A step-by-step guide for proper validationPart 4 of 4: This final installment completes our series on method validation, continuing the discussion of the mechanics of the process.
Read the Complete Series
- Part 1, PFQ February-March 2010: Described high-performance liquid chromatography (HPLC) procedures for demonstrating that a method is stability indicating through the use of forced degradation studies and evaluation of peak purity using a photo diode array UV detector.
- Part 2, PFQ June-July 2010: Defined each of the other method validation components—selectivity; linearity and range; accuracy and recovery; assay precision; intermediate precision; limit of detection; limit of quantitation; ruggedness, robustness, and comparative studies—with the primary focus on assay methods.
- Part 3, PFQ October-November 2010: Began the description of how to perform method validation for each validation component.
As with linearity, prepare a series of five standards that span a range of 50% to 150% of the analyte working range except that each standard will contain placebo in an amount proportional to the placebo/active ratio of the drug product for which the method is being validated. For example, if a tablet weighing 500 mg contains 50 mg of active pharmaceutical ingredient (API), 450 mg is placebo.
Using the method being validated (assume it is an HPLC method), inject each spiked standard three times. For each spiked standard—50%, 75%, 100%, 125%, 150% of the working standard, respectively—each containing a proportional amount of placebo, calculate the area unit’s percent relative standard deviation (%RSD) to determine injection precision at each level. Use the mean values for active concentration at each of the five levels, computing the percent recovery from the placebo at each level.
In most cases, it is desirable to have an injection precision, in terms of area unit RSD, of less than 2% for each standard level, i.e., 50% to 150% of the working concentration. Recovery (accuracy) limits vary with method requirements, but are usually considered acceptable if the recovery of spiked active from the placebo at each level is between 98% and 102%.
Assay PrecisionMultiple (six) sample preparations are made from a single homogeneous sample, and the six separate preparations are assayed, using the method under validation, versus a freshly prepared standard. Precision between individual assay results is calculated and expressed as %RSD. An assay precision of not more than 2% is generally considered acceptable for assays. Some applications, such as residual solvent determinations or trace analysis, will have different acceptance criteria.
Limit of DetectionThe limit of detection (LOD) of the analytical method is determined by comparing the test results obtained from samples with known concentrations of analyte against those of blank samples and establishing the minimum level of analyte that can be detected. There are a number of ways to express LOD, including a multiple of noise level, a minimum area count %RSD, or a fixed percentage of the lower limit of the linearity curve (50% lowest level for example). LOD is of little importance for assay determinations but is of great importance to applications such as impurity analyses or determination of trace solvent levels.
Experimentally, LOD can be determined by serial dilution of a working standard until the sample peak is indistinguishable from baseline noise.
Limit of QuantitationThe limit of quantitation (LOQ) of the analytical method is the lowest level at which analyte can be reliably measured. Some common definitions of LOQ: three times the LOD and a level at which the %RSD of injection precision is less than 5%. As with LOD, LOQ can be determined by serial dilutions of a standard. LOQ is important in determination of trace components such as impurities and residual solvents.
Ruggedness and Intermediate PrecisionThe ruggedness of an analytical method is determined by analyzing multiple samples from homogeneous lots. Samples from the same lot are assayed in hextuplicate (six times) using six sample preparations (six assays). Separate sets of six assays from the same homogeneous sample are performed by different chemists on different days, using different columns, different instruments (if possible), and different standard preparations. The %RSD of each set of assay results for each chemist should be no greater than 2.0, and the pooled %RSD for all 12 assays should be no greater than 2.0.
As a point of information, ruggedness (inter-lab precision) refers to performing each of the two assay sets in different labs, whereas intermediate precision (intra-lab precision) refers to performing each of the two assay sets in the same lab.
RobustnessRobustness is determined by observing how a method responds to slight variations in normal operating parameters. In HPLC methods, for instance, this could be a change in flow rate, column length, column temperature, or mobile phase concentration. A simple way of doing this is by performing assay precision under various conditions that vary slightly from the method parameters.
For example, if a method calls for a 30 cm column, a 1.0 mL/minute flow rate, a column temperature of 30 degrees C and a mobile phase consisting of 80 parts water and 20 parts methanol, one could perform an assay precision (six replicate assays) at 0.8 mL/minute, 1.0 mL/minute and 1.2 mL/min, at column temperatures of 28 degrees C, 30 degrees C, and 32 degrees C, at method conditions on a 25 cm column and at mobile phase ratios of 78/22, 80/20 and 82/18 water/methanol. One set of sample preparations is used for all these experiments (six weighings and a ton of injections). The %RSD of the assay results at slightly varied conditions should be no greater than the maximum allowed under normal method operating conditions.
Acceptance CriteriaThe validation protocol must include acceptance criteria for each validation parameter. The criteria are the performance requirements of the method—a yardstick against which the method’s validity is measured and which will vary depending on the intended application. The typical acceptance criteria for a drug product assay method are cited in each of the validation parts described above. Refer to part three in this series for information on stability indication, selectivity, and linearity and range.
Well, this story is fictional, but glycerin from China contaminated with anti-freeze is real. How can we as consumers avoid buying contaminated or substandard drug products? Do you know if the products sitting in your medicine chest or kitchen cabinets are safe to use? How can you be sure? This book uncovers and documents questionable manufacturing practices used by drug companies, mostly generic companies, that could have serious health consequences for American consumers who purchase the drugs produced by those companies.
No one wants to use drug products that are contaminated with dangerous chemicals, foreign matter such as metal and glass, or harmful bacteria, or wants to purchase and use drugs that are mislabeled or have the wrong strength, or that have been made using shoddy manufacturing procedures in dirty equipment by untrained workers. Yet, these things are commonplace, particularly with over-the-counter drugs.
The Validation ReportOnce the validation has been executed, a validation report is prepared and submitted for approval. The elements described below should be included in the validation report.
- Summary: The summary is a simple statement about the results of the validation study. For example, “The method for the assay of product XYZ by HPLC was found to be accurate, precise, selective, linear, and stability indicating, and thus suitable for its intended use.”
- Analytical Validation Data: Analytical data should be presented in tabular and graphical form for ease of evaluation. The data presentation should show analytical results for each validation parameter, plus residuals and all calculations used to derive results from laboratory data. Be sure to include all graphs, curves, and copies of raw data (chromatograms and notebook pages).
- Discussion: Describe the outcome of the validation in detail. The discussion/ conclusions should deal with any problems that were encountered and should include a rationale for accepting or rejecting the validation. Any experiments or failing results that were repeated and then accepted need to be explained and justified.