A multiple-site laboratory evaluation of three on-site urinalysis drug-testing devices

Presented are findings from a multisite laboratory evaluation comparing on-site urinalysis drug-test results to results from Syva EMIT immunoassay and gas chromatography-mass spectrometry (GC-MS). Three laboratories participated in the NHTSA-funded project. Specimens were tested for amphetamines, benzodiazepines, cocaine, cannabinoids, and opiates. Each laboratory selected 20 urines that tested positive for a single drug/drug class and 20 that tested negative to challenge the on-site drug-testing devices. Qualitative and quantitative GC-MS confirmations were performed to ensure that all positive samples contained the target drug(s)/metabolite(s) and that all negative samples did not contain the target analytes. EZ-SCREEN, ONTRAK, and TRIAGE on-site test kits were selected for evaluation. On-site false-positive results, in which GC-MS-verified negative urine samples gave positive on-site results, were rare. Two such errors were recorded with both EZ-SCREEN and TRIAGE. Cross-reactivity from samples containing GC-MS-verified high concentrations of alternate drugs was also rare. One cross-reactive error was recorded while testing for cocaine with EZ-SCREEN, a second while testing for benzodiazepines with ONTRAK, and a third while testing for cocaine with ONTRAK. The EZ-SCREEN kit did not appear to adhere to a cutoff concentration as demonstrated by the number of samples that contained low concentrations of the target drugs that tested positive with this device. A significant finding of this study was that comparing on-site test device results with those of EMIT for samples with drug concentrations near the reporting cutoff was very complex. It required a thorough knowledge of the performance of each device, EMIT, and GC-MS. It also required an investigation of each discrepant result-a consideration not taken in many previous evaluations of on-site testing devices. Compared with current federal guidelines for workplace urinalysis testing, more donor samples would screen positive for cannabinoids and cocaine by the on-site devices than by EMIT immunoassay. However, fewer would be reported as positive because most contained GC-MS-determined drug concentrations lower than the federal confirmation and reporting limits.