利用统计学方法消除检验延误对大鼠致癌试验凝血指标的干扰

Application of statistical method in eliminating interference of testing delays on coagulation parameters of rat carcinogenic test

目的利用析因分析、偏相关和多元回归分析判断与消除检验延误(时间因素)对大鼠致癌试验凝血指标的干扰。方法对照、低、中、高剂量4个实验组共168只Wistar雌性大鼠致癌试验期满后(104周)以磁珠法检测凝血酶原时间(PT)、活化部分凝血活酶时间(APTT)和凝血酶时间(TT)。采用析因分析方法并结合偏相关和多元回归分析对采血后3~24 h内完成检测的数据进行分析。结果析因分析显示,APTT分别受时间和剂量的独立影响,TT受时间影响并与染毒剂量存在交互作用,PT亦受时间影响。偏相关和多元回归分析显示,APTT、TT分别与时间有相关性,并入选APTT、TT的回归模型;PT与时间无相关性,亦无法建立回归模型,由此判断因检验延误延长了APTT和TT,产生干扰的时间节点为5 h。剔除问题数据后对照、低、中、高4个实验组APTT均值分别下降了1.85%、4.56%、20.08%和30.45%,TT分别下降了4.01%、6.78%、11.76%和16.64%。结论联合使用析因分析、偏相关和多元回归分析可识别和消除检验延误对大鼠致癌试验凝血指标APTT和TT的干扰。

Objective To distinguish and eliminate the interference of testing delays（ time factors） on coagulation parameters in rat carcinogenic test using statistical method. Methods In four groups of low,middle and high dose,and control,168 female Wistar rats were tested for prothrombin time（ PT）,activated partial thromboplastic time（ APTT） and thrombin time（ TT） by magnetic beads method after finishing the carcinogenic test（ 104 weeks）. After blood samples were collected and tested within 3-24 h,the data were analyzed using factor analysis method combining with partial correlation and multiple regression analysis. Results The factor analysis showed that APTT was independently affected by time and dose,TT was influenced by time and interacted with the dose of exposure,and PT was also influenced by time. Partial correlation and multiple regression analyses showed that APTT and TT were separately correlated with time and the time was selected into the regression model of APTT and TT,respectively; PT and time had no correlation and can＇t establish a regression model. Based on the statistical results it is judged that testing delay had affect APTT and TT,and the time threshold was 5 h. After removing the problematic data,mean values of APTT in four groups of low,middle and high dose,and control decreased by 1. 85%,4. 56%,20. 08%,and 30. 45%,and those of TT decreased by 4. 01%,6. 78%,11. 76%,and16. 64%,respectively. Conclusion The combined use of factor analysis,partial correlation analysis and multiple regression analysis can distinguish and eliminate the interference of testing delays on coagulation parameters APTT and TT in rat carcinogenic test.