the parent, and historical
the target of the investigation. AND NMR ANALYSIS data (i.e. forced degradation study data)
can all aid in the strategy development.
In the event that MS data are incon- Also, it may be beneficial to assess any
clusive, further investigation will be subtle changes that may have occurred
required to identify the unknown peak. including:
At this point, it is suggested that NMR
be employed to determine the structure

3. LIFE SCIENCE I TECHNICAL BULLETIN 3
• Changes in raw material suppliers the real amount present is known, a plan
• Changes in manufacturing location forward can be developed to resolve
or process the issues to: accept or reject the batch,
• Any deviations or issues during the amend the specification to control the
manufacturing process impurity level, and potentially release the
• Changes in the container closure product to market.
system or supplier
CONCLUSION
Lastly, the discovery of an unknown peak
usually causes a great deal of stress. The The identification of an unknown peak
resolution of the unknown peak issue is a valuable capability to have in ones
will be time critical and there will be troubleshooting arsenal. It requires plan-
pressure to complete the investigation in ning as well as the assimilation of many
an expedited manner. It is important to pieces of data. The combination of DAD,
understand that this type of an investiga- low and high resolution MS, MS/MS, and
tion takes time and there are many tech- NMR data will contribute to the success-
nical hurdles that may arise during the ful structural elucidation. Each situation
course of the work. A reasonable time will offer unique challenges and consid-
estimate for the full study is anywhere erations, so the overall strategy can not
from a few days to ~4 weeks. be a “one size fits all” approach. Different
investigational tools may be needed and
However, performing a due diligence of alternate strategies may be mapped out,
all the available information prior to start- but in general, using this guidance as a
ing the work and following the strategy basis for initiating the anomalous peak
outlined in Figure 1, helps to ensure a investigation will afford the framework
successful outcome and efficient utiliza- needed for expeditious success.
tion of time and resources. Once the
identification of the peak is obtained and
FIGURE 1: FLOW CHART FOR DETERMINING THE STRATEGY OF
UNKNOWN PEAK INVESTIGATIONS
UNKNOWN PEAK
OBSERVED
YES ID/SOURCE OF PEAK
DETERMINED?
PERFORM OOS
INVESTIGATIONS NO
PEAK COLLECTION
YES
ID/SOURCE OF PEAK CONCLUDE
DETERMINED? INVESTIGATION
NO
YES DETERMINE
ENOUGH MATERIAL
EXTINCTION
DAD ANALYSIS TO WEIGH?
COEFFICIENT
NO
YES
YES
METHOD MS
NMR
COMPATIBLE?
NO
MAKE COMPATIBLE OTHER
AND PERFORM ID/SOURCE OF PEAK NO INVESTIGATION
METHOD DETERMINED? REQUIRED
DEVELOPMENT PERFORM LC-MS,
LC-MS/MS AND
HIGH RES MS

4. LIFE SCIENCE I TECHNICAL BULLETIN 4
To receive future articles on current trends and regulatory updates, subscribe to SGS’ Life Science News at www.sgs.com/lss_subscribe
CONTACT INFORMATION
EUROPE ASIA NORTH AMERICA
BELGIUM INDIA CANADA
+32 10 42 11 11 +91 44 2254 2601 + 1 905 364 3757
be.pharmaqc@sgs.com in.pharmaqc@sgs.com ca.pharmaqc@sgs.com
FRANCE (PARIS) SINGAPORE USA (FAIRFIELD, NJ)
+33 1 41 06 95 93 +65 677 53 034 + 1 888 747 8782
fr.pharmaqc@sgs.com sg.pharmaqc@sgs.com us.pharmaqc@sgs.com
FRANCE (POITIERS) CHINA USA (LINCOLNSHIRE, IL)
+33 (0) 5 49 57 04 04 +86 21 6115 2197 +1 847 821 8900
clinicalresearch@sgs.com cn.pharmaqc@sgs.com us.pharmaqc@sgs.com
GERMANY (BERLIN) TAIWAN USA (WEST CHESTER, PA)
+49 30 3460 7500 +886 2 2299 3279 ext 2500 + 1 610 696 8210
de.pharmaqc@sgs.com tw.pharmaqc@sgs.com us.pharmaqc@sgs.com
GERMANY (FREIBURG)
+49 761 6116 7760 WWW.SGS.COM/PHARMAQC
de.pharmaqc@sgs.com
GERMANY (TAUNUSSTEIN)
+49 6128 744 245
de.pharmaqc@sgs.com
SWITZERLAND (GENEVA)
+41 22 794 8374
pharmaqc@sgs.com
UK (WOKINGHAM)
+44 (0) 1189 896940
pharmaqc@sgs.com