B-302 Increased False-Positives vs Increased Interference Flags: Using the AU 5800 to Compare Two Ethyl Glucuronide Assays

Abstract Background Ethyl glucuronide (EtG) is a metabolite of ethanol that can be detected in urine for up to three to five days following consumption of alcohol. Urine specimens are routinely screened for EtG to assess ethanol exposure for compliance and/or abuse with presumptive positive samples being confirmed using quantitative Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS). LC-MS/MS is costly and labor intensive. Therefore, providing a sensitive and specific screening option for EtG in urine is beneficial, with it being more cost-effective, efficient, and widely available. Evaluation of the performance characteristics of different EtG assays can aid in identifying an assay that most closely agrees with LC-MS/MS. Methods To evaluate agreement, 100 residual urine specimens (approximately 50 positive/50 negatives identified using ARK Ethyl Glucuronide Assay) were thawed, mixed, and analyzed. Specimen were analyzed using ARK and Immunalysis Ethyl Glucuronide Urine HEIA on the Beckman Coulter AU 5800. Discrepant results were investigated using LC-MS/MS. Additionally, 15 false-positive samples (identified positive by ARK but confirmed negative by LC-MS/MS) were evaluated on both ARK and Immunalysis assays. Lastly, 7 samples were analyzed on ARK and Immunalysis because of the inability to quantitate by LC-MS/MS due to interfering substance. Results were evaluated using EP Evaluator and Microsoft Excel. Results Of the 100 samples analyzed, 8 samples were identified as discrepant between the two immunoassays. These 8 discrepant samples tested positive on ARK and negative on Immunalysis, resulting in an 83.3% positive agreement and a 100.0% negative agreement between ARK and Immunalysis. For these 8 discrepant samples, LC-MS/MS confirmed 7 samples negative and 1 sample positive suggesting that using ARK would result in 14.6% of presumptive positives confirming negative. Of the initial 100 samples, 7 samples resulted with interference flags by Immunalysis, indicating the presence of an unknown interfering substance within the sample. Of these samples with interference flags, 3 were among the 8 discrepant samples. Evaluation of identified false-positive samples (n = 15) by ARK resulted in 90.9% negative agreement for Immunalysis compared to 0.0% negative agreement by ARK. Of these 15 false-positive samples tested, 3 had interference flags identified using Immunalysis. Between agreement samples (n = 100) and false-positive samples (n = 15), a total of 10 interference flags were identified by Immunalysis, 6 of which were observed to have discrepant results between ARK and Immunalysis. ARK identified no interference flags. Interference was confirmed for 7 samples using LC-MS/MS, with Immunalysis detecting interference in 57% of those samples, and ARK reported no interference flags. In total, LC-MS/MS results were obtained for 37 samples, of which, only 2 false-positive and 1 false-negative samples were observed using Immunalysis. Conclusion While the assays offered by ARK and Immunalysis are similar in that they screen for EtG, we observed the accuracy between the two assays to be quite different, when compared to LC-MS/MS. Using ARK reagents would result in greater number of presumptive positives. Of note, the instrument flags accompanying Immunalysis results appear to be indicative of an interfering substance that ARK is not detecting and then subsequently reports as (falsely) positive.