Leito Signe, Mölder Kadi, Künnapas Allan, Herodes Koit, Leito Ivo
Institute of Chemical Physics, University of Tartu, Jakobi 2, 51014 Tartu, Estonia; State Agency of Medicines, Nooruse 1, 50411 Tartu, Estonia.
J Chromatogr A. 2006 Jul 14;1121(1):55-63. doi: 10.1016/j.chroma.2006.04.032. Epub 2006 Jun 6.
An ISO GUM measurement uncertainty estimation procedure was developed for a liquid-chromatographic drug quality control method-assay of simvastatin in drug formulation. In quantification of uncertainty components several practical approaches for including difficult-to-estimate uncertainty sources (such as uncertainty due to peak integration, uncertainty due to nonlinearity of the calibration curve, etc.) have been presented. Detailed analysis of contributions of the various uncertainty sources was carried out. The results were calculated based on different definitions of the measurand and it was demonstrated that unequivocal definition of the measurand is essential in order to get rigorous uncertainty estimate. Two different calibration methods - single-point (1P) and five-point (5P) - were used and the obtained uncertainties and uncertainty budgets were compared. Results calculated using 1P and 5P calibrations agree very well. The uncertainty estimate for 1P is only slightly larger than with 5P calibration.
针对液相色谱法测定药物制剂中辛伐他汀的药物质量控制方法，开发了一种符合ISO GUM的测量不确定度评估程序。在不确定度分量的量化过程中，提出了几种包含难以估计的不确定度来源（如峰积分引起的不确定度、校准曲线非线性引起的不确定度等）的实用方法。对各种不确定度来源的贡献进行了详细分析。结果是基于被测量的不同定义计算得出的，结果表明，为了获得严格的不确定度估计，明确被测量的定义至关重要。使用了两种不同的校准方法——单点（1P）和五点（5P）——并比较了获得的不确定度和不确定度预算。使用1P和5P校准计算的结果非常吻合。1P的不确定度估计仅比5P校准略大。
