Influence of errors in sampling time and in activity measurement on the single sample clearance determination

Introduction Plasma clearance rate of 51 Cr-EDTA estimated by using one blood sample is commonly used for the calculation of glomerular filtration rate. Aim To estimate the error on single-sample clearance determination induced by errors in sampling time and activity measurement, and to compare it with the error observed on the clearance determination obtained using the slope-intercept method. Methods Forty-five adult patients were chosen from a data base of 51 Cr-EDTA plasma clearance values determined by using two blood samples taken around 2 and 4 h. Patients were selected in such a way as to include clearances from 30 ml.min -1 to 155 ml.min -1 , with steps of 3 ml.min -1 . Based on the slope and the intercept of the slope with the y-axis, the plasma concentration at exactly 2 and 4 h was determined. Normally distributed random errors were then introduced in the sampling time (SD of 0, 1 and 2 min) as well as in the activity measurement (SD of 0,1, 2 and 5%). Then, clearance was calculated using two single-sample methods (i.e. the algorithms of Groth and Tauxe), and the slope-intercept method, which requires two blood samples. For each setting, the simulation was repeated 200 times. The effects on clearance of a random error on the time sampling and/or the activity measurement were then evaluated. Results The error on single-sample clearance induced by a 2 min error in sampling time associated with a 5% error in activity measurement was negligible. For all clearance levels, the SD of the error on the calculated clearance was less than 3.8 ml min-1. Whatever algorithm was chosen, the errors on the single-sample clearance were systematically lower than those observed with the slope-intercept method, for the whole clearance range. Conclusion Errors in sampling time and in activity measurement induced only a very small error on the single-sample EDTA clearance, which is systematically lower compared to that observed on the slope-intercept method using two blood samples.