The Dose Makes the Poison!

Paracelsus highlighted that “the dose makes the poison,” which is the fundamental tenant of modern toxicology. That is to say, toxic substances are harmless at very small doses, and conversely a relatively harmless substance can be deadly if over-consumed. Building on this tenet, Fritz Haber stated that concentration and time are important variables in the toxicity of many materials. Thus, the incidence and/or severity of toxicity is dependent on the total exposure, which is to say, exposure concentration (c) × duration time (t) of exposure (c × t).

Haber’s Law, with appropriate caveats (1), is often used in defining exposure guidance for toxic substances, but it results in absolute exposure limits (≤ x µg/day), rather than the more familiar relative exposure limits (≤ 0.2 percent). Haber’s law is applied in the setting of safe limits for less than lifetime (LTL) exposure of mutagenic impurities (MI) in pharmaceuticals (ICH M7) (2). The LTL limits have a wide range – 120 µg/day (≤30 days), 20 µg/day (≥1-12 months), 10 µg/day (≥1-10 years), 1.5 µg/day (≥10-lifetime) – and are applicable in both clinical development and for marketed products.

From an analytical chemistry perspective, MIs are often challenging to detect and control. Due to their reactive nature, they generally have poor stability in the analytical matrix. Very high sensitivities are typically required (often ppm or even ppt levels). As a class, MIs have very diverse structural features and physicochemical properties, which often preclude a generic analytical strategy. Selectivity is typically challenging as high levels of the active pharmaceutical ingredient (API) will be present, along with the MI(s), which can cause significant matrix interference. The issue is magnified in drug products, where, in addition to the API, there will be excipients that can also potentially interfere. Lastly, there can be interference from the impurities that are present in both API and excipients. Notably, the synthetic route for the API is often evolving and knowledge of likely impurity profiles are limited during early development. The MI methods need to be rapidly developed and aligned with aggressive development timelines (3).

Method selection is often based on the volatility of the MI and demands a reasonably big toolbox. For volatile but thermally stable MIs, gas chromatography (GC) is favored, with direct injection, headspace or derivatization. Detection is typically via mass spectrometry – either using single ion monitoring mode (SIM) or MS/MS selective reaction monitoring mode (SRM). For non-volatile analytes, high performance liquid chromatography (HPLC) is used – once again typically coupled with MS (SIM) or MS/MS (SRM) detection. There is a role for HPLC-UV detection when MIs are controlled via in-process tests. And in addition to standard reverse phase HPLC, hydrophilic interaction liquid chromatography (HILIC) for polar alkylating agents and ion chromatography for a limited sub-set of MIs (for example, hydrazine) have found favor (3).  MI methods, during early development, are typically not fully validated (as per International Council for Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use guidelines, ICH Q2(R1)) (4), but are used as limit tests, with validation involving specificity, accuracy/recovery and LOD.

ICH M7 defines four different control strategies that are applicable to MIs that are present in APIs. The first approach is control of the MIs on the API specification (option 1). This test can be included on the specification as a limit test, but is generally not favored by the pharmaceutical industry as it involves transfer of sophisticated, sensitive HPLC-MS or HPLC-MS/MS methods into production. The MI impurity can also be specified as a control test in a registered starting material, reagent, and so on (option 2), and is used for materials that are incorporated near the completion of the synthetic pathway, where purging arguments may not be justifiable. The MI can also be included as an in-process control (option 3). Here, there is knowledge of the downstream purging ability of the synthetic process (often supplemented by spike and purge experiments), that allow upstream controls, often at much higher levels (for example, 0.5 percent).

Finally, theoretical, non-empirical arguments can be made that the MI is so reactive and is introduced sufficiently far enough upstream from the final API that there is no likelihood of it carrying over into the API (option 4). So, in conclusion, the introduction of ICH M7 for the control of MIs, although challenging from an analytical chemistry perspective, has been well received by the pharmaceutical industry and contributes to our overall efforts to ensure patient safety.