BackgroundImmunoassays provide a rapid tool for the screening of drugs-of-abuse (DOA). However, results are presumptive and confirmatory testing is warranted. To reduce associated cost and delay, laboratories should employ assays with high positive and negative predictive values (PPVs and NPVs). Here, we compared the results of urine drug screens on cobas 6000 (cobas) and ARCHITECTi2000 (ARCHITECT) platforms for six drugs against LC-MS/MS to assess the analytical performance of these assays. MethodsEighty nine residual urine specimens, which tested positive for amphetamine, THC-COOH, benzoylecgonine, EDDP, opiates and/or oxycodone during routine drug testing, were stored frozen until later confirmation by LC-MS/MS. Immunoassays were performed on cobas and ARCHITECT using a split sample. A third aliquot from these samples was tested by LC-MS/MS to assess the percentage of false positive, false negative, true positive and true negative results and calculate the PPVs and NPVs for each immunoassay. ResultsThe PPVs of THC-COOH and EDDP assays were 100% on both platforms. Suboptimal PPVs were achieved for oxycodone (cobas, 57.1% vs ARCHITECT, 66.7%), amphetamine (77.8 vs. 100%), opiates (80.0 vs. 84.6%) and benzoylecgonine (88.9 vs. 84.2%) assays. The NPV was 100% for cobas and ARCHITECT oxycodone assays. Lower NPVs were achieved for THC-COOH (cobas, 28.6% vs ARCHITECT, 25.0%), EDDP (72.7% for both assays), benzoylecgonine (74.4% vs 73.8%), amphetamine (83.3% vs 82.8%) and opiates (100% vs 85.3%). ConclusionOverall, cobas and ARCHITECT urine drug screens have comparable analytical performance. Confirmatory testing is warranted for positive test results especially for oxycodone, amphetamine, opiates and cocaine. Negative drug screen results must be interpreted with caution especially for THC-COOH, EDDP, benzoylecgonine, amphetamine and opiates.