HPLC

A number of models have been proposed for the prediction of various types of chromatographic retention times based on chemical structures. These models are typically limited in their applicability, either in terms of the molecules for which accurate retention times can be performed, in terms of the type of chromatographic method, or both. Recently, a system for structure-based prediction of retention time for generic chromatographic methods was devised. It uses the concept of a "federation of local models" to give accurate prediction for diverse structures, even for gradient methods. This paper will describe further accuracy improvements through the incorporation of Abraham's parameter prediction. The technique involves automated detection of when these parameters are applicable, and incorporates them accordingly. A diverse group of structures were studied under reversed-phase and HILIC methods. The accuracy of prediction will be compared, both with and without the use of this technology.

Quality by Design (QbD) has become popular within the pharmaceutical industry and its application to process analytical technology (PAT) and drug manufacturing is rapidly increasing. At the heart of QbD lies several principles, all leading toward the goal of building in quality right from the earliest stages of drug discovery, including knowledge retention, elimination of errors, and increased experimental scope. To be successful, QbD must be applied at every stage in the development and manufacturing process; however, one can consider processes within drug discovery on an individual basis to apply QbD principles. One such process is the development of chromatographic methods for impurities and degradant studies. Ensuring both robustness and optimization from an efficiency standpoint is time-consuming and difficult. Generally method development is carried out using a trial-and-error approach, and requires a large amount of manual data interpretation. However, QbD principles can be directly applied to this process, resulting in better, more robust separations. Furthermore, quality is achieved when QbD is applied to the stability study process by producing well developed, traceable, error free results, and finally into the drug product by bringing novel products to market faster, and more cost effectively.

The impurity profile of a drug candidate continuously changes during the development process. New impurities are introduced by (minor) changes in the synthetic route or formulation design. Some of them may have toxic or other undesirable properties, jeopardizing the safety and integrity of the drug candidate. Since chromatographic methods may fail to detect or resolve new impurities, the development and use of 'orthogonal' methods has become standard practice to minimize the risk of non-detected or co-eluting analytes. Peak tracking and data handling, however, may be labour-intensive, and become a serious bottleneck, if many samples, impurities and orthogonal methods are involved.

This paper will present a new HPLC-based orthogonal screening system which maximizes the power of detection and resolution while minimizing the time needed for peak tracking and data review. The impurity profile obtained with a given chromatographic method is compared with the chemo-metrically-extracted profile obtained with an orthogonal method, and any discrepancies are automatically highlighted for the user. Within just a few seconds, the chromatographer can find out if new impurities are present in a sample. This presentation will discuss the various orthogonal methods employed, and highlight the strengths and limitations of the chemo-metric detection system with real-life samples and data.