The European Medicines Agency’s Q&A on nitrosamine impurities is no longer a single document. Since the July 2023 revision that amended Question 10, the guidance has grown into a main Q&A anchored by three appendices, and each appendix is refreshed on its own cadence. The main document sits at Revision 23, last updated 10 October 2025. Appendix 1, the running list of established acceptable intakes, was last refreshed on 24 June 2026 with 13 additions and 7 revisions. Appendix 2 fixes the carcinogenic potency categorisation approach (CPCA). Appendix 3 fixes the enhanced Ames test (EAT).

For analytical labs, the structural change matters more than any single number. The Q&A is now a living framework, and the framework has moved from “measure and control” toward “categorise, then measure and control at the limit that the category assigns.” The measurement stakes that the framework produces - single-digit and sub-ppb intake limits for many N-nitrosamines - are what push method development squarely into LC-MS/MS, and increasingly into GC-MS/MS with high-resolution options where LC ionisation is unfavourable.

What the appendix structure now enforces

Appendix 2 puts every nitrosamine into a potency category based on structural features, and each category maps to a default acceptable intake (AI). This replaces the older habit of deriving AIs from a small pool of surrogates. In practice, most novel nitrosamine drug substance-related impurities (NDSRIs) can now be assigned an AI at desk-study speed instead of waiting for compound-specific carcinogenicity data.

Appendix 3 gives a route out of over-conservative CPCA categorisation. The enhanced Ames test protocol - refined with FDA National Center for Toxicological Research input - uses specific tester strains, higher metabolic activation, and a top-dose regime chosen to detect the weakly mutagenic behaviour typical of N-nitrosamines. A negative EAT result supports a higher AI than the CPCA default; a positive one confirms the default or drives it lower.

Appendix 1 is where the categorisation lands as a number an analyst can use. The 24 June 2026 update added N-nitroso variants of brensocatib, dabigatran etexilate amide and remibrutinib, among others. Seven substances - including N-nitroso-folic acid, N-nitroso-labetalol and N-nitroso-meglumine - had their AIs revised. Earlier refreshes had already re-evaluated NDIPA, NEIPA and NMBA upward under CPCA against the previous defaults. Reading Appendix 1 quarterly is now a compliance activity, not a curiosity.

Where this leaves HPLC and MS methods

The Q&A itself does not prescribe an instrumental approach. It defines the target, and the target sets the method. For high-dose products, acceptable intakes in the tens of ng/day translate into concentration limits in the low-ppb range in the drug substance, with several substances falling below 33 ppb and specific NDSRIs sitting in the 26.5 ng/day band or lower. Those numbers rule most single-quadrupole and diode-array approaches out of the game and place the burden on tandem mass spectrometry.

For pharmaceutical QC laboratories, three method families are doing the practical work:

  • LC-MS/MS with electrospray ionisation, running in multiple-reaction-monitoring mode, is the default for polar N-nitrosamines and for NDSRIs that share the drug substance’s ionisation behaviour. Recent published methods report limits of quantification in the ppb and sub-ppb range with recoveries between 95 and 105 percent and relative standard deviations under 2 percent.
  • LC-MS/MS with atmospheric-pressure chemical ionisation (APCI) is used for the sartan-related nitrosamines - NDMA, NDEA, NDIPA, NEIPA - where ESI response can be poor or matrix-suppressed. APCI is the historical default for the valsartan/losartan/irbesartan combinations that triggered the original 2018 recalls.
  • GC-MS and GC-MS/MS remain the reference for volatile nitrosamines (NDMA, NDEA) when the matrix suits headspace or direct injection. For most tablet and solution matrices, LC has displaced GC, but the technique is retained where GC gives a cleaner background than LC after solid-phase extraction.

None of this is new science. What has changed is the population of analytes the method must cover. Where a validated method three years ago targeted five or six named nitrosamines, a 2026 method is expected to be extendable to any NDSRI that Appendix 1 or the manufacturer’s own risk assessment throws at it. Method development documentation now leans heavily on the ICH Q14 analytical-procedure lifecycle framework, which formalises the extension logic and the change-control pathway.

What this means for method transfer and continued verification

Method transfer between sites was already the harder half of nitrosamine work. Sub-ppb performance is instrument-specific: two triple quadrupoles from the same vendor can differ by a factor of two in signal-to-noise on the same standard. Under the current EMA framework, marketing authorisation holders are expected to demonstrate not only that the method meets its own validation acceptance criteria at each site, but that its limits are appropriate against the current Appendix 1 AI - which may have moved since the method was validated.

Multi-site rollouts are now planning around this in two ways. Some are locking the method to a nominal AI 20-30 percent below the current Appendix 1 value, on the assumption that CPCA re-evaluation is more likely to tighten than loosen limits. Others are treating Appendix 1 as an input to a periodic method-suitability review, tied to the site’s product quality review cycle. The lifecycle question - how to keep a validated method aligned with a moving regulatory target - is the same one that shows up for chemometric models under the comparative FDA/EMA/PMDA framework.

The European Pharmacopoeia’s parallel move

The European Pharmacopoeia Commission has been trimming N-nitrosamine sections from individual active-substance monographs, reasoning that monograph-level specification duplicates - and can conflict with - the Q&A framework. Monographs increasingly limit themselves to process-related and degradation impurities; nitrosamine control lives in the marketing authorisation dossier, tested to the AI Appendix 1 assigns.

The practical consequence for QC laboratories is that a pharmacopoeial monograph is no longer a sufficient anchor for a nitrosamine method. The anchor is the Q&A and its appendices, refreshed on the EMA’s cadence, and the method’s validation package must reference the current AI at each site release.

Reading the guidance today

For a laboratory building or revising a nitrosamine control strategy in the second half of 2026, the practical reading order is: Appendix 2 to understand what CPCA does to a novel structure; Appendix 3 to understand when to invest in an enhanced Ames study; and Appendix 1 to find the number the method has to reach. The main Q&A defines the reporting and change-control expectations that surround the appendices. All four documents are updated separately, and none of them is finished.