De Beer T R M, Baeyens W R G, Vermeire A, Broes D, Remon J P, Vervaet C
Laboratory of Drug Quality Control, Department of Pharmaceutical Analysis, Ghent University, Harelbekestraat 72, B-9000 Ghent, Belgium.
Anal Chim Acta. 2007 Apr 25;589(2):192-9. doi: 10.1016/j.aca.2007.03.002. Epub 2007 Mar 12.
An alternative fast and non-destructive validated Raman spectroscopic analytical procedure, requiring no sample preparation, was compared with the industrially applied HPLC reference method (Pfizer Manufacturing Belgium) for the quantitative determination of medroxyprogesterone acetate (MPA) in DepoProvera suspensions (150 mg mL(-1), Pfizer). The Raman calibration model was developed by plotting the peak intensity of the baseline-corrected and normalized spectral band (corrected by external standard measurements) between 1595 and 1620 cm(-1) against known MPA concentrations in standards. At this band, no spectral interferences from the suspension medium are observed. The most suitable model for the calibration data (straight line or higher order polynomial) was determined by evaluating the fit and predictive properties of the models. In a second step, the developed Raman spectroscopic analytical method was validated by calculating the accuracy profile on the basis of the analysis results of validation samples. Furthermore, based on the data of the accuracy profile, the measurement uncertainty was determined. Finally, as the aim of the alternative method is to replace the destructive, time-consuming HPLC method, requiring sample preparation, it needs to be demonstrated that the new Raman method performs at least as good as the HPLC method. Therefore, the performance (precision and bias) of both methods was compared. A second order polynomial calibration curve through the calibration data supplies the best predictive properties and gives an acceptable fit. From the accuracy profile, it was concluded that at the target concentration (150 mg mL(-1)), 95 out 100 future routine measurements will be included within the acceptance limits (5%). Comparison of the alternative method with the reference method at the target concentration indicates that the Raman method performs at least as good as the HPLC method for precision (repeatability and intermediate precision) and bias. The fast and non-destructive Raman method hence provides an alternative for the destructive and time-consuming HPLC procedure.
将一种无需样品制备的快速、无损且经过验证的拉曼光谱分析方法,与工业上应用的HPLC参考方法(比利时辉瑞制药)进行比较,用于定量测定醋酸甲羟孕酮(MPA)在醋酸甲羟孕酮注射液(150 mg mL⁻¹,辉瑞)中的含量。拉曼校准模型通过绘制经基线校正和归一化后的光谱带(通过外标测量校正)在1595至1620 cm⁻¹之间的峰强度与标准品中已知MPA浓度的关系来建立。在此波段,未观察到来自悬浮介质的光谱干扰。通过评估模型的拟合度和预测性能,确定了最适合校准数据的模型(直线或高阶多项式)。第二步,基于验证样品的分析结果计算准确度轮廓,对所开发的拉曼光谱分析方法进行验证。此外,根据准确度轮廓的数据确定测量不确定度。最后,由于替代方法的目的是取代需要样品制备的破坏性、耗时的HPLC方法,因此需要证明新的拉曼方法至少与HPLC方法表现一样好。因此,比较了两种方法的性能(精密度和偏差)。通过校准数据的二阶多项式校准曲线具有最佳的预测性能,并给出了可接受的拟合度。从准确度轮廓得出结论,在目标浓度(150 mg mL⁻¹)下,未来100次常规测量中有95次将包含在接受限(5%)内。在目标浓度下将替代方法与参考方法进行比较表明,拉曼方法在精密度(重复性和中间精密度)和偏差方面至少与HPLC方法表现一样好。因此,快速、无损的拉曼方法为破坏性且耗时的HPLC程序提供了一种替代方法。