Influence of time interval and number of blood samples on the error in renal clearance determination using a mono-exponential model: a Monte Carlo simulation.

摘要

Mono-compartmental analysis based on 2- and 4-h blood samples (BS) of 51Cr-EDTA (EDTA, ethylenediaminetetraacetic acid) is commonly used for the calculation of the glomerular filtration rate (GFR). The purposes of this study were to estimate the magnitude of error in clearance induced by errors in the time of sampling and activity measurement; to estimate the impact of changing the interval between the BS; and to assess the influence of a higher number of BS in reducing the error. A model of mono-exponential curves based on a finite number of BS was created. Normally distributed random errors were introduced in the time of sampling and activity measurement. In a first step, three different time intervals were used; in a second step, seven different numbers of BS were used, all taken between 120 and 240 min. For each setting, the random errors were successively introduced 200 times and the coefficients of variation (CV) of the calculated clearances were determined. Variable errors in clearance were induced by errors in the time of sampling and activity measurement. In general, the observed errors were higher for high and low clearance, with lower errors for moderately reduced clearances. The errors in indicating the time of sampling played an important role for high clearance, whereas the errors in activity measurements led to important errors for low clearance. Prolonging the interval from 1 to 2 h resulted generally in an important decrease in error, except in the range 60-100 ml x min(-1). Prolonging the interval from 2 to 3 h resulted in only a small additional decrease in error, except for very low clearance. Errors in indicating the time of sampling and in activity measurements induce errors in clearance determination. These errors cannot be significantly reduced by simply increasing the number of BS or by prolonging the interval between the samples. It is probably better, in most cases, to keep using the 2-4-h method and to take extreme care when indicating the time of sampling and when measuring the activity, instead of increasing the number of samples or lengthening the procedure.