Stokvold Anders, Dyrstad Knut, Libnau Fred Olav
Chemical Development, Oslo Plant, Nycomed Imaging AS, P.O. Box 4220, Nydalen, N-0401, Oslo, Norway.
J Pharm Biomed Anal. 2002 Jun 1;28(5):867-73. doi: 10.1016/s0731-7085(01)00668-9.
The purpose of this study was to build a best possible NIR prediction model for monitoring of water content in a freeze-dried drug product. The best pre-treatments of the NIR spectra were found to be: transforming from reflection to absorption, baseline correction in the 1845-2165 nm area and a maximum normalisation in the same area. These pre-treatments resulted in a model with the following attributes: SEP of 0.08% (w/w) and one PLS factor, the latter indicating a robust model. The limit of quantification was calculated to 0.24% (w/w). During the stability study an increase in water content in the freeze-dried drug product was revealed, which were found to depend on storage time and temperature. It is believed that the water is derived from the stoppers. The highest increase was found for storage at 40 degrees C, and was estimated to be 0.04% points a month by weight, from an initial value of about 0.25% (w/w).
本研究的目的是建立一个用于监测冻干药品水分含量的最佳近红外预测模型。发现近红外光谱的最佳预处理方法为：从反射转换为吸收、在1845 - 2165 nm区域进行基线校正以及在同一区域进行最大归一化。这些预处理得到了一个具有以下特性的模型：SEP为0.08%（w/w）且有一个偏最小二乘法（PLS）因子，后者表明该模型稳健。定量限计算为0.24%（w/w）。在稳定性研究期间，发现冻干药品中的水分含量增加，且这取决于储存时间和温度。据信水分来自瓶塞。在40摄氏度储存时水分增加最多，从初始值约0.25%（w/w）开始，估计每月重量增加0.04个百分点。
