基于Excel软件实现四参数logistic模型在生物活性测定中的规范化应用

Normalized application of four-parameter logistic model in bioassay based on Microsoft Excel

目的在生物活性测定中建立规范的基于四参数logistic模型分析流程,并利用Excel软件实现其拟合及计算功能。方法以各国药典均收录的重组人粒细胞集落刺激因子(recombinant human granulocyte colony-stimulating factor,rh G-CSF)生物活性测定研究数据的分析为例,参照国外药典该制品量反应平行线测定法要求,建立基于四参数logistic模型的规范分析流程(模型拟合、可靠性测验、相对效价及置信区间估计)及配套的Excel分析模板,比较其与生物统计分析软件SoftMax及PLA分析结果的一致性。同时采用《欧洲药典》收载的四参数回归计算法实例数据对其进行验证。结果研究数据经Excel分析模板分析,Excel的四参数logistic模型拟合与Softmax及PLA对各参数的估计值基本一致;Excel可靠性测验结果与Softmax及PLA相应结果基本一致;3种软件计算相对效价的95%置信区间均为98.3%~109.7%,其相对置信区间均为94.6%~105.6%,结果一致。实例数据的Excel曲线拟合结果、效价及置信区间计算结果与《欧洲药典》报告结果一致,可靠性测验结果与PLA一致。结论建立的生物活性测定分析流程涵盖的可靠性测验及置信区间估计为实验结果有效性及实验误差范围的判断提供了依据,且建立的Excel计算模板分析结果准确可靠,为检验人员在分析软件的选择上提供了依据。

Objective To establish a standardized analysis process based on four-parameter logistic model in the bioassay and to realize the fitting and calculation functions by using Excel software. Methods Based on the bioassay data analysis of recombinant human granulocyte colony-stimulating factor(rhG-CSF)and the requirements for parallel line assay of the product in different pharmacopoeias,a normalized analysis process based on four-parameter logistic model,including model fitting,test for validity,as well as estimations of relative potency and confidence interval,was established. The matched analysis template was set up by Excel software, and the results were compared with those obtained by internationally recognized statistical analysis software such as SoftMax and PLA. The analysis template in Excel was further validated by comparing its analysis result with reported result of example in European Pharmacopeia. Results The results of four-parameter logistic model fitting and test for validity by Excel software based on selected bioassay data of rh G-CSF was consistent with that by SoftMax and PLA analyses. The 95% confidence intervals of relative potencies calculated by the three software were 98. 3% ~ 109. 7%,while the relative confidence intervals were 94. 6% ~ 105. 6%,which were in agreement. The analysis results by Excel software was also comparable with the reported result in European Pharmacopeia,while the test result for validity was consistent with that by PLA software. Conclusion The tests of validity and confidenc e interval estimation included in the analysis procedure based on four-parameter logistic model provide a basis for the judgment on whether the experimental results are valid and errors are acceptable. The results of Excel analysis template are reliable,which provide a support for the selection of analysis software.