Ranitidine - Investigations into the Root Cause for the Presence of N-Nitroso- N, N-dimethylamine in Ranitidine Hydrochloride Drug Substances and Associated Drug Products

Article abstract of DOI:10.1021/acs.oprd.0c00462

The presence of low levels of N-nitroso-N,N-dimethylamine (NDMA) in ranitidine hydrochloride drug products has been reported by regulatory agencies. GlaxoSmithKline undertook a root cause analysis to investigate this observation using contemporaneous, highly sensitive analytical methodologies. The root cause analysis suggested that the presence of NDMA results from a slow degradation of the ranitidine molecule. Analysis using suitably isotopically labeled ranitidine hydrochloride confirmed the formation of NDMA solely from an intermolecular reaction of ranitidine hydrochloride without involvement of impurities. Factors that influence the rate of degradation include heat, humidity, and the crystal morphology of ranitidine hydrochloride with the material exhibiting a columnar habit showing a slower rate of degradation.

Full text of DOI:10.1021/acs.oprd.0c00462

pubs.acs.org/OPRD
Article
RanitidineInvestigations into the Root Cause for the Presence of
N‑Nitroso‑N,N‑dimethylamine in Ranitidine Hydrochloride Drug
Substances and Associated Drug Products
Fiona J. King, Andrew D. Searle,* and Michael W. Urquhart
Cite This: https://dx.doi.org/10.1021/acs.oprd.0c00462
Read Online
Metrics & More
Article Recommendations
ACCESS
ABSTRACT: The presence of low levels of N-nitroso-N,N-dimethylamine (NDMA) in ranitidine hydrochloride drug products has
been reported by regulatory agencies. GlaxoSmithKline undertook a root cause analysis to investigate this observation using
contemporaneous, highly sensitive analytical methodologies. The root cause analysis suggested that the presence of NDMA results
from a slow degradation of the ranitidine molecule. Analysis using suitably isotopically labeled ranitidine hydrochloride confirmed
the formation of NDMA solely from an intermolecular reaction of ranitidine hydrochloride without involvement of impurities.
Factors that influence the rate of degradation include heat, humidity, and the crystal morphology of ranitidine hydrochloride with
the material exhibiting a columnar habit showing a slower rate of degradation.
KEYWORDS: ranitidine hydrochloride, N-nitroso-N,N-dimethylamine (NDMA), degradation
nitrosation source,6 and although 1 does contain a tertiary
amine, it does not contain an obvious nitrosation source given
that the nitro group is in a higher oxidation state; therefore, the
formation of NDMA from ranitidine was not reasonably
INTRODUCTION
■
Nitrosamines are potential mutagenic materials labeled as being
in the “cohort of concern” within the ICH M7 guideline
“Assessment and Control of DNA Reactive (Mutagenic)
Impurities in Pharmaceuticals to Limit Carcinogenic Risk”.1 In
predicted via a favorable mechanistic reaction pathway. We
have undertaken an in-depth scientific evaluation to determine
the root cause for the presence of NDMA in ranitidine and
associated drug products using contemporaneous, highly
sensitive, and specific analytical methodologies with the results
described below.
June 2018, health authorities became aware that levels of certain
nitrosamines were present within the valsartan product from one
supplier.2 Following this observation, there was a requirement
that all sartan drugs, angiotensin II receptor antagonists, were
assessed for levels of the reported nitrosamines.3 The non sartan
drug, pioglitazone, was later discovered to contain specific
nitrosamines. Since the initial marketing of ranitidine, advances
in analytical technology, especially in mass spectrometry, have
significantly increased the sensitivity of detection for these trace-
level impurities which may have precluded their discovery earlier
in the product life cycle.
ASSESSMENT OF THE SYNTHETIC ROUTE
■
It has been suggested that NDMA could be formed or
introduced during the synthetic process.7 However, looking at
the synthetic pathway to ranitidine hydrochloride 1 shown in
Scheme 1, the only reasonably predicted source of nitrite was
sodium nitrite, which may be used during the manufacture of the
reagent nitromethane.8 Nitromethane 5 was one of the starting
materials for the nitro-olefin 6. This source of nitrite is four
synthetic stages from the introduction of the dimethylamine
function, and these stages included isolations and purifications.
Additionally, the nitromethane reagent was purified through
distillation. Therefore, there is a negligible likelihood for nitrite
to be present within this reagent.
In September 2019, low levels of N-nitroso-N,N-dimethyl-
amine (NDMA) were found within ranitidine products from a
number of manufacturers.4 GlaxoSmithKline (GSK) have
conducted an investigation into the source and formation of
NDMA, and the initial findings are reported within.
Ranitidine 1 was patented by Allen and Hanburys in 19775
and initially marketed in 1981 under the trade name Zantac for
the treatment of peptic ulcer disease.
Received: October 14, 2020
In order for a nitrosamine to form, a suitably reactive
secondary or tertiary amine must be present together with a
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 1. Manufacturing Process to Ranitidine Hydrochloride 1
Figure 1. Ph. Eur. and USP impurities for ranitidine hydrochloride.
Dimethylamine could react with an appropriate nitrosating
source to form NDMA and is used in the stage 1 process, but
there is no source of nitrite considered likely to be available.
Dimethylamine is not reasonably predicted to be present as an
impurity within the ranitidine drug substance given that it is used
five stages prior to end product isolation within the synthetic
process. In addition, there are numerous opportunities for any
residual levels to be purged9 through both solubility (i.e.,
purging in mother liquors) and volatility particularly as
dimethylamine has a low boiling point (b.p., 7 °C) and
intermediate 4 is purified by high-vacuum distillation. In order
to confirm this assessment, we screened both intermediate 4 and
an aqueous solution of ranitidine free base 7 taken after the
coupling reaction with 6 for NDMA content using liquid
chromatography−high-resolution mass spectrometry (LC−
HRMS), and none was detected (<0.25 mcg/g) in either
material. The use of LC rather than gas chromatography (GC)
follows the advice from the Food and Drug Administration
(FDA), who have published that for ranitidine, GC-based
methods have been observed to elevate NDMA levels in tested
materials through thermal degradation of the sample and that
LC is the preferred approach for accurate analysis.10
The risk that secondary amines or nitrosating agents or
nitrosamines directly were being introduced unintentionally
through recycled solvents was also considered for suppliers of
ranitidine to GSK. Although several of the process solvents are
recycled, recycling is discrete to the stage in which the solvent is
used, and the solvent is never mixed with solvents derived from
B
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 2. Observed degradation pathways for ranitidine hydrochloride.
other processes. There is therefore no opportunity for cross
contamination. It is also important to note that no dipolar
aprotic amide solvents, such as N,N-dimethylformamide or N-
methylpyrrolidinone, are used within the process which could
introduce nitrosamines through cross contamination.11
All drug product formulations were assessed for the potential
for formation of NDMA, and it was concluded that during
manufacture, there was no identified risk as none of the
excipients used contain any nitrosation agents. All excipients are
tested as per current United States Pharmacopeia/National
Formulary (USP/NF) and European Pharmacopeia (Ph. Eur.)
standards, and based on a review of their chemical structures, no
amine groups are anticipated to be present, although the
presence of trace nitrites cannot be ruled out.12
The known and potential impurities from ranitidine hydro-
chloride were also assessed. There are 11 impurities listed within
the Ph. Eur. and USP, and these are shown in Figure 1. While the
loss of the dimethylamine function from ranitidine 1 could
theoretically be a concern from interaction with adventitious
nitrosating agents, it is apparent that this group has not been lost
in any of the known and potential impurities.
hydrochloride 1 cannot be considered to be reasonably
predicted, and is not a major degradation pathway, an
investigation into the formation of low levels of NDMA within
ranitidine hydrochloride 1 was instigated. The results from this
investigation are discussed below.
NDMA CONTENT OF RANITIDINE DRUG
SUBSTANCES AND DRUG PRODUCTS
■
The NDMA content of drug substance batches manufactured by
three different suppliers, including GSK, was determined (using
methods 1 and 2). These batches were manufactured between
2010 and 2019, and the NDMA content is shown in Figure 3.
Some of these results exceeded the upper validated range of
the analytical method (>9.9 mcg/g), but extrapolated
Additionally, the hydrolysis13 and solid-state stability14 of
ranitidine hydrochloride have been published, and a summary of
the degradation pathways for ranitidine is shown in Figure 2.
Ranitidine would appear stable under strongly acidic conditions,
leading to the protonated drug 18. From exposures to pH 2 and
above, ranitidine 1 can be degraded under certain conditions,
which results in a range of degradation products. What is clear
from the observed hydrolytic degradation of ranitidine is that
there are no identified products whereby the dimethylamine
group has been liberated, which again highlights that the
formation of the nitrosamine NDMA is not likely to be a major
pathway.
Having reviewed existing information and reaching the
conclusion that the formation of NDMA from ranitidine
Figure 3. NDMA content of drug substance batches (mcg/g) vs batch
date of manufacture (tested between October and December 2019).
C
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
concentrations are shown in order to demonstrate the range of
values observed.
Details of samples studied, which were manufactured either by
supplier 2 using one of the two different processes or by GSK,
are shown in Table 2. The storage of two samples was extended
The NDMA contents of batches of drug products from four
different formulations were also tested during the same period
(using methods 1 and 2). The level of NDMA of the input drug
substance was not tested prior to the secondary formulation.
Table 1 shows the range of NDMA contents from all batches of
drug substances and drug products tested.
Table 2. Details of Ranitidine Hydrochloride Drug Substance
Samples Studied
date of
manufacture
(DOM)
process
(supplier 2 only)
date of age at testing
source
testing
(months)
Table 1. Content Range Observed in Ranitidine-Containing
Drug Substances and Product Batches Tested
supplier 2
process 1
process 1
process 2
Aug 2019
Nov 2019
Dec 2019
Dec 2014
Nov
3
2019
supplier 2
supplier 2
GSK
Dec
2019
Dec
2019
Dec
2019
1
no. of batches
tested
NDMA content range observed
a
material
(mcg/g)
<1
60
b
drug substance
film-coated
tablets
effervescent
tablets
IV solution
syrup
134
119
0.11−>9.9
0.24−7.6
0.52−2.1
0.24−1.1
32
up to 115 days to determine the impact of prolonged storage.
The samples were tested to determine NDMA content under
accelerated stability conditions, and the results are listed in
Table 3.
The samples tested broadly fell into two distinct groups; the
two batches from supplier 2, process 1 generated about 7 mcg/g
NDMA after 7 days, with the two other samples remaining at <1
mcg/g at this same time point. This general trend continued for
the two samples stored for 115 days. There was no evidence that
the rate of NDMA formation significantly slowed, which would
be observed if the formation was due to reaction with a trace
impurity which was consumed by the reaction and therefore
depleted.
Differences between these samples were considered, which
could explain the differences in stability. The initial focus was on
the supplier 2 batches where the manufacturing process differed
and significant differences in the rate of NDMA formation were
observed for process 1 relative to process 2. The manufacturing
processes for these batches were identical other than the final
salt formation/crystallization step (Table 4).
Part 2 Determination of Physical Properties. In order to
interrogate morphological differences resulting from different
processing conditions, samples were analyzed using a variety of
solid-state techniques. Scanning electron microscopy (SEM)
images are shown in Figure 4.
Significant differences in morphology were apparent in SEM
images. The supplier 2, process 1 batch had an irregular
morphology containing small spherical particulates, which was
consistent with a process that proceeded via an oil phase,
whereas the supplier 2, process 2 batch exhibited a columnar
crystal habit. X-ray powder diffraction (XRPD) analysis
confirmed that both samples were of the same polymorphic
form (form 2). Having observed these significant differences in
physical properties, further tests were conducted (Table 5).
The irregular supplier 2, process 1 material was shown to
adsorb water and had a lower onset of decomposition
temperature, both suggesting minor differences in physical
properties. The surface area differences were likely related to
differences in particle size observed in the SEM images. There
was no difference in the true density of the batches, which
indicated that the crystal packing and form were the same for
both, consistent with XRPD and spectroscopy data.
55
24
<0.68−2.7
a
b
All results are reported relative to the ranitidine free base. The
result is above the highest calibration standard.
The following trends were observed from analysis of these
data
• There is an indication of a trend that the NDMA content
does increase with the age of the batch which would
indicate that NDMA could be forming via a slow
degradation reaction. Lowest levels are observed in
materials made in 2019 and the highest levels in materials
made in 2014. This indicated that NDMA could be
formed by an unknown reaction path that resulted in a
slow degradation process over time. There does however
remain a degree of variability in the data set. One reason
for this could be that these samples were sourced from
retained samples stored at many different production sites
and therefore small differences in storage conditions
could create an additional variable.
• Batches supplied by GSK in 2014 had lower levels of
NDMA than would have been anticipated when
compared to other batches of a similar age. The rate of
formation of NDMA in these samples was calculated to be
approximately 0.04 mcg/g per year, indicating that the
reaction forming NDMA must be highly disfavored in the
material and vary depending on the supplier or the specific
process used.
• The range of NDMA contents in all four drug product
formulations was smaller than that for drug substances.
• When both drug product and the input drug substance
batches have been tested within a period of a few weeks, in
almost all cases, the NDMA content of the drug substance
is higher than that of the batch of drug products made
from the drug substance batch. This indicates that the rate
of NDMA formation decreases or ceases once the product
is formulated.
On the basis of these observations, the drug substance became
the focus for the root cause investigation into the presence of
NDMA in ranitidine.
Part 1 Short-Term Stability Studies. In order to study the
relative stability of drug substance batches under accelerated
conditions, a small number of batches were stored in closed vials
and then held at 60 °C for 7 days (uncontrolled humidity).
A solid-state NMR study was conducted to study the location
of solvent molecules in the crystal lattice. This was done by
recrystallizing batches using the solvent ratios used in the
supplier 2, process 1 and supplier 2, process 2 methods using
D
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
a
Table 3. NDMA Content (mcg/g) of Ranitidine Hydrochloride Stored at 60 °C
batch detail
day
supplier 2, process 1 DOM Aug
2019 NDMA (mcg/g)
supplier 2, process 1 DOM Nov
2019 NDMA (mcg/g)
supplier 2, process 2 DOM Dec
2019 NDMA (mcg/g)
GSK DOM Dec 2014
NDMA (mcg/g)
0
3
7
46
115
0.33
6.3
7.7
not tested
not tested
1.05
ND
3.6
7.4
38
66
1.06
ND
<LOQ (0.17)
0.2
0.36
0.39
not tested
not tested
0.03
0.28
2
3.3
0.04
rate of formation over 7 days
(mcg/g per day)
a
Tested using method 3 (115 day samples tested using method 4).
Table 4. Comparison of Ranitidine Hydrochloride Drug
Substance Isolation Processes
solvent peak. The actual amounts of residual isopropanol (IPA)
and methanol were higher for the process 1 sample, which was
consistent with residual solvent analysis performed on samples
of the process 1 material (0.28% w/w IPA, 0.2% w/w methanol)
and the process 2 material (<0.01% w/w IPA and 0.02% w/w
methanol).
Available data for the GSK drug substance were largely
consistent with the supplier 2, process 2 material. An SEM image
for the GSK-produced material is shown in Figure 5. As the GSK
manufacturing process included a final granulation step, the
columnar habit, while distinguishable by SEM, contained
agglomerated particulates as well.
a
supplier 2,
process 1
supplier 2,
process 2
parameter
GSK
IPA
a
solvent composition
IPA/MeOH
(approx. 70:30)
IPA/MeOH
(approx. 40:60)
a b
,
total solvent volume
water concentration
8.7
3.6%
3.5
2.5
ac
,
not added to the
process
3.4%
crystallization
temperature
50 °C 10−25 °C
10−25 °C
observed rate of NDMA slow
fast
slow
d
formation
As a part of this study, we operated the GSK process in the
laboratory and similarly confirmed that it did not proceed via an
oil phase.
a
Values may vary slightly from batch to batch because of the use of
some solvents/water in the pH adjustment during the crystallization
step. Volume of the solvent as a ratio to the input weight of the
b
c
Part 3 Accelerated Stability Study. The impact of
temperature and humidity on the stability of both the supplier
2, process 1 and supplier 2, process 2 drug substance batches was
studied (Table 6). Batches were stored for up to 21 days
between 50 and 80 °C and 10 and 80% relative humidity (RH).
For the supplier 2, process 2 batch, the physical appearance of
the sample and, in some cases, the high-performance liquid
chromatography (HPLC) purity of the sample were also
recorded. As GSK ceased manufacturing drug substances in
2014, there was an insufficient sample available to assess GSK-
manufactured ranitidine hydrochloride in the same manner.
The key observations made from accelerated stability studies
reported in Table 6 were as follows: both an elevation in
temperature and RH increased the rate of formation of NDMA.
ranitidine free base. % v/v water in the overall crystallization mixture.
d
While the process details in Table 4 also suggest that the presence of
water favors the formation of the more stable morphology, additional
information not reported herein from supplier 1 where water was
included in the final crystallization demonstrated that in this case, the
less stable morphology was produced. Other variables such as total
solvent quantity, composition, and crystallization temperature may
also contribute to the observed morphology.
deuterated solvents (d8 isopropanol, d4 methanol, and
deuterium oxide) and analyzing isolated solids by solid-state
NMR. It was possible to determine that in both isolated solids,
the solvents were adsorbed on the surface and not included in
the crystal lattice by assessment of the spinning side bands of the
Figure 4. SEM images for supplier 2, process 1 (bottom row) and process 2 (top row) batches.
E
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 5. Physical Property Testing of Ranitidine Hydrochloride
result
test
purpose
process 1
process 2
dynamic vapor sorption
thermogravimetric analysis
differential scanning calorimetry
XRPD
specific surface area
helium pycnometry
study uptake and retention of water
change in mass vs temperature
change in heat flow vs temperature
polymorphic form
0.063% w/w moisture uptake at 70% RH
decomposition initiation <150 °C
0.043% w/w moisture uptake at 70% RH
decomposition initiation >150 °C
no significant differences
form 2
form 2
0.24 m2/g
1.32 g/cm3
0.40 m2/g
1.32 g/cm3
measure true density
Raman and FTIR spectroscopy
no significant differences
Part 4 Hypothesis to Link the Process, Physical
Properties, and Stability. The control of morphology
unexpectedly appeared to be a key factor for the control of the
rate of NDMA formation in the ranitidine drug substance and
particularly the avoidance of what appears to be a progression of
the crystallization through an oil phase. This can be achieved by
controlling the concentration and temperature of the crystal-
lization process, by increasing the methanol concentration of the
solvent mixture, or by using IPA alone under more dilute
conditions (the GSK process). Control of these variables
appears to favor the desired columnar crystal habit which
subsequently forms NDMA at a slower rate on stability testing,
although it is not possible to prevent formation completely.
We hypothesize that this may be due to differences in surface
properties between columnar habit (process 2) and process 1
particles as observed with the differences in residual solvent
adsorption. The process 1 material also retains and adsorbs more
water when compared to the columnar material, resulting in a
faster degradation rate and as reported here; the presence of
water is a factor that increases the rate of NDMA formation.
While the crystal habit and morphology work and the
corresponding stability work elucidated the reasons for
differences in NDMA formation rates across different material
pedigrees, the work did not uncover the actual reaction path
through which the reaction progressed, which are further
interrogated in subsequent sections.
Figure 5. SEM image for GSK ranitidine hydrochloride captured at a
×1500 magnification.
A similar finding has recently been reported for ranitidine drug
substances and tablets.15 For samples where the material
remained as a solid, the rate of NDMA formation was lower
for the supplier 2, process 2 batch (relative to supplier 2, process
1), which was consistent with the short-term stability study. No
significant other chemical degradation was observed for the solid
samples.
At high RH, deliquescence was observed for all samples. This
resulted in a more significant initial increase in formation of
NDMA (up to 150 mcg/g) with no distinction between supplier
2, process 1 and supplier 2, process 2 samples, but the level of
NDMA then reduced from the peak. The deliquescence was
accompanied with a significant level of chemical degradation of
ranitidine hydrochloride (up to a 69% area by HPLC), so
secondary degradation of NDMA also likely occurred.
Part 5 NDMA Being Formed by Reaction of an
Impurity with Ranitidine Hydrochloride. The possibility
that NDMA could be formed by reaction between an impurity
present in the drug substance and ranitidine hydrochloride was
considered during the investigation.
The oxime impurity (Ph. Eur. impurity G, structure 14 in
Figure 2) was considered as a theoretical nitrosating agent
because of isomerization via a nitroso intermediate.
This study further supported the importance of the process
and resulting morphological differences in the rate of NDMA
formation. Crystals that exhibited the columnar morphology
were more stable at low RH, but the interaction with water at
higher RH resulted in increased rates of NDMA formation as the
change from a crystalline solid to oil occurred. Once
deliquescence had occurred, and the material was liquid, there
was no further increase in NDMA content, although it is not
known whether this was due to the degradation pathway no
longer taking place or due to a secondary degradation reaction.
The increased formation of NDMA under the most stressed
conditions further supported alignment with the FDA
recommendation for using LC−MS over GC−MS for the
accurate determination of NDMA content of samples.
However, when ranitidine hydrochloride containing 2.7%
area impurity G was subjected to the short-term stability test (3
days, 60 °C, 10% RH), it did not lead to any increased formation
of NDMA.
Although nitrite was not considered a likely impurity in
ranitidine hydrochloride, we demonstrated that the addition of
sodium nitrite or nitrous acid led to an instantaneous reaction,
resulting in the formation of typical acid-catalyzed degradation
products and trace levels of NDMA. However, this reaction
profile was not consistent with the slow reaction to generate
F
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
a
Table 6. Accelerated Stability Monitoring of Ranitidine Hydrochloride
temperature (°C)
RH (%)
10
batch ID
process 2
test
initial
day 3
day 10
day 21
c
50
NDMA content (mcg/g)
total impurities by HPLC (% area)
description
NDMA content (mcg/g)
total impurities by HPLC (% area)
description
NDMA content (mcg/g)
NDMA content (mcg/g)
total impurities by HPLC (% area)
description
NDMA content (mcg/g)
NDMA content (mcg/g)
total Impurities by HPLC (% area)
description
NDMA content (mcg/g)
NDMA content (mcg/g)
total Impurities by HPLC (% area)
description
NDMA content (mcg/g)
NDMA content (mcg/g)
total impurities by HPLC (% area)
description
NDMA content (mcg/g)
NDMA content (mcg/g)
total impurities by HPLC (% area)
description
NDMA content (mcg/g)
NDMA Content (mcg/g)
total impurities by HPLC (% area)
description
ND
0.17
NT
ND
0.17
NT
0.33
ND
0.17
NT
0.33
ND
0.17
NT
0.33
ND
0.17
NT
0.33
ND
0.17
NT
0.33
ND
0.17
NT
0.33
ND
0.17
NT
0.06
0.13
0.23
NT
white solid
8.61
white solid
b
50
60
75
30
process 2
22.0
NT
24.0
orange oil
brown oil
b
b
process 1
process 2
29.0
47.1
0.19
0.80
0.46
NT
white solid
pale-yellow solid
b
b
60
70
70
80
80
65
10
80
25
50
process 1
process 2
68.4
50.9
12.5
b
b
31.8
NT
47.6
dark-brown oil
dark-brown oil
b
b
process 1
process 2
26.5
43.3
0.65
1.6
0.35
NT
off-white solid
yellow sticky solid
3.5
b
process 1
process 2
55.7
22.2
b
b
63.8
NT
69.0
dark-brown oil
dark-brown oil
b
b
process 1
process 2
59.8
1.8
NT
120.2
94.0
b
30.9
yellow solid
dark-brown oil
b
b
process 1
process 2
128.1
12.1
148.91
29.9
20.11
68.2
dark-brown oil
dark-brown oil
a
b
c
NDnot detected; NTnot tested. Above the highest calibration standard (9.9 mcg/g). Below the limit of quantitation (0.12 mcg/g). Tested
using method 1.
NDMA over a prolonged period. The reaction of nitrous acid
with ranitidine has been studied previously, and the only
product observed was the nitrolic acid derived from nitrosation
at N,N′-dialkyl-2-nitroethene-1,1-diamine.16
Scheme 2. Proposed Formation of +7 NDMA from
Isotopically Labeled Ranitidine Hydrochloride
Neither the oxime impurity nor sodium nitrite presence as
impurities was therefore considered to be implicated as the cause
of the degradation to generate NDMA.
The formation of NDMA from ranitidine during wastewater
treatment with chloramine is reported.17 Although this was
considered during the root cause analysis, should chloramine be
an impurity in water used during manufacture, the nondetected
levels of NDMA in freshly manufactured ranitidine hydro-
chloride suggested that this was not the root cause.
Part 6 Isotopically Labeled Study. To further examine the
degradation pathway, M + 7 ranitidine hydrochloride was
prepared incorporating stable isotope labels in both the
dimethylamino and nitro functional groups. The synthesis is
shown in Scheme 3 using readily available hexadeuterodimethyl-
amine and 15N nitromethane as the starting materials. The
NDMA generated would be the +7 isotopologue (Scheme 2).
A sample of the M + 7 material (purity by HPLC of 99.81%
area) was recrystallized with an equal amount of unlabeled
ranitidine hydrochloride (purity by HPLC of 99.95% area),
producing a blended batch with a combined purity by HPLC of
99.95% area in order to study whether the mechanism of NDMA
formation proceeds via an intermolecular or intramolecular
process. Formation of +1 and +6 NDMA isotopologues would
confirm an intermolecular process, whereas formation of solely
NDMA and the +7 NDMA isotopologues would indicate an
intramolecular process.
Both the M + 7 material and the blended batch exhibited the
same columnar morphology shown previously in Figure 4.
Samples of M + 7 ranitidine hydrochloride and the 1:1
labeled/unlabeled blend were heated in an oven at 60 °C. The
results of the NDMA analysis are shown in Table 7. Samples
were tested using method 5. Quantitative determination of the
NDMA isotopologues was not possible because standards of the
G
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 3. Preparation of +7 Ranitidine Hydrochloride
Heating M + 7 ranitidine hydrochloride afforded only the +7
isotopologue of NDMA. The absence of other isotopologues
implies that there were no external sources of either the
dimethylamino or nitrosating agent, thus providing conclusive
evidence that NDMA formation occurred via a degradation
mechanism involving the d6 dimethylamino and 15N nitro
groups.
The formation of four isotopologues of NDMA upon heating
the 1:1 blend of the labeled and unlabeled materials for 25 days
provided conclusive evidence that an intermolecular reaction to
form NDMA had occurred with the dimethylamino and nitroso
groups originating from different ranitidine molecules (Scheme
4). If the reaction had proceeded via an intramolecular pathway,
only parent NDMA and the +7 isotopologue would be formed.
It was observed that for +7 NDMA, the response for the
unblended sample after 25 days was approximately 4 times that
of the same peak in the blended sample. This is an indication that
the four different isotopologues were formed in equal amounts,
which is what would be expected for the intermolecular reaction.
The exact chemical mechanism by which NDMA is formed has
not been elucidated, and there is no obvious single pathway that
Table 7. NDMA Content (mcg/g) Following Stress Testing
of Isotopically Labeled Ranitidine Hydrochloride and
Labeled Ranitidine Hydrochloride Blended with the
a
Unlabeled Material
days stored at
+7 NDMA
content
+6 NDMA
content
+1 NDMA
content
NDMA
content
60 °C
labeled ranitidine hydrochloride unblended
0
7
25
ND
strong signal
strong signal
ND
ND
ND
ND
ND
ND
ND
ND
ND
labeled ranitidine hydrochloride blended with unlabeled ranitidine
hydrochloride
0
ND
ND
ND
ND
7
25
weak signal
present
ND
present
weak signal
present
ND
present
a
Each isotopologue had a different response according to the method
used.
different isotopologues of NDMA were not available. The results
are expressed by the presence of an isotopologue and the relative
strength of the signal observed.
Scheme 4. Products from Intermolecular and Intramolecular Pathways to Form NDMA
H
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
can explain its presence. Further work is required to confirm this
pathway.
Mass spectrometry: Waters QDA, alternate-scan positive and
negative electrospray ionization, scan range = 100−1000 AMU,
scan frequency = 5 Hz.
High-Pressure Liquid Chromatography (HPLC) LC
Method C. Agilent 1200 Waters Xterra MS C18 (100 × 4.6
mm, 3.5 mc m) at 35 °C, a flow rate of 1.5 mL/min, an injection
volume of 10 mcL, a UV detection of 230 nm, A: 0.05 M
potassium phosphate monobasic pH 7.1/acetonitrile 98:2 (v/
v), B = 0.05 M potassium phosphate monobasic pH 7.1/
acetonitrile 78:22 (v/v), gradient 0−100% B over 10 min, held
for 5 min, 0−100% A over 1 min, held for 4 min.
CONCLUSIONS
■
We have demonstrated that NDMA (N-nitrosodimethylamine)
is formed in ranitidine drug substances because of an
intermolecular degradation reaction of ranitidine molecules
that occurs primarily in the solid state. The exact reaction
mechanism is yet to be determined and requires further work to
confirm. The degradation proceeds without participation of any
extraneous impurities. NDMA is not detected in freshly
prepared batches of drug substances, thus demonstrating that
it is not formed as a synthetic impurity in the process or
introduced as a contaminant. NDMA levels in drug substances
start to increase at a slow rate from the point of manufacture;
both elevated temperature and RH contribute to an increase in
the rate of degradation. The rate of NDMA formation in the
drug substance stored at ambient temperature and low humidity
indicates that the reaction pathway is disfavored and would not
be detected without the use of highly sensitive and sophisticated
analytical methods.
The rate of NDMA formation in the drug substance and the
crystal morphology are linked, with the columnar crystals
exhibiting the slowest rate of formation. This morphology is
produced when the drug substance crystallization does not
proceed via an oil phase. The reason for the difference in NDMA
formation rates between the batches exhibiting different
morphologies is not certain, but it could be related to differences
in surface properties which promote interactions with water,
resulting in faster degradation.
[5-(Bis(2H3methyl)aminomethyl)furan-2-yl]methanol 22.
Furan-2-ylmethanol (4.7 mL, 5.3 g) and 40% aqueous
formaldehyde (6.5 mL, 2.6 g) were added to a stirred solution
of (2H3methyl)2 amine hydrochloride (5.0 g) in water (10 mL)
at room temperature. The solution was warmed to 50 °C and
stirred under nitrogen overnight. The cooled mixture was
diluted with 2 N hydrochloric acid (30 mL) and ethyl acetate
(30 mL), the layers were separated, and the aqueous layer was
washed with ethyl acetate (2 × 30 mL).
The aqueous layer was cooled in an ice/water bath, with
stirring, and treated with 50% aqueous sodium hydroxide
solution to pH 14. The resulting solution was extracted with
ethyl acetate (3 × 40 mL), additional 50% aqueous sodium
hydroxide solution was added to the aqueous layer then
extracted with ethyl acetate (40 mL). The combined organic
phase was washed with brine (50 mL) and evaporated to dryness
under reduced pressure to afford a red-brown oil (6.98 g, 80%).
1H NMR (400 MHz, chloroform-d): δ 3.43 (s, 2H), 4.57 (s,
2H), 6.14 (d, J = 2.93 Hz, 1H), 6.20 (d, J = 3.42 Hz, 1H); LCMS
(LC method A): 97% PAR, MS (m/z, positive ion): 162.1 [M +
H]+.
2-[(5-{[Bis(2H3methylamino)methyl]furan-2-yl}methyl)-
thio]ethan-1-amine 23. Concentrated hydrochloric acid (12
mL) was added to cysteamine hydrochloride (4.82 g) and 22
(6.85 g), and the mixture was stirred at 50−55 °C under
nitrogen for 2.5 h and then cooled to room temperature and
diluted with water (30 mL). The pH was adjusted to 14 by the
addition of 50% aqueous sodium hydroxide solution, with
cooling, and the solution was extracted with ethyl acetate (3 ×
50 mL). The combined organic phase was washed with brine (3
× 40 mL) and evaporated to dryness under reduced pressure.
The product was purified using silica gel column chromatog-
raphy eluting with dichloromethane, methanol, and triethyl-
amine [97.5:2.5:1 v/v (1600 mL); 95:5:1 v/v (800 mL);
92.5:7.5:1 v/v (800 mL)] to afford 23 as a clear, golden oil (5.2
g, 56%). 1H NMR (400 MHz, chloroform-d): δ 2.61 (t, J = 6.36
Hz, 2H), 2.83 (t, J = 6.40 Hz, 2H), 3.43 (s, 2H), 3.70 (s, 2H),
6.09−6.14 (m, 2H); LCMS (LC method A): 100% PAR, MS
(m/z, positive ion) 221.3 [M + H]+.
EXPERIMENTAL SECTION
■
General. All starting materials, solvents, and reagents were
obtained from commercial sources and were used as they were
received. (2H3methyl)2amine hydrochloride was purchased
from Aldrich, and 15N nitromethane was purchased from
Cambridge Isotope Laboratories. All reactions were carried out
in an Easymax reactor or oven-dried reaction vessels.
1H NMR spectra were recorded on 400, 600, and 700 MHz
spectrometers at 300 K and are reported as chemical shifts (δ) in
parts per million (ppm). Tetramethylsilane and water in D2O
were used as a reference.
Melting points were determined using a Stanford Research
Systems OptiMelt and are uncorrected.
HRMS (m/z) was performed using a Fusion Tribrid mass
spectrometer (Thermo Fisher) equipped with a heated
electrospray ionization ion source.
Ultraperformance Liquid Chromatography−Mass
Spectrometry (LC−MS). Ultraperformance liquid chromatog-
raphy (UPLC) was conducted on a Waters Acquity CSH C18
column (a 50 mm × 2.1 mm i.d., a 1.7 μm packing diameter) at
40 °C, a flow rate of 1 mL/min, an injection volume of 0.3 μL,
and UV detection = a summed signal from wavelengths of 210−
350 nm.
LC-method A elution: A: 10 mM ammonium bicarbonate in
water adjusted to pH 10 with ammonia solution, B = acetonitrile,
gradient 0−97% B over 1.45 min, held for 0.4 min.
LC-method B elution: A: 0.1% v/v solution of formic acid in
water, B = 0.1% v/v solution of formic acid in acetonitrile,
gradient 3−97% B over 1.5 min, held for 0.4 min.
(Z)-N-Methyl-1-(methylthio)-2-[15N]nitroethen-1-amine
24. [15N]nitromethane (0.95 g) was added to a solution of
methyl isothiocyanate (1.04 g) in dimethyl sulfoxide (DMSO)
(6.5 mL) at 5−6 °C under a nitrogen atmosphere, followed by a
line wash with DMSO (0.78 mL). A solution of potassium
hydroxide (0.902 g) in water (0.56 mL) was added to the
mixture in three portions, maintaining the temperature at 1−11
°C. The mixture was warmed to 25 °C over 15 min, stirred at 25
°C for 35 min, and then cooled to −2 °C.
Dimethyl sulfate (1.5 mL) was added to the mixture,
maintaining the temperature at 1−3 °C, over 30 min and then
warmed to 12 °C and stirred for 1 h 40 min. Water (9.2 mL) was
added dropwise over 30 min, maintaining the temperature at
I
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
12−16 °C, and then, the mixture was cooled and stirred at 1−2
°C overnight. The suspension was filtered, and the cake was
washed with chilled water (2 mL) and dried under vacuum at
ambient temperature with a nitrogen blanket to furnish 24 as a
pale-yellow solid (1.09 g, 51.6%). 1H NMR (400 MHz,
chloroform-d): δ 2.45 (s, 3H), 3.13 (d, J = 5.17 Hz, 3H), 6.60
(d, J = 1.48 Hz, 1H), 10.48 (br s, 1H); LCMS (LC method B):
99% PAR, (m/z, positive ion) 150.0 [M + H]+
+ H]+ calcd for C13H172H6O3N315NS, 322.1832; found,
322.1830. Mass spectrometry: average isotope incorporation =
6.99, no evidence of M0 above a 0.1% intensity. HPLC (LC
method C): retention time 6.74 min, 99.81 area %.
(E/Z)-N-(2-{[(5-{[Bis(2H3 methyl)amino]methyl}furan-2-yl)-
methyl]thio}ethyl)-N′-methyl-2-[15N]nitroethene-1,1-dia-
mine Hydrochloride 21, Blended with (E/Z)-N-[2-({5-
[(Dimethylamino)methyl]furan-2-yl}methyl)thio}ethyl]-N′-
methyl-2-nitroethene-1,1-diamine Hydrochloride 1. 2-Prop-
anol (16 mL) was added over 45 min to a stirred solution of M
+7 labeled ranitidine hydrochloride 21 (1 g) and unlabeled
ranitidine hydrochloride 1 (1 g) in methanol (8 mL) at 43−46
°C under a nitrogen atmosphere. The mixture was stirred at 47−
48 °C for 30 min and then cooled to 3−5 °C over 2 h and stirred
at 3−5 °C for 2 h.
(E/Z)-N-(2-{[(5-{[Bis(2H3methyl)amino]methyl}furan-2-yl)-
methyl]thio}ethyl)-N′-methyl-2-[15N]nitroethene-1,1-dia-
mine 25. A solution of 23 (4.65 g) in water (10 mL) was added
dropwise over 3 min to a stirred suspension of 24 (3.1 g) in water
(10 mL) at room temperature under a flow of nitrogen, with the
nitrogen exhaust passed through a bleach scrubber before
venting to the fume cupboard. When the addition was complete,
the dropping funnel was washed through into the reaction vessel
with water (5 mL). The mixture was stirred and warmed to 35−
40 °C, stirred for 2 h, and then allowed to cool to room
temperature overnight.
The slurry was filtered; the cake was washed with cold 2-
propanol (2 × 15 mL), air-dried, and then dried under vacuum
at 30 °C for 17 h to afford the title product as a white solid (1.77
1
g, 88.5%). mp 135.4−135.9 °C (dec.). H NMR (700 MHz,
DMSO-d6): δ 2.68 (s), 2.70 (br s), 2.89 (br s), 3.36 (br s), 3.49
(br s), 3.87 (s), 4.32 (br s), 6.38−6.47 (m), 6.54 (br s), 6.66 (d, J
= 2.30 Hz), 7.48 (br s), 7.61 (br s), 9.92 (br s), 10.04 (br s),
11.03 (br s). MS: average isotope incorporation = 3.47, M0 =
49.7%. HPLC (LC method C): retention time 6.17 min, 49.12
area %; retention time 6.54 min 50.83 area %.
The mixture was extracted with chloroform (2 × 10 mL), and
the combined organic layers were washed with water (20 mL).
The organic phase was diluted with methyl isobutyl ketone
(MIBK, 10 mL) and concentrated under reduced pressure to
approximately 15 mL. MIBK (15 mL) was added and
concentrated under reduced pressure to approximately 15 mL
three times. The resulting solution was diluted with MIBK (15
mL) and stirred, with cooling in an ice/water bath. The cold,
stirred solution was seeded with a few grains of the unlabeled
ranitidine free base. MIBK (5 mL) was added to the resulting
slurry, and the mixture was stirred for an hour and filtered. The
solid was washed with ice-cold MIBK (3 × 7 mL) and dried
under vacuum at room temperature to afford 25 as an off-white
solid (4.6 g, 68%). 1H NMR (400 MHz, D2O): δ 2.84 (br s, 2H),
2.93 (br s, 3H), 3.43 (t, J = 6.36 Hz, 2H), 3.52 (s, 2H), 3.84 (s,
2H), 6.31 (s, 2H), 6.85 (br s, 1H). LCMS (LC method A): (m/
z, positive ion): 322.3 [M + H]+. HPLC (LC method C):
retention time 6.75 min, 99.74 area %.
APPENDIX
■
X−Analytical NDMA Test Method Summaries
The NDMA content of batches of ranitidine drug substances has
been measured by the following validated analytical methods,
developed and run on “state-of-the-art” analytical equipment,
ensuring that appropriate sensitivity and specificity were
achieved to support these investigations. All method conditions
have been based on or shown to be equivalent to the FDA
recommended conditions.18 The use of LC instead of GC was
chosen specifically to avoid the degradation of ranitidine and
subsequent formation of NDMA that has previously been
observed during the use of headspace-GC-based methods.
Method 1. UPLC−HRMS analysis was performed using a
Waters Acquity UPLC system (Milford, MA) coupled to a Sciex
API4000 triple quadruple mass spectrometer (Warrington, UK).
Instrument parameters, data acquisition, and data processing
were controlled by Sciex Analyst 1.4 software (Warrington, UK).
The analysis was run on a Phenomenex Synergi Hydro RP 150
× 2 mm, 4 μm column (Torrance, CA) heated to 40 °C and
eluted using deionized water and methanol with a 0.1% (v/v)
formic acid gradient of increasing methanol from 2 to 8% over
8.5 min, followed by an increase of methanol to 95% over 0.1
min and followed by a 0.9 min hold at 95% methanol and 5 min
equilibration at 2% methanol with a flow rate of 0.3 mL/min,
and the eluent was diverted to waste before 3 min and after 5
min.
The mass spectrometer was operated in positive mode
electrospray, and data were obtained in the MS/MS mode with
an isotopically labeled internal standard. Precursor/product
transitions monitored were as follows: NDMA 75.048/43.000*,
75.048/58.000, and 75.048/44 and d6-NDMA (internal stand-
ard) 81.078/46.067* and 81.078/64.058 (* indicates transition
used for quantitation), with the precursor ion mass set to m/z =
75.048. The method achieved a limit of detection of 0.02 mcg/g
and a limit of quantification (LOQ) of 0.07 mcg/g for NDMA in
the ranitidine drug substance.
(E/Z)-N-(2-{[(5-{[Bis(2H3 methyl)amino]methyl}furan-2-yl)-
methyl]thio}ethyl)-N′-methyl-2-[15N]nitroethene-1,1-dia-
mine Hydrochloride 21. A solution of 5−6 M hydrogen
chloride in 2-propanol (2.3 mL) was added dropwise with
stirring to a solution of 25 (3.95 g) in a mixture of 2-propanol
(7.1 mL) and methanol (3.55 mL) at room temperature. The
mixture was seeded with a few grains of unlabeled ranitidine
hydrochloride and stirred at room temperature for 2 h. The solid
was filtered, washed with ice-cold 2-propanol (4 × 2.5 mL), and
dried under vacuum at room temperature overnight to afford
crude 21 as a buff-colored solid (3.95 g, 90%).
2-Propanol (30.8 mL) was added over 45 min to a stirred
solution of crude 21 (3.85 g) in methanol (15.4 mL) at 46−50
°C under a nitrogen atmosphere; the resulting solution was
cooled to 40 °C; 0.3 mL was removed and triturated, and the
resulting solid was added to the solution. The mixture was
stirred at 40 °C for 1 h and then cooled to 5 °C over 2 h and
stirred overnight.
The slurry was filtered, and the cake was washed with cold 2-
propanol (2 × 20 mL), air-dried, and then dried under vacuum
at 29 °C for 6.5 h and then at ambient temperature to afford 21
as an off-white solid (3.44 g, 88%). mp 134.4−134.8 °C (dec.).
1H NMR (600 MHz, DMSO-d6): δ 2.60−2.95 (m, 5H), 3.22−
3.55 (m, 2H + H2O), 3.87 (s, 2H), 4.31 (s, 2H), 6.36−6.59 (m,
1H), 6.41 (br s, 1H), 6.65 (d, J = 3.08 Hz, 1H), 7.27−7.57 (m,
1H), 9.80−10.15 (m, 1H), 10.72 (br s, 1H). HRMS (m/z): [M
J
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Equivalence between the method conditions described above
and the FDA method conditions18 has been demonstrated.
Method 2. Ultra-HPLC and HRMS (UHPLC−HRMS)
analysis was performed using a Thermo Scientific Ultimate 3000
UHPLC system (Waltham, MA) coupled to a Thermo Fisher
Scientific Q Exactive Plus mass spectrometer (Waltham, MA).
Instrument parameters, data acquisition, and data processing
were controlled by Thermo Fisher Scientific Xcalibur software
(Waltham, MA).
The analysis was run on an ACE chromatography C18-AR 50
× 4.6 mm, 3 μm column (Aberdeen, UK) heated to 30 °C and
eluted using deionized water and acetonitrile with a 0.1% (v/v)
formic acid gradient of increasing acetonitrile from 5 to 20%
over 2 min, followed by a further increase to 100% over 7 min
and followed by a 2 min hold at 100% acetonitrile and 5 min
equilibration at 5% acetonitrile with a flow rate of 0.5 mL/min,
and the eluent was diverted to waste between 3.2 and 14 min.
The mass spectrometer was operated in positive mode
electrospray; the conditions were as follows: capillary temper-
ature400 °C, source heater temperature350 °C, Sheath gas
flow55 (arbitrary units), aux gas flow15 (arbitrary units),
sweep gas flow0 (arbitrary units), source voltage3.5 kV,
FTMS MSn AGC target20,000, and FTMS MSn max ion
time100 ms.
Clara, CA) coupled to a Thermo Fisher Scientific Q Exactive
Plus mass spectrometer (Waltham, MA). Instrument parame-
ters, data acquisition, and data processing were controlled by
Thermo Fisher Scientific Xcalibur software (Waltham, MA).
The analysis was run on an ACE chromatography C18-AR 50
× 4.6 mm, 3 μm column (Aberdeen, UK) heated to 30 °C and
eluted using deionized water and acetonitrile with a 0.1% (v/v)
formic acid gradient of increasing acetonitrile from 5 to 100%
over 7 min, followed by a 2 min hold at 100% acetonitrile and 5
min equilibration at 5% acetonitrile with a flow rate of 0.5 mL/
min, and the eluent was diverted to waste between 3.2 and 13
min.
The mass spectrometer was operated in positive mode
electrospray; the conditions were as follows: capillary temper-
ature400 °C, source heater temperature350 °C, sheath gas
flow55 (arbitrary units), aux gas flow15 (arbitrary units),
sweep gas flow0 (arbitrary units), source voltage3.5 kV, S-
Lens RF level 70%, FTMS AGC target200,000, and FTMS
MSn max ion time1000 ms.
Data were obtained in the PRM mode. The precursor ion
mass was set to m/z = 75.0553 with a 1.5 m/z unit isolation
window and subjected to 80% HCD CE excitation. The product
ion scan range was m/z 50−95, with an Orbitrap resolution of
35,000. A mass chromatogram was obtained for m/z 75.0553
with a 15 ppm window, smoothed using a 7-point Gaussian
smooth.
Data were obtained in the MS/MS mode. The precursor ion
mass was set to m/z = 75.0553 with a 1.5 m/z unit isolation
window and subjected to 80% HCD CE excitation. The product
ion scan range was m/z 50−95, with an Orbitrap resolution of
35,000. A mass chromatogram was obtained for m/z 75.0553
with a 15 ppm window. The method achieved a limit of
detection of 0.01 mcg/g and an LOQ of 0.03 mcg/g for the
ranitidine drug substance.
Method 3. HPLC−HRMS analysis was performed using an
Agilent Technologies 1260/1290 modular LC system (Santa
Clara, CA) coupled to a Thermo Fisher Scientific Velos Plus
Orbitrap hybrid mass spectrometer (Waltham, MA). Instrument
parameters, data acquisition, and data processing were
controlled by Thermo Fisher Scientific Xcalibur software
(Waltham, MA).
The analysis was run on an ACE chromatography C18-AR 50
× 4.6 mm, 3 μm column (Aberdeen, UK) heated to 30 °C and
eluted using deionized water and acetonitrile with a 0.1% (v/v)
formic acid gradient of increasing acetonitrile from 5 to 100%
over 7 min, followed by a 2 min hold at 100% acetonitrile and 5
min equilibration at 5% acetonitrile with a flow rate of 0.5 mL/
min, and the eluent was diverted to waste between 3.2 min and
13 min.
Method 5 NDMA Fusion Method (Multi-Isotopologue).
HPLC−HRMS analysis was performed using an Agilent
Technologies 1290 infinity II LC system (Santa Clara CA)
coupled to a Thermo Fisher Scientific Fusion hybrid mass
spectrometer (Waltham, MA). Instrument parameters, data
acquisition, and data processing were controlled by Thermo
Fisher Scientific Xcalibur software (Waltham, MA).
The analysis was run on an ACE chromatography C18-AR 50
× 4.6 mm, 3 μm column (Aberdeen, UK) heated to 30 °C and
eluted using deionized water and acetonitrile with a 0.1% (v/v)
formic acid gradient of increasing acetonitrile from 5 to 100%
over 7 min, followed by a 2 min hold at 100% acetonitrile and 5
min equilibration at 5% acetonitrile with a flow rate of 0.5 mL/
min, and the eluent was diverted to waste between 3.2 and 13
min.
The mass spectrometer was operated in positive mode
electrospray; the conditions were as follows: capillary temper-
ature400 °C, source heater temperature450 °C, sheath gas
flow20 (arbitrary units), aux gas flow8 (arbitrary units),
sweep gas flow0 (arbitrary units), source voltage2.5 kV, RF
lens 50%, normalized AGC target400%, and maximum
injection time246 ms, with data obtained in the MS/MS
mode.
A precursor ion mass of m/z = 75.0553 precursor selection
(m/z 75.0553 for the unlabeled, m/z 76.0523 for the 15N label,
m/z 81.0929 for the d6 label, and m/z 82.0900 for the d6-15N
label isotopologues) was achieved using quadrupole isolation
with an isolation window of 0.9 m/z units and multiplexing of all
isotopologues. The product ion scan range was m/z 50−95, with
an Orbitrap resolution of 120,000. A mass chromatogram with a
15 ppm window was obtained for each isotopologue at the same
m/z as the precursor.
The mass spectrometer was operated in positive mode
electrospray; the conditions were as follows: capillary temper-
ature350 °C, source heater temperature350 °C, sheath gas
flow50 (arbitrary units), aux gas flow15 (arbitrary units),
sweep gas flow0 (arbitrary units), source voltage3.5 kV, S-
Lens RF level 70%, FTMS MSn AGC target50000, and
FTMS MSn max ion time1000 ms.
Data were obtained in the MS/MS mode. The precursor ion
mass was set to m/z = 75.0553 with a 0.5 m/z unit isolation
window and subjected to 80% HCD CE excitation. The product
ion scan range was m/z 50−250, with an Orbitrap resolution of
100,000. A mass chromatogram was obtained for m/z 75.0553
with a 15 ppm window, smoothed using a 7-point Gaussian
smooth.
AUTHOR INFORMATION
Corresponding Author
■
Method 4. HPLC−HRMS analysis was performed using an
Agilent Technologies 1260/1290 modular LC system (Santa
Andrew D. Searle − Medicines Research Centre,
GlaxoSmithKline, Stevenage, Hertfordshire SG1 2NY, U.K.;
K
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(10) https://www.fda.gov/media/130801/download (accessed Sep-
tember 2019).
(11) https://www.who.int/medicines/publications/drugalerts/
InformationNote_Nitrosamine-impurities/en/ (accessed November
2019).
(12) Wu, Y.; Levons, J.; Narang, A. S.; Raghavan, K.; Rao, V. M.
Reactive Impurities in Excipients: Profiling, Identification and
Mitigation of Drug-Excipient Incompatibility. AAPS PharmSciTech
2011, 12, 1248.
(13) Haywood, P. A.; Martin-Smith, M.; Cholerton, T. J.; Evans, M. B.
Isolation and Identification of the Hydrolytic Degradation Products of
Ranitidine Hydrochloride. J. Chem. Soc., Perkin Trans. 1 1987, 951.
(14) Teraoka, R.; Otsuka, M.; Matsuda, Y. Effects of Temperature and
Relative Humidity on the Solid-state Chemical Stability of Ranitidine
Hydrochloride. J. Pharm. Sci. 1993, 82, 601.
(15) Abe, Y.; Yamamoto, E.; Yoshida, H.; Usut, A.; Tomita, N.;
Kanno, H.; Masada, S.; Yokoo, H.; Tsuji, G.; Uchiyama, N.;
Hakamatsuka, T. Temperature-Dependent Formation of N-Nitro-
sodimethylamine during the Storage of Ranitidine Reagent Powders
and Tablets. Chem. Pharm. Bull. 2020, 68, 1008−1012.
(16) Brittain, R. T.; Harris, D. M.; Martin, L. E.; Poynter, D.; Price, B.
J. Safety of Ranitidine. Lancet 1981, 318, 1191.
orcid.org/0000-0002-5079-5302; Email: andy.d.searle@
gsk.com
Authors
Fiona J. King − Medicines Research Centre, GlaxoSmithKline,
Stevenage, Hertfordshire SG1 2NY, U.K.
Michael W. Urquhart − Medicines Research Centre,
GlaxoSmithKline, Stevenage, Hertfordshire SG1 2NY, U.K.;
orcid.org/0000-0002-8626-0123
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00462
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
́
(17) Roux, J. L.; Gallard, H.; Croue, J.-P.; Papot, S.; Deborde, M.
■
Environ. Sci. Technol. 2012, 46, 11095. Shen, R.; Andrews, S. A. Water
Res. 2013, 47, 802. Jeon, D.; Kim, J.; Shin, J.; Hidayat, Z. R.; Na, S.; Lee,
Y. NDMA Formation by Chloramination of Ranitidine: Kinetics and
Mechanism. J. Hazard Mater. 2016, 318, 802.
The authors thank the following for their contributions: Geoff
Badman and Darren Smith for the synthesis of +7 ranitidine
hydrochloride, John Hayler for helping to prepare this
manuscript and supporting laboratory experiments, Andrew
Dominey for supporting laboratory experiments, Emily Great-
orex for physical property testing, James Wickens for NDMA
testing, Tran Pham for solid-state NMR support, and Tamas
Balogh for running accelerated stability monitoring studies.
(18) LC−HRMS Method for the Determination of NDMA impurity
in Ranitidine Drug Substance or Drug Product, FDA, 09/13/2019.
(USFDA, Liquid chromatography-high resolution mass spectrometry
(LC−HRMS) method for the determination of NDMA in ranitidine
drug substance and drug product, https://www.fda.gov/media/
130801/download).
REFERENCES
■
(1) ICH Harmonised Guidance. Assessment and Control of DNA
Reactive (Mutagenic) Impurities in Pharmaceuticals to Limit Potential
Carcinogenic Risk M7 R1, March 2017.
(2) https://www.ema.europa.eu/en/news/ema-reviewing-
medicines-containing-valsartan-zhejiang-huahai-following-detection-
impurity-some (accessed May 2018).
(3) https://www.ema.europa.eu/en/medicines/human/referrals/
angiotensin-ii-receptor-antagonists-sartans-containing-tetrazole-group
(accessed September 2018).
(4) https://www.ema.europa.eu/en/documents/press-release/ema-
review-ranitidine-medicines-following-detection-ndma_en.pdf (ac-
cessed September 2019).
(5) Allen & Hanbury’s Ltd.; Price, B. J.; Clitherow, J. W.; Bradshaw, J.
Aminoalkyl furan derivatives. GB 1565966 A, 1977, which claims
priority to applications filed in 1976.
(6) Finar, I. L. Organic Chemistry Volume 1: The Fundamental
Principals, 3rd ed.; Longmans, Green and Co. Ltd., 1959; p 300.
Anselme, J.-P. The Organic Chemistry of N-Nitrosamines: A Brief
Review. ACS Symposium Series; American Chemical Society, 1979.
Saavedra, J. E. Recent Synthetic Applications of N-Nitrosamines and
Related Compounds. Org. Prep. Proced. Int. 1987, 19, 85.
(7) Kodamatani, H.; Sugihara, K.; Tanisue, T.; Kanzani, R.; Tomiyasu,
T. Contamination, decomposition and formation of N-nitrosodime-
thylamine in water samples at the ng/L level of determination. Anal. Sci.
2020, 36, 1393−1399.
(8) Aitken, R. A.; Aitken, K. M. Product class 1: Synthesis of
nitroalkane. Sci. Synth. 2010, 41, 9. : Process for the preparation of
nitromethane from dimethyl sulfate and sodium nitrite Yang, S.; Xiong,
X.; Shi, S.; Yu, D.; Assignee Hubei Yuanda Fuchi Pharmaceutical
Chemicals Co., Ltd., PR China. Novel method for producing
nitromethane. CN 102659602 A, 2012.
(9) Teasdale, A.; Elder, D.; Chang, S.-J.; Wang, S.; Thompson, R.;
Benz, N.; Sanchez Flores, I. H. Risk assessment of genotoxic impurities
in new chemical entities: strategies to demonstrate control. Org. Process
Res. Dev. 2013, 17, 221.
L
https://dx.doi.org/10.1021/acs.oprd.0c00462
Org. Process Res. Dev. XXXX, XXX, XXX−XXX