Objective To evaluate the effects of analytic variations in creatintine measurements on estimated glo -merular filtration rate (eGFR) and the classification of chronic kidney disease ( CKD).Methods A total of 13 157 patients, including inpatients, outpatients and health individual , were enrolled, whose creatinine range from upper reference limit (URL) -12% ×URL to URL +12% ×URL.There were 9 886 males and 3 271 fe-males with an age range of 20 ~89 years.The results of sCr incremented or decremented 4%, 8%, 12% and orig-inal results were divided into 7 groups.The effects of different degree of analytic variation in sCr measurement on eGFR and classification of CKD using eGFR were evaluated .Results When sCr was in the range of URL ±12%×URL, creatintine increased with age, on the contrary, eGFR decreased with age (P <0.05).The mean bias of the results of eGFR compared with the original results of eGFR was increased with the analytic variation of sCr . When the results of sCr had a reverse analytic bias of 12%, the results of eGFR had a forward bias of 16.73%. When sCr results had a forward analytic bias of 12%, the male patients in stage G3 (eGFR in 30 ~59 ml/min/1.73 m 2 ), according to the classification of CKD using eGFR, increased about 20%; while women in this stage could increase by nearly 30%.When sCr results had a reverse analytic bias of 12%, the percentage of female pa-tients in stage G3 could decrease from 38.65% to 10.42%; while male patients could decrease by 7%.Conclu-sion The analytic variation of creatinine change in laboratory testing tolerance range can cause large shift in the distribution of eGFR, which can cause change in the classification of patients .The correct eGFR report relies on the accurate detection of sCr.Routine reporting of eGFR alongside creatintine should pay attention to the detection accuracy of sCr.

A total of 306 fresh human serum samples were randomly selected. Alb concentrations were measured by bromcresol green ( BCG) method, modified bromcresol purple ( mBCP) method, and immunoturbidmetic assay ( ITA) , respectively. GA was measured by an enzymatic method. GA value was expressed as the percentage of GA in the total serum Alb [ GA%(%) =GA/Alb?100%] . When Alb≥40 g/L, the clinical differences between BCG and mBCP, BCG and ITA, and mBCP and ITA were not significant, and there was no statistical difference between GA% BCG and GA% mBCP(P=0. 537); when Alb<40 g/L, BCG had statistical difference between mBCP and ITA ( P<0. 01 ) , and GA% BCG was significantly lower than GA% mBCP ( P<0. 01 ) . No obvious clinical signifi-cance of Alb concentrations was observed measured by BCG, mBCP, and ITA when Alb≥40 g/L. There was no difference in different Alb methods for GA% calculation. When Alb<40 g/L, the consistency of mBCP and ITA was better than that of BCG and ITA, which indicated that mBCP method may be more suitable for the assay of gly-cated albumin value ( GA%) , whereas the GA% may be underestimated if using BCG method for the determination of Alb.