An analytical quality by design-guided method development for the quantification of N-nitroso-fluoxetine to support mitigation strategies for a fluoxetine drug product

Abstract

This work focused on N-nitroso-fluoxetine, a N-nitrosamine drug substance-related impurity (NDSRI) formed in fluoxetine-containing products via nitrosation pathways. N-nitroso-fluoxetine was previously detected above the established safety threshold of 100 ng/day, leading to temporary withdrawal of a marketed product in Belgium. An Analytical Quality by Design (AQbD) approach supported the development of a reversed-phase LC-(ESI+)-MS/MS method. Design of experiments identified aqueous mobile phase pH and gradient time as critical method parameters affecting detection or quantification limits and chromatographic separation performance, respectively.

Adequate separation was achieved using Acquity® HSS T3 guard-column and column with a run time of 22 min. The method was validated according to ICH Q2(R2) guideline using a combined approach based on total error. It enables accurate and precise quantification of N-nitroso-fluoxetine from 0.09 to 4.75 ppm (injected concentration in ng/mL) while simultaneously screening for other small-molecule N-nitrosamines in both drug substance and finished products. The validated method was applied to assess risk mitigation measures implemented by the manufacturer, including the use of low-nitrite excipients, desiccant-integrated aluminum/aluminum blister packaging, and reduced shelf-life.

In the updated formulation, N-nitroso-fluoxetine contents were 0.58 ppm at release and increased to 1.15 ppm after 21 months of storage at 25 °C/60% RH, remaining below the safety threshold of 1.67 ppm calculated according to the EMA “Call for Review” procedure. Together, these modifications have effectively controlled N-nitroso-fluoxetine formation. Therefore, the present study demonstrates a robust AQbD-guided LC-MS/MS method, contributes to NDSRI analysis, and supports science- and risk-based control strategies for these emerging impurities.

Introduction

In medicinal products, N-nitrosamine impurities can be broadly classified into three categories: (i) generic small-molecule N-nitrosamines, which may originate from multiple sources across different drug products; (ii) N-nitrosamine drug substance-related impurities (NDSRIs), which are specific to a given active pharmaceutical ingredient (API) and its related substances; and (iii) N-nitrosamine drug linker-related impurities (NDLRIs), which are encountered in antibody drug conjugates. Since the emergence of the so-called N-nitrosamine crisis in the pharmaceutical industry, NDSRIs have continued to be identified. The first reported case, N-nitroso-varenicline, resulted in voluntary product recalls in 2021 [1]. According to data updated by the European Medicines Agency (EMA) in March 2026, NDSRIs now account for approximately 92% of the N-nitrosamine impurities identified in pharmaceutical products [2]. This observation highlights the progressive shift of regulatory and scientific attention toward NDSRIs as the crisis has evolved.

In line with this trend, the present study investigates a case of an NDSRI detected on the Belgian market: N-nitroso-fluoxetine, which is specifically associated with fluoxetine-containing formulations. Fluoxetine is a widely prescribed antidepressant. Several fluoxetine-based products have been recalled worldwide following the detection of N-nitroso-fluoxetine at levels exceeding the recommended acceptable intake limit [3], [4], [5].

Although NDSRIs generally lack carcinogenicity data and their acceptable intake limits are therefore typically assigned using the carcinogenic potency categorization approach [6], [7], the acceptable intake for N-nitroso-fluoxetine was established at 100 ng/day based on a structure-activity relationship/read-across approach, with the TD₅₀ of 4-(methylnitrosamino)−1-(3-pyridyl)−1-butanone (NNK) as the point of departure, given its positive results in in vivo mutagenicity studies [8], [9]. More recently, due to the limited carcinogenicity data available for N-nitroso-fluoxetine, Vogel et al. investigated its mutagenic potential. Their study demonstrated that, following metabolic activation by cytochrome P450 enzymes (CYP2B6, CYP2C19, and CYP3A4), N-nitroso-fluoxetine generates higher levels of DNA methylation adducts and induces more DNA strand breaks than N-nitrosodimethylamine (NDMA), a well-characterized small-molecule N-nitrosamine [10]. In light of the associated mutagenic risk, the Federal Agency for Medicines and Health Products in Belgium mandated marketing authorization holders to assess the levels of N-nitroso-fluoxetine in fluoxetine-containing products. After N-nitroso-fluoxetine amounts exceeding established safety thresholds were detected, Fluoxone® Divule® 20 mg was temporarily withdrawn from the Belgian market in 2023 [11], [12], [13].

The principles of quality by design (QbD) were first introduced to the pharmaceutical industry by the United States Food and Drug Administration (FDA) [14]. QbD is defined in the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Q8(R2) as a systematic approach to development that begins with predefined objectives, emphasizing product and process understanding and process control [15]. In combination with ICH Q9(R1) guideline, QbD strengthens science- and risk-based regulatory decision-making [15], [16].

Unlike the traditional quality by testing paradigm, in which product quality is assessed primarily through end-product testing, QbD proposes that quality should be built into the process by design [15]. This philosophy promotes enhanced understanding and control of manufacturing processes, thereby reducing the risk of out-of-specification products and improving overall product quality. Consequently, QbD not only minimizes financial risks for manufacturers by limiting batch rejections, but also safeguards patient health by ensuring the consistent production of high-quality medicines.

The concept of analytical quality by design (AQbD) extends QbD principles to analytical procedure development, as outlined in ICH Q14 guideline. AQbD shares the advantages of QbD while providing a structured alternative to minimal method development strategies. Method development phases are often time-consuming due to extensive trial-and-error experimentation. By integrating prior knowledge, risk assessment, design of experiments (DoE), and a lifecycle-oriented continuous improvement mindset, AQbD enhances analytical robustness and regulatory flexibility. Specifically, AQbD incorporates systematic knowledge and risk management to achieve two primary objectives: (i) improving method robustness to reduce the risk of inadequate analytical performance and erroneous results; and (ii) enabling continuous method improvement throughout the product lifecycle [17].

Compared with development approaches based on one-factor-at-a-time (OFAT) experimentation, AQbD represents a more scientific and risk-based strategy that relies on DoE. The OFAT approach varies a single factor while holding all others constant, limiting the ability to identify factor interactions and often requiring a large number of experiments. This limitation may compromise method robustness and increase the risk of analytical failure. In contrast, DoE evaluates multiple factors simultaneously, allowing the assessment of main effects, interactions, and quadratic effects while reducing experimental effort, time, and cost [18].
To date, most published studies have focused on the determination of small-molecule N-nitrosamines [19], [20]. Although NDSRIs have attracted increasing regulatory and scientific attention, relatively few analytical methods have been reported for their determination [21], [22]. The Taiwan FDA and the German official medicines control laboratory (OMCL) have published liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods for the quantification of N-nitroso-fluoxetine, achieving quantification limits of approximately 0.1 ppm in the drug substance (equivalent to 1 ng/mL) and in immediate-release tablets (equivalent to 0.1 ng/mL), respectively [23], [24].

The acceptable limit for N-nitroso-fluoxetine in the studied product is obtained by dividing its acceptable intake limit (100 ng/day) by the maximum daily dose of fluoxetine (60 mg/day), yielding a specification of 1.67 ppm. Applying the AQbD framework, this work developed a robust LC-MS/MS method capable of quantifying N-nitroso-fluoxetine down to 0.1 ppm (or ng/mL) while simultaneously screening for other small-molecule N-nitrosamines in fluoxetine drug substance and drug products. The study provides a valuable contribution to NDSRI determination and supports robust, science-based control strategies for these emerging impurities.

Download the full article as PDF here An analytical quality by design-guided method development for the quantification of N-nitroso-fluoxetine to support mitigation strategies for a fluoxetine drug product

or continue reading here

Chemicals and Reagents

Certified reference stock solutions (500 µg/mL) of NDMA, N-nitrosodi-n-propylamine (NDPA), N-nitrosodiisopropylamine (NDIPA), N-nitrosoethylisopropylamine (NEIPA), N-nitrosodi-n-butylamine (NDBA), N-nitroso-diethylamine (NDEA), and N-nitroso-N-methyl−4-aminobutyric acid (NMBA) were obtained from the European Directorate for the Quality of Medicines and HealthCare (Strasbourg, France). A certified reference stock solution (1 mg/mL) of N-nitroso−1,2,3,6-tetrahydropyridine (NTHP) was supplied by the United States Pharmacopeia (Rockville, MD, USA). Reference stock solutions (1 mg/mL) of NNK and N-nitrosomethylphenylamine (NMPA) were purchased from Sigma-Aldrich/Merck (St. Louis, MO, USA). Reference stock solutions (1 mg/mL) of N-nitrosopyrrolidine (NPYR), and N-nitroso-piperidine (NPIP) were purchased from Cambridge Isotopes Laboratories (Tewksbury, MA, USA). Reference standards of N-nitrosomorpholine (NMOR), N-nitrosodiphenylamine (NDPhA), N-nitroso-diethanolamine (NDELA), and fluoxetine hydrochloride were obtained from TCI (Tokyo, Japan). N-nitroso-fluoxetine was purchased from Merck (Kyiv, Ukraine). Deuterated N-nitroso-fluoxetine-d5 (stable isotope-labeled standard) was obtained from Toronto Research Chemicals (Vaughan, Canada).

UPLC-MS grade methanol, water, formic acid, and ammonium acetate were purchased from Biosolve (Dieuze, France). A 25% ammonia solution was obtained from VWR (Rosny-sous-Bois, France). Fluoxetine hydrochloride drug substance, excipients (Table S1), and Fluoxone® Divule® 20 mg used in the work were supplied by a pharmaceutical industry partner.

Yue Zhang, Pierre-Yves Sacré, Bruno Streel, Eric Ziemons, Philippe Hubert, Cédric Hubert, An analytical quality by design-guided method development for the quantification of N-nitroso-fluoxetine to support mitigation strategies for a fluoxetine drug product, Journal of Pharmaceutical and Biomedical Analysis, Volume 282, 2027, 117692, ISSN 0731-7085, https://doi.org/10.1016/j.jpba.2026.117692.

Read also our introduction article on Quality by Design (QbD) here: