Analytical methods performance attributes

For any new excipients, APIs or drug products (where new does not necessarily mean novel, but new to the receiving site) there are additional testing criteria, e.g. supplier audits, third-party contract laboratory audits, analytical method transfers, sample management/tracking, etc. For those key excipients, where there is on-site historical experience, it still behoves both parties to check whether the local grade/supplier used by the CMO is equivalent to that used by the supplier (Worsham, 2010). There are many examples of differences in excipient physical properties, e.g. particle size, which have been attributed to different excipient sources that could ultimately impact on the performance of those excipients in formulated products (Frattini and Simioni, 1984 Dansereau and Peck, 1987 Phadke et al., 1994 Lin and Peck, 1994). 

Once the analytical method is validated for accuracy at the laboratory scale, it can be used to obtain extensive information on process performance (blend homogeneity, granulation particle size distribution, and moisture content) under various conditions (blender speed, mixing time, drying air temperature, humidity, volume, etc.). Statistical models can then be used to relate the observable variables to other performance attributes (e.g., tablet hardness, content uniformity, and dissolution) in order to determine ranges of measured values that are predictive of acceptable performance. 

An alternative interesting classification approach has been proposed within the Codex Committee for Residues of Veterinary Drugs in Foods (20). In this approach, methods are classified according to their performance attributes. This alternative approach defines methods by the level of analytical detail or information provided concerning the amount and nature of the analyte of interest, and identifies three levels. 

Performance characteristic A functional quality that can be attributed to an analytical method, for example, specificity, accuracy, trueness, precision, repeatability, reproducibility, recovery, detection capability, and ruggedness. 

Not needed for nonspecific analytical methods (e.g., total organic caibon). All of the compound detected must be attributed to the target analyte unless rationale supporting an alternative has been documented, "in cases where reproducibility has been performed, intermediate precision is not needed. 

The evaluation study will determine the attributes (bias, precision, specificity, limits of detection) of the immunoassay. Bias testing (systematic error) will be conducted by measuring recoveries of the analyte added to matrices of interest. Replicate analysis will be performed on blind replicates or split levels (e.g., Youden pairs). A minimum number of replicates will be performed to provide statistically meaningful results. The number of replicates will be determined by the intended purpose of the immunoassay as well as the documented method performance of the comparative method. 

A representative sample should be taken for the purpose of performing a retest. The company s overall policy, intentions, and approach to validation, including the validation of production processes, cleaning procedures, analytical methods, in-process control test procedures, computerized systems, and persons responsible for design, review, approval, and documentation of each validation phase, should be documented. The critical parameters/attributes should normally be identified during the development stage or from historical data, and the necessary ranges for the reproducible operation should be defined. This should include ... 

An analytical procedure is in statistical control when the variation among the observed results can be attributed to a constant system of chance causes. It is assumed that an analytical method is in a state of statistical control when the distribution of results can be approximated by the normal distribution around the conventional true value ( Xr) with a standard deviation ( t). When a measurement system is in statistical control the data have statistical predictability and several statistical calculations can be performed for its evaluation and documentation. 

In summary, the ATP example shown here describes critical method performance requirements, that is, those characteristics that have a direct impact on the ability of a method to quantitate an analyte. LC-method attributes such as peak resolution, linearity, and efficiency, are not included. Although they are important features of an LC method and should be evaluated and incorporated into the control strategy as required, they do not provide a direct measiu-ement of the ability of a method to accurately quantitate an analyte and should not appear in the ATP. 

Chapter 4 covers the site selection and site controlling phase. Consequently, it deals with the assessment of individual production sites based on primarily qualitative criteria. Alternative Multiple Attribute Decision Analysis methods are reviewed and a decision support model employing the Analytic Hierarchy Process, which can be used both for site selection problems and as a controlling tool to perform site portfolio rankings of entire production networks, is proposed. Experiences from application in industry are reported. 

---