In the THC capsule and Sativex experiment, results below the quantification limit were assigned one tenth of that limit for statistical analysis. Those assigned values were not measured trace concentrations.
The reporting boundary was analyte-specific
Fourteen cannabis users completed five treatment sessions involving swallowed THC, THC/CBD mouth spray or placebo. Oral fluid was collected before dosing and for 10.5 hours afterward. The assay had different quantification limits for parent cannabinoids and THC-COOH.
The authors used one tenth of each analyte’s quantification limit for below-limit observations in statistical analyses. This supplied a numerical input where the assay did not provide an acceptable quantitative result.
Substitution and measurement differ
An assigned value below the limit should not be quoted as a precisely measured concentration in a participant. It also should not be silently changed to zero when describing the published analysis. Either choice can matter to distributions, transformed data and estimated relationships near the reporting boundary.
Separate this rule from the ratio analysis
The paper calculated cannabinoid ratios only for positive analyte pairs. Its handling of below-limit concentrations was therefore not a license to generate a ratio for every specimen. An interpretation should specify which calculation is being discussed and which samples contributed. These analytical rules do not turn a saliva value into evidence of impairment or establish that an undetected capsule dose was never taken.
Related reading: cannabinoid assay interpretation.
Research cited
Lee and colleagues (2012). Can oral fluid cannabinoid testing monitor medication compliance and/or cannabis smoking during oral THC and oromucosal Sativex administration?. DOI: 10.1016/j.drugalcdep.2012.10.011.
Leave a Reply