Page:NIOSH Manual of Analytical Methods - 8326.pdf/5

 S-Benzylmercapturic acid and S-phenylmercapturic acid in urine: METHOD 8326, Issue 1, dated 20 May 2014 - Page 5 of 10

concentration levels investigated. For each batch run, the experimental trial consisted of three samples at four different concentration levels. The recovery and precision for each level (n = 9 samples) are displayed in Table 2. The relative standard deviations (RSD) observed ranged from 2.0 to 5.0%. These accuracy and precision numbers fall well within parameters established for bioanalytical methods [7]. The second recovery study was performed using a urine substitute, fortified over three separate experimental batch runs, and these data are presented in Table 3. Average recoveries were between 99 and 109% for the two analytes over the four spiked concentration levels investigated. For each batch run, the experimental trial consisted of three samples at four different concentration levels. The recovery for each level (n = 9 samples) is displayed in Table 3. As seen in the RSD column of Table 3, precision was much worse for the urine substitute recovery study. This is due to a higher level of ion suppression in the electrospray source. This was especially pronounced at the lower spiked concentration levels (1 and 2 ng/mL). The highest result for the RSD of the 9 samples fortified at 1 ng/L BMA was 29%. While the accuracy numbers meet standard criteria, the precision values at the lower concentrations do not [7]. The use of a urine substitute for QC samples is not recommended for this method. The third recovery study used urine specimens from twelve non-exposed volunteers; six were smokers and six were non-smokers. Most participants’ specimens had base levels for both BMA and PMA as are shown in Table 4. Again, reasonable accuracy and precision were demonstrated for a 6 ng/mL equivalent level spike; individual recoveries ranged from 95 to 109% for BMA and 101 to 123% for PMA (Table 4). No interferences were detected in the unspiked urine from the 12 volunteer specimens. This study illustrates two important points: 1. Levels of both metabolites can be found in the urine of nonexposed individuals, which is why aqueous standards are used in this method, and 2. It is important to ascertain smoking status and other non-occupational exposures that may cause elevated levels of these metabolites. Linearity. All calibration curves used during the development of this method were linear and had correlation coefficients of 0.99 and greater. The concentration range was equivalent to 0.5 to 50 ng/mL BMA and PMA for the extraction of 4.0 mL urine. Calibration curves were run at the beginning and end of all sample batch runs; calibration curve slope drift was found to be minimal. Specificity. The optimized chromatographic conditions developed for this method, along with the tandem mass spectrometric detector, proved to be specific and showed no major interferences. The mass transition ions of BMA and PMA chosen in this method had the greatest response and were the predominant daughter ions. Robustness. Two C18 HPLC columns (Zorbax Rx-C18, Agilent Technologies, Santa Clara, CA, USA) from the same manufacturer but from different manufacturing lots were used during the recovery studies. Accuracy and precision were not affected; therefore, the method was found to be reproducible with any normal functioning C18 HPLC column. Recovery results from individual volunteer urine specimens spiked with the analytes indicate that the method is accurate and not significantly affected by individual urine specimen matrix differences during analyte extraction. The urine substitute was found to cause problems with precision, especially at lower levels; this was attributed to an increase in ion suppression over human urine within the electrospray source. Stability. Sample stability was evaluated, whereby a six-day stability study was conducted on the final chromatographic sample solution. BMA and PMA were stable at 8 °C (the autosampler temperature) and at room temperature in the absence of light. A light stability experiment was carried out by storing this solution in a clear glass vial at room temperature on the window sill, thus creating a worst-case scenario. After 1 day of storage in light under these conditions, BMA and PMA had mean assay values of 75 and 72% (n = 3), respectively, when compared to solutions of freshly prepared reference standards. After three days of light exposure, extensive degradation was noticed; BMA mean assay values had fallen to 9% (n = 3) of the original level and PMA had fallen to 16% (n = 3). After six days of exposure to light, both analytes were nearly completely degraded. Although both analytes benefit from the use of

NIOSH Manual of Analytical Methods (NMAM), Fifth Edition