Unexpected increase in false-positive results following benzodiazepine immunoassay reformulation

Benzodiazepine immunoassays are commonly used for urine drug screening but have historically demonstrated poor detection of glucuronidated metabolites, resulting in false-negative results. A reformulated Roche benzodiazepine immunoassay (BNZ2) incorporating β-glucuronidase was developed to improve detection of glucuronidated benzodiazepines. This study evaluated the impact of BNZ2 implementation on false-positive rates in a tertiary medical center.

A retrospective observational study compared urine benzodiazepine screening results from 12 months before and after BNZ2 implementation. Screening was performed using Roche BENZ Plus or BNZ2 reagents on the Roche Cobas Pro c503 analyzer. Presumptive positive results underwent reflex confirmatory testing by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Amphetamine screening served as an internal control.

With BENZ Plus, 363/4513 (8.0%) screens were presumptive positive, compared with 827/4665 (17.7%) following BNZ2 implementation. Confirmatory testing demonstrated false-positive rates of 1.2-1.5% with BENZ Plus and 5.9-7.8% with BNZ2 (p < 0.001). Medication review for patients with false-positive samples identified some common prescriptions but could not identify a causal agent. Amphetamine false-positive rates remained stable across periods (16.2% vs 16.3%, p > 0.05). Following BNZ2 implementation, lorazepam became the most frequently confirmed benzodiazepine (36.7%, n = 269), with substantial increases in lorazepam-only positive samples.

BNZ2 reformulation improved detection of glucuronidated benzodiazepines, particularly lorazepam, but was associated with a significant increase in false-positive results. In this population, 1 in 8 additional presumptive positive samples on BNZ2 was a false positive. These findings demonstrate a trade-off between improved analytical sensitivity and reduced specificity, resulting in increased confirmatory testing requirements and potential downstream effects on laboratory workflow and clinical decision-making.