Biological monitoring of low level exposure to pollutants is a very challenging analytical activity, and the quality of results is difficult to assess, especially when a certified reference material is unavailable. The aim of this work was to evaluate the reliability of the assays used to measure urinary benzene (Benz-U) and S-phenylmercapturic acid (SPMA), by applying an internal quality control protocol. Urine spot samples from 705 subjects who were either members of the general urban population, gasoline station attendants, or refinery plant workers were assayed for Benz-U and SPMA, using GC/MS and LC/MS/MS, with quantification limits of 15 ng/L and 0.10 μg/L. The median Benz-U concentration was 263 ng/L (60-2789 ng/L, 5th-95th percentile), and the median SPMA concentration was 0.19 μg/L (<0.1-2.5 μg/L, 5th-95th percentile). Linearity of both assays was good, but a less-than-proportional response was found for SPMA concentrations below 1 μg/L. Between-run precision and accuracy for Benz-U concentration determination were assessed using quality controls at 120 ng/L and 1000 ng/L and were 10.3% and 4.8%, and 104.8% and 98.9%, respectively; while the precision and accuracy for SPMA concentration determination at 0.3 μg/L, 2.5 μg/L, and 20 μg/L were 40.3%, 6.2%, and 6.2%, and 48.3%, 96.3%, and 98.8%, respectively. Precision, estimated using duplicates of unknown samples, was 13.4% for Benz-U and 26.5% for SPMA analyses. Control charts for the means of the slope of the linear calibration curve of Benz-U showed good stability of the means over a five-year period. For SPMA, a two-laboratory comparison revealed acceptable agreement between ln-transformed data pairs, with a slope of the linear regression of 0.863 (confidence interval 0.774-0.952), null intercept, and a Pearson's r value of 0.844. Reliable results were obtained for Benz-U analyses over the entire concentration range, and for high and medium SPMA levels. However, the determination of SPMA concentrations at levels close to the limit of quantification was less reliable.
Copyright © 2009 Elsevier B.V. All rights reserved.