As certain as the change of seasons, new DUI laws appear on every legislative docket. This year we Washingtonians get a new ballot measure as well. I-502 would legalize marijuana and set a limit for how much can be in your blood while driving (5 nano grams).
Unfortunately, the DUI forensic science issue of impaired driving has become muddled with the politics of the legalization of marijuana. While legalization may be the smart move for society at this point, setting a 5 nanogram THC per se limit is a mistake. Its just too early to know what the standard should be. Once we set a 5 nanogram limit politics will prevent any further adjustment in the name of science.
The backers of Initiative 502 claim that the 5 nanograms of THC level for the DUI law separates "impaired" from those who aren't. They claim that THC impairs driving skills and that the impairment can last for a few hours after smoking. They cite a 2007, Society for the Study of Addiction Addiction (102, 1910–1917) article., Developing Limits For Driving Under Cannabis.
The article has a powerful list of authors: Franjo Grotenhermen, Gero Leson, Günter Berghaus, Olaf H. Drummer, Hans-Peter Krüger, Marie Longo, Herbert Moskowitz, Bud Perrine, Johannes G. Ramaekers, Alison Smiley & Rob Tunbridge
The authors come from the following prestigious universities and organizations: Nova-Institut, Hürth, Germany, Leson and Associates, Berkeley, USA, University of Cologne, Germany, Victorian Institute of Forensic Medicine and Department of Forensic Medicine, Monash University, Melbourne, Australia, University of Würzburg, Germany,Drug and Alcohol Services South Australia, Parkside, Australia, University of California, Los Angeles, USA, Vermont Alcohol Research Centre, USA, Maastricht University, the Netherlands, Human Factors North Inc.,Toronto, Canada and Transport Research Laboratory, Crowthorne, UK.
The article simply reviews current literature on the issue of THC limits for drivers. Nothing new is discovered or reported - its a literature review. Unfortunately, the main conclusion of the article is not what I-502 backers claim: that 5 nanograms is a proper limit based on reason. There is one small reference to the 5 nanogram level and a correlation to fatalities, which are always difficult to control for different variables. Here are some of the articles other findings.
Using current scientific evidence on cannabis-induced impairment of psychomotor skills and the related accident risk, this paper suggests a range of 7–10 ng/ml THC in the serum for an initial non-zero per se limit. It offers reasonably reliable separation of drivers whose driving is in fact impaired by cannabis from those who are not impaired. Inadequate evidence from epidemiological studies renders this limit preliminary and suggests the need for review and possibly revision in the future. Our findings also suggest that using a zero limit for legal determination of impairment by cannabis, which in practice corresponds to the limit of detection for THC in blood, would classify inaccurately many drivers as driving under the influence of, and being impaired by, the use of cannabis.
The main conclusion of the article is that we don't have a scientific basis to conclude what any legal standard should be. The I-502 proponents of the 5 ng standard cite the 2004 Drummer study (Dummer O.H., Gerostamoulos J., Batziris H., Chu M., Caplehor J., Robertson M. D. et. al. The Involvement of Drugs in Drivers of Motor Vehicles Killed in Australian Road Traffic Crashes. Accid Anal Prev 2004 ; 36: 239-48):
Evidence from the few meaningful epidemiological studies on cannabis use and driving is insufficient for deriving a risk-based per se limit for DUIC. While based on too few cases of drivers who had used cannabis and not other drugs, the study by Drummer et al. suggests that a serum THC concentration of 12–16 ng/ml may correspond to the same accident risk as a BAC of 0.05% .
But later Drummer is dismissed by the authors as too statistically small of a study for a legal basis to be founded upon.
As the study by Drummer et al. was based on only 58 cases whose blood samples contained only THC and no other indicators of drugs, the above considerations do not yield a statistically acceptable basis for an enforceable per se limit. The latter would require epidemiological data from a far larger number of cases.
The uncertainty or confidence intervals associated with all of the literature reviewed are huge. Sometimes the ranges came very close to the THC level itself - 3.4 ng as cited in the article by the German Society of Toxicological Forensic Chemistry as an appropriate safety margin. I thought we might be moving above the zero tolerance level with I-502 but it doesn't appear that way if the range of uncertainty or confidence interval swallows up the limit itself.
From lab to lab there is variability on the measurement of THC in blood. A laboratory with a documented internal accuracy of 2.5 ng/ml for THC would report samples exceeding 7.5 ng/ml as in violation of a 5 ng/ml per se limit.
Another area of variability is the given time-lapse between smoking and blood testing, the correlation between a smoked THC dose of THC and the resulting THC blood concentration shows considerable inter- and intra-individual variability. The Grotenhermen article suggests:
According to the pharmacokinetic model of Sticht & Käferstein, which was used in the above meta-analysis to estimate THC blood concentrations, a male weighing 70 kg and smoking a THC dose of 19 mg will, after 3 hours, present with a serum concentration of 4.9 ng/ml with a confidence interval of 3.1–7.7 ng/ml . To minimize false positive test results among drivers with THC concentrations at the upper end of this range without being impaired, a risk-based lower serum limit of 7 ng/ml, rather than 4.2 ng/ml, is thus suggested. Secondly, enforceable legal limits for DUI must consider the effects of analytical errors made during blood analysis.
The Grotenhermen article treats the epidemiological study by Drummer et al. approval but suggests that THC concentrations indicate elevated accident risk at levels higher than indicated by experimental studies. In other words, experimental studies show that some moderate amounts of marijuana improve driving. The article says:
This may be due to the more pronounced adaptive behaviours (slowing down, reduced risk-taking) observed with cannabis-affected drivers in driving simulator and on-road studies, both of which represent more closely real-life conductions. In that case, comparison of experimental studies for alcohol and THC, respectively, would result in systematically lower per se limits for THC than derived from epidemiological studies.
...an increase in serum THC concentrations beyond 5 ng/ml further impairs automatic functions while performance of tasks requiring cognitive control remains stable up to concentrations of 10 ng/ml . This supports the above-mentioned observations from driving studies that drivers under the influence of cannabis may compensate consciously for some of the impairment of their automatic performance, for example by reducing speed or keeping more distance.
Unfortunately, I-502 got it wrong. While the article cited by I-502 may suggest a basis for a per se limit it clearly says this idea is preliminary. Little evidence exists that supports the proposed 5 ng standard. "The few epidemiological studies on cannabis use and driving is insufficient for deriving a risk-based per se limit for DUIC."