New Peak is Confirmed: Now what – Potential Sources
In Part 2, we will address identifying potential sources of the new discovered peak. This step is critical to the design of the comprehensive investigation. Before jumping into the creation of an investigation plan, please ensure that the observed new peak is real and reproducible (see Part 1 of this series for details).
At a very high level, this new observed peak is either a degradant, a leachable or a contaminate. The appearance of this new peak does not automatic translate in a new degradant product related to one’s active pharmaceutical ingredient (API) or drug product.
To better understand the situation and prevalence of this new identified peak, conduct a historical review of previous data generated using the method in which the peak was observed. Determine if the peak was observed previous but overlooking during the review of the data. It is important to understand if the analysis where the new peak was identified is the only incidence where the peak was seen. If the new peak is seen in multiple data sets, determine if the peak is associate with multiple lots of API and /or product. Also examine the data for correlations with the following parameters:
Observation in multiple lots tested in the same analysis (e.g., HPLC run)
Procedure used for sample preparation
Instrument and/or column used
Defined interval of time
Lot number of component or reagent used in testing
Analyst performing test
These correlations can provide direction for the investigation, such as, suggesting that source is related to the performance of analytical test and not the actual product. As a result, the investigation may focus on analytical testing first. Note: A word of caution here; complete the next step below before beginning any laboratory work!
The source of this new peak can be total unrelated to the API or the drug product. Before developing any investigation plan, it is best to brainstorm a comprehensive list of all possible potential sources that could possibly contain or generate this peak. When building the list consider the following area:
Measurement/Test Method
o Procedure
o Equipment/Analytical Column
o Glassware
o Solvents/Reagents/Consumables
o Preparation/Aged of Preparation
o Testing Site
Manufacturing Process
o Formulation/compounding steps
o Filling
o Packaging
o In-Process Testing
o Layout/Rooms Used
o Gowning Practices
Equipment
o Formulation Tanks/Mixers
o Balances
o Transfer Lines
o Filler
o Disposable components: Filters/Tubing/etc.
o Storage
o Cleaning Logs/Cleaning Practices
Raw materials
o Excipients
o Solvents
o Packaging Components
o Vendors/Vendor qualification
o Sampling Practices
Environment
o Laboratory
o Manufacturing Suite
o Warehouse
o Cleanliness
o Monitoring Practices
o Temperature/Humidity
People
o Analysts
o Operators
o Warehouse staff
o Training
o Hygiene
o Management priorities
Be thorough as the ultimate source may surprise you. The above list is not intended to be all inclusive, but a starting point for the brain-storming session. A fishbone diagram (basic template shown below) is a convenient total for laying out brainstormed list of potential causes, allowing the relationships to be examined. Often the fishbone diagram will reveal independences of certain steps in the overall process across different main ‘bones’ on one’s diagram. For example, for the method to function as expected, the instrument must perform properly with the correct mobile. For the instrument to perform properly, it must be cleaned, maintained and calibrated to the vendor provided specifications.
Below are a couple of case studies to illustrate the point:
Case Study 1: An unknown peak at a relative retention time of 1.25 (RRT=1.25) was identified during the analysis of the related compound for the drug product. A review of current and historical data showed the presence of the peak in multiple drug product lots in the same analysis and in different analyses using the same method. The amount observed was independent of drug product age and storage temperature. It was also discovered that unknown peak at RRT = 1.25 was observed determination of assay but ignored as the assay method focused on the quantitation of parent compound only. The peak at RRT = 1.25 was absence in the diluent and standard injections. It should be noted that the assay and related compound methods shared a common sample preparation for the drug product. Furthermore, the standards and drug product samples were prepared differently. The preparation of the drug product samples contained a centrifugation and filtration step. The method diluent was non-aqueous and consisted of entirely of an alcohol. The testing laboratory was convinced that the source of the peak was related to the drug product. The testing laboratory had the story partially correct; the drug product was a factor but the unknown peak at RRT=1.25 was not related to the actual drug product formulation. It turns out the unknown peak is an analytical artifact derived from the sample preparation step for the drug product. The source of the peak was the syringe used to filter the supernatant from the centrifugation step. The syringe contained a black rubber tip on the plunger. The unknown peak at RRT=1.25 was extracted from the black rubber by the method diluent. By changing the syringe used to all-plastic syringe (no rubber), the unknown peak at RRT=1.25 was eliminated.
Case Study 2: Again, an unknown peak was observed early in the chromatogram (RRT = ~0.45) during the analysis of related compounds in the drug product. The unknown peak eluted very close to the solvent front. A historical review of existing data reveal that this peak had been seen in previous drug product samples analyzed for both assay and related compounds over a specific period of time. The unknown peak (RRT =0.45) was observed at various levels in release and stability samples. The amount of the unknown peak was independent of time and often was not observed at accelerated storage conditions. Although present in the release samples, the unknown peak at RRT=0.45 was not observed in any of the in-process bulk assay samples which were analyzed by the very same method. The peak was not in the bulk formulation but was introduced to the drug product during the filling process. Thus, the potential source of the peak was narrow down to the filling line or the packaging components. After an extensive investigation, unknown peak was isolated form the bottle, and identified to be a naturally occurring flavonoid which is commonly used in food to impart a cinnamon flavor. For a period of months, the bottles were contaminated with this naturally occurring flavonoid. The ultimate source was cross-contamination from a flavoring agent being stored the same warehouse as the bottles.
In closing, be open-minded when evaluating potential sources for pesky unknown peaks as the source is often not what it was thought to be. Next time, we will review investigation strategies and the process used for identification of unknown peaks.
Written By: Bette Monnot-Chase
Comments