Low-Level Impurities Move to the Forefront of Industry Concerns

AAPS, San Antonio (Oct. 30)-Although the analysis of impurities in drug substances and drug products is not a new topic, new regulations are increasing concern in the industry. Ganapathy Mohan, PhD, associate vice-president of global analytical services department of Sanofi-Aventis (Malvern, PA, www.sanofi-aventis.us) pointed out at the 20th Annual Meeting of the American Association of Pharmaceutical Scientists that sophisticated technologies have made it possible to detect genotoxic impurities (GTIs), which can cause direct DNA damage and so require higher safety margins and lower limits than most other types of impurities. Moreover, the CMC implications of low-level impurities have raised questions about how specifications and acceptance criteria should be set as well as the regulatory impact of low-level impurities.

During the symposium titled “Low-Level Impurities and Their Impact on Pharmaceutical Development," Mohan noted, “The synthesis of pharmaceuticals involves reactive materials that lead to the production of low-level impurities that may have toxicological profiles than the active pharmaceutical ingredient. These reactions may produce genotoxic impurities or the materials themselves may be genotoxic.”

Although ICH Q3 describes how to set specifications for impurities, the argument over setting realistic limits for low-level impurities continues. “The ICH guidelines are insufficient for dealing with genotoxic impurities,” he said. In June 2006, the European Medicines Agency (EMEA) developed the first regulatory guideline on genotoxic impurities for new actives and will become effective in Jan. 1, 2007. In addition to providing key definitions, the EMEA guideline uses the threshold of toxicological concern (TTC) approach, a concept first introduced by the food industry, which is currently at approximately 1.5 micrograms per person per day or higher in the case of life-threatening diseases or short-term exposure. A better strategy for handling genotoxic impurities, says Mohan, is a “staged” TTC approach, a concept presented in a Pharmaceutical Research and Manufacturers' Association (PhRMA) position paper in February 2006 that takes into account the dose and the dosing frequency.

According to Mohan, analysts can detect low-level impurities better than before because the technologies have evolved. At first, calorimetric, titrimetric, and wet chemistry techniques were used, followed by chromatographic and spectroscopic techniques, for example, high-performance liquid chromatography (LC), mass spectrometry (MS), LC-MS, and CE. Analysts now use near infrared, atomic absorption, ICP, and X-ray fluorescence. With faster techniques, analysts can achieve “superior” detection limits from percentage values to parts per million, parts per billion, and even parts per trillion levels, says Mohan.

There are CMC challenges, however. Some of the factors of concern involve chemical development (chemistry), pharmaceutical development (formulation), and analytical sciences (developing methods that are capable of detecting low level impurities and how to validate these methods). “Synthetic processes, the control of impurities, and analytical methods evolve as compounds progress through clinical development. To address these concerns, we need a good grasp of process chemistry to proactively assess low-level impurities,” he said. From a formulation perspective, analysts need to determine the impact of timelines for early formulation development. In selecting the dosage form, evaluating low-level impurities and their impact on quality is critical. And in determining quality attributes, companies must figure out when to set specifications and at what stage in the development process.

One analytical challenge is determining how far back in the synthesis one should look for impurities and deciding whether a simple HPLC analysis sufficient for analytical assay development or if a more sophisticated technology such as LC–MS is necessary. Companies also must determine whether it is feasible to ask for validated, sufficiently sensitive analytical methods for all potential low-level impurities and at which stage of development should they be available. He warns, however, that tight specifications at early stages of development can have “severe impacts on many aspects of development.” Finally, companies should determine whether genotoxic impurities should be eliminated when possible, even when they fall below the TTC level. The answer depends on the cost to eliminate and at what level and on the indication (e.g., oncology).

Concluding his presentation, Mohan pointed out, “The risk of genotoxic impurities to patients is probably small in early development when exposed populations are small and the duration of exposure is short. Also, companies need to take a balanced approach when setting limits for potential low-level impurities that takes into account the benefits, the risk to patients, and the development phase.”

GTI detection strategies

Sandeep Modi, PhD, director of CMC Documentation and Pharmaceutical Development Strategic Operations at the Pharmaceutical Research Institute of Bristol-Myers Squibb Company (BMS, New Brunswick, NJ, www.bms.com) continued the discussion GTIs by providing actual case studies and approaches taken at BMS. First, although there are no FDA guidelines for industry, manufacturers can turn to several resources when dealing with GTIs. Specifically, ICH Q3 (R1) provides a table of impurities in new drug substances. “However, ICH has not provided specific limits” for GTIs, said Modi. The EMEA guidance discusses the limit as “the level below which you expect will not pose a significant carcinogenic or toxicological risk,” and that the limit is based “on lifelong exposure” to the drug, reiterating the concept of a staged TTC approach proposed by PhRMA.

To identify potential GTIs in active pharmaceutical ingredients (APIs), researchers should be aware of the structure of the API, noting that there are several “alerting structures” such as aromatic and alkyl groups. The problem, notes Modi, is that these structures are common. Ames testing might provide a better indication: Modi referred to the work of Dugger et al. (2005) and Carey et al. (2005), who point out that approximately 20–25% of all intermediates can be expected to test Ames positive.

GTI detection also may affect current analytical methods, notes Modi, because GTIs must be controlled at levels significantly lower that those typically reported using HPLC.

Modi presented two cases studies observed at BMS dealing with impurities. In the first, researchers found two impurities in a drug substance during the pilot stage. They then conducted spiking experiments with the objective of determining how to set the acceptance criteria. Their strategies involved two steps: evaluation of vendor-sourced key intermediates and impurity tracking from key intermediates to the API, including using GC–MS and LC–MS techniques rather than GC–FID and LC–UV to achieve early impurity detection and evaluate the fate and tolerance of the impurities.

The second case study involved nine compounds that led to the drug substance. Of these, three compounds had at least 10 detected impurities and tested Ames positive. After further evaluation, BMS researchers conducted HPLC identification tests on one compound (identified as BMS-152), using several analytical methods, including LC–MS, LC–UV, and LC–MS, for which they developed a set of process control strategies.

Concluding his presentation, Modi suggested that the industry consult with FDA, take into account the disease severity (e.g., life-threatening) and therapy, and the duration and exposure of treatment. In the meantime, he pointed out that “If it’s Ames positive, it’s a GTI,” noting as well that 1.5 microgram per day is a nominal limit.

Regulatory perspective on low-level impurities

Until there is regulatory guidance for low-level impurities, the industry can turn to other documents pertaining to pharmaceutical impurities in general. Thirunellai G. Venkateshwaran, PhD, associate director of global regulatory affairs, CMC, at Wyeth Research (Collegeville, PA, www.wyeth.com) reviewed the acceptance limits presented in the ICH guidances Q3A(R), which provides thresholds for drug substance impurities, and ICH Q3B(R), which provides thresholds for drug product impurities. Venkateshwaran also reviewed the three impurity classifications (organic, inorganic, and residual solvents). In addition, ICH Q3C presents the three classification of residual solvents (Class I–III). “Knowing the origin of these impurities facilitates their control and accelerates development,” he said. “Profiling impurities and degradation products throughout the product development cycle provides information that can be used to establish acceptance criteria at the time of filing. And, knowledge of the degradation pathways, product development process, and stability testing, should be used to characterize degradation product pathways.”

Venkateshwaran concluded his presentation by providing clinically relevant specifications and applications to impurities, including a discussion of design space, control space, and control strategies.

Follow-up questions from the attendees covered several key points, including the question of whether it is reasonable to base GTI detection solely on the Ames test and whether an in vivo test should be used as well. The panel pointed out that EMEA says that the Ames 2 in vitro test is a “driver” for GTI identification and researchers “should not stop the compound solely as a result of a Ames positive test. But this test can serve to indicate further monitoring is needed as the development moves forward. The decision will then be based on additional data, which may include in vivo tests with pharmacokinetic results.”

Another attendee pointed out that the problem of determining whether the level of impurity will be of safety concern during scale up is much more complex in the case of an unknown impurity because the relative response factor (which may not be one-to-one) also is unknown. The panel agreed and pointed out that one common approach considers impurities concentrations less than 0.1% as “low-level” impurities. Levels lower than 0.03% will not usually arouse concerns. Once the level reaches around 0.05%, however, there may be a problem when it is time to scale up the process.