Now that 29 states have passed medical marijuana laws and nine have legalized cannabis for recreational use, there is growing concern about the potential for more impaired drivers on the road. Already states that have legalized cannabis have experienced more traffic accidents and fatalities with cannabis detected (1).

However, epidemiological data review of fatal traffic accidents has found only a small increase in the risk of a fatal crash when only cannabis was present. More alarming was the significant positive synergistic effect for the risk of a fatal crash when cannabis was combined with alcohol (2).

A natural expectation is that whole blood testing for driving under the influence (DUI) of cannabinoids, along with per se legislation like that for ethanol, would be implemented when cannabis is legalized. Yet cannabis DUI testing has proven to be much more complex than ethanol. As a result, laboratorians’ expertise is essential to ensuring testing protocols that meet the needs of the moment while not repeating the mistakes of the past in testing for other drugs.

Advantages of Oral Fluid in DUI Cases

Compared to whole blood, oral fluid has several advantages in DUI cases: Oral fluid is quicker and less invasive to collect at the roadside, and this easy collection means that the amount of drug measured more closely reflects a true snapshot of cannabis exposure at the time of suspected impairment. This is important because the average time for DUI suspects to have their blood drawn after a traffic stop is about 11/2 hours.

But cannabinoids in whole blood can decrease by as much as 90% in 90 minutes (3). The rapid disappearance of THC in whole blood, combined with the problems of obtaining this sample in an appropriate time frame, makes oral fluid testing attractive.

It is important to note that no research has established the minimum concentration of cannabinoids present in any matrix that can be used to determine impaired driving, and any cutoff set in the future that definitively establishes impairment will likely be controversial. Moreover, a single sample, whether oral fluid or blood, is not able to determine chronic versus acute exposure nor re-cent versus remote use.

Published research on establishing an oral fluid concentration associated with impaired driving has identified several cannabinoids as potential markers of recent use: cannabidiol (CBD), cannabinol (CBN), Δ9-tetrahydrocannabinol (THC), cannabigerol (CBG), Δ9-tetrahydrocannabinolic acid (THCA-A), and tetrahydrocannabivarin (THCV).

The following have been not detected or present be-low 1 ng/mL: 11-hydroxy-Δ9-THC (11-OH-THC), 11-nor-9-carboxy-Δ9-THC (THC-COOH), 11-nor-9-carboxy-Δ9-THC-glucuronide (THC-COOH-gluc), and Δ9-THC-glucuronide. While THC-COOH has been detected in oral fluid using certain collection devices, the levels have been in the ng/L range, making routine measurement difficult.

Remaining Limitations

Oral fluid is not a perfect solution. No published research has demonstrated a significant association between oral fluid and whole blood cannabinoid concentrations. This reinforces the need to establish separate limits for identifying impairment. The temporal, qualitative, and quantitative differences in cannabinoids found in oral fluid need to be established for each route of administration (i.e. smoking, vaporizing, and ingesting).

More studies also must tackle whether different routes of administration have effects on both concentration and type of cannabinoids present in oral fluid. Dutch coffee shop experiments have demonstrated passive contamination of oral fluid from secondhand smoke above proposed cutoff guidelines (5 ng/mL THC) for up to 2 hours, but these studies did not test whether the non-smokers were impaired following exposure (4).

Another concern is the lack of standardized collection protocols along with standardized analytical processes. The collection devices themselves have proprietary extraction/stabilization buffers that require matrix depletion and extraction of cannabinoids prior to analysis. In addition, collection pads and plastic tubes are known to absorb drug, which could result in falsely lowered results.

Stability of the compounds is also a concern. Certain cannabinoids require that samples be processed within 2 months of collection (5). Oral fluid collected for DUI analysis includes a stabilization buffer, but the volume is low, usually <3 mL. Most analytical methods require 1 milliliter of this mixture, and if a retest is requested, the volume of sample and stability of the cannabinoids can become problematic.

Another drawback of oral fluid is the problem of short sample collected at the roadside. The collection devices themselves are fairly accurate and consistent in collecting the correct volume, but dry mouth is a common side effect of cannabis use. When samples are collected shortly after smoking, the collection device may be unable to collect a sufficient volume within 10 minutes. These short samples are still valuable for testing, but cannabinoid concentrations will be elevated due to the concentration calculation being performed as if the correct (larger) volume had been collected.

A potential workaround would be to determine how much sample was collected gravimetrically and provide results based on ng of THC per mass (g) of oral fluid collected, rather than a default volume. Unfortunately, deuterium labeled internal standards are not commercially available for all the cannabinoids. These internal standards are essential for accurate quantification.

For instance, THCV and THCA-A have been proposed as markers of recent use (6), but without deuterium labeled internal standards, these markers are more qualitative than quantitative. As more pure cannabinoid standards become available, they will be incorporated into testing methods, but accurate quantitative values will require analogous deuterated internal standards.

Sticky Problems in Practice

My experience working with oral fluid has been a result of an ongoing research project at the University of California, San Diego and the Center for Medicinal Cannabis Research. I established and validated a LC-MS/MS procedure to quantify all the previously listed cannabinoids in oral fluid. One well-known problem with cannabinoids is that they are very “sticky” compounds.

During the method development phase, we used a glass Hamilton syringe to aliquot calibrators and quality controls to minimize their exposure to plastic pipette tips. When validating this procedure, we encountered significant imprecision for THCA-A at the lower end of our analytical measurement range. Initial troubleshooting suggested this was caused by carryover from a Hamilton syringe used to spike calibrators and controls.

We investigated the carryover by first pipetting a solution of 10 ng/mL standards in methanol with a Hamilton syringe according to our protocol. Second, we washed the Hamilton syringe five times with a (1:1:1:1) mixture of methanol, acetonitrile, isopropanol, and water. Third, we wiped the plunger with tissue paper and reassembled the syringe. Lastly, we analyzed a methanol fraction. This process, labeled as Fraction 1 in Figure 1, was repeated 11 times (55 total washes and 11 plunger wipes of the syringe).

Methanol fractions chromatogram figure

What was amazing to observe was that CBN and THC had returned back to baseline after the first wash cycle, yet for THCA-A in the first fraction, the carryover peak area was equivalent to roughly 0.8 ng/mL and only diminished roughly 13% for each subsequent wash with the last fraction containing a peak area equivalent to approximately 0.2 ng/mL.

Ultimately, the solution to this problem was simply to switch to disposable pipette tips, which restored our linearity and precision for THCA-A. The takeaway message is that certain cannabinoids adhere to surfaces much more than others, and those differences can introduce a major and unexpected bias in your method.

Next Steps

If oral fluid is to supplant blood collection, there needs to be harmonization among states and testing centers on devices and testing methods. Laboratories also need for proficiency testing surveys and methods to be agreed upon prior to defending results in criminal cases.

Oral fluid does have the potential to be used by itself in DUI cases based on its ease of use, better temporal association, and reproducible results. The keys to success will be overcoming short sample collections, demonstrating a strong correlation between concentration and impairment, and harmonization among laboratories through proficiency testing, standard reference materials, and methods.

Philip Sobolesky, PhD, is a clinical chemistry fellow in the Center for Advanced Laboratory Medicine at the University of California, San Diego. Email:


  1. Li G, Bradya J, Chen Q. Drug use and fatal motor vehicle crashes: A case-control study. Accid Anal Prev 2013;60:205–10.
  2. Chihuri S, Li G, Chen Q. Interaction of marijuana and alcohol on fatal motor vehicle crash risk: A case–control study. Inj Epidemiol 2017;4:8.
  3. Milman G, Schwope D, Gorelick D, et al. Cannabinoids and metabolites in expectorated oral fluid following controlled smoked cannabis. Clin Chim Acta 2012;413:765-70.
  4. Moore C, Coulter C, Uges D, et al. Cannabinoids in oral fluid following passive exposure to marijuana smoke. Forensic Sci Int 2011; 212:227-30.
  5. Scheidweiler KB, Andersson M, Swortwood MJ, et al. Long-term stability of cannabinoids in oral fluid after controlled cannabis administration. Drug Test Anal 2016. 9:143-7.
  6. Fabritius M, Chtioui H, Battistella G, et al. Comparison of cannabinoid concentrations in oral fluid and whole blood between occasional and regular cannabis smokers prior to and after smoking a cannabis joint. Anal Bioanal Chem 2013;405:9791-803.


CLN's Focus on Mass Spectrometry is supported by Waters Corporation.

waters corporation logo