By Clifford Nilsen, CSSBB
Editor’s Note: This is the second in a series of articles on analytical method validation (the first part appeared on p. 30 of our February/ March issue). The next article will review how to perform the various parts of analytical method validation and will provide sample acceptance criteria for each validation parameter.
This issue, we continue to discuss analytical method validation, which is the process of demonstrating through laboratory studies that an analytical method is suitable for its intended use. We will define each of the other method validation components, which include selectivity, linearity and range, accuracy and recovery, assay precision, intermediate precision, limit of detection, limit of quantitation, ruggedness, robustness, and comparative studies.
Selectivity: Selectivity (specificity) is often expressed as the degree of bias obtained by analyzing samples that contain added impurities, degradation products, related chemical compounds, or placebo ingredients against samples without added substances. The bias, if any, is the difference between the two groups of samples. This concept will make more sense when we reach the “How to Do It” section in part three of this series.
Linearity and Range Plus Injection Precision: The linearity of an analytical method is its ability to produce test results that are proportional to the concentration of the analyte in sample solutions, within a specified range, such as 50% to 150% of the working concentration. Linearity is usually expressed as the variance around the slope of a linear regression line. The importance of linearity depends upon how wide ranging the method is. Linearity data is evaluated using both Pearson’s correlation coefficient (r) and the significance of the correlation (r2). Injection precision is evaluated using descriptive statistics.
Accuracy and Recovery: The accuracy of an analytical method is defined by how close the test result is to the true value (analytical result versus actual value). The accuracy may be expressed as the percent recovery of known amounts of analyte that have been added to a placebo. It is a measure of the exactness of an analytical method. Accuracy will be evaluated using descriptive statistics in concert with regression techniques to determine whether the method accuracy has or does not have any statistically significant differences over a predetermined range of concentrations.
Accuracy, in cases where it is not possible to spike a placebo with active, such as with API (active pharmaceutical ingredient) assays, can be determined using the residuals on the linear regression line.
Assay Precision: The precision of an analytical method is the degree of agreement among individual test results when the procedure is applied repeatedly to multiple samplings of a homogeneous sample. The precision of an analytical method is usually expressed as the percent relative standard deviation of the individual test results.
Limit of Detection: The limit of detection (LOD) is the lowest concentration of analyte in the target matrix detectable at the most sensitive instrument settings that can be discriminated from background noise to a 95% confidence level. This is only meaningful for trace analysis such as impurity testing, related compounds, and residual solvents. It is generally not a requirement for assay methods.
Limit of Quantitation: The limit of quantitation (LOQ) is the minimum level of analyte in the target matrix that can be quantitated at the 95% confidence level. It is often defined as the lowest analyte level that can produce an injection precision of no greater than 5.0% for N=6 injections. LOQ is often used for analyzing impurities in bulk drug substances, degradation products, related compounds, and trace analyses such as residual solvents. Both LOD and LOQ are commonly determined based on visual evaluation, signal-to-noise approach, or the standard deviation of the response and the slope.1
Intermediate Precision (Intra-laboratory Precision): The intermediate precision of an analytical method is the degree of reproducibility of test results obtained by analyzing the same samples under a variety of normal test conditions. The method should not be prone to day-to-day or place-to place conditions, different analysts, instruments, reagent lots, elapsed assay times, assay temperatures, days, lots of mobile phase, standard preparations, or columns, for example. Intermediate precision is a measure of how test results vary under normal, expected operational conditions within the same laboratory and from analyst to analyst. Suitability of intermediate precision is determined with tests for equal variance and hypothesis testing using a two-sample T-test.
Robustness: The robustness of a method is a measure of its capacity to remain unaffected by small but deliberate normal variations in method parameters. It provides an indication of the method’s reliability during normal usage, in addition to that demonstrated by the intermediate precision portion of the validation. Examples of such variations are temperature, mobile phase composition, and flow rate. This robustness study will be conducted via a statistically based design of experiment, either full or fractional factorial, depending upon the number of experiments we wish to run.
Comparative Study: This can be a comparison between an official method, such as a United States Pharmacopeia (USP) method, and an in-house method to demonstrate equivalency. It also applies to demonstrating the suitability of a method for its intended use. For in-house methods, the entire validation comprises the comparison. For USP methods, accuracy and recovery studies will usually suffice. Caution: If a USP method exists but you opt to use your own method, and if a conflict arises, the USP method (or any regulatory method) will generally prevail.
A comparative study can also be an inter-laboratory precision study. This is the same as an intermediate precision measure, except that the second chemist is located in a different laboratory rather than in the same one. Inter-laboratory precision is commonly used for method transfers from one laboratory to another. Data can be evaluated for comparability using two-sample T-tests or linear model one-way analysis of variance, depending upon the number of data sets involved.
The next installment will begin to discuss how to perform the various parts of an analytical method validation and will offer sample acceptance criteria for each validation parameter. The “How to Do It” section will include calculations for linearity, residuals, accuracy and recovery, and robustness. I will demonstrate and explain Six Sigma statistical methods using example data from actual method validation studies.
Reference
- International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Guideline Q2(R1): Validation of analytical procedures: text and methodology. Geneva, Switzerland; 1994. Available at: www.ich.org/LOB/media/MEDIA417.pdf. Accessed June 13, 2010.
