Chemistry Reference
In-Depth Information
Precision
Accuracy
Limit of Detection
Limit of Quantitation
Method
Validation
Specificity
Linearity
Range
Robustness
FIgure 4.1
AMV analytical performance characteristics.
4.3.1 A
ccurAcy
Accuracy is the measure of exactness of an analytical method, or the closeness of
agreement between an accepted reference value and the value found in a sample.
Established across the range of the method, accuracy is measured as the percent of
analyte recovered by the assay. For the drug substance, accuracy measurements are
obtained by comparison of the results to the analysis of a standard reference material,
or by comparison to a second, well-characterized method. For the assay of the drug
product, accuracy is evaluated by the analysis of synthetic mixtures spiked with known
quantities of components. For the quantification of impurities, accuracy is determined
by the analysis of samples (drug substance or drug product) spiked with known
amounts of impurities (if impurities are not available, see specificity, Section 4.3.3).
Table 4.1 illustrates the results for a typical accuracy study. To document accu-
racy, the guidelines recommend that data be collected from a minimum of nine
determinations over a minimum of three concentration levels covering the specified
range (i.e., three concentrations [e.g., 50%, 100%, and 150% of target value], three
replicates each). The data should be reported as the percent recovery of the known,
added amount, or as the difference between the mean and true value with confi-
dence intervals (±1 SD). In Table 4.1, data is shown relative to 100%, and the mean
recovery for
n
= 9 samples is 98.69% with %RSD = 0.28%. In this example, both the
accuracy and precision pass the predefined acceptance criteria of 98% to 102% and
≤2%, respectively.
In most cases, accuracy requires a priori knowledge of the level of analyte
present in simulated drug products, or in synthetic mixtures spiked with known
amounts or quantities of the components of interest (analytes). When standards
are available, it is sufficient to demonstrate that results are unaffected by using
spiked samples. However, in situations where the actual level is not known, or if
standards of the analyte of interest are not available, then accuracy can only be
