Objective To study the role of clinical patient samples in verifying consistency of clinical results between different reagent lots.Methods The results of 5 clinical samples,Roche quality control materials and Biorad quality control materials for 13 clinical routine biochemical tests were obtained by using both old and new reagent lot,and performed the statistics.The bias of 13 measurands were computed after changing reagent lots,and also were judged according to the 1/3 limits of CLIA' 88 defined acceptable error,whether the bias was conform to clinical requirements or not.Results QC material results had no significant difference between two reagent lots (P ＞ 0.05).The results of ALT,UA,TC and ApoA for patient samples had significant difference (P ＜ 0.05),but the bias of these items were within the 1/3 limits of CLIA' 88 defined acceptable error,which were 2.57％,4.41％,1.81％ and 3.84％ respectively,in accordance with clinical requirement.Conclusions As there is matrix effect of quality control materials,the results for QC materials should not be used to verify the consistency of results for patient samples when changing reagent lots.To overcome this situation,we should use quality control materials together with clinical samples to compare consistency of results after changing reagent lots.