Tolerability of Risk: A commentary on the nitrosamine

contamination issue

Journal: Journal of Pharmaceutical Sciences

F
Manuscript ID 20-1285

Article Type: General Commentary

Date Submitted by the

23-Nov-2020
Author:

Complete List of Authors: Elder, David; David P Elder Consultancy, CMC

Johnson, George; Swansea University, Swansea University Medical
School
Snodin, David; Xiphora Biopharma Consulting, Nonclinical Drug
Development

Cancer, Deoxyribonucleic acid (DNA), Dose-response, Drug standards,

Keywords:
Mutation(s), Nonlinear regression(s), Toxicity, Toxicology, safety
based limits, ppb, risk assessment, clinical safety, metabolism,
Tolerability of Risk: A commentary on the nitrosamine contamination issue

D.P. Elder, G. Johnson, D.J. Snodin

Abstract

For decades, regulators have grappled with different approaches to address the issue of control of
impurities. Safety-based limits, such as permissible daily exposure (PDE), acceptable intake (AI),
threshold of toxicological concern (TTC) and less than lifetime limits (LTL) have all been used. For many
years these safety-based limits have been recognized as virtually safe doses (VSDs). Recently, however,
many regulatory agencies are seeking to impose limits for N-nitrosamine impurities, which are
significantly below the VSD. This commentary will discuss the evolution of safety-based limits for
impurities, provide an overview of the valsartan N-nitrosamine contamination issue and review the
toxicology of N-nitrosamines. The outcome of a lessons-learned exercise on sartan medications
undertaken by the European Medicines Agency (EMA) will also be discussed. The review will also
highlight the many analytical challenges inherent with controlling impurities to ppb-based limits. The
use of highly sensitive, low ppb limits, methods may lead to future issues of batch rejection, based on
false positives. Regulators initially viewed the N-nitrosamine risk as being insufficient to prompt
immediate product discontinuation and patients were specifically advised to continue using their
affected medication. Patients were also informed that exposure to N-nitrosamines is extremely
common via food and drinking water.

Introduction

Impurities and degradation products that are by-products of the manufacture and storage of medicinal
products afford no benefits to the patient, only potential risk1 with certain exceptions including closely
related impurities in proteins, peptides, or semi-synthetic products, e.g. oligonucleotides, beta-lactams,
etc., where these analogues frequently retain beneficial biological activity and are often controlled as
families or classes of impurities.2 Regulators have long grappled with different approaches to address
this intrinsic issue. Strategies have included risk-based assessments, e.g. ICH Q3D(R1)3, ICHM7(R1)4,
avoidance, e.g. N-nitrosamine impurities5, as low as reasonably practicable (ALARP)6, or safety-based
limits, including permissible daily exposure (PDE) [employed for residual solvents in ICH Q3C(R6)7 and
elemental impurities in ICH Q3D(R1)3], threshold of toxicological concern (TTC), less than lifetime (LTL)
limits or acceptable intakes (AIs) [used for mutagenic impurities in ICH M7(R1)4]; and sometimes a
combination of all of these approaches [ICH M7(R1)4].

However, many industry commentators have reflected that since some level of impurities is an
inevitable outcome of the synthesis of complex, multifunctional active pharmaceutical ingredients
(APIs), the complete avoidance of risk from impurities is not feasible.6 Similarly, whilst meaningful
measures are implemented to minimize degradation within a drug product during storage, breakdown
products are still likely to be encountered owing to the unavoidable presence of moisture and oxygen
within the storage environment, leading to the two most common degradation pathways, i.e.
hydrolysis and oxidation.8 As such, there needs to be an underlying acceptance that there will be low,
but controlled levels of residual impurities/degradation products arising from the synthesis of drug-
substance or drug-product manufacture and storage, i.e. a ‘tolerability of risk’ approach.9 Toxic
impurities, particularly those with mutagenic potential, pose special challenges, but even here a
combination of risk-based strategies and virtually safe doses (VSDs), i.e. TTC, LTL or AI, can be used to
safely control the risk to patients. On the other hand, managing extremely toxic impurities, such as
some N-nitrosamines, does challenge the VSD orthodoxy and may reintroduce concepts such as
avoidance and ALARP that were employed prior to the inception of ICH M7(R1)4, i.e. the period
between 2009-2014.6 This review will focus on the evolution of safety-based guidance for impurities
[with a special focus on those belonging to the Cohort of Concern (CoC), including an evaluation of their
metabolism and toxicology]. In addition, discussion is focused on what happened during the N-
nitrosamine contamination crisis, the reasons for its occurrence and measures undertaken to address
the issue. A critique of EMA’s lessons-learned review and a commentary on some of the proposed
changes to existing guidance will be provided. Finally, the review will discuss whether the responses to
the crisis are proportionate given the significant daily exposure to N-nitrosamines from diet and
drinking water and whether the apparent reluctance to advise patients to discontinue impacted
medicines is appropriate.

Evolution of Safety Based Guidance for Pharmaceutical Impurities

Impurities
Impurities are controlled mainly by quality guidelines within the framework of the International
Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH): i.e.,
ICH Q3A(R2)10, Q3B(R2)11, Q3C(R6)7, Q3D(R1)3, Q3E12, Q5C13, Q6A14, Q6B14. Other guidance is
contained in multi-disciplinary guidelines, i.e., ICH M3(R2)15 and M7(R1)4, and impurities in anti-cancer
medications are controlled by reference to ICH S916,17. ICH Q3C(R6)7 was the first of the impurity
guidelines that provided safety-based guidance on allowable limits of common residual solvents. This
guideline uses either PDE limits and/or the corresponding concentration-based limits, i.e. ppm - but
the ppm limits are more restrictive being based on a standard dose of 10g . Interestingly, Class 1
solvents (“to be avoided”) use only ppm-based limits, i.e. benzene < 2 ppm; but some commentators
have argued that a PDE-based limit would introduce more flexibility and consistency.18 Class 2 solvents
(“to be limited”) use specific PDE and ppm limits interchangeably, i.e. acetonitrile 4.1mg/day (PDE) or
410 ppm. Finally, Class 3 solvents (“low toxic potential”) use a default class-based PDE capped at
50mg/day, corresponding to a concentration limit of 5000 ppm limit. Actual PDEs for some Class 3
solvents are many multiples of the class-based PDE; for example; acetic acid 3200 mg/day; acetone
210 mg/day; formic acid 180 mg/day; heptane 840 mg/day. Nevertheless, the concept of ALARP is
sometimes introduced by quality reviewers for Class 3 solvents (although not specifically mandated by
regulatory agencies); process-capability arguments may be used to reduce class 3 solvent levels well
below the ICH Q3C(R6)7 safety- and/or class-based limits.17

ICH Q3D(R1)3 defines limits of residual elemental impurities. ICH Q3D(R1)3 also uses PDEs and
establishes values for different routes of administration, i.e. oral, inhalation and parenteral, and
introduces a risk-based approach. There is no expectation to tighten PDE limits based on process
capability considerations provided that PDE values are not exceeded. However, specific elements can
be removed from the drug substance specification if they can be shown to be absent routinely at levels
equivalent to 30% of the designated PDE [cf. ICH Q3C(R6)7 and ICHM7(R1)4].

Finally, ICHM7(R1)4 defines limits of mutagenic impurities (MIs) within a drug substance or drug
product. MIs are controlled on the assumption that they can potentially cause cancer in man. A TTC
approach, originally based on food safety considerations, was developed to describe an acceptable
intake for any novel MI. This equates to a generic allowable exposure of 1.5 μg/day based on a
carcinogenic risk of 1 in 100,000, which is viewed as a VSD. The control strategies for MIs are based on
established risk-assessment approaches.

The acceptable excess risk of carcinogenic potential is considered greater for early development
phases, i.e. 1 in 1,000,000, because the acceptable risk in healthy human volunteers is lower than
that in patients, although the total exposure duration is limited, i.e. LTL. Thus, for example, if the
acceptable intake is 1.5 µg/day for lifetime exposure, the LTL limits can be increased to a daily intake of
120 µg (< 1 month), 20 µg (> 1-12 months) or 10 µg (> 1-10 years’ treatment duration). These LTL limits
are applicable to both clinical development and for commercial products. In the latter case, ICH
M7(R1) provides guidance in note 7 on those drug classes that can in theory utilize LTL limits. For
example, the 120 µg/day limit for treatment duration of < 1 month; can be used for those drugs used in
”emergency procedures (antidotes, anesthesia, acute ischemic stroke), actinic keratosis, and the
treatment of lice”. In contrast, for commercial products, the lower cancer risk level of 1 in 100,000 is
considered acceptable for lifetime exposure, i.e. 70 years. MIs are then classified into five different
categories in order of decreasing regulatory concern:
• Class 1, known mutagenic carcinogens. Control at or below compound-specific limit, i.e. AIs
or PDEs. For example, hydrazine has an AI of 39μg/day, based on a TD50 linear extrapolation
(ICHM7(R1)) and aniline has a threshold-based PDE of 720μg/day. Interestingly, AIs but not PDEs can
be adjusted using the LTL factors.
• Class 2, known [or (Q)SAR-predicted] mutagens with unknown carcinogenic potential.
Control at or below allowable limits, i.e. LTL or TTC

• Class 3, contain alerting structures (not related to API) with no supporting mutagenicity data.
Control at or below allowable limits, i.e. LTL or TTC. Or conduct bacterial mutagenicity assay; if non-
mutagenic = Class 5, if mutagenic = Class 2
• Class 4, contain alerting structures related to known structures or the specific API, which are
non-mutagenic. Treat as non-mutagenic impurity, i.e. use default ICH Q3A(R2)10/Q3B(R2)11 limits

• Class 5, contain no alerting structures, or alerting/mutagenic but non-carcinogenic. Treat as

non-mutagenic impurity, i.e. use default ICH Q3A(R2)/Q3B(R2) limits ICH M7(R1)4 includes four control
options for API MIs; of these only one stipulates control of the MI in the API specification (option 1).
Options 2 and 3 identify some levels of in-process control; in contrast, option 4 is focused on process
understanding alone, typically purge-factor arguments.19

The control strategies for MIs in oncology products are different since the latter are outside the scope
of ICHM7(R1)4 given that the risk/benefit ratio in advanced-cancer patients differs from that in the
normal population. As such, MIs in oncology products can be controlled at higher levels (normally
based on ICH Q3A(R2)10/3B(R2)11 criteria). For other indications when the drug substance is itself
genotoxic at therapeutic concentrations and may be expected to be associated with an increased
cancer risk, a similar approach on impurities is recommended in ICH M7(R1)4. Conversely, this
pragmatic risk/benefit based approach is not routinely applied to other (rare) diseases where life
expectancy is low.

Finally, a sub-class of MIs described as the CoC, including N-nitroso compounds, aflatoxin-like and alkyl
azoxy compounds, were deliberately excluded from TTC or AI-based control strategies as they were
considered to be highly toxic (typically potent mutagenic carcinogens). This is despite the fact that
not all N-nitrosamines, for example N-nitrosodiphenylamine and N-nitroso-2,6-dimethylpiperidine, are
mutagenic. In addition, Thresher et al.20 recently concluded that nearly one-fifth of N-nitrosamines
were considered to be non-carcinogenic in nature.
Metabolic Activation of N-nitrosamines

N-Nitrosamines are generally considered to be indirect-acting mutagens requiring cytochrome P450

(CYP2E1) activation by hydroxylation of the carbon atom in the α-position relative to the nitroso
moiety.21 There are some reports indicating that radical mechanisms are involved in the initial α-
hydroxylation reaction. Further reactions produce an alkyl diazonium ion that is thought to be the
proximate mutagen.

Toxicity of N-nitrosamines

Dose-response functions in genetic toxicology are often reported to be non-linear and thresholded,
which applies for example to direct-acting alkylating agents, polyaromatic hydrocarbons (PAHs),
aneugens and topoisomerase inhibitors.22-27 Indeed, ICH M7(R1)4 recognized this fact: “The existence
of mechanisms leading to a dose response that is nonlinear or has a practical threshold is
increasingly recognized, not only for compounds that interact with non-DNA targets but also for

DNA-reactive compounds, whose effects may be modulated by, for example, rapid detoxification
before coming into contact with DNA, or by effective repair of induced damage.” The point of
(PoD) approach is a useful technique for defining acceptable exposure limits in humans.28 .
departure 50
The PoD is the point on a dose-response curve established from experimental data from which
extrapolation below which the PoD may be employed, in conjunction with the application of
uncertainty factors, for low-dose risk assessment and determination of an acceptable exposure level,
or reference dose. The BMDL10 (benchmark dose modelling) approach, leading to a determination of an
appropriate PoD, is considered a more realistic starting point for excess cancer risk assessments than
the methodology using linear extrapolation of the dose that results in a 50% increased tumor incidence
over background (TD50).4 The latter approach assumes a straight line from the TD50 to the background
cancer frequency at zero dose, and is considered to be more conservative and therefore tends to
overestimate the real risk; but the TD50 approach (based on values listed in the CPDB - Carcinogenic
Potency Database)29 – has become the preferred potency metric for risk assessments under ICH
M7(R1)4.

Initially, most agencies set AI-based interim limits for N-nitrosamines (NDMA and NDEA) calculated in
accord with ICHM7(R1)4 using TD50 metrics. A read-across approach was utilized for some N-
nitrosamines considered to possess similar alkyl substituents, i.e. NDIPA, NEIPA, NMBA. For other N-
nitrosamines, i.e. MNP, CPNP, a lower class-specific limit of 18ng/day (derived from the Lhasa
carcinogenic potency database) can be used as a default option, although this limit is extremely
conservative given that the N-nitrosamine with the highest potency, NDEA, has a higher AI limit of
26.5ng/day)30 (see Table 1 for naming, abbreviations, structures and allowable intakes of the various
N-nitrosamines).

The CHMP in their updated questions-and-answers update of the article 5(3) review have also
indicated, without a clear explanation, that the use of LTL limits, a mainstay of ICH M7(R1)4, particularly
in early drug development, is not permitted for N-nitrosamines.30

Linear Extrapolation

The risk assessments for NDMA and NDEA to date have been based on the default linear-extrapolation
approach of the cancer potency data.31 Due to the variation in response in each dose group, it is
common for both the linear and non-linear models to provide a comparable fit to the data. With these
substances, we are fortunate to have a ‘mega-study’ (rodent carcinogenicity bioassay) for each32-33, in
that an extended dose range was tested, with an increased number of replicates per dose (15 doses
plus control at n=60). However, even with such an extensive data set, the conclusion of the expert
authors in Zeilmaker et al.34 is that the latent variable models (LVM-E4) that has linear behavior at low
doses, and log-probit model which is sub-linear at low doses, both adequately describe the available
dose-response data. This illustrates that even for these excellent datasets, it cannot be decided
whether the dose-response at the lower end is in fact linear or not.34 Note also plotting on a log-scale
produces sublinear curves (where the slope increases with dose) at the low range, whereas the curves
are linear on the untransformed dose scale.34 [The plot shown in Figure 1 using a log-transformed dose
scale seems to suggest that liver tumor incidence at the lower end of the dose scale is essentially
equivalent to the spontaneous incidence in untreated animals.]

Due to these intricacies of dose-response modelling, it was considered prudent to assess these data
further. The Zeilmaker et al.34 risk assessment was focused on utilizing the data for a margin of
exposure (MOE) calculation. This is an improved assessment compared to the linear back-
extrapolation approach.

The conclusion of linear low-dose modelling for this data set is interesting. The BMD approach is
used to apply a set of suitable models to the data, in order to define a small but measurable increase
above the background or a specified increase in incidence.

Interestingly, when the CHMP’s SWP (Committee for Medicinal Products for human use Safety Working
Party) reviewed the issue of safe levels for the N-nitrosamines in November 2018 they were initially
divided.35 Many committee members (the risk minimization sub-group) thought that control of N-
nitrosamines could be approached in a similar manner to any other mutagenic impurity, based on the
extremely conservative approaches already utilized in ICHM7(R1)4. There was no particular evidence
that NDMA and NDEA were “fundamentally different from other mutagenic carcinogens, which are
covered by the TTC framework in ICH M7(R1)4, besides being more potent”. This increased toxicity can
be evaluated by “defining compound-specific thresholds based on carcinogenicity data and by linear
extrapolation” resulting in an AI – the strategy subsequently adopted for the interim safety-based limits
(see Table 1). As such, there is no requirement for a “no threshold” approach. However, the more
conservative part of SWP (the risk avoidance sub-group) deemed that “there is no need to establish AIs
for impurities in the first place as the contamination risk is considered to be avoidable by avoiding
certain manufacturing processes”.35 However, as ongoing reform of the “Delaney Clause” in the US
highlighted 36, total avoidance or absence of any contaminant is totally impracticable6; and as
regulatory agencies have come to realize this, they have needed to put in place limits to control N-
nitrosamines in medicinal products. However, rather than continue to use the interim AIs, which are
recognized as virtually safe doses, they have implemented methodology-based limits that are entirely
driven by analytical capability and unrelated to classical safety-based limits – a clear case of not
tolerating even minimal risk (see N-nitrosamines methodology).

In addition to the numerous short-chain alkyl N-nitrosamines that have been identified during this
crisis, two N-nitrosovalsartan impurities were also identified. Both were found to be Ames-negative,
presumably because metabolic activation is significantly impaired by steric hindrance. Other Ames-
negative N-nitrosamines include L-proline and hydroxyproline derivatives.

Clinical Safety

The contamination of sartans and other medicinal products with N-nitrosamines was possibly the
principal medicine-related public-health episode of the last decade and the biggest issue since the
heparin contamination scandal of 2007.5,37, 38 What concerned and puzzled industry in equal measures
was how did this issue occur and why the risk was not identified earlier? N-Nitrosamines are part of
the so-called CoC (in fact a subset of N-nitroso compounds). According to perceived wisdom, the
structural classes within the CoC were unlikely to be seen during the routine synthesis of active
pharmaceutical ingredients (APIs).39 In addition, broad, risk-based assessment processes, to evaluate
the potential for formation of mutagenic impurities, were in place throughout industry [ICH M7(R1)4].
Therefore, the announcement in June 2018 that the EMA was “reviewing medicines containing
valsartan drug substance supplied by Zhejianjg Huahai following the detection of an impurity” came as
a significant surprise to many40, particularly, as the impurity of concern was found to be an N-
nitrosamine, i.e. NDMA (N-nitrosodimethylamine) and this impurity had formed as “a result of a
change in the manufacturing process” during the period between 2011-2013.41-43

The principal initial message to patients from world-wide regulatory agencies was that medicines
containing valsartan (which is used for high blood pressure and heart failure) from specific companies
were being recalled; that patients should be given alternative treatments, but it was essential that they
not discontinue their medicines unless directed to do so by healthcare practitioners. Importantly,
patients were told that there was no immediate risk and that there was a higher risk from terminating
treatments. A frequent complaint from patients and patient groups was the absence of clear and
unambiguous information from regulatory agencies’ websites or indeed from healthcare professionals
“as to whether their own medicines were affected”.31 This concern was shared by community
pharmacists who were critical of the absence of detailed instructions for healthcare professionals from
regulatory agencies, and particularly the lack of alternative treatments in the event of recalls being
initiated. FDA44 also sought to downplay the risks of N-nitrosamine contamination as they indicated in
early August 2018, that “NDMA is a known environmental contaminant. For context, it is found in
water and foods including meats, dairy products and vegetables”. Consumer data from Finland and The
Netherlands45 show that typical adult dietary exposure to NDMA from food and beverages is around
100 ng/day (cf. AI limit of 96ng/day), thus providing a MOE of 16200 (> 10,000) based on the SCCS's
(Scientific Committee on Consumer Safety) reference point.5

In early August 2018, EMA’s CHMP provided an initial risk estimate for patients exposed to NDMA
via contaminated valsartan from Zhejiang Huahai.40 They indicated that: ‘Following a preliminary
evaluation, EMA estimates that there could be one extra case of cancer for every 5,000 patients
taking the affected medicines at the highest valsartan dose (320 mg) every day for 7 years. This is
based on average levels of this impurity detected in the active substance from Zhejiang Huahai (60
parts per million).’ FDA also issued similar reassurance stating there “could be one extra case of
cancer for every 5,000 patients over 4 years”.46 FDA updated this risk to one extra case of cancer for
every 8,000 patients over 4 years in early 2019.47 In August of the same year, FDA further downgraded
the risk, stating that “In reality, the vast majority of patients exposed to NDMA through ARBs
(angiotensin II receptor blockers) received much smaller amounts of the impurity than this worst-case
scenario, and, since not all ARBs are affected, it’s very likely that a patient taking an ARB for four years
would not have always received one of the affected products”.48 Global regulatory agencies also put
the risk in the context of the lifetime risk of cancer in their territories and highlighted NDMA potential
exposure from food and water sources.

During 2019, EMA published an updated risk assessment based on a potential “worst case scenario”,
where they assessed joint exposure to the highest observable levels of NDEA for 4 years, between the
years 2011 to 2015, and to NDMA for 6 years, between the years 2012 to 2018. This resulted in a
cumulative theoretical excess risk of cancer of 29.5/100,000 or 1/3390, i.e. 0.029% using ICH M7(R1)
approaches.49 This was then compared to the lifetime risk of cancer in the EU of approximately 50%,
and on that basis, this excess risk was considered to be very low. Excess risk refers to the excess rate of
cancer associated with exposure to these substances.

Most recently, FDA have strongly re-affirmed the underlying risk/benefit message, linked to
contamination of two antibacterial drugs used to treat tuberculosis. FDA indicated that: “To mitigate
or avoid shortages and to help ensure patients have access to these necessary medicines, FDA will
not object to certain manufacturers temporarily distributing rifampin containing 1-methyl-4-
nitrosopiperazine (MNP) or rifapentine containing 1-cyclopentyl-4-nitrosopiperazine (CPNP) above the
acceptable intake limits until they can reduce or eliminate the impurities”. However, FDA made
it clear that levels of MNP must be below 5 ppm in rifampin and rifapentin must contain no more
than 14 ppm of CPNP. This is compared to the AI levels for these N-nitrosamines that are 0.16 ppm of
MNP and 0.1 ppm of CPNP, respectively. Tellingly, FDA reiterated that, “FDA will not object to
these higher exposures to maintain patient access to these life-saving medications” and “… the risk
of not taking the medicine outweighs any potential risk from MNP or CPNP”.50

However, there was still confusion and alarm amongst patient groups. One Canadian study51 showed
increased non-adherence following valsartan recalls with commensurate higher levels of
hospitalization. The authors stated that, ‘Patients may have been willing to risk the short-term
potential of uncontrolled hypertension to avoid ingesting a potential carcinogen.’51

A recent Danish epidemiological study52 reported on the outcome of a clinical study assessing the
 increased cancer risk from patients taking valsartan products that were potentially contaminated
with NDMA. The final study cohort included over 5000 people, who were followed for an average of
years. The authors presumed any valsartan process changes leading to NDMA contamination
were introduced during the period of 2012/2013. They further assumed that all subsequent medicinal
products supplied post-process change would be likely contaminated with NDMA, i.e. a worst case
scenario. The study found 104 cancer outcomes in the placebo group and 198 in the active group. The
adjusted hazard ratio for overall cancer findings was 1.09 with no indication of a dose-response
relationship between the various active groups (P=0.70). For individual cancer outcomes, increased
cancer risk was seen for colorectal and uterine cancers. The authors recommended that longer-term
studies were needed to fully assess any long-term increased cancer risk. A British Medical Journal (BMJ)
editorial53 also called for longer-term assessment of affected patients. However, CHMP did not support
these proposals.31 Firstly, they contended that “theoretical risk of cancer was very low and was itself
based on a worst-case scenario”. Secondly, cancer screening methodologies carry additional risks for
patients. Thirdly, there was still “considerable uncertainty as to which organs or tissues could be at risk
from cancer”.3

Valsartan Contamination Issue

Evolution of a Crisis

NDMA contamination of valsartan products led to a rapid EU product recall of affected batches from
the Chinese generic supplier, Zhejianjg Huahai (ZHP).54 EMA findings of NDMA contamination of
valsartan API from ZHP had world-wide ramifications. The Taiwanese Food and Drug Administration
(TFDA) notified other regulators of the detection of NDMA in valsartan APIs manufactured by two
other Chinese generic companies, Zhejiang Tianyu (ZTP) and Zhuhai Rundu Pharma (ZRP) Pharma.
Based on information received from EMA, FDA reported issues with Valsartan from the same Chinese
supplier, ZHP.55 On 9 August 2018, the FDA also confirmed an N-nitrosamine recall of valsartan
medicinal products from Hetero Labs Ltd in India.56 This was the first indication that the N-nitrosamine
issue was truly global as these products were only sold to US consumers. By the end of August 2018,
FDA indicated that sixteen suppliers of Valsartan API to the US market had been implicated.

As a result of this flurry of findings and ad hoc inter-agency communications, an international

‘Angiotensin II Receptor Blockers (ARB) International Strategic Group’ was formed.57 This group
comprised Health Canada (coordinator), the EMA, US FDA, Japan’s Ministry of Health, Labour and
Welfare/Pharmaceuticals and Medical Devices Agency (MHLW), Australia’s Therapeutic Goods
Administration (TGA), Singapore’s Health Science Authority (HSA), and Swissmedic. The role of the
group was to coordinate activities within the various regulatory agencies and to ensure that there was
visibility and awareness of each other’s activities. The main areas of collaboration included the
determination of risks, assessment of testing methodology, GMP site inspections, and aligned public
communication. Efforts were also made to lessen duplication by having regulators conduct either
concurrent or joint regulatory inspections. The FDA and Health Canada performed several concurrent
inspections of sartan manufacturing sites in affected areas, whilst Health Canada raised the potential of
participating in an EU inspection. Although this international strategic group was highly effective in
coordinating international efforts, the absence of the regulators from India [Central Drugs Standard
Control Organization (CDSCO)] and China [National Medical Product Administration (NMPA)] was, in
retrospect, a deficiency.31

Then, in late August 2018, ZHP informed European regulators of the presence of a second N-
nitrosamine, NDEA, in some batches of its valsartan API. Thereafter, the European regulatory network
was informed on 14 September of trace amounts of NDEA, in another sartan, losartan, from Hetero
Labs in India. EDQM (European Directorate for the Quality of Medicines and Healthcare) informed
European Agencies on 17th September 2018 of the detection of traces of NDEA, in irbesartan, from
another Indian generic API manufacturer, Aurobindo Pharma Limited. This was quickly followed by
identification of NDEA contamination in others sartans, i.e., losartan and irbesartan, from two Indian
generic companies, Hetero Labs and Aurobindo Pharma Limited, respectively.31 In parallel, FDA
reported the presence of NDEA in batches of valsartan API.58 FDA subsequently found low levels NDEA
in Irbesartan from ScieGen59 and losartan in a combination product, losartan/hydrochlorothiazide
manufactured by Sandoz.60

This action prompted European regulatory agencies to risk assess all tetrazole-containing sartans, i.e.
valsartan, candesartan, irbesartan, losartan and olmesartan.61 In mid-September 2019, EMA sent a
formal request to all MAHs for medicinal products “containing chemically synthesized active
substances requiring them to evaluate the risk of N-nitrosamines being present in their products and to
take appropriate risk mitigation measures”.62 Health agencies across the globe followed suite, with
requests to evaluate the risk (step 1), of the presence of N-nitrosamine impurities in all human
pharmaceutical products currently on the market; i.e. EMA, 26 Sep 2019 (Article 31: Sartan containing
tetrazoles)63, Health Canada 02 October 201964, Swissmedic 15 November 201965. MAHs (Marketing
Application Holders) were given 6 months to redress this problem, but because of Covid-19
considerations this was extended to 12 months. Step 2 was confirmatory testing of those products at
high risk of contamination and step 3 was implementation of any modifications to the marketing
authorisation to minimise the issue. The interim limits for the assessment were the AIs derived from
ICHM7(R1)4. WHO also issued an update on the N-nitrosamine issue recommending similar risk-based
strategies to those outlined above.66 A transition period of two years was established within EU, i.e. up
to 02 April 2021; thereafter a technical-based limit of 30 ppb will be implemented.

EMA also highlighted issues of NDMA contamination of pioglitazone manufactured by Hetero Labs in
India during January 2019 – the first non-sartan to be implicated in the crisis.62 No recalls were
implemented as the levels were below the interim AI. This incident was also the first time an N-
nitrosamine was found in an API where its formation did not occur in the final synthetic step; nitrite
sources would need to be carried over to the subsequent synthetic stage, before reacting with a
secondary amine – probably arising from dimethylformamide (DMF), a solvent used in the process.
Other possible root causes were also evaluated, including solvent recycling, which potentially could
lead to DMF being contaminated with NDMA. EDQM reviewed all CEP (Certificate of Suitability to the
monographs of the European Pharmacopoeia) applications and in April 2019, both EMA and NCAs
(National Competent Authorities) within EU requested that all pioglitazone Marketing Authorisation
Holders (MAHs) that were using specified reagents, e.g. NaNO2 in their manufacturing processes check
their products and processes for N-nitrosamine contamination.67
In March 2019, FDA identified a new N-nitrosamine impurity, N-Nitroso-N-methyl-4-aminobutyric acid
(NMBA) in losartan from Hetero Labs in India.68 During April 2019, FDA sought to re-assure US patients
by indicating that there were 40 ARB (angiotensin II receptor blocker) medications where FDA
assessments had concluded that they do not contain any known nitrosamine impurities.69

In the middle of 2019, a fourth N-nitrosamine, N-nitrosomethylphenylamine (NMPA) was found in

valsartan from another Indian generic manufacturer, Divi’s Labs. The levels observed were below the
proposed AI for NMPA.31 In September 2019, Valisure issued a Citizens Petition in the US relating to
ranitidine. They indicated that “Valisure tests all batches of all its medications for quality and
consistency issues and through such tests detected extremely high levels of N-nitrosodimethylamine
(“NDMA”), a probable human carcinogen, in every lot tested, across multiple manufacturers and
dosage forms of the drug ranitidine”.70 They contrasted the typical levels of NDMA found in ranitidine
tablets of 3,000,000 ng/ tablet, i.e. 3 mg/tablet, with the AI limit of 96 ng/day of NDMA for this potent
mutagen. FDA conducted a review and did indeed find elevated levels of NDMA, but these levels were
much lower than reported by Valisure and this was attributed to a methodology issue (see N-
Nitrosamine Methods).71 During September 2019, EMA also initiated an Article 31 review for all
ranitidine products after testing demonstrated NDMA contamination.57 This resulted in several EU
national authorities initiating product recalls. On 12 September, 2019 following the detection of NDMA
in excess of the interim AI, all CEPs for ranitidine hydrochloride were suspended by EDQM .72 There is
some limited evidence that NDMA may form from the degradation of ranitidine itself with higher levels
seen over its shelf life. It is still not clear whether NDMA can also be formed via metabolism in vivo.
EMA has therefore decided that, “Given the uncertainties, the CHMP has recommended a
precautionary suspension of these medicines in the EU”.72

Based on all of these events, EMA arrived at the view that an Article 5(3) review for other medicines
was required. In mid-September 2019, EMA a sent a formal request to all MAHs for medicinal products
“containing chemically synthesised active substances requiring them to evaluate the risk of N-
nitrosamines being present in their products and to take appropriate risk mitigation measures”.62 As a
result of these findings in several different classes of medicinal products, Health Agencies across the
globe initiated requests to evaluate the risk (step 1), of the presence of N-nitrosamine impurities in all
human pharmaceutical products currently on the market.

EMA initiated a separate Article 5(3) review (step 2, i.e. confirmatory testing) assessment of metformin
products in late 2019. FDA indicated that they were aware of low levels of NDMA in metformin drug
products, highlighted by other regulatory agencies, but that the levels were “within the range that is
naturally occurring in some foods and in water .68 FDA subsequently analyzed metformin tablets from
16 batches from seven different companies and reported that all batches were in compliance with
FDA’s limit for NDMA.74

Valisure issued a second Citizens Petition relating to high levels of NDMA that they had found in
metformin products.75 FDA subsequently conducted a review, and found lower levels of NDMA than
reported by Valisure in metformin API and various drug products. The issue was again methodology-
related (see N-nitrosamine methods).76 However, FDA did find levels of NDMA above the AI in
metformin extended-release products and they subsequently requested voluntary recalls of these
products.77 The issue was attributed to the drug product from five suppliers and not the input API.
Other manufacturers of metformin extended-release products were not impacted.
In the early part of 2020, FDA found NMDA in a second H2-receptor blocker drug, i.e. Neftazidine.78 In
June 2020, Health Canada acknowledged the increased likelihood for NDMA formation in doxylamine
succinate API.79Historically, NDMA had been found in doxylamine succinate antihistamine
preparations.80

EMA completed the review under Article 5(3) (i.e. EC 726/2004) in June 2020 that provided guidance to
MAHs on how to avoid the presence of nitrosamine impurities in human medicines. CHMP asked MAHs
in July 2020 to assess all chemical medicinal products, and for the first time, all biological medicinal
products used in man for the potential presence of N-nitrosamines and to test affected products at risk
following the CHMP guidance. This review of biological products complements the similar review of
chemically synthesized active substances (ongoing since September 2019).81 In the middle of 2020,
EMA published a very comprehensive “Lessons Learnt” review. The main findings are critiqued
separately in this commentary.

In the latter half of 2020, a new class of drugs (rifamycin antibiotics, i.e. rifampin (rifampicin in EU) and
rifapentine) were implicated and a further two, novel, N-nitrosamines, i.e. MNP and CPNP were found.
In parallel, the EU also investigated the presence of MNP, in some batches of drug substances used in
rifampicin drug products.82

A schematic of the different types of N-nitrosamines identified thus far, in different medicinal products,
i.e. sartans, pioglitazone, metformin and anti-tuberculosis (TB) products, as well as the different
suppliers (where known) is provided in scheme 1.

Scheme 1: Timelines for Identification of Various N-Nitrosamine Impurities (impurity type, product type
and supplier)83

For both of these tuberculosis drugs, FDA clearly articulated that, “ …the risk of not taking the medicine
outweighs any potential risk from MNP or CPNP”. FDA, allowed higher exposures of these N-
nitrosamines (i.e. higher than their designated AIs) “to maintain patient access to these life-saving
medications”. FDA also indicated that “they would determine on a case-by-case basis whether these
drugs should be released for distribution”.84 EMA gave similar guidance, “The risk to patients from not
taking their rifampicin medicines far outweighs any potential risk from MNP. Healthcare professionals
should therefore continue to prescribe rifampicin medicines as normal in accordance with the product
information”.85
EMA also indicated that they will address any future issues arising from findings in other drug classes,
e.g. metformin86 and rifamycin 85 antibiotics under separate article 5(3) reviews. A schematic of the EU
regulatory response to N-nitrosamine crisis is given in scheme 2; as it can appear confusing to non-EU
nationals.

Scheme 2: EU Regulatory Response to N-nitrosamine Crisis

How did this contamination occur?

As there were no N-nitrosamine issues with the originator valsartan product, Diovan, attention
rapidly focused on the synthetic processes used by ZHP. This generic company replaced tributyltin azide
with sodium azide in the tetrazole ring-formation reaction. This, in turn,
necessitated the use of NaNO2 as an azide-quenching agent (see Scheme 3).87

Scheme 3: Comparison of Penultimate Stage of Synthetic Pathway for Valsartan (ZHP vs. Novartis)87

However, excess nitrite under acidic conditions can also form nitrous acid. This was then free to react
with secondary amines, e.g. dimethylamine, diethylamine, to form the corresponding N-nitrosamine,
i.e. NDMA or NDEA, respectively. It would appear that these volatile dialkyl amines are formed as
impurities in DMF (dimethylformamide) which is utilised as the solvent in the reaction. Although, the
thermal instability of DMF is well documented in the literature88, it was not assessed in the mutagenic
risk assessment. Since the tetrazole ring-forming reaction is the penultimate stage of valsartan
synthesis, there was little opportunity to purge these volatile N-nitrosamines. In contrast, if the
tetrazole ring system is introduced earlier in the synthesis, as with candesartan, there is greater
opportunity to effectively remove these impurities89, even when they need to be controlled to very low
levels, i.e. NDMA, AI of 96ng/day and NDEA, AI of 26 ng/day.

In contrast, levels of NDMA in ranitidine were likely due to the inherent thermal/hydrolytic
instability of ranitidine; as the molecule contains both nitrite and dimethylamine (DMA)
fragments, then NDMA could form due to intra-molecular rearrangements. Nizatidine is similarly
unstable and can form NDMA on storage.90 N-Nitrosamines other than NDMA and NDEA, arising from
different alkyl amines have also being reported, i.e. NMBA.91 Recently, FDA reported that certain
rifamycin antiobiotics, i.e. rifampin and rifapentine have shown N-nitrosamine contamination leading to
two additional analytes, i.e. MNP and CPNP.84

It is now abundantly clear that the N-nitrosation reaction can occur between nitrous acid (or residual
nitrites in an acidic environment) and any small chain alkyl secondary or tertiary amine; indeed, in
some cases reactions have been observed with amine drug substances, e.g. N-nitroso valsartan.

This is summarized in Table 2.

In addition, outsourcing of recycling of solvents, reagents or catalysts by third parties resulted in
additional N-nitrosamine contamination, i.e. all good manufacturing practice (GMP) deviations. FDA
issued a warning letter to Lantech Pharmaceuticals Limited in Telangana, India, for GMP violations.
Lantech is a contract solvent recovery facility used for many valsartan drug substance manufacturing
operations. FDA determined that recovered solvent was contaminated with NDEA; the company had
not assessed the risks linked with their processes and did not effectively assess those impurities.73
However, as a result of contamination arising during recycling there has been a significant increase in
waste solvents, with commensurate environmental implications.

NDBA and NMPA were also found to arise from their corresponding amines, which in turn could be
impurities in the phase transfer catalyst, tetrabutylammonium bromide (TBAB). Finally, contamination
also originated from unsatisfactory cleaning and cross-contamination in large multi-purpose facilities.
FDA recently sent a warning letter to Maylan, Andhra Pradesh, India for “failure to adequately clean
equipment and utensils”, linked with the ARB-contamination issue. 92 Another cause of N-nitrosamine
contamination has been linked with primary blister packaging. The root-cause assessment identified
reactions between amine components within the printing ink and nitrocellulose in the lidding foil
leading to the formation of N-nitrosamines. These volatile N-nitrosamine impurities could then
contaminate drug product during the heat-sealing part of blistering operations. As nitrocellulose lidding
is a very common component in blistering operations it could potentially impact on many products.
However, there is a relatively easy fix for this problem by replacing nitrocellulose lidding components
and the levels of NDMA/NDEA found were below the alerting limits.31

N-Nitrosamine contamination has been linked to residual NDMA levels in water; that is, potable,
pharmacopoeial-grade and other bespoke grades, i.e. water for injection. NDMA contamination can
arise as a by-product of disinfection processes using chloramination, i.e. the reaction of
monochloramine with dimethylamine, a common contaminant in water. It appears that ranitidine
(and other members of its class) show the most significant potential for NDMA formation from water
that was disinfected with chloramine and ozonolysis.93 In addition, NDMA can be formed as a by-
product of chlorination.

NDMA can also arise as a by-product of anion-exchange treatment for water softening. NDMA is
generally destroyed by ultraviolet irradiation during water treatment. The existing WHO guideline
for drinking water ascribes a limit for NDMA of 0.1 ppb arising from different sources in the
environment.94-95 Some parts of Europe and the US have in place lower limits. As a result of its high
aqueous solubility, it is very unlikely that potable or process water constitutes a significant source of
NDMA contamination for APIs. Similarly, water-based processes in secondary manufacturing, i.e. wet
granulation, film-coating, spray drying, lyophilisation or preparation of aqueous solutions are deemed
to be low risk, due to the extremely low levels on N-nitrosamines.31

The final cause of potential contamination arises from residual nitrites in water96 and excipients97.
These residual nitrites could subsequently react with amines (APIs or impurities), reagents and many
solvents, to form N-nitrosamines. Drinking water, and other types of water, are likely contaminated
with nitrite. The WHO has assigned a limit for nitrite in drinking water of 3 mg/L (3μg/g or 3 ppm) as
nitrite ion. Nitrosation of amines in APIs or impurities would be very slow due to low reagent
concentration and high activation energy. As such, suitable conditions for N-nitrosamine formation
would require elevated temperatures and acidic conditions.

N-Nitrosamines arising from water are not considered to be a likely source of contamination of drug
substances. Nitrosation arising from degradation processes in API/drug products are theoretically
possible but extremely unlikely. Disinfected water currently cannot be eliminated as a causative factor,
but it seems unlikely. Ashworth et al.96 used established “mechanistic and kinetic studies of amine
nitrosation to assess the risk that traces of nitrite in the water” utilized during drug substance
manufacturing could generate significant amounts of N-nitrosamines. They concluded that the
concentrations of nitrite in water, that is used for drug substance manufacture are very low, i.e. <0.01
μg/mL, and hence will not generate significant levels of N-nitrosamines by reacting with basic
secondary amines with pKa values greater than 9.5.
Nitrites as reactive impurities are likely to occur in many excipients at ppm levels.97 These include
lactose fast floR, PVP (polyvinyl pyrrolidone), pre-gelatinised starch, croscarmellose sodium and sodium
starch glycollate. Nitrites probably arise during processing.

When reviewing the N-nitrosamine contamination issue holistically, the only example of where
suspicion is focused on the product (rather than the API) are the Metformin ER (extended release)
medications. No issues were seen with the API or corresponding metformin IR (instant release)
products, only the ER products. As such, residual nitrites in the polymeric excipients reacting with
amine sources in the formulation could be the cause, but at this stage that is conjecture. Similarly, it is
not currently clear how the rifamycin antibiotics, i.e. rifampin and rifapentine are contaminated with
MNP and CPNP, respectively.

N-Nitrosamine Analytical Method

European Union

The Official Medicines Control Laboratories (OMCL) network within the EU reacted rapidly to the sartan
crisis. In retrospect it does seem strange that there was initial confusion over the intended analytical
target range.31 Snodin & Elder5 commented that it was illogical that many of the initial
OMCL methods were not aligned with one another with respect to limit of quantitation (LOQ), when it
was clear that the AI for NDMA in valsartan was 0.3 ppm, i.e. 96ng/day. By September 2019, six
methods had been developed by the OMCL consortium.31 None of the OMCL laboratories initially
published method validation as per ICH Q2(R2). As such, many commentators were skeptical that the
HPLC-UV method (high performance liquid chromatography – ultra violet detection) for example,
developed by ANSM (French National Agency for Medicines and Health Product Safety) with a
claimed LOQ of 0.3 ppm could reproducibly analyze samples at these very low levels.98

However, three methods from the OMCL network did use hyphenated GC (gas chromatography) or
HPLC (high performance liquid chromatography) methodology: an HS-GC-MS method (head space-gas
chromatography-mass spectroscopic detection) from PALG (Public Analysts in Galway), with an
NDMA LOQ of 0.04ppm99; an APCI-UHPLC-MS/MS method (atmospheric pressure chemical ionization-
ultra-high performance liquid chromatography - mass spectroscopic detection) from CVUA
(Chemischen und Veterinäruntersuchungsamt)100 with an NDMA LOQ in API of 0.2 ppm and in
drug product of 0.1ppm, and an NDEA LOQ in API of 0.08 ppm and in drug product of 0.04ppm;
finally, the HPLC-MS/MS method (high performance liquid chromatography – mass spectroscopic/mass
spectroscopic) from LGL (Bayerisches Landesamt für Gesundheit und Lebensmittelsicherheit)101 with an
NMBA LOQ of 28.6ppb. Although the network subsequently identified these three hyphenated
techniques as their “methods of choice” in the development of the Ph.Eur. general chapter 2.4.36 (see
below), it is still unclear from the OMCL Network webpage, which methods are considered most
appropriate.102

The European Pharmacopoeia (Ph. Eur.) in collaboration with the OMCL network is developing a new
general chapter on the analysis of N-nitrosamine impurities in active substances (2.4.36).103 The
chapter focuses on the analysis of N-nitrosamines in ARBs (sartans) containing a tetrazole group, for
which there are currently five Ph. Eur. monographs (valsartan, losartan potassium, candesartan
cilexetil, irbesartan and olmesartan medoxomil). Manufacturers of these five ARBs have to implement a
control strategy for N-nitrosamines, and from April 2021 the batches of API they produce must not
contain quantifiable levels, corresponding to < 0.03 ppm, i.e. 30 ppb, of the two main N-nitrosamine
impurities, i.e. NDMA and NDEA.

General chapter (2.4.36) proposes three procedures using either GC-MS, GC-MS/MS and HPLC-
MS/MS covering seven N-nitrosamine impurities: N-nitroso-dimethylamine (NDMA), N-nitroso-
diethylamine (NDEA), N-nitroso-dibutylamine (NDBA), N-nitroso-N-methyl-4-aminobutyric acid
(NMBA); N-nitroso-diisopropylamine (NDiPA), N-nitroso-ethylisopropylamine (NEiPA) and N-nitroso-
dipropylamine (NDPA). The three listed procedures have been validated as limit tests with a target
concentration of 30 ppb, for the designated APIs. Ph. Eur. reflected that: “It was considered important
to include a varied set of procedures using different instruments, thus covering the needs of many
quality control laboratories in Europe and beyond”.103

As a result of a “Lessons Learnt” exercise, EMA have indicated that in the future, EDQM (European
Directorate of Quality of Medicines) will coordinate testing priorities, rather than the individual
OMCLs. In addition, all individual OMCLs will have adequate resources to test for trace levels of MIs,
including modern, hyphenated chromatographic methodologies, i.e. GC-MS and HPLC-MS.31
Page 17 of 37 Journal of Pharmaceutical Sciences

US Food and Drug Administration & USP

In contrast, FDA focused their efforts far more intensely. The original HS-GC-MS method developed
for NDMA in APIs had an LOQ of 0.3 ppm.104 A more sensitive HS-GC-MS for both NDMA and NDEA
was subsequently required and had LOQ limits of 0.1 and 0.05 ppm, respectively.105

During April 2019, FDA issued a combined direct injection GC-MS/MS method for NDMA, NDEA, NEIPA,
NDIPA, and NDBA to address the fact that other N-nitrosamine impurities had being found in valsartan
drug substance and drug products.106 The method showed excellent sensitivity for all four N-
nitrosamines in the API (LOQ 0.05 ppm) and drug product (LOQ 0.05 ppm). This was supplemented with
a headspace GC-MS/MS method for NDMA, NDEA, NEIPA, NDIPA. 107

FDA also issued an HPLC-HR-MS method (high performance liquid chromatography-high resolution-
mass spectroscopic detection) for six N-nitrosamine impurities (NDMA, NDEA, NEIPA, NDIPA, NDBA
and NDBA) in losartan drug substance and drug product at sub-ppm levels (LOQ 0.05 ppm, range: 0.05-
5 ppm).108 This was the first method capable of directly detecting and quantifying NMBA. The
FDA indicated that “the method may also be capable of testing for these six impurities in other ARB
drug substances and drug products pending verification and/or validation”.

FDA also developed and validated a RapidFire-MS/MS method for screening of nitrosamine impurities
NDMA, NDEA, NEIPA, NDIPA, NDBA and NMBA in losartan potassium API batches.109 The method was
determined to be accurate, precise, specific and linear over the corresponding analytical ranges, with
LOQs of 25ppm (NDMA), 50 ppm (NDEA), 0.1 ppm (NEIPA), 0.25 ppm (NDIPA), 0.1 ppm (NDBA) and 0.1
ppm (NMBA), respectively. FDA commented that: “This method would not be sufficient for batch
release purposes for verifying that NDMA or NDEA is not present in drugs intended for human use”.

The findings from the Valisure Citizens Petition on ranitidine70, which cited very high levels of NDMA
in ranitidine tablets, i.e. 3mg/tablet, prompted another FDA method re-assessment as it became clear
that these elevated levels were a result of ranitidine degradation within the GC injection port leading to
artifactually high levels of the analyte. The FDA subsequently developed a HPLC-HR-MS method for
NDMA in ranitidine with an LOQ of 0.033 ppm.71

A second Valisure Citizens Petition75 on high levels of NDMA in metformin products, i.e. API, IR (instant
release) and ER prompted FDA to caution users on potential method-interference issues. FDA rapidly
sought to repeat the analysis of the affected batches. They found that re-testing the samples using two
orthogonal methods demonstrated that FDA’s results were lower than those reported by Valisure. This
was attributed to interference from DMF in the method and that unless a high accuracy MS was used
during data acquisition and processing then over-estimation of NDMA was likely (the molecular weight
difference between DMF and NDMA is only 0.0016 amu or 21 ppm). FDA subsequently reported that
the application of very sensitive methods for routine testing of N-nitrosamines in novel matrices will
not be straightforward.76

The United States Pharmacopeia (USP) have published a draft general chapter <1469> Nitrosamine
Impurities in Pharmacopeial Forum. 110 They report two procedures that can be employed. Procedure
one employs HPLC-HR-MS methodology to quantify six nitrosamines, i.e. NDMA, NDEA, NDIPA, NEIPA,
NMBA and NDBA, with a sensitivity of 1ng/mL (i.e. 1ppb). Procedure two employs HS-GC-MS/S
methodology for the quantitation of four nitrosamines, i.e. NDMA, NDEA, NDIPA, NEIPA and has a
sensitivity of 4 ng/mL i.e. 4ppb).
 Journal of Pharmaceutical Sciences Page 18 of 37

Other Contributors

There is a Swissmedic limit test for the determination of N-nitrosamines (NDMA, NDEA, EIPNA, DIPNA,
DPNA and DBNA) by GC-MS/MS, which is validated for the five impacted ARBs, i.e. valsartan losartan,
irbesartan, olmesartan and candesartan. The method highlights that for other APIs or finished
products, that in-situ validation, with a focus on extraction, specificity and quantification will be
required.111 Health Canada developed a method to detect and quantify the nitrosamine impurities
NDMA and NDEA in ARBs. The procedure is performed by GC-MS-MS using direct injection. The method
cautions that if interference is observed further validation may be required.112

The TFDA issued a series of methods over a 2-year period.113-115 In August 2018 they reported an HPLC-
MS/MS method for NDMA and NDEA. This was initially developed for valsartan, before being extended
to the other affected sartans, ranitidine and metformin APIs and drug products.113 During 2019 they
released an HPLC-MS/MS method for N-nitroso-N-methyl-4-aminobutyric acid (NMBA) in affected
sartan drug substance and their drug products.114 In late 2019 they issued an HPLC-MS/MS method for
twelve N-nitrosamines, i.e. NDMA, NDEA, NDiPA, NDPA, NEIPA, NMBA, NMEA, N-
Nitrosodiethanolamine (NDELA), N-Nitrosomorpholine (NMOR), N-Nitrosopiperidine (NPIP), N-
Nitrosopyrrolidine (NPYR), N-Nitrosodiisopropanolamine (NDiPLA), in all medicines.115

In addition, many instrument manufacturers, academic institutions and companies have published
their own methodologies. However, despite the plethora of available methodologies it is still unclear
to many industrial users which are the best method(s), particularly against a background of ever lower,
“technically”-based limits, i.e. 30 ppb in EU. There appears to be no published account of technology
transfer of these highly sensitive methods between different laboratories to demonstrate that they are
fit for purpose in other laboratories. Equally, the huge demand for ppb-based methods is threatening
to overwhelm world-wide capacity; indeed many CMOs (contract manufacturing organizations) are
diverting capacity from bio-analysis to satisfy demand. In addition, there is no clear picture as to
whether these ppb-based methods are sufficiently rugged and robust to satisfy ongoing demands.

We already know from the Valisure experiences70,75 that a method may be inappropriate for the
intended matrix, i.e. direct injection GC-MS method for NDMA in ranitidine (originally developed for
sartans) or that the method will over-estimate these analytes due to interference or lack of sufficient
mass accuracy within the MS detector, i.e. HPLC-HR-MS method75 for NDMA in metformin products. In
the latter case, a second orthogonal method was required to confirm the correct result.76 This
requirement for a second, highly sensitive orthogonal method, will again negatively impact on
capacity considerations.

Lastly, it has become apparent that implementing extremely sensitive methodologies, i.e. GC-MS(n) or HPLC-
MS(n) to monitor levels of N-nitrosamine impurities in medicinal products at these very low limits is fraught with
uncertainty. The major concern is that there are a limited number of global laboratories and support staff who
can routinely operate this type of methodology and produce reliable data. As with all trace analysis applications,
contamination is a very real issue. The worry is that we will see many more issues of methodology-derived
specification failures, i.e. false positives arising from co-eluting contaminant peaks, e.g. DMF, where analytical
issues result in acceptable product being recalled from the market. In addition to the financial impact, these
product recalls also affect the global supply chain affecting patient access to critical medicines; further
undermining the public’s confidence in the health system.
Page 19 of 37 Journal of Pharmaceutical Sciences

Control Strategies

In the short term most agencies have set AI-based limits using ICHM7(R1).4 In the EU, companies
were then requested to “make necessary changes to their API-manufacturing processes to minimize
nitrosamine contamination”, i.e. avoidance, and they were given a 2-year transition period to
accomplish this; thereafter, they are required to meet much stricter limits, i.e. 0.03 ppm (30 ppb),
which is based on ALARP principles, set on the basis of method sensitivity of those methods validated
by OMCL laboratories.31 So the situation has evolved from control strategies based on safety-based
limits, i.e. AIs, to ALARP-based and finally to concentration-based limits established using the technical
feasibility of the methodology rather than a virtually safe dose.

The USP in its new general chapter <1469> Nitrosamine impurities has a section on control
strategies.116 The ultimate goal of the control strategy is “ensuring that levels of nitrosamines, if their
presence could not be totally avoided, are at or below the provisional acceptable intake (AI)”. The
section provides guidance on how to achieve the goal using risk assessment, “the components of DP
should be assessed for the potential to form nitrosamines or be contaminated with nitrosamines.” The
section includes a high-level process flow to facilitate the generation of a nitrosamine impurity
control strategy.

What Can be Done?

We are currently awaiting after-action reviews (AARs) from many of the major regulatory authorities,
including FDA, PMDA, Health Canada, Swissmedic, etc. However, EMA have issued their “Lessons
Learnt” document.31 The recommendations are fairly wide-ranging and cover many different aspects of
manufacturing, packaging and distribution as well as proposing changes to many of the underpinning
ICH quality guidance documents, i.e. ICH Q7117 and ICH Q9118, and multi-disciplinary guidance
documents, i.e. ICHM7(R1)4. A series of summaries are provided in Table 3-7.
Page 25 of 37 Journal of Pharmaceutical Sciences

Conclusions

The identification of NDMA contamination in valsartan from a Chinese supplier in mid-2018 was the
beginning of an extremely challenging 24-month period for the pharmaceutical industry. Within a short
period of time, the crisis had impacted all sartans medicines containing a tetrazole ring system,
global sartan manufacturers (particularly in India and China) and highlighted that the issue went
beyond simple NDMA contamination – potentially, nitrosamines originating from all short-chain
secondary or tertiary amines can be implicated. Thereafter, other compound classes, e.g. H2
(histamine-2) blockers, i.e. ranitidine, nizatadine, aminoglyside antibiotics, i.e. rifampine as well as
other compounds, e.g. metformin and pioglitazone were also implicated. In addition to intrinsic
contamination resulting from the reaction between secondary/tertiary amines and nitrous
acid/nitrites, investigations also highlighted GMP concerns arising from contaminated
solvents/reagents, degradation of some APIs, e.g. ranitidine, niftazidine and issues with packaging,
i.e. lidding foil containing nitrocellulose and excipient concerns, e.g. metformin extended-release
drug products. The focus has also recently been extended to biologics/diagnostics, at least in the EU.

So concerned were many global regulators than several mandated risk assessments for all medicinal
products in their territories, e.g. Health Canada, Swissmedic, EMA. There were global calls for
avoidance: “there is no need to establish AIs for impurities in the first place as the contamination
risk is considered to be avoidable by avoiding certain manufacturing processes”.49 Nevertheless, it
rapidly became clear that in order to manage the situation in the short-term interim AI limits were
necessary. In the longer term, rather than continue to use these interim AIs, which are recognized as
virtually safe doses, many regulators have implemented technically based limits, i.e. 30 ppb in EU,
that are totally driven by analytical capability and unrelated to classical safety-based limits – a clear
case of not tolerating even minimal risk. These strict limits will not apply to oncology products, where
currently, due to the markedly different risk/benefit considerations for cancer patients, mutagenic
impurities can be controlled to higher limits than those articulated in ICHM7(R1)4. Regulatory agencies
are silent as to whether other diseases with reduced life expectancy, e.g. rare diseases will be similarly
allowed to implement higher limits. Interestingly, FDA is allowing levels of N-nitrosamines in
aminoglyside antibiotics, i.e. rifampine above the designated AI-based limits “tomaintain patient access
to these life-saving medications”.

In addition, it has become apparent that implementing extremely sensitive methodology, i.e. GC-MS(n)
or HPLC-MS(n) to monitor medicinal products at these very low limits is fraught with uncertainty.
Valisure, a US company has issued two Citizens Petitions during this crisis: for rantidine70 and
metformin75. In both cases Valisure reported over-inflated levels of NDMA in these products,
unnecessarily heightening public unease. The major concern is that there are a limited number of
global laboratories and support staff who can routinely operate this type of methodology and produce
reliable data. The worry is that we will see many more issues of methodology-derived specification
failures, where analytical issues result in acceptable product being recalled from the market, further
undermining the public’s confidence in the health system.

In Europe, the CHMP in their updated questions and answers summary of the article 5(3) review have
also indicated, with little convincing evidence, that the use of less than lifetime (LTL) limits, a mainstay
of ICH M7(R1)4, particularly in early drug development, is not permitted.30 Thus, ppb methodology will
need to be employed for sub-optimal processes for early-phase compounds with a high likelihood of
attrition. USP has indicated that, “Nitrosamines are common chemicals in water and foods including
cured and grilled meats, dairy products and vegetables. Everyone is exposed to some level of
nitrosamines”.116 ICH M7(R1)4 also indicated that, “Higher acceptable intakes may be justified when
human exposure to the impurity will be much greater from other sources e.g., food, or endogenous
metabolism (e.g., formaldehyde). Against this background it is surprising that USP (and other regulatory
bodies) should indicate that, “However, their presence in medicines, even at trace level is considered
unacceptable because nitrosamine impurities are probable human carcinogens”.116

At the same time, throughout this crisis regulators have sought to ease patient safety concerns by
restating that N-nitrosamine exposure is widely encountered via food and water and that it was
much more important that patients continued to take their medicines. Recently, FDA43 reiterated that
 Journal of Pharmaceutical Sciences Page 26 of 37

patients taking aminoglyside antibiotics contaminated with N-nitrosamines should continue taking
their medicines based on a risk/benefit assessment. Longer term cancer risk assessments based on
taking N-nitrosamine contaminated valsartan for up to 6 years showed an excess cancer risk of 1/3390.
In a BMJ editorial53 there was a call for longer-term assessment of exposed patients. Nevertheless,
CHMP did not support these proposals. They argued that “theoretical risk of cancer was very low and
was itself based on a worst-case scenario”.31 The lay observer has a right to be confused. Either, the
theoretical risk of excess risk of cancer is low, in which case AI-based limits, which constitute virtually
safe doses are appropriate or the risk to patients is significant and product recalls, avoidance and limits
significantly below the AI are applicable. One thing is certain, it cannot be both scenarios.

Risk-based assessments and ppb chemistry knowledge are intrinsically contradictory in nature and as
FDA indicated, “neither regulators nor industry fully understood how NDMA and NDEA could form
during this particular manufacturing process”.47 In addition, risk is often in the eye of the beholder
particularly when a significant process-chemistry risk masks the potential for another underlying risk.
The valsartan contamination problem arose because NaNO2 was used as an azide-quenching agent to
address the significantly more hazardous (in the minds of the risk assessors) formation of hydrazoic
acid, which poses a dangerous explosive risk when shocked or heated.

Lastly, although some of the proposed fixes for the problems seem logical and even-handed, many
more reflect regulators’ unease in risk-based resolutions; they want analytical data to re-assure
themselves. This crisis highlights a fundamental unwillingness to tolerate risk from pharmaceutical
products, even when we are exposed to the same risk on a daily basis through diet and drinking water.

Abbreviations
PDE (permissible daily exposure); AI (acceptable intake); less than lifetime (LTL) limits; TTC (threshold of
toxicological concern); VSD (virtually safe dose); ALARP (as low as reasonably practicable); Active Pharmaceutical
Ingredients (APIs); International Council for Harmonisation of Technical Requirements for Pharmaceuticals for
Human Use (ICH); Cohort of Concern (CoC); Quantitative Structure Activity Relationships ((Q)SAR); Mutagenic
Impurities (MI ); Potentially Mutagenic Impurities (PMIs); Poly Aromatic Hydrocarbons (PAHs); DeoxyRibonucleic
acid (DNA); Cytochrome(CYP); Benchmark Dose Modelling (BMDL10); Point of Departure (PoD); Dose that
results in a 50% increased tumor incidence over background median toxic dose (TD50); Carcinogenic Potency
Database (CPDB); NMDA (N-nitrosodimethylamine); N-Nitrosodiethylamine (NDEA); N-Nitroso-N-methylamino
butyric acid (NMBA); N-Nitrosodipropylamine (DPNA); N-Nitrosodiisopropylamine (DIPNA); N-
Nitrosoethylisopropylamine (EIPNA); N-nitroso-N-methylaniline (NMPA); 1-methyl-4-nitrosopiperazine (MNP); 1-
cyclopentyl-4-nitrosopiperazine (CPNP); N-Nitrosodiethanolamine (NDELA); N-Nitrosomorpholine (NMOR); N-
Nitrosopiperidine (NPIP); N-Nitrosopyrrolidine (NPYR); N-Nitrosodiisopropanolamine (NDIPLA); Latent Variable
Models (LVM); Margin of Exposure (MOE); SCCS's (Scientific Committee on Consumer Safety); Committee for
Medicinal Products for Human use Safety Working Party (CHMP’s SWP); US Food and Drug Administration (FDA);
European Medicines Agency (EMA); Central Drugs Standard Control Organization (CDSCO); National Medical
Product Administration (NMPA); Japan’s Ministry of Health, Labour and Welfare/Pharmaceuticals and Medical
Devices Agency (MHLW); Australia’s Therapeutic Goods Administration (TGA); Singapore’s Health Science
Authority (HSA); British Medical Journal (BMJ); Taiwanese Food and Drug Administration (TFDA);
Dimethylformamide (DMF); Angiotensin II Receptor Blocker (ARB); Certificate of Suitability to the monographs of
the European Pharmacopoeia (CEP);European Directorate of Quality of Medicines (EDQM); Marketing
Authorisation Holders (MAHs); NCAs (National Competent Authorities); Dimethylamine (DMA);
Diisopropylethylamine (DIPEA); trimethylamine (TEA); diethylamine (DEA);: 4-methylaminobutryic acid (MBA);
N,N- dimethylaniline (N,N-DMA); dibutylamine (DBA); tributylamine (TBA); Good Manufacturing Practice (GMP);
Tetrabutylammonium bromide (TBAB); World Health Authority (WHO); European Pharmacopoeia (Ph. Eur.);
United States Pharmacopeia (USP); United States (US); Official Medicines Control Laboratories (OMCL); French
National Agency for Medicines and Health Product Safety (ANSM); Gas Chromatography (GC); High Performance
Liquid Chromatography (HPLC); Head Space- Gas Chromatography-Mass Spectroscopy (HS-GC-MS); Public
Analysts in Galway (PALG); Atmospheric Pressure Chemical Ionization- Ultra-High Performance Liquid
Chromatography - Mass Spectroscopy (APCI-UHPLC-MS/MS); Chemischen und Veterinäruntersuchungsamt
Page 27 of 37 Journal of Pharmaceutical Sciences

(CVUA); High Performance Liquid Chromatography - Mass Spectroscopy/Mass Spectroscopy (HPLC-MS/MS);

Bayerisches Landesamt für Gesundheit und Lebensmittelsicherheit (LGL); Limit of Quantitation (LOQ); High
Performance Liquid Chromatography-High Resolution-Mass Spectroscopy (HPLC-HR-MS); Parts –per-Million
(ppm); IParts –per-Billion (ppb); Instant release (IR); Extended Release (ER); Contract Manufacturing
Organisations (CMOs); After-Action Reviews (AARs); Marketing Application Authorisation (MAA); Questions &
Answers (Q&As); Active Substance Master File (ASMF); Common Technical Document (CTD); European Union
(EU); Zhejianjg Huahai Pharma (ZHP); Zhejiang Tianyu Pharma (ZTP); Zhuhai Rundu Pharma (ZRP).

Acknowledgements

The authors wish to acknowledge the useful comments and suggestions made by the reviewers in the
compilation of this manuscript.

Conflict of Interest

The authors have received no external funding to support writing this commentary.
 Journal of Pharmaceutical Sciences Page 28 of 37

References

1. Jacobson-Kram, D., McGovern, T. Toxicological overview of impurities in pharmaceutical

products. Adv. Drug Delivery Rev. 59, 2007, 38-42.
2. EMA, 2012. Guideline on setting specifications for related impurities in antibiotics. 30 June 2012.
EMA/CHMP/CVMP/QWP/199250/2009 corr.
https://www.ema.europa.eu/en/documents/scientific-guideline/adopted-guideline-setting-
specifications-related-impurities-antibiotics_en.pdf. Accessed on 19 July 2020.
3. ICH Q3D(R1), 2019. ICH guideline Q3D (R1) on elemental impurities. Current Step 4 Version Dated 22
March 2019.
4. ICH M7(R1), 2017. Assessment and Control of DNA Reactive (Mutagenic) Impurities in
Pharmaceuticals to Limit Potential Carcinogenic Risk. Current Step 4 Version Dated 31 March 2017.
5. Snodin DJ, Elder DP. Short commentary on NDMA (N-nitrosodimethylamine) contamination
of valsartan products. Reg Toxicol Pharmacol. 2019; 103: 325–329.
6. Elder DP, Teasdale A. Is Avoidance of Genotoxic Intermediates/Impurities Tenable for Complex,
Multistep Syntheses? Org Proc Res Dev. 19, 2015,1437-1446.
7. ICH Q3C(R6). ICH guideline Q3C (R6) on impurities: guideline for residual solvents, Step 5. 9
August 2019. EMA/CHMP/ICH/82260/2006. Committee for Human Medicinal Products.
8. Baertschi SW, Alsante KM, Reed RA. Stress testing the chemistry of drug degradation. In
Pharmaceutical Stress Testing. Predicting drug degradation, Eds. Baertschi SW, Alsante KM,
Reed RA. 2nd edition, Informa Healthcare, New York and London, 2011. pp.49-141.
9. Bouder, F., Slavin, D., Lofstedt, R.E. In The Tolerability of Risk: A New Framework for Risk
Management. May 26, 2009, Routledge, pp. 1-160.
10. ICH Q3A(R2). Impurities in new drug substances. Current Step 4 version dated 25 October 2006.
11. ICH Q3B(R2). Impurities in new drug products. Current Step 4 version dated 2 June 2006.
12. ICH Q5C, 1995. Quality of biotechnological products: Stability testing of
biotechnological/biological products. Current Step 4 version dated 30 November 1995
13. ICH Q6A, 1999. Specifications: Test procedures and acceptance criteria for new drug
substances and new drug products: Chemical Substances Q6A. Current Step 4 version dated 6 October
1999.
14. ICH Q6B, 1999. Specifications: Test procedures and acceptance criteria for
biotechnological/Biological products Q6B. Current Step 4 version dated 10 March 1999.
15. ICH M3(R2), 2009. Guidance for non-clinical safety studies for the conduct of human clinical
trials and marketing authorisation for pharmaceuticals M3(R2). Current Step 4 version dated 11 June
2009.
16. ICH S9. Nonclinical evaluation for anticancer pharmaceuticals. Current Step 4 version dated 29
October 2009.
17. Elder DP. Exposure based limits for controlling impurities, Am Pharm Rev. 2017; 20(4): 10-15.
18. Snodin, D.J., McCrossen, S.D. Guidelines and pharmacopoeial standards for pharmaceutical
impurities: overview and critical assessment. Reg Toxicol Pharmacol. 2012; 63(2): 298-312.
19. Teasdale A, Elder D, Chang S-J, Wang S, et al. Risk assessment of genotoxic impurities in new
chemical entities: Strategies to demonstrate control. Org Proc Res Dev. 2013; 17: 221-230
20. Thresher A, Foster R, Ponting DJ, Stalford SA, Tennant RE, Thomas R. Are all nitrosamines
concerning? A review of mutagenicity and carcinogenicity data. Reg Toxicol Pharmacol. 2020; 116:
10474921. Chowdhury G, Calcutt MW, Nagy LD, Guengerich P. Oxidation of Methyl and Ethyl
Nitrosamines by Cytochromes P450 2E1 and 2B1. Biochem. 2012; 51(50): 9995-10007.
22. Thybaud V, Aardema M, Clements J, Dearfield K, Galloway S, Hayashi M et al. Strategy for
genotoxicity testing: hazard identification and risk assessment in relation to in vitro testing. Mutat Res.
2007; 627: 41-58.
 Journal of Pharmaceutical Sciences Page 26 of 37

23. Elhajouji A, Lukamowicz M, Cammerer Z, Kirsch-Volders M. Potential thresholds for

genotoxic effects by micronucleus scoring. Mutagenesis 2011; 26: 199-204.
24. Lynch A, Harvey J, Aylott M, Nicholas E, et al. Investigations into the concept of a threshold
for topoisomerase inhibitor-induced clastogenicity. Mutagenesis 2003; 18: 345-353.
25. Zair ZM, Jenkins GJ, Doak SH, Singh R, Brown K, Johnson GE. N-methylpurine DNA
glycosylase plays a pivotal role in the threshold response of ethyl methanesulfonate-induced
chromosome damage. Tox Sci 2011; 119: 346-358.
26. Health Canada. The Assessment of Mutagenicity. Health Protection Branch Mutagenicity
Guidelines. Health Protection Branch Genotoxicity Committee, Department of National Health and
Welfare, Canada. Environ Mol Mutagen 1993; 21: 15-37.
27. Braithwaite E, Wu X, Wang Z . Repair of DNA lesions induced by polycyclic aromatic hydrocarbons in
human cell-free extracts: involvement of two excision repair mechanisms in vitro. Carcinogenesis 1998;
19: 1239-1246.
28. MacGregor JT Frötschl R, White PA, Crump KS, et al. WGT report on quantitative approaches to
genotoxicity risk assessment II. Use of point-of-departure (PoD) metrics in defining acceptable
exposure limits and assessing human risk. Mut Res. 2015; 783: 66-78.
29. Fitzpatrick RB. CPDB: Carcinogenic Potency Database. Med Ref Serv Q. Fall 2008; 27(3): 303-311.
30. EMA. 2020a. Procedure under Article 5(3) of Regulation EC (No) 726/2004 Nitrosamine
impurities in human medicinal products. EMEA/H/A-5(3)/1490. In: Agency EM, editor.
31. EMA, 2020b. Lessons learnt from presence of N-nitrosamine impurities in sartan medicines.
https://www.ema.europa.eu/en/documents/report/lessons-learnt-presence-n-nitrosamine-
impurities-sartan-medicines_en.pdf. Accessed on 16 October, 2020.
32. Peto R, Gray R, Brantom P, Grasso P. Dose and time relationships for tumor induction in the
liver and esophagus of 4080 inbred rats by chronic ingestion of N-nitrosodiethylamine or N-
nitrosodimethylamine. Cancer Res. 1991a; 51(23 Pt 2): 6452-6469.
33. Peto R, Gray R, Brantom P, Grasso P. Effects on 4080 rats of chronic ingestion of N-
nitrosodiethylamine or N-nitrosodimethylamine: a detailed dose-response study. Cancer Res. 1991b;
51(23 Pt 2): 6415-6451.
34. Zeilmaker MJ, Bakker MI, Schothorst R, Slob W. Risk Assessment of N-nitrosodimethylamine
Formed Endogenously after Fish-with-Vegetable Meals. Toxicol.Sci. 2010; 116(1): 323-335.
35. EMA, 2019a. Assessment Report. EMA/217823/2019, 14 February 2019.
36. Merrill RA. Food safety regulation: Reforming the Delaney Clause. Annu Rev Public Health. 1997;
18: 313–340.
37. Szajek AY, Chess E, Johansen K, et al. The US regulatory and pharmacopoeia response to the global
heparin contamination crisis. Nat Biotechnol. 2016; 34(6): 625-630.
38. Charoo NA, Ali AA, Buha SK, Rahman Z. Lessons learnt from the recall of Valsartan and
otherangiotensin II receptor blockade drugs containing NDMA and NDEA impurities. AAPS
PharmSciTech. 2019; 20(5): 166.
39. Delaney EJ. 2007. An Impact Analysis of the Application of the Threshold of Toxicological
Concern Concept to Pharmaceuticals. Regul Toxicol Pharmacol. 2007; 49(2): 107–124.
40. EMA, 2018a. Update on review of recalled valsartan medicines: preliminary assessment of
possible risk to patients. EMA/526031/2018, 02 August 2018.
41. Christensen J. https://www.kyma.com/health/common-heart-drug-recalled-in-22-countries- for-
possible-cancer-link/764127741. 2018. Accessed on 20 July 2020.
42. Perscheid M. How Did N-Nitroso-Dimethylamine (NDMA) Get into Valsartan?
https://www.linkedin.com/pulse/how-did-n-nitroso-dimethylamine-ndma-mm048709-get-moritz-
perscheid. Accessed on 20 July 2020.
43. Shanley, A. Sartan Recalls Beg the Question: Is Compendial Impurity Testing Enough? Pharm Tech
Eur. 2018; 30(1): 38–42.
 Journal of Pharmaceutical Sciences Page 28 of 37

44. FDA, 2018a. Update- FDA updates recalled valsartan- containing product information and. presents
NDMA levels in some foods, 20 August 2018. https://www.fda.gov/drugs/drug-
safety-and-availability/fda-updates-and-press-announcements-angiotensin-ii-receptor-blocker-arb
recalls-valsartan-losartan#5f16904a2e9db. Accessed on 04 July 2020.
45. Herrmann SS, Duedahl-Olesen L, Christiansen T, Olesen PT, Granby K. Dietary exposure to volatile
and non-volatile N-nitrosamines from processed meat products in Denmark. Food Chem Toxicol. 2015;
80: 137-143.
46. FDA, 2018b. Update- Analysis of N-nitrosodimethylamine (NDMA) levels in recalled valsartan
in the U.S. 27 July 2018. https://www.fda.gov/drugs/drug-safety-and-availability/fda-
updates-and-press-announcements-angiotensin-ii-receptor-blocker-arb-recalls-valsartan-
losartan#5f16904a2edb2. Accessed on 04 July 2020.
47. Gottlieb S, Woodcock J. FDA Statement on the FDA’s ongoing investigation into valsartan and ARB
class impurities and the agency’s steps to address the root causes of the safety
issues. 25 January 2019a. https://www.fda.gov/news-events/press-announcements/fda-
statement-fdas-ongoing-investigation-valsartan-and-arb-class-impurities-and-agencys-steps.
Accessed on 18 July 2020.
48. Woodcock J., 2019. Statement on the agency’s ongoing efforts to resolve safety issue with
ARB medications. 28 August, 2019. https://www.fda.gov/news-events/press-
announcements/statement-agencys-ongoing-efforts-resolve-safety-issue-arb-medications.
Accessed on 18 July 2020.
49. EMA, 2019b. EMA advises companies on steps to take to avoid nitrosamines in human
medicines. 26/09/2019. https://www.ema.europa.eu/en/news/ema-advises-companies-
steps-take-avoid-nitrosamines-human-medicines . Accessed on 04 July 2020.
50. FDA, 2020a. FDA works to mitigate shortages of rifampin and rifapentine after manufacturers find
nitrosamine impurities. 8/26/2020. https://www.fda.gov/drugs/drug-safety-and-availability/fda-works-
mitigate-shortages-rifampin-and-rifapentine-after-manufacturers-find-nitrosamine. Accessed on 15th
October 2020.
51. Jackevicius CA, Krumholz HM, Chong A, Koh M, Ozaki AF, Austin PC, Udell JA, Ko DT. Population
impact of generic valsartan recall. Circulation. 2020; 141(5): 411-413. Epub 2019
Nov 11. https://www.ices.on.ca/Publications/Journal-Articles/2019/November/Population-impact-of-
generic-valsartan-recall. Accessed on 17 July 2020.
52. Pottegård, A., Kristensen, K.B., Ernst, M.T., Johansen, N.B., Quartarolo, P., Hallas, J., BMJ 2018; 362
k3851.
53. Banzi R, Bertele V. Regulatory response to contaminated valsartan. BMJ 2018; 362: k3855.
54. EMA, 2018b. EMA Reviewing Medicines Containing Valsartan from Zhejianjg Huahai following the
Detection of an Impurity. EMA/459276/2018, 5 July 2018
EMA, 2019c. Sartan medicines: companies to review manufacturing processes to avoid
presence of nitrosamine impurities. 01/02/2019.
https://www.ema.europa.eu/en/news/sartan-medicines-companies-review-manufacturing-
processes-avoid-presence-nitrosamine-impurities. Accessed date: 20 July 2020.
55. FDA, 2018c. FDA Announces Voluntary Recall of Several Medicines Containing Valsartan
Following Detection of an Impurity. July 13, 2018.
https://www.fda.gov/newsevents/newsroom/pressannouncements/ucm613532.htm.
Accessed date: O5 July 2020.
56. Camber Pharma, 2018. Camber Pharmaceuticals, Inc. Issues Voluntary Nationwide Recall of
Valsartan Tablets, USP, 40mg, 80mg, 160mg and 320mg Due to The Detection of Trace
Amounts of N-Nitrosodimethylamine (NDMA) Impurity, Found in an Active Pharmaceutical
Ingredient (API). 08/08/2018. https://www.fda.gov/safety/recalls-market-withdrawals-
safety-alerts/camber-pharmaceuticals-inc-issues-voluntary-nationwide-recall-valsartan-tablets-usp-
40mg-80mg-160mg. Accessed on 18 July 2020.
57. Teixeira T, Kweder SL, Saint-Raymond A. Are the European Medicines Agency, US Food and
Drug Administration, and Other International Regulators Talking to Each Other? Clin Pharmacol Ther.
2020: 107(3): 507–513.
58. EMA, 2019e. CHMP List of questions. To be addressed by the marketing authorisation
holders for ranitidine-containing medicinal products. Referral under Article 31 of Directive 2001/83/EC.
19 September 2019. EMA/CHMP/506572/2019.
https://www.ema.europa.eu/en/documents/referral/ranitidine-article-31-referral-chmp-
Page 29 of 37 Journal of Pharmaceutical Sciences

list-questions_en.pdf. Accessed date: 20 July 2020.

59. FDA, 2018d. Update – FDA releases method for detection of both NDMA and NDEA. 11
October 2018. https://www.fda.gov/drugs/drug-safety-and-availability/fda-updates-and-
press-announcements-angiotensin-ii-receptor-blocker-arb-recalls-valsartan-
losartan#5f16904a2e431. Accessed on 18 July 2020.
60. FDA, 2018e. Update- FDA alerts patients and health care professionals to ScieGen’s Irbesartan
recall due to NDEA. 30 October 2018. https://www.fda.gov/drugs/drug-safety-and-availability/fda-
updates-and-press-announcements-angiotensin-ii-receptor-blocker-arb-recalls-valsartan-
losartan#5f16904a2e1d7. Accessed on 18 July 2020.
61. FDA, 2018f. Update – FDA alerts patients and health care professionals to Sandoz’s Losartan
Potassium and Hydrochlorothiazide recall of one lot due to NDEA. 09 November 2018.
https://www.fda.gov/drugs/drug-safety-and-availability/fda-updates-and-press-announcements-
angiotensin-ii-receptor-blocker-arb-recalls-valsartan-losartan#5f16904a2e14a. Accessed on 20 July
2020.
62. EMA, 2018c. Valsartan: Review of Impurities Extended to Other Sartan Medicines. 21/09/18.
https://www.ema.europa.eu/news/valsartan-review-impurities-extendedother-sartan-medicines.
Accessed date: 20 July 2020.
63. EMA, 2019d. Update on nitrosamine impurities: EMA continues to work to prevent
impurities in medicines. 26 April 2019, EMA/241020/2019.
64. EMA, 2019f. EMA advises companies on steps to take to avoid nitrosamines in human medicines,
26 Sep 2019 and subsequent Covid-19 extension, 04 May 2020.
https://www.ema.europa.eu/en/news/ema-advises-companies-steps-take-avoid-nitrosamines-human-
medicines. Accessed on 20 July 2020.
65. Health Canada, 2019. Questions and Answers. Request to evaluate the risk of the presence N-
nitrosamine impurities in human pharmaceutical products. 02 October 2019 and subsequent Covid-19
extension, 26 March 2020
66. Swiss Medic, 2019. Potential nitrosamine contamination: request to perform a risk evaluation. 19
November 2019 and subsequent Covid-19 extension, 03 April 2020.
https://www.swissmedic.ch/swissmedic/en/home/news/mitteilungen/aufforderung-zlinhaberinnen-
ham.html. Accessed on 11 July 2020.
67. WHO, 2019. Essential medicines and health products. Information Note Nitrosamine
Impurities. Update on nitrosamine impurities.
https://www.who.int/medicines/publications/drugalerts/InformationNote_Nitrosamine-
impurities/en/. Accessed on 17 July 2020.
68. Brennan Z. 2019. EMA Urges Pioglitazone Manufacturers to Check for Nitrosamine
Impurities. 26 April 2019. Regulatory Focus. https://www.raps.org/news-and-articles/news-
articles/2019/4/ema-urges-pioglitazone-manufacturers-to-check-for. Accessed date: 20 July
2020.
69. FDA, 2019a. FDA provides update on its ongoing investigation into ARB drug products;
reports on finding of a new nitrosamine impurity in certain lots of losartan and product
recall, 01 March 2019. https://www.fda.gov/news-events/press-announcements/fda-
provides-update-its-ongoing-investigation-arb-drug-products-reports-finding-new-
nitrosamine. Accessed on 20 July 2020.
70. Gottlieb S, Woodcock J., 2019b. FDA Statement on the agency’s list of known nitrosamine-free
valsartan and ARB class medicines, as part of agency’s ongoing efforts to resolve ongoing safety issue,
04 April, 2019. https://www.fda.gov/news-events/press-announcements/fda-statement-agencys-list-
known-nitrosamine-free-valsartan-and-arb-class-medicines-part-agencys. Accessed on 20 July 2020.xs
71. Valisure Citizen Petition on Ranitidine, September 9, 2019: https://www.valisure.com/wp-
content/uploads/Valisure-Ranitidine-FDA-Citizen-Petition-v4.12.pdf. Accessed on 04 July 2020.
72. FDA, 2019b. Liquid Chromatography-High Resolution Mass Spectrometry (LC-HRMS) Method for
the Determination of NDMA in Ranitidine Drug Substance and Drug Product. 09/13/2019.
https://www.fda.gov/media/130801/download. Accessed on 04 July 2020.
73. EMA, 2020c. Suspension of ranitidine medicines in the EU. Update of 26 June 2020.
EMA/231394/2020. Rev.1. https://www.ema.europa.eu/en/documents/referral/ranitidine-
article-31-referral-suspension-ranitidine-medicines-eu_en.pdf. Accessed date: 20 July 2020.
74. Woodcock, 2019b. Statement from Janet Woodcock, M.D., director of FDA’s Center for Drug
Evaluation and Research, on impurities found in diabetes drugs outside the U.S. December 05, 2019.
https://www.fda.gov/news-events/press-announcements/statement-janet-woodcock-md-director-
fdas-center-drug-evaluation-and-research-impurities-found. Accessed on 04 July 2020.
75. FDA, 2020b. Statement. FDA posts laboratory results for NDMA levels in metformin, 03 February
2020. https://www.fda.gov/drugs/drug-safety-and-availability/fda-updates-and-press-announcements-
ndma-metformin#5f153851dc977. Accessed on 20 July 2020.
76. Valisure, 2020. Re: Valisure Citizen Petition on Metformin, 02 March 2020.
https://www.valisure.com/wp-content/uploads/Valisure-FDA-Citizen-Petition-on-Metformin-v3.9.pdf.
Accessed on 20 July 2020.
77. Yang J, Marzan TA, Ye W, Sommers CD, et al. A cautionary tale: Quantitative LC-HRMS
analytical procedures for the analysis of M-nitrosodimethylamine in metformin. AAPS J.
2020; 22(89):1-8.
78. FDA, 2020c. FDA Alerts Patients and Health Care Professionals to Nitrosamine Impurity Findings in
Certain Metformin Extended-Release Products. https://www.fda.gov/news-
events/press-announcements/fda-alerts-patients-and-health-care-professionals-nitrosamine-impurity-
findings-certain-metformin. Accessed on 04 July 2020.
79. FDA, 2020d. FDA Updates and Press Announcements on NDMA in Zantac (ranitidine). FDA is
alerting patients and health care professionals to Amneal Pharmaceutical’s voluntary recall
of nizatidine. 16 April 2020. https://www.fda.gov/drugs/drug-safety-and-availability/fda-updates-and-
press-announcements-ndma-zantac-ranitidine. Accessed on 04 July 2020.
80. Health Canada, 2020. Subject: Risk of NDMA formation in doxylamine succinate- containing
product, 19 July 2020.
81. Nitrosamine impurities. Guidance for Marketing Authorisation Holds.
https://www.ema.europa.eu/en/human-regulatory/post-authorisation/referral-
procedures/nitrosamine-impurities#rifampicin-medicines-section. Accessed on 18 December 2020.
82. Nitrosamine impurities. Rifampicin medicines. https://www.ema.europa.eu/en/human-
regulatory/post-authorisation/referral-procedures/nitrosamine-impurities#rifampicin-medicines-
section. Accessed 18 December 2020.
83. Boon K. N-Nitrosamine Update. Regulatory Hot Topics V. Joint Pharmaceutical Symposium, London,
October 2020.
84. FDA, 2020e. FDA works to mitigate shortages of rifampin and rifapentine after manufacturers find
nitrosamine impurities. FDA not objecting to rifapentine with CPNP at or below 20 ppm remaining on
the market. Update 10/29/2020. https://www.fda.gov/drugs/drug-safety-and-availability/fda-works-
mitigate-shortages-rifampin-and-rifapentine-after-manufacturers-find-nitrosamine. Accessed on 30
October 2020.
85. Nitrosamine Impurities. Metformin containing medicines. https://www.ema.europa.eu/en/human-
regulatory/post-authorisation/referral-procedures/nitrosamine-impurities#metformin-containing-
medicines-section. Accessed 18 December 2020.
86. Nitrosamine Impurities. Rifampicin containing medicines. https://www.ema.europa.eu/en/human-
regulatory/post-authorisation/referral-procedures/nitrosamine-impurities#rifampicin-medicines-
section. Accessed on 20 December 2020.
87. Shanley A. 2018. After Valsartan Recalls, Regulators Grapple with Nitrosamine Contamination in
APIs. September 26, 2018. Pharm Tech https://www.pharmtech.com/view/after-valsartan-recalls-
regulators-grapple-nitrosamine-contamination-apis. Accessed on 11 July 2020.
88. Allmendinger T, Bixel D, Clarke A, Di Geronimo, L, et al. Carry Over of Impurities: A Detailed
Exemplification for Glycopyrrolate (NVA237). Org Proc Res Dev. 2012; 16(11): 1754–1769.
89. Burns M, Teasdale A, Elliot E, Barber C. Controlling a cohort: Use of Mirabilis™ based purge
calculations to understand Nitrosamine-related risk and control strategy options. Org Proc Res Dev. op-
2020-00264b. Submitted for publication.
90. FDA, 2019c. Laboratory analysis of ranitidine and nizatidine products. 01
November2019.https://www.fda.gov/drugs/drug-safety-and-availability/laboratory-tests-ranitidine.
Accessed on 20 July 2020.
91. FDA, 2019d. FDA updates table of interim limits for nitrosamine impurities in ARBs. Update
Page 31 of 37 Journal of Pharmaceutical Sciences

2/28/2019. https://www.fda.gov/drugs/drug-safety-and-availability/fda-updates-and-press-
announcements-angiotensin-ii-receptor-blocker-arb-recalls-valsartan-losartan#5f16904a2cdf. Accessed
on 18 July 2020.
92. FDA, 2019e. Update – FDA warns Mylan for cGMP violations, 13 November, 2019.
https://www.fda.gov/drugs/drug-safety-and-availability/fda-updates-and-press-announcements-
angiotensin-ii-receptor-blocker-arb-recalls-valsartan-losartan#5f16904a2cdfe. Accessed on 18 July
2020.
93. Lv J, Wang L, Li Y. Characterization of N-nitrosodimethylamine formation from the ozonation of
ranitidine. J Environ Sci (China). 2017; Aug; 58: 116-126. https://pubmed.ncbi.nlm.nih.gov/28774600/.
Accessed on 18 July 2020.
94. WHO, 2008. N-Nitrosodimethylamine in Drinking-water. Background document foe development of
WHO Guidelines for Drinking-water Quality.
https://www.who.int/water_sanitation_health/dwq/chemicals/ndma_2add_feb2008.pdf. Accessed on
20 December 2020.
95. WHO Guidelines for Drinking-water Quality.
https://www.who.int/water_sanitation_health/dwq/chemicals/ndma_2add_feb2008.pdf?ua=1.
Accessed on 30 October 2020.
96. Ashworth I, Dirat O, Teasdale A, Whiting M. Potential for the Formation of N-Nitrosamines During
the Manufacture of Active Pharmaceutical Ingredients: An Assessment of the Risk Posed by Trace
Nitrite in Water. Org Proc Res Dev. op-2020-00224w. Submitted for publication.
97. Wu Y, Levons J, Narang AS, Raghaven K, et al. Reactive Impurities in Excipients: Profiling,
Identification and Mitigation of Drug–Excipient Incompatibility. AAPS PharmSciTech. 2011; 12(4):
1248–1263.
98. ANSM (French National Agency for Medicines and Health Product Safety), 2018.
Determination of NDMA in valsartan active substances and finished products by HPLC/UV,
Method reference : 18A0399-02. https://www.edqm.eu/sites/default/files/omcl-method-
determination-ndma-valsartan-ansm-september2018.pdf. Accessed on 21 July 2020.
99. PALG (Public Analyst’s Laboratory Galloway, Ireland), 2018. Determination of NDMA (HS-GC-
MS), Method 3/30 Issue No. 1. https://www.edqm.eu/sites/default/files/omcl-ndma-
method-palg-ie-september2018.pdf. Accessed on 21 July 2020.
100. CVUA (CHEMISCHES UND VETERINÄRUNTERSUCHUNGSAMT KARLSRUHE), 2018. Test method for
the determination of NDMA by LC/MS/MS in Valsartan finished products.
https://www.edqm.eu/sites/default/files/omcl-method-determination-ndma-valsartan-cvua-
september2018.pdf. Accessed on 21 July 2020.
101. LDL (Bayerisches Landesamt für Gesundheit und Lebensmittelsicherheit ), 2018. LC-MS/MS
Method ´190321_NMBA_APCI_POS_waste´ (Shimadzu HPLC + Sciex QTrap 5500):
https://www.edqm.eu/sites/default/files/de_lgl_losartan_method_parameters_nmba_lcms. pdf.
Accessed on 21 July 2020.
102. OMCL, 2020. Ad-hoc projects of the OMCL Network. https://www.edqm.eu/en/ad-hoc-projects-
omcl-network. Accessed on 30 October 2020.
103. Ph. Eur., 2020. Last chance to comment: analysis of N-nitrosamine impurities.
https://www.edqm.eu/en/news/last-chance-comment-analysis-n-nitrosamine-impurities. Accessed on
21 July 2020.
104. FDA, 2018g. GC/MS Headspace Method for Detection of NDMA in Valsartan Drug Substance. 16
August 2018. https://www.fda.gov/downloads/Drugs/DrugSafety/UCM618053.pdf. Accessed on 11
July 2020.
105. FDA, 2018h. Combined direct injection N-nitrosodimethylamine (NDMA) and N-
nitrosodiethylamine (NDEA) Impurity Assay by GC/MS.
https://www.fda.gov/downloads/Drugs/DrugSafety/UCM623578.pdf. Accessed on 11 July 2020.
106. FDA, 2019f. Combined Direct Injection N-Nitrosodimethylamine (NDMA), N-
Nitrosodiethylamine (NDEA), N-Nitrosoethylisopropylamine (NEIPA), N-
Nitrosodiisopropylamine (NDIPA), and N-Nitrosodibutylamine (NDBA) Impurity Assay by GC-
MS/MS. 04/19/2019. https://www.fda.gov/media/123409/download. Assessed on 21 July
2020.
107. FDA, 2019g. Combined Headspace N-Nitrosodimethylamine (NDMA), N-Nitrosodiethylamine
(NDEA), N-Nitrosoethylisopropylamine (NEIPA), and N-Nitrosodiisopropylamine (NDIPA) Impurity Assay
by GC-MS/MS. 29/04/2020. https://www.fda.gov/media/124025/download. Accessed on 16 October
2020.
108. FDA, 2019h. Liquid Chromatography-High Resolution Mass Spectrometry (LC-HRMS) Method for
the Determination of Six Nitrosamine Impurities in ARB Drugs. 05/21/19.
https://www.fda.gov/media/125478/download. Assessed on 16 October 2020.
109. FDA, 2019i. Development and validation of a RapidFire-MS/MS method for screening of
nitrosamine carcinogen impurities N-Nitrosodimethylamine (NDMA), N-Nitrosodiethylamine (NDEA),
N-Nitrosoethylisopropylamine (NEIPA), N-Nitrosodiisopropylamine (NDIPA), N-Nitrosodibutylamine
(NDBA) and N-Nitroso-N-methyl-4-aminobutyric acid (NMBA) in ARB drugs. 07/24/2019.
https://www.fda.gov/media/125477/download. Assessed on 16 October 2020.
110. PF, 2020. General Chapter Prospectus: <1469> Nitrosamine Impurities. 24 April 2020.
https://www.uspnf.com/notices/nitrosamine-impurities-gc-prospectus-20200424. Accessed on 16
October 2020.
111. SwissMedic, 2020. Prüfvorschrift 31_PV_171 Nitrosamine mittels GC-MS/MS. Prüfvorschrift
31_PV_171 Nitrosamine mittels GC-MS/MS.
112. Health Canada, 2020. Determination of N-Nitrosodimethylamine (NDMA) and N-
Nitrosodiethylamine (NDEA) by GC-MS-MS (Direct Injection) in Sartan Finished Products and Drug
Substances. https://www.canada.ca/en/health-canada/services/drugs-health-
products/compliance-enforcement/information-health-product/drugs/angiotensin-receptor-
blocker.html#a4. Accessed on 17 October 2020.
113. Taiwanese FDA, 2020a. Determination of N-Nitrosodimethylamine and N-
Nitrosodiethylamine in Medicines. Issued Aug 10, 2018, Revised, Nov 14, 2018, Revised, Sep
27, 2019, Revised, Dec 12, 2019. https://www.fda.gov.tw/ENG/siteList.aspx?sid=10360.
9 Accessed on 17th October 2020.
114. Taiwanese FDA, 2020b. Determination of N-Nitroso-N-Methyl-4-Aminobutyric Acid
in Sartan Drug Substances and Drug Products. Issued Apr 10, 2019, Revised, May 22, 2019,
Revised, Jul 28, 2020. https://www.fda.gov.tw/ENG/siteList.aspx?sid=10360. Accessed on
17th October 2020.
115. Taiwanese FDA, 2020c. Method of Test for Nitrosamines in Medicines – Multiple Analysis. Issued.
Dec 12, 2019, Aug 24, 2020. https://www.fda.gov.tw/ENG/siteList.aspx?sid=10360. Accessed on 17th
October 2020.
116. Biba E, 2020. USP general chapter <1469> Nitrosamine Impurities.
http://kpbma.or.kr/api_board/download/seminar/454Accessed on 16 October 2020.
117. ICH Q7, 2000. Good manufacturing practice for active pharmaceutical ingredients. Current Step 4
version dated 10 November 2000.
118. ICH Q9, 2005. Quality risk management. Current Step 4 version dated 9 November 2005.
119. EC, 2012. EudraLex. The Rules Governing Medicinal Products in the European Union, Volume 4. EU
Guidelines for Good Manufacturing Practice for Medicinal Products for Human and Veterinary Use.
Chapter 7, Outsourced Activities. SANCO/AM/sl/ddg1.d.6(2012)860362.
https://ec.europa.eu/health/sites/health/files/files/eudralex/vol-4/vol4-chap7_2012-06_en.pdf.
Accessed on 20 October 2020.
120. PubChem, 2020. Hydrazoic Acid. https://pubchem.ncbi.nlm.nih.gov/compound/Hydrazoic-
acid#section=GHS-Classification. Accessed on 20 July 2020.
121. Wittenberger SJ. Recent developments in tetrazole chemistry. A review. Org Prep Proced Int.
1994; 26: 499-531
122. N, Hart RJ, Walters CL, et al. 1984. Susceptibilities of drugs to nitrosation under standardized
chemical conditions. Food Chem Toxicol 1984;22: 269–27
Page 33 of 37 Journal of Pharmaceutical Sciences

Figure legend:

Figure 1. Dose-response relationship for total liver tumors after lifelong exposure to NDMA via
drinking water (milligram per kilogram per day). Data: Peto et al. 32. Lines: fitted LVM-E4 model
for males and females, assuming parameters a and c identical but b different among sexes. The
horizontal dotted line indicates the 10% extra risk level (compared with the estimated
background response of 7.3%), the two vertical dotted lines the associated doses (BMDs) for
females (triangles) and males (circles). Extra risk is defined as additional risk divided by the
non-affected fraction in the control. Note that dose is plotted on log-scale, resulting in
sublinear curves at the low range, which are however linear on the untransformed dose scale
(Zeilmaker et al.)34

Table legends:

Table 1: Allowable Intakes (AI) for N-nitrosamines30

Table 2: Source of Amine leading to N-nitrosamine formation31

Table 3: Commentary on EMA’s recommendations on ICH Q7 from Lessons Learnt 31 after

action review

Table 4: Commentary on EMA’s recommendations on ICH Q9 from Lessons Learnt 31 after

action review

Table 5: Commentary on EMA’s recommendations on ICH M7(R1) from Lessons Learnt31 after action
review

Table 6: Commentary on EMA’s recommendations on EU’s Guideline on Chemistry of Active

Substances from Lessons Learnt31 after action review

Table 7: Commentary on EMA’s recommendations on EMA GMP guidelines from Lessons Learnt31 after
action review