Roadside Oral Fluid THC Tests Are Good at Detecting Any Cannabis Use but Poor at Predicting Blood Levels

Oral fluid THC reliably detected presence of THC in blood (71.2% sensitivity, 97.7% specificity). However, at commonly used cutoffs, oral fluid was less sensitive and less specific for detecting blood THC above per se limits like 5 ng/mL. This would produce many false positives in random driver testing.

Systematic review and meta-analysis of studies comparing oral fluid THC with blood THC. Individual-level data obtained from study authors and pooled for analysis. Over 18,000 paired samples analyzed. Sensitivity and specificity calculated at various oral fluid cutoff values for detecting different blood THC concentration thresholds.

Many countries use roadside oral fluid tests to enforce cannabis-impaired driving laws. This study shows these tests are useful for identifying recent cannabis use but unreliable for determining whether drivers exceed legal blood THC limits, with major implications for enforcement fairness.

The disconnect between oral fluid THC (which reflects recent exposure) and blood THC (which correlates better with impairment) creates a fundamental problem for per se driving laws. The authors recommend against random oral fluid screening where per se blood limits exist, but support its use for investigating already-suspicious drivers.

Paired samples came from various study conditions (controlled dosing, roadside testing, clinical). The relationship between blood THC and actual driving impairment is itself imperfect. Different oral fluid collection devices may perform differently.