Blood and oral fluid cannabinoid levels varied widely between frequent and occasional cannabis smokers

Frequent users had higher residual blood THC and were more likely to test positive before even smoking. Per se blood THC limits (up to 5 ng/mL) offered limited usefulness as biomarkers of recent use. Blood CBN at 1 ng/mL cutoff had 100% specificity but only 31.4% sensitivity. Oral fluid THC at 10 ng/mL showed the best overall performance (99.7% specificity, 82.4% sensitivity) for detecting use within 3 hours, but was still detectable in 23.2% of participants 4.4 hours post-smoking.

Randomized trial of 191 frequent (4+ times/week) and occasional (<4/week) cannabis users who smoked placebo, 5.9%, or 13.4% THC cannabis ad libitum. Blood, oral fluid, and breath samples collected before and up to 6 hours after smoking. Ten cannabinoids measured in oral fluid, 8 in blood, THC in breath via LC-MS/MS.

Several states have adopted per se THC driving limits, but this study shows these limits perform poorly at identifying recent use, especially in frequent users who may test positive without having used recently. This has direct implications for roadside testing fairness.

The fundamental problem with cannabis impaired driving laws is that no biomarker cleanly separates "recently used and likely impaired" from "used days ago and not impaired." This study quantifies that gap and suggests oral fluid may be better than blood, but neither is sufficient without behavioral assessment.

Only tested smoked cannabis, not edibles or other routes. Laboratory conditions may not reflect real-world use. Detection windows may vary with different cannabis products. CBN levels may vary by cannabis strain.