Prasad Narasimha N G, Muddukrishna Shammana N
Department of Mathematical and Statistical Sciences, University of Alberta, T6G 2G1, Edmonton, AB, Canada.
Biosyn Corporation, 5939 Darwin Court, Suite 114, Carlsbad, CA, 92008, USA.
Trials. 2017 Mar 14;18(1):123. doi: 10.1186/s13063-017-1843-5.
Retrospective analysis of Center effect of the multi-center trial conducted to compare Immucothel (KLH Immunotherapy drug product) with Mytomycin-C (MM) concluded that efficacy evaluation of the drug product may be impacted by physician's subjective performance of Transurethral resection of bladder tumor (TURBT).
A randomized trial was performed in 18 hospitals (clinical centers) and a total of 553 recruited, 283 patients under KLH arm and 270 patients under MM. An initial statistical analysis of efficacy comparisons between KLH and MM based on log-rank test performed for each center (hospital) showed 6 hospitals out of 18 hospitals a p-value of <0.05 and remaining 12 hospitals showed a p-values of >0.05. No association was observed between number of patients analysed and the associated p-values across hospitals. Final statistical analyses were carried out under each drug product using Kaplan-Meier survival analysis along with log-rank test after combining all eligible patients data for 6 hospital group and 12 hospital group.
Median recurrence free survival (RFS) times (in weeks) showed statistical significance (p-value = 0.03) between two groups of hospitals under KLH arm, while similar median values showed no statistical significance (p-value = 0.05).
Center effect with respect to median RFS values under KLH was more pronounced than under MM. Under the presence of such center effect, for reasons other than product related effects, concluding superiority of one drug product over another may create confounding bias conclusions in multi-center clinical trials. In the above cited clinical trial study, physician's prior experience on TURBT might have contributed to center effect in examining efficacies of KLH and MM. Similar observation was also noted in the literature on studies dealing with TURBT and in other clinical studies.
Data set used in this study is based on previously documented clinical trial in the literature: See (Lammers et al., J Clin Oncol 30:2273-9, 2012) and Acknowledgments.
一项旨在比较免疫疗法药物Immucothel（KLH免疫治疗药物）与丝裂霉素C（MM）的多中心试验的中心效应回顾性分析得出结论，该药物产品的疗效评估可能受到医生经尿道膀胱肿瘤切除术（TURBT）主观操作的影响。
在18家医院（临床中心）进行了一项随机试验，共招募了553名患者，283名患者接受KLH治疗，270名患者接受MM治疗。基于对数秩检验对每个中心（医院）进行的KLH与MM疗效比较的初步统计分析显示，18家医院中有6家医院的p值<0.05，其余12家医院的p值>0.05。各医院分析的患者数量与相关p值之间未观察到关联。在合并6家医院组和12家医院组的所有符合条件的患者数据后，使用Kaplan-Meier生存分析和对数秩检验对每种药物产品进行最终统计分析。
KLH治疗组两组医院之间的无复发生存期（RFS）中位数时间（以周为单位）显示出统计学意义（p值 = 0.03），而相似的中位数未显示出统计学意义（p值 = 0.05）。
KLH治疗下的RFS中位数中心效应比MM治疗下更明显。在存在这种中心效应的情况下，由于与产品无关的其他原因，得出一种药物产品优于另一种药物产品的结论可能会在多中心临床试验中产生混淆性偏差结论。在上述临床试验研究中，医生对TURBT的既往经验可能导致了在检查KLH和MM疗效时的中心效应。在关于TURBT的研究文献以及其他临床研究中也有类似的观察结果。
本研究中使用的数据集基于文献中先前记录的临床试验：见（Lammers等人，《临床肿瘤学杂志》30:2273 - 9，2012年）及致谢部分。
